content
stringlengths 0
113k
| doi
stringlengths 14
40
| pagenum
stringlengths 9
9
|
|---|---|---|
Submitted to
58
Figure 9. (Measuring the conductance properties of oligoprolines using redox-act ive
ferrocene and cyclic voltammetry [30]. The experimental cur ve shows the exponential
reduction of the electron transfer rate with increaing le ngth of oligoproline.)
Table 1. (Reported activation energy for various proteins. Reprinted in part with permission
from [25]. Copyright 1969, American Institute of Physics)
!
"
" #
$% & ' ('' "
( ' ' "
)
"
# (
*
+
"
#
#
#%
*
#
| 8547d7962e2428a6d1d13df9b8aba7cff60d3747 | page_0057 |
D1210–D1216 NucleicAcidsResearch,2018,Vol.46,Databaseissue Publishedonline20October2017
doi:10.1093/nar/gkx957
FlavorDB: a database of flavor molecules
Neelansh Garg1,2,†, Apuroop Sethupathy1,3,†, Rudraksh Tuwani1,4,†, Rakhi NK5,†,
Shubham Dokania1,6,†,A r v i n dI y e r1,†, Ayushi Gupta1,†, Shubhra Agrawal1,†,
Navjot Singh1,6,†, Shubham Shukla1,7,†, Kriti Kathuria1,8,†, Rahul Badhwar5, Rakesh Kanji5,
Anupam Jain5, Avneet Kaur1, Rashmi Nagpal1and Ganesh Bagler1,*
1Center for Computational Biology, Indraprastha Institute of Information Technology (IIIT -Delhi), New Delhi, India,
2University School of Information, Communication and Technology, Guru Gobind Singh Indraprastha University, New
Delhi, India,3Ashoka University, Sonepat, Haryana, India,4Sri Venkateswara College, Delhi University, New Delhi,
India,5Department of Bioscience and Bioengineering, Indian Institute of Technology Jodhpur, Jodhpur, India,6Delhi
Technological University, New Delhi, India,7Northern India Engineering College, Guru Gobind Singh Indraprastha
University, New Delhi and8Maharaja Agrasen College, Delhi University, New Delhi, India
Received August 15, 2017; Revised September 18, 2017; Editorial Decision October 05, 2017; Accepted October 06, 2017
ABSTRACT
Flavor is an expression of olfactory and gusta-
tory sensations experienced through a multitude of
chemical processes triggered by molecules. Beyondtheir key role in defining taste and smell, flavor
molecules also regulate metabolic processes with
consequences to health. Such molecules presentin natural sources have been an integral part of
human history with limited success in attempts
to create synthetic alternatives. Given their utilityin various spheres of life such as food and fra-
grances, it is valuable to have a repository of flavor
molecules, their natural sources, physicochemicalproperties, and sensory responses. FlavorDB ( http:
//cosylab.iiitd.edu.in/flavordb ) comprises of 25,595
flavor molecules representing an array of tastes andodors. Among these 2254 molecules are associated
with 936 natural ingredients belonging to 34 cate-
gories. The dynamic, user-friendly interface of the re-source facilitates exploration of flavor molecules for
divergent applications: finding molecules matching
a desired flavor or structure; exploring molecules of
an ingredient; discovering novel food pairings; find-
ing the molecular essence of food ingredients; as-sociating chemical features with a flavor and more.
Data-driven studies based on FlavorDB can pave the
way for an improved understanding of flavor mecha-nisms.INTRODUCTION
Flavor is a complex, multi-sensory human experience with
a rich evolutionary history ( 1). Molecules form the chemi-
calbasisofflavorexpressedprimarilyviagustatoryandol-
factory mechanisms. The perception of flavor arises from
interaction of flavor molecules with the biological machin-ery and could be perceived as an emergent property of a
complex biochemical system. While some components of
thispuzzlehavebeenunearthed,aholisticviewofthisphe-nomenon still eludes us ( 2–5). Taking a data-centric ap-
proachcanprovideasystemsperspectiveofflavorsensation
byofferingways todiscern itskey features.
Flavors derived from natural sources have shaped culi-
nary habits throughout human history. Analogous to vari-
ations in regional languages, cultures have evolved varia-tionsinthewaytheycook.Traditionalrecipecompositions
encode ingredient combinations that are not only palat-
ablebutappetizing.Heuristicassociationsbetweenmolecu-lar properties and perception of flavors provide indications
towards its chemical basis ( 1). For example, combinations
of aliphatic esters play a major role in many fruit flavors.Ketones are known to impart metallic flavors in oxidized
butter, and monoterpenoids provide the characteristic fla-
vors of many herbs and spices. However, such knowledgeremains largely unstructuredand incomprehensive.
FlavorDB was created with the aim of integrating mul-
tidimensional aspects of flavor molecules and representingtheir molecular features, flavor profiles and details of nat-
ural source (Figure 1). FooDB, one of the efforts in simi-
lardirection,compilesmoleculesfromfoodingredients;al-
though its focus is not on chemical basis of flavor or fla-
vorpairing( http://foodb.ca ).Flavornetisanotherresource,
which provides a list of flavor molecules and their odor
*Towhomcorrespondenceshouldbeaddressed.Tel:+911126907443;Email:bagler@iiitd.ac.in;ganesh.bagler@gmail.com
†Theseauthors contributed equally to this workas firstauthors.
Present address:Ganesh Bagler,Centerfor Computational Biology, Indraprastha Institute of Information Technology(IIIT-Delhi),New Delhi 1100 20, India.
C/circlecopyrtTheAuthor(s) 2017. Published by Oxford University Presson behalf of NucleicAcidsResearch.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http: //creativecommons.org /licenses /by-nc /4.0/), which
permits non-commercial re-use, distribution, and reproduction in any medium, provided the original work is properly cited. For commercial re-use, please contact
journals.permissions@oup.comDownloaded from https://academic.oup.com/nar/article/46/D1/D1210/4559748 by Princeton University user on 03 April 2023
| b3a12df31f90e41e517c42fc831ad9f8f08465f6 | page_0000 |
NucleicAcidsResearch,2018,Vol.46,Databaseissue D1211
Figure 1. FlavorDB is a seamless amalgamation of ‘entity space’ and ‘flavor space’. The resource provides a comprehensive dataset along with a user-
friendlyinterfaceandinterlinkedsearchenginesforexploringtheflavoruniverse.
profiles, but does not furnish information of their natural
sources(6).Otherattemptsinthisdirectionhavefocusedon
compilationofdataspecifictoaspectsofflavors:tastessuch
as bitter (BitterDB) and sweet (SuperSweet), and volatile
compounds of scents (SuperScent) ( 7–9). Certain others
havetargetednutritionalfactors(NutriChem),polyphenols
(Phenol-Explorer)andthemedicinalvalueoffood( 10–13).
Among other sources, FlavorDB collates information
from FooDB, Flavornet, SuperSweet and BitterDB to cre-
ate a comprehensive repository of flavor molecules, fla-
vorprofiles,physicochemicalpropertiesandnaturalsources(Section S1, Supplementary Data). Compared to FooDB
whichhas2816flavorcompounds,FlavorDBcovers25,595
flavor molecules compiled from Fenaroli’s Handbook ofFlavor Ingredients and literature survey in addition to in-
tegrating data from all the above-mentioned sources. Fla-
vorDB spans across 34 ingredient categories covering 936
ingredients of which 190 are unique. One of the features
which sets FlavorDB apart from similar resources is that itpresentsinformationinahierarchyoffoodcategory,ingre-dients, flavor molecules and their flavor profile, and chem-
ical descriptors including functional groups and physico-chemical properties. Through an extensive repertoire of in-
gredients and their constituent flavor molecules, FlavorDB
also provides a tool for experimenting with flavor pairing.Thus, it offers an integrative platform for exploring the fla-
vor space withthehelp of afeature-richvisual interface.
FlavorDB combines different dimensions of flavor con-
stitutingthe‘entityspace’and‘flavorspace’(Figure 1).The
former incorporates facets of ingredients which are enti-
ties from natural sources often used in food, whereas thelatter represents molecules responsible for flavor sensation
andtheirdescriptors.Bybringingrelevantinformationun-
der a single umbrella, FlavorDB provides a comprehensivedataset backed by a user-friendly interface, creative visual-
izations, and interlinked search engines for exploring fea-
tures that contribute to the sensation of flavor. Thus, Fla-
vorDBpavesthewayforanimprovedunderstandingoffla-
vor perception arising out of complex interplay of flavorcompoundswithbiologicalsystemsandalliedapplications.Downloaded from https://academic.oup.com/nar/article/46/D1/D1210/4559748 by Princeton University user on 03 April 2023
| b3a12df31f90e41e517c42fc831ad9f8f08465f6 | page_0001 |
D1212NucleicAcidsResearch,2018,Vol.46,Databaseissue
DATABASE OVERVIEW
FlavorDB is a resource with extensive coverage of 25,595
flavor molecules (Figure 1). Among molecules listed in
the database, 2254 have been reported to be found in 936
natural entities /ingredients. These natural ingredients have
further been classified into 34 categories, and mapped to
527 distinct natural sources. An additional 13,869 com-
poundswereidentifiedassynthetic.Fortheremaining9472molecules,nospecificsourcecouldbeascertained.Thefea-
tures provided as part of the detailed molecular and flavor
profiles of these compounds have an impact on their tasteand odor through gustatory and olfactory sensory mecha-
nisms.
FlavorDB offers a user-friendly interface for querying
and browsing flavor molecules, entities /ingredients, natu-
ral sources, as well as performing flavor pairing. Interac-
tive data visualizations such as the flavor network and in-terlinked search options are provided to retrieve relevant
information. Apart from searching via textual query or by
drawing the chemical structure, FlavorDB also provides a
‘Visual Search’. Using this a user can interactively browse
through the ingredient categories to access correspondingnaturalentitiesandsubsequentlyobtaindetailsoftheirfla-
vor molecules.
For any flavor molecule, the resource also facilitates
lookup for structurally similar molecules within the
database as well as those commercially available from ex-
ternal sources (ZINC ( 14)). Thus, through a blend of the
entity and flavor space along with a dynamic interface and
visualizations,FlavorDBprovidesawidespectrumofinfor-
mation facilitatinginsights intothe flavor universe.
DATA COMPILATION
One of the primary motivations behind the creation of this
resource was to map the space of molecules critical for the
sensations of taste and smell. To begin with, a list of in-gredients was created using ( 15–17), FooDB ( http://foodb.
ca) and arXiv preprint arXiv:1502.03815, 2015. Each of
the 936 ingredients were then manually classified into 34categories: Additive, Animal Product, Bakery, Beverage,
Beverage Alcoholic, Beverage Caffeinated, Cereal, Maize,
Dairy, Dish, Essential Oil, Fish, Seafood, Flower, Fruit,Berry, Fruit Citrus, Fruit Essence, Fungus, Herb, Meat,
Legume, Nut, Seed, Plant, Plant Derivative, Spice, Veg-
etable, Cabbage, Vegetable Fruit, Vegetable Gourd, Veg-etable Root, Vegetable Stem, and Vegetable Tuber. Each
entity was also mapped to its natural source, with a to-
tal of 527 unique sources being identified. The detailsof entities and their natural sources, related images, and
scientific classification were obtained from Wikipedia us-
ing Python’s BeautifulSoup4 library ( https://www.crummy.
com/software/BeautifulSoup )andMediaWiki’sactionAPI
(MediaWikiThe Free WikiEngine).
Thedataofflavormoleculesforeachoftheseingredients
werecompiledviaflavorresourcessuchasFenaroli’shand-
bookofflavoringredients,previouslyreporteddata( 16,17),
FooDB ( http://foodb.ca ), arXiv preprint arXiv:1502.03815
and literature survey (Also see Section S2, Supplementary
Data). Common names, scientific name and synonyms ofingredients were used to query PubMed to obtain articles
that reported their flavor molecules. Flavor molecules as-
sociated with entities /ingredients were thus curated from
existing sources ( 15–17) (FooDB; http://foodb.ca ,a r X i v
preprint arXiv:1502.03815, 2015) and compiled manually.
Molecules from Flavornet, BitterDB and SuperSweet were
furtherincludedalongwiththeirflavorprofiles( 6,7,9).Ad-
ditionally,informationfor33tastereceptors(Sweet,Bitter,
Sour,andUmami)and1068odorreceptorsisalsoavailable
inFlavorDB.Foreachreceptor,weprovideitsUniprotID,name,involvement in taste,and Uniprot link ( 18).
The chemical identifiers of molecules were obtained
fromvarioussources(FooDB; http://foodb.ca )(6,
7,9,15,16)
(AlsoseeSectionS2,SupplementaryData),andwerestan-
dardized to procure their CAS (Chemical Abstract Ser-
vice) numbers. CAS numbers were then mapped to theircorresponding PubChem IDs, as the former are often de-
generate with multiple CAS numbers pointing to the same
molecule, and some pointing to multiple molecules. Thus,
PubChemIDwasusedastheuniqueprimarykeyforevery
flavormolecule.UsingthePubChemID,compoundidenti-fiers(suchascommonname,IUPAC,CanonicalSMILES),
physicochemical properties and 2D images were obtained
fromPubChemRESTAPI( https://pubchem.ncbi.nlm.nih.
gov/pug
rest/PUG REST.html ). The flavor profile of the
molecule (Flavor Profile, FEMA Flavor Profile, FEMA
Number, Taste, and Odor) was created by compiling infor-mationfromFooDB( http://foodb.ca ),Flavornet( 6),Super-
Sweet(9),BitterDB( 7) and PubChem.
Further 2D /3D, ADMET and physicochemical proper-
ties as well as Mol2 files for all 25,595 molecules were ob-
tained using Discovery Studio 4.0 (DS4.0; Accelrys Inc.).
ThefunctionalgroupswereobtainedusingCheckmolsoft-ware(19).Functionalgroupreferstoanatom,oragroupof
atoms that have similar chemical properties whenever they
occurindifferentcompounds( 20).Thus,itdefinesthechar-
acteristicphysicalandchemicalpropertiesoffamiliesofor-
ganic compounds.
Please refer to Supplementary Figures S1, S2, S3 and S4
in Supplementary Data for the statistics of entities, cate-
gories,flavormolecules,theirflavorsandfunctionalgroups.
DATABASE ARCHITECTURE AND WEB INTERFACE
FlavorDBfacilitateseasycomprehensionofcomplexinter-
relations among flavor molecules, entities /ingredients and
theirnaturalsources(Figure 2;alsoseeFigureS5ofSupple-
mentary Data). Interactive data visualizations and a wide
varietyofuser-friendlysearchesprovidequickaccesstode-
sired information. The following utilities and applicationsinFlavorDBenablevisualexplorationsof‘flavorspace’and
‘entityspace’ toget insights into theflavor universe.
The flavor network
Flavor Network visualizes the graph of flavor-sharing
across all entities /ingredients. To make it easier to observe
flavor-sharing within and across categories, the entities are
groupedcategory-wiseandarespacedoutalongthecircum-
ference. To address dense pattern of interrelationships due
to abundance of sharing, the backbone network showingDownloaded from https://academic.oup.com/nar/article/46/D1/D1210/4559748 by Princeton University user on 03 April 2023
| b3a12df31f90e41e517c42fc831ad9f8f08465f6 | page_0002 |
NucleicAcidsResearch,2018,Vol.46,Databaseissue D1213
Figure 2. Schematic of FlavorDB user interface highlighting features for searching and graphical navigation of data. ( 1) Flavor Molecule Search, ( 2)
Entity /IngredientSearch,( 3)NaturalSourceSearch,( 4)FlavorPairing,( 5)AdvancedSearch,( 6)Molecular&FlavorProfile(includingfeaturesforfinding
relatedentities,searchingforstructurallysimilarmolecules,externallink-outsanddatadownload),( 7)MoleculeoftheDay,( 8)TheFlavorNetworkand
(9)VisualSearch.Downloaded from https://academic.oup.com/nar/article/46/D1/D1210/4559748 by Princeton University user on 03 April 2023
| b3a12df31f90e41e517c42fc831ad9f8f08465f6 | page_0003 |
D1214NucleicAcidsResearch,2018,Vol.46,Databaseissue
statistically significant edges is depicted ( 21). Clicking on
an entity shows its association with other entities by virtue
ofsharedflavormolecules,therebyenablingsearchforsimi-
laritiesamongseeminglydisparateentities.TheFlavorNet-work was implemented with the D3.js JavaScript library
(https://d3js.org ).
Visual search
VisualSearch,implementedwiththeGoogleChartslibrary
(http://developers.google.com/chart ), provides an interac-
tive way of exploring FlavorDB. At the top of the hierar-chy,itdisplaysall34ingredientcategoriesasboxes.Thesize
of each box is determined by the number of unique flavor
molecules present in that category. Greater the number ofmolecules, larger the size of the box. Clicking through any
ofthecategories,onecannavigatetoitsconstituentingredi-
ents and their respective flavor profiles. Thus, visual searchenables multi-level,open-endedexploration of FlavorDB.
Molecular search
By virtue of extensive molecular and flavor features pro-
vided in FlavorDB, ‘Molecular Search’ forms a key query
mechanism. It facilitates querying on the basis of a host
of features including Common Name, Functional Group,
FEMA Flavor, Molecular Weight, Number of Hydrogen
BondDonors /Acceptors,andTypeofMolecules(Natural,
Synthetic, Unknown). Additionally, the JSME Molecule
Editor enables search based on structural similarity ( 22).
The editor facilitates creation of molecules using SMILES,MOLorSDFfiles.Structurallysimilarcompoundscanalso
befoundusingthe‘SearchSimilarinFlavorDB’buttonpro-
vided on themolecules’ profile page.
Molecularsearchyieldsmatchingflavorcompoundswith
detailed‘molecularandflavorprofile’.A3Dvisualizationof
the molecule is provided with the JSmol library along withanexternallinktoPubChem.Furtheranoptionfordown-
loadingthemoleculeindifferentformats(MOL2,SMILES,
2Dimages)isalsoavailable.Theflavormoleculescanbefil-tered using search and sort functionality provided by ‘data
tables’ plugin.
The algorithm for performing structural similarity com-
putes molecular fingerprints (FP2) of all flavor molecules
using OpenBabel ( 11). For any molecular structure that is
queried, its fingerprint is computed using an OpenBabelprotocol and is compared with the database using the Tan-
imoto coefficient of structural similarity. Molecules with at
least 30% similarity are returned. FlavorDB also facilitatesbrowsing all 25,595 flavor molecules by doing a null search
(no constraints; all query fields empty). ‘Molecule of the
Day’ feature offers a peek into the flavor universe, fromwheretheuser can start exploring theresource.
Advanced search
Advanced search provides an option for refined search by
queryingFlavorDBdataonthebasisofavarietyofmolec-
ular properties (number of rings, rotatable bonds, energy,
surfaceareaetc.)apartfromthoseprovidedinbasicsearch.
For numeric fields either a range or discrete values can be
provided as aquery.Entity and natural source search
These searches facilitate querying on the basis of
‘Entity /Ingredient Name’ as well as ‘Category Name’
to retrieve detailed information of entity, category, natural
source, scientific classification, Wiki page link, synonyms,images, and flavor molecules associated with the entity.
Additionally, we facilitate search based on synonyms,
knowing that many ingredients are often known by avariety of alternative names. For example, to search for
‘Eggplant’, one may as well search by either ‘Aubergine’ or
‘Brinjal’. All text search fields are assisted by jQuery UIautocomplete.
Flavor pairing
‘Flavor Pairing’ (also known as ‘Food Pairing’) is a
heuristic with empirical evidence ( 15,16), (arXiv preprint
arXiv:1505.00155, 2015), used for finding ingredient pairs
thatareexpectedtogowelltogetherinarecipe /foodprod-
uct.Onthefoundationofextensiverepertoireof936entities
and their constituent flavor molecules (2254), ‘Flavor Pair-
ing’ tool provides a powerful engine for spanning through>437,000 pairings to reveal flavor profile overlaps across
categories and disparate entities. The app provides pairing
results through interactive visualization for easy compre-hension. Detailed analysis yields the number of shared fla-
vor molecules between the queried entity and all the other
entities having at least one shared flavor molecule. The re-sults can be further filtered using data tables to probe the
listof shared flavor molecules.
Webservertechstack
FlavorDB has been designed as a Relational Database us-
ing MySQL ( https://www.mysql.com ). The webserver has
been built using the Python web development framework,Django ( https://www.djangoproject.com ). Django has a
built in ORM (Object Relational Mapper) for querying
the database, thus optimizing queries and making it eas-
ier to perform complex queries, apart from reducing the
development period. The front-end has been built using
HTML, CSS and JavaScript. The jQuery, Bootstrap, D3.jsand Google Charts libraries were used to add to the func-
tionality of FlavorDB. An Apache HTTP Server has been
used to route requests to the Django app and to enabledatacompressionforfasterpageloadtimes.Thesiteisbest
viewedinlatestversionsofGoogleChrome,Firefox,Opera,
InternetExplorer, and Microsoft Edge.
EXAMPLES
Below we provide a few case studies illustrating the utility
ofFlavorDB forvarious applications.
Applications for Flavor /Food pairing
The food pairing principle suggests that ingredients which
tastesimilartendtobeusedtogetherinrecipes( 23).Histor-
ically practiced on a trial-and-error basis, food pairing has
reliedheavilyonhumanjudgmentandtheintuitionoffood
connoisseurs. Evidence-based understanding of rules thatDownloaded from https://academic.oup.com/nar/article/46/D1/D1210/4559748 by Princeton University user on 03 April 2023
| b3a12df31f90e41e517c42fc831ad9f8f08465f6 | page_0004 |
NucleicAcidsResearch,2018,Vol.46,Databaseissue D1215
dictate food choices can facilitate informed experiments to
pair ingredients in a recipe. While many Western cuisines
are reported to be characterized with ‘uniform food pair-
ing’, Indian cuisine tends to follow ‘contrasting food pair-ing’pattern( 15,16)(arXivpreprintarXiv:1505.00155,2015,
arXiv preprint arXiv:1502.03815, 2015). Based on data of
ingredients and flavor molecules, FlavorDB offers an appbywhichuserscanexperimentwithfoodpairingofdesired
ingredients.
Let’s consider the case where a user wants to find a sub-
stitute for oregano. Using the Flavor Network, one can
quickly get an overview of ingredients that share flavor
molecules with oregano. Via the Flavor Pairing app, it canbe observed that oregano and thyme have the most num-
berofcommonmolecules.Hence,byvirtueofuniformfood
pairing,thymecanbeapossiblesubstitutefororegano.ThisexampledemonstratestheutilityofdifferentfeaturesofFla-
vorDBformakinginformeddecisionsonflavorpairingand
choiceofingredients.
Searching fordrug-like compounds
Find molecules structurally similar to a compound satisfy-
ing Lipinski’s rule of five. This example demonstrates use
ofFlavorDB’s‘AdvancedSearch’and‘StructuralSimilarity
Search’ features to find molecules matching desired chemi-calproperties.Lipinski’sruleisaheuristictoevaluatedrug-
likeness; the suitability of a chemical compound to have
pharmacological or biological activity making it a likelycandidate for orally activedrug. It states that, an orally ac-
tive drug has no more than one violation of the following
criteria:nomorethanfivehydrogenbonddonors;nomorethan10hydrogenbondacceptors;Molecularmasslessthan
500 Da; an octanol–water partition coefficient (log P) not
>5.
Using the above mentioned criteria in the ‘Advanced
Search’,onecansearchformoleculesthatsatisfyLipinski’s
rule. This query can be refined further by using the struc-turalsearchprovidedthroughtheJSMEtool,tofinddrug-
like compounds that are structurally related to a certain
compound.As an example,onemayspecifytheconditionsfor Lipinski’s rule in the ‘Advanced Search’ and draw the
chemicalstructureforPyridine.Theresultingsearchreturns
allmoleculesthatarepotentiallysuitableforbeingtestedasa drug, ranked in descending order of structural similarity
withPyridine.
Finding flavor molecules similar to any desired compound
Allicin (SMILES: O =S(SC\C=C)C\C=C) is one of the
primary molecules present in garlic. Using the structural
search, one can get a list of similar molecules in FlavorDB.The top result for this example is Diallyl Disulfide with
66.7% similarity. Interestingly, as reflected in its ‘Molecu-
larandFlavorProfile’,DiallylDisulfidehasa‘sharp,garlictaste’. Thus it can be a possible substitute for Allicin and
used as a scaffold to create a synthetic garlic flavor. One
may also search for flavor molecules matching a particular
FEMA flavor term and /or functional group among natu-
rally occurringas well as syntheticcompounds.Exploring flavor properties of ingredients
Wasabiisempiricallyknowntohaveapungentodor.Using
the ‘Entity Search’ option, it can be discovered that there
are five known flavor molecules for Wasabi in FlavorDB.
Of these, one molecule, Allyl Isothiocyanate is reported toexhibit characteristics of a ‘very pungent’ odor. It can be
speculatedthatthepungentsmellofwasabiisprimarilydue
to thepresenceofAllyl Isothiocyanate.
TheaboveexamplesillustratehowFlavorDBanditsfea-
tures can be interactively used to make data-driven deci-
sions that caterto various aspects offlavor.
SUMMARY AND OUTLOOK
Study of molecules and mechanisms involved in flavor sen-
sation has been of interest for its applications for food and
fragrances ( 1–13)(http://www.pherobase.com ). Contribut-
ingtotheeffortsoncompilationofmoleculesfromfoodandtheirflavors,FlavorDBprovidesadetailedperspectiveinto
the ‘entity space’, ‘flavor space’ and the latent connections
between the two. In doing so, it lays down the foundationfor conducting data driven analysis which can aid in build-
ing applications meant for molecular gastronomy, culinary
food pairing, novel recipe generation, aroma blending andpredictingodor from chemicalfeatures ( 24–26).
Despite our best efforts, FlavorDB is not an exhaustive
repositoryofallflavormoleculesandingredients.Ourdataon flavor molecules of an ingredient is limited by the avail-
ability of information about them from literature survey.
Similarly,theingredientsrepresentedinthedatabasearenot
exhaustiveinthemselvesastheirchoiceislimitedbythere-
portsofflavorcompounds.Also,atpresentthedatabasefo-cuses on flavor profiles of natural ingredients and thus, the
cruxofourflavorspaceismadeupofflavormoleculesfrom
natural sources.
In future, we intend to increase the coverage of flavor
moleculesandlookfortheirlatenteffectsonhumanhealth.
Our endeavor is to integrate aspects of flavor profiles with
thoseofentitiesandmoleculesthathashithertobeenunex-
plored.
AVAILABILITY
FlavorDB is available at http://cosylab.iiitd.edu.in/flavordb
SUPPLEMENTARY DATA
Supplementary Dataare available at NAROnline.
ACKNOWLEDGEMENTS
G.B. thanks the Indraprastha Institute of Information
Technology (IIIT-Delhi) for providing computational fa-
cilities and support. N.G., A.S., R.T., S.D., N.S., S.S. andK.K. were Summer Research Interns in Dr. Bagler’s lab at
theCenterforComputationalBiology,andarethankfulto
IIIT-Delhi for the support and fellowship. A.I., A.G. andS.A., M.Tech. (Computational Biology) students, thank
IIIT-Delhiforthefellowship.R.N.K.,R.B.andR.K.thank
the Ministry of Human Resource Development, Govern-
ment of India and Indian Institute of Technology Jodhpur
for thesenior researchfellowship.Downloaded from https://academic.oup.com/nar/article/46/D1/D1210/4559748 by Princeton University user on 03 April 2023
| b3a12df31f90e41e517c42fc831ad9f8f08465f6 | page_0005 |
D1216NucleicAcidsResearch,2018,Vol.46,Databaseissue
FUNDING
The open access publication charge for this paper has been
waived byOxfordUniversityPress- NAR.
Conflictofintereststatement. None declared.
REFERENCES
1. Fisher,C.andScott,T.(1997) FoodFlavours:BiologyandChemistry
RoyalSocietyofChemistry .
2. Shepherd,G.(2013) Neurogastronomy–HowtheBrainCreatesFlavor
andWhyitMatters ColumbiaUniversityPress.
3. Malnic,B.,Hirono,J.,Sato,T.andBuck,L.B.(1999)Combinatorial
receptorcodesforodors. Cell,96,713–723.
4. Mouritsen,O.G.(2015)Thescienceoftaste. Flavour,4,1–2.
5. Newcomb,R.D.andOhla,K.(2013)Thegeneticsandneuroscienceof
flavour.Flavour,2,17.
6. Arn,H.andAcree,T.E.(1998)Flavornet:adatabaseofaroma
compoundsbasedonodorpotencyin naturalproducts. Dev.Food
Sci.,40,27.
7. Wiener,A.,Shudler,M.,Levit,A.andNiv,M.Y.(2012)BitterDB:a
databaseofbittercompounds. NucleicAcidsRes. ,40,413–419.
8. Dunkel,M.,Schmidt,U.,Struck,S.,Berger,L.,Gruening,B.,
Hossbach,J.,Jaeger,I.S.,Effmert,U.,Piechulla,B.,Eriksson,R. etal.
(2009)SuperScent–adatabaseofflavorsandscents. NucleicAcids
Res.,37,291–294.
9. Ahmed,J.,Preissner,S.,Dunkel,M.,Worth,C.L.,Eckert,A.and
Preissner,R.(2011)SuperSweet-Aresourceon naturalandartificial
sweeteningagents. NucleicAcidsRes. ,39,D377–D382.
10. Scalbert,A.,Andres-Lacueva,C.,Arita,M.,Kroon,P.,Manach,C.,
Urpi-Sarda,M.andWishart,D.(2011)Databasesonfoodphytochemicalsandtheirhealth-promotingeffects. J.Agric.Food
Chem.,59,4331–4348.
11. Rothwell,J.A.,Perez-Jimenez,J.,Neveu,V.,Medina-Rem ´on,A.,
M’Hiri,N.,Garc ´ıa-Lobato,P.,Manach,C.,Knox,C.,Eisner,R.,
Wishart,D.S. etal.(2013)Phenol-Explorer3.0:amajorupdateofthe
Phenol-Explorerdatabasetoincorporatedataontheeffectsoffood
processingonpolyphenolcontent. Database ,2013,bat070.12. Jensen,K.,Panagiotou,G.andKouskoumvekaki,I.(2015)
NutriChem:asystemschemicalbiologyresource toexplorethemedicinalvalueofplant-basedfoods. NucleicAcidsRes .,43,
D940–D945.
13. Neveu,V.,Perez-Jimenez,J.,Vos,F.,Crespy,V.,duChaffaut,L.,
Mennen,L.,Knox,C.,Eisner,R.,Cruz,J.,Wishart,D. etal.(2010)
Phenol-Explorer:anonlinecomprehensivedatabaseonpolyphenol
contentsin foods. Database ,2010,bap024.
14. Sterling,T.andIrwin,J.J.(2015)ZINC15––liganddiscoveryfor
everyone.J.Chem.Inf.Model. ,55,2324–2337.
15. Jain,A.,Rakhi,N.K.andBagler,G.(2015)Analysisoffoodpairingin
regionalcuisinesofIndia.
PLoSOne ,10,e0139539.
16. Ahn,Y.-Y.,Ahnert,S.E.,Bagrow,J.P.andBarab ´asi,A.-L.(2011)Flavor
networkandtheprinciplesoffoodpairing. Sci.Rep.,1,196.
17. Burdock,G.A.(2010)Fenaroli’shandbookofflavoringredients.
18. Wasmuth,E.V.andLima,C.D.(2016)UniProt:theuniversalprotein
knowledgebase. NucleicAcidsRes. ,45,1–12.
19. Haider,N.(2010)Functionalitypatternmatchingasanefficient
complementarystructure /reactionsearchtool:anopen-source
approach.Molecules ,15,5079–5092.
20. IUPAC(2014)Compendiumofchemicalterminology.
21. Serrano,M.A.,Boguna,M.andVespignani,A.(2009)Extractingthe
multiscalebackboneofcomplexweightednetworks. Proc.Natl.
Acad.Sci.U.S.A. ,106,6483–6488.
22. Bienfait,B.andErtl,P.(2013)JSME:a freemoleculeeditorin
JavaScript. J.Cheminform. ,5, 1–6.
23. Blumenthal,H.(2008)TheBigFatDuckCookbookBloomsbury
Publishing.
24. Keller,A.,Gerkin,R.C.,Guan,Y.,Dhurandhar,A.,Turu,G.,Szalai,B.,
Mainland,J.D.,Ihara,Y.,Yu,C.W.,Wolfinger,R. etal.(2017)
Predictinghumanolfactoryperceptionfromchemicalfeatures ofodormolecules. Science,355,820–826.
25. Spence,C.,Hobkinson,C.,Gallace,A.andFiszman,B.P.(2013)A
touchofgastronomy. Flavour,2, 14.
26. This,H.(2002)Moleculargastronomy. Angew.Chem.Int.Ed. ,41,
83–88.Downloaded from https://academic.oup.com/nar/article/46/D1/D1210/4559748 by Princeton University user on 03 April 2023
| b3a12df31f90e41e517c42fc831ad9f8f08465f6 | page_0006 |
Molecules 2011 , 16, 202-220; doi:10.3390/molecules16010202
molecules
ISSN 1420-3049
www.mdpi.com/journal/molecules
Review
Small Molecule Inhibitors as Countermeasures for Botulinum
Neurotoxin Intoxication
Bing Li 1,*, Norton P. Peet 1, Michelle M. Butler 1, James C. Burnett 2, Donald T. Moir 1 and
Terry L. Bowlin 1
1 Microbiotix, Inc., One Innovati on Drive, Worcester, MA 01605, USA
2 Target, Structure-Based Drug Di scovery Group, SAIC-Frederick, Inc ., National Cancer Institute at
Frederick, 1050 Boyles Street , Frederick, MD 21702, USA;
E-Mail: burnettjames@mail.nih.gov (J.C.B.)
* Author to whom correspondence should be addressed; E-Mail: bli@microbiotix.com;
Tel.: +1-508-757-2800; Fax: +1-508-757-1999.
Received: 30 November 2010; in revised form: 20 December 2010 / Accepted: 29 December 2010 /
Published: 30 December 2010
Abstract: Botulinum neurotoxins (BoNTs) are the most potent of known toxins and are
listed as category A biothreat agents by the U. S. CDC. The BoNT-mediated proteolysis of
SNARE proteins inhibits the exocytosis of acetylcholine into neuromuscular junctions,
leading to life-threatening flaccid paraly sis. Currently, the onl y therapy for BoNT
intoxication (which results in the diseas e state botulism) includes experimental
preventative antibodies and long-te rm supportive care. Therefore, th ere is an urgent need to
identify and develop inhibitors that w ill serve as both prophylactic agents and post-
exposure ‘rescue’ therapeutics. This review focuses on recent progress to discover and
develop small molecule inhibitors as therap eutic countermeasures for BoNT intoxication.
Keywords: botulinum neurotoxin; inhibitor; drug discovery
1. Introduction
Botulinum neurotoxins (BoNTs), secreted by the anaerobic spore-forming bacterial Clostridia
species botulinum, baratii, and butyricum, are the most poisonous of known biological toxins [1,2],
and as a result are listed as cat egory A biothreat agents by the Un ited States Centers for Disease OPEN ACCESS | 4fd91a400fdbb432e4d634acfcf42609ede5f4cf | page_0000 |
Molecules 2011 , 16
203
Control and Prevention (CDC). BoNTs can be easily produced and may be delivered by either aerosol
route, [2-4] or through contamination of the f ood supply. Consequently, th ese toxins represent a
serious threat to both m ilitary personnel and civ ilians [5-7]. Moreover, si nce both BoNT/A (Botox™)
and BoNT/B (Myobloc™) are available commerciall y, and are now used for cholinergic nerve and
muscle dysfunction therapy, as well as cosmetic treat ments [8-16], it is likely that overdose, misuse
and/or adverse side effects [17] may result in systemic toxin expos ure. The currently available BoNT
toxoid vaccine, as well as experime ntal preventative antibodies, cannot counter these toxins in the
neuronal cytosol. This is an important point, as it is likely that individuals will seek medical attention
only after clinical symptoms of intoxication manifest ( i.e., life-threatening paralysis). Currently,
critical care mechanical ventilation is the only tr eatment option once neurons have been intoxicated
and diaphragm muscles cease to function. However, long-term mechanical ventilation would be
impractical for the treatment of a large population of intoxicated indi viduals. Therefore, there is an
urgent need to iden tify and develop low molecula r weight, non-peptidic inhib itors that w ill serve as
both prophylactic agents and post-ex posure ‘rescue’ therapeutics.
There are seven botulinum neurotoxin (BoNT) sero types (A-G), which possess different tertiary
structures and significant sequence divergence. Of the seven BoNT se rotypes, A, B, and E are known
to cause human botulism [18,19], with BoNT/A a nd BoNT/B exhibiting the longest durations of
activity in the neuronal cytosol ( i.e., from several weeks to months, depending on the severity of the
poisoning [20-22]). Hence, the vast majority of research to develop inhibitors to counter BoNT
intoxication post-neuronal internal ization has focused on the BoNT/A and BoNT/B light chains (LCs).
Once inhaled into the lungs or ingested into the ga strointestinal tract, BoNTs are transcytosed across
the respiratory epithelium or mucosa into the blood stream, where they can enter the intr acellular space
prior to accessing peripheral cho linergic nerve endings. St ructurally, BoNTs are synthesized as single
polypeptide chains that undergo bacterial or host- mediated cleavage resulting in a 100 kDa heavy
chain (HC) component and a 50 kDa light chain (LC) component. These two components, which
compose the biologically active holotoxin, are connected by a disulfide bridge until reaching the
reducing cytosolic environment of the neuronal target cells [23,24] . The LC is a zinc-dependent
endopeptidase. The intoxication of cells involves a stepwise sequence of cell surface binding, receptor-
mediated endocytosis, pH-induced translocation, and cytosolic metalloendopr otease activity [24]. The
HC serves as a delivery system for the proteolyti c LC by binding to neurons and transporting the LC
into the cytosol via endosomes. Each BoNT LC cl eaves a component of the soluble N-ethylmaleimide-
sensitive factor attachment protein receptor (SNAR E) proteins [25,26], which are responsible for
transporting acetylcholine into neuromuscular junctions. BoNT serotypes A and E cleave SNAP-25
(synaptosomal-associated protein (2 5 kDa)) [27], serotypes B, D, F, and G cleave VAMP-2 (vesicle-
associated membrane protein, also referred to as synaptobrevin) [ 28-32], and serotype C cleaves both
SNAP-25 and syntaxin 1 [33]. BoNT-mediated cleav age of any one of the three SNARE proteins
terminates the function of autonomic nerves via the inhibition of acetylcholin e release, which produces
flaccid paralysis. Once diaphragm muscle s are affected, suffocation results. | 4fd91a400fdbb432e4d634acfcf42609ede5f4cf | page_0001 |
Molecules 2011 , 16
204
2. Crystal Structures of Botulinum Neurotoxins
The structures of BoNT proteins have been fairly well characterized. The crystal structures of the
BoNT/A, B, and E holotoxins have been reported [23,34,35], and a hand ful of LC crystal structures
and receptor binding domains of BoNT HCs are also available [36-43]. The BoNT proteins contain
three functional domains: the bindi ng domain, the translocation domain, and the catalytic domain. The
LC folds into the catalytic domain and functions as a Zn-dependent endopeptidase that cleaves
SNARE proteins in the neuronal cytosol. The C-te rminus of the HC forms the binding domain, which
targets the cell surface, while the N-terminus of the HC is involved in the tr anslocation of the toxin
across the neuronal membrane [44-46] (Figure 1). Based on sequence and functional similarity, it was
originally believed that the three-dimensional struct ures of BoNTs would also be similar. Indeed, the
structures of the individual func tional domains in sero types BoNT/A, B and E are similar; however,
the overall domain arrangements are different [35] . In the BoNT/A and BoNT/B holotoxins, the three
domains are arranged in a linear fashion, with the tr anslocation domain in the center (Figure 1, left
panel). However, in the BoNT/E holotoxin, both the binding domain and the catalytic domain are on
the same side of the translocation domain, and all three domains mutually interact with one another
(Figure 1, right panel). This unique association may result in the faster rate of internalization and
translocation observed for the BoNT/E, and thus explai ns the faster intoxicati on rate of BoNT/E with
respect to other BoNT serotypes.
Figure 1. Cartoon representations of two types of three-domain organizations of BoNT
holotoxins, ( a) BoNT/A, left panel, and ( b) BoNT/E, right panel. The two structures were
obtained from the Protein Data Bank (PDB codes: 3BTA for BoNT/A and 3FFZ for
BoNT/E).
Comparison of the x-ray structures of the BoNT /A holotoxin and the Bo NT/B holotoxin indicate
that the enzymes’ tran slocation domain protective belts, which wrap around the LCs of both
holotoxins, possess different orientations. Specificall y, in the BoNT/A, the protective belt completely
obstructs the active site and inhi bits substrate binding. Thus, the Bo NT/A LC is catalytically active | 4fd91a400fdbb432e4d634acfcf42609ede5f4cf | page_0002 |
Molecules 2011 , 16
205
only after translocation into the ne uronal cytosol and dissociation from the HC. Experimental evidence
indicates that the BoNT/A LC, after separation fr om the rest of the holot oxin, is the actual active
component [47]. In contrast, the or ientation of the BoNT/B protective belt allows for active site access
relative to the BoNT/A holotoxin [3 9], making the BoNT/B LC catalytic ally active prior to reduction
of the disulfide bond.
