ReviewsinComputationalChemistry,Volume17.EditedbyKennyB.Lipkowitz,DonaldB.Boyd Copyright(cid:1)2001JohnWiley&Sons,Inc. ISBNs:0-471-39845-4(Hardcover);0-471-22441-3(Electronic) Reviews in Computational Chemistry Volume 17 Reviews in Computational Chemistry Volume 17 Edited by Kenny B. Lipkowitz and Donald B. Boyd NEWYORK (cid:1) CHICHESTER (cid:1) WEINHEIM (cid:1) BRISBANE (cid:1) SINGAPORE (cid:1) TORONTO Designationsusedbycompaniestodistinguishtheirproductsareoftenclaimedastrademarks. InallinstanceswhereJohnWiley&Sons,Inc.,isawareofaclaim,theproductnamesappear ininitialcapitalorALLCAPITALLETTERS.Readers,however,shouldcontacttheappropriate companiesformorecompleteinformationregardingtrademarksandregistration. Copyright(cid:1)2001byJohnWiley&Sons,Inc.Allrightsreserved. Nopartofthispublicationmaybereproduced,storedinaretrievalsystemortransmittedinany formorbyanymeans,electronicormechanical,includinguploading,downloading,printing, decompiling,recordingorotherwise,exceptaspermittedunderSections107or108ofthe1976 UnitedStatesCopyrightAct,withoutthepriorwrittenpermissionofthePublisher.Requeststothe PublisherforpermissionshouldbeaddressedtothePermissionsDepartment,JohnWiley&Sons, Inc.,605ThirdAvenue,NewYork,NY10158-0012,(212)850-6011,fax(212)850-6008, E-Mail:[email protected]. Thispublicationisdesignedtoprovideaccurateandauthoritativeinformationinregardtothe subjectmattercovered.Itissoldwiththeunderstandingthatthepublisherisnotengagedin renderingprofessionalservices.Ifprofessionaladviceorotherexpertassistanceisrequired,the servicesofacompetentprofessionalpersonshouldbesought. ISBN0-471-22441-3 ThistitleisalsoavailableinprintasISBN0-471-39845-4. FormoreinformationaboutWileyproducts,visitourwebsiteatwww.Wiley.com. Preface The aphorism ‘‘Knowledge is power’’ applies to diverse circumstances. Anyonewhohasclimbedanorganizationalladderduringacareerunderstands thisconceptandknowshowtoexploitit.Theproblemforscientists,however, is that there may exist too much to know, overwhelming even the brightest intellectual. Indeed, it is a struggle for most scientists to assimilate even a tiny part of what is knowable. Scientists, especially those in industry, are under enormous pressure to know more sooner. The key to using knowl- edge to gain power is knowing what to know, which is often a question of what some might call, variously, innate leadership ability, intuition, or luck. Attemptstomanagespecializedscientificinformationhavegivenbirthto thenewdisciplineofinformatics.Thebranchofinformaticsthatdealsprimar- ily with genomic (sequence) data is bioinformatics, whereas cheminformatics deals with chemically oriented data. Informatics examines the way people work with computer-based information. Computers can access huge ware- housesofinformationintheformofdatabases.Effectiveminingofthesedata- bases can, in principle, lead to knowledge. In the area of chemical literature information, the largest databases are producedbythe ChemicalAbstractsService(CAS) ofthe American Chemical Society (ACS). As detailed on their website (www.cas.org), their principal databases are the Chemical Abstracts database (CA) with 16 million docu- ment records (mainly abstracts of journal articles and other literature) and the REGISTRY database with more than 28 million substance records. In an earlier volume of this series,* we discussed CAS’s SciFinder software for mining these databases. SciFinder is a tool for helping people formulate queries and view hits. SciFinder does not have all the power and precision ofthecommand-linequerysystemofCAS’sSTN,asoftwaresystemdeveloped earliertoaccesstheseandotherCASdatabases.ButwithSciFinderbeingeasy *D.B.BoydandK.B.Lipkowitz,inReviewsinComputationalChemistry,K.B.Lipkowitz andD.B.Boyd,Eds.,Wiley-VCH,NewYork,2000,Vol.15,pp.v–xxxv.Preface. v vi Preface touseandwithfavorableacademicpricingfromCAS,nowmanyinstitutions have purchased it. ThisvolumeofReviewsinComputationalChemistryincludesanappen- dix with a lengthy compilation of books on the various topics in computa- tional chemistry. We undertook this task because as editors we were occasionally asked whether such a listing existed. No satisfactory list could befound,sowedevelopedourownusingSciFinder,supplementedwithother resources. Wewereanticipatingnotbeingabletoretrieveeverybookwewerelook- ing for with SciFinder, but we were surprised at how many omissions were encountered. For example, when searching specifically for our own book ser- ies,ReviewsinComputationalChemistry,severaloftheexistingvolumeswere