The Wittig olefination reaction and modifications involving phosphoryl-stabilized carbanions.pdf

Subscriber access provided by ADOLOR CORPORATIONChemical Reviews is published by the American Chemical Society.1155 SixteenthStreet N.W.,Washington,DC 20036The Wittig olefination reaction and modifications involvingphosphoryl-stabilized carbanions.Stereochemistry,mechanism,and selected synthetic aspectsBruce E.Maryanoff,and Allen B.ReitzChem.Rev.,1989,89(4),863-927 DOI:10.1021/cr00094a007Downloaded from http:/pubs.acs.org on January 30,2009More About This ArticleThe permalink http:/dx.doi.org/10.1021/cr00094a007 provides access to:Links to articles and content related to this articleCopyright permission to reproduce figures and/or text from this articleChem.Rev.1989,89,863-927 863 The Wittig Olefination Reaction and Modifications Involving Selected Synthetic Aspects Phosp horyl-Sta bilized Carbanions.Stereochemistry,Mechanism,and BRUCE E.MARYANOFF and ALLEN B.REITZ Chemical Research Department,Janssen Research Foundation,Spring House,Pennsylvania 19477 Received May 23,1988(Revised Manuscript Received January 17,1989)Contents I.Introduction 11.Phosphonium Ylides A.Stereochemistry and Mechanism 1.1,2-Oxaphosphetanes and Betaines as 2.Nonstabilized Ylides 3.Stabilized and Semistabilized Ylides 4.General Discussion Stereochemistry 1.Wittig Reactions with Anomalous 2.Selected Synthetic Applications Intermediates B.Selected Synthetic Aspects Involving Stereochemistry(1979-1987)111.Phosphoryl-Stabilized Carbanions A.Phosphonate Carbanions 1.Mechanistic Aspects 2.Preparation of Phosphonate Reagents 3.Different Types of Phosphonates in 4.Newer Reaction Technologies 5.Intramolecular Reactions B.Phosphine Oxide Carbanions C.Other Phosphoryl and Thiophosphoryl Synthesis Carbanions IV.Concluding Remarks A.Phosphonium Ylides B.Phosphoryl-Stabilized Carbanions V.Acknowledgments VI.References and Notes 863 863 864 864 868 874 876 88 1 88 1 896 90 1 902 902 903 904 910 91 1 912 916 916 916 917 917 917 I.Introduction There was a time in organic chemistry when the olefination of ketones and aldehydes was faced with some trepidation.Because of limited synthetic meth-ods,as recently as 30 years ago,the chemist had to contend with two isomer problems,that of double-bond position and that of double-bond geometry.Landmark paperslia published by Wittig and co-workers in the early 1950s disclosed a means for the preparation of alkenes with unambiguous positioning of the double bond,based on the reaction of aldehydes or ketones with phosphonium ylides(eq 1).Because of its effec-tiveness and generality,the Wittig reaction became widely used and thereby changed the course of olefin synthesis for all time.3 Indeed,the development of the Wittig reaction helped to usher in the modern era of 0009-266518910789-0863$06.50/0 organic synthesis,wherein positional selectivity,ste-reoselectivity,and chemoselectivity are of paramount importance to,and under the sensitive and responsive control of,the synthetic pratitioner.The 1960s witnessed major advances in the Wittig reaction and in Wittig-style olefinations.The stereo-chemistry and mechanism of the Wittig reaction were investigated,and a complementary reaction involving phosphoryl-stabilized carbanions was developed.Al-though several reviews have documented the state of the Wittig and related reactions,up to as recently as 1985,$17 key recent facets,especially in the areas of stereochemistry and mechanism,have inspired us to compose this article.Our emphasis will be placed on information added to this topic from 1978 to the present.Also,we will present new synthetic highlights from this period of time to provide a full,up-to-date discussion.This review will be limited to reactions of aldehydes and ketones;it will not deal with ester-or amide-type substrates.*I I.Phosphonium Yiides The conventional Wittig reaction entails the reaction of a phosphonium ylide with an aldehyde or a ketone(eq 1).This olefination method has enjoyed wide-R R O+(R73P=C/Y-”X;R+(R73P=0(I)aldehyde phosphorus alkenes phosphlne or ketone gide oxide spread prominence and recognition because of its sim-plicity,convenience,and efficiency.