The active sites of BoNT LCs contain a common HE XXH (X is any amino acid) zinc-binding motif
with two His residues and one Glu residue ligating the zinc [30-32] , the catalytic water molecule
provides the fourth zinc ligand (the structure of active site of the BoNT/A LC is shown in Figure 2).
All seven BoNT serotype LCs contain one zinc at om, with the exception of the BoNT/C LC, which
possesses two zinc atoms [48]. Superimposition of the x-ray structures of the catalytic clefts of the BoNT/A, B and E LCs indicates that their compositions and geometries are essentially identical [38].
Interestingly, BoNT LC seque nce identities range from 31 −59%, while sequence similarities range
from 52−75% [49]. However, even though the active sites of BoNT LCs are highly homologous,
serotype-specific inhibitors , which preferentially target different LC active sites, have been identified
[50-54].
Figure 2. Active site of the BoNT/A LC (p repared from PDB code 2IMC) [55].
The seven BoNT LC serotypes exhibit unique subs trate selectivity and cleavage site specificity.
Hence, the virtually identical structures of the ac tive sites of the neurotoxins suggest that substrate
recognition is not dictated by the enzymes’ catalytic clefts [38]. Rather, and as evidenced by the crystal
structure of the BoNT/A LC:SNAP- 25 complex which comprises 4,840 Å2 of the enzyme-substrate
interface area, the substrate recognition locations of BoNT LCs are discontinuous and distal to the
catalytic site. Further evidence suppo rting this hypothesis is the large size of the substrate recognition
requirement for the BoNT/A LC ( i.e., a minimum of 17 substrate amin o acids are required for SNAP-
25 cleavage) [56], which is unusual for meta lloproteases. Based on the BoNT/A LC:SNAP-25
complex, two exosites, termed the α-exosite and the β-exosite, are required fo r substrate recognition
[38]. In addition, a Cys165 site (we have termed it as the ϒ-exosite), which is adjacent to the active | 4fd91a400fdbb432e4d634acfcf42609ede5f4cf | page_0003 |
Molecules 2011 , 16
206
site, has also been proposed for small molecule ligand binding [57]. The putative exosites of the
BoNT/F LC, which were identified based on th e crystal structure and molecular modeling, are
different from that of the BoNT/A LC and the BoNT /B LC [41]. In general, the substrate selectivity
and cleavage site specific ities of the BoNT proteas es have provided the base s for the discovery and
development of serotype specific BoNT inhibitors.
3. Approaches to BoNT Inhibition
To counter the BoNT threat, se veral different approaches are currently being explored. While
vaccines will likely play a role in biodefense [58,59] , the development of therapeutic approaches that
are effective both pre- and post-exposure are essentia l. In particular, vaccine s are useless for the post-
exposure protection of previously unvaccinated indi viduals, and the identifica tion and inoculation of
all members of large, at-risk popu lations prior to exposure is probl ematic. Therapeutic approaches
under development include the following: (a) anti-Bo NT antibodies, with the most effective strategy
involving the simultaneous administration of thr ee monoclonal antibodies. The antibodies bind the
BoNT/A with non-overlapping reacti vity, and provide potent protectio n against toxin challenge in
mice [60]; (b) soluble versions of the BoNT/B a nd the BoNT/G receptors (“receptor decoys”),which
function as anti-toxins in cell cu lture and in mice [61,62] (however, in the case of the BoNT/B, the
anti-toxin effect requires the co-adm inistration of gangliosides); and (c) small molecule approaches to
BoNT inhibition. This final categor y is the focus of this review.
4. Broad-Spectrum Small Molecule BoNT Inhibitors
Numerous attempts to develop broad spectrum anta gonists of BoNTs have me t with limited success.
One approach is to inhibit toxin-cell interactions by targeting carbohydrates in general, and sialic acid
in particular, on cellular receptors. This approach is based on the premise that the cellular receptors for
the toxins may not be identical, but may possess cer tain elements of commonality. Lectins, which are
large glycoproteins that are highly specific for thei r sugar moieties, and which exert their effects by
preventing the binding of the BoNT HC to the membrane receptor [63], were found to antagonize
several serotypes. Finally, no small molecules that inhibit toxin-cell recepto r interactions have
been identified.
BoNT-induced muscle paralysis involves holotoxin translocation and subse quent release from an
acidic endosome. In th e early 1980s, the trit erpenoid toosendanin ( 1, Figure 3), was reported to protect
monkeys against BoNT/A, BoNT/B and BoNT/E-indu ced death in a dose-depe ndent fashion [64-67].
In a spinal cord cell-based assay, toosendanin comple tely inhibits SNAP-25 cleavage at concentrations
above 200 nM, and partial inhibition can be observed with concentrations as low as 8 nM for BoNT/A
and 40 nM for BoNT/E. Single molecule channel e xperiments have demonstr ated that toosendanin
exhibits an unprecedented dual mode of action within the protein-condu cting channel, acting both as a
cargo-dependent inhibitor of translo cation and as a cargo-free channel activator [68]. To elucidate the
mechanistic nature of its anti-B oNT properties, several toosendani n analogs have been prepared by
semisynthetic approaches (Figure 3) [69,70]. Only the THF-toosendanin analog 2 exhibited similar
activity, indicating that the furan ring of toosendanin can be modifi ed. However, the epoxide moiety
on the five-membered-ring still seems to be importa nt for anti-BoNT activity, as replacement of the | 4fd91a400fdbb432e4d634acfcf42609ede5f4cf | page_0004 |
Molecules 2011 , 16
207
epoxide moiety with a thermodynamically more stable ketone resulted in inactive compound 3. Ketone
4 and deacetylated compound 5 also lack activity against BoNTs.
Figure 3. Chemical structures of toosendanin ( 1) and its analogs.
The total synthesis of toosendanin ( 1) and its analogs would allow for a more thorough assessment
of the structure-activity relationshi ps (SARs) associated with this chemotype; however, due to the
complexity of the toosendanin stru cture, and the unlikely possibility that a total synthesis will be
developed, an alternative Functi on-Oriented Synthesis (FOS) [71] strategy has been applied to
determine the structural features important for th e anti-BoNT activity of this compound. The principle
of FOS is that the functi on of a biologically active lead structur e can be emulated, tuned, or possibly
improved by replacement with simpler scaffolds desi gned to encompass the key activity-determining
structural features of the natural product. To this end, a CD-ring fragment with an epoxide moiety on
the five-member-ring [69] and tw o epimers of an AB-ring fragment [70] (Figure 4) have been
synthesized and tested.
In a rat spinal cord cellular assay (RSC), additio n of two epimers of the AB-ring fragment at 1 mM
concentrations did not prevent BoNT/A induced SNAP-25 cleavage in prim ary neuronal cells, while
200 µM toosendanin ( 1) resulted in complete inhibition of BoNT/A activity. In a mouse lethality assay
(MLA), intravenous administration of the synthesi zed CD-ring fragment com pound did not protect or
prolong the mean time to death, wh ile at the same concentration t oosendanin extended time to death
7.1 h. No in vitro assay data was reported for the CD-ring fragment.
The endosome acidification proce ss is required for BoNT-induced muscle paralysis. This is
evidenced by the fact that amm onium chloride and methylamine hydr ochloride exhibit concentration-
and time-dependent antagonism of the onset of neuromuscular bl ockade by BoNT/A, /B, /C, and
tetanus toxin [72,73]. However, these amines act solely by antagonizing the internalization of the | 4fd91a400fdbb432e4d634acfcf42609ede5f4cf | page_0005 |
Molecules 2011 , 16
208
toxins by inhibiting endosome acidi fication, since they neither inactiv ate the toxins, nor irreversibly
change tissue function at concentrations that antagonize the onset of BoNT-induced paralysis.
Figure 4. AB-ring and CD-ring fragments generated from a function-or iented synthesis
(FOS) strategy.
O
OAcOO
H
H
HOAcOHO
O
FOSFOSAcO
OO
HO
H
HOO
AcO OH+HO
H
HOO
AcO OHCD ring of toosendanin
toosendanin (1)
ABring o ftoosendanin
Following the same logic, cellular ATPase is require d for the acidification of endocytotic vesicles.
Thus, the inhibition of a vesicle H+-ATPase could result in the antago nism of a broad-spectrum of
BoNTs. To this end, bafilomycin A (Figure 5, compound 6), an ATPase inhibito r, has been shown to
be a universal antagonist of BoNTs A-G, as well as tetanus toxin [74].
Figure 5. Structures of BoNT inhibitors 6 (bafilomycin A 1) and 7 (lomofungin).
A natural product, lomofungin (7, Figure 5) [75], was identified as an inhibitor of the BoNT/A LC
(Ki value of 6.7 ± 0.7 µM) from the high-throughput sc reening (HTS) of a drug library. The screening
was conducted by The Scripps Research Institute . Lomofungin displayed classical noncompetitive
inhibition kinetics and wa s not mutually exclusive when examined in tandem with an active site
inhibitor (2,4-dichlorocinnamic hydroxama te) [76] and a noncompetitive inhibitor, D-chicoric acid.
These data suggest that lomofungin binds to a different ligand binding site on the BoNT/A LC. In the | 4fd91a400fdbb432e4d634acfcf42609ede5f4cf | page_0006 |
Molecules 2011 , 16
209
same report, lomofungin was reported to display weak BoNT/B LC inhibition (IC 50 ≥ 50 µM) in a
FRET-based assay.
5. Small Molecule BoNT/A LC Inhibitors
BoNT LCs induce neuronal paralysi s via the specific proteolysis of SNARE proteins. Therefore,
inhibiting BoNT LC activity has been the major focu s of research efforts to discover and develop
selective therapeutic agents. Several ligand binding si tes have been identified within the BoNT/A LC
substrate binding domain, including the active site, α- and β-exosites, and the Cys165 site ( ϒ-exosite).
Active site inhibitors usually exhibit competitive kinetics vs the SNAP-25 substrate. Co-crystal
structures of 2,4-dichlorocinamic hydroxamate (Figure 6, compound 8) and L-arginine hydroxamate
(Figure 6, compound 9) in complex with the BoNT/A LC ha ve shown that the hydroxamates bind in
the enzyme’s active site, with the cinnamyl side ch ain oriented toward the 370 loop, and the catalytic
water molecule (which ligates the zinc ion), displaced by the hydroxamate moiety (Figure 7).
Figure 6. Structures of BoNT/A LC inhibitors.
The hydroxyl oxygen of the hydroxamate moiety coor dinates the catalytic zinc ion. However, a
dramatic conformational change is observed for the 370 loop in comp aring the two complexes. In the
complex LC: 9, the side chain of Phe369 is withdrawn from the catalytic cleft and the side chain of
Asp370 is exposed, thereby allowing it to form a sa lt bridge with the guanidinium functionality of
compound 9. Rational design based on this co-crystal data resulted in the identification of chiral
compound 8a, which possesses an R- configuration at the β-carbon. Compound 8a displayed a K i value
of 0.16 μM ± 0.002 µM [77]. Moe et al . [78] have reported a series of mercaptoacetamides, for
example compound 8b, which provide low μM anti-BoNT/A LC activity. The SAR of the | 4fd91a400fdbb432e4d634acfcf42609ede5f4cf | page_0007 |
Molecules 2011 , 16
210
mercaptoacetamides is very similar to that of the cinnamic hydroxamates, suggesting that these
inhibitors also bind in the BoNT/A LC catalytic cleft.
Figure 7. (a) Co-crystal structure of BONT/A LC: 8 (PDB code: 2IMA). ( b) Co-crystal
structure of BONT/A LC: 9 (PDB code: 2IMB).
NSC-240898 (Figure 6, compound 10), a bisamidine compound, has been identified as a BoNT/A
LC inhibitor from screening a diversity set of small molecules from the National Cancer Institute’s
Open Repository (using a high throughput FRET-based enzyme assay) [79,80]. Chemical optimization
studies of this lead structure have been conducted at both the University of Pittsburgh [81] and
Microbiotix, Inc. [54].
Figure 8. Proposed binding mode for inhibitor 12 (shown in green stick model). Oxygen
atoms are red, nitrogen atoms are blue, hydrogen atoms are white, and Zn is cyan. The
BoNT/A LC is rendered ribbon. (Reproduced from Li, B. et al. J. Med. Chem .; published
by American Chemical Society [54].)
| 4fd91a400fdbb432e4d634acfcf42609ede5f4cf | page_0008 |
Molecules 2011 , 16
211
One of the most potent analogs of this chemotype is compound 12 (Figure 6), which possesses an
IC50 value of 2.5 µM in a FRET- based enzyme assay. Hence, 12 is 4.4-fold more pot ent than the lead
structure (IC 50 = 11 µM), 3-fold more potent than cinnamic acid hydroxamate 8 (IC 50 = 8.9 µM), and
11.2-fold more selective for the Bo NT/A LC than anthrax lethal f actor (also a metalloprotease).
Compound 11, another analog of NSC-240898, possesses IC 50 values of 12.5 µM and 9.4 µM in a
FRET-based and an HPLC-based assay, respectiv ely, and has shown prot ection against BoNT/A-
induced cleavage of SNAP-25 in both rat and chicken neuronal cell -based assays [82]. Bisamidine 10
and its analogs are competitive BoNT/A LC inhib itors, and molecular modeli ng studies suggest that
they bind in the active site of the BoNT/A LC and do not directly interact with the catalytic zinc ion
(Figure 8).
Recently, an in silico screening campaign, coupled with bioc hemical assays, has been used to
identify BoNT/A LC inhibitors. From this study, quinolinol derivative CB7969312 (Figure 9), has
been reported as a potent inhibitor that pr otects against neuromuscular block in an ex vivo mouse
phrenic nerve hemidiaphram assay (EC 50 = 0.5 µM).
Figure 9. Active site BoNT/A LC inhibitors.
This compound also significantly neutralizes the BoNT/A holotoxin in N2a cells [83]. Biochemical
analysis of the inhibition and binding of this qui nolinol compound with the BoNT/A LC suggest that it
exhibits atypical noncompetitive kine tics [84]. A molecular docking stu dy suggests that the quinolinol
compound binds within a large hydrophobic pocket in the BoNT/A LC active site, and that the
hydroxyquinoline moiety purportedly binds the catalytic zinc. Pang et al . [85] reported a series of
competitive, active site inhibitors based on pyrrole and thiophene structures using synthesis-based
computer-aided molecular design. The most pot ent compound (AHP) displayed BoNT/A inhibition
with a K i value of 0.76 ± 0.17 µM and an IC 50 value of <1 µM, and a thiophene-based compound | 4fd91a400fdbb432e4d634acfcf42609ede5f4cf | page_0009 |
Molecules 2011 , 16
212
(F4H) demonstrated 100% and 70% protect ion of mice against BoNT/A at 5ED 50 within periods of
two and four half-lives, respectivel y, of the inhibitors [86]. In th e study, a single dose of inhibitor was
administered IP (concentration = 2 mg/kg) pre-BoNT challenge.
Several non-Zn-chelating small mo lecule (non-peptidic) BoNT/A LC inhibitors (SMNPIs) have
been described based on pharmacophore-based desi gn [80,87-89]. SMNPIs are continually integrated
into the pharmacophore to both develop three-dime nsional (3D) search queries to discover novel
SMNPI chemotypes and guide the ra tional design of more potent SMNP I derivatives. Employing this
iterative approach, new chemotypes including diazachrysene (Figure 10, compound 13) and
phenylterephthalamide (Figure 10, compound 14), have been identifie d. The pharmacophore model
has guided design and synthesis of compound 15 (Figure 10), which possesses a K i value of 0.572 µM
(± 0.041 µM) vs. the BoNT/A LC.
Figure 10. Small molecule BoNT/A LC inhibitors identified using phamacophore-based design.
A series of benzylidene cyclopentenedione-based inhibitors has been reported to inhibit the
BoNT/A LC metalloprotease by putatively formi ng a covalent bond with the enzyme [90]. Among
these inhibitors, compound 16 (Figure 11) displayed a kinact/KI value of 520 M−1s−1 and SNAP-25
cleavage was significantly decreased at concentrations of 600 µM in a primary rat spinal cord neuron
assay. Unfortunately, compound 16 is highly bound to serum and is reactive with glutathione. Such a
poor pharmacokinetic profile prevents further developm ent of this type of inhibitor. Researchers at
Microbiotix have identified benzimidazole compound 17 (Figure 11) from high throughput, FRET-
based screening of compound libraries against the BoNT/A LC. Compound 17 inhibited the BoNT/A | 4fd91a400fdbb432e4d634acfcf42609ede5f4cf | page_0010 |
Molecules 2011 , 16
213
LC with an IC 50 value of 7.2 µM in the FRET-based assa y (and 10 µM in an HPLC-based assay);
however, no cell-based activity was observed.
Figure 11. Covalent inhibito rs of BoNT/A LC.
Chemical optimization of the benzimidazole acryloni trile series of BoNT/A LC inhibitors yielded
compound 18, which possesses an IC 50 value of 26 µM. Moreover, compound 18 displays 58%
protection of SNAP-25 cleavage at an inhib itor concentration of 30 µM [57]. Silhar et al. reported that
a natural product isolated from Echinacea, D-chicoric acid ( 19) (Figure 12), inhibits BoNT/A LC
activity by binding to an e xosite, and displays noncompetitive pa rtial inhibition of the LC with a
submicromolar inhibition consta nt [76]. In a combination study, D-chicoric acid was synergistic with
competitive BoNT/A LC inhibitor 2,4-dichlorocinnamic hydroxamic acid ( 8).
Figure 12. Chemical structure of D-chicoric acid ( 19).
6. Small Molecule BoNT/B LC Inhibitors
Known zinc chelators, such as bis(5-amidi no-2-benzimidazolyl)methane (BABIM, Figure 13,
compound 20), have been shown to be weak BoNT/B LC inhibitors (IC 50 = 5−10 µM) [91]. X-ray co-
crystal structures of BA BIM in complex with the BoNT/B LC have been reported [39,91]. As shown
in the latter published BABIM:BoNT/B structure, two inhibitor molecules ar e bound to the holotoxin.
One molecule of BABIM enters through a cleft formed between th e translocation domain and the
catalytic domain. Another molecule of BABIM sits in the cleft formed between the translocation
domain and the binding domain, sugge sting that there are two pathways for the inhibitor to enter the
toxin. The two inhibitor molecules do not bind to the enzyme’s catal ytic zinc, as was observed in a
previously reported BoNT/B LC:BABIM structure [39]. The co-crystal structures suggest that, in the
presence of inhibitor, the environm ent of the active site rearranges, a nd the catalytic zi nc is gradually
removed from the active site and transported to a different site of the protein. ICD 1578 ( 22, Figure 12),
a human leukocyte elastase inhib itor, was reported to inhibi t the BoNT/B LC with an IC 50 value of | 4fd91a400fdbb432e4d634acfcf42609ede5f4cf | page_0011 |
Molecules 2011 , 16
214
27.6 µM in a FRET-based assay using a 50-mer synapt obrevin peptide as substrate [92]. However, no
enzyme selectivity data has been presented for compound 22. To date, BoNT/B LC specific small
molecule inhibitors suitable for cl inical use have not been reported.
Figure 13. Small molecule BoNT/B LC inhibitors.
H2NNH
NH
N
RNHNHNNH2
20R=C H2(BABIM)
21R=C OO
OCl
nPr-O
N
HNHO
22(ICD 1578)
7. Summary
As described above, large molecule biologics su ch as monoclonal antibodi es and receptor decoys
may be effective in mouse models or cellular syst ems if they are administered simultaneously with
BoNT; however, they have not been tested for, and are unlikely to be useful for, post-exposure therapy
(i.e., after BoNT penetration into the neuronal cytosol) . Peptide-based inhibitors suffer from a similar
limitation because their large molecu lar size and metabolic instability limit their ability to reach the
BoNT endopeptidase within neurons. Therefore, sma ll molecule, non-peptidic inhibitors offer the best
opportunity for the development of post-exposure therapeutics. Howeve r, none of the reported agents
demonstrate adequate therapeutic utility and none have shown protection in mice due to limited
efficacy, poor cell membrane permeability, cytotoxici ty, or poor pharmacokinetic properties, although
some of them may have prolonged the mean time to death. Thus, more dr ug-like small molecule
botulinum inhibitors, which are pote nt, effective, safe, and possess suitable absorpti on, distribution,
metabolism, excretion, and toxicity (ADMET) prof iles are urgently needed. The key would be to
further refine the design strategy to develop analogs of the lead molecules with improved solubility or
ADMET properties. Viable biological assays to target different phases of BoNT intoxication are also
needed to identify inhibitors wi th novel mechanisms of action. Th e development of high-throughput
screening enzymatic and cell-based assays, as well as structure-based drug design approaches, are
valuable for the identification of inhibitor ‘leads ’ for further optimization. Finally, the validation of in
vivo models of BoNT intoxication will be important for the determination of compounds that will be of
clinical use.
References
1. Arnon, S.S.; Schechter, R.; Inglesby, T.V.; Hend erson, D.A.; Bartlett, J.G.; Ascher, M.S.; Eitzen,
E.; Fine, A.D.; Hauer, J.; Layton, M.; Lillibridge, S.; Osterholm, M.T.; O'Toole, T.; Parker, G.;
Perl, T.M.; Russell, P.K.; Swerdlow, D.L.; Tonat, K. Botulinum toxin as a biological weapon:
Medical and public health management. JAMA 2001 , 285, 1059−1070. | 4fd91a400fdbb432e4d634acfcf42609ede5f4cf | page_0012 |
Molecules 2011 , 16
215
2. Paddle, B.M. Therapy and prophyla xis of inhaled bi ological toxins. J. Appl. Toxicol. 2003 , 23,
139−170.
3. Burnett, J.C.; Henchal, E.A.; Schmaljohn, A.L. ; Bavari, S. The evolvi ng field of biodefense:
Therapeutic developments and diagnostics. Nat. Rev. Drug Discov. 2005 , 4, 281−297.
4. Burnett, J.C.; Schmidt, J.J.; McGrath, C.F. ; Nguyen, T.L.; Hermone, A.R.; Panchal, R.G.;
Vennerstrom, J.L.; Kodukula, K.; Zaharevitz, D.W.; Gussio, R.; Bavari , S. Conformational
sampling of the botulinum neurotoxin serotype A li ght chain: Implications for inhibitor binding.
Bioorg. Med. Chem. 2005 , 13, 333−341.
5. Josko, D. Botulin toxin: A weapon in terrorism. Clin. Lab. Sci. 2004 , 17, 30–34.
6. Clarke, S.C. Bacteria as potential tools in bioterrorism, with an em phasis on bacterial toxins. Br. J.
Biomed. Sci. 2005 , 62, 40−46.
7. Hicks, R.P.; Hartell, M.G.; Nichols, D.A.; Bh attacharjee, A.K.; van Hamont, J.E. Skillman, D.R.
The medicinal chemistry of botu linum, ricin and anthrax toxins. Curr. Med. Chem. 2005 , 12,
667−690.
8. Burnett, J.C.; Henchal, E.A.; Schmaljohn, A.L. ; Bavari, S. The evolvi ng field of biodefense:
Therapeutic developments and diagnostics. Nat. Rev. Drug Discov. 2005 , 4, 281−297.
9. Shukla, H.D.; Sharma, S.K. Clostridium botulinum: A bug with beauty and weapon. Crit. Rev.
Microbiol. 2005 , 31, 11−18.
10. Comella, C.L.; Pullman, S.L. Botuli num toxins in neurological disease. Muscle Nerve 2004 , 29,
628−644.
11. Glogau, R.G. Review of the use of botulinum toxin for hyperhidrosis and cosmetic purposes. Clin.
J. Pain 2002 , 18 (6 Suppl.), S191-S197.
12. Marks, J.D. Medical as pects of biologic toxins. Anesthesiol. Clin. N. Amer. 2004 , 22, 509−532.
13. Montecucco, C.; Molgo, J. Botulinal neurotoxins: Revival of an old killer. Curr. Opin.
Pharmacol. 2005 , 5, 274−279.
14. Bhidayasiri, R.; Truong, D.D. Expanding use of botulinum toxin. J. Neurol. Sci. 2005 , 235, 1−9.
15. Bigalke, H.; Rummel, A. Me dical aspects of toxin weapons. Toxicology 2005 , 214, 210−220.
16. Foster, K.A. A new wrinkle on pain relief: Re-e ngineering clostridial neurotoxins for analgesics.
Drug Discov. Today 2005 , 10, 563−569.
17. Cote, T.R.; Mohan, A.K.; Polder, J.A.; Walto n, M.K.; Braun, M.M. Botulinum toxin type A
injections: Adverse events reported to the US Food and Drug Administration in therapeutic and
cosmetic cases. J. Am. Acad. Dermatol. 2005 , 53, 407−415.
18. Arnon, S.S.; Schechter, R.; Inglesby, T.V.; Hend erson, D.A.; Bartlett, J.G.; Ascher, M.S.; Eitzen,
E.; Fine, A.D.; Hauer, J.; Layton, M.; Lillibridge, S.; Osterholm, M.T.; O'Toole, T.; Parker, G.;
Perl, T.M.; Russell, P.K.; Swerdlow, D.L.; Tonat, K. Botulinum toxin as a biological weapon:
Medical and public health management. JAMA 2001 , 285, 1059−1070.
19. Foran, P.G.; Mohammed, N.; Lisk, G.O.; Nagw aney, S.; Lawrence, G.W.; Johnson, E.; Smith, L.;
Aoki, K.R.; Dolly, J.O. Evaluation of the therapeu tic usefulness of botulinum neurotoxins B, C1,
E, and F compared with the long lasting type A. Basis for distin ct durations of inhibition of
exocytosis in central neurons. J. Biol. Chem. 2003 , 278, 1363−1371.
20. Greenfield, R.A.; Brown, B.R.; Hutchins, J.B.; Iandolo, J.J.; Jack son, R.; Slater, L.N.; Bronze,
M.S. Microbiological, biological, and chem ical weapons of warfare and terrorism. Am. J. Med.
Sci. 2002 , 323, 326−340. | 4fd91a400fdbb432e4d634acfcf42609ede5f4cf | page_0013 |
Molecules 2011 , 16
216
21. Rosenbloom, M.; Leikin, J.B.; Vogel, S.N.; Ch audry, Z.A. Biological and chemical agents: A
brief synopsis. Am. J. Ther. 2002 , 9, 5−14.
22. Meunier, F.A.; Lisk, G.; Sesard ic, D.; Dolly, J.O. Dynamics of motor nerve terminal remodeling
unveiled using SNARE-cleaving botulinum toxins: th e extent and duration are dictated by the
sites of SNAP-25 truncation. Mol. Cell. Neurosci. 2003 , 22, 454−466.
23. Lacy, D.B.; Tepp, W.; Cohen, A.C. ; DasGupta, B.R.; Stevens, R.C. Crystal structure of botulinum
neurotoxin type A and im plications for toxicity. Nat. Struct. Biol. 1998 , 5, 898−902.
24. Simpson, L.L. Identification of the major steps in botulinum toxin action. Annu. Rev. Pharmacol.
Toxicol. 2004 , 44, 167−193.
25. Singh, B.R. Intimate details of the most poisonous poison. Nat. Struct. Biol. 2000 , 7, 617−619.
26. Turton, K.; Chaddock, J.A.; Acharya, K.R. Bo tulinum and tetanus neur otoxins: Structure,
function and therapeutic utility. Trends Biochem. Sci. 2002 , 27, 552−558.
27. Binz, T.; Blasi, J.; Yamasaki, S.; Baumeister , A.; Link, E.; Sudhof, T.C.; Jahn, R.; Niemann, H.
Proteolysis of SNAP-25 by types E and A botulinal neurotoxins. J. Biol. Chem. 1994 , 269,
1617−1620.
28. Schmidt, J.J.; Stafford, R.G. Fluorogenic subs trates for the protease activities of botulinum
neurotoxins, serotypes A, B, and F. Appl. Environ. Microbiol. 2003 , 69, 297−303.
29. Schiavo, G.; Malizio, C.; Trimble, W.S.; Polv erino de Laureto, P.; M ilan, G.; Sugiyama, H.;
Johnson, E.A.; Montecucco, C. Botulinum G neurot oxin cleaves VAMP/synaptobrevin at a single
Ala-Ala peptide bond. J. Biol. Chem. 1994 , 269, 20213−20216.
30. Schiavo, G.; Benfenati, F.; Poulain, B.; Rosse tto, O.; de Polverino, L.P.; DasGupta, B.R.;
Montecucco, C. Tetanus and botulinum-B neur otoxins block neurotransmitter release by
proteolytic cleavage of synaptobrevin. Nature 1992 , 359, 832−835.
31. Schiavo, G.; Rossetto, O.; Catsicas, S.; de Po lverino, L.P.; DasGupta, B.R.; Benfenati, F.;
Montecucco, C. Identification of the nerve terminal targets of botulinum ne urotoxin serotypes A,
D, and E. J. Biol. Chem. 1993 , 268, 23784−23787.
32. Schiavo, G.; Shone, C.C.; Rossetto, O.; Alexande r, F.C.; Montecucco, C. Botulinum neurotoxin
serotype F is a zinc endopeptid ase specific for VAM P/synaptobrevin. J. Biol. Chem. 1993 , 268,
11516−11519.
33. Blasi, J.; Chapman, E.R.; Yamasa ki, S.; Binz, T.; Niemann, H.; Ja hn, R. Botulinum neurotoxin C1
blocks neurotransmitter release by means of cleaving HPC-1/syntaxin. EMBO. J. 1993 , 12,
4821−4828.
34. Swaminathan, S.; Eswaramoorthy, S. Structural analysis of the catalytic and binding sites of
Clostridium botulinum neurotoxin B. Nat. Struct. Biol. 2000 , 7, 693−699.
35. Kumaran, D.; Eswaramoorthy, S.; Furey, W.; Navaza, J.; Sax, M.; Swaminathan, S. Domain
organization in Clostridium botulinum neurotoxin type E is unique: Its implication in faster
translocation. J. Mol. Biol. 2009 , 386, 233−245.
36. Zuniga, J.E.; Schmidt, J.J.; Fenn, T.; Burnett, J.C.; Arac, D.; Gussio, R.; Stafford, R.G.; Badie,
S.S.; Bavari, S.; Brunger, A.T. A potent pepti domimetic inhibitor of botulinum neurotoxin
serotype A has a very different conformation than the SNAP-25 substrate. Structure 2008 , 16,
1588−1597. | 4fd91a400fdbb432e4d634acfcf42609ede5f4cf | page_0014 |
Molecules 2011 , 16
217
37. Silvaggi, N.R.; Wilson, D.; Tzipori, S.; Allen, K. N. Catalytic features of the botulinum neurotoxin
A light chain revealed by high resolution st ructure of an inhibitory peptide complex. Biochemistry
2008 , 47, 5736−5745.
38. Breidenbach, M.A.; Brunger, A.T. Substrat e recognition strategy for botulinum neurotoxin
serotype A. Nature 2004 , 432, 925−929.
39. Hanson, M.A.; Stevens, R.C. Co-crystal st ructure of synaptobrevin-II bound to botulinum
neurotoxin type B at 2.0 Å resolution. Nat. Struct. Biol. 2000 , 7, 687−692.
40. Agarwal, R.; Eswaramoorthy, S.; Kumaran, D.; Binz , T.; Swaminathan, S. Structural analysis of
botulinum neurotoxin type E catalytic domain and its mutant Glu212-->Gln reveals the pivotal
role of the Glu212 carboxylate in the catalytic pathway. Biochemistry 2004 , 43, 6637−6644.
41. Agarwal, R.; Binz, T.; Swaminathan, S. Structur al analysis of botulinum neurotoxin serotype F
light chain: implications on subs trate binding and inhibitor design. Biochemistry 2005 , 44,
11758−11765.
42. Arndt, J.W.; Yu, W.; Bi, F.; Stevens, R.C. Crysta l structure of botulinum neurotoxin type G light
chain: Serotype divergen ce in substrate recognition. Biochemistry 2005 , 44, 9574−9580.
43. Arndt, J.W.; Chai, Q.; Christian, T.; Stevens, R. C. Structure of botulinum neurotoxin type D light
chain at 1.65 Å resolution: repercussi ons for VAMP-2 substrate specificity. Biochemistry 2006 ,
45, 3255−3262.
44. Montecucco, C.; Schiavo, G. Structure and function of tetanus and botulinum neurotoxins. Q. Rev.
Biophys. 1995 , 28, 423−472.
45. Krieglstein, K.G.; DasGupta, B. R.; Henschen, A.H. Covalent st ructure of botulinum neurotoxin
type A: Location of sulfhydryl groups, and disulf ide bridges and identifi cation of C-termini of
light and heavy chains. J. Protein. Chem. 1994 , 13, 49−57.
46. Sagane, Y.; Watanabe, T.; Kouguchi, H.; Suna gawa, H.; Inoue, K.; Fujinaga, Y.; Oguma, K.;
Ohyama, T. Dichain structure of botulinum neurotoxi n: Identification of cleav age sites in types C,
D, and F neurotoxin molecules. J. Protein. Chem. 1999 , 18, 885−892.
47. Gul, N.; Smith, L.A.; Ahmed, S.A. Light chain separated from the rest of the type a botulinum
neurotoxin molecule is the mo st catalytically active form. PLoS One 2010 , 5, e12872.
48. Schiavo, G.; Shone, C.C.; Bennett, M.K.; Sche ller, R.H.; Montecucco, C. Botulinum neurotoxin
type C cleaves a single Lys- Ala bond within the carboxyl-term inal region of syntaxins. J. Biol.
Chem. 1995 , 270, 10566−10570.
49. Lebeda, F.J.; Cer, R.Z.; Mudunur i, U.; Stephens, R.; Singh, B.R.; Alder, M. The zinc-dependent
protease activity of the botulinum neurotoxins. Toxins 2010 , 2, 978−997.
50. Hicks, R.P.; Hartell, M.G.; Nichols, D.A.; Bh attacharjee, A.K.; van Hamont, J.E.; Skillman, D.R.
The medicinal chemistry of botu linum, ricin and anthrax toxins. Curr. Med. Chem. 2005 , 12,
667−690.
51. Willis, B.; Eubanks, L.M.; Dickerson, T.J.; Ja nda, K.D. The strange case of the botulinum
neurotoxin: Using chemistry and biology to modulate the most deadly poison. Angew. Chem. Int.
Ed. 2008 , 47, 8360−8379.
52. Burnett, J.C.; Schmidt, J.J.; Stafford, R. G.; Panchal, R.G.; Nguyen, T.L.; Hermone, A.R.;
Vennerstrom, J.L.; McGrath, C.F.; Lane, D.J.; Sa usville, E.A.; Zaharevitz, D.W.; Gussio, R.;
Bavari, S. Novel small molecule inhibitors of botulinum neurot oxin A metallopr otease activity.
Biochem. Biophys. Res. Commun. 2003 , 310, 84−93. | 4fd91a400fdbb432e4d634acfcf42609ede5f4cf | page_0015 |
Molecules 2011 , 16
218
53. Burnett, J.C.; Opsenica, D.; Sriraghavan, K.; Panchal, R.G.; Ruthel, G.; Hermone, A.R.; Nguyen,
T.L.; Kenny, T.A.; Lane, D.J.; McGrath, C.F.; Schmidt, J.J.; Vennerstrom, J.L.; Gussio, R.; Solaja, B.A.; Bavari, S. A refined pharmacophor e identifies potent 4-am ino-7-chloroquinoline-
based inhibitors of the botulinum ne urotoxin serotype A metalloprotease. J. Med. Chem. 2007 , 50,
2127−2136.
54. Li, B.; Pai, R.; Cardinale, S.C.; Butler, M.M.; Peet, N.P.; Moir, D.T.; Bavari, S.; Bowlin, T.L.
Synthesis and biological ev aluation of botulinum neurot oxin A protease inhibitors. J. Med. Chem.
2010 , 53, 2264−2276.
55. Silvaggi, N.R.; Boldt, G.E.; Hixon, M.S.; Kenne dy, J.P.; Tzipori, S.; Ja nda, K.D.; Allen, K.N.
Structures of clostridium botu linum neurotoxin serotype A light chain complexed with small-
molecule inhibitors highlight active-site flexibility. Chem. Biol. 2007 , 14, 533−542.
56. Schmidt, J.J.; Bostian, K.A. Proteolysis of s ynthetic peptides by type A botulinum neurotoxin. J.
Protein Chem. 1995 , 14, 703−708.
57. Cardinale, S.C.; Butler, M.M.; Ruthel, G.; Nuss, J.E.; Wanner, L.M.; Li, B.; Pai, R.; Peet, N.P.;
Bavari, S.; Bowlin, T.L. Novel benzimidazole inhibitors of botulinum neurotoxin/A display
enzyme and cell based potency. The Botulinum J. 2010 , in press.
58. Park, J.B.; Simpson, L.L. Progress toward de velopment of an inhalation vaccine against
botulinum toxin. Expert Rev. Vaccines 2004 , 3, 477−487.
59. Byrne, M.P.; Smith, L.A. Development of vaccines for prevention of botulism. Biochimie 2000 ,
82, 955−966.
60. Nowakowski, A.; Wang, C.; Powers, D.B.; Amer sdorfer, P.; Smith, T.J.; Montgomery, V.A.;
Sheridan, R.; Blake, R.; Smith, L.A.; Marks, J.D. Potent neutralization of botulinum neurotoxin
by recombinant oligoclonal antibody. Proc. Natl. Acad. Sci. USA 2002 , 99, 11346−11350.
61. Rummel, A.; Karnath, T.; Henke, T.; Bigalke, H.; Binz, T. Synaptotagmins I and II act as nerve
cell receptors for botulinum neurotoxin G.
J. Biol. Chem. 2004 , 279, 30865−30870.
62. Dong, M.; Richards, D.A.; Goodnough, M.C.; Te pp, W.H.; Johnson, E.A.; Chapman, E.R.
Synaptotagmins I and II mediate entry of botulinum neurot oxin B into cells. J. Cell Biol. 2003 ,
162, 1293−1303.
63. Bakry, N.; Kamata, Y.; Simpson, L.L. Lectins from Triticum vulgaris and Limax flavus are
universal antagonists of botulinum neurotoxin an d tetanus toxin. J. Pharmacol. Exp. Ther. 1991 ,
258, 830−836.
64. Li, P.-Z.; Zhou, J.; Miao, W.-Y.; Ding, F. -H.; Meng, J.-Y.; Jia, G.-R.; Li, J.-F.; Ye, H.-J.; He, X.-
Y.; Chen, M.-Y.; Huang, Z.-M. Therapeutic e ffect of toosendanin on animal botulism. Tradit.
Herb. Drugs 1982 , 13, 28−30.
65. Shi, Y.-L.; Li, M.-F. Biologica l effects of toosendani n, a triterpenoid ex tracted from Chinese
traditional medicine. Prog. Neurobiol. 2007 , 82, 1−10.
66. Shi, Y.-L.; Wang, Z.-F. Cure of experimental botulism and antibotulismic effect of toosendanin.
Acta Pharmacol. Sin. 2004 , 25, 839−848.
67. Zou, J.; Miao, W.-Y.; Ding, F.-H.; Meng, J.-Y.; Ye , H.-J.; Jia, G.-R.; He, X.-Y.; Sun, G.-Z.; Li,
P.-Z. The effect of toosendanin on monkey botulism. J. Tradit. Chin. Med. 1985 , 5, 29−30.
68. Fischer, A.; Nakai, Y.; Eubanks, L.M.; Clancy, C.M.; Tepp, W.H.; Pellett, S.; Dickerson, T.J.;
Johnson, E.A.; Janda, K.D.; Montal, M. Bimodal m odulation of the botulinum neurotoxin protein-
conducting channel. Proc. Natl. Acad. Sci. USA 2009 , 106, 1330−1335. | 4fd91a400fdbb432e4d634acfcf42609ede5f4cf | page_0016 |
Molecules 2011 , 16
219
69. Nakai, Y.; Tepp, W.H.; Dickerson, T.J.; Johnson, E.A.; Janda, K.D. Functio n-oriented synthesis
applied to the anti-botulinum natural product toosendanin. Bioorg. Med. Chem. 2009 , 17,
1152−1157.
70. Nakai, Y.; Pellett, S.; Tepp, W.H.; Johnson, E.A.; Janda, K.D. Toosendanin: Synthesis of the AB-
ring and investigations of its an ti-botulinum properties (Part II). Bioorg. Med. Chem. 2010 , 18,
1280−1287.
71. Wender, P.A.; Verma, V.A.; Paxton, T.J.; Pillo w, T.H. Function-oriented synthesis, step
economy, and drug design. Acc. Chem. Res. 2008 , 41, 40−49.
72. Simpson, L.L. The interaction between aminoqui nolines and presynaptica lly acting ne urotoxins.
J. Pharmacol. Exp. Ther. 1982 , 222, 43−48.
73. Simpson, L.L. Ammonium chloride and met hylamine hydrochloride antagonize clostridial
neurotoxins. J. Pharmacol. Exp. Ther. 1983 , 225, 546−552.
74. Simpson, L.L.; Coffield, J.A.; Bakry, N. I nhibition of vacuolar ad enosine triphosphatase
antagonizes the effects of cl ostridial neurotoxins but not phospholipase A2 neurotoxins. J.