not‘‘hit.’’Moreover,thesewerenotconsecutiveomissionslikeVolumes2–5, but ratherthey were missing sporadically. Clearly, something aboutthe data- base is amiss. Whereas experienced chemistry librarians and information specialists may fully appreciate the limitations of the CAS databases, a less experienced user may wonder: How punctilious are the data being mined by SciFinder? Certainly, for example, one could anticipate differences in spelling like Mueller versus Mu¨ller, so that typing in only Muller would lead one to not finding the former name. The developers of SciFinder foresaw this problem, and the software does give the user the option to look for names that are spelled similarly. Thus, there is some degree of ‘‘fuzzy logic’’ implemented in the search algorithms. However, when there are misses of information that should be in the database, the searches are either not fuzzy enough or there may be wrong or incomplete data in the CAS databases. Presumably, theseerrorswere generatedbytheCASstaffduring theprocessofdataentry. In any event, there are errors, and we were curious how prevalent they are. To probe this, we analyzed the hits from our SciFinder searches. Three kindsoferrorswereconsidered:(1)wrong,meaningtherewerefactualerrors in an entry which prevented the citation from being found by, say, an author search (although more exhaustive mining of the database did eventually uncover the entry); (2) incomplete, meaning that a hit could be obtained, but there were missing pieces of data, for example, the publisher, the city of publication, the year of publication, or the name of an author or editor; (3) spelling, meaning that there were spelling or typographical errors apparent in the entry, but the hit could nevertheless be found with SciFinder. In our study, about 95% of the books abstracted in the CA database were satisfac- tory; 1% had errors that could be ascribed to the data being wrong, 3% had incomplete data,and1% hadspellingerrors.Theseerrorratesare lower lim- its. There almost certainly exist errors in spellings of authors’ names or other errors that we did not detect. Concerning the wrong entries, most of them were recognized with the help of books on our bookshelves, but there are probably others we did not notice. Many errors, such as missing volumes of Preface vii a series, became evident when books from the same author or on the same topic were listed together. If we noticed a variation of the spelling of an author’s name from year to year or from edition to edition, especially when Russian and Eastern European names are involved, we classified these entries as being wrong if the infraction is serious enough to give a wrong outcome in a search. If one is looking for books by I. B. Golovanov and A. K. Piskunov, for example, one needs to search also for Golowanow and Piskunow, respectively. The userdiscoversthatthespellingoftheirco-authorchangesfromN.M.Sergeev to N. M. Sergejew! Should the user write Markovnikoff or Markovnikov? (Bothspellingscanbefoundincurrentundergraduateorganicchemistrytext- books.) More of the literature is being generated by people who have non- English names. But even for very British names, such as R. McWeeney and R. McWeeny, there are misspellings in the CAS database. Perhaps one of the more frequent occurrences of misspellings and errors is bestowed on N. Yngve O¨hrn. Some of the CAS spellings include: N. Yngve Oehrn, Yngve Ohrn,YnaveOhrn,andevenYngveOehru!Therealsomaybeerrorsconcern- ing the publishing houses, some not very familiar to American readers. For example, aside from variability in their spellings, the Polish publisher Panst- woweWydawnictwoNaukowe(PWN)isenteredasPANinoneoftheentries of W. Kolos’ books, whereas the others are PWN. Some of this analysis might be considered ‘‘nit-picking,’’ but an error is certainlyseriousifitpreventsauserfromfindingwhatisactuallyinthedata- base. Our exercises with SciFinder suggest that it would be helpful if CAS strengthened their quality control and standardization processes. Cross- checking and cleaning up the spellings in their databases would allow users to retrieve desired data more reliably. It would also enhance the value of the CAS databases if missing data were added retrospectively. So, what level of data integrity is acceptable? The total percentage of errors we found in our study was 5%. Is this satisfactory? Is this the best wecanhopefor?Hopefullynot,especiallyasmorepeoplebecomedependent