“14 Yet,despite such venerable attributes,the attractiveness of the Wittig reaction in synthesis may often hinge on effective stereocntrol.JJJ High selectivity for(2)-or(E)-alkenes is available,depending on the particular cir-cumstances,such as the type of ylide,type of carbonyl compound,or reaction conditions.8J1 Phosphorus ylides have been loosely classified ac-cording to their general reactivity.“Stabilized”ylides have strongly conjugating substituents(e.g.,COOMe,CN,or S0,Ph)on the ylidic carbon and usually favor the production of)-alkenes,“semistabilized”(or“moderated”)ylides bear mildly conjugating substitu-ents(e.g.,Ph or allyl)and often give no great preference one way or the other,and“nonstabilized”ylides lack such functionalities and usually favor(2)-alkenes.Of course,there are notable,if not glaring,exceptions to these generalized stereoselectivities,some of which will 0 1989 American Chemical Society 864 Chemical Reviews.1989.Vol.89.No.4 Maryanoff and Rekz persisted with respect to this high preference for con-trathermodynamic(2)-alkenes in,for example,reactions of triphenylphosphorus nonstabilized ylides with al-dehydes.This characteristic has attracted the curiosity of chemists for decades and stimulated attempts to arrive at a truly satisfying mechanistic explanation.The other two classes of ylides are also interesting from a mechanistic standpoint.For example,one may wonder:Is the strong preference for the(E)-alkenes with many stabilized ylides a consequence of kinetic or thermodynamic control?To define the source of such stereocontrol,organic chemists have resorted to mech-anistic studies and the pursuit of reaction intermediates.These two subjects will be addressed in section 1I.A.1.1,2-Oxaphosphetanes and Betaines as Intermediates Regarding intermediates in the reaction,Wittig first mentioned a four-membered cyclic phosphorane(a 1.2-oxaphosphetane)early on$however,he soon came to favor a zwitterionic phosphorus betaine(eq 2)?,21 P Bruce Maryanoff was born in Philadelphia,PA,in 1947 and has reslded in that region for most of his life.He received his B.S.(1969)and Ph.D.(1972,with Professor Robert Hutchins)degrees from Drexei University.Afler postdoctoral studies with Professor K u l Mklow at Princeton University(1972-1974).ha jo4ned the staff of McNeil Laboratories(renamed McNeil aceutical in 1980).Following reorganization of the Johnson 8 Johnson pharmaceutical Sector in 1987.he became part of the Belgium-headquarlered Janssen Research Foundation(JRF)Worldwue.He now holds the position of Distinguished Research Fellow in the laboratories of JRF-US in Spring House.PA.His major research interests have entailed stereochemistry and mechanisms of organic reactions.conformational analysis.monosaccharide chemistry,heterocyclic chemistry.selective reduction processes.and drugs for treating disorders of the central nervous system.He has published over 80 scientific papers,is an inventor on 15 US.Patents.and was recipient 01 the 1984 Section Award of the American Chemical Society,Philadelphia Section.He dedicates ais review to his wife Cynthia.f Ailen Reitz was born in Aiameda.CA.in 1956.He received his B.A.degree(1977)from the University of California at Santa Barbara and his Ph.D.degree(1982)from the University of CalC fornia at San Diego,working with Professor Murray Goodman.Afler a 1-year postdoctoral stint with Dr.Maryanoff at McNeil Pharmacautiii.ha was appointed to the medicinal chemistry staff.He is currently a Principal Scientist in the Janssen Research Foundation at Spring House.PA.His major research interests include development of new synthetic methods.stereocontrol in cyclization reactions.and synthesis of monosaccharides for ther-apeutic applications.H e has ca.30 scientific publications to his credit.He dedicates this review to Evelyn,his wife of 10 years.with whom he has two children.Darryl and Meredith.emerge in the subsequent discourse.A.Stereochemistry and Mechanism The nonstabilized class of phosphorus ylides is par-ticularly significant mechanistically in that the ther-modynamically less stable(2)-alkene is often produced preferentially.8112.9.20 In fact,a certain mystique has R,P=CHR/+RCHO+2 It-0 X p l W h t O S This view gained broad acceptance in the mid and by 1970 the mechanism of the Wittig reaction was commonly expressed in terms of two steps:(1)nucleophilic addition of the phosphorus ylide to the carbonyl compound to give a betaine species and(2)irreversible decomposition of the betaine to give alkene and phosphine oxide(eq 2).9,22-26 Although the 1,2-oxaphosphetane