Pharmacol. Exp. Ther. 1994 , 269, 256−262.
75. Eubanks, L.M.; Silhar, P.; Salzameda, N.T. ; Zakhari, J.S.; Xiaochuan, F.; Barbieri, J.T.;
Shoemaker, C.B.; Hixon, M.S.; Janda, K.D. Identif ication of a natural pr oduct antagonist against
the botulinum neurotoxin light chain protease. ACS Med. Chem. Lett. 2010 , 1, 268−272.
76. Silhar, P.; Capkova, K.; Salzameda, N.T.; Barb ieri, J.T.; Hixon, M.S.; Janda, K.D. Botulinum
neurotoxin A protease: Di scovery of natural pro duct exosite inhibitors. J. Am. Chem. Soc. 2010,
132, 2868−2869.
77. Stowe, G.N.; Silhar, P.; Hixon, M.S.; Silvaggi , N.R.; Allen, K.N.; Moe, S.T.; Jacobson, A.R.;
Barbieri, J.T.; Janda, K.D. Chirality holds th e key for potent inhibition of the botulinum
neurotoxin serotype a protease. Org. Lett. 2010 , 12, 756−759.
78. Moe, S.T.; Thompson, A.B.; Smith, G.M.; Fredenburg, R.A.; Stein, R.L.; Jacobson, A.R.
Botulinum neurotoxin serotype A inhibitors: Small molecule mercaptoacetamide analogs. Bioorg.
Med. Chem. 2009 , 17, 3072−3079.
79. Burnett, J.C.; Ruthel, G.; Stegmann, C.M.; Panchal, R.G.; Nguyen, T.L.; Hermone, A.R.;
Stafford, R.G.; Lane, D.J.; Kenny, T.A.; McGrath, C.F.; Wipf, P.; Stahl, A.M.; Schmidt, J.J.;
Gussio, R.; Brunger, A.T.; Bavari, S. Inhibition of metalloprotease botulinum serotype A from a
pseudo-peptide binding mode to a small mol ecule that is active in primary neurons. J. Biol. Chem.
2007 , 282, 5004−5014.
80. Burnett, J.C.; Wang, C.; Nuss, J.E.; Nguyen, T.L.; Hermone, A.R.; Schmidt, J.J.; Gussio, R.;
Wipf, P.; Bavari, S. Pharmacophore-guided lead optimization: The rational design of a non-zinc
coordinating, sub-micromolar inhibitor of the botulinum neurotoxin sero type A metalloprotease.
Bioorg. Med. Chem. Lett. 2009 , 19, 5811−5813.
81. Wang, C.; Widom, J.; Petronijevic, F.; Burnett, J.C.; Nuss, J.E.; Bavari, S.; Gussio, R.; Wipf, P.
Synthesis and biological evaluati on of inhibitors of botulinum neurotoxin metalloprotease.
Heterocycles 2009 , 79, 487−520.
82. Butler, M. M.; Cardinale, S. C. ; Li, B.; Pai, R.; Ruthel, G.; Nuss, J. E.; Wanner, L. M.; Park, J.-B.;
Rich, C.; Basu, A.; Mills, D.; Williams, J.D.; P eet, N.P.; Moir, D.; Bavari, S.; Bowlin, T.L.
Unpublished results. | 4fd91a400fdbb432e4d634acfcf42609ede5f4cf | page_0017 |
Molecules 2011 , 16
220
83. Roxas-Duncan, V.; Enyedy, I.; Montgomery, V.A. ; Eccard, V.S.; Carrington, M.A.; Lai, H.; Gul,
N.; Yang, D.C.H.; Smith, L.A. Identification and bi ochemical characterization of small molecule
inhibitors of clostridium botulinum neurotoxin seroptype A. Antimicro. Agents Chemother. 2009 ,
53, 3478−3486.
84. Lai, H.; Feng, M.; Roxas-Duncan, V.; Dakshanamu rthy, S.; Smith, L. A.; Yang, D.C.H. Quinolinol
and peptide inhibitors of zinc protease in botulinum neurotoxin A: Effects of zinc ion and peptides
on inhibition. Arch. Biochem. Biophys. 2009 , 491, 75−84.
85. Pang, Y.-P.; Vummenthala, A.; Mishra, R.K.; Park, J.G.; Wang, S.; Davis, J.; Millard, C.B.;
Schmidt, J.J. Potent new small molecule i nhibitor of botulinum neurotoxin serotype A
endopeptidase developed by synthesis-ba sed computer-aided molecular design. PLos One 2009 ,
4, e7730.
86. Pang, Y.-P.; Davis, J.; Wang, S.; Park, J.G.; Namb iar, M.P.; Schmidt, J.J.; Millard, C.B. Small
molecules showing significant pr otection of mice against botuli num neurotoxin serotype A. PLoS
One 2010 , 5, e10129.
87. Hermone, A.R.; Burnett, J.C.; Nuss, J.E.; Tressler, L.E.; Nguyen, T.L.; Solaja, B.A.;
Vennerstrom, J.L.; Schmidt, J.J.; Wipf, P.; Bava ri, S.; Gussio, R. Three-dimensional database
mining identifies a unique chemotype that unite s structurally divers e botulinum neurotoxin
serotype A inhibitors in a three-zone pharmacophore. ChemMedChem 2008 , 3, 1905−1912.
88. Nuss, J.E.; Dong, Y.; Wanner, L.M.; Ruthel, G.; Wipf, P.; Gussio, R.; Vennerstrom, J.L.; Bavari,
S.; Burnett, J.C. Pharmacophore refinement guides the rational desi gn of nanomolar-range
inhibitors of the botulinum neur otoxin serotype A metalloprotease. ACS Med. Chem. Lett. 2010 ,
1, 301−305.
89. Solaja, B.A.; Opsenica, D.; Smith, K.S.; Milhous, W.K.; Terzic, N.; Opsenica, I.; Burnett, J.C.;
Nuss, J.; Gussio, R.; Bavari, S. Novel 4-aminoqui nolines active against chloroquine-resistant and
sensitive P. falciparum strains that also inhibit botulinum serotype A. J. Med. Chem. 2008 , 51,
4388−4391.
90. Capkova, K.; Hixon, M.S.; Pellett, S.; Barbieri, J.T.; Johnson, E.A.; Janda, K.D. Benzylidene
cyclopentenediones: First irreve rsible inhibitors against bot ulinum neurotoxin A's zinc
endopeptidase. Bioorg. Med. Chem. Lett. 2009 , 20, 206−208.
91. Eswaramoorthy, S.; Kumaran, D.; Swaminathan, S. A novel mechanism for clostridium
botulinum neurotoxin inhibition. Biochemistry 2002 , 41, 9795−9802.
92. Adler, M.; Nicholson, J.D.; Cornille, F.; Hack ley, B.E., Jr. Efficacy of a novel metalloprotease
inhibitor on botulinum neurotoxin B activity. FEBS Lett. 1998 , 429, 234−238.
Sample Availability: Not available.
© 2010 by the authors; licensee MDPI, Basel, Switz erland. This article is an open access article
distributed under the terms and conditions of the Creative Commons Attribution license
(http://creativecommons .org/licenses/by/3.0/). | 4fd91a400fdbb432e4d634acfcf42609ede5f4cf | page_0018 |
Building one molecule from a reservoir of two atoms
L. R. Liu,123J. D. Hood,13Y . Yu,123J. T. Zhang,123N. R. Hutzler,123y
T. Rosenband,2and K.-K. Ni123
1Department of Chemistry and Chemical Biology, Harvard University,
Cambridge, Massachusetts, 02138, USA
2Department of Physics, Harvard University, Cambridge, Massachusetts, 02138, USA
3Harvard-MIT Center for Ultracold Atoms, Cambridge, Massachusetts, 02138, USA
yCurrent address: Division of Physics, Mathematics, and Astronomy,
California Institute of Technology, Pasadena, CA, 91125, USA
To whom correspondence should be addressed; E-mail: ni@chemistry.harvard.edu
April 25, 2018
Chemical reactions typically proceed via stochastic encounters between re-
actants. Going beyond this paradigm, we combine exactly two atoms into a
single, controlled reaction. The experimental apparatus traps two individual
laser-cooled atoms (one sodium and one cesium) in separate optical tweez-
ers and then merges them into one optical dipole trap. Subsequently, photo-
association forms an excited-state NaCs molecule. The discovery of previously
unseen resonances near the molecular dissociation threshold and measure-
ment of collision rates are enabled by the tightly trapped ultracold sample
of atoms. As laser-cooling and trapping capabilities are extended to more el-
ements, the technique will enable the study of more diverse, and eventually
more complex, molecules in an isolated environment, as well as synthesis of
1arXiv:1804.04752v2 [physics.atom-ph] 23 Apr 2018 | a54ea132ff254451ee01a8d4a3674004c6e7baf3 | page_0000 |
designer molecules for qubits.
Chemical reactions proceed through individual collisions between atoms or molecules. How-
ever, when performed in stochastic ensembles, the individual reaction probabilities are observed
as averages. Crossed molecular beams reduce the thermal velocity dispersion to probe elemen-
tary reaction processes based on single collision events, illuminating many aspects of reaction
dynamics ( 1–4). In quantum degenerate gases, cooled to temperatures below 1K, the quan-
tum motional degrees of freedom play a critical role in the reaction ( 5–7). Comparisons of
such experimental reaction rates with theoretical models currently underpin our understanding
of reactions at the most elementary level ( 8–10 ).
To further improve the specificity and precision of reaction steps ( 11–13 ), individual particle
control is needed, similar to pioneering atom-positioning experiments with scanning tunneling
microscopes ( 14), but untethered from surfaces. By controlling individual particles via laser
cooling and optical trapping, molecules may be constructed atom by atom, while maintaining
specific internal and external quantum states.
Herein, we realize chemistry in the minimum number regime, where precisely two atoms
are brought together to form one molecule with the aid of a photon. We achieve this by using
movable optical tweezers, where individual atoms of different elements (here Na and Cs) are
isolated, cooled, manipulated, and eventually combined into a single optical tweezer. With
exactly two atoms in an optical tweezer, we can observe their collisions. We can also perform
single molecule spectroscopy in the gas phase by optically exciting the atom pair on a molecular
transition, thereby realizing the chemical reaction Na+Cs!NaCs. Subsequent imaging of
Na and Cs fluorescence distinguishes between four possible experimental outcomes: both, only
one, or no atoms are detected in the tweezer, the latter indicating a reaction has occurred. We
chose NaCs for the demonstration because it possesses a large molecular fixed-frame dipole
moment of 4.6 Debye ( 15), making it a strong candidate for a molecular qubit in a future
2 | a54ea132ff254451ee01a8d4a3674004c6e7baf3 | page_0001 |
quantum computing architecture.
We began by preparing laser-cooled Na and Cs atoms at a few hundred K in overlapped
magneto-optical traps (MOTs) in a vacuum chamber ( 10 8Pa). The MOTs serve as cold atom
reservoirs for loading single atoms into tightly focused optical tweezer traps ( 16). After loading,
the MOTs are extinguished. A schematic of the apparatus is shown in Fig. 1A. The NA=0.55 mi-
croscope objective focuses two different wavelengths of light, 700 nm and 976 nm, to waists of
0.7m and 0.8m radius. Due to the difference in Na and Cs polarizabilities, the 700 nm wave-
length light attracts Na and repels Cs, while 976 nm light attracts Cs five times more strongly
than Na ( 17), enabling us to manipulate Na and Cs independently as depicted in Fig. 2A. A
typical trap depth of 1 mK is achieved for 5 mW of tweezer power.
When tightly confined identical atoms are illuminated with near-atomic-resonant light, light-
assisted pairwise collisions result in either zero or one final atom in the trap ( 16,18 ). Single atom
loading succeeds approximately half of the time ( 19). However, the large light shifts for Na in a
700 nm wavelength tweezer would normally prevent atom cooling, and consequently, efficient
atom loading. We eliminate this light shift for Na by alternating the tweezer and cooling beams
at a rate of 3 MHz ( 20). Subsequently, Na, followed by Cs, are imaged and polarization gra-
dient cooled to 70 K and 10K respectively. To determine whether an atom is in the optical
tweezer, the fluorescence photoelectron counts from each atom in a region of interest (Fig. 1B)
are compared to a threshold (Fig. 1C). The fluorescence histograms indicate that the cases of
zero or one atom can be distinguished with a fidelity better than 99.97 %. We find that in 33%
of cases we load a single Na and a single Cs atom side-by-side. In 18% of cases, no atoms
are loaded, and the rest of the time either a single Na or a Cs atom is loaded (Fig. 1B). The
experiment, which repeats at 3 Hz, records initial and final fluorescence images to determine
survival probabilities for different stages of the molecule formation process.
Once single atoms have been loaded in separate traps, they need to be transported to the same
3 | a54ea132ff254451ee01a8d4a3674004c6e7baf3 | page_0002 |
location for molecule formation. Optical tweezers have been used to move single atoms while
maintaining atomic internal state coherence ( 21) and to merge two indistinguishable atoms by
coherent tunneling into one tweezer ( 22). Here we adiabatically transport and merge two differ-
ent atoms, Na and Cs, into the same tweezer, as depicted in Fig. 2A by using optical tweezers at
two different wavelengths. The trap depths are adjusted by changing the beam intensities, and
the positions are steered by applying different radio frequencies to the respective acousto-optic
deflectors (AODs) (Fig. 1A).
For the merge sequence, the 700 nm tweezer containing Na is kept stationary while the
976 nm tweezer containing Cs is moved to overlap the atoms (Fig. 2A, panel I - III). Following
the merge, the 700 nm tweezer is extinguished adiabatically to leave both atoms in the 976 nm
tweezer (Fig. 2A, panel IV). We design this merge trajectory such that i) Cs is deeply confined
at all times, and ii) the double-well potential imposed on Na is sufficiently asymmetric to avoid
a near-degenerate ground state. This process is time-reversible, which enables us to image the
atoms separately and determine survival probability.
Because the 700 nm tweezer is extinguished for 1 ms after the merge, while the 976 nm
tweezer is always active, the Na atom escapes unless the two tweezers are overlapped at the
end of the merge sequence, whereas the Cs atom is always trapped. Fig. 2B shows the result
obtained when scanning the endpoint of the 976 nm tweezer trajectory. The height of the Na
survival peak at 0 m of 94(1)% is near the re-imaging survival probability of 96%.
Having demonstrated adiabatic transport and merging of two species into a tight tweezer, we
turn to their collisions. Isolated collisions between two atoms do not usually result in molecule
formation due to the need to simultaneously conserve momentum and energy. However, the
atoms can change their hyperfine states after colliding, and the exothermic hyperfine-spin-
changing collisions impart enough kinetic energy ( 100mK) to the atoms to eject them from
the tweezer (1mK depth) ( 23).
4 | a54ea132ff254451ee01a8d4a3674004c6e7baf3 | page_0003 |
Generally, a given initial trap occupancy can evolve into 4 possible outcomes following an
experiment: i) both atoms, ii) no atoms, iii) only Cs, and iv) only Na remain in the trap. Single-
atom images from each repetition allow us to post-select on any of these cases and separate 1-
and 2-body processes, giving both lifetimes from a single dataset (Fig. 3). For example, when
Na and Cs are both present (effective pair density of n2= 21012cm 3(24)), and prepared
in a mixture of hyperfine spin states, they are both rapidly lost loss= 8(1) ms, where loss
is the 1=etime of exponential decay. This yields a loss rate constant = 510 11cm3=s.
In contrast, if the atoms are both optically pumped into the lowest energy hyperfine levels,
conservation of energy prevents the change of hyperfine states, and the atom lifetime increases
to 0.63(1) s, similar to the rate of hyperfine-state relaxation for Cs due to off-resonant scattering
of the tweezer light ( 25). When only one atom is present, 1-body loss due to collisions with
background gas limits the lifetime to 5 s.
Because of the rapid 2-body loss for mixed hyperfine states, we optically pump each atom
into its lowest energy hyperfine state to maintain a long-lived sample of co-trapped Na and Cs
atoms. We then perform photoassociation (PA) of the atoms to form an excited state molecule,
realizing a single instance of the chemical reaction Na +Cs!NaCs. When illuminating the
atoms with resonant PA light, an electronically excited state molecule may form (Fig. 4A) and
then rapidly decay to the ground state. The molecule does not scatter imaging light, causing
molecule formation to manifest as simultaneous loss of both Na and Cs atoms. The bottom
panel of Fig. 4B shows these loss resonances as the frequency of the PA light is scanned below
the dissociation threshold.
Our optical tweezer architecture offers a number of advantages for PA measurements over
previous methods with bulk samples ( 26). The ability to precisely define the initial reagents
eliminates contributions from other reaction processes such as Cs 2formation or 3-body loss.
The combination of the high effective pair density ( 24)n2= 31012cm 3, afforded by the
5 | a54ea132ff254451ee01a8d4a3674004c6e7baf3 | page_0004 |
tweezer confinement, and high PA light intensity of 3kW=cm2yields fast PA rates. The high
contrast measurements of single-atom loss result in near-unity molecule detection efficiency
and avoid the need for ionization detection ( 26).
We scan the 200 MHz frequency-broadened PA light from 30 to 250 GHz below the Cs
atomic D2 line (6S 1=2- 6P 3=2). We take steps of 200 MHz with 100ms pulse duration, and
average over approximately 100 repetitions at each data point. An absolute accuracy of 1 GHz is
set by the wavemeter. During PA, the Cs atom could be promoted into the upper hyperfine level
due to off-resonant scattering of the PA beam, which would lead to spin-changing collisional
loss. We counteract this effect by simultaneously optically pumping Cs into the lower hyperfine
level with a separate beam.
The ability to detect molecule formation via atom loss with high efficiency allows us to
probe NaCs* vibrational levels near the dissociation threshold, including resonances that have
not been previously observed (Fig. 4). According to ab initio calculations of NaCs* with spin-
orbit coupling ( 27), five molecular potentials converge to the Cs (6P 3=2) + Na (3S 1=2) asymptote
(Fig. 4A):B11,c3+
=0;1, andb3
=0;2. Of these, only the c3+
1levels have previously been
observed in the near-threshold regime ( 28), and our measurement agrees to within 1 GHz. To
identify the vibrational progressions, we fit the LeRoy-Bernstein (LB) dispersion model ( 29) to
our observed resonances. Near threshold, the vibrational quantum number v0(v0= 1is the
highest bound state) is related to the binding energy by
Ev0= 1
C1=2
6"
2h2
1=2 (7=6)
(2=3)(v0 v0
0)#3
; (1)
whereis the reduced mass, and his the reduced Planck’s constant. We extract the C6dis-
persion coefficients that characterize the 1=r6component of the potentials and v0
0, which is an
offset between -1 and 0.
Fitting to the positions of our observed c3+
1resonances gives v0= 0:79andC6=
6 | a54ea132ff254451ee01a8d4a3674004c6e7baf3 | page_0005 |
8:5(6)103a.u. (in atomic units), in agreement with the theoretical value C6= 7:96
103a.u. ( 30). From the remaining loss resonances, we identify two additional progressions
(B11andc3+
0) withC6= 1:42(33)104a.u. andC6= 1:47(26)104a.u. (Fig. 4B). Both
values are near the theoretical value of C6= 1:83104a.u. ( 30). Our state labels correspond to
the molecular wavefunctions in the near-threshold regime and differ from the labels in Ref. ( 30)
due to an avoided crossing as noted in Ref. ( 31). Here the assignment of the c3+
1progression is
based on previous observation of the same resonances ( 28), whileB11continues a previously
observed sequence ( 31). The remaining progression corresponds to c3+
0, because this is the
only other compatible state. We interpret the photoassociation spectrum as clear evidence for
molecule formation, because the resonance peaks appear exclusively as simultaneous loss of Na
and Cs, and the resonance frequencies agree with independent measurements.
Our technique can in principle be extended beyond the simple bialkalis demonstrated here,
and to produce deeply bound molecules. Molecules in a single quantum state could be created
by coherent transfer ( 32, 33 ) of atoms prepared in the motional ground state ( 34–37 ). Dipolar
molecules trapped in a configurable array of optical tweezers ( 38, 39 ) would constitute a new
type of qubit for quantum information processing ( 40) and an important resource to explore
quantum phases ( 41, 42 ).
References
1. D. R. Herschbach, Angewandte Chemie International Edition in English 26, 1221 (1987).
2. Y . T. Lee, Angewandte Chemie International Edition in English 26, 939 (1987).
3. A. B. Henson, S. Gersten, Y . Shagam, J. Narevicius, E. Narevicius, Science 338, 234
(2012).
4. W. E. Perreault, N. Mukherjee, R. N. Zare, Science 358, 356 (2017).
7 | a54ea132ff254451ee01a8d4a3674004c6e7baf3 | page_0006 |
5. S. Ospelkaus, et al. ,Science 327, 853 (2010).
6. K.-K. Ni, et al. ,Nature 464, 1324 (2010).
7. M. H. G. de Miranda, et al. ,Nature Physics 7, 502 (2011).
8. K. Liu, Annual Review of Physical Chemistry 52, 139 (2001).
9. X. Yang, Annual Review of Physical Chemistry 58, 433 (2007). PMID: 17105413.
10. A. Klein, et al. ,Nature Physics 13, 35 EP (2016).
11. L. Ratschbacher, C. Zipkes, C. Sias, M. K ¨ohl,Nature Physics 8, 649 EP (2012).
12. S. A. Moses, et al. ,Science 350, 659 (2015).
13. P. Puri, et al. ,Science 357, 1370 (2017).
14. D. M. Eigler, E. K. Schweizer, Nature 344, 524 EP (1990).
15. P. J. Dagdigian, L. Wharton, The Journal of Chemical Physics 57, 1487 (1972).
16. N. Schlosser, G. Reymond, I. Protsenko, P. Grangier, Nature 411, 1024 (2001).
17. M. S. Safronova, B. Arora, C. W. Clark, Phys. Rev. A 73, 022505 (2006).
18. A. Fuhrmanek, R. Bourgain, Y . R. P. Sortais, A. Browaeys, Phys. Rev. A 85, 062708 (2012).
19. P. Sompet, A. V . Carpentier, Y . H. Fung, M. McGovern, M. F. Andersen, Phys. Rev. A 88,
051401 (2013).
20. N. R. Hutzler, L. R. Liu, Y . Yu, K.-K. Ni, New Journal of Physics 19, 023007 (2017).
21. J. Beugnon, et al. ,Nat Phys 3, 696 (2007).
8 | a54ea132ff254451ee01a8d4a3674004c6e7baf3 | page_0007 |
22. A. M. Kaufman, et al. ,Science 345, 306 (2014).
23. B. Ueberholz, S. Kuhr, D. Frese, D. Meschede, V . Gomer, Journal of Physics B: Atomic,
Molecular and Optical Physics 33, L135 (2000).
24. See supplementary material.
25. R. A. Cline, J. D. Miller, M. R. Matthews, D. J. Heinzen, Opt. Lett. 19, 207 (1994).
26. K. M. Jones, E. Tiesinga, P. D. Lett, P. S. Julienne, Reviews of Modern Physics 78, 483
(2006).
27. M. Korek, S. Bleik, A. R. Allouche, The Journal of Chemical Physics 126, 124313 (2007).
28. A. Grochola, et al. ,Phys. Rev. A 84, 012507 (2011).
29. R. L. Roy, R. Bernstein, Journal of Chemical Physics 52, 3869 (1970).
30. M. Marinescu, H. Sadeghpour, Physical Review A 59, 390 (1999).
31. A. Grochola, P. Kowalczyk, W. Jastrzebski, Chemical Physics Letters 497, 22 (2010).
32. K. Bergmann, H. Theuer, B. W. Shore, Rev. Mod. Phys. 70, 1003 (1998).
33. L. R. Liu, et al. ,ArXiv:1701.03121 (2017).
34. C. Monroe, et al. ,Phys. Rev. Lett. 75, 4011 (1995).
35. X. Li, T. A. Corcovilos, Y . Wang, D. S. Weiss, Phys. Rev. Lett. 108, 103001 (2012).
36. A. M. Kaufman, B. J. Lester, C. A. Regal, Phys. Rev. X 2, 041014 (2012).
37. Y . Yu, et al. ,ArXiv:1708.03296 (2017).
9 | a54ea132ff254451ee01a8d4a3674004c6e7baf3 | page_0008 |
38. D. Barredo, S. de L ´es´eleuc, V . Lienhard, T. Lahaye, A. Browaeys, Science 354, 1021
(2016).
39. M. Endres, et al. ,Science 354, 1024 (2016).
40. D. DeMille, Phys. Rev. Lett. 88, 067901 (2002).
41. N. Y . Yao, M. P. Zaletel, D. M. Stamper-Kurn, A. Vishwanath, arXiv preprint
arXiv:1510.06403 (2015).
42. B. Sundar, B. Gadway, K. R. Hazzard, arXiv preprint arXiv:1708.02112 (2017).
43. C. Tuchendler, A. M. Lance, A. Browaeys, Y . R. Sortais, P. Grangier, Physical Review A
78, 033425 (2008).
Acknowledgments
We thank R. Gonz ´alez-F ´erez, P. Julienne, D. DeMille, and C. Regal for discussions. K.-K. N.
thanks D. S. Jin for encouragement to pursue the research presented here. Funding: This work is
supported by the Arnold and Mabel Beckman Foundation, as well as the AFOSR Young Inves-
tigator Program, the NSF through the Harvard-MIT CUA, and the Alfred P. Sloan Foundation.
Author contributions: L. R. L., J. D. H., Y . Y ., J. T. Z., N. R. H., T. R., K.-K. N. performed
the experiment. L. R. L., J. D. H., T. R., K.-K. N. analyzed the data and wrote the manuscript.
Competing interests: None. Data and materials availability: All data are supplied in the paper
and supplementary material.
Supplementary materials
Materials and Methods
Supplementary Text
10 | a54ea132ff254451ee01a8d4a3674004c6e7baf3 | page_0009 |
Fig. S1
Table S1
11 | a54ea132ff254451ee01a8d4a3674004c6e7baf3 | page_0010 |
AOD
(Na)
dichroic
mirror
imaging lens
objectiveCCD
UHV glass cell
AOD
(Cs)A)
mirror
976 nm tweezer
700 nm tweezerdichroic
mirror
NA = 0.55B)
C)1 µm
from laser
from laser
852 nm
fluorescence
589 nm
fluorescence
Figure 1: Dual-species single atom trapping and imaging. A ) Schematic of the setup. Optical
tweezer atom trapping beams (700 nm and 976 nm wavelengths) are independently steered
by acousto-optic deflectors, expanded by telescopes, and then combined on a dichroic mirror
before being focused by the objective into a glass cell. Fluorescence from trapped Na and Cs
atoms is collected through the objective onto the CCD camera. B) Fluorescence images of
single Na and Cs atoms. Length scale of 1m is indicated. Cs (top panels) and Na (bottom
panels) are imaged sequentially in the same field of view. The four possible cases are shown
with their initial loading probabilities: no atoms, a single Na atom, a single Cs atom, both
Na and Cs atoms. Dashed blue (Cs) and orange (Na) boxes indicate the region of interest for
determining presence of atoms. C) Histogram of Cs (blue) and Na (orange) fluorescence. The
bimodal distribution shows clear separation between zero- and one-atom peaks. Red dashed
lines indicate the threshold that is used to determine the presence of an atom.
12 | a54ea132ff254451ee01a8d4a3674004c6e7baf3 | page_0011 |
A) B)
●
●●●●●●●●●●
●●
●●●●
■
■
■
■■■■■■
■
■
■
■■
■
■
■■
-4 -2 0 2 40.20.40.60.81.0
Tweezer Radial Position(μm)Survival probability
Cs
Na●
150 kHz132 kHz150 kHz 289 kHz
IV
5.3 msCs
Na
●
■ ■■ ■
●-4 -3 -2 -1 0 1 2-2-10
Tweezer Radial Position(μm)III
3.7 ms
●■ ■
-2-10II
2.2 ms
●
■ ■-2-10I
0 ms
●
■ ■-2-10Trap Depth(mK)Figure 2: Merging single Na and Cs atoms, which are initially separated by 3 m, into one
tweezer. A ) 1-D cuts of the combined, time-varying 700 nm and 976 nm tweezer potentials for
both atoms during the merge sequence. Na and Cs are represented by dots that track the minima
of their potentials (orange for Na and blue for Cs). Overlaid are graphics of the optical tweezers.
Radial trap frequencies are labeled in the first and last panels (axial trap frequencies are roughly
6 times smaller). Panels I-III depict the merging process. In panel IV , the 700 nm tweezer
has been extinguished and only the 976 nm tweezer remains. B) Measured survival probability
of Na and Cs after the sequence depicted in (A), followed by separating the tweezers through
a reverse sequence to image the atoms. The two atoms are merged into the same tweezer at
the survival maximum for Na. Error bars denote the Wilson score interval. The dashed lines
represent the survival rates due to imperfect re-imaging.
13 | a54ea132ff254451ee01a8d4a3674004c6e7baf3 | page_0012 |
Starting with both atoms Starting with one atom
lower HF
0.63(1) s
5.1(3) s
5.3(1) sNa only
Cs onlymixed HF
8(1) msno atomsboth atomsNa only
Cs only
Figure 3: Collisions of Na and Cs. The hold time in the merged trap is varied to measure the
evolution of trap occupancy due to various collision mechanisms. Post-selection on initial and
final trap occupancies allows us to distinguish 1- and 2-body processes. The fastest timescales
are indicated next to the thick fitted curves. The fits are explained in the supplementary ma-
terial. Left: For both atoms in a mixture of hyperfine states, the loss is dominated by rapid
2-body hyperfine-state-changing collision induced loss. Center : For both atoms in their lowest
hyperfine states, the loss is explained by 2-body hyperfine state changing collisions that follow
off-resonant scattering of trap light. In these two panels, different markers denote the final trap
occupancy. Right : One-body loss gives background gas limited lifetime of about 5 s for both
atoms. Here, we post-select on empty final tweezers and markers denote initial trap occupancy.
14 | a54ea132ff254451ee01a8d4a3674004c6e7baf3 | page_0013 |
4 6 8 10
Internuclear distance (Å)
-0.50.51.5
Energy (104cm-1)
A
PA
+
+
X1ΣB1Π
a3ΣA1Σc3Σ
b3Π
Probability of empty trap
PA detuning (GHz)Na Cs
v=-5 v=-6 v=-7 v=-8 v=-9
c3ΣB1Π1
10c3Σ
BFigure 4: Photoassociation Spectroscopy of NaCs. A) NaCs molecular potentials as a func-
tion of internuclear distance ( 27). Photoassociation (PA) light excites the ground state atoms
to vibrational levels of the NaCsexcited molecular potentials, from which they mostly de-
cay to vibrationally excited electronic ground state molecules (squiggly line). The long range
asymptotes of the excited state potentials (dominated by van der Waals interactions in the het-
eronuclear molecules) correspond to one of two cases: ground state Na colliding with excited
Cs in either the lower energy 6P1=2(D1 line) or higher energy 6P3=2state (D2 line). B) The
probability of single Na (orange), Cs (blue), and joint Na+Cs (red) atoms evolving to the “no
atoms” detection channel, as the PA light is detuned from the Cs D2 line dissociation threshold
at 351730 GHz. When both atoms are initially loaded into the tweezer (lower panel), 2-body
loss resonances appear due to molecule formation. As a validation of our method, we check that
no loss resonances are observed when only one atom is present (upper panel). The positions of
the loss resonances are fitted with the LB dispersion model in eq. 1 to identify three different
potentials and fit the respective C6dispersion coefficients. The expected resonance positions
based on these fits are marked by vertical lines as indicated in the legend. Except for at v=-7, the
RMS deviation of the fitted dispersion curve from the measured frequencies are 0.3, 0.6, and 0.8
GHz for the c31, c30, and B11states, respectively. At v=-7, a crossing of molecular energy
levels causes the measured spectrum to deviate from the prediction based on eq. 1. Unassigned
lines in the spectrum are likely due to rotational and hyperfine structure and pre-dissociating
potentials.
15 | a54ea132ff254451ee01a8d4a3674004c6e7baf3 | page_0014 |
MATERIALS AND METHODS
Experimental apparatus
All experiments are performed in an epoxy-bonded quartz vacuum chamber maintained at
<10 8Pa by a getter/ion pump. Six alkali dispensers for Cs and Na are suspended inside the
chamber, 5 inches from the MOT region. We maintain a constant current of 2A and 4 A through
two of them to maintain sufficient Cs and Na vapor pressures the experiments presented.
The 700 nm optical tweezer is derived from a cavity-locked Titanium sapphire laser. The
976 nm tweezer is derived from a free-running distributed Bragg reflector (DBR) laser.
The light for the Na MOT is derived from a frequency doubled Raman fiber amplifier that
is seeded by a 1178 nm external cavity diode laser (ECDL). The cooling and the repumping
frequencies are generated from the same laser by sending it through a 1.7 GHz acousto optical
modulator. Optical pumping for Na is provided by another 1178 nm ECDL that is frequency
doubled with a PPLN waveguide. Optical pumping on the D1 instead of D2 line is necessary to
avoid unwanted off-resonant cycling transitions in the Na D2 line due to the small excited-state
hyperfine splitting. All lasers for Na are locked via saturated absorption spectroscopy to a vapor
cell.
The light for the Cs MOT is derived from two 852 nm DBR lasers. The first is locked to
the Cs D2 line via saturated absorption spectroscopy, while the second is referenced to the first
with a phase lock, providing repumping and cooling frequencies respectively. Optical pumping
for Cs is provided by the same light.
At the vacuum chamber, 2 mW of Cs MOT and 5 mW of Na MOT light are expanded
to 6 mm beam diameter and combined before being directed into the chamber in a 6-beam
configuration. The gradient field for both MOTs, which are formed simultaneously, is 9 G/cm.
A 0.55 NA achromatic objective points at the chamber from between two of the horizontal MOT
beams. A custom dichroic separates resonant fluorescence from the far detuned tweezers for
S1 | a54ea132ff254451ee01a8d4a3674004c6e7baf3 | page_0015 |
both species simultaneously.
We image atoms onto an EMCCD with a magnification factor of 16. The total efficiency
from atomic fluorescence to photoelectron counts is about 4 %. To determine the presence of
atoms, we image them with resonant light. The signal is 104counts/s, at imaging times of 1
ms and 10 ms for Na and Cs respectively.
Na and Cs effective pair density
The “effective pair density” n2is defined as the probability of finding a single Na and Cs
atom per unit volume (eq. S1)
n2=1Z
11Z
11Z
1nCs(x;y;z )nNa(x;y;z )dxdydz (S1)
To get the individual atomic density distributions nNa(x;y;z )andnCs(x;y;z ), we assumed
the atoms occupy a thermal ensemble in a 3-dimensional harmonic oscillator potential with trap
frequencies (132, 123, 24) kHz for Na and (150, 140, 28) kHz for Cs, as measured by parametric
heating. The temperature during the collision measurements was measured by a release and
recapture technique ( 43) and found to be 90 K and 42K for Cs and Na, respectively, giving
n2= 21012cm 3. For the PA measurements the Cs temperature was 28 K, givingn2=
31012cm 3.
Na and Cs 1- and 2-body collisions
To obtain the fits in Fig. 3, we use the model depicted in Fig. S1. This yields the system
of differential equations eq. S2 for the time dependence of each tweezer occupation state. The
boundary conditions are the initial populations of each state (which can be read off directly
from the data) and the fact that all population should end up in (0,0) at long times.
Single atom images and post-selection allow us to isolate individual branches of Fig. S1.
The 1-body processes ((1,0) to (0,0) and (0,1) to (0,0)) feature only a single exponential decay
and are fitted first to obtain 1=kCs= 5:3(1) s and 1=kNa= 5:1(3) s, (Fig. 3, Right). These
S2 | a54ea132ff254451ee01a8d4a3674004c6e7baf3 | page_0016 |
rates are then fixed and the losses out of (1,1;L) are fitted to obtain 1=k2s= 0:63(1) s (Fig. 3,
Center). Finally, this rate is fixed as well and the losses out of (1,1;M) are fitted to obtain
1=k2f= 8(1) ms (Fig. 3, Left).
d
dt2
66664P00(t)
P01(t)
P10(t)
P11;L(t)
P11;M(t)3
77775=2
666640kNakCs k2s k2f
0 kNa 0 kCs kCs
0 0 kCskNa kNa
0 0 0 k2s kCs kNa 0
0 0 0 0 k2f kCs kNa3
777752
66664P00(t)
P01(t)
P10(t)
P11;L(t)
P11;M(t)3
77775
(S2)
For the measurements of 2-body collisions, the Cs and Na temperatures are measured to be
90K and 42K respectively, giving n2= 2:31012cm 3. This yields a loss rate constant
= 510 11cm3=s.
NaCs* Photoassociation spectroscopy
The NaCs photoassociation spectroscopy data presented in Fig. 4 are tabulated in Table S1.txt.
The data are organized in columns. The first column is the PA laser detuning in GHz. The next
columns are probability, followed by associated error bar, for the outcomes (Cs,Na) = (1,1) to
(0,0), (1,1) to (0,1), (1,1) to (1,0), (1,1) to (1,1), (0,1) to (0,0), and (1,0) to (0,0).
S3 | a54ea132ff254451ee01a8d4a3674004c6e7baf3 | page_0017 |
NaCs
kNa kCs
k2s
kCs kNa(1,1;L)
kNa kCs
k2f
kCs kNa(1,1;M)
(0,1)(0,1) (1,0) (1,0)
(0,0) (0,0)Figure S1: Model for 2-body collisions of Na and Cs. Four possible tweezer occupation states
exist: (1,1) both Cs and Na; (0,1) only Na; (1,0) only Cs; (0,0) empty. Transitions between
states are depicted by arrows with associated rates: 1-body Cs loss kCs, 1-body Na loss kNa,
slow 2-body loss k2s, fast 2-body loss k2f. Single atom images allow us to directly detect
transitions between any two of these states, thereby determining the rates k. (1,1) is further split
into two components: L, where both Na and Cs are in their lowest hyperfine states; and M, any
other combination of hyperfine states.
S4 | a54ea132ff254451ee01a8d4a3674004c6e7baf3 | page_0018 |
Relativistic and Electron
Correlation Effects in Moleeules
and Solids | 70a07c1ae882c92ddf7b508f57ecb89aee790c0d | page_0000 |
NATO ASI Series
Advanced Science Institutes Serles
Aseries presenting the results of activities sponsored by the NA TO Science
Committee, which aims at the dissemination of advanced sCientific and technological
knowledge, with a view to strengthening links between scientific communities.
The series is published by an international board of publishers in eonjunetion
with the NATO Seientific Affairs Division
A Ufe Seiences
B Physics
C Mathematical and Physical Seiences o Behavioral and Social Sciences
E Applied Sciences
F Computer and Systems Sciences
G Ecological Sciences
H Cell Biology
I Global Environmental Change
Recent Volumes in this Sertes Plenum Publishing Corporation
New Vork and London
Kluwer Aeademic Publishers
Dordrecht, Boston, and London
Springer~Verlag
Berlin, Heidelberg, New Vork, London,
Paris, Tokyo, Hong Kong, and Barcelona
Volume 317 -Solid State Lasers: New Developments and Applications
edited by Massimo Inguscio and Richard Wallenstein
Volume 318 -Relativistic and Electron Correlation Effeets in Moleeules and Solids
edited by G. L. Malli
Volume 319 -Staties and Dynamics of Alloy Phase Transformations
edited by Patrice E. A. Turchi and Antonios Gonis
Volume 320 -Singular Limits of Dispersive Waves
edited by N. M. Ercolani, I. R. Gabitov, C. D. Levermore, and
D.Serre
Volume 321 -Topics in Atomic and Nuelear Collisions
edited by B. Remaud, A. Calboreanu, and V. Zoran
Volume 322 -Techniques and Concepts of High Energy Physies VII
edited by Thomas Ferbel
Volume 323 -Soft Order in Physical Systems
edited by V. Rabin and R. Bruinsma
Volume 324 -On Three Levels: Micro-, Meso-, and Maero-Approaches in Physies
edited by Mark Fannes, Christian Maes, and Andre Verbeure
Series B: Physics | 70a07c1ae882c92ddf7b508f57ecb89aee790c0d | page_0001 |
Relativistic and Electron
Correlation Effects in Moleeules
and Solids
Edited by
G. L. Malli
Simon Fraser University
Burnaby, British Columbia, Canada
Springer Science+Business Media, LLC | 70a07c1ae882c92ddf7b508f57ecb89aee790c0d | page_0002 |
Proceedings of a NATO Advanced Study Institute on
Relativistic and Electron Correlation Effects in Molecules and Solids,
held August 10-21, 1992,
at the University of British Columbia, Vancouver, British Columbia, Canada
NATO-PCO-DATA BASE
The electronic index to the NATO ASI Series provides full bibliographical references (with
keywords and/or abstracts) to more than 30,000 contributions from international scientists
published in all sections of the NATO ASI Series. Access 10 the NATO-PCO-DATA BASE is
possible in two ways:
-via online FILE 128 (NATO-PCO-DATA BASE) hosted by ESRIN, Via Galileo Galilei,
1-00044 Frascati, Italy
-via CD-ROM "NATO Science and Technology Disk" with user-friendly retrieval software in
English, French, and German (©WTV GmbH and DATAWARE Technologies, Inc. 1989). The
CD-ROM also contains the AGARD Aerospace Database.