on databases and the rate of production of data becomes ever faster. Clearly, there is a need for a system that will better validate data being entered in the most used CAS databases. It is desirable that the quality of the databases increases at the same time as they are mushrooming in size. A Tribute Manyprominentcolleagueswhohaveworkedincomputationalchemis- try have passed away since about the time this book series began. These include (in alphabetical order) Jan Almlo¨f, Russell J. Bacquet, Jeremy K. Burdett, Jean-Louis Calais, Michael J. S. Dewar, Russell S. Drago, Kenichi Fukui, Joseph Gerratt, Hans H. Jaffe, Wlodzimierz Kolos, Bowen Liu, Per- OlovLo¨wdin,AmatzyaY.Meyer,WilliamE.Palke,BernardPullman,Robert viii Preface Rein, Carlo Silipo, Robert W. Taft, Antonio Vittoria, Kent R. Wilson, and MichaelC.Zerner.*Thesescientistsenrichedthefieldofcomputationalchem- istryeachinhisownway.Threeoftheseindividuals(Almlo¨f,Wilson,Zerner) were authors of past chapters in Reviews in Computational Chemistry. Dr.MichaelC.ZernerdiedfromcanceronFebruary2,2000.Othertri- butes have already been paid to Mike, but we would like to add ours. Many readers of this series knew Mike personally or were aware of his research. Mike earned a B.S. degree from Carnegie Mellon University in 1961, an A.M. from Harvard University in 1962, and, under the guidance of Martin Gouterman, a Ph.D. in Chemistry from Harvard in 1966. Mike then served his country in the United States Army, rising to the rank of Captain. After postdoctoralworkinUppsala,Sweden,wherehemethiswife,heheldfaculty positions at the University of Guelph, Canada, and then at the University of Florida. At Gainesville he served as department chairman and was eventually nameddistinguishedprofessor,apositionheldbyonly16otherfacultymem- bers on the Florida campus. Probably, Mike’s research has most touched other scientists through his development of ZINDO, the semiempirical molecular orbital method and *Afterthisvolumewasinpress,thefieldofcomputationalchemistrylostatleastfourmore highlyesteemedcontributors:G.N.Ramachandran,GildaH.Loew,PeterA.Kollman,and DonaldE.Williams.Wealongwithmanyothersgrievetheirdemise,butremembertheir contributionswithgreatadmiration.ProfessorRamachandranlenthisnametotheplotsfor displayingconformationalanglesinpeptidesandproteins.Dr.LoewfoundedtheMolecular ResearchInstituteinCaliforniaandappliedcomputationalchemistrytodrugs,proteins,and othermolecules.ShealongwithDr.JoyceJ.Kaufmanwereinfluentialfiguresinthebranch ofcomputationalchemistrycalledbyitspractitioners‘‘quantumpharmacology’’duringthe 1960sand1970s.ProfessorKollman,likemanyinourfield,beganhiscareerasaquantum chemistandthenexpandedhisintereststoincludeotherwaysofmodelingmolecules.Peter’s workinmoleculardynamicsandhisAMBERprogramarewellknownandhelpedshapethe fieldasitexiststoday.ProfessorWilliams,anauthorofachapterinVolume2ofReviewsin ComputationalChemistry,wasfamedforhiscontributionstothecomputationofatomic chargesandintermolecularforces.Drs.Ramachandran,Loew,andWilliamswereblessed withlongcareers,whereasPeter’swascutshortmuchtooearly. AlthoughseveralofPeter’sstudentsandcollaboratorshavewrittenchaptersforReviews inComputationalChemistry,Peter’sassociationwiththebookserieswasareviewhewrote aboutVolume13.AsatributetoPeter,wewouldliketoquoteafewwordsfromthisbook review,whichappearedinJ.Med.Chem.,43(11),2290(2000).Whilealwaysobjectivein hisevaluation,Peterwasalso generousin praiseof theindividualchapters(‘‘abeautiful pieceofpedagogy,’’‘‘timelyandinteresting,’’‘‘valuable,’’and‘‘anenjoyableread’’).Hehad theseadditionalcommentswhichweshalltreasure: ThisvolumeofReviewsinComputationalChemistryisofthesame very high standard as previous volumes. The editors have played a keyroleincarvingoutthedisciplineofcomputationalchemistry,hav- ingorganizedaseminalsymposiumin1983andhavingservedasthe chairmenofthefirstGordonConferenceonComputationalChemistry in1986.Thus,theyhaveabroadperspectiveonthefield,andthearti- clesinthisandpreviousvolumesreflectthis. WewouldliketoaddthatPeterwasaninvitedspeakerattheSymposiumonMolecular Mechanics (held in Indianapolis in 1983) and was co-chairman of the second Gordon ResearchConferenceonComputationalChemistryin1988.AswepointedoutinthePre- faceofVolume13(p.xiii)ofthisbookseries,noonehadbeencitedmorefrequentlyin ReviewsofComputationalChemistrythanPeter.Peter—andtheothers—willbemissed. Preface ix program for calculating the electronic structure of molecules. To relieve the burdenofprovidingusersupport,Mikeletasoftwarecompanycommercialize it,anditiscurrentlydistributedbyAccelrys(ne´eMolecularSimulations,Inc.) In addition, a version of the ZINDO method has been written separately by scientists at Hypercube in their modeling software HyperChem. Likewise, ZINDO calculations can be done with the CAChe (Computer-Aided Chemis- try) software distributed by Fujitsu. Several thousand academic, government, andindustriallaboratorieshaveusedZINDOinoneformoranother.ZINDO is even distributed by several publishing companies to accompany their text- books, including introductory texts in chemistry. Mikepublishedover225researcharticlesinwell-respectedjournalsand 20bookchapters,oneofwhichwasinthesecondvolumeofReviewsinCom- putational Chemistry. It still remains a highly cited chapter in our series. In addition, Mike edited 35 books or proceedings, many of which were asso- ciated with the very successful Sanibel Symposia that he helped organize with his colleagues at Florida’s Quantum Theory Project (QTP). If you have never organized a conference or edited a book, it may be hard to realize how much work is involved. Not only was Mike doing basic research, teaching (includingatworkshopsworldwide),andservingonnumerousuniversitygov- ernance and service committees, he was also consulting for Eastman Kodak, Union Carbide, and others. A little known fact is that Mike is a co-inventor of eight patents related to polymers and polymer coatings. Mike’s interests and abilities earned him invitations to many meetings. He attended four Gordon Research Conferences (GRCs) on Computatio- nal Chemistry (1988, 1990, 1994, and 1998).* Showing the value of cross- fertilization,Mikesubsequentlybroughtsomeofthetopicsandideasofthese GRCstotheSanibelSymposia.MikealsolongedtoserveaschairoftheGRC. The GRCs are organized so that the job of chair alternates between someone from academia and someone from industry. The participants at each biennial conference elect someone to be vice-chair at the next conference (two years later),andthenthatpersonmovesuptobecomechairfouryearsaftertheelec- tion.Mikewasacandidatein1988and1998,whichwereyearswhennonin- dustrial participants could run for election. He and Dr. Bernard Brooks (NationalInstitutesofHealth)wereelectedco-vice-chairsin1998.Sadly,Mike diedbeforehewasabletofulfillhisdream.AttheGRCinJuly2000,y tributes werepaidtoMikebyDr.TerryR.Stouch(Bristol-MyersSquibb),Chairman, andbyDr.Brooks.Inaddition,Dr.JohnMcKelvey,Mike’scollaboratordur- ing the Eastman Kodak consulting days, beautifully recounted Mike’s many fine accomplishments. Ourscienceofcomputationalchemistryowesmuchtothecontributions of our departed friends and colleagues. *D.B.BoydandK.B.Lipkowitz,inReviewsinComputationalChemistry,K.B.Lipkowitz andD.B.Boyd,Eds.,Wiley-VCH,NewYork,2000,Vol.14,pp.399–439.Historyofthe GordonResearchConferencesonComputationalChemistry. ySeehttp://chem.iupui.edu/rcc/grccc.html. x Preface This Volume As with our earlier volumes, we ask our authors to write chapters that canserveastutorialsontopicsofcomputationalchemistry.Inthisvolume,we havefourchapterscoveringarangeofissuesfrommoleculardockingtospin– orbit coupling to cellular automata modeling. Thisvolumebeginswithtwochaptersondocking,thatis,theinteraction and intimate physical association of two molecules. This topic is highly ger- mane to computer-aided ligand design. Chapter 1, written by Drs. Ingo Muegge and Matthias Rarey, describes small molecule docking (to proteins primarily).Theauthorsputthedockingproblemintoperspectiveandprovide a brief survey of docking methods, organized by the type of algorithms used. Theauthorsdescribetheadvantagesanddisadvantagesofthemethods.Rigid dockingincludinggeometrichashingandposeclusteringisdescribed.Tomo- delnaturemoreclosely,onereallyneedstoaccountforflexibilityofbothhost and guest during docking. The authors delineate the various categories of treating flexible ligands and explain how each works. Then an evaluation of how to handle protein flexibility is given. Docking of molecules from combi- natoriallibrariesisdescribednext,andthevalueofconsensusscoringiniden- tifying potentially interesting bioactive compounds from large sets of molecules is pointed out. Of particular note in Chapter 1 are explanations of the multitude of scoring functions used in this realm of computational chemistry: shape and chemical complementary scoring, force field scoring, empiricalandknowledge-basedscoring,andsoon.Theneedforreliablescor- ing