The CD-ROM can be ordered through any member of the Board of Publishers or through
NATO-PCO, Overijse, Belgium.
L1brary of Congress Cataloglng-in-Publlcat1on Data
Relativistic and electron correlation effects in moleeules and sol ids
! edited by G.L. Malli.
p. cm. -- (NATO ASI series. Series B. Physics ; v. 318)
"Published in cooperation with NATO Scientific Affairs Division,"
"Proceedings of a NATO Advanced Study Institute on Relativistic
and Electron Correlatlon Effects in Moleeules and Sol lds. held
August 10-21. 1992. at the University 01 British Columbla.
Vancouver. British Columbia. Canada"--T.p. verso.
Includes bibllographical references and index.
1. Electron configuration--Congresses. 2. Solid state physics-
-Congresses. 3. Molecules--Congresses. 4. Electronic structure-
-Congresses. 5. Wave functions--Congresses. I. Malli, G. L.
II. North Atlantic Treaty Organization. SCientific Affairs
Division. II!. NATO Advanced Study Institute on Relativistic and
Electron Correlation Effects in Moleeules and Solids (1992
VanCDuver. B.C.) IV. Series.
QCI76.8.E4R45 1993
530.4' 11--dc20
ISBN 978-1-4899-1342-5 ISBN 978-1-4899-1340-1 (eBook)
DOI 10.1007/978-1-4899-1340-1
©1994 Springer Science+Business Media New York
Originally published by Plenum Press, New York in 1994.
Softcover reprint 01' the hardcover 1 st edition 1994
All rights reserved 93-50141
CIP
No part of this book may be reproduced, stored in a retrieval system, or transmitted in any
form or by any means, electronic, mechanical, photocopying, microfilming, recording, or
otherwise, without written permission from the Publisher | 70a07c1ae882c92ddf7b508f57ecb89aee790c0d | page_0003 |
PREFACE
The NATO Advanced Study Institute (ASI) on "R@lativistic and Electron Correlation
Effects in Molecules and Solids", co-sponsored by Simon Fraser University (SFU) and the
Natural Sciences and Engineering Research Council of Canada (NSERC) was held Aug 10-
21, 1992 at the University of British Columbia (UBC), Vancouver, Canada. A total of 90
lecturers and students with backgrounds in Chemistry, Physics, Mathematics and various
interdisciplinary subjects attended the ASI.
In my proposal submitted to NATO for financial support for this ASI, I pointed out
that a NATO ASI on the effects of relativity in many-electron systems was held ten years
ago, [See G.L. Malli, (ed) Relativistic Effects in Atoms, Molecules and Solids, Plenum
Press, Vol B87, New York, 1983]. Moreover, at a NATO Advanced Research Workshop
(ARW) on advanced methods for molecular electronic structure "an assessment of state-of
the-art of Electron Correlation ... " was carried out [see C.E. Dykstra, (ed), Advanced
Theories and Computational Approaches to the Electronic Structure of Molecules, D.
Reidel Publishin~ Company, Vol C133, Dordrecht, The Netherlands 1984]. However,
during the last five years, it has become clear that the relativistic and electron correlation
effects must be included in the theoretical treatment of many-electron molecules and solids
of heavy elements (with Z > 70). Molecules and clusters containing heavy elements are
of crucial importance in a number of areas of Chemistry and Physics such as nuclear fuels,
catalysis, surface science, etc.
Since both the relativistic and electron correlation effects are expected to be very
pronounced, it is mandatory to treat both these effects accurately for systems of heavy
elements. Therefore, there is an urgent need for a concerted effort by leading scientists
working in both areas, to delineate and diffuse the state-of-the-art theoretica1 and
computational strategies so as to enable scientists to mount a joint attack on these
gargantuan problems. This ASI was, therefore, arranged to allow scientists from the
interdisciplinary areas of non-relativistic and relativistic quantum chemistry, molecular and
solid-state physics, etc., to assemble and discuss the various aspects of the effects of
relativity and electron correlation on the electronic structure, bonding, physical and
chemical properties of molecules and solids, especially those involving heavy elements.
My proposal to hold the ASI in Vancouver, Canada, received enthusiastic support
from all the colleagues who were approached to present lectures at the proposed ASI. It
was emphasized to all the invited lecturers that the main purpose of the ASI was to present
a systematic and coherently structured teaching program at the advanced level with the
object of disseminating currently available knowledge. The ASI would also help establish
contacts among scientists from various NATO countries working or planning to work in
areas of the ASI.
v | 70a07c1ae882c92ddf7b508f57ecb89aee790c0d | page_0004 |
This volume contains most of the invited lectures presented at the NATO ASI on
"Relativistic and Electron Correlation Effects in Molecules and Solids". It was decided not
to 'referee' the submitted manuscripts, and therefore the entire responsibility of the content
of the lectures rests with the authors.
I hope very much that the various lectures presented at this Advanced Study Institute
will be of as great interest to the reader as they were to the participants in the ASL It is,
however, impossible in this volume to convey the highly stimulating atmosphere of the
lectures, panel discussions, tutorials, informal discussions, formal get-togethers, etc., at the
ASI.
It was my pleasure to act as the Director of the ASI and I am most grateful to my
esteemed colleagues on the organising committee and the invited lecturers for their
enthusiastic support and valuable advice on various matters relating to the ASI. My special
heartfelt thanks go to the students without whose interest and participation the ASI would
not have been possible.
I sincerely thank Dr. L. V. da Cunha, Director, NATO Advanced Study Institute
Programme and his staff, especially Ms. Alison Trapp who helped me throughout the
various stages of the ASI. The generous funding of the ASI by NATO, SFU and NSERC
is gratefully acknowledged. IBM and CRA Y corporations are thanked for the financial
contributions to the ASI.
I am very grateful to Ms. Vanessa Mah for the invaluable help and assistance she
rendered as the ASI Secretary. My special personal thanks go to Ms. Sharon Beever on
the Departmental Secretarial Staff for her dependable and continuous help.
Finally, I cannot fully express my appreciation for the unfailing support and
understanding given to me by my wife Uma, and my daughters Sarada and Shivani, during
my Directorship of the Advanced Study Institute.
G.L. Malli
vi | 70a07c1ae882c92ddf7b508f57ecb89aee790c0d | page_0005 |
CONTENTS
Relativistic and Electron Correlation Effects in Molecules of Heavy Elements
G.L. Malli
Polyatomic Molecular Dirac-Hartree-Fock Calculations with Gaussian Basis Sets: 17
Theory, Implementation and Applications
K.G. Dyall
Molecular Electronic Structure Calculations based on the Dirac-Coulomb-(Breit) 59
Hamiltonian
W.C. Nieuwpoort, P.J.C. Aerts, and L. Visscher
Electronic Structure of Molecules, Clusters and Surfaces using Ab lnitio Relativistic 71
Effective Core and CorelValence Polarization Potentials
W.C. Ermler and M.M. Marino
Configuration Interaction Wave Functions 105
E.R. Davidson
Full Configuration Interaction and M011er-Plesset Theory 133
N.C. Handy
A Discussion of Some Aspects of the MCSCF Method 161
R. Shepard
Electron Correlation in Molecules Using Direct Second Order MCSCF 179
HJ.A. Jensen
Algebraic Approach to Coupled Cluster Theory 207
J. Paldus
Correlated and Non-Correlated Wave Functions for Organometallics 283
M.-M. Rohmer, M. Costas, R. Erenwein, J.-Y. Kempf, M.Ulmschneider,
P. de Vaal, T. Leininger, G.-H. Jeung, R. Wiest, and Marc Benard
Modem Tools for Including Electron Correlation in Electronic Structure Studies: 315
Hondo and Chem-Station
M. Dupuis, S. Chin, and A. Marquez
vii | 70a07c1ae882c92ddf7b508f57ecb89aee790c0d | page_0006 |
Multiconfigurational Perturbation Theory 339
R.B. Murphy and R.P. Messmer
Density Functional Theory, The Modem Treatment of Electron Correlations 367
E.K.U. Gross and S. Kurth
Density Functional Theory, Its Gaussian Implementation and Applications to 411
Complex Systems
D.R. Salahub, M. Castro and E.I. Proynov
An Introduction to GUGA in the Columbus Program System 447
R. Shepard
The Unitary Group Approach in Context 461
M. Schlesinger and R.D. Kent
Participants 471
Index 477
viii | 70a07c1ae882c92ddf7b508f57ecb89aee790c0d | page_0007 |
6 Springer Series in Chemical Physics
Edited by Fritz Peter Schafer "-------------' | 097e6afcaa9de7602f6ffdc106e4be6d59f77c7b | page_0000 |
Springer Series in Chemical Physics
Editors: V. I. Goldanskii R. Gomer F. P. Schafer 1. P. Toennies
Volume 1 Atomic Spectra and Radiative Transitions
By I. I. Sobelman
Volume 2 Surface Crystallography by LEED Theory, Computation
and Structural Results
By M. A. Van Hove, S. Y. Tong
Volume 3 Advances in Laser Chemistry
Editor: A. H. Zewail
Volume 4 Picosecond Phenomena
Editors: C. V. Shank, E. P. Ippen, S. L. Shapiro
Volume 5 Laser Spectroscopy Fundamentals and Techniques
By W. Demtroder
Volume 6 Laser-Induced Processes in Molecules Physics and Chemistry
Editors: K. L. Kompa, S. D. Smith
Volume 7 Excitation of Atoms and Broadening of Spectral Lines
By I. I. Sobelman, L. A. Vainshtein, E. A. Yukov | 097e6afcaa9de7602f6ffdc106e4be6d59f77c7b | page_0001 |
Laser-Induced Processes
in Molecules
Physics and Chemistry
Proceedings of the European Physical Society
Divisional Conference at Heriot -Watt University
Edinburgh, Scotland, September 20-22, 1978
Editors
K.L.Kompa and S.D. Smith
With 196 Figures
Springer-Verlag Berlin Heidelberg New York 1979 | 097e6afcaa9de7602f6ffdc106e4be6d59f77c7b | page_0002 |
Professor Dr. Karl Ludwig Kompa
Projektgruppeflir Laserforschung, Max-Planck-GeseIlschaft zur Forderung der Wissenschaften e.V.
D-8046 Garching bei Miinchen, Fed. Rep. of Germany
Professor Stanley Desmond Smith, Ph.D. D.Sc. FRS
Department of Physics, Heriot-Watt University, Riccarton,
Edinburgh EH14 4AS, Scotland
Series Editors:
Professor Vitalii I. Goldanskii
Institute of Chemical Physics, Academy
of Sciences, Vorobyevskoye Chaussee 2-b
Moscow V-334, USSR
Professor Robert Gomer
The James Franck Institute
The University of Chicago, 5640 Ellis A venue
Chicago, IL 60637, USA Professor Dr. Fritz Peter Schafer
Max-Planck-Institut flir Biophysikalische
Chemie, D-3400 Gottingen-Nikolausberg
Fed. Rep. of Germany
Professor Dr. J. Peter Toennies
Max-Planck-Institut flir Stromungsforschung
BottingerstraBe 6-8
D-3400 Gottingen, Fed. Rep. of Germany
Conference Chairmen:
Conference Organizer: Professor S. D. Smith FRS and Professor K. L. Kompa
Dr. R.G. Harrison, Heriot-Watt University
Conference Treasurer: H.A. MacKenzie, Heriot-Watt University
Sponsors: Imperial Chemical Industries Ltd.; Spectra Physics Inc.;
Edinburgh Instruments Ltd.
Programme and Advisory Committee: M.J. Berry, B. Bolger, B. Burlamacchi, J. Ernest,
S. Kimel, K.L. Kompa, S. Leach, V.S. Letokhov, S.D. Smith,
I.J. Spalding, O. Svelto, J.J. Turner, H. Walther, R.N. Zare
ISBN-I3: 978-3-642-67256-9 e-ISBN-I3: 978-3-642-67254-5
DOl: 10.1007/978-3-642-67254-5
This work is subject to copyright. All rights are reserved, whether the whole or part of the material is
concerned, specifically those of translation, reprinting, reuse of illustrations, broadcasting, reproduction by
photocopying machine or similar means, and storage in data banks. Under § 54 of the German Copyright
Law where copies are made for other than private use, a fee is payable to the publisher, the amount of the
fee to be determined by agreement with the publisher.
© by Springer-Verlag Berlin Heidelberg 1979
Softcover reprint of the harcover I st edition 1979
The use of registered names, trademarks, etc. in this publication does not imply, even in the absence of a
specific statement, that such names are exempt from the relevant protective laws and regulations and
therefore free for general use. | 097e6afcaa9de7602f6ffdc106e4be6d59f77c7b | page_0003 |
Preface
This conference on both the physics and chemistry of laser-induced processes
in molecules was organized by the Quantum Electronics Divisional Board of
the European Physical Society whose membership is given on p.367. The confer
ence aim, to mix physicists and chemists interested in this exciting field
both from Europe and further afield, was well fulfilled by the attendance
of around 250 participants and the submission of about 100 papers, which dre
presented here.
Numerous people at both the Physics Department, Heriot-Watt University,
Edinburgh, and at the Projektgruppe fUr Laserforschung, MPI, Garching, con
tributed hard work to the organization; in addition to Dr. Bob Harrison,
who bore the biggest burden with conspicuous success, we particularly thank
Hugh MacKenzie, Richard Dennis and last but not least Miss Joanne Askham and
the secretaries in Edinburgh together with Frau Doris Maischberger and the
secretaries in Garching.
December 1978 K.L. Kompa S.D. Smith | 097e6afcaa9de7602f6ffdc106e4be6d59f77c7b | page_0004 |
Conren~
Part I. Study of Lasers and Related Techniques Suitable for Applications
in Chemistry and Spectroscopy
Rare Gas Halogen Lasers and Photochemical Applications.
By S.D. Rockwood................................................... 3
Group VI Molecular Photolytic Dissociation Studies Using Rare Gas
Halide Lasers. By M.C. Gower, A.J. Kearsley, and C.E. Webb ......... 8
Broadly Tunable UV Source Based on Stimulated Raman Scattering.
By V. Wi 1 ke and W. Schmi dt ......................................... 12
Laser-Induced Intermolecular and Intramolecular Energy Transfer
Processes. By S. Spei ser ........................................... 15
A New Method for Exposing Mammalian Cells to Intense Laser Radiation
Using the Evanescent Fields Created in Optical Waveguides.
By H.L. Cox, Jr. .................................................. 19
Time-Dependent Kinetic-Thermodynamic Description of a Nonequilibrium
Molecular System: The HF Chemical Laser. By A. Ben-Shaul, O. Kafri,
and R.D. Levine.................................................... 22
Low-Temperature Performance of Roto-Vibrational Molecular Lasers.
By J.M. Green...................................................... 26
HBr Laser Emission at 4 ~m Based on the Chemical Generation of
Bromine Atoms. By S.J. Arnold and K.D. Foster " .................... 29
Tunable Infrared Generation from Four-Wave Mixing in Room-Temperature
Germanium. By S.R. Butcher, R.G. Harrison, and R.A. Wood ........... 32
Part II. Spectroscopic Studies with and Related to Lasers
High-Resolution Double-Resonance Spectroscopy of SF6.
By P.F. Moulton and A. Mooradian ................................... 37
VII | 097e6afcaa9de7602f6ffdc106e4be6d59f77c7b | page_0005 |
High-Resolution Coherent Raman Spectroscopy of Methane and Intensity
of CARS vs Pressure in Gases. By J. -P. Boqui 11 on ................... 43
Molecular Spectroscopy of NO and C2H4 Using Step-Tunable IR Lasers
and the External Resonator Controlled Spin-Flip Raman Laser.
By G. Crowder, R.B. Dennis, H.A. MacKenzie, M.H. Mozolowski,
and M. Sioudi ...................................................... 46
Photophysics of Gaseous UF6 in the UV: Energy Ba'lance Through
Quantum Efficiency Measurements. By O. de Witte, R. Dumanchin,
J.P. Gauyacq, and M. Michon ........................................ 49
Optical Dephasing in Molecular Iodine. By R.G. Brewer and
S.S. Kano .......................................................... 54
An Efficient Algorithm for the Study of Nonlinear Resonant
Propagation of Two Concomitant Optical Pulses Interacting with
a Three-Level Atomic System. By F.P. Mattar and J.H. Eberly ........ 61
The Influence of Hyperfine Coherence and of Elastic Collisions on
the Circular Polarisation of Emission from Li2. By M.D. Rowe and
A.J. McCaffery..................................................... 66
Collisional Depolarization and Rotational Energy Transfer of the
7Li2B1rru -Li2S1/2 System Using Laser-Induced Fluorescene.
By C.R. Vidal ...................................................... 69
Saturated Absorption Experiments on a Molecule Dressed by a Radio-
frequency Fi e 1 d. By E. Ari mondo and P. Glori eux .................... 71
dc Magnetic Field Effects on Polarized Monochromatic Laser Radiation
Absorption in Electronic Transitions: A Simple Model.
By M. Bernardini, A. Fubini, and G. Sacerdoti ...................... 74
Effect of Two-Photon Saturation on Ordinary Raman Scattering.
By A.D. Wilson-Gordon and H. Friedmann 78
Predissociation in the A'A" State of HNO.
By R.N. Dixon and M. Noble......................................... 81
Photoassociation of Heavy Metal Excimers: Spectroscopic,
Kinetic, and Laser Applications. By D.J. Ehrlich and
R .M. Osgood, Jr. ................................................... 85
Dye-Laser-Induced Emission from the KXe Molecule. By C.R. Webster and
F. Rostas .......................................................... 89
VIII | 097e6afcaa9de7602f6ffdc106e4be6d59f77c7b | page_0006 |
Laser-Induced Emission from the NaK Molecule in a Supersonic Beam.
By E.-J. Breford and F. Engelke .................................... 92
The Singlet-Triplet Energy Separation of CH2 and Spectroscopy of the
alAI State. By D. Feldmann, K. Meier, R. Schmiedl, H. Zacharias,
and K.H. Welge..................................................... 95
Time-Resolved Fluorescence Spectroscopy Using Pulsed Laser Excitation.
By K.P. Ghiggino, A.J. Roberts, and D. Phillips .................... 98
Ultrafast Spectroscopy with a Streak Camera: Excited State Spectros-
copy and Kinetics of Coumarin Derivatives. By J. Schulz-Hennig
and A. MUller ...................................................... 101
Laser-Induced Processes in Phthalocyanines. By J. McVie, R.S. Sinclair,
and T. G. Truscott .................................................. 104
Fluorescence Decay-Time Measurement of Rhodamine 6G and Rhodamine B
in Different Solutions. By H. Alobaidi, F. Alberkdar, Z. Hafidh,
and S. Alalkawy .................................................... 108
Laser Flash Spectroscopy of 4-Nitrostilbenes and Thioindigo Dyes;
Configuration of the Triplet State in Solution. By H. Garner and
D. Schulte-Frohl inde ............................................... 111
Part III. Multiphoton Excitation, Dissociation and Ionization
Laser-Induced Unimolecular Reactions. By N. Bloembergen and
E. Yablonovitch .................................................... 117
Time and Intensity Dependence of the Infrared Absorption of SF6:
Measurements with an Injection-Locked Single Mode TEA CO2 Laser.
By S.D. Smith, W.E. Schmid, F.M.G. Tablas, and K.L. Kompa .......... 121
IR Absorption of Highly Excited SF6.
By W. Fui3 and J. Hartmann .......................................... 128
Absorption of Intense Laser Radiation by SF6 and S2F10'
By J.L. Lyman ...................................................... 131
Collisional Effects in the Multiple-Photon Infrared Laser Pumping of
Polyatomic Molecules. By G.P. Quigley .............................. 134
Coherent Pulse Propagation Effects in Multilevel Molecular Systems.
By C.D. Cantrell, W.H. Louisell, and J.F. Lam ...................... 138
IX | 097e6afcaa9de7602f6ffdc106e4be6d59f77c7b | page_0007 |
The Influence of Intensity and Laser-Energy Fluence in Unimolecular
Reactions Induced by Monochromatic Infrared Radiation (URIMIR).
By M. Quack ........................................................ 142
On the Dynamics of Multiphoton Dissociation of Polyatomic
Molecules II. Application to 03. By K.D. Hansel .................... 145
Nonthermal Theory of Threshold Behaviour of Collisionless
Multiphoton Dissociation. By H. Friedmann .......................... 149
An Anharmonic Model for Molecular Photo-Excitation.
By P.G. Harper ..................................................... 152
Dynamic Stark Splitting of Multiphoton Absorption Resonances.
By P.L. Knight ..................................................... 155
Polarization Selection Rules for Two-Photon Processes.
By B.R. Marx and L. Allen
Doppler-Free Raman Spectroscopy and Suppression of Molecular
Doppler-Broadened Transitions Induced by Laser Field. 158
By A.K. Popov and L.N. Talashkevich ................................ 161
Laser-Controlled Unimolecular and Bimolecular Processes:
Field-Dependent Rate "Constants". By A.M.F. Lau .................... 163
Laser-Induced Predissociation of Diatomic and Polyatomic Molecules
by the Photo-Cata lyti c Effect. By A. M. F. Lau ....................... 167
Possibility of Hole Burning in Single Quantum Power Spectrum Due to
Autler-Townes Splitting. By J.V. Moloney and F.H.M. Faisal ......... 171
Computation of Rovibrational Multiphoton Spectra: Application to CO.
By J.V. Moloney and F.H.M. Faisal .................................. 173
Crossed Laser and Molecular Beam Study of Multiphoton Dissociation
of C2F5Cl. By D.J. Krajnovich, A. Giardini-Guidoni, Aa.S. Sudb~,
P.A. Schulz, Y.R. Shen, and Y.T. Lee ............................... 176
Study of Primary Characteristics of Multiple IR Photon Excitation
and Dissociation of CF3I. By V.N. Bagratashvili, V.S. Doljikov,
V.S. Letokhov, and E.A. Ryabov ..................................... 179
Photofragment Spectroscopy of Multiple Photon Dissociation by Laser
Induced Fluorescence. By R. Schmiedl, R. Bottner, H. Zacharias,
U. Meier, D. Feldmann, and K.H. Welge .............................. 186
x | 097e6afcaa9de7602f6ffdc106e4be6d59f77c7b | page_0008 |
Laser-Excited Fluorescence Studies of IR Multiple Photon Dissociation
Fragments: NH2(2B1) from CH3NH2 and NH3. By G. Hancock,
R.J. Hennessy, and T. Villis ....................................... 190
Time Behavior of CN Radicals in Laser-Initiated Thermal
Isomeriza.tion of CH3NC. By D.S. Bethune, J.R. Lankard, M.M.T. Loy,
J. Ors, and P.P. Sorokin ........................................... 193
Laser-Induced Decomposition of Methanol: A Comparative Study Using
Pulsed HF and CO2 Lasers. By R. Bhatnagar, P.E. Dyer,
and G.A. Oldershaw ................................................. 195
IR Laser Sensitised Chemical Reactions of Polyatomic Molecules.
By C. Steel, V. Starov, P. John, R.G. Harrison, and R. Leo ......... 198
Excitation of Ammonia in the Megawatt Region Using a CO2 Laser.
By V. Starov, C. Steel, S. Butcher, R.G. Harrison, P. John,
and R. Leo ......................................................... 201
Reaction Between CH3N02 and H2 Induced by CO2 Laser Pulses.
By M. Neve de Mevergni es and P. Fettwei s ........................... 205
Multiphoton Dissociation in Formaldehyde.
By G. Koren and U. P. Oppenhei m ..................................... 209
Multiphoton Vacuum UV Photodissociation of Simple Polyatomic
Molecules. By A.P. Baronavski, V.M. Donnelly, and J.R. McDonald .... 213
Mass Selective Two-Photon Ionization of a Polyatomic Molecule.
By U. Boesl, H.J. Neusser, and E.W. Schlag ......................... 219
Part IV. Laser Control of Chemical Reactions
Laser-Induced Chemical Processes: Reactions with Oriented Reagents.
By R.N. Zare ....................................................... 225
Laser-Induced Fluorescence Study of the Reactions of F Atoms with
CH3I, CF3I and IC1. By L. Stein, J. Wanner, H. Figger,
and H. Walther ..................................................... 232
A Laser-Induced Fluorescence Study of the Reaction
Ca + CC14-+ CaCl + CC13. By A. Schultz and W. Schmidt .............. 236
cw Laser-Induced Fluorescence Study of BA + S02: Vibronic Distribution
of BaO as a Function of Collisional Energy. By R. Dirscherl
and H.U. Lee ....................................................... 239
XI | 097e6afcaa9de7602f6ffdc106e4be6d59f77c7b | page_0009 |
Laser Studies of the Relaxation and Reaction of Species in Defined
Quantum States. By I.W.M. Smith .................................... 242
+ + Calculated Infrared Photodissociation Cross Section of H2 and HD
By P. Fournier, B. Lassier-Govers, and G. Comtet 247
Pulsed Molecular Beam Study of Ethylene Dimer Photodissociation with
a CO2 Laser. By M.A. Hoffbauer, W.R. Gentry, and C.F. Giese ........ 252
Photodissociation Lifetimes of Van der Waals Complexes.
By J.A. Beswick and J. Jortner ..................................... 255
Towards Resonant Single Photon Dissociation of SF6.
By J.P. Astruc, R. Barb~, and J.P. Schermann ....................... 258
Photon-Enhanced Dissociative Electron Attachment in SF6.
By C.L. Chen and P.J. Chantry ...................................... 261
Laser Specific Versus Thermal Reactions.
By S. Kimel ........................................................ 265
Deuterium Enrichment by cw Vibrational Photochemistry of Methane-
Economic Considerations. By T.J. Manuccia and D.S.Y. Hsu ........... 270
Laser-Induced Chemical Reaction of BC13 with CH4.
By B. Schramm .................................. , ................... 274
Study of the cw Laser-Induced Reaction of CH3CF2Cl.
By R.N. Zitter, D.F. Koster, A. Cantoni, and J. Pleil .............. 277
IR Photochemistry in an Electronically Excited State.
By H. Stafast, J. Opitz, and J.R. Huber ............................ 280
Single-Photon Infrared Photochemistry: Wavelength and Temperature
Dependence of the Quantum Yield for the Laser-Induced Ionization
of Water. By D.M. Goodall, R.C. Greenhow, B. Knight,
J.F. Holzwarth, and W. Frisch
A Laser Study of the Cage Effect in High-Pressure Gases:
Iodine and Bromine Photodissociation. By H. Hippler, K. Luther, 283
M. Maier, J. Schroeder, and J. Troe ................................ 286
Isotope Selective Molecular Spectroscopy and Production of
Isotopically Pure Molecules with a Dye Laser. By U. Boesl,
H.J. Neusser, and E.W. Schlag ...................................... 290
XII | 097e6afcaa9de7602f6ffdc106e4be6d59f77c7b | page_0010 |
Selective Photoaddition of Iodine Chloride to Acetylene: Pressure,
Buffer, and Wavelength Dependence of Enrichment.
By M. Stuke and E.E. Marinero ...................................... 294
Laser-Induced Photolysis of Ozone Reactions of O(3P), O(lD) and
1 02( t::.). By I. Arnold and F.J. Comes ................................ 298
Time-Resolved Studies of Reactions of I (2P1/2).
By K.-H. Stephan and F.J. Comes .................................... 301
Infrared Laser-Induced Photochemistry in the Solid State.
By M. Poliakoff .................................................... 304
Part V. Molecular Relaxation
Ultrafast Vibrational Relaxation of Polyatomic Molecules.
By W. Kaiser and A. Laubereau
Generation of Time-Correlated Picosecond Laser Pulses and Their
Application for Rapid Sampling of Optical Relaxation Phenomena. 313
By S. Schneider, E. Lill, P. Hefferle, and F. Dorr ................. 321
Electronic to Vibrational Energy Transfer and Infrared Lasers.
By R.J. Donovan .................................................... 324
Inelastic Collisions in a Potentially Reactive System: Rotational
Energy Transfer in the A 2A1 Excited State of NH2 Induced by
Collisions with H Atoms. By R.N. Dixon and D. Field ................ 329
Vibrational Relaxation of Ethylene Excited with a Parametric
Oscillator. By J. Hager, W. Krieger, T. RUegg, and H. Walther ...... 333
Vibrational Relaxation of HF (v = 3, 4) by HF, H2, D2, CO2, and
Isobutene. By D.J. Douglas and C.B. Moore .......................... 336
Study of Energy Transfer in Methane by Excitation of Fundamental,
Overtone, and Combination Bands. By P. Hess and C.B. Moore ......... 339
Line Selective Excitation of Ethylene with CO2 Laser Light and
Vibrational Relaxation. By X. de Hemptinne and D. De Keuster 342
Vibrational Energy Transfer at Low Temperatures: CD3F in Rare Gas
and Nitrogen Matrices. By L. Abouaf-Marguin, B. Gauthier-Roy,
and F. Legay ....................................................... 345
Laser-Induced Vibrational Fluorescence in Matrix Isolated Molecules.
By H. Dubost ....................................................... 348
XIII | 097e6afcaa9de7602f6ffdc106e4be6d59f77c7b | page_0011 |
Quenching of N02 (2Bl) and (2B2).
By D. Haaks and U. Schurath ........................................ 352
Quenching of C2(a3Hu) Produced in an Intense Infrared Laser Field.
By S.V. Filseth, G. Hancock, J. Fournier, and K. Meier ............. 356
Time-Resolved Geometrical Optics in Molecular Pumping Experiments.
By R.T. Bailey, F.R. Cruickshank, D. Pugh, and W. Johnstone ........ 359
Energy Transfer in Biologically Interesting Molecules.
By A. Anders ....................................................... 362
Index of Contri butors ................................................. 365
The Quantum Electronics Division of the
European Physical Society -Information ............................ 367
XIV | 097e6afcaa9de7602f6ffdc106e4be6d59f77c7b | page_0012 |
Data in Brief 42 (2022) 108280
Contents lists available at ScienceDirect
Data in Brief
journal homepage: www.elsevier.com/locate/dib
Data Article
Developing a dataset for the expected
anthropogenic mercury release in China in
response to the Minamata convention on
mercury
Habuer a , ∗, Takeshi Fujiwara a , Masaki Takaoka b
a Graduate School of Environmental and Life Science, Okayama University, 3-1-1 Tsushima Naka Kita-Ku, Okayama
700-8530, Japan
b Graduate School of Engineering, Kyoto University, C-1-3 Nishikyo-ku, Kyoto 615-8540, Japan
a r t i c l e i n f o
Article history:
Received 6 February 2022
Revised 5 May 2022
Accepted 9 May 2022
Available online 15 May 2022
Keywords:
Anthropogenic activity
Mercury release
Minamata convention on mercury
Technology transformation a b s t r a c t
This paper contains supplementary data in support of a re-
search paper published [1] regarding the expected anthro-
pogenic mercury release in China in response to the Mina-
mata Convention on Mercury (MCM). The dataset provided
within this article contains a set of excel spreadsheets. Each
spreadsheet contains filtered (collected) and analysed data,
i.e., parameters, collected data, calculated and summarized
results for mercury distribution by the category of mineral
production, intentional uses, secondary metal production, ex-
traction and combustion, and waste treatment in a specific
year. The collected (filtered) data in this article consist of
the input factor (IF), activity rate data (ARD), output scenario
(OS), initial distribution factor (iDF), and redistribution factor
(rDF). IF was from the default IF in the United Nations En-
vironment Programme (UNEP) Toolkit Level 2 and published
scientific papers. ARD was obtained from the U.S. Geolog-
ical Survey database, China Statistical Yearbooks, and pub-
lished scientific papers. The OS content was from the de-
fault OS in the UNEP Toolkit Level 2 and published scien-
tific papers. iDF was from the default distribution factor (DF)
in the UNEP Toolkit Level 2 and published scientific papers.
DOI of original article: 10.1016/j.jclepro.2021.129089
∗Corresponding author.
E-mail address: habuer@okayama-u.ac.jp .
https://doi.org/10.1016/j.dib.2022.108280
2352-3409/© 2022 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND
license ( http://creativecommons.org/licenses/by-nc-nd/4.0/ ) | 229a5963aa3649b5a9ad714c5764b07ddd8918d5 | page_0000 |
2 Habuer, T. Fujiwara and M. Takaoka / Data in Brief 42 (2022) 108280
rDF was from published scientific paper. The mercury input
was calculated using IF and ARD. The mercury release to dif-
ferent media in the initial distribution step was calculated
using the mercury input and iDF. The release of mercury to
the final sinks in the redistribution step was calculated us-
ing the amount of sector-specific treatment/disposal, product
or by-product, and rDF. The dataset with combination of the
collected (filtered) and analyzed data can contribute to an
understanding of differences in anthropogenic mercury re-
lease before and after implementation of the MCM, especially
considering technology transformation in China. Government
policymakers involved in hazardous waste management, es-
pecially those working on MCM, and engineers and scientists
interested in hazardous waste management may benefit from
these data. The data can be used for identifying the environ-
mental impact of anthropogenic mercury release before and
after the MCM in China. The data can facilitate the creation
of strategic management policies for mercury as the MCM is
implemented in China.
©2 0 2 2 The Authors. Published by Elsevier Inc.
This is an open access article under the CC BY-NC-ND
license ( http://creativecommons.org/licenses/by-nc-nd/4.0/ )
Specifications Table
Subject Environmental Engineering
Specific subject area Waste Management and Disposal
Type of data Table
How the data were acquired Chinese statistical data published by the National Bureau of Statistics of the
People’s Republic of China, and electronic yearbooks in both Chinese and
English can be accessed freely on the Bureau of Statistics website. Mineral
Yearbooks by the National Minerals Information Center in the United States
can be accessed freely on the U.S. Geological Survey (USGS) website. Toolkits
for identifying and quantifying mercury release, reference reports and revised
inventory level 2 reports are provided by the United Nations Environment
Programme (UNEP) Chemicals, and can be downloaded freely. Other secondary
data can be acquired from published scientific papers.
Data format Collected
Filtered
Analyzed
Description of data collection The collected and filtered data in this article consist of input factors (IF),
activity rate data (ARD), output scenarios (OS), initial distribution factors (iDF),
and redistribution factors (rDF). IF was obtained from the default IF in the
UNEP Toolkit Level 2 and published scientific papers. ARD was from the USGS
database, China Statistical Yearbooks, and published scientific papers. The
types of OS were from the default OS in UNEP Toolkit Level 2 and published
scientific papers. iDF was from the default DF in UNEP Toolkit Level 2, and
published scientific papers. The rDF was from published scientific paper.
Data source location USGS database: USGS, 2016-2020. Minerals Yearbooks, National Minerals
Information Center.
https://www.usgs.gov/centers/nmic/minerals- yearbook- metals- and- minerals .
Minerals Yearbooks of the National Minerals Information Center (U.S.
Geological Survey), U.S.
China Statistical Yearbook: NBSC, 2016-2020. China Statistical Yearbook
2016–2020 (in both Chinese and English).
http://www.stats.gov.cn/tjsj/ndsj/ . National Bureau of Statistics of the People’s
Republic of China, Beijing, China.
( continued on next page ) | 229a5963aa3649b5a9ad714c5764b07ddd8918d5 | page_0001 |
Habuer, T. Fujiwara and M. Takaoka / Data in Brief 42 (2022) 108280 3
Toolkit Level 2: United Nations Environment Programme (UNEP) Chemicals,
2017. Toolkit for Identification and Quantification of Mercury Releases,
Reference Report and Revised Inventory Level 2.
https://wedocs.unep.org/handle/20.500.11822/30684 . Report by Chemicals,
UNEP, Switzerland.
Published scientific papers: [4–8 , 11]
Data accessibility Habuer; Fujiwara, Takeshi; Takaoka, Masaki (2022), “Dataset for the expected
anthropogenic mercury release in China between 2016 to 2019”, Mendeley
Data, V2
doi: 10.17632/tjzm2gdntf.2
Related research article Habuer, T. Fujiwara, M. Takaoka, The response of anthropogenic mercury
release in China to the Minamata Convention on Mercury: a hypothetical
expectation, J. Clean. Prod. 323 (2021) 129089.
doi: 10.1016/j.jclepro.2021.129089
Value of the Data
• Standardized data collection (filtering) and accounting method are important for precise
identifying a time-series anthropogenic mercury release. The dataset with combination of
the collected (filtered) and analyzed data can contribute to an understanding of differences
in anthropogenic mercury release before and after implementation of the Minamata Conven-
tion on Mercury (MCM), especially considering technology transformation in China.
• The data provided can contribute to reduce duplication of effort for relevant data collection.
• Government policymakers involved in hazardous waste management, especially those work-
ing on the MCM, and engineers and scientists interested in hazardous waste management
may benefit from these data.
• The data can be used for comparing the environmental impact of anthropogenic mercury
release before and after implementation of the MCM in China.
• The data can facilitate the creation of strategic management policies for mercury as the MCM
is implemented in China.
1. Data Description
The data described in this section could be found in Ref. [2] . The data comprises six spread-
sheets: “Data for 2016”, “Data for 2017”, “Data for 2018_BAU”, “Data for 2018_ACR”, “Data for
2019_BAU”, and “Data for 2019_ACR”. Each spreadsheet contains the following worksheets:
• Worksheet 1 (title “Intro”): a title page with references.
• Worksheet 2 (title “Dataset caption”): the dataset numbers and captions.
• Worksheet 3 (title “Dataset S1”): a summary of mercury releases in the initial distribution
(iD) step. Category C1 has two subcategories, of which subcategory C1.1 contains eight sub-
categories. Category C2 has two subcategories, of which subcategories C2.1 and C2.2 have
two and five subcategories, respectively. Category C4 has four subcategories: C4.1 four sub-
categories; C4.2 five subcategories; and C4.4 two subcategories. Category C5 has two subcat-
egories each of which (C5.1 and C5.2) has two subcategories. The name of each source cate-
gory is provided. The calculation results are given as the release of mercury (R Hg ) from each
category and R Hg to different media, such as “air”, “water”, “land”, “stock”, “general waste”,
and “sector-specific treatment/ disposal”.
• Worksheet 4 (title “Dataset S2”): a summary of all mercury releases from the five categories
in the initial and redistribution (rD) steps. According to Habuer et al. [1] , the final sinks
include (1) air, (2) water, (3) land, (4) stock, and (5) stabilization, and intermediate reservoirs
include “general waste” and “sector-specific treatment/ disposal”. The recovered amount and
total mercury released to the natural environment by category in the iD and rD stages are
also provided. | 229a5963aa3649b5a9ad714c5764b07ddd8918d5 | page_0002 |
4 Habuer, T. Fujiwara and M. Takaoka / Data in Brief 42 (2022) 108280
• Worksheet 5 (title “Dataset C1 (Mineral production)”): this contains parameters, collected
and filtered data, and the results of the mercury distribution by mineral production cate-
gory. This category includes two subcategories: virgin metal and mineral production (C1.1)
and cement production (C1.2). Subcategory C1.1 has eight subcategories: “mercury (primary)
extraction and initial processing”, “gold (and silver) extraction with mercury amalgamation
processes (GEMA)/from whole ore”, “zinc extraction and initial processing/production of zinc
from concentrates”, “copper extraction and initial processing/production of copper from con-
centrates”, “lead extraction and initial processing/production of lead from concentrates”, “in-
dustrial gold smelting”, “aluminum extraction and initial processing/alumina production from
bauxite”, and “primary ferrous metal production”. The default input factors in the United Na-
tions Environment Programme (UNEP) Toolkit Level 2 [3] have been provided, and the input
factors (IF), distribution factors (DF), and output scenarios (OS) applied for the calculation; if
there were no remarks, the default values from the UNEP Toolkit Level 2 were used. The ac-
tivity rate data (ARD) data and calculation results for amounts input and released to different
media have also been provided.
• Worksheet 6 (title “Dataset C2 (Intentional uses)”): this contains the parameters, collected
and filtered data, and results of mercury distribution by intentional use category, which has
two subcategories: uses in industrial processes (C2.1) and in consumer products (C2.2). Cate-
gory C2.1 contains two subcategories: “chlor-alkali production with mercury-technology” and
“vinyl chloride monomer (VCM) production with mercury catalyst”. Category C2.2 contains
five subcategories: “thermometers /production”, “electrical switches and relays/production”,
“light sources/production”, “batteries/ production”, and “dental mercury-amalgam fillings
/preparation of fillings at dental clinics”. When applying IF, DF, and OS to the calculation,
if there were no remarks, the default values from the UNEP Toolkit Level 2 were used. The
ARD data and the calculation results for amounts input and released to different media have
also been provided.
• Worksheet 7 (title “Dataset C3 (Secondary metal.)”): a database of parameters, collected and
filtered data, and results of mercury distribution by secondary metal production category.
The input amount was from a published scientific paper [4] , and DF was from the default
value in UNEP Toolkit Level 2.
• Worksheet 8 (title “Dataset C4 (Extraction.)”): a database of parameters, collected and fil-
tered data and results of mercury distribution by the extraction and combustion category.