functions underlies the role that docking can play in the discovery of ligands for pharmaceutical development. ThefirstchaptersetsthestageforChapter2whichcoversprotein–protein docking.Drs.LutzP.EhrlichandRebeccaC.Wadepresentatutorialonhow to predict the structure of a protein–protein complex. This topic is important becauseasweentertheeraofproteomics(thestudyofthefunctionandstruc- ture of gene products) there is increasing need to understand and predict ‘‘communication’’ between proteins and other biopolymers. It is made clear at the outset of Chapter 2 that the multitude of approaches used for small molecule docking are usually inapplicable for large molecule docking; the generation of putative binding conformations is more complex and will most likely require new algorithms to be applied to these problems. In this review, the authors describe rigid-body and flexible docking (with an emphasis on methods for the latter). Geometric hashing techniques, confor- mational search methodologies, and gradient approaches are explained and put into context. The influence of side chain flexibility, backbone confor- mational changes, and other issues related to protein binding are described. Contrasts and comparisons between the various computational methods are made, and limitations of their applicability to problems in protein science are given. Preface xi Chapter 3, by Dr. Christel Marian, addresses the important issue of spin–orbit coupling. This is a quantum mechanical relativistic effect, whose impact on molecular properties increases with increasing nuclear charge in a waysuchthattheelectronicstructureofmoleculescontainingheavyelements cannotbedescribedcorrectlyifspin–orbitcouplingisnottakenintoaccount. Dr.Marianprovidesahistoryandthequantummechanicalimplicationsofthe Stern–Gerlach experiment and Zeeman spectroscopy. This review is followed by a rigorous tutorial on angular momenta, spin–orbit Hamiltonians, and transformationsbasedonsymmetry.Tipsandtricksthatcanbeusedbycom- putational chemists are given along with words of caution for the nonexpert. Computational aspects of various approaches being used to compute spin– orbiteffectsare presented,followedbyasection oncomparisons ofpredicted and experimental fine-structure splittings. Dr. Marian ends her chapter with descriptions of spin-forbidden transitions, the most striking phenomenon in which spin–orbit coupling manifests itself. Chapter 4 moves beyond studying single molecules by describing how one can predict and explain experimental observations such as physical and chemical properties, phase transitions, and the like where the properties are averagedoutcomesresultingfromthebehaviorsofalargenumberofinteract- ing particles. Professors Lemont B. Kier, Chao-Kun Cheng, and Paul G. Seyboldprovideatutorialoncellularautomatawithafocusonaqueoussolu- tion systems. This computational technique allows one to explore the less- detailed and broader aspects of molecular systems, such as variations in speciespopulationswithtimeandthestatisticalandkineticdetailsofthephe- nomenon beingobserved.Themethodologycantreatchemicalphenomenaat a level somewhere between the intense scrutiny of a single molecule and the averaged treatment of a bulk sample containing an infinite population. The authorsprovideabackgroundonthedevelopmentanduseofcellularautoma- ta, their general structure, the governing rules, and the types of data usually collected from such simulations. Aqueous solution systems are introduced, and studies of water and solution phenomena are described. Included here are the hydrophobic effect, solute dissolution, aqueous diffusion, immiscible liquids and partitioning, micelle formation, membrane permeability, acid dis- sociation,andpercolationeffects.Theauthorsexplainhowcellularautomata areusedforsystemsoffirst-andsecond-orderkinetics,kineticandthermody- namic reaction control, excited state kinetics, enzyme reactions, and chroma- tographic separation. Limitations of the cellular automata models are made clearthroughout.Thiskindofcoarse-grainedmodelingcomplementstheideas consideredintheotherchaptersinthisvolumeandpresentsthebasicconcepts needed to carry out such simulations. Lastly, we provide an appendix of books published in the field of com- putationalchemistry.The number islarge, morethan1600. Rather thansim- plypresenting all these books inone longlist sortedby authororby date, we have partitioned them into categories. These categories range from broad