This category includes four subcategories: coal combustion and use (C4.1), mineral oil ex-
traction, refining and use (C4.2), natural gas extraction, refining, and use (C4.3), and biomass
combustion (C4.4). Category C4.1 has four subcategories: “coal combustion in power plants”,
“coal combustion in coal fired industrial boilers”, “coke production”, and “residential coal
use”. Category C4.2 has five subcategories: “extraction and uses”, “oil combustion facilities”,
“transportation and other uses than residential heating and other oil combustion facilities”,
“residential heating”, and “other oil combustion facilities”. Category C4.4 has two subcate-
gories: “use of biomass” and “charcoal combustion”. When applying IF, DF, and OS to the
calculation, if there were no remarks, the default values from the UNEP Toolkit Level 2 were
used. The ARD data and the calculation results for amounts input and released to different
media have also been provided.
• Worksheet 9 (title “Dataset C5 (Waste treatment)”): a dataset of parameters, collected and
filtered data, and results of mercury distribution for the waste treatment category. This has
two subcategories: waste incineration (C5.1) and municipal sewage and informal landfilling
(C5.2). Category C5.1 has two subcategories: “incineration of municipal/general waste” and
“incineration of hazardous waste”. Category C5.2 has two subcategories: “informal dumping
of general waste” and “municipal sewage system/treatment”. When applying IF, DF, and OS
to the calculation, if there were no remarks, the default values from the UNEP Toolkit Level 2
were used. The ARD data and calculation results for amounts input and released to different
media have also been provided. | 229a5963aa3649b5a9ad714c5764b07ddd8918d5 | page_0003 |
Habuer, T. Fujiwara and M. Takaoka / Data in Brief 42 (2022) 108280 5
• Worksheet 10 (title “Dataset R1”): a dataset for mercury redistribution to different media.
The wastes, products, and by-products from five categories are provided. The redistribution
factors (rDF) were from a published scientific paper [5] .
• Worksheet 11 (title “Dataset U1” in spreadsheet named “Data for 2016”): a dataset for an
output of tornado analysis. The most sensitive input variables are listed. The downside and
upside values of the input and total input are also provided.
2. Experimental Design, Materials and Methods
2.1. Definition of source categories and final sinks
The five source categories and 33 leading subcategories in the initial distribution step and
33 subcategories in the redistribution step were defined according to the UNEP Toolkit Level
2 [3] and published scientific papers [4 , 5] . The final sinks were (1) air, (2) water, (3) land, (4)
stock, and (5) stabilization and intermediate reservoirs “general waste” and “sector-specific treat-
ment/disposal” were defined according to UNEP Toolkit Level 2 [3] and published scientific pa-
pers [5 , 6] .
2.2. Data collection
The collected and filtered data in this article consist of IF, ARD, type of OS, iDF, and rDF. IF
was the default input factor in the UNEP Toolkit Level 2 and published scientific papers [5 , 7 , 8] .
ARD was from the U.S. Geological Survey database [9] , China Statistical Yearbooks [10] , and pub-
lished scientific papers [4 , 5 , 11] . The types of OS were from the default OS in UNEP Toolkit Level
2 and published scientific papers [5 , 7 , 8] . The DF was from the default DF in UNEP Toolkit Level 2
and a published scientific paper [5] . rDF was from a published scientific paper [5] . More details
are provided in the spreadsheets [2] .
2.3. Quantification of mercury releases
The input of mercury (I Hg ) was calculated using IF and ARD. The releases of mercury to dif-
ferent media in the initial step were calculated using the I Hg and DF. The calculation algorithm
is detailed in Habuer et al. [1] . The R Hg to the final sinks in the rD step were calculated using
the sector-specific treatment/disposal, products or by-products, and rDF. Then, a substance flow
analysis of mercury was performed based on quantified input and output data, and distribution
data. STAN (SubsTance flow ANalysis) 2.6 freeware was used to identify the principal release
sources and visually present the distribution routes in the related research article [1] .
2.4. Uncertainty and sensitivity analysis
Since the default IFs in UNEP Toolkit Level 2 have wide ranges, the total inputs contain uncer-
tainty. This uncertainty was analyzed using a Monte Carlo method in Oracle Crystal Ball (OCB)
software. To determine the contribution to the total uncertainty, tornado analysis was conducted
using OCB. A tornado chart is useful for deterministic sensitivity analysis, i.e., comparing the rel-
ative importance of variables. In a tornado chart of the input variables, the upper bars repre-
sent the greatest contributors to the variability of the outcome, and therefore what the decision
maker should focus on. | 229a5963aa3649b5a9ad714c5764b07ddd8918d5 | page_0004 |
6 Habuer, T. Fujiwara and M. Takaoka / Data in Brief 42 (2022) 108280
Ethics Statements
The authors declare that creation of these data did not involve the use of human or animal
subjects, nor data collection from social media platforms.
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal rela-
tionships that could have appeared to influence the work reported in this paper.
Data Availability
Dataset for the expected anthropogenic mercury release in China between 2016 to 2019
(Original data) (Mendeley Data).
CRediT Author Statement
Habuer: Conceptualization, Methodology, Software, Investigation, Formal analysis, Data cura-
tion, Writing – original draft; Takeshi Fujiwara: Supervision, Writing –r e v i e w & editing; Masaki
Takaoka: Visualization, Validation, Writing –r e v i e w & editing.
Acknowledgments
This work was supported by JSPS KAKENHI Grant No. JP21K17895. A part of research
was conducted under the Environment Research and Technology Development Funds (JP-
MEERF20S20601).
References
[1] Habuer, T. Fujiwara, M. Takaoka, The response of anthropogenic mercury release in China to the Minamata conven-
tion on mercury: a hypothetical expectation, J. Clean. Prod. 323 (2021) 129089, doi: 10.1016/j.jclepro.2021.129089 .
[2] Habuer, T. Fujiwara, M. Takaoka “Dataset for the expected anthropogenic mercury release in China between 2016 to
2019”, Mendeley Data (2022), V2, doi: 10.17632/tjzm2gdntf.2 .
[3] United Nations Environment Programme (UNEP) ChemicalsToolkit for Identification and Quantification of Mercury
Releases, UNEP, 2017 Reference Report and Revised Inventory Level 2 Report. by Chemicals .
[4] Habuer , Y.J. Zhou , M. Takaoka , Time-series analysis of excess mercury in China, J. Mater. Cycles Waste Manag. 20
(3) (2018) 1483–1498 .
[5] M.L. Hui , Q.R. Wu , S.X. Wang , S. Liang , L. Zhang , F.Y. Wang , M. Lenzen , Y.F. Wang , L.X. Xu , Z.T. Lin , H. Yang , Y. Lin ,
T. Larssen , M. Xu , J.M. Hao , Mercury flows in China and global drivers, Environ. Sci. Technol. 51 (1) (2017) 222–231 .
[6] Habuer , T. Fujiwara , M. Takaoka , Anthropogenic mercury release flow in China, Chem. Eng. Trans. 83 (2021) 7–13 .
[7] K. Liu , S. Wang , Q. Wu , L. Wang , Q. Ma , L. Zhang , G. Li , H. Tian , L. Duan , J. Hao , A highly resolved mercury emission
inventory of Chinese coal-fired power plants, Environ. Sci. Technol. 52 (4) (2018) 2400–2408 .
[8] K. Liu , Q. Wu , L. Wang , S. Wang , T. Liu , D. Ding , Y. Tang , G. Li , H. Tian , L. Duan , X. Wang , X. Fu , X. Feng , J. Hao ,
Measure-specific effectiveness of air pollution control on China’s atmospheric mercury concentration and deposition
during 2013-2017, Environ. Sci. Technol. 53 (15) (2019) 8938–8946 .
[9] USGS, Minerals Yearbooks 2016-2020, National Minerals Information Center, 2021 https://www.usgs.gov/centers/
nmic/minerals- yearbook- metals- and- minerals accessed 6.23 .
[10] NBSC, China Statistical Yearbook 2016-2020 2016-2020, National Bureau of Statistics of the People’s Republic of
China, 2021 (in Both Chinese and English) ed.: National Bureau of Statistics of the People’s Republic of China. ac-
cessed 1.18 .
[11] M.D. Liu , Q.R. Zhang , Y. Luo , R.P. Mason , S.D. Ge , Y.P. He , C.H. Yu , R.N. Sa , H.L. Cao , X. Wang , L. Chen , Impact of
water-induced soil erosion on the terrestrial transport and atmospheric emission of mercury in China, Environ. Sci.
Technol. 52 (12) (2018) 6945–6956 . | 229a5963aa3649b5a9ad714c5764b07ddd8918d5 | page_0005 |
Food security has become an issue of global impor -
tance, and major price spikes for staples such as rice
and wheat have occurred in recent years. These price
spikes are partly due to the impact of plant diseases,
such as the spread of a new strain of the wheat stem
rust pathogen from East Africa into the Middle East1.
This has sparked an increased focus on improving
approaches to crop protection. The most effective and
environmentally sensitive approach to disease preven -
tion involves breeding crop plants for resistance. Indeed,
plant breeders have been using ‘resistance’ genes to con -
trol diseases in crop plants for almost 100 years, and the
effectiveness of this strategy sparked early genetic stud -
ies that defined ‘gene-for-gene’ relationships between
host resistance genes and pathogen virulence factors2.
However, only through recent molecular studies has
it become apparent that host resistance genes encode
components of the plant immune system that confer
the capacity to recognize and respond to specific path -
ogens. Plant immunity depends on cell-autonomous
events; these events are related to the innate immune
system in animals3 but plants have a much bigger rec -
ognition repertoire to compensate for their lack of an
adaptive immune system. Ongoing research is revealing
the recognition capacity of the plant immune system,
and concurrent studies on pathogen biology are begin -
ning to unravel how these organisms manipulate host
immunity to cause disease. The recent convergence of
these two fields has dramatically changed our percep -
tion of plant–pathogen interactions and is providing
new approaches for crop protection.Microbial plant pathogens almost always occupy
extracellular niches. Despite this, the nutrients that
enable pathogen growth are derived from host cells, and
the host cytoplasm and organelles are important sites of
molecular interaction. Plants have evolved two strategies
to detect pathogens4,5 (FIG. 1). On the external face of the
host cell, conserved microbial elicitors called pathogen-
associated molecular patterns (PAMPs) are recognized
by receptor proteins called pattern recognition receptors
(PRRs)6. PAMPs are typically essential components of
whole classes of pathogens, such as bacterial flagellin or
fungal chitin. Plants also respond to endogenous mol -
ecules released by pathogen invasion, such as cell wall
or cuticular fragments called danger-associated molec -
ular patterns (DAMPs). Stimulation of PRRs leads to
PAMP-triggered immunity (PTI). The second class of per -
ception involves recognition by intracellular receptors of
pathogen virulence molecules called effectors ; this recog -
nition induces effector-triggered immunity (ETI). This mode
of recognition leads to co-evolutionary dynamics between
the plant and pathogen that are quite different from PTI
as, in stark contrast to PAMPs, effectors are characteris -
tically variable and dispensable. Extreme diversification
of ETI receptors and pathogen effectors both within and
between species is the norm, whereas some PRR func -
tions are conserved widely across families. Generally, PTI
and ETI give rise to similar responses, although ETI is
qualitatively stronger and faster and often involves a form
of localized cell death called the hypersensitive response
(HR). PTI is generally effective against non-adapted
pathogens in a phenomenon called non-host resistance, *Commonwealth Scientific
and Industrial Research
Organisation (CSIRO),
Division of Plant Industry,
GPO BOX 1600, Canberra,
Australian Capital T erritory
2601, Australia.
‡Research School of Biology,
Australian National
University, RN Robertson
Building, Biology Place,
Acton, Australian Capital
T erritory 0200, Australia.
e-mails: peter .dodds @csiro.au;
john.rathjen@anu.edu.a u
doi:10.1038/nrg2812
Published online 29 June 2010Elicitors
Molecules that induce (‘elicit’)
an immune defence response.
In the context of this Review,
this term is used to refer to
both pathogen-associated
molecular patterns (PAMPs)
and effectors.
Pathogen-associated
molecular patterns
Any of a number of
conserved, usually structural,
molecules common to
pathogen organisms.Plant immunity: towards an integrated
view of plant–pathogen interactions
Peter N. Dodds* and John P . Rathjen‡
Abstract | Plants are engaged in a continuous co-evolutionary struggle for dominance
with their pathogens. The outcomes of these interactions are of particular importance
to human activities, as they can have dramatic effects on agricultural systems.
The recent convergence of molecular studies of plant immunity and pathogen
infection strategies is revealing an integrated picture of the plant–pathogen interaction
from the perspective of both organisms. Plants have an amazing capacity to recognize
pathogens through strategies involving both conserved and variable pathogen elicitors,
and pathogens manipulate the defence response through secretion of virulence
effector molecules. These insights suggest novel biotechnological approaches to
crop protection.REVIEWS
nATuRE REvIEwS | Genetics ADvAncE OnlInE Publ IcATIO n | 1Nature Reviews Genetics | AOP , published online 29 June 2010; doi:10.1038/nrg2812
© 20 Macmillan Publishers Limited. All rights reserved 10 | 6f570033c1d07467276ad185bf1521cd736c5f3f | page_0000 |
Nature Re views | GeneticsPAMP s
NB-LRREffecto r
Effecto r
Plant cellPTI respons e
ETI respons e
Fungus/
oomy ceteExtracellular spac e
HaustoriumPilus
PRR BAK1BacteriumPattern recognition
receptors
Plasma membrane-localized
receptors that recognize
the presence of pathogen-
associated molecular
patterns (PAMPs) in the
extracellular environment.
PAMP-triggered immunity
The plant defence response
elicited by pathogen-
associated molecular
pattern (PAMP) recognition.
Effectors
Proteins secreted by
pathogens into host cells
to enhance infection.
Many of these function to
suppress PAMP-triggered
immunity responses.whereas ETI is active against adapted pathogens. However
these relationships are not exclusive and depend
on the elicitor molecules present in each infection.
Here, we provide an overview of the plant PTI and
ETI systems, highlighting recent advances and identify -
ing key gaps in our understanding of these processes. we
consider the roles of PRRs in initial pathogen perception,
our expanding knowledge of pathogen effectors and their
roles in suppressing PTI responses, the nature of effector
recognition and the downstream responses to pathogen
perception. Finally, we discuss briefly how this knowledge
is beginning to feed back into the agricultural context
that originally spawned the study of plant immunity.
Extracellular recognition by PRRs
PRRs have been reviewed recently7, so here we discuss
some important principles and recent findings relat -
ing to key proteins in the process of recognition of
extracellular pathogen molecules.Pattern recognition receptors . Known PRRs fall into
one of two receptor classes: transmembrane receptor
kinases and transmembrane receptor-like proteins,
the latter of which lack any apparent internal signal -
ling domain7. Recent work has shown that endoplasmic
reticulum quality-control mechanisms are crucial for
PRR biogenesis (BOX 1). The receptor kinase gene family
has undergone huge expansion in plants: for exam -
ple, about 610 members are present in the Arabidopsis
thaliana genome, and many of these are responsive
to biotic stresses8. The receptor-like protein class has
57 members in A. thaliana9. The expansion of these
families is in contrast to the situation in animals, which
possess 12 Toll-like receptors that fulfil an equivalent
role to PRRs in plants10.
The PAMPs recognized by plants are multifarious
and include proteins, carbohydrates, lipids and small
molecules, such as ATP6. Recognition of PAMPs is
best understood in the case of the A. thaliana recep -
tor kinase F lAGE llIn SEnSInG 2 ( FlS2), which
binds bacterial flagellin directly and then assembles
an active signalling complex. Although the PAMP
concept encompasses the idea that all PAMPs should
be recognized by all species, this has been found to
not always be the case, as perception of the bacterial
elongation factor EF -Tu is apparently restricted to the
brassicaceae11. Similarly, the Xa21 receptor in rice pro -
vides race-specific resistance to the bacterial pathogen
Xanthomonas oryzae , and was recently shown to act as
a PRR for a novel sulphonated bacterial protein termed
Ax21 (ReF. 12).
BAK1 , a central regulator of PAMP -triggered immu -
nity. Most known PRRs require the leucine-rich
repeat ( lRR) receptor kinase bRASSI nOSTEROID
InSEnSITI vE 1-ASSO cIATED KI nASE 1 ( bAK1) for
function13,14 (FIG. 2). An exception is the fungal chitin
receptor cHITI n ElIcITOR RE cEPTOR KI nASE 1
(cERK 1)15,16, which also responds to an unknown bac -
terial PAMP17. bAK1 is part of a family of five somatic
embryogenesis receptor kinase (SERK) members and
is also known as SERK3. It is not yet known whether
other SERK family members have redundant roles in
immune signalling. bAK1 does not have a direct role
in elicitor perception, but F lS2 rapidly forms a complex
with bAK1 after elicitation. This interaction results in
phosphorylation of both proteins, which peaks 30–60
seconds after elicitor treatment18. bAK1 also has a
role in the perception of other elicitors, probably
also through heterodimerization with PRRs in the
lRR-receptor kinase family.
As such, bAK1 is a central regulator of plant immu -
nity and consequently the target of several pathogen
virulence effector molecules19 (see below). Despite
this, A. thaliana plants containing a null muta -
tion in the bak1 gene are actually marginally more
resistant to biotrophic pathogens, although they are
slightly more susceptible to necrotrophic pathogens20.
These phenotypes may be related to a deregulated
cell death phenotype that has been described in the
bak1 mutants20,21.Figure 1 | the principles of plant immunity. Bacterial plant pathogens propagate
exclusively in the extracellular spaces of plant issues. Most fungal and oomycete
pathogens also extend their hyphae into this space, although many also form
specialized feeding structures, known as haustoria, that penetrate host cell walls but
not the plasma membrane. Other fungi extend invasive hyphae into plant cells, but
again do not breach the host membrane. Molecules released from the pathogens into
the extracellular spaces, such as lipopolysaccharides, flagellin and chitin (pathogen-
associated molecular patterns (PAMPs)) are recognized by cell surface pattern
recognition receptors (PRRs) and elicit PAMP-triggered immunity (PTI). PRRs
generally consist of an extracellular leucine-rich repeat (LRR) domain (mid-blue),
and an intracellular kinase domain (red). Many PRRs interact with the related protein
BRASSINOSTEROID INSENSITIVE 1-ASSOCIATED KINASE 1 (BAK1) to initiate the PTI
signalling pathway. Bacterial pathogens deliver effector proteins into the host cell
by a type-III secretion pilus, whereas fungi and oomycetes deliver effectors from
haustoria or other intracellular structures by an unknown mechanism. These
intracellular effectors often act to suppress PTI. However, many are recognized
by intracellular nucleotide-binding (NB) -LRR receptors, which induces effector-
triggered immunity (ETI). NB -LRR proteins consist of a carboxyl -terminal LRR domain
(light blue), a central NB domain (orange crescent) that binds ATP or ADP (yellow oval),
and an amino -terminal Toll, interleukin -1 receptor, resistance protein (TIR) or
coiled-coil (CC) domain (purple oval).REVIEWS
2 | ADvAncE OnlInE Publ IcATIO n www.nature.com/reviews/genetics
© 20 Macmillan Publishers Limited. All rights reserved 10 | 6f570033c1d07467276ad185bf1521cd736c5f3f | page_0001 |
Effector-triggered immunity
The plant defence response
elicited by effector recognition.
Biotrophic
Biotrophic pathogens
propagate in living plant tissue
and generally do not cause
necrosis as a result of infection.
They use various means, such
as haustoria production, to
extract nutrients from host cells.
Necrotrophic
Necrotrophic pathogens
actively induce necrosis in
infected tissues, often through
the production of toxins, and
obtain nutrients from the
dead host tissue.
T ype-III secretion system
A syringe-like structure
produced by many plant
and animal pathogen bacteria
that allows direct secretion
of effector proteins from
the bacterial cytoplasm into
host cells.One potential regulator of the F lS2–bAK1 com -
plex is the cytoplasmic protein kinase bOTRYTIS-
InDucED KI nASE 1 ( bIK1). bIK1 was identified as
a potential regulator because bik1 is upregulated after
pathogen or elicitor treatment of A. thaliana leaves22.
bIK1 interacts with both F lS2 and bAK1 before elicita -
tion and seems to dissociate from the complex after elic -
itation. In vitro , bAK1 phosphorylates bIK1 and bIK1
phosphorylates both F lS2 and bAK1. In vivo ,
bIK1 becomes phosphorylated 5–10 min after treat -
ment with flagellin23; this phosphorylation peaks
after the F lS2–bAK1 phosphorylation. confusingly,
bik1 mutant A. thaliana plants are more resistant
to Pseudomonas syringae infection than wild-type
A. thaliana plants22 as a result of them overproduc -
ing the defence hormone salicylic acid (SA), but they
are also more susceptible to infection with the necro -
trophic fungal pathogen Botrytis cinerea . Despite this,
deficiencies in F lS2-mediated immune responses
could be measured in these plants23. These contrast -
ing results make it difficult to ascribe a clear function
to bIK1 in plant immunity, and further studies will
be required.
Virulence activities of pathogen effectors
Successful pathogens are able to suppress PTI responses
and thereby multiply and cause disease. They achieve sup -
pression through the deployment of ‘effector’ proteins.
Studies of bacterial phytopathogens have provided
most of our understanding of effector strategies
and mechanisms. Individual phytopathogen strains
encode 20–30 effectors, which are highly regulated and
secreted directly into the host cytoplasm by a dedicated needle structure, the type-III secretion system (TTSS)24.
The repertoire of individual effectors varies dra -
matically among closely related bacterial strains, and
effectors themselves act redundantly and are appar -
ently interchangeable25; examples of such effectors
are discussed below. Many effectors interfere directly
with PTI responses26, and bacterial mutants that lack
the TTSS system are non-pathogenic. Interestingly,
a number of examples show that transgenic overex -
pression of an individual type-III effector in the host
plant restores the ability of such bacterial mutants to
grow27,28, suggesting that bacterial pathogenicity only
requires suppression of PTI. However, contributions
of as yet undefined mechanisms to other processes,
such as nutrient acquisition, cannot be excluded.
Bacterial effector functions. bacterial effectors have
molecular or enzymatic activities that specify both
their ability to modify host targets and their intracel -
lular recognition by ETI receptors29 (see below). The
redundancy among effectors is illustrated by the unre -
lated P . syringae effectors AvrPto and AvrPto b, which
both target the F lS2–bAK1 complex. Although the
models for how suppression works conflict in molec -
ular detail19,30,31, it is generally accepted that AvrPto b
uses a dual strategy for kinase suppression: its amino -
terminal kinase-targeting domain is sufficient to sup -
press flagellin responses, and its carboxy -terminal
E3 ligase domain can tag interacting kinase proteins
with ubiquitin to direct them for degradation32,33.
AvrPto b is known to target five host kinases of the
Pto/interleukin receptor-associated kinase (IRAK)
class32, but because this clade is hugely expanded in
plants8, there are probably many more such targets.
likewise, AvrPto suppresses multiple PRR receptor
kinases, perhaps by acting as a kinase inhibitor19,30,34.
Overall, these effectors seem to be part of a bacterial
strategy that targets host kinases nonspecifically.
A further example of overlapping effector func -
tions involves the host protein RPM1-I nTERA cTInG
PROTEI n 4 (RIn4), which is targeted by the P . syringae
effectors Avr b, AvrRPM1 and AvrRpt2 through dif -
ferent molecular strategies35,36. Recently, it was shown
that the P . syringae effector HopF2 may also target
RIn4 (ReF. 37). Overexpression of HopF2 prevented
degradation of RI n4 by the protease AvrRpt2 but
did not alter the interactions of RI n4 with AvrRPM1
or Avr b. bacteria that lack HopF2 have increased
growth on lines that lack RI n4, suggesting that RI n4
could indeed be a target for virulence, but an indirect
cause for this observation was not ruled out. RI n4
is a negative regulator of both PTI and ETI28,38, and
also interacts with the plasma membrane H+-ATPases
AHA1 and AHA2 to enhance stomatal openin g39, a
key event during bacterial pathogenicity on leaves.
Thus it is not clear how targeting of RI n4 by multiple
effectors would enhance bacterial virulence, as disrup -
tion of RI n4 should actually restrict pathogenicity.
However, the number of effectors involved in this
process is consistent with RI n4 being an important
virulence target. Box 1 | Pattern recognition receptor biogenesis
Most eukaryotic membrane proteins undergo quality control during folding and
maturation in the endoplasmic reticulum (ER), a process termed ER‑QC114.
A number of recent studies show that the biogenesis of a pattern recognition
receptor (PRR), the EF‑Tu receptor (EFR), is regulated by this mechanism115–119.
After secretion into the ER, proteins are modified at glycosylable Asn residues by
an oligosaccharyltransferase complex, which covalently attaches a complex
polysaccharide containing three terminal glucose residues. The glucose moieties
are subsequently trimmed by glucosidases I and II. A single glucose residue is
added back by UDP ‑glucose:glycoprotein glucosyltransferase (UGGT) near regions
of protein disorder. Monoglucosylated proteins interact with the lectins calnexin
(CNX) or calreticulin (CRT) to retain misfolded substrates in the ER. In this way,
UGGT acts as a folding sensor, and glycosylation is intimately related to protein
maturation. Terminally misfolded proteins are degraded.
Another ER folding pathway is based on the chaperone BiP (a form of heat shock
protein 70 (Hsp70)). Unfolded proteins undergo cycles of BiP binding and release,
which is regulated by Hsp40 co ‑chaperones containing J domains (for example, the
ERdj protein). Forward genetic screens showed that Arabidopsis thaliana genes
encoding glucosidase II, UGGT, CRT3, ERdj3B and ERD2b are required for EFR
function and accumulation. In addition, STT3A, a subunit of the oligosaccharyl ‑
transferase complex, was necessary for EFR biogenesis. Finally, STROMAL ‑
DERIVED FACTOR 2 (SDF2) resides in a protein complex with ERdj3B and BiP , and
was also required for EFR maturation. Plants with mutations in these genes are
generally more susceptible to pathogens, indicating that EFR is not the only
immune protein that is governed by ER‑QC. However, neither FLAGELLIN SENSING 2
(FLS2) nor CHITIN ELICITOR RECEPTOR KINASE 1 (CERK1) function is significantly
affected in these mutants.REVIEWS
nATuRE REvIEwS | Genetics ADvAncE OnlInE Publ IcATIO n | 3
© 20 Macmillan Publishers Limited. All rights reserved 10 | 6f570033c1d07467276ad185bf1521cd736c5f3f | page_0002 |
Nature Re views | GeneticsPRR BAK1 PRR BAK1Bacterium Apoplast
PRR BAK1 P P
CDPKsBIK1a
bc
Plant cell MAPKs
Haustoria
(sing. haustorium.) Specialized
structures produced by
some fungal and oomycete
pathogens. Haustoria extend
through the plant cell wall
and expand in the host cell.
They remain surrounded by
a host-derived membrane
and hence are topologically
extracellular and separated
from the host cytoplasm.
Hemibiotrophic
Hemibiotrophic pathogens
incorporate aspects of both
biotrophic and necrotrophic
infection strategies. Often
this involves an initial
biotrophic infection phase
during which the pathogen
spreads in host tissue,
followed by a necrotrophic
phase during which host
cell death is induced.
NB-LRR proteins
A class of intracellular
receptor proteins containing
nucleotide-binding (NB) and
leucine-rich repeat (LRR)
domains that recognize specific
pathogen effectors.It is important to note that not all effectors target PTI.
One example of an alternative bacterial effector strategy
is given by the transcription activator-like (TA l) effec -
tors of Xanthomonas spp., which are transcription fac -
tors that induce the expression of specific host genes,
some of which contribute to symptom developmen t40.
unlike AvrPto and AvrPto b in Pseudomonas spp., TA l
effectors do not seem to act redundantly because several
of them are essential for virulence. They interact specifi -
cally with a site in the target gene promoters through a
central tandem repeat region that forms a D nA-binding
domain41–43. Strikingly, two hypervariable amino acid
residues in each repeat specify interaction with a charac -
teristic nucleotide in the effector recognition site. Thus,
the nucleotide sequence of the target D nA can be pre -
dicted by the amino acid sequence of the tandem repeat
domain. biotechnologically this is significant because it
enables precise modification of gene expression in vivo,
including turning this system against Xanthomonas spp.
by engineering Avr bs3-responsive elements (known as
uPA sites), upstream of active resistance genes44. In
nature, this strategy has been pre-empted in some plant
species: target sites for certain TA l effectors have been
incorporated upstream of the resistance genes Bs3 and
Xa27 in pepper and rice, respectivel y45,46.
Eukaryotic effectors. Data on eukaryotic effectors and
their functions are sparse in comparison with data on
bacterial effectors. both fungal and oomycete patho -
gens produce effectors that are secreted through the
endomembrane system and are subsequently delivered
into host cells by unknown mechanisms47,48. Oomycete
effectors characteristically contain the internal motif Arg-X-leu-Arg (RX lR, in which X represents any
amino acid), which is required for delivery into plant
cells. Genome sequencing of Phytophthora infestans49,
the Irish potato famine pathogen, revealed 563 RX lR
effector genes. Seventy of these genes are under diver -
sifying selection and only 16 share orthologues in the
genomes of 2 other sequenced Phytophthora spp., which
indicates that very strong selection processes act on
these effectors. A further 196 effectors of a separate class
(known as crinkler proteins) are encoded by P . infestans .
Such generalized identification of fungal effector genes
has been restricted by the lack of conserved motifs to
aid genome interrogation, but genome analysis of sev -
eral fungal pathogens predicts that they have complex
and diversified secretomes50,51. The massive expansion
in eukaryotic effector repertoires relative to bacterial
effector repertoires may suggest a requirement for more
diverse effector functions by eukaryotic pathogens,
possibly to support their more specialized nutrient
acquisition strategies.
Some data support roles of P . infestans effectors in
suppression of immunity52; for example, Avr3a sup -
presses elicitor-induced cell death through interac -
tion with the host cMPG1 E3 ligase53, but in general
very little is known about effector functions in fungi
or oomycetes. However, many other potential roles
remain, such as establishment of the pathogenic niche
through development of the haustoria feeding struc -
tures and manipulation of host cell death during the
hemibiotrophic lifestyle.
Sedentary nematode pathogens of plants form pro -
longed associations with roots, in which they induce the
formation of novel host structures, such as multinucle -
ate giant cells, from which they feed using a specialized
proboscis called a stylet. The stylet also delivers salivary
secretions into host cells; proteomic analysis of saliva
from one such species, Meloidogyne incognita , identi -
fied 486 potential effector proteins54. Ongoing genom -
ics analyses of such species will identify many more and
help in elucidating the pathogenic strategies of these fas -
cinating organisms. In addition, viral pathogens encode
specific suppressors of the small R nA pathway to pre -
vent degradation of their genomes and/or abrogation of
viral gene expression55.
Overall, our understanding of effector proteins and
their host targets is at a very early stage. Sophisticated
biochemical screens for host protein targets that interact
with the diverse suites of pathogen effectors are likely to
lead to the identification of important components of
host defence mechanisms and teach us more about host
immune pathways and pathogenicity strategies.
Intracellular effector recognition
ETI is the second pathogen-sensing mechanism in
plants and is based on intracellular recognition of effec -
tor proteins4,5. Recognition events are mostly mediated
by a class of receptor proteins that contain nucleotide-
binding ( nb) domains and lRRs (FIG. 1). Plant NB-LRR
proteins confer resistance to diverse pathogens, includ -
ing fungi, oomycetes, bacteria, viruses and insects. nb
and lRR domains are also present in nOD-like immune Figure 2 | Formation of active pattern recognition receptor complexes.
a | Infectious pathogens, such as bacteria, shed pathogen-associated molecular
patterns (PAMPs; pink, yellow and purple shapes) into the apoplast, where they are
recognized by specific pattern recognition receptors (PRRs). b | Immediately after
ligand binding, the PRR forms an active complex with BRASSINOSTEROID
INSENSITIVE 1-ASSOCIATED KINASE 1 (BAK1). c | This results in transphosphorylation
(indicated by P) of the respective kinase domains of the PRR and BAK1. Signalling via
this active complex can be mediated directly by BOTRYTIS-INDUCED KINASE 1
(BIK1), or by mitogen-activated protein kinases (MAPKs) or calcium-dependent
protein kinases (CDPKs). This is a generalized model that is based on FLAGELLIN
SENSING 2 (FLS2), the receptor for bacterial flagellin.REVIEWS
4 | ADvAncE OnlInE Publ IcATIO n www.nature.com/reviews/genetics
© 20 Macmillan Publishers Limited. All rights reserved 10 | 6f570033c1d07467276ad185bf1521cd736c5f3f | page_0003 |
Nature Re views | Geneticsa Direct b Guar d/decoy c Bait
receptors ( nlRs), which are involved in PAMP induction
of innate immunity responses in animals3,56, and in the
animal apoptotic factors apoptotic protease-activating
factor 1 (APAF1) and cell death protein 4 ( cED4). Many
plant nb-l RR proteins also contain an n-terminal TIR
(Toll, interleukin -1 receptor, resistance protein) domain
related to the intracellular signalling domain of animal
Toll-like receptors10. A second common class of nb-l RR
proteins contain an n-terminal domain with a coiled-
coil ( cc) domain, whereas others have no conserved
n-terminal region.
Direct and indirect recognition. nb-lRR proteins can
recognize pathogen effectors either directly by physical
association or indirectly through an accessory protein
that is part of an nb-l RR protein complex (FIG. 3). In
general, direct recognition has been demonstrated by
yeast two-hybrid (Y2H) assays, in some cases supported
by in vitro protein interaction assays. For example, the
rice cc-nb -lRR Pi-ta protein binds to the Magnaporthe
grisea effector AvrPita both in vitro and in Y2H assays57.
The flax TIR- nb-lRR l and M proteins also interact
in Y2H assays with the Melampsora lini fungal effec -
tors Avr l567 and AvrM, respectivel y58–61. These pairs of
receptor and effector proteins show evidence of strong
diversifying selection and are characterized by high
levels of sequence polymorphism between alleles in
the host and pathogen populations, respectively, with
these variants showing different recognition specifici -
ties. This is likely to be the result of antagonistic co-
evolution between the interacting components in the
host and pathogen.
Indirect effector recognition has been observed in
a number of cases. In the best-described models, the
effector interaction is mediated by an accessory protein
that is a pathogen virulence target or a structural mimic
of one. The effector induces a change in the accessory
protein that enables the accessory to be recognized by
the nb-l RR protein62. This strategy neatly sidesteps the
evolutionary advantage of the faster evolving pathogen,
as the host takes advantage of the pathogen’s virulence
strategy to drive the recognition. Three conceptual models have been proposed to describe these mecha -
nisms (FIG. 3). The ‘guard’ model postulates that nb-l RR
proteins guard an accessory protein (or guardee) that
is targeted and modified by pathogen effectors63. This
model is exemplified by the A. thaliana RIn4 protein.
RIn4 forms exclusive complexes with the nb-l RR pro -
teins RPM1 and RESISTA ncE TO PSE uDOMO nAS
SYRI nGAE 2 (RPS2)36,64. Degradation of RI n4 by the
protease effector AvrRpt2 de-represses RPS2, whereas
Avrb or AvrRPM1-mediated phosphorylation of
RIn4 activates RPM1 (ReFS 35,36) . Thus, modification
of RI n4 by the effectors explains how an individual
nb-l RR (in this case, RPM1) can recognize more than
one effector.
However, the guard model postulates that RI n4 is a
virulence target of the effectors, which is as yet unproven
(see also above). Also, this model creates an evolutionary
problem: RI n4 should evolve to avoid binding to the
effector proteins in the absence of RPS2 and RPM1, but
in their presence, selection will favour effector binding to
promote recognition5. To solve this problem, the ‘decoy’
model was proposed62, in which duplication of the
effector target gene or independent evolution of a target
mimic could relax evolutionary constraints and allow the
accessory protein to participate solely in effector percep -
tion. This situation is exemplified by the tomato nb-l RR
protein Prf, which forms a complex with the accessory
protein Pto kinase65. Pto kinase is closely related to the
kinase domains of F lS2 and cERK1, which are targets of
AvrPto and AvrPto b32,66. Thus, Pto provides the recogni -
tion capability for Prf, and this drives diversification of
the Pto family to broaden the spectrum of recognition
capability67. In the decoy model, the accessory protein
specializes in perception of the effector by the nb-l RR
protein but has no other function. This fails to explain
the requirement for Pto kinase activity in Prf activation68
and the clear role of RI n4 in defence responses. A fur -
ther modification of the decoy concept is the bait-and-
switch model69, which envisages a two-step recognition
event. First, an effector interacts with the accessory
‘bait’ protein associated with an nb-l RR, and then a
subsequent recognition event occurs between the effec -
tor and nb-l RR protein to trigger signalling. That
is, the nb-l RR protein interacts with an effector target
(the bait) to facilitate direct recognition of the pathogen
effector, rather than recognizing the modified target as
envisaged in the guard model.
It is important to remember that these models are
generalizations based on limited specific examples, none
of which is yet fully understood. Thus, although they are
useful conceptual tools, they are unlikely to adequately
describe all effector recognition events and can be
restrictive. For instance, in addition to providing effec -
tor recognition, Pto seems to participate actively with
Prf in a highly evolved co-regulatory relationship65,68.
The massive diversity in effector and receptor biol -
ogy suggests that many variations on these themes,
and probably other novel recognition events, are likely
to occur. For example, the Pto kinase phosphorylates
the effector AvrPto b, leading to inactivation of its
intrinsic E3 ligase activity70; this is an intriguing and Figure 3 | Models of direct and indirect recognition. Plant nucleotide-binding
(NB) -leucine-rich repeat (LRR) receptors can recognize pathogen effectors by
either direct or indirect mechanisms. a | In direct recognition, the effector (green)
triggers immune signalling by physically binding to the receptor (purple, orange,
yellow and blue; see FIG. 1 for a description of the receptor). b | In the guard and
decoy models, the effector modifies an accessory protein (red), which may be its
virulence target (guard model) or a structural mimic of such a target (decoy model).
The modified accessory protein is recognized by the NB -LRR receptor. c | Under the
bait model, interaction of an effector with an accessory protein facilitates direct
recognition by the NB -LRR receptor.REVIEWS
nATuRE REvIEwS | Genetics ADvAncE OnlInE Publ IcATIO n | 5
© 20 Macmillan Publishers Limited. All rights reserved 10 | 6f570033c1d07467276ad185bf1521cd736c5f3f | page_0004 |
so far unique example of the host taking the bacteria’s
virulence strategy for its own. Interestingly, several
examples have been described recently in which two
different nb-l RR genes are required for recognition
of specific effector proteins71–75, providing a further
challenge for recognition models. The A. thaliana TIR-
nb-lRR genes RPS4 and RESISTANT TO RALSTONIA
SOLANACEARUM 1 (RRS1) are arranged in divergent
tandem configuration within the major recognition
gene complex MR c-J on chromosome 5 (ReF. 76). RPS4
confers immunity to P . syringae through recognition of
the effector AvrRps4, and RRS1 recognizes the Ralstonia
solanacearum effector PopP2; new data show that both
of these genes need to be expressed together for recog -
nition of these effectors and for resistance to the fungus
Colletotrichum higginsianu m71. Genetic data suggest
that the encoded proteins act in the same pathway,
potentially as members of a protein heterocomplex.
There are nine other examples of coordinate nb-l RR
gene arrangement in the A. thaliana genome, and
numerous other examples of nb-l RR genes working
together have been described in various species. How
these proteins function together remains unknown.
NB-LRR activation. One of the remaining challenges
is to understand how effector recognition leads to
nb-l RR activation, and whether the activation mecha -
nisms are the same for different recognition systems.
broadly, the nb-l RR is a conserved multidomain
switch that translates diverse direct or indirect pathogen
signals into a general immune response69. numerous
genetic studies have shown that the lRR domain often
controls recognition specificity77–80, with the implica -
tion that the lRR mediates effector interaction in these
systems. However, these studies have necessarily been
conducted on nb-l RRs that belong to diversified fami -
lies, including some that are known to directly interact
with their cognate effectors. by contrast, lRR domains
of nb-l RRs that participate in indirect recognition are
often conserved, and it is not clear what part the lRR
domain plays in these cases. It is possible that direct and
indirect recognition mechanisms involve fundamentally
different nb-l RR activation processes.
In the absence of an effector trigger, nb-l RR pro -
teins are maintained in a restrained conformation. In
some indirect recognition systems, negative regulation
of the nb-l RR by an accessory protein is released by
effectors, and this is sufficient for activation of ETI81.
This constitutes a simple paradigm that may occur
widely. In other cases, the nb-l RR is autoinhibited75;
that is, intramolecular interactions hold the protein in
an inactive conformation until disrupted by the pres -
ence of the effector. This may be a general feature of
direct recognition events. nucleotide binding by the
nb domain seems to be crucial for the function of all
plant nb-l RR proteins75, and signal activation may
involve an exchange of ATP and ADP in the bind -
ing site82. biochemical analysis of nb-l RR proteins
and their complexes has proven difficult but is cru -
cial to advance our understanding of these complex
activation events.Animal nb-containing proteins, such as nlRs and
the apoptotic factors APAF1 and cED4, self-oligomerize
through their centrally located nb domain after activa -
tion, thereby forming an active signalling platform83.
In this state, an n-terminal interaction domain (such
as a caspase recruitment domain ( cARD), pyrin
domain or baculovirus inhibitor ( bIR) domain) is
made accessible for signalling adaptor proteins, which
initiate the downstream signalling pathways leading
to inflammatory response or apoptosis84. Similarly,
the tobacco n protein oligomerizes in the presence
of p50; oligomerization is dependent on a functional
nb domain and also seems to involve the n-terminal
TIR domain85. Interestingly, tomato Prf exists in an
oligomeric complex before stimulation with AvrPto
or AvrPto b67.
Similar to the n-terminal domains of mammalian
nOD proteins, there is evidence that the TIR domain
provides the downstream signalling capability for plant
TIR- nb-lRR proteins. For instance, deletion or point
mutations of the TIR domain from the tobacco n pro-
tein block HR induction downstream of the oligomer -
ization event85. Furthermore, overexpression of the
isolated TIR domains of several TIR- nb-lRR proteins
is sufficient to trigger an HR86,87. The TIR domains of
Toll-like receptors are activated by dimerization trig -
gered by extracellular PAMP recognition10, so it is pos -
sible that effector-induced R protein oligomerization
enables TIR activation through induced proximity. For
some non-TIR nb-l RRs, overexpression of the cc-nb -
ARc fragments can trigger plant defence signalling,
whereas the cc domains alone do not88–90. In tobacco,
the n-terminal portion of the nb domain of the Rx
protein (which confers resistance to potato virus X)
is sufficient to induce cell death89.
Signalling pathways and downstream responses
A number of cellular events associated with both PTI
and ETI are known, essentially as correlative phenom -
ena. These include a rapid influx of calcium ions from
external stores, a burst of active oxygen species, acti -
vation of mitogen-activated protein kinases (MAPKs),
reprogramming of gene expression, deposition of cal -
losic cell wall appositions at sites of attempted infection
and, often, localized cell death (HR). There is extensive
overlap among the gene expression profiles elicited by
most PAMPs6. PTI and ETI gene expression signatures
are largely similar, suggesting that the responses are
the same overall but vary in magnitude91. One of the
big gaps in our understanding of plant immunity is
in the signalling pathways that operate immediately
downstream of PRR and nb-l RR protein activation.
Genetic screens have had very limited success in iden -
tifying signalling components, and the components of
these pathways remain mostly elusive. Several of the
partially understood pathways are described below.
Kinase signalling. One topic that has received a lot
of attention is MAPK signalling. MAPK pathways
are ubiquitous signal transduction components in
eukaryotes and transfer signals from extracellular REVIEWS
6 | ADvAncE OnlInE Publ IcATIO n www.nature.com/reviews/genetics
© 20 Macmillan Publishers Limited. All rights reserved 10 | 6f570033c1d07467276ad185bf1521cd736c5f3f | page_0005 |
receptors to cellular responses. A MAPK cascade typi -
cally consists of a modular complex consisting of a MAPK
kinase kinase (MAPKKK), which phosphorylates a
MAPK kinase (MAPKK), which phosphorylates a MAPK.
These pathways regulate the activity of various substrates,
such as transcription factors and protein kinases.
Importantly, MAPK cascades have been impli -
cated in both PTI and ETI92. A putative MAPK cas -
cade that acts downstream of flagellin perception has
been characterized in A. thaliana . It comprises the
MAPKKs MKK4 and MKK5 upstream of the MAPKs
MPK3 and MPK6, and leads to activation of wRKY-
type transcription factors. The cascade culminates in
the expression of defence genes93. congruently, con -
stitutively active MKK4 and MKK5 confer resistance
to infection by P . syringae in A. thaliana . Previously,
the MAPKKK MEKK1 was thought to be part of
this cascade93 but more recent evidence indicates
that this is unlikely, as mekk1 mutant plants are not
compromised in activation of MPK3 and MPK6 trig -
gered by the flagellin peptide flg22 (ReF. 94). Rather,
MEKK1 seems to act at the apex of a cascade com -
prising MEKK1, MKK1, MKK2 and MPK4, which is
also activated by flg22 treatment. MPK3 and MPK6
are also activated by other PAMPs6. MPK6 activates
ethylene biosynthesis through phosphorylation of
1-AMI nOcYclOPROPA nE-1-cARbOXY lIc AcID
SYnTHASE (A cS6) on flg22 perception95. Moreover,
ERF104, an ethylene response factor, is a known MPK6
substrate96. The MPK6–ERF104 interaction is rapidly
lost in response to flg22, presumably allowing the
liberated ERF104 to access target genes and activate
ethylene signalling during PTI96.
In a recent paper, Sheen and colleagues defined an
alternative pathway based on activation of calcium-
dependent protein kinases ( cDPKs)97. using a func -
tional genomics approach, they defined a subclade
of A. thaliana cDPKs that are required for F lS2-
dependent immunity. This pathway acts mostly inde -
pendently of the MAPK pathway, as judged by gene
expression assays, but antagonistic and synergistic
effects were also observed. This model is consistent with
the observation that calcium channel inhibitors abrogate
most immune responses elicited by microbe-associated
molecular patterns (MAMPs) or effectors.
Effector-triggered immunity signalling. Most of the
genes identified in genetic screens for suppressors of
ETI are either genes specific to the recognition system
used in the screen, such as recognition accessory pro -
teins, or members of a chaperone complex required for
the function of many nb-l RR proteins98. Only a cou -
ple of genuine signalling proteins have been identified.
EnHAncED DISEASE S uScEPTI bIlITY 1 (EDS1)
is required for signalling of all TIR- nb-lRRs tested to
date, suggesting that it acts specifically in TIR domain
signalling99. However, it is not clear what intermediaries
connect EDS1 and TIR- nb-lRRs. Similarly, the inte -
gral plasma membrane protein nOn-RAcE-SPE cIFIc
DISEASE RESISTA ncE 1 ( nDR1) is required for
signalling from some cc-nb -lRRs (which are all membrane associated), but again the connecting steps
are unknown100. The lack of success with genetic screens
for signalling components could suggest that there
are few essential elements in ETI signalling, and
there is a possibility that redundant signalling pathways
operate in parallel. biochemical approaches for identi -
fying signalling components interacting with activated
nb-l RR proteins will be necessary to uncover further
steps in these pathways.
Despite the difficultly in identifying components,
an interesting model for ETI signalling has been pro -
posed recently. In this model, nb-l RRs relocate to the
nucleus on activation and interact with nuclear factors
to trigger changes in gene expression. For instance,
the tobacco n protein, barley M lA10 protein and
A. thaliana RPS4 proteins localize to both the cell
cytoplasm and nucleus, and nuclear accumulation is
required for their function101–103. However, only a small
fraction of these nb-l RR proteins is present in the
plant cell nucleus. It can also be difficult to distinguish
interactions related to recognition from those related
to signalling. For example, A. thaliana RRS1 -R inter -
acts with the Ralstonia solanacearum effector PopP2
in the nucleus104,105. In addition, although a putative
nuclear localization signal is required for the func -
tion of full-length RPS4 protein, it is not present in
the constitutively active TIR domain of RPS4 (ReF. 87).
To date, no signalling partners common to different
nb-l RR proteins have been identified in the nucleus.
It will be interesting to discover the extent to which
nuclear localization explains the signalling activity of
these examples, and whether this is a general feature
of all nb-l RR receptors.
Downstream responses. Some of the downstream
responses to ETI and PTI are better understood than the
signalling pathways. The SA and jasmonic acid (JA)–
ethylene (ET) hormone pathways are important regula -
tors of defence-gene expression106. These two pathways
act antagonistically to some extent, with SA involved in
resistance to biotrophic pathogens and JA–ET involved
in responses to necrotrophic pathogens and chewing
insects. However, although there are substantial dif -
ferences in the gene expression outputs of these path -
ways, and several genes act as specific markers for the
activation of either the SA or JA–ET pathways, there
is also considerable overlap between them. Recently,
Tsuda et al.107 found complex interactions between SA
and JA–ET signalling in a detailed combinatorial study
using multiple mutants blocked in different pathways.
The SA and JA–ET pathways seemed to act synergisti -
cally in PTI to amplify the response. This may explain
why many pathogen effectors are able to suppress PTI
by interacting with different targets; because the signal
itself is relatively weak, blocking just one component
is sufficient to substantially perturb the response.
However, the ETI response is stronger and involves
redundant activities of SA and JA–ET pathways107.
Thus, even in the absence of SA signalling, the JA–ET
response contributes to maintaining a substantial
level of pathogen resistance. These compensatory REVIEWS
nATuRE REvIEwS | Genetics ADvAncE OnlInE Publ IcATIO n | 7
© 20 Macmillan Publishers Limited. All rights reserved 10 | 6f570033c1d07467276ad185bf1521cd736c5f3f | page_0006 |
interactions may simply result from the higher signal
flux in ETI, and probably make this response more
robust against pathogen interference.
Despite, or perhaps because of, the number of gene
expression changes resulting from PTI and ETI acti -
vation, the key changes that result in prevention of
pathogen growth are not clear in any disease system.
Perhaps the many responses each have such minor
effects that individual contributions are difficult to
quantify, but it is also likely that different aspects of
the response are effective against different types
of pathogens. Some recent work is beginning to
unravel specific responses with significant effects on
pathogen invasion. For instance, the PE nETRATIO n 2
(PEn2) and PE n3 proteins of A. thaliana are involved
in preventing cellular penetration by powdery mildew
fungi108,109. PEn2 is a hydrolytic enzyme that produces a
glucosinylate compound from an inactive precursor and
PEn3 is an A bc transporter that seems to be involved
in secretion of this molecule at the site of fungal
attack110,111. Although these activities may have a direct
antimicrobial role, both proteins are also required for
the deposition of callose at the infection site and encase -
ment of powdery mildew haustoria, suggesting a more
subtle regulatory role in blocking infection. The Lr34
gene of wheat, which has been widely used in agriculture
because it confers broad spectrum resistance against leaf
and stripe rust fungi as well as powdery mildew, was
also recently shown to encode an A bc transporter112.
RPw8 in A. thaliana is another protein that provides
broad spectrum resistance to powdery mildew fungi. It
is targeted to a host-derived membrane that surrounds the fungal infection structures and also acts to enhance
the callosic encasement of the fungal haustorium113.
Conclusions
The current synthesis of plant–pathogen molecular
interactions provides a strong conceptual framework
for understanding how these organisms coexist. Plants
have evolved innate immune systems that recognize
the presence of potential pathogens and initiate effec -
tive defence responses, whereas successful pathogens
have evolved effector proteins that can suppress host
immune responses. Furthermore, effectors can them -
selves act as elicitors and can be disabled by the host.
Overall, the pathogenic niche is highly evolved and
carefully monitored by both participants.
Despite the advances in characterization of individ -
ual molecular interactions and their consequences, our
understanding of host–pathogen molecular co-evolution
is poorly developed. The dominant synthesis, as cur -
rently understood, invokes a molecular arms race.
However, this area is in fact relatively unexplored and
both hosts and pathogens have generally not evolved
rapid mechanisms to generate massive diversity in
pathogen elicitors or host receptors, respectively. It is
crucially important for the deployment of existing and
novel resistance genes in agriculture that we advance
our knowledge in this area to aid predictions of how
changes in selection parameters will affect the evolution
of pathogens, at both microscale and population levels.
This need for a better understanding of co-evolution is
particularly true for breakthrough technologies, such as
the deployment of new PAMP-recognition specificities
in crop species (BOX 2).
Moreover, there are many fundamental molecular
questions about which we are still ignorant, such as
what are the distinct and common signalling compo -
nents of PTI and ETI? How are nb-l RR proteins acti -
vated by effector recognition? what are the induced
host components and compounds downstream of
pathogen perception that effect immunity? what are
the targets of effectors, and how does the deployment
of these effectors maximize the pathogenic niche? And
what are the effector delivery mechanisms of fungi
and oomycetes? widespread genome sequencing of
both host and pathogen genomes will facilitate the iden -
tification of effector proteins, the genome-wide analysis
of dynamic effector expression patterns and the identifi -
cation, through proteomics and gene homology, of host
target proteins. The immediate technological impact of
next-generation sequencing will open up the study
of important non-model host–pathogen systems, such
as wheat rusts and the black sigatoka disease of banana.
One promising avenue is to exploit the diversity of
plant species to access useful pathogen receptors from
sexually incompatible host plants, which will expand
the resource of resistance genes that can be transferred
into agricultural species. However, filling in many of the
current gaps in knowledge will require the application
of biochemical, structural and cell biology approaches
to unravel the molecular events associated with receptor
activation and downstream signalling pathways. Box 2 | Novel agricultural applications
Plant breeders have long recognized the importance of resistance genes for
preventing disease in crop plants. Many of these genes have now been found to
encode effector ‑triggered immunity (ETI) receptors, and we know that pathogens
can evolve to overcome these genes through loss or alteration of the effectors
that are recognized. The careful deployment of resistance genes in crop plants,
particularly by using multiple effective receptors in combination and by
selecting target effectors that have crucial virulence functions, should allow
more durable resistance.
Many nucleotide ‑binding (NB)‑leucine ‑rich repeat (LRR) genes have now been
cloned, and this can facilitate their application in agriculture either through
conventional breeding approaches, in which the cloned sequences are used as
molecular markers, or through transgenic means. Widespread genome
sequencing of plant pathogens is now yielding long lists of effector proteins that
could be recognized by plant immune receptors, and these can now be screened
against wild relatives of crop plants to identify new sources of resistance. This
approach has been useful already in identifying new sources of resistance to the
potato blight pathogen Phytophthora infestans in wild potatoes120. Pathogen ‑
associated molecular pattern (PAMP) ‑triggered immunity (PTI) receptors are
typically not variable within species and thus have not contributed widely to
traditional breeding efforts. However, the transfer of these receptors among
species has tremendous potential to deliver durable resistance, as the recognition
components are highly conserved among pathogens. Although pathogens that
are adapted to a particular host plant may be adept at suppressing the pattern
recognition receptors (PRRs) of that host, their effectors might not recognize
PRRs from other host plants. For instance, the Arabidopsis thaliana EF‑Tu receptor
occurs only in the Brassicaceae family, and transfer of this gene into tomato
provided good resistance against various bacterial pathogens121.REVIEWS
8 | ADvAncE OnlInE Publ IcATIO n www.nature.com/reviews/genetics
© 20 Macmillan Publishers Limited. All rights reserved 10 | 6f570033c1d07467276ad185bf1521cd736c5f3f | page_0007 |
N. Babkovskaia and J. Poutanen: Water masers in late-type stars 961
the dust is Tas=235 K and Tac=500 K, respectively. Taking
NH2=108cm−3and water fraction fH2O=6×10−3(Jeong
et al. 2003), the resulting gas temperature is T=360 K. From
Fig. 10, we getτm≈8f o r H=1014cm which gives us the
absorption coefficientαm
L=τm/H=−8×10−14cm−1.F o rt h e
projected disk radius of R=16 AU, the coherent length in the
disk is about S=1.2R=3×1014cm. Using Eq. (9), we get
τ=Sαm
L/√
π=−13.5. However, for H=5×1013cm (e.g.
away from the disk central plane), the inversion is larger by a
factor 1.5 (see Fig. 11). This then gives the maximum maser
optical depth of−20 which is larger than the observed lower
limit. Even for a much lower water content fH2O=10−3we still
getτ=−14. Thus, collisional energy exchange with the dust
can provide the gas heating which is necessary for the masers
to operate.
3.5.2. Masers from AGB winds
Water masers from AGB stars are observed in their expand-
ing envelopes. Four low-mass late-type stars (IKTau, U Ori,
RT Vir and U Her) have been mapped by MERLIN re-
cently (Bains et al. 2003). The data show that maser radiation
comes from individual maser clouds with the apparent size of
2−4 AU (Richards et al. 1999) and the filling factor of only
∼0.01. Masers are observed at typical distance from the star of
10−70 AU. The total maser photon production rate (luminos-
ity) is about (1−8)×1042s−1. With the total number of clouds
varying between 14 and 286 depending on a source, one can
estimate a single cloud luminosity of less than 1041s−1.F o ra
spherical cloud with the radius of ∼2.5 AU, the maser emissiv-
ityΦvaries between∼0.2a n d∼0.6c m−3s−1.
Let us consider a gas-dust cloud of radius H=4×1013cm
in a radiation field of the star with T∗=3000 K and W=10−4
(i.e. at a distance of about 50 AU for a 1 AU stellar radius), and
other parameters are the same as in Sect. 3.5.1. The gas temper-
ature depends on the heating which can be provided by colli-
sions with dust. We consider heating (a) due to the gas thermal
motion (see Eq. (23)) and (b) due to the drift of grains through
the gas (see Eq. (31)). If only heating due to thermal motions is
considered, the gas temperature is about 360 K and it lies be-
tween temperatures of the cold and hot dust. The maser optical
depth isτm≈−5 (see dotted curve in Fig. 15). From our model
we also can compute the average maser photon emissivity (see
e.g. Elitzur 1991)
Φ=guAul|∆nul|NH2O|Sul|Km
2, (32)
which is about 2×10−3cm−3s−1, i.e. two orders of magnitudes
lower than observed.
When drift heating corresponding to the velocity of only
Vd=2k m s−1is considered, the gas temperature becomes
much larger than temperatures of the dust. The maser optical
depth increases by a factor of three and the resulting power
by two orders of magnitude reaching Φ= 0.4c m−3s−1.T h i s
increase occurs because maser pumping is the most e ffective
when gas is hotter than the dust (see, for example, BP04).
We can conclude that gas heating by collisions with dust
(due to thermal motion) is not su fficient to produce observedFig. 15. Dependence of the maser optical depth, maser photon emis-
sivityΦ[photon cm−3s−1] and temperatures of the gas and dust on a
height zwithin a slab ( z=0 in the slab center). Solid curves show the
case of the drift heating of the gas given by Eq. (31) and dotted curves
represent the case of heating due to thermal motions (see Eq. (23)).
Dashed and dot-dashed curves represent the temperatures of the as-
tronomical silicate and amorphous carbon, respectiv ely, calculated for
W=10−4andT∗=3000 K. Other parameters are NH2=108cm−3,
fH2O=6×10−3,fd=10−2,H=4×1013cm,a=0.01µm, drift
velocity Vd=2k ms−1.
maser luminosity, while additional drift heating by moving dust
is capable of explaining the masers in AGB winds.
4. Summary
We have considered the maser e ffect in a medium consisting of
a mixture of gas (hydrogen and water vapor) and dust of various
types. The gas and dust temperatures and level populations of
water molecule are calculated self-consistently from the system
of population balance equations and thermal balance equations
for the gas and dust in the radiation field of a late-type star.
When dust of different types is present, the gas interact-
ing with the grains can be heated by one type of dust and is
cooled by another. The gas temperature then takes an interme-
diate value. Radiative cooling by water and the presence of hot
dust strongly influence the water molecule energy level popula-
tions and therefore should be taken into account in calculating
the maser effect. We find that for a small slab thickness Hthe
inversion appears because of the de Jong (1973) mechanism,
while for large H, the maser can be pumped by radiation from
the dust, whose temperature di ffers from that of the gas.
The maser strength depends on the combination of dust
types. When the medium is optically thick to the line radiation, | 10.1051_0004-6361:20053609 | page_0012 |
962 N. Babkovskaia and J. Poutanen: Water masers in late-type stars
the inversion of the 6 16→523maser level populations appears
only for combinations of silicates with carbon (or graphite).
The main cycle of maser pumping is 5 235.5µm−→ 6E
347.6µm−→
74321µm−→ 616. The upward transition in this cycle is dominated
by radiation from the hot dust, while the heat sink is realized byphoton absorption by the cold dust. Combinations of water ice
with any other dust type produces no inversion because of the
gap in the ice absorption coe fficient near 30µm. Thus, masers
operating on the di fference between the gas and ice tempera-
tures (Deguchi 1981; BP04) are extinguished by the hot dustradiation.
Strong masers can also be produced if there is a size distri-
bution of the dust grains. The maser e ffect appears due to the
de Jong mechanism for all discussed dust types, if slab half-
thickness H<∼10
15cm. We find that the maser disappears at
H<∼1011cm, because the gas temperature becomes too low
due to water cooling. For graphite, the inversion exists also at
H>∼1015cm, where the maser pumping cycle appears to be
the same as in the two-dust maser model.
We show that the hot-cold dust model is able to reproduce
the strength of water masers observed from a disk around thecompanion of the carbon star in the binary system V778 Cyg.
However, the masers in the winds of AGB stars require an ad-
ditional source of heating, for example due to friction between
drifting dust grains and the gas.
Acknowledgements. This work was supported by the Magnus
Ehrnrooth Foundation, the Finnish Graduate School for Astronomyand Space Physics (N.B.), and the Academy of Finland (J.P.). We
are grateful to Dmitrii Nagirner for the code computing K-a n d
L-functions, Ryszard Szczerba for p roviding dust absorption coe ffi-
cients and Seppo Alanko for useful discussions.
References
Babkovskaia, N., & Poutanen, J. 2004, A&A, 418, 117 (BP04)Bains, I., Cohen, R. J., Louridas, A., et al. 2003, MNRAS, 342, 8
Bertie, J. E., Labbe, H. J., & Whalley, E. 1969, J. Chem. Phys., 50,
4501
Bolgova, G. T., Strelnitskii, V . S., & Shmeld, I. K. 1977, Soviet
Astronomy, 21, 468
Chandra, S., Kegel, W. H., Varshalovich, D. A., & Albrecht, M. A.
1984a, A&A, 140, 295
Chandra, S., Varshalovich, D. A., & Kegel, W. H. 1984b, A&AS, 55,
51
Collison, A. J., & Watson, W. D. 1995, ApJ, 452, L103
Cooke, B., & Elitzur , M. 1985, ApJ, 295, 175Danchi, W. C., Bester, M., Degiacomi, C. G., Greenhill, L. J., &
Townes, C. H. 1994, AJ, 107, 1469
David, P., & Pegourie, B. 1995, A&A, 293, 833de Jong, T. 1973, A&A, 26, 297
de Jong, T. 1977, A&A, 55, 137
Deguchi, S. 1977, PASJ, 29, 669Deguchi, S. 1981, ApJ, 249, 145
Downes, D., Genzel, R., Becklin, E. E., & Wynn-Williams, C. G.
1981, ApJ, 244, 869
Elitzur, M. 1991, Astronomical Mase rs (Dordrecht: Kluwer Academic
Publishers)
Engels, D. 1994, A&A, 285, 497
Engels, D., & Leinert, C. 1994, A&A, 282, 858Goldreich, P., & Kwan, J. 1974, ApJ, 191, 93
Goldreich, P., & Scov ille, N. 1976, ApJ, 205, 144
Goldsmith, P. F., & Langer, W. D. 1978, ApJ, 222, 881Green, S. 1980, ApJS, 42, 103
Green, S., Maluendes, S., & McLean, A. D. 1993, ApJS, 85, 181
Groenewegen, M. A. T. 1994, A&A, 290, 531Hartquist, T. W., Dalgarno, A., & Oppenheimer, M. 1980, ApJ, 236,
182
Hollenbach, D., & McKee, C. F. 1979, ApJS, 41, 555Hudgins, D. M., Sandford, S. A., Allamandola, L. J., & Tielens,
A. G. G. M. 1993, ApJS, 86, 713
H u m p h r e y s ,E .M .L . ,Y a t e s ,J .A . ,G r a y ,M .D . ,F i e l d ,D . ,&B o w e n ,
G. H. 2001, A&A, 379, 501
Jeong, K. S., Winters, J. M., Le Bertre, T., & Sedlmayr, E. 2003, A&A,
407, 191
Jura, M. 1996, ApJ, 472, 806
Kegel, W. H. 1975, A&A, 44, 95
Laor, A., & Draine, B. T. 1993, ApJ, 402, 441Mathis, J. S., Rumpl, W., & Nordsieck, K. H. 1977, ApJ, 217, 425
Neufeld, D. A., Lepp, S., & Melnick, G. J. 1995, ApJS, 100, 132
Neufeld, D. A., & Melnick, G. J. 1987, ApJ, 322, 266Richards, A. M. S., Yates, J. A., & Cohen, R. J. 1999, MNRAS, 306,
954
Rosen, B. R., Moran, J. M., Reid, M. J., et al. 1978, ApJ, 222, 132Rouleau, F., & Martin, P. G. 1991, ApJ, 377, 526
Shakura, N. I., & Sunyaev, R. A. 1973, A&A, 24, 337
Sobolev, V . V . 1960, Moving envelopes of stars (Cambridge: Harvard
University Press)
Strelnitskij, V . S. 1977, Soviet Astronomy, 21, 381
Szczerba, R., Szymczak, M., Babkovskaia, N., et al. 2005, A&A,
submitted [ arXiv:astro-ph/0504354 ]
Tielens, A. G. G. M., & Hollenbach, D. 1985, ApJ, 291, 722
Toth, R. A. 1991, J. Opt. Soc. Am. B, 8, 2236Wallin, B. K., & Wats on, W. D. 1997, ApJ, 476, 685
Yamamura, I., Dominik, C., de Jong, T., Waters, L. B. F. M., &
Molster, F. J. 2000, A&A, 363, 629
Yates, J. A., Field, D., & Gray, M. D. 1997, MNRAS, 285, 303 | 10.1051_0004-6361:20053609 | page_0013 |
ARTICLE
Received 18 Mar 2015 |Accepted 23 Apr 2015 |Published 24 Jun 2015
Imaging an aligned polyatomic molecule
with laser-induced electron diffraction
Michael G. Pullen1,*, Benjamin Wolter1,*, Anh-Thu Le2, Matthias Baudisch1, Michae ¨l Hemmer1, Arne Senftleben3,
Claus Dieter Schro ¨ter4, Joachim Ullrich4,5, Robert Moshammer4, C.D. Lin2& Jens Biegert1,6,7
Laser-induced electron diffraction is an evolving tabletop method that aims to image ultrafast
structural changes in gas-phase polyatomic molecules with sub-Ångstro ¨m spatial and
femtosecond temporal resolutions. Here we demonstrate the retrieval of multiple bondlengths from a polyatomic molecule by simultaneously measuring the C–C and C–H bondlengths in aligned acetylene. Our approach takes the method beyond the hitherto achievedimaging of simple diatomic molecules and is based on the combination of a 160 kHzmid-infrared few-cycle laser source with full three-dimensional electron–ion coincidence
detection. Our technique provides an accessible and robust route towards imaging ultrafast
processes in complex gas-phase molecules with atto- to femto-second temporal resolution.DOI: 10.1038/ncomms8262 OPEN
1ICFO-Institut de Ciencies Fotoniques, Mediterranean T echnology Park, Castelldefels (Barcelona) 08860, Spain.2J. R. Macdonald Laboratory, Department of
Physics, Kansas State University, Manhattan, Kansas 66506-2604, USA.3Universita ¨t Kassel, Institut fu ¨r Physik und CINSaT, Heinrich-Plett-Strasse 40,
Kassel 34132, Germany.4Max-Planck-Institut fu ¨r Kernphysik, Saupfercheckweg 1, Heidelberg 69117, Germany.5Physikalisch-Technische Bundesanstalt
(PTB), Bundesallee 100, Braunschweig 38116, Germany.6Department of Physics and Astronomy, University of New Mexico, 1919 Lomas Boulevard NE,
Albuquerque, New Mexico 87131, USA.7ICREA-Institucio ´Catalana de Recerca i Estudis Avanc ¸ats, Barcelona 08010, Spain. * These authors contributed
equally to this work. Correspondence and requests for materials should be addressed to M.G.P. (email: michael.pullen@icfo.eu).
NATURE COMMUNICATIONS | 6:7262 | DOI: 10.1038/ncomms8262 | www.nature.com/naturecommunications 1
&2015 Macmillan Publishers Limited. All rights reserved. | 10.1038_ncomms8262 | page_0000 |
Dynamic imaging of chemical reactions or biological
functions is one of the grand challenges of science1,2.
These processes are typically triggered by sub-Ångstro ¨m-
scale events that are initiated on the few-femtosecond timescale.
An imaging method that could achieve the required
spatiotemporal resolutions would provide revolutionary insights
into the connection between molecular structure at critical
transition points and barrier heights; hallmark examples are
transition states3, rapid dynamics in the vicinity of conical
intersections4or proton migration and isomerization5. The
capability of imaging the motions of the atomic constituents
during these processes with few-femtosecond temporal and sub-
Ångstro ¨m spatial resolutions therefore represents a paradigm
shift in modern physics and chemistry.
Ultrafast electron diffraction (UED) is capable of resolving
atomic positions with sub-Ångstro ¨m resolution6, however, the
achievable temporal resolution is currently limited to hundreds of
femtoseconds mainly due to Coulomb repulsion in the electron
bunch. Such temporal resolution is not sufficient to resolve the
initiation reactions and ultrafast changes of the prototypical
processes mentioned above. Current developments therefore aim
at reducing space charge7or using relativistic electron bunches8.
X-ray diffraction methods9currently suffer from spectrotemporal
jitter and are only available at large-scale facilities. These
restraints have motivated the development of new dynamical
imaging techniques, largely for the gas phase, such as chirped-
encoded recollisions10, photoelectron holography11, femtosecond
photoelectron diffraction12, Coulomb explosion imaging13and
laser-assisted electron diffraction14.
Clearly, versatile laboratory-scale tabletop methods that
provide the combined spatial and temporal resolutions would
signify a breakthrough, especially for the imaging of gas-phase
molecular dynamics. Laser-induced electron diffraction (LIED) is
such a method and is based on probing an objects structure using
its own electrons that are rescattered during strong-field-induced
recollisions15–17. This process is depicted in Fig. 1 where the
longitudinal and transverse momenta are defined as k||¼kyand k?¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffi
k2
xþk2
zp
, respectively. Coherent subcycle elastic
scattering of the electron wavepacket with attosecond (single
pulse) to femtosecond (pulse train) resolution retains structural
information of the ionic species in the resultant diffraction
pattern18–21. The challenge lies in simultaneously fulfilling the
extremely stringent conditions for LIED in order to extract
structural information; these are as follows: (i) achieving high
recollision energies despite a small fraction of target ionization,
(ii) achieving core penetrating collisions and sufficient
momentum transfer with the scattered electron, (iii) driving
recollision in the quasi-static regime to enable extraction of field-
free diffraction data from the photoelectron momentum spectra.
When these conditions are met, the method of molecular
structure retrieval is similar to conventional electron or X-ray
diffraction, with the added benefit of femtosecond temporal
resolution of the driving laser. In general, state of the art near-
infrared lasers cannot fulfil these combined conditions, however,
investigations have still been undertaken for homonuclear
diatomic molecules such as O 2(ref. 19). Recently, these
conditions were satisfied with B2mm lasers and structural
retrieval of N 2and O 2molecules was demonstrated22,23. Spatial
resolutions of 0.05 Å were reported, which were sufficient to
image a 0.1 Å contraction of the simple O 2molecule during the
B5 fs it takes an electron to rescatter. This result established the
potential of LIED as a dynamical imaging technique with sub-
Ångstro ¨m spatial and few-femtosecond temporal resolutions.
To harness the combined temporal and spatial resolutions of
LIED and apply it to polyatomic molecules (that is, systems with
three or more atoms that exhibit full prototypical molecular
dynamics) requires addressing a decisive and unresolved issue,
namely the fact that launching the recollision (imaging) electron
initiates molecular distortion and eventually fragmentation.
Therefore, a certain portion of the detected electrons serves as
an unwanted background that can make imaging difficult or even
impossible. This problem can be resolved through ion–electron
coincidence detection and the retrieval of the doubly differential
cross-section. This ensures unambiguous imaging of the mole-
cular structure, or fragments, of interest. In addition to this major
concern, there are other experimental obstacles that must be
overcome. First, complex molecules commonly have ionization
energies around and below 10 eV, which necessitates the use of
mid-infrared driving lasers in order to avoid ionization satura-
tion. Mid-infrared sources also have the added benefit that
electrons with the required energies are liberated at lower
intensities, which results in less distortion of the molecule.
Second, because each constituent atom has a unique scattering
cross-section, a careful selection of the electron-scattering
parameters ensures that they all contribute significantly to the
scattering and hence facilitates the simultaneous determination of
multiple bond lengths. Third, to resolve the increased structural
complexity, it is highly beneficial for the gas ensemble to be
anisotropically distributed with respect to the molecular axis in
order to remove averaging effects20.
Here we meet all of these challenges through a combination of
experimental methodologies. A unique home-built optical para-
metric chirped pulse amplification (OPCPA) source provides
1.7mm and 3.1 mm pulses at a repetition rate of 160 kHz (ref. 24)
and with excellent long-term stability. The 1.7 mm light is used to
impulsively align the target molecule, while the 3.1 mm light
induces electron rescattering. The lower efficiency of the
rescattering process25,26at longer wavelengths is more than
compensated for by the two orders of magnitude higher
repetition rate of our source compared with typical 1 kHz
systems. Equally as important is the reaction microscope
(ReMi) detection system that allows a careful selection of the
relevant channels (over both electron energies and scattering
Molecular jet E
B
Ion
detector
Helmholtz coilsAlignEmit
Image
/afii9835y
zxElectron
detector
/afii9825BW negative
BW negative
BW negative30
31
32
30
30A
31
31A
32
Figure 1 | Laser-induced electron diffraction from aligned C 2H2molecules
using a mid-infrared OPCPA source and a reaction microscope. The
cartoon film shows the procedure. ( a) The C 2H2molecules are pre-aligned
by focusing the 1.7 mm pump pulse (blue) into a molecular jet. ( b) The
3.1mm pulse (red) is used to generate high-energy electrons that
subsequently rescatter off the parent ion. ( c) The rescattered electrons
carry structural information of the parent ion that is contained in the
detected angular momentum distributions. The anticollinear electric ( E)
and magnetic ( B) fields guide the charged fragments towards opposing
position-sensitive detectors.ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms8262
2 NATURE COMMUNICATIONS | 6:7262 | DOI: 10.1038/ncomms8262 | www.nature.com/naturecommunications
&2015 Macmillan Publishers Limited. All rights reserved. | 10.1038_ncomms8262 | page_0001 |
angles) from the doubly differential cross-section in
coincidence27. We validate our unique experimental approach
by simultaneously imaging the C–C and C–H bond lengths of
aligned polyatomic molecule acetylene (C 2H2). This establishes
LIED as a methodology for dynamically visualizing larger and
heteronuclear molecular structures. We chose acetylene as the test
molecule since it is heteronuclear, readily alignable, linear and
symmetric so that orientation is not required, and its bond
lengths are accurately known. More importantly, however, is the
fact that acetylene is a prototypical organic molecule in which the
dynamics associated with isomerization, proton migration,
internal vibrational redistribution of energy and conical
intersections can be studied in the future using LIED. The
measured bond lengths lie within o5% of the expected acetylene
cation equilibrium distances of 1.25 and 1.08 Å (ref. 28),
respectively, for both molecular alignments.
Results
Extraction of molecular structure . The procedure for extracting
structural information from aligned (see Supplementary Note 1
and Supplementary Figs 1 and 2) C 2H2using LIED is outlined in
Fig. 2. Figure 2a shows the momentum distribution of all
electrons detected in coincidence with all positive fragments
after the ionization of C 2H2with our mid-infrared source.
Following the quantitative rescattering theory (see Supplementary
Note 2), the molecular differential cross-section (DCS) is
extracted by sweeping the scattering angle ( yr) around the cir-
cumference of a circle with radius equal to the momentum of the
rescattered electron ( kr). The influence of the ionizing laser fieldmust be considered; consequently, the origin of the circle is given
by the vector potential ( Ar) at the time of rescattering (see the
Supplementary Note 3 and Supplementary Fig. 3). Each circle
represents rescattering by different electron energies. The
extracted experimental molecular DCS ( sM) is combined with
the theoretical atomic DCS ( sA), which is calculated using the
independent atom model (see the Supplementary Note 4), for the
same electron energy and emission angle to calculate the mole-
cular contrast factor (MCF) MF¼(sM–sA)/sA. The MCFs are
typically presented as a function of the momentum transfer
q¼2krsin(yr/2) experienced by the rescattered electrons.
Aw2-based fitting routine is used to compare the experimentally
obtained MCF to theoretical predictions (see the Supplementary
Note 5 and Supplementary Fig. 4).
Full-particle coincidence detection in three dimensions . The
coincidence detection capability of the ReMi is crucial for accu-
rate retrieval of polyatomic molecular structure from the
experimental MCF. To develop the time-resolving capabilities of
LIED, it is important that we ensure the scattering pattern
originates from the fragmentation channel of interest only. To
highlight this point we present the time-of-flight (TOF) spectrum
of all the detected positively charged fragments in Fig. 2b. The
main peak near 4.2 ms is the acetylene cation (C 2H2þ) investigated
in this manuscript, and it constitutes B10% of the total number
of detected fragments. The inset shows a close-up of this peak and
the black-shaded area represents the region that the electrons
associated with C 2H2þare extracted from. Many other fragments
can be observed and identified in the TOF and each of these peaks
4a
4 442
2 220
0 00–2
–4
–6 66
–4 –2 1 3 5k⊥ (a. u.)
K/p35 (a. u.)106106
105105
104
103
102
101107
106
105
104
103
102
101
100
0 100 200 300 400 500
Ion counts
Electron ener gy (eV)3.2 3.7 4.2 4.7 5.2 5.7
Momentum transfer (Å–1)Time-of-flight ( μs)
MCF0.3
0.2
0.1
0.0
–0.1
–0.2
–0.3All electrons All electrons
C2H2+ electrons C2H2+ electrons
C2H2+ fit
–10% fit
+10% fit4.1 4.2 4.3
H+H2+C2H22+C2H2+
C2H+
CH2+
C2+C+50 eV50 eV
100 eV 100 eV
ArAr/afii9835r/afii9835rkrkrC2H2+Electrom countsb
cd
Figure 2 | Method to extract structural information from the momentum distributions. (a) Logarithmically scaled momentum distribution of electrons
corresponding to all ionic fragments. The circles represent the scattering of electrons with the same energy at different angles. ( b) The detected ion TOF
showing the numerous fragments created during the strong-field interaction. The inset shows the peak corresponding to the C 2H2þion near 4.2 ms and the
shaded region represents the window of ions that the C 2H2þelectrons are taken from. ( c) The electron kinetic energy distribution for the C 2H2þion (black)
and for all possible fragmentation processes (blue). ( d) An extracted MCF for the acetylene cation (black circles) as well as for electrons from all fragments
(blue squares). The solid black curve shows the best fit, which matches very well with the cation channel. The MCFs for ±10% changes in the C 2H2
molecular lengths (dashed curves) highlight the sensitivity of the LIED technique. The s.d. error bars are derived from Poissonian statistics.NATURE COMMUNICATIONS | DOI: 10.1038/ncomms8262 ARTICLE
NATURE COMMUNICATIONS | 6:7262 | DOI: 10.1038/ncomms8262 | www.nature.com/naturecommunications 3
&2015 Macmillan Publishers Limited. All rights reserved. | 10.1038_ncomms8262 | page_0002 |
has associated electrons. Figure 2c shows the measured electron
kinetic energy spectrum for all fragments (blue) and for the
C2H2þfragment only (black). An order of magnitude difference
in the number of detected electrons is visible over the entire
spectral range (see Methods and Supplementary Fig. 5). It is these
omnipresent extra electrons that serve as an unwanted back-
ground signal and are detrimental to structure retrieval without
coincidence detection. Figure 2d summarizes this decisive point
by comparing the MCFs retrieved when analysing electrons
corresponding to all fragments (blue) and from C 2H2þonly
(black). The C 2H2þdata result in an MCF that compares well
with the equilibrium acetylene structure, which indicates that, in
the case of acetylene, launching the recollision electron does not
cause detrimental differences between the neutral and ionic
species within the short recollision time. On the other hand, using
electrons from all fragmentation channels results in a dramati-
cally different MCF that cannot be accurately fitted (blue squares)
and fails in retrieving the C 2H2þbond lengths. We validate with
this analysis that electron–ion coincidence detection is a pre-
requisite for the application of LIED to larger molecules. The high
sensitivity of LIED to the exact molecular structure is also illu-
strated in Fig. 2d by the dramatic change induced in the MCF by
a 10% contraction (green) or expansion (red) of the molecule.
Simultaneous measurement of multiple bond lengths .T o
visualize complex molecules, LIED needs to be able to retrieve
multiple bond lengths between different atomic species. We
demonstrate that our implementation of LIED fulfills this pro-
mise and is even able to image the (typically) elusive hydrogen
atom by exploiting the fact that we measure the full doubly
differential cross-section. At the tens of keV electron energies
used in UED, a hydrogen atom has a scattering cross-section ( sH)
that is typically much less than that of a C atom ( sC). Figure 3a
presents the sH/sCratio29for a 25 keV electron as a function of
the scattering angle (green curve). The ratio is maximal for
scattering angles between 0 and 5 /C176that are typical in UED
(shaded region) but even in this range sH/sCo0.05. The electron
energies used in LIED result in much higher values of the sH/sC
ratio, as is also presented in Fig. 3a for 50 eV (red curve) and
100 eV (blue curve) electrons, because of a minimum in the C
atom differential cross-section. Both of the presented energies
have wide angular regions where sH/sC40.10 and a peak of
sH/sC40.50 is observed near 80 /C176for 50 eV electrons. The shaded
regions represent the much wider scattering angles for which the
LIED technique is valid.To take advantage of the favourable cross-section ratio
available to LIED, we confirm that we can simultaneously
measure both the C–H and C–C bonds. The MCFs extracted for
both molecular alignments after scattering of 60 eV electrons
are presented in Fig. 3b. For aligned molecules (blue squares)
the best theoretical fit (dashed blue curve) from the w2-fitting
routine results in bond lengths of DA;60
CC¼1:28/C60:13˚A and
DA;60
CH¼1:04/C60:10˚A, while for anti-aligned molecules
(red circles) the same procedure results in estimates of
DAA ;60
CC¼1:33/C60:13˚A and DAA ;60
CH¼1:15/C60:12˚A. Here the
notation Dalignment ;energy
bondis used for the results of the individual fits
to refer to energy-specific bond lengths. The estimated bond
lengths agree well with the known values28, and the accuracy of
each fit is B10 pm, which is an order of magnitude lower than
the de Broglie wavelength of the scattering electrons ( lE¼1.3 Å).
The positions of the MCF extrema and zero crossings, as well as
the peak-to-peak modulation, are very sensitive to changes in the
bond lengths and the molecular alignment. It is the sensitivity to
these parameters that is utilized to monitor sub-Ångstro ¨m
changes in molecular structure. The two MCFs presented in
Fig. 3b show some differences such as the position of the
minimum near q¼3.5 Å/C01, which is closer to zero for anti-
aligned acetylene, and the modulation amplitude, which is smaller
in the aligned case. Depending on the target and the degree to
which it is aligned, molecular alignment or anti-alignment can
lead to larger differences in the peak-to-peak amplitude of the
MCFs, which is beneficial for structural imaging. These results
confirm that LIED can simultaneously extract multiple bond
lengths from complex polyatomic molecules with high accuracy.
Temporal resolution of LIED . Next, we illustrate the possible
attosecond temporal resolution30of the technique in Fig. 4. We
measure the doubly differential cross-section, which permits
retrieving the C–C and C–H bond lengths as a function of the
rescattering electron energy. On the basis of operating mid-
infrared LIED in the quasi-static limit we can invoke the classical
rescattering model to associate a specific time to the measured
electron-rescattering energy. The top axis in Fig. 4 shows the
corresponding return time for each electron energy and indicates
that a temporal resolution below 100 as could be achieved by
analysing at different rescattering energies. We further elaborate
that the measured energy range can also be used to establish an
unprecedented level of confidence and redundancy for the
retrieved bond length. The extracted DAA
CC,DAA
CH,DA
CCand DA
CH
values are consistent with the estimated ionic equilibrium values
100
10–1
10–2
0 30 60 90 120 150 180Cross-section ratio ( /afii9846H//afii9846C)
Scattering angle( °)LIEDab
50 eV
100 eV
UED
25 keV
2.5 3.0 3.5 4.0 4.5 5.0
Momentum transfer (Å–1)5.5 6.0Anti-aligned
Fit
FitAligned0.3
0.2
0.1
0.0
–0.1
–0.2
–0.3
–0.4
MCF
Figure 3 | Simultaneous extraction of multiple bond lengths from polyatomic molecules. (a) The ratio of the H and C scattering cross-sections as a
function of electron-scattering angle for typical energies used in LIED (50 and 100 eV) and CED/UED (25 keV). The ratios are much higher for the energie s
relevant to LIED and are also applicable over a much wider angular range (shaded regions). ( b) Blue squares (red circles) show the experimental molecular
contrast factor that results from the scattering of 60 eV electrons by aligned (anti-aligned) molecules. The best theoretical fits (dashed lines) all ow the
accurate extraction of the C–H and C–C bond lengths from both alignments. The s.d. error bars are derived from Poissonian statistics.ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms8262
4 NATURE COMMUNICATIONS | 6:7262 | DOI: 10.1038/ncomms8262 | www.nature.com/naturecommunications
&2015 Macmillan Publishers Limited. All rights reserved. | 10.1038_ncomms8262 | page_0003 |
(dashed black lines)28over the investigated energy range. As no
significant structural rearrangements are expected after acetylene
is ionized from a neutral to a cation28, fitting a horizontal line to
the energy-dependent bond length estimates will yield an overall
estimate of the C–C and C–H bond lengths. This fitting results in
estimates of RAA
CC¼1:24/C60:04˚A and RAA
CH¼1:10/C60:03˚A for
anti-aligned molecules while the same analysis with aligned
molecules results in bond lengths of RA
CC¼1:26/C60:04˚A and
RA
CH¼1:05/C60:03˚A. This method amounts to performing two-
dimensional fitting over both electron energy and scattering
angle, which is not possible with other techniques, and highlights
the accuracy of the LIED method.
Discussion
In summary, we demonstrate a robust method for the retrieval of
multiple bond lengths from an aligned polyatomic molecule with
mid-infrared LIED, which we validate by accurately determining
the structure of acetylene. The use of a ReMi in combination with
a home-built 160 kHz mid-infrared OPCPA exploits coincidence
detection together with the measurement of the doubly differential
elastic scattering cross-section in the quasi-static regime. This
unique capability enables imaging the hydrogen atom by selection
of a suitable scattering energy range for which the relative cross-
sections contribute comparably. An excellent bond length
confidence level is achieved due to the large range of rescattering
energies for which structural information is measured. Our
method demonstrates a clear path to exploit the intrinsic atto-
to femto-second temporal resolution of LIED for the imaging of
complex molecules. Finally, we provide a solution to selective
imaging of the multiple fragmentation pathways that are
inherently created when launching the recollision electron in
polyatomic molecules. This capability is an enabling step towards
time-resolved imaging and permits accurate retrieval of the
geometrical structure from the fragment of interest only. Our data
already contain structural information on the ultrafast isomeriza-
tion and deprotonation of acetylene and on multiply charged ions,
which we aim to investigate in future work. The technique
provides an accessible and robust route towards probing ultrafastprocesses in complex gas-phase molecules by combining attose-
cond and collision physics towards realising the molecular movie.
Methods
Mid-infrared OPCPA source .The OPCPA-based source used in this work has
been presented previously24. It provides 6.5-cycle (70 fs) mid-infrared pulses at a
repetition rate of 160 kHz. The 3.1 mm radiation is derived from difference
frequency generation of 1.55 and 1.05 mm pulses in a magnesium oxide doped
periodically poled lithium niobate (MgO:PPLN) crystal. It is subsequently chirped
and parametrically amplified in four cascaded OPA stages before the pulse is
compressed to 70 fs in a grating compressor. After focussing with a 50 mm
parabolic mirror an estimated peak intensity of 5.5 /C21013Wc m/C02was reached.
This corresponds to a ponderomotive energy of UP¼50 eV and a Keldysh
parameter of g¼0.34 for C 2H2, which has an ionization potential of 11.4 eV.
Owing to the quadratic scaling of the maximum rescattering electron energy with
laser wavelength ( Emaxpl2), our laser can generate much more energetic electrons
compared with a ubiquitous 800 nm Ti:Sapphire laser. These electrons are a basic
requirement for LIED as they penetrate deep into the core of the molecule, thereby
revealing structural information. The low mid-infrared photon energy also ensures
that complex molecular targets, which typically have low ionization energies, are
not in the ionization saturation regime for the required high intensities. The high
160 kHz repetition rate ensures a high data accumulation rate and more than
compensates for the lower mid-infrared rescattering probability compared with the
typical 800 nm, 1 kHz systems. The phase-coherent signal output of the OPCPA at
1.7mm is utilized to induce impulsive molecular alignment, as is discussed below.
This radiation has a pulse duration of B98 fs and also operates at a repetition rate
of 160 kHz. Both outputs have high stability with power fluctuations less than a
percent being typical over the course of the data acquisition period.
Reaction microscope detection system .We utilize a ReMi detection system to
detect the high-energy rescattered electrons. For a thorough overview of the func-
tion and capabilities of ReMis see ref. 27. A cold and thin molecular jet is formed by
supersonically expanding gas into vacuum and subsequent skimming. The gas is
ionized in the interaction region and the resultant charged particles are guided
towards opposing position-sensitive microchannel plate detectors by homogenous
electric ( E) and magnetic ( B) fields. Momentum distributions of both ions and
electrons are then extracted from the position and time of detection. To detect high-
energy electrons in three dimensions, fields of E¼51 V cm/C01andB¼39 G are
chosen. The scaling of momentum resolution ( qk) with field strengths has been
discussed previously27, and our calculations show that the momentum resolutions
are comparable to the momentum integration ranges ( Dk) used to obtain the
angularly resolved DCSs (see Supplementary Fig. 3 for more information), that is,
qkEDk. For example, a 50 eV electron-scattering at an angle of yr¼50/C176has a
calculated momentum resolution of Bqk¼±15.3%, which is almost the same as
the integration range of Dk¼±15% used at that point. Therefore, the detection
resolution limit of the ReMi is not a limiting factor to this work.
Rescattering time (fs)
9.10a
8.95 8.80 8.65 8.50 8.35 9.10 8.95 8.80 8.65 8.50 8.35
1.7
1.5
1.3
1.1
0.9
0.7Bond length (Å)1.7
1.5
1.3
1.1
0.9
0.7
50 60 70 80 90 100 50 60 70 80 90 100
Electron energy (eV)RRAA
= 1.10 = 1.10 ± ± 0.03 Å0.03 ÅRAA
= 1.24 = 1.24 ± ± 0.04 Å0.04 Å RA
CCCC = 1.26 = 1.26 ± ± 0.04 Å0.04 Å
RA
CHCH = 1.05 = 1.05 ± ± 0.030.03 ÅHH HHHH
HH
HH
HHHH HH CC CCCCC
CC
CCCCCC CC E ECC
CHb
cd
Figure 4 | Accurate C 2H2bond length extraction. The C–C (C–H) bond length estimates are presented as a function of the scattering electron energy and
rescattering time in the top (bottom) quadrant. The expected equilibrium values of the acetylene cation are also shown (dashed black lines). The valu es of
the best horizontal fits for each bond are displayed in the respective panels. See Supplementary Fig. 4 for details about the bond length estimate error bars.NATURE COMMUNICATIONS | DOI: 10.1038/ncomms8262 ARTICLE
NATURE COMMUNICATIONS | 6:7262 | DOI: 10.1038/ncomms8262 | www.nature.com/naturecommunications 5
&2015 Macmillan Publishers Limited. All rights reserved. | 10.1038_ncomms8262 | page_0004 |
ReMis possess a number of decisive benefits over typical TOF spectrometers for
LIED experiments, such as the direct extraction of the full three-dimensional
electron momentum distribution and the capability of detecting electrons incoincidence with ions, which allows a simple removal of unwanted electron counts
in post-processing. Figure 2c in the main text presents electron energy spectra for
electrons associated with all of the detected ionic fragments (blue) and for electrons
associated with the C
2H2þion only (black). A large difference between the two
curves can be observed for the entire spectral range. It is important not to includethese unwanted electrons (which originate from different ionization and
fragmentation processes) in the data analysis as they interfere with bond length
determination.
Molecular fragmentation.In Supplementary Fig. 5a,b we present the measured
ionic TOFs for the simple diatomic molecule O 2and for the polyatomic molecule
C2H2, respectively. The O 2TOF shows a clear lack of fragments with only the
single and double ions significantly contributing, while the C 2H2TOF is full of
other ionic fragments. In Supplementary Fig. 5c,d the electrons corresponding to
the main single ion (black curves) for each TOF are present along with the elec-
trons corresponding to all ionic fragments (coloured curves). An order of mag-
nitude difference is observed over the entire spectrum in the case of C 2H2, while in
the O 2case the single ion electrons make up 480% of those detected. This is the
reason that electron–ion coincidence detection apparatuses are required to perform
LIED on polyatomic molecules.
References
1. Zewail, A. H. & Thomas, J. M. 4D Electron Microscopy: Imaging in Space and
Time (Imperial College Press, 2009).
2. Chapman, H. N. X-ray imaging beyond the limits. Nat. Mater. 8,299–301
(2009).
3. Zhong, D. & Zewail, A. H. Femtosecond real-time probing of reactions. 23.
studies of temporal, velocity, angular, and state dynamics from transition states
to final products by femtosecond-resolved mass spectrometry. J. Phys. Chem. A
102, 4031–4058 (1998).
4. Domcke, W. & Yarkony, D. R. Role of conical intersections in molecular
spectroscopy and photoinduced chemical dynamics. Annu. Rev. Phys. Chem.
63,325–352 (2012).
5. Jiang, Y. H. et al. Ultrafast extreme ultraviolet induced isomerization of
acetylene cations. Phys. Rev. Lett. 105, 263002 (2010).
6. Ihee, H. et al. Direct imaging of transient molecular structures with ultrafast
diffraction. Science 291, 458–462 (2011).
7. Lahme, S., Kealhofer, C., Krausz, F. & Baum, P. Femtosecond single-electron
diffraction. Struct. Dyn. 1,034303 (2014).
8. Sciaini, G. & Miller, R. J. D. Femtosecond electron diffraction: heralding the era
of atomically resolved dynamics. Rep. Prog. Phys. 74,096101 (2011).
9. Ku ¨pper, J. et al. X-ray diffraction from isolated and strongly aligned gas-phase
molecules with a free-electron laser. Phys. Rev. Lett. 112, 083002 (2014).
10. Baker, S. et al. Probing proton dynamics in molecules on an attosecond time
scale. Science 312, 424–427 (2006).
11. Krasniqi, F. et al. Imaging molecules from within: ultrafast angstro ¨m-scale
structure determination of molecules via photoelectron holography using free-electron lasers. Phys. Rev. A 81,033411 (2010).
12. Boll, R. et al. Femtosecond photoelectron diffraction on laser-aligned
molecules: towards time-resolved imaging of molecular structure. Phys. Rev. A
88,061402(R) (2013).
13. Xu, H., Okino, T. & Yamanouchi, K. Tracing ultrafast hydrogen migration in
allene in intense laser fields by triple-ion coincidence momentum imaging.
J. Chem. Phys. 131, 151102 (2009).
14. Morimoto, Y., Kanya, R. & Yamanouchi, K. Laser-assisted electron
diffraction for femtosecond molecular imaging. J. Chem. Phys. 140, 064201
(2014).
15. Zuo, T., Bandrauk, A. D. & Corkum, P. B. Laser-induced electron diffraction: a
new tool for probing ultrafast molecular dynamics. Chem. Phys. Lett. 259,
313–320 (1996).
16. Lein, M., Marangos, J. P. & Knight, P. L. Electron diffraction in above-threshold
ionization of molecules. Phys. Rev. A 66,051404(R) (2002).
17. Lin, C. D., Le, A. T., Chen, Z., Morishita, T. & Lucchese, R. Strong-field
rescattering physics-self-imaging of a molecule by its own electrons. J. Phys. B
43,122001 (2010).
18. Ray, D. et al. Large-angle electron diffraction structure in laser-induced
rescattering from rare gases. Phys. Rev. Lett. 100, 143002 (2008).19. Meckel, M. et al. Laser-induced electron tunneling and diffraction. Science 320,
1478 (2008).
20. Xu, J., Chen, Z., Le, A. T. & Lin, C. D. Self-imaging of molecules from
diffraction spectra by laser-induced rescattering electrons. Phys. Rev. A 82,
033403 (2010).
21. Okunishi, M., Niikura, H., Lucchese, R. R., Morishita, T. & Ueda, K. Extracting
electron-ion differential scattering cross sections for partially aligned molecules
by laser-induced rescattering photoelectron spectroscopy. Phys. Rev. Lett. 106,
063001 (2011).
22. Blaga, C. I. et al. Imaging ultrafast molecular dynamics with laser-induced
electron diffraction. Nature 483, 194–197 (2012).
23. Xu, J. et al. Diffraction using laser-driven broadband electron wave packets..
Nat. Commun. 5,4635 (2014).
24. Thai, A., Hemmer, M., Bates, P. K., Chalus, O. & Biegert, J. Sub-250-mrad,
passively carrier-envelope-phase-stable mid-infrared OPCPA source at high
repetition rate. Opt. Lett. 36,3918–3920 (2011).
25. Tate, J. et al. Scaling of wave-packet dynamics in an intense midinfrared field.
Phys. Rev. Lett. 98,013901 (2007).
26. Austin, D. R. & Biegert, J. Strong-field approximation for the wavelength
scaling of high-harmonic generation. Phys. Rev. A 86,023813 (2012).
27. Ullrich, J. et al. Recoil-ion and electron momentum spectroscopy: reaction-
microscopes. Rep. Prog. Phys. 66,1463–1545 (2003).
28. Dopfer, O., Olkhov, R. V., Mladenovic ´, M. & Botschwina, P. Intermolecular
interaction in an open-shell p-bound cationic complex: IR spectrum and coupled
cluster calculations for C 2H2þ-Ar. J. Chem. Phys. 121, 1744–1753 (2004).
29. Jablonski, A., Salvat, F. & Powell, C. J. NIST Electron Elastic-Scattering Cross-
Section Database, Version 3.2, SRD 64 (National Institute of Standards and
Technology, 2010).
30. Corkum, P. B. & Krausz, F. Attosecond science. Nat. Phys. 3,381–387 (2007).
Acknowledgements
We acknowledge support from the Spanish Ministerio De Economia Y Competitividad
(MINECO) through ‘Plan Nacional’ (FIS2011-30465-C02-01) and the Catalan Agenciade Gestio ´d’Ajuts Universitaris i de Recerca (AGAUR) with SGR 2014–2016. This
research has been supported by Fundacio ´Cellex Barcelona, LASERLAB-EUROPE grant
agreement 228334 and COST Action MP1203. B.W. was supported by AGAUR with a
PhD fellowship (FI-DGR 2013–2015). M.G.P. is supported by the ICFONEST þpro-
gramme, partially funded by the Marie Curie cofunding of Regional, National andInternational Programmes—COFUND (FP7-PEOPLE-2013- COFUND) action of theEuropean Commission, the ‘Severo Ochoa’ Program of the Spanish Ministry of Economy
and Competitiveness and ICFO. A.-T.L. and C.D.L. are supported by the Chemical
Sciences, Geosciences, and Biosciences Division, Office of Basic Energy Sciences, Officeof Science, U.S. Department of Energy (DOE) under Grant No. DE-FG02-86ER13491.We thank Dr Michele Sclafani for helpful discussions.
Author contributions
J.B. conceived the experimental investigation. M.G.P., B.W., M.B. and M.H. acquired the
experimental data. A.-T.L. and C.D.L. provided theoretical support. A.S., C.D.S., J.U. and
R.M. provided experimental support. M.G.P., B.W., A.-T.L., C.D.L. and J.B. wrote themanuscript.
Additional information
Supplementary Information accompanies this paper at http://www.nature.com/
naturecommunications
Competing financial interests: The authors declare no competing financial interests.
Reprints and permission information is available online at http://npg.nature.com/
reprintsandpermissions/
How to cite this article: Pullen, M. G. et al. Imaging an aligned polyatomic molecule
with laser-induced electron diffraction. Nat. Commun. 6:7262 doi: 10.1038/ncomms8262
(2015).
This work is licensed under a Creative Commons Attribution 4.0
International License. The images or other third party material in this
article are included in the article’s Creative Commons license, unless indicated otherwisein the credit line; if the material is not included under the Creative Commons license,users will need to obtain permission from the license holder to reproduce the material.To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms8262
6 NATURE COMMUNICATIONS | 6:7262 | DOI: 10.1038/ncomms8262 | www.nature.com/naturecommunications
&2015 Macmillan Publishers Limited. All rights reserved. | 10.1038_ncomms8262 | page_0005 |
Citation: Aqeel, I.; Bilal, M.; Majid,
A.; Majid, T. Hybrid Approach to
Identifying Druglikeness Leading
Compounds against COVID-19 3CL
Protease. Pharmaceuticals 2022 ,15,
1333. https://doi.org/10.3390/
ph15111333
Academic Editors: Jean-Pierre
Bazureau and Dhimant Desai
Received: 24 September 2022
Accepted: 25 October 2022
Published: 28 October 2022
Publisher’s Note: MDPI stays neutral
with regard to jurisdictional claims in
published maps and institutional affil-
iations.
Copyright: © 2022 by the authors.
Licensee MDPI, Basel, Switzerland.
This article is an open access article
distributed under the terms and
conditions of the Creative Commons
Attribution (CC BY) license (https://
creativecommons.org/licenses/by/
4.0/).
pharmaceuticals
Article
Hybrid Approach to Identifying Druglikeness Leading
Compounds against COVID-19 3CL Protease
Imra Aqeel1
, Muhammad Bilal1
, Abdul Majid1and T uba Majid2,*
1Biomedical Informatics Research Lab, Department of Computer & Information Sciences, Pakistan Institute of
Engineering & Applied Sciences, Nilore, Islamabad 45650, Pakistan
2Experimental Continuum Mechanics Research Group, Department of Mechanical and Process Engineering,
ETH Zurich, 8092 Zürich, Switzerland
*Correspondence: tmajid@ethz.ch
Abstract: SARS-CoV-2 is a positive single-strand RNA-based macromolecule that has caused the
death of more than 6.3 million people since June 2022. Moreover, by disturbing global supply chains
through lockdowns, the virus has indirectly caused devastating damage to the global economy. It is
vital to design and develop drugs for this virus and its various variants. In this paper, we developed
an in silico study-based hybrid framework to repurpose existing therapeutic agents in finding drug-
like bioactive molecules that would cure COVID-19. In the first step, a total of 133 drug-likeness
bioactive molecules are retrieved from the ChEMBL database against SARS coronavirus 3CL Protease.
Based on the standard IC50, the dataset is divided into three classes: active, inactive, and intermediate.
Our comparative analysis demonstrated that the proposed Extra Tree Regressor (ETR)-based QSAR
model has improved prediction results related to the bioactivity of chemical compounds as compared
to Gradient Boosting-, XGBoost-, Support Vector-, Decision Tree-, and Random Forest-based regressor
models. ADMET analysis is carried out to identify thirteen bioactive molecules with the ChEMBL
IDs 187460, 190743, 222234, 222628, 222735, 222769, 222840, 222893, 225515, 358279, 363535, 365134,
and 426898. These molecules are highly suitable drug candidates for SARS-CoV-2 3CL Protease. In
the next step, the efficacy of the bioactive molecules is computed in terms of binding affinity using
molecular docking, and then six bioactive molecules are shortlisted, with the ChEMBL IDs 187460,
222769, 225515, 358279, 363535, and 365134. These molecules can be suitable drug candidates for
SARS-CoV-2. It is anticipated that the pharmacologist and/or drug manufacturer would further
investigate these six molecules to find suitable drug candidates for SARS-CoV-2. They can adopt
these promising compounds for their downstream drug development stages.
Keywords: SARS-CoV-2; 3C-like protease; drug repurposing; regression model; bioactive molecules;
molecular docking
1. Introduction
Novel coronavirus (nCoV-19) is a rapidly spreading pandemic. The International Com-
mittee on Taxonomy of Viruses (ICTV) officially named severe acute respiratory syndrome
coronavirus 2 (SARS-CoV-2) on February 11, 2020 [ 1]. At first, coronavirus-2 appeared
in December 2019 in Asia and then spread out worldwide. A total of 228 countries and
more than 500 million people got infected. SARS-CoV-2 is like MERS-CoV and SARS-CoV .
Both these viruses have caused severe acute respiratory syndrome. There are seven strains
of Alpha and Beta coronaviruses in human coronaviruses. HCoV-229E and HCoV-NL63
belong to the type of alpha-coronaviruses. On the other hand, HCoV-HKU1, HCoV-OC43,
SARS-CoV , MERS-CoV , and SARS-CoV-2 belong to beta-coronaviruses [ 2].COVID-19 virus
is a single-strand ribonucleic acid (ssRNA) virus that consists of multiple structural and
non-structural proteins. The structural proteins have four different types: spike (S), mem-
brane (M), envelope (E), and nucleocapsid (N) proteins. However, non-structural proteins
Pharmaceuticals 2022 ,15, 1333. https://doi.org/10.3390/ph15111333 https://www.mdpi.com/journal/pharmaceuticals | 10.3390_ph15111333 | page_0000 |
Pharmaceuticals 2022 ,15, 1333 2 of 21
contain sixteen different types, named NSP1, NSP2, NSP3 . . ., and NSP16. These proteins
are mainly more responsible for spreading out SARS-CoV-2 than other types of proteins.
Consequently, these proteins are considered potential targets to prevent SARS-CoV-2, espe-
cially the 3C-like protease (3CLproor Mpro), which is crucial for replication [ 3]. Figure 1a
shows a visual model of SARS-CoV-2 with all the constituent proteins, and Figure 1b
depicts its large genome size of 29.9 kb, starting from 50to 30. This virus has the inherent
capability of auto-reproduction into sixteen different types of non-structural proteins.
Pharmaceuticals 2022 , 15, x FOR PEER REVIEW 2 of 22
structural and non-structural proteins. The stru ctural proteins have four different types:
spike (S), membrane (M), envelope (E), an d nucleocapsid (N) proteins. However, non-
structural proteins contain sixteen differ ent types, named NSP1, NSP2, NSP3…, and
NSP16. These proteins are mainly more resp onsible for spreading out SARS-CoV-2 than
other types of proteins. Consequently, these pr oteins are considered potential targets to
prevent SARS-CoV-2, especially the 3C-like protease (3CLpro or Mpro), which is crucial for
replication [3]. Figure 1a shows a visual model of SARS-CoV-2 with all the constituent
proteins, and Figure 1b depicts its large genome size of 29.9 kb, starting from 5 ʹ to 3ʹ. This
virus has the inherent capability of auto-reproduction into sixteen different types of non-
structural proteins.
Figure 1. (a) SARS-CoV-2 with constituent proteins, ( b) related genome detailed information.
Upon entrance into the host cell, the viral genome is translated to produce two over-
lying polyproteins named pp1a and pp1b [4]. During the proteolytic activity, these poly-
proteins are excised from the 3CL protease (3CLpro, also known as the Main protease
(Mpro)). These proteins work with a papain-like protease to slice the polyproteins to pro-
duce a total of sixteen function al nonstructural proteins (NSP s). It was reported that the
eleven slicing sites of polyprotein 1ab we re shared and operated by only the 3CLpro of
SARS, and no other human protease was involved in the slicing process [4]. To initiate
viral replication, the viral replication transcription complex (RTC) is assembled by the
sliced NSPs.
The computational drug discovery process has become a crucial strategy to develop
the drug against COVID-19. It can be an effect ive tool to save money and reduce the time
for drug discovery/repurposing [5]. Recently, machine learning (ML) approaches have been employed for data modeling and drug discovery applications. Various online med-
ical databases that contain sufficient information related to bioactive molecules are avail-
able. This has made it possible to employ the ML approaches-based QSAR model to
Figure 1. (a) SARS-CoV-2 with constituent proteins, ( b) related genome detailed information.
Upon entrance into the host cell, the viral genome is translated to produce two overly-
ing polyproteins named pp1a andpp1b [4]. During the proteolytic activity, these polypro-
teins are excised from the 3CL protease (3CLpro, also known as the Main protease (Mpro)).
These proteins work with a papain-like protease to slice the polyproteins to produce a total
of sixteen functional nonstructural proteins (NSPs). It was reported that the eleven slicing
sites of polyprotein 1ab were shared and operated by only the 3CLproof SARS, and no
other human protease was involved in the slicing process [ 4]. To initiate viral replication,
the viral replication transcription complex (RTC) is assembled by the sliced NSPs.
The computational drug discovery process has become a crucial strategy to develop
the drug against COVID-19. It can be an effective tool to save money and reduce the time
for drug discovery/repurposing [ 5]. Recently, machine learning (ML) approaches have
been employed for data modeling and drug discovery applications. Various online medical
databases that contain sufficient information related to bioactive molecules are available.
This has made it possible to employ the ML approaches-based QSAR model to quickly
develop vaccines for the COVID-19 pandemic [ 6]. Due to stringent storage requirements,
this vaccine is rather difficult to transport and warehouse. Moreover, successful virus vacci-
nations for humans and animals are seriously hampered by vaccine-associated increased
illness [ 7]. This has shown that people are not as receptive to getting vaccinated as they
are to taking drugs [ 8]. On the other hand, underdeveloped countries suffered the most
from the pandemic, with the official death tolls of India and Brazil, at the time of writing
this manuscript, being 525,000 and 672,000, respectively. Since the start of this pandemic, | 10.3390_ph15111333 | page_0001 |
1. Singh, R. P . et al. in Advances in Agronomy
(ed. Sparks, D. L.) Vol. 98 Ch. 5, 271–309
(Elsevier, London, 2008).
2. Flor, H. H. Current status of gene -for-gene concept .
Annu. Rev. Phytopathol. 9, 275–296 (1971).
3. Ausubel, F. M. Are innate immune signaling pathways
in plants and animals conserved? Nature Immunol. 6,
973–979 (2005).
4. Chisholm, S. T ., Coaker, G., Day, B. & Staskawicz, B. J.
Host–microbe interactions: shaping the evolution of
the plant immune response . Cell 124, 803–814
(2006).
5. Jones, J. D. & Dangl, J. L. The plant immune system .
Nature 444, 323–329 (2006).
6. Boller, T . & Felix, G. A renaissance of elicitors:
perception of microbe-associated molecular patterns
and danger signals by pattern-recognition receptors .
Annu. Rev. Plant Biol. 60, 379–406 (2009).
7. Zipfel, C. Pattern-recognition receptors in plant innate
immunity . Curr. Opin. Immunol. 20, 10–16 (2008).
8. Lehti-Shiu, M. D., Zou, C., Hanada, K. & Shiu, S. -H.
Evolutionary history and stress regulation of plant
receptor-like kinase/pelle genes . Plant Physiol. 150,
12–26 (2009).
9. Wang, G. et al. A genome-wide functional investigation
into the roles of receptor-like proteins in Arabidopsis.
Plant Physiol. 147, 503–517 (2008).
10. Gay, N. J. & Gangloff, M. Structure and function of T oll
receptors and their ligands . Annu. Rev. Biochem. 76,
141–165 (2007).
11. Zipfel, C. et al. Perception of the bacterial PAMP EF -T u
by the receptor EFR restricts Agrobacterium -mediated
transformation . Cell 125, 749–760 (2006).
12. Lee, S. -W. et al. A type I-secreted, sulfated peptide
triggers XA21-mediated innate immunity . Science
326, 850–853 (2009).
Although most PRRs provide subtle protection, this
paper shows that a single PAMP receptor can confer
effective disease resistance to a bacterial pathogen.
13. Chinchilla, D. et al. A flagellin-induced complex of the
receptor FLS2 and BAK1 initiates plant defence .
Nature 448, 497–500 (2007).
14. Heese, A. et al. The receptor-like kinase SERK3/BAK1
is a central regulator of innate immunity in plants .
Proc. Natl Acad. Sci. USA 104, 12217–12222 (2007).
This study identified of BAK1 as a central regulator
of plant immunity.
15. Miya, A. et al. CERK1, a LysM receptor kinase, is
essential for chitin elicitor signaling in Arabidopsis.
Proc. Natl Acad. Sci. USA 104, 19613–19618 (2007).
16. Wan, J. et al. A LysM receptor-like kinase plays a
critical role in chitin signaling and fungal resistance in
Arabidopsis. Plant Cell 20, 471–481 (2008).
17. Gimenez-Ibanez, S., Ntoukakis, V. & Rathjen, J.
The LysM receptor kinase CERK1 mediates bacterial
perception in Arabidopsis. Plant Signal. Behav. 4,
539–541 (2009).
18. Schulze, B. et al. Rapid heteromerization and
phosphorylation of ligand-activated plant
transmembrane receptors and their associated kinase
BAK1 . J. Biol. Chem. 285, 9444–9451 (2010).
19. Shan, L. et al. Bacterial effectors target the common
signaling partner BAK1 to disrupt multiple MAMP
receptor-signaling complexes and impede plant
immunity . Cell Host Microbe 4, 17–27 (2008).
20. Kemmerling, B. et al. The BRI1-associated kinase 1,
BAK1, has a brassinolide-independent role in plant
cell-death control . Curr. Biol. 17, 1116–1122 (2007).
21. He, K. et al. BAK1 and BKK1 regulate brassinosteroid-
dependent growth and brassinosteroid-independent
cell-death pathways . Curr. Biol. 17, 1109–1115 (2007).
22. Veronese, P . et al. The membrane-anchored
BOTRYTIS-INDUCED KINASE1 plays distinct roles in
Arabidopsis resistance to necrotrophic and biotrophic
pathogens . Plant Cell 18, 257–273 (2006).
23. Lu, D. et al. A receptor-like cytoplasmic kinase, BIK1,
associates with a flagellin receptor complex to initiate
plant innate immunity . Proc. Natl Acad. Sci. USA 107,
496–501 (2010).
24. Cunnac, S., Lindeberg, M. & Collmer, A.
Pseudomonas syringae type III secretion system
effectors: repertoires in search of functions . Curr. Opin.
Microbiol. 12, 53–60 (2009).
25. Kvitko, B. H. et al. Deletions in the repertoire of
Pseudomonas syringae pv. tomato DC3000 type III
secretion effector genes reveal functional overlap
among effectors . PLoS Pathog. 5, e1000388 (2009).
A clear, genetic demonstration that bacterial
effectors work redundantly. This explains why
individual deletions of effector genes often have
minor phenotypes.26. Zhou, J. -M. & Chai, J. Plant pathogenic bacterial
type III effectors subdue host responses . Curr. Opin.
Microbiol. 11, 179–185 (2008).
27. Hauck, P ., Thilmony, R. & He, S. Y. A Pseudomonas
syringae type III effector suppresses cell wall-based
extracellular defense in susceptible Arabidopsis plants .
Proc. Natl Acad. Sci. USA 100, 8577–8582 (2003).
28. Kim, M. G. et al. Two Pseudomonas syringae type III
effectors inhibit RIN4-regulated basal defense in
Arabidopsis. Cell 121, 749–759 (2005).
29. Grant, S. R., Fisher, E. J., Chang, J. H., Mole, B. M. &
Dangl, J. L. Subterfuge and manipulation: type III
effector proteins of phytopathogenic bacteria .
Annu. Rev. Microbiol. 60, 425–449 (2006).
30. Xiang, T . et al. Pseudomonas syringae effector AvrPto
blocks innate immunity by targeting receptor kinases .
Curr. Biol. 18, 74–80 (2008).
This paper shows that the bacterial effector
AvrPto targets receptor kinases, providing the
intellectual basis for current decoy models of
indirect pathogen recognition.
31. Gohre, V. et al. Plant pattern-recognition receptor
FLS2 is directed for degradation by the bacterial
ubiquitin ligase AvrPtoB . Curr. Biol. 18, 1824–1832
(2008).
32. Gimenez-Ibanez, S. et al. AvrPtoB targets the LysM
receptor kinase CERK1 to promote bacterial virulence
on plants . Curr. Biol. 19, 423–429 (2009).
33. Rosebrock, T . R. et al. A bacterial E3 ubiquitin ligase
targets a host protein kinase to disrupt plant
immunity . Nature 448, 370–374 (2007).
34. Xing, W. et al. The structural basis for activation of
plant immunity by bacterial effector protein AvrPto .
Nature 449, 243–247 (2007).
35. Axtell, M. J. & Staskawicz, B. J. Initiation of RPS2-
specified disease resistance in Arabidopsis is coupled
to the AvrRpt2-directed elimination of RIN4 . Cell 112,
369–377 (2003).
36. Mackey, D., Belkhadir, Y., Alonso, J. M., Ecker, J. R. &
Dangl, J. Arabidopsis RIN4 is a target of the type III
virulence effector AvrRpt2 and modulates RPS2-
mediated resistance . Cell 112, 379–389 (2003).
37. Wilton, M. et al. The type III effector HopF2Pto
targets Arabidopsis RIN4 protein to promote
Pseudomonas syringae virulence . Proc. Natl Acad.
Sci. USA 107, 2349–2354 (2010).
38. Marathe, R. & Dinesh-Kumar, S. P. Plant defense:
one post, multiple guards?! Mol. Cell 11, 284–286
(2003).
39. Liu, J. et al. RIN4 functions with plasma membrane
H+-ATPases to regulate stomatal apertures during
pathogen attack . PLoS Biol. 7, e1000139 (2009).
40. Kay, S. & Bonas, U. How Xanthomonas type III effectors
manipulate the host plant . Curr. Opin. Microbiol. 12,
37–43 (2009).
41. Kay, S., Hahn, S., Marois, E., Hause, G. & Bonas, U.
A bacterial effector acts as a plant transcription
factor and Induces a cell size regulator . Science 318,
648–651 (2007).
42. Boch, J. et al. Breaking the code of DNA binding
specificity of TAL-T ype III effectors . Science 326,
1509–1512 (2009).
43. Moscou, M. J. & Bogdanove, A. J. A simple cipher
governs DNA recognition by TAL effectors . Science
326, 1501 (2009).
44. Römer, P., Recht, S. & Lahaye, T. A single plant
resistance gene promoter engineered to recognize
multiple TAL effectors from disparate pathogens .
Proc. Natl Acad. Sci. USA 106, 20526–20531
(2009).
45. Gu, K. et al. R. gene expression induced by a type-III
effector triggers disease resistance in rice . Nature
435, 1122–1125 (2005).
46. Romer, P . et al. Plant pathogen recognition mediated
by promoter activation of the pepper Bs3 resistance
gene. Science 318, 645–648 (2007).
47. Kamoun, S. Groovy times: filamentous pathogen
effectors revealed . Curr. Opin. Plant Biol. 10,
358–365 (2007).
48. Panstruga, R. & Dodds, P . N. T errific protein traffic:
the mystery of effector protein delivery by filamentous
plant pathogens . Science 324, 748–750 (2009).
49. Haas, B. J. et al. Genome sequence and analysis of the
Irish potato famine pathogen Phytophthora infestans.
Nature 461, 393–398 (2009).
50. Dean, R. A. et al. The genome sequence of the rice
blast fungus Magnaporthe grisea. Nature 434,
980–986 (2005).
51. Kämper, J. et al. Insights from the genome of the
biotrophic fungal plant pathogen Ustilago maydis.
Nature 444, 97–101 (2006).52. Hogenhout, S. A., Van der Hoorn, R. A. L., T erauchi, R.
& Kamoun, S. Emerging concepts in effector biology of
plant-associated organisms . Mol. Plant Microbe
Interact. 22, 115–122 (2009).
53. Bos, J. I. et al. Phytophthora infestans effector AVR3a
is essential for virulence and manipulates plant
immunity by stabilizing host E3 ligase CMPG1 .
Proc. Natl Acad. Sci. USA 107, 9909–9914 (2010).
54. Bellafiore, S. et al. Direct identification of the
Meloidogyne incognita secretome reveals proteins
with host cell reprogramming potential . PLoS Pathog.
4, e1000192 (2008).
55. Voinnet, O. RNA silencing as a plant immune system
against viruses . T rends Genet. 17, 449–459 (2001).
56. Girardin, S. E., Philpott, D. J. & Lemaitre, B.
Sensing microbes by diverse hosts. Workshop on
pattern recognition proteins and receptors . EMBO
Rep. 4, 932–936 (2003).
57. Jia, Y., McAdams, S. A., Bryan, G. T ., Hershey, H. P . &
Valent, B. Direct interaction of resistance gene and
avirulence gene products confers rice blast resistance .
EMBO J. 19, 4004–4014 (2000).
58. Catanzariti, A. M., Dodds, P . N., Lawrence, G. J.,
Ayliffe, M. A. & Ellis, J. G. Haustorially expressed
secreted proteins from flax rust are highly enriched for
avirulence elicitors . Plant Cell 18, 243–256 (2006).
59. Dodds, P . N., Lawrence, G. J., Catanzariti, A. M.,
Ayliffe, M. A. & Ellis, J. G. The Melampsora lini
AvrL567 avirulence genes are expressed in haustoria
and their products are recognized inside plant cells .
Plant Cell 16, 755–768 (2004).
60. Catanzariti, A. M. et al. The AvrM effector from flax
rust has a structured C -terminal domain and interacts
directly with the M resistance protein . Mol. Plant
Microbe Interact. 23, 49–57 (2010).
61. Dodds, P . N. et al. Direct protein interaction underlies
gene-for-gene specificity and coevolution of the flax
resistance genes and flax rust avirulence genes .
Proc. Natl Acad. Sci. USA 103, 8888–8893 (2006).
An elegant demonstration of how direct recognition
of pathogen effectors works at the molecular level
and drives antagonistic co-evolution.
62. van der Hoorn, R. A. & Kamoun, S. From guard to
decoy: a new model for perception of plant pathogen
effectors . Plant Cell 20, 2009–2017 (2008).
63. Dangl, J. L. & Jones, J. D. Plant pathogens and
integrated defence responses to infection . Nature
411, 826–833 (2001).
64. Mackey, D., Holt, B. F., Wiig, A. & Dangl, J. L.
RIN4 interacts with Pseudomonas syringae
type III effector molecules and is required for
RPM1-mediated resistance in Arabidopsis. Cell 108,
743–754 (2002).
65. Mucyn, T . S. et al. The tomato NBARC-LRR protein Prf
interacts with Pto kinase in vivo to regulate specific
plant immunity . Plant Cell 18, 2792–2806 (2006).
66. Zipfel, C. & Rathjen, J. P. Plant immunity:
AvrPto targets the frontline . Curr. Biol. 18,
R218–R220 (2008).
67. Gutierrez, J. R. et al. Prf immune complexes of
tomato are oligomeric and contain multiple Pto-like
kinases that diversify effector recognition . Plant J. 61,
507–581 (2009).
68. Mucyn, T . S., Wu, A. J., Balmuth, A. L., Arasteh, J. M.
& Rathjen, J. P . Regulation of tomato Prf by Pto-like
protein kinases . Mol. Plant Microbe Interact. 22,
391–401 (2009).
69. Collier, S. M. & Moffett, P . NB -LRRs work a ‘bait and
switch’ on pathogens . T rends Plant Sci. 14, 521–529
(2009).
70. Ntoukakis, V. et al. Host inhibition of a bacterial
virulence effector triggers immunity to infection .
Science 324, 784–787 (2009).
71. Narusaka, M. et al. RRS1 and RPS4 provide a dual
resistance-gene system against fungal and bacterial
pathogens . Plant J. 60, 218–226 (2009).
An as-yet-unique example showing how the host
can inactivate a pathogen effector, leading to its
recognition and host immunity.
72. Sinapidou, E. et al. T wo TIR:NB:LRR genes are
required to specify resistance to Peronospora
parasitica isolate Cala2 in Arabidopsis. Plant J. 38,
898–909 (2004).
73. Loutre, C. et al. Two different CC -NBS-LRR genes
are required for Lr10-mediated leaf rust resistance
in tetraploid and hexaploid wheat . Plant J. 60,
1043–1054 (2009).
74. Lee, S. K. et al. Rice Pi5-mediated resistance to
Magnaporthe oryzae requires the presence of two coil
ed-coil-nucleotide -binding -leucine-rich repeat genes .
Genetics 181, 1627–1638 (2009).REVIEWS
nATuRE REvIEwS | Genetics ADvAncE OnlInE Publ IcATIO n | 9
© 20 Macmillan Publishers Limited. All rights reserved 10 | 6f570033c1d07467276ad185bf1521cd736c5f3f | page_0008 |
75. T akken, F. L. & T ameling, W. I. T o nibble at plant
resistance proteins . Science 324, 744–746 (2009).
76. Birker, D. et al. A locus conferring resistance to
Colletotrichum higginsianum is shared by four
geographically distinct Arabidopsis accessions . Plant J.
60, 602–613 (2009).
77. Dodds, P . N., Lawrence, G. J. & Ellis, J. G. Six amino
acid changes confined to the leucine-rich repeat
β-strand/ β-turn motif determine the difference
between the P and P2 rust resistance specificities in
flax. Plant Cell 13, 163–178 (2001).
78. Ellis, J. G., Lawrence, G. J., Luck, J. E. & Dodds, P . N.
Identification of regions in alleles of the flax rust
resistance gene L that determine differences in
gene -for-gene specificity . Plant Cell 11, 495–506
(1999).
79. Shen, Q. H. et al. Recognition specificity and RAR1/
SGT1 dependence in barley Mla disease resistance
genes to the powdery mildew fungus . Plant Cell 15,
732–744 (2003).
80. Rairdan, G. J. & Moffett, P . Distinct domains in the
ARC region of the potato resistance protein Rx
mediate LRR binding and inhibition of activation .
Plant Cell 18, 2082–2093 (2006).
81. Belkhadir, Y., Nimchuk, Z., Hubert, D. A., Mackey, D.
& Dangl, J. L. Arabidopsis RIN4 negatively regulates
disease resistance mediated by RPS2 and RPM1
downstream or independent of the NDR1 signal
modulator and is not required for the virulence
functions of bacterial type III effectors AvrRpt2 or
AvrRpm1 . Plant Cell 16, 2822–2835 (2004).
82. T ameling, W. I. et al. Mutations in the NB -ARC domain
of I-2 that impair ATP hydrolysis cause autoactivation .
Plant Physiol. 140, 1233–1245 (2006).
83. Inohara, N. et al. An induced proximity model for
NF-κB activation in the Nod1/RICK and RIP signaling
pathways . J. Biol. Chem. 275, 27823–27831 (2000).
84. Shaw, M. H., Reimer, T ., Kim, Y. G. & Nunez, G.
NOD-like receptors (NLRs): bona fide intracellular
microbial sensors . Curr. Opin. Immunol. 20, 377–382
(2008).
85. Mestre, P . & Baulcombe, D. C. Elicitor-mediated
oligomerization of the tobacco N disease resistance
protein . Plant Cell 18, 491–501 (2006).
86. Frost, D. et al. T obacco transgenic for the flax rust
resistance gene L expresses allele-specific activation of
defense responses . Mol. Plant Microbe Interact. 17,
224–232 (2004).
87. Swiderski, M. R., Birker, D. & Jones, J. D.
The TIR domain of TIR -NB-LRR resistance proteins
is a signaling domain involved in cell death induction .
Mol. Plant Microbe Interact. 22, 157–165 (2009).
88. Bendahmane, A., Farnham, G., Moffett, P . &
Baulcombe, D. C. Constitutive gain -of-function
mutants in a nucleotide binding site-leucine rich
repeat protein encoded at the Rx locus of potato .
Plant J. 32, 195–204 (2002).
89. Rairdan, G. J. et al. The coiled-coil and nucleotide
binding domains of the potato Rx disease resistance
protein function in pathogen recognition and
signaling . Plant Cell 20, 739–751 (2008).
90. T ao, Y., Yuan, F., Leister, R. T ., Ausubel, F. M. &
Katagiri, F. Mutational analysis of the Arabidopsis
nucleotide binding site -leucine -rich repeat resistance
gene RPS2. Plant Cell 12, 2541–2554 (2000).
91. T ao, Y. et al. Quantitative nature of Arabidopsis
responses during compatible and incompatible
interactions with the bacterial pathogen Pseudomonas
syringae. Plant Cell 15, 317–330 (2003).
92. Pitzschke, A., Schikora, A. & Hirt, H. MAPK cascade
signalling networks in plant defence . Curr. Opin. Plant
Biol. 12, 421–426 (2009).93. Asai, T . et al. MAP kinase signalling cascade in
Arabidopsis innate immunity . Nature 415, 977–983
(2002).
94. Suarez-Rodriguez, M. C. et al. MEKK1 is required for
flg22-induced MPK4 activation in Arabidopsis plants .
Plant Physiol. 143, 661–669 (2007).
95. Liu, Y. & Zhang, S. Phosphorylation of
1-aminocyclopropane -1-carboxylic acid synthase by
MPK6, a stress-responsive mitogen-activated protein
kinase, induces ethylene biosynthesis in Arabidopsis.
Plant Cell 16, 3386–3399 (2004).
96. Bethke, G. et al. Flg22 regulates the release of an
ethylene response factor substrate from MAP kinase 6
in Arabidopsis thaliana via ethylene signaling .
Proc. Natl Acad. Sci. USA 106, 8067–8072 (2009).
97. Boudsocq, M. et al. Differential innate immune
signalling via Ca2+ sensor protein kinases . Nature
464, 418–422 (2010).
98. Shirasu, K. & Schulze-Lefert, P . Complex formation,
promiscuity and multi-functionality: protein
interactions in disease-resistance pathways.
T rends Plant Sci. 8, 252–258 (2003).
99. Wiermer, M., Feys, B. J. & Parker, J. E. Plant immunity:
the EDS1 regulatory node . Curr. Opin. Plant Biol. 8,
383–389 (2005).
100. Day, B., Dahlbeck, D. & Staskawicz, B. J.
NDR1 interaction with RIN4 mediates the differential
activation of multiple disease resistance pathways in
Arabidopsis. Plant Cell 18, 2782–2791 (2006).
101. Burch-Smith, T . M. et al. A novel role for the TIR
domain in association with pathogen-derived elicitors .
PLoS Biol. 5, e68 (2007).
102. Shen, Q. H. et al. Nuclear activity of MLA immune
receptors links isolate-specific and basal disease-
resistance responses . Science 315, 1098–1103 (2007).
103. Wirthmueller, L., Zhang, Y., Jones, J. D. & Parker, J. E.
Nuclear accumulation of the Arabidopsis immune
receptor RPS4 is necessary for triggering EDS1-
dependent defense . Curr. Biol. 17, 2023–2029
(2007).
The key paper underlying the hypothesis that an
active fraction of plant NB -LRR proteins resides in
the plant cell nucleus.
104. Deslandes, L. et al. Physical interaction between
RRS1 -R, a protein conferring resistance to bacterial
wilt, and PopP2, a type III effector targeted to the
plant nucleus . Proc. Natl Acad. Sci. USA 100,
8024–8029 (2003).
105. Bernoux, M. et al. RD19, an Arabidopsis cysteine
protease required for RRS1 -R-mediated resistance,
is relocalized to the nucleus by the Ralstonia
solanacearum PopP2 effector . Plant Cell 20,
2252–2264 (2008).
106. Bari, R. & Jones, J. D. Role of plant hormones in plant
defence responses . Plant Mol. Biol. 69, 473–488
(2009).
107. T suda, K., Sato, M., Stoddard, T ., Glazebrook, J. &
Katagiri, F. Network properties of robust immunity in
plants . PLoS Genet. 5, e1000772 (2009).
A network approach to plant immunity shows
complex interactions between defence hormone
signalling pathways acting in both PTI and ETI.
108. Lipka, V. et al. Pre- and postinvasion defenses both
contribute to nonhost resistance in Arabidopsis.
Science 310, 1180–1183 (2005).
109. Stein, M. et al. Arabidopsis PEN3/PDR8, an ATP
binding cassette transporter, contributes to nonhost
resistance to inappropriate pathogens that enter by
direct penetration . Plant Cell 18, 731–746 (2006).
110. Bednarek, P . et al. A glucosinolate metabolism
pathway in living plant cells mediates broad-spectrum
antifungal defense . Science 323, 101–106 (2009).111. Clay, N. K., Adio, A. M., Denoux, C., Jander, G. &
Ausubel, F. M. Glucosinolate metabolites required for
an Arabidopsis innate immune response . Science
323, 95–101 (2009).
112. Krattinger, S. G. et al. A putative ABC transporter
confers durable resistance to multiple fungal pathogens
in wheat . Science 323, 1360–1363 (2009).
The first cloning of a broad spectrum resistance
gene in wheat that is active against rusts and
mildews. This study opened the way for the genetic
manipulation of crop cultivars.
113. Wang, W., Wen, Y., Berkey, R. & Xiao, S.
Specific targeting of the Arabidopsis resistance protein
RPW8.2 to the interfacial membrane encasing the
fungal Haustorium renders broad-spectrum resistance
to powdery mildew . Plant Cell 21, 2898–2913 (2009).
114. Anelli, T . & Sitia, R. Protein quality control in the early
secretory pathway . EMBO J. 27, 315–327 (2008).
115. Häweker, H. et al. Pattern recognition receptors
require N -glycosylation to mediate plant immunity .
J. Biol. Chem. 285, 4629–4636 (2010).
116. Li, J. et al. Specific ER quality control components
required for biogenesis of the plant innate immune
receptor EFR . Proc. Natl Acad. Sci. USA 106,
15973–15978 (2009).
117. Lu, X. et al. Uncoupling of sustained MAMP receptor
signaling from early outputs in an Arabidopsis
endoplasmic reticulum glucosidase II allele . Proc. Natl
Acad. Sci. USA 106, 22522–22527 (2009).
118. Nekrasov, V. et al. Control of the pattern-recognition
receptor EFR by an ER protein complex in plant
immunity . EMBO J. 28, 3428–3438 (2009).
119. Saijo, Y. et al. Receptor quality control in the
endoplasmic reticulum for plant innate immunity .
EMBO J. 28, 3439–3449 (2009).
120. Vleeshouwers, V. G. A. A. et al. Effector genomics
accelerates discovery and functional profiling of potato
disease resistance and Phytophthora infestans
avirulence genes . PLoS ONE 3, e2875 (2008).
121. Lacombe, S. et al. Interfamily transfer of a plant pattern-
recognition receptor confers broad-spectrum bacterial
resistance . Nature Biotech. 28, 365–369 (2010).
A groundbreaking paper showing the potential for
interfamily transfer of PRRs to provide broad
spectrum disease protection in crop species.
Acknowledgements
We apologize to those authors whose work could not be cited
owing to space limitations. J.P.R. is an Australian Research
Council Future Fellow. Work in P.N.D.’s laboratory is funded
by the Australian Research Council, the US National Institutes
of Health and the Grains Research and Development
Corporation. We thank J. Ellis and B. Staskawicz for helpful
discussions.
Competing interests statement
The authors declare no competing financial interests.
DATABASES
TAIR: http://www.arabidopsis.or g
RPS4 | RRS1
UniProtKB: http://www.uniprot.or g
BAK1 | BIK1 | CERK1 | FLS2 | RIN4 | RPS2
FURTHER INFORMATION
Peter N. Dodds’ homepage:
http://www.csiro.au/people/Peter.Dodds.html
John P . Rathjen’s homepage:
http://biology.anu.edu.au/Staff/Profiles/PS/Rathjen
All links Are Active in the online pd FREVIEWS
10 | ADvAncE OnlInE Publ IcATIO n www.nature.com/reviews/genetics
© 20 Macmillan Publishers Limited. All rights reserved 10 | 6f570033c1d07467276ad185bf1521cd736c5f3f | page_0009 |
Review of the literature and suggestions for the design
of rodent survival studies for the identification of compounds
that increase health and life span
Stephen Richard Spindler
Received: 2 September 2010 /Accepted: 21 February 2011 /Published online: 22 March 2011
#The Author(s) 2011. This article is published with open access at Springerlink.com
Abstract Much of the literature describing the search
for agents that increase the life span of rodents was
found to suffer from confounds. One-hundred-six
studies, absent 20 contradictory melatonin studies, ofcompounds or combinations of compounds were
reviewed. Only six studies reported both life span
extension and food consumption data, thereby ex-cluding the potential effects of caloric restriction. Six
other studies reported life span extension without a
change in body weight. However, weight can be anunreliable surrogate measure of caloric consumption.
Twenty studies reported that food consumption or
weight was unchanged, but it was unclear whetherthese data were anecdotal or systematic. Twenty-nine
reported extended life span likely due to induced
caloric restriction. Thirty-six studies reported no effecton life span, and three a decrease. The remaining studies
suffer from more serious confounds. Though still widely
cited, studies showing life span extension using short-lived or “enfeebled ”rodents have not been shown to
predict longevity effects in long-lived animals. We
suggest improvements in experimental design that willenhance the reliability of the rodent life span literature.First, animals should receive measured quantities of
food and its consumption monitored, preferably daily,
and reported. Weights should be measured regularly and
reported. Second, a genetically heterogeneous, long-lived rodent should be utilized. Third, chemically
defined diets should be used. Fourth, a positive control
(e.g., a calorically restricted group) is highly desirable.Fifth, drug dosages should be chosen based on surrogate
endpoints or accepted cross-species scaling factors.
These procedures should improve the reliability of thescientific literature and accelerate the identification of
longevity and health span-enhancing agents.
Keywords Longevity therapeutics .CR mimetics .
Geroprotectors .Health span .Life span .Longevity .
Drug discovery .Pharmaceutical testing
Introduction
There are presently no authentic longevity therapeu-
tics. Such compounds would intervene in the processof aging to extend mean and/or maximum life span,
maintain physiological function, and mitigate the
onset and severity of a broad spectrum of age-related diseases in mammals. Such drugs might
engage the pathways used by caloric, methionine,
and phenylalanine restriction, and the longevity-enhancing mutations (reviewed in Spindler 2009 ).
The terms “CR mimetics ”and “geroprotectors ”have
been used to describe such compounds (Weindruch etAGE (2012) 34:111 –120
DOI 10.1007/s11357-011-9224-6
Electronic supplementary material The online version of this
article (doi:10.1007/s11357-011-9224-6) contains
supplementary material, which is available to authorized users.
S. R. Spindler ( *)
Department of Biochemistry, University of California,
Riverside, CA 92521, USA
e-mail: spindler@ucr.edu | 82cfe32fe3b91f79483b71434b7fe2f6ad5a63d0 | page_0000 |
al.2001 ; Roth et al. 2001 ; Cao et al. 2001 ; Anisimov
1982 ; Lippman 1981 ). In this report, we will use the
general term “longevity therapeutics. ”
While a full understanding of the mechanisms of
aging will greatly facilitate the development anddeployment of longevity therapeutics, drug discovery
and development have a long history of using
surrogate assays for identifying therapeutics, oftenwith little knowledge or understanding of the etiology
of the diseases for which the therapeutics were
intended (discussed in Spindler 2006 ). Indeed, most
of the medications currently in our armamentarium
were discovered using surrogate assays. Thus, the
development and refinement of surrogate assays forlongevity therapeutics should speed their identification.
There have been multiple methods used in the
attempt to identify such compounds. For example, we
and others have utilized genome-wide microarray
studies of treated mice to identify potential therapeu-tics (Barger et al. 2008 ; Spindler and Dhahbi 2007 ;
Spindler and Mote 2007 ; Spindler 2006 ; Dhahbi et al.
2005 ; Corton et al. 2004 ). Another approach, which
will be discussed here, is the direct assays of
compounds for their effects on the life span of
rodents.
Longevity assays using genetically normal, healthy
rodents
In mice, a number of natural mutations, gene knock-
outs, and overexpressed transgenes are known to
extend longevity and increase health span (Selman et
al.2008 ; Taguchi et al. 2007 ; Kurosu et al. 2005 ;
Holzenberger et al. 2003 ; Flurkey et al. 2001 ;
Coschigano et al. 2000 ; Zhou et al. 1997 ; Brown-
Borg et al. 1996 ). Thus, potential therapeutic targets
for life span extension exist in mammals. However,
no robustly effective, safe, and widely recognized
longevity therapeutics exist at present. The likelyreason that such drugs have not been identified is that
we have not mounted an effective search for them.
Life span studies in rodents have been used in thissearch (Table 1and Electronic supplementary material
(ESM) Table 1). More recently, this literature benefits
from the improved levels of hygiene used in animalhusbandry (e.g., see Sebesteny 1991 ). For example,
several older studies in Table 1appear to report data
consistent with the presence of infectious agents inthe rodent colony (Ferder et al. 1993 ; LaBella and
Vivian 1978 ; Sperling et al. 1978 ). Despite these
improvements, the design and implementation of
rodent life span studies could be improved further.
Table 1and ESM Table
1summarize and evaluate
all of the rodent life span studies we found using
repeated key word searches of the online databases. In
ESM Table 1, under the heading “Evaluation, ”we
present our evaluation of the study results. ESM
Table 1presents 106 life span studies performed with
healthy rodents. We excluded from this table 20rodent life span studies performed with melatonin,
which are contradictory in their results and which
have been reviewed elsewhere (Anisimov et al. 2006 ).
Despite the fact that the effects of caloric restric-
tion on life span were described 76 years ago (McCay
et al. 1935 ), drug screening studies which regulate or
measure food consumption are rare. We found only
six studies which measured food consumption andalso found life span extension (Liang et al. 2010 ;
Caldeira da Silva et al. 2008 ; Cai et al. 2007 ; Stoll et
al.1997 ; Yen and Knoll 1992 ; Cotzias et al. 1977 ).
These were deprenyl fed to Syrian hamsters (Stoll et
al.1997 ); deprenyl and Dinh lang root extract fed to
mice (Yen and Knoll 1992 ); dinitrophenol fed to
normal mice of a short-lived strain (Caldeira da Silva
et al. 2008 );
L-dopa fed to male mice (Cotzias et al.
1977 ); marine collagen peptides fed to Sprague –
Dawley rats (Liang et al. 2010 ); and reduced
advanced glycation end product-containing standard
mouse diet fed to mice (Cai et al. 2007 ). These are the
only studies in the literature showing an increase in
rodent longevity for which the potential effects of
“voluntary ”CR on life span can be confidently
excluded. Four studies which controlled or measured
caloric intake found no change in life span with
various treatments (Smith et al. 2010 ; Spindler and
Mote 2007 ; Lee et al. 2004 ; Pugh et al. 1999b ).
Six other studies found life span extension and
reported the effects of the treatments on body weightas a surrogate measure of food consumption (Table 1
and ESM Table 1). However, there are demonstrated
instances in which a discordance was found betweenbody weight and food consumption, making body
weight a potentially unreliable surrogate measure of
caloric consumption (see below). These treatmentsare: coenzyme Q10 administered orally to male
Wistar rats fed a diet high in polyunsaturated fatty
acids (Quiles et al. 2004 );Ginkgo biloba extract112 AGE (2012) 34:111 –120 | 82cfe32fe3b91f79483b71434b7fe2f6ad5a63d0 | page_0001 |
administered orally to male F344 rats (Winter 1998 );
green tea polyphenols administered in drinking waterto male C57BL/6 (B6) mice (Kitani et al. 2007 ); 2-
mercaptoethanol administered orally in food to male
BC3F1 mice (Heidrick et al. 1984 ); PBN fed to B6
male mice (Saito et al. 1998 ); and piperoxane
administered by injection to F344 rats (Compton et
al.1995 ).
Twenty studies found extended life span, but
potential CR effects cannot be excluded based on
the data available (Table 1and ESM Table 1). Many
of these reports include statements to the effect that
no change in body weight (most common) or food
intake (rarely) occurred, but no data or analysis areshown. No indication is given of whether the data
were anecdotal or systematic, when and how manytimes during the study the measurements were taken,
or what statistical methods were used to analyze the
data. These uncertainties, coupled with the potentialfallibility of weight as a biomarker for food consump-
tion (see below), make these studies less persuasive.
Twenty-nine other studies report life span exten-
sion by treatments, but the body weight and/or food
consumption data presented in the publication suggest
that induced voluntary CR was responsible for thelongevity effects observed. Of the remaining studies,
nine would be difficult to repeat because the
composition, preparation, or mode of delivery of theTable 1 Summary appraisal of the published life span studies using healthy rodents
106 separate life span studies where compounds were administered to normalarodents (less 20 contradictory melatonin studies)b
6 studies found life span extension and showed food consumption was not responsible by measuring it
Deprenyl administered orally to female hamsters
Deprenyl and Dinh lang root extract administered to miceDinitrophenol administered to a short-lived, normal mouse strain
L-dopa administered orally to male mice
Marine collagen peptides extended the mean life span of Sprague –Dawley rats
Reduced advanced glycation end products present in standard rodent diet
6 studies found life span extension and reported no change in weight, with data shown or details given (this list excludes studies which
showed no change in food consumption listed above)
Coenzyme Q10 administered orally to male Wistar rats a diet high in polyunsaturated fatty acidsGinkgo biloba administered orally to F344 rats
Green tea polyphenols administered orally to mice
2-Mercaptoethanol administered orally to mice
PBN administered orally to micePiperoxane administered by injection to rats
20 studies report LS extension but potential CR effects cannot be excluded
Body weight and/or food consumption called “unchanged ”, but no data given or data given but not analyzed statistically (e.g., it
remains unclear whether the data are anecdotal or systematic, when and how many times during the study measurements were
taken, the means and standard deviations of the measurements, and what statistical methods were used to analyze the data?)
29 studies report results that are likely due to induced “voluntary ”CR
Body weights or food consumption were less than those of controls or neither was reported
36 studies report no effect on life span
3 studies report reduced LS9 studies would be difficult to repeat or have methodological or reporting confounds that render their data of uncertain significance
Only English language publications were reviewed
bNormal in this context means the animals had no known genetic defect leading to an artificially decreased life span and were not
given a physical or chemical treatment to stress the animals and shorten their life span
bIf a publication reports the testing of a compound or compounds using more than one group of animals, each test was listed and
counted separately. If a compound was tested in more than one publication, these studies are counted separately. If a compound haddifferential effects on the lifespan of mice of different strains in a single report, these effects were counted under multiple categories.AGE (2012) 34:111 –120 113 | 82cfe32fe3b91f79483b71434b7fe2f6ad5a63d0 | page_0002 |
treatment agents are published in difficult to obtain
journals or are not reported.
Food consumption should be measured
Body weight is often used in longevity studies as a
surrogate measure of caloric consumption (Table 1
and ESM Table 1). The vast majority of the studies
using body weight as a surrogate do not report the
methods used or the results obtained (ESM Table 1).
Thus, the reader cannot know whether the conclu-
sions drawn used systematic or anecdotal measures.
The number of animals weighed, the number of timesthey were weighed, and the statistics used are not
reported. Such problems are evident in two reports
from the NIA Interventions Testing Program (NIH-ITP; Harrison et al. 2009 ; Strong et al. 2008 ; Miller et
al.2007 ). While the studies are unusually robust in
many aspects of experimental design, including large
cohorts of genetically heterogeneous mice of both
sexes tested at multiple sites, most of their reportsgive no details regarding body weight measurements
(Harrison et al. 2009 ; Strong et al. 2008 ; Miller et al.
2007 ). Thus, NIH-ITP investigators have reported that
the same concentration of rapamycin fed to HET3
mice produced either no effect on body weight
(methods and data unspecified; Harrison et al. 2009 )
or a 6% or 10% decrease in body weight (for females
and males, respectively; Miller et al. 2011 ). Thus, it is
possible that the mice in the first study experienced anundetected reduction in body weight. It also is unclear
whether the reductions in body weight found in the
second study were due to reduced caloric intake.Thus, “voluntary CR ”may have played a role in the
longevity effects observed. While one may seek
further information from these investigators at thistime, our publications are likely to outlive us.
Body weight is an unreliable surrogate measure of
caloric intake. Both dietary
L-dopa and dietary
dinitrophenol reduce body weight without changing
food consumption (Caldeira da Silva et al. 2008 ;
Cotzias et al. 1977 ). A drug-induced discordance
between body weight and food intake may not be
uncommon. We found five agents or combination of
agents that significantly decreased body weight andfour agents or combination of agents which signifi-
cantly increased the body weight of mice fed
isocalorically (unpublish ed results). For example,mice fed food supplemented with four doses of
nordihydroguaiaretic acid (NDGA) experienced anapproximately dose-responsive decrease in body
weight without a corresponding decrease in food
consumption (Fig. 1). Food was packed in 1-g pellets
and fed daily. Food intake for each of these groups was
carefully monitored and recorded. Any uneaten food,
even when masticated and dropped into the bedding,was readily identifiable by shape, color, and texture.
Quantitatively, the group fed the highest dose of
NDGA weighed the same or less than a 20% calorie-restricted (20% CR) group at most times during the
study (Fig. 1). Others have reported, without showing
data, that mice consuming NDGA-supplemented dietsad libitum have no change in body weight relative to
controls (Strong et al. 2008 ). Thus, it is possible that
the mice in this published study maintained their body
weight by increasing food consumption. Feeding
measured quantities of food and monitoring of itsconsumption ensures that life span data are not
confounded by changes in caloric consumption. This
reduces the likelihood of CR-related changes in lifespan (Merry 2002 ; Compton et al. 1995 ).
Monitoring of both food consumption and body
weight will identify instances in which a compoundproduces a discordance between them. Drug-induced
changes in activity, metabolic rate, or intestinal
absorption of calories might lead to such a discor-dance, which would not be detected by monitoring of
only body weight. Once detected, a discordance can
be investigated further using measurements of spon-taneous activity, metabolic rate, and absorption of
calories (e.g., Westbrook et al. 2009 ; Adams et al.
2006 ). Thus, measured feeding coupled with body
weight monitoring is a much more robust approach to
life span studies than body weight monitoring alone.
Methods for isocaloric feeding
In the author ’s experience, measuring food con-
sumption is less difficult and expensive than it is
sometimes assumed to be. In an ongoing longevitystudy involving 2,400 mice, measured feeding is
~9% of total costs. To deliver a known amount of
food to each cage conveniently, we use the methoddescribed by Weindruch and colleagues (Pugh et al.
1999a ). The food (AIN-93M) and any additional
components are cold-packed into 1-g pellets by Bio-114 AGE (2012) 34:111 –120 | 82cfe32fe3b91f79483b71434b7fe2f6ad5a63d0 | page_0003 |
Serv (Frenchtown, NJ). These round pellets are
conveniently scooped into a 1.6-cm inner diameterPlexiglas tube fitted with a commercially available
plastic cap. Tubes cut to different lengths are used to
deliver different numbers of pellets to the cages.
If a supplemented diet is under-consumed, flavoring
can be added, the supplement can be changed to an
agent with similar actions, the supplement concentrationin the food can be reduced, or, if desired, the amount of
food given to a control group can be decreased to that of
the test agent. We slightly underfeed all the mice in ourstudies to insure that all food is eaten.
Healthy, long-lived rodents, such as an F1 hybrid
or a more genetically heterogeneous mouse
should be used for compound screening
During our survey of the literature, we found many
reports of life span-based compound screening per-
formed with short-lived or enfeebled rodents (data not
shown). By “enfeebled, ”we mean natural or selected
rodent lines that have genetic (or possibly epigenetic)
changes that reduce longevity and health relative to
their unaltered parental or control strains. For exam-ple, many studies utilized senescence-accelerated
prone mouse strains (SAMP1 through SAMP9) to
rapidly screen for longevity therapeutics (Rodriguezet al. 2008 ; Li et al. 2007 ; Umezawa et al. 2000 ;
Boldyrev et al. 1999 ; Kumari et al. 1997 ; Edamatsu et
al.1995 ; Zhang et al. 1994 ). SAMP mice suffer from
the early onset of a spectrum of age-related patholo-
gies which abbreviate their life span. We found only
one study in which the effectiveness of an agent wastested in both an SAMP mouse (SAMP8) and in one
of its associated control mouse strains (SAMPR1;
Fig. 1 Isocaloric feeding of diets containing NDGA reduced
body weight without altering food consumption. The left axis
shows the mean bimonthly weights of dietary groups fed AIN-93M diet with no additional additives ( empty square ) or AIN-
93M diet containing NDGA at 1.5-g/kg diet ( empty triangle ),
2.5 g/kg diet ( empty diamond ), 3.5-g/kg diet ( empty hexagon ),
and 4.5-g/kg diet (
); a 20% CR diet ( empty downturned
triangle ); or a 40% CR diet ( circle ). The mice were shifted
from chow feeding to the defined diets at 12 months of age.
The right axis shows the percentage of the kilocalories fed to
each group of mice which were actually consumed for the
group fed AIN-93M diet with no additives ( filled square ); AIN-
93M diet containing NDGA at 1.5-g/kg diet ( filled triangle ),
2.5-g/kg diet ( filled diamond ), 3.5-g/kg diet ( filled hexagon ),
and 4.5-g/kg diet ( ); a 20% CR diet ( filled downturned
triangle ); or a 40% CR diet ( filled circle ). The symbols
representing food consumption are superimposed in the figure,making them difficult to distinguish because the mice ateessentially all their food. Error bars and symbols for statistical
significance were omitted for the sake of clarity. The body
weights were significantly different than controls, as judged bythe non-parametric Mann –Whitney test, for the NDGA 1.5-g/kg
diet group at 22 months ( P<0.01), 24 months ( P<0.001),
26 months ( P<0.01), 28 months ( P<0.05), and 30 months ( P<
0.01); for the 2.5-mg/kg diet group at 18 months ( P<0.01), 20 –
26 months ( P<0.001), and 28 months ( P<0.01); for the 3.5-
mg/kg diet group at 20 and 22 months ( P<0.01), 24 and
26 months ( P<0.001), and 28 and 30 months ( P<0.01); and for
the 4.5-mg/kg diet group at 16 months ( P<0.01) and 18 –
30 months ( P<0.001). The mice were shifted from chow
feeding to the defined diets at 12 months of age. These studies
used male B6C3F1 mice (Harlan Breeders, Indianapolis)randomly assigned to treatment groups at 3 weeks of age. At12 months of age, the mice were shifted from ad libitum chow
feeding (Diet no. 5001, Purina Mills, Richmond, IN) to daily
feeding with either 13.3 kcal/day per mouse of control diet(AIN-93M, Diet no. F05312; Bioserv, Frenchtown, NJ) or daily
feeding with an identical quantity of control diet supplemented
with the indicated concentrations of NDGA. The 20% CRgroup was shifted from ad libitum chow feeding to 11 kcal/dayper mouse of AIN-93M 20% Restricted Diet (Diet no. F06298,
Bioserv). The 40% CR group was shifted from ad libitum chow
feeding to 11 kcal/day per mouse of AIN-93M 20% RestrictedDiet for 2 weeks and thereafter to 7.46 kcal/day of AIN-93M40% Restricted Diet (Diet no. F05314, Bioserv). The diets for
the 20% and 40% calorically restricted groups were fortified so
the mice received fewer calories in the form of carbohydratethan the other groups, but approximately equal amounts of fat,
protein, vitamins, and minerals. All mice were fed the amounts
indicated daily. Food consumption was monitored at the time offeeding, and any food left was noted and removed. With rareexceptions, all food was eaten each day. The drugs were mixed
with powered diet and cold-pressed into 1-g pellets by Bio-serv.
The food was stored moisture free at 4°C until used. The micedrank acidified (pH 4.0) tap water ad libitum and weremaintained on a 12-h light/dark cycle at 22°C. Cohorts of 296
negative control mice and 36 CR and treated mice were utilizedAGE (2012) 34:111 –120 115 | 82cfe32fe3b91f79483b71434b7fe2f6ad5a63d0 | page_0004 |
Zhang et al. 1994 ). In this study, a botanical which
extend the life span of SAMP8 mice did not extend thelife span of the control strain. Similarly, resveratrol was
reported to extend the life span of high fat-fed, obese,
and diabetic mice (Baur et al. 2006 ). While this article
has been cited by many as evidence that resveratrol can
extend mammalian life span, the results did not
translate to healthy mice (Pearson et al. 2008 ). Thus,
screening agents in enfeebled rodents has not yet been
shown to facilitate the identification of compounds
which extend the life span of healthy animals.
For these reasons, studies designed to identify
longevity therapeutics should utilize long-lived mice,
such as an F1 hybrid or more genetically heteroge-neous mouse. F1 hybrid mice, which are widely
available, are genetically heterozygous at all loci for
which their parents are heteroallelic. They are more
disease- and stress-resistant and have larger litters and
longer life spans than their inbred parental lines(Flurkey et al. 2009 ). HET3 mice, which are produced
using a four-way crossbreeding scheme, are more
genetically heterogeneous than F1 mice and are usedby the NIH-ITP. However, they are more difficult to
produce and have shorter life spans than some F1
mice. For example, B6C3F1 mice have a mean lifespan of about 915 days (Spindler and Mote 2007 ;
Pugh et al. 1999b ; Smith and Walford 1977 ), while
HET3 mice have a mean life span of about 800 days(Strong et al. 2008 ). Longer life spans are usually
regarded as signs of greater vigor. Outbred mice,
which are even more genetically heterogeneous thanHET3 mice, are more vigorous and less expensive
than inbred mice (Flurkey et al. 2009 ). However, they
have the disadvantage of being genetically undefined.Because each mouse is genetically unique, study
results can be more varied and thus more difficult to
reproduce.
Chemically defined diets should be used
for gerontological research
There are three general categories of rodent diets:
cereal-based (non-purified), purified, and chemical-
l yd e f i n e d( K o z u le ta l . 2008 ;R e e v e se ta l . 1993 ;
American Institute of Nutrition ad hoc Committee onStandards for Nutritional Studies 1977 ). The major-
ity of the studies summarized in ESM Table 1
appeared to have used non-purified or purifiedcereal-based diets. However, cereal-based diets are
often variable in compositio n (American Institute of
Nutrition ad hoc Committee on Standards for
Nutritional Studies 1977 ), and this variability, and
the presence of trace contaminants, can stronglyinfluence experimental results (Kozul et al. 2008 ;
Jensen and Ritskes-Hoitinga 2007 ; Allred et al.
2004 ; Thigpen et al. 2004 ,2003 ). For example,
Prolab-RMH 1000 rodent chow contains appreciable
quantities of polychlorinated dibenzo- p-dioxins and
dibenzofurans, probably from pesticide residues(Schecter et al. 1996 ). Purina Laboratory Rodent
diet 5001 (LRD-5001) contains high concentrations
of methylmercury and a mixture of inorganic andorganic arsenic compounds at a concentration 36
times the EPA-recommended level for drinking
water (Kozul et al. 2008 ; Weiss et al. 2005 ). The
specifications for diets such as NIH-31 allow
manufacturers to use any of a number of sources ofprotein, including fish meal, a possible source of
arsenic and other contaminants, or soy, a possible
source of pesticide residue. Thus, purified, defineddiets are preferable.
Use of a positive control is highly desirable
Many life span studies are published without the
benefit of a positive control group, such as a 40% CR
group. If none of the compounds tested in a study
extend life span, the possibility cannot be excludedthat the rodents would not respond to a longevity
treatment under the study conditions. Few reviewers
would endorse the publication of negative biochem-ical data without the inclusion of a positive control to
show that the assay was working. This should be
similarly important for rodent life span studies.
Dosages of agents tested in rodents
The dosages at which potential therapeutics are tested
in rodents must balance a number of competingtheoretical and practical issues. Ideally, one would
like to know that a therapeutic level of the agent is
maintained throughout a life span study. Of course,the ideal therapeutic level of an agent is not known
for most life span studies. Furthermore, food intake,
body volume, intestinal absorption, and metabolism116 AGE (2012) 34:111 –120 | 82cfe32fe3b91f79483b71434b7fe2f6ad5a63d0 | page_0005 |
may change with age. Monitoring the blood levels of
an agent throughout a life span study would bedifficult and expensive. Group sizes which would
make rodents available for testing throughout the
study are often impractical.
Several approaches can mitigate these limitations.
Published studies with well-defined treatment endpoints
can be used to estimate dosages. In this way, one can bereasonably certain that a therapeutic level of the agent is
achieved. Initial signs that a dosage is too high, such as
reduced food intake or inattention to grooming, can beused to adjust dosages “on the fly. ”Where rodent
studies cannot be found, equivalent rodent dosages can
be calculated from human dosages using default cross-species scaling factors (Reagan-Shaw et al. 2008 ;U S
EPA 2005 ; Rhomberg and Lewandowski 2004 ;D o u r -
son et al. 1992 ,1996 ; Dourson and Stara 1983 ). These
scaling factors are often used to set and access drug
dosages in human and animal studies (e.g., Chalastaniset al. 2010 ). Empirically, small animals have been
found to require larger dosages per gram body weight
than larger animals. These differences are due topharmacokinetic differences (e.g., rates of uptake,
metabolism, and clearance of compounds) and to
pharmacodynamic differences (e.g., rates of damageto macromolecules, cellular repair and regeneration,
signaling cascades, and proliferative responses) be-
tween small and large animals. One widely usedscaling formula increases the human dosage in milli-
grams per kilogram body weight/day by tenfold to
obtain the equivalent mouse dosage. Another scalingfactor also in use is based on the 3/4 power of body
weight [i.e., (milligrams/kilogram body weight)
3/4/
day], which leads to equivalent mouse dosages thatare about sevenfold higher than the equivalent human
dosages. While these calculations were initially devel-
oped for chemotherapeutics, they are also used asstarting points when human dosages must be extrapo-
lated from preclinical rodent data (Chalastanis et al.
2010 ).
Summary: the preferred design for testing
potential longevity therapeutics using mouse life
span studies
B a s e do nt h ei n f o r m a t i o nr e v i e w e da b o v e ,w e
recommend a number of design parameters essential
or highly desirable for rodent life span assays: (1) Thediets should be fed in measured amounts and
consumption monitored. Body weight should bemonitored regularly. These measurements and their
statistical analysis should be reported. (2) A long-
lived, healthy rodent strain should be used, preferablyan F1 or further outcrossed strain. (3) Chemically
defined diets should be used. They ensure the greatest
degree of reproducibility and avoid the confoundsintroduced by contaminants or compositional vari-
ability. (4) Use of a positive control is highly
desirable. Without a positive control, negative resultsare of questionable significance. We use a 40% CR
control, which also allows us to calibrate the effects of
a treatment (e.g., Fig. 1). (5) Dosages can be chosen
using treatment endpoints gleaned from the literature
or, where necessary, from human dosages using
accepted cross-species scaling factors. Use of these
methods will produce a more reliable literature on
which to base further studies.
Acknowledgments The author would like to thank Mehgan
Hassanzadah and Patricia Mote for their help in the preparationof this manuscript.
Open Access This article is distributed under the terms of the
Creative Commons Attribution Noncommercial License whichpermits any noncommercial use, distribution, and reproduction
in any medium, provided the original author(s) and source are
credited.
References
Adams SH, Lei C, Jodka CM, Nikoulina SE, Hoyt JA, Gedulin
B, Mack CM, Kendall ES (2006) PYY[3-36] administra-
tion decreases the respiratory quotient and reducesadiposity in diet-induced obese mice. J Nutr 136:195 –201
Allred CD, Allred KF, Ju YH, Clausen LM, Doerge DR,
Schantz SL, Korol DL, Wallig MA, Helferich WG
(2004) Dietary genistein results in larger MNU-induced,estrogen-dependent mammary tumors following ovariec-
tomy of Sprague –Dawley rats. Carcinogenesis 25:211 –
218
American of Nutrition ad hoc Committee on Standards for
Nutritional Studies (1977) Report of the American
Institute of Nutrition ad hoc Committee on Standards for
Nutritional Studies. J Nutr 107:1340 –1348
Anisimov VN (1982) Carcinogenesis and aging. III. The role of
age in initiation and promotion of carcinogenesis. Exp
Pathol 22:131 –147
Anisimov VN, Popovich IG, Zabezhinski MA, Anisimov SV,
Vesnushkin GM, Vinogradova IA (2006) Melatonin asantioxidant, geroprotector and anticarcinogen. Biochim
Biophys Acta 1757:573 –589AGE (2012) 34:111 –120 117 | 82cfe32fe3b91f79483b71434b7fe2f6ad5a63d0 | page_0006 |
Subsets and Splits