was widely considered to be a transi-tion state between betaine and final products,rather than a distinct intermediate,two reviews were careful to present the oxaphosphetane as a possible interme-diate.?,9 Greater weight had been placed on the dipolar be-taine intermediate because of certain experimental observations:(1)the formation in situ of stable adducts between betaines and lithium halide salts,(2)the trapping of betaines as-hydroxy phosphonium salts by addition of acid at low temperature,and(3)the pronounced effect of lithium salts on alkene stereo-hemistry.2 However,in 1973 Vedejs reported for the first time that oxaphosphetanes are the sole observable intermediates by 31P NMR spectroscopy in conventional reactions of nonstabilized ylides at low temperature.n Vedejs positive observations,along with the lack of evidence for uncomplexed.betaines,revo-lutionized impressions about the Wittig reaction mechanism for most organic chemists.Subsequent work by the Vedejs group,reported in 1981:O estab-lished 1,Z-oxaphosphetanes as principal intermediates in a variety of reactions involving nonstabilized phos-phorus ylides and aldehydes or ketones.In the 1980s,Maryanoff and co-workers extended the oxaphosphe-The Wittig Olefination Reaction tane paradigm by detecting and quantitating the short-lived diastereomeric intermediates in Wittig re-actions of nonstabilized ylides and aldehyde.-In general,the 31P NMR signal for pentacoordinate phosphorus in oxaphosphetanes occurs far upfield(e.g.,from-50 to-80 ppm)relative to the reference(at 0 ppm),while the signal for tetracoordinate phosphorus in a betaine would be expected to occur downfield(e.g.,from 10 to 50 ppm).The relative importance of oxaphosphetanes vs be-taines as intermediates has been a persistent concern.To date,true betaines have never been observed di-rectly in any Wittig reaction.The precipitates formed in certain lithium salt reactions22 are really betaine-lithium halide adducts,which should not be confused with salt-free(i.e.,uncomplexed)betaines.Such complexes can arise by the addition of a lithium salt(mild Lewis acid)across the P-0 bond of a preformed oxaphosphetane,20 as opposed to direct formation.By the same token,the production of 0-hydroxy phos-phonium salts on treatment of Wittig reactions with acid at low temperature can be attributed to oxaphos-phetanes,which are readily cleaved by addition of HX across the P-0 bond.20v2231 Even in cases where the betaine must be generated first,such as in deprotona-tion of a 0-hydroxy phosphonium salt with base(eq 3),only oxaphosphetane species have been noted by NMR s p e c t r o c o p y.(3)R R Chemical Reviews,1989,Vol.89,No.4 865 way,not involving betaine species.The oxaphosphe-tane,which formed through a very small energy barrier,was a local minimum on the energy surface.MNDO calculations have been performed on the re-action of H3P=CHMe or Me3P=CHMe with Me-CH0.1bJ2 A transition state entailing advanced C-C bond formation was deemed most germane,others be-ing much higher in energy.In this model,a P-O gauche transition state was clearly preferred to a P-0 anti one(by at least 4 kcal/mol).Betaines were found to be much higher in energy than oxaphosphetanes(by ca.20 kcal/mol).(a)Decomposition of Oxaphosphetanes.In principle,oxaphosphetanes can fragment in two directions:to ylide and aldehyde(retro-Wittig reaction)or to alkene and phosphine oxide.In practice,both of these pro-cesses have been recorded,and their relative proportion appears to be dependent on oxaphosphetane structure,particularly the substituents appended to the ring,and on reaction conditions,such as in response to the presence of lithium salts.202223313437-40 On the whole,these reaction pathways represent a dynamic state that is balanced by the relative rates for the various pro-cesses(e.g.,see section II.A.2.b).Failure to detect re-versal experimentally does not necessarily mean that it is nonexistent,just that its rate is noncompetitive with the forward reaction,the facility of which poses an obstacle to complete elucidation of the Wittig re-action mechanism in many cases.Generally,oxaphosphetanes are thermally unstable;they readily disintegrate to alkene and phosphine oxide below room temperature.Reasonable decomposition rates for various oxaphosphetanes derived from non-stabilized ylides have been documented at-30 to 0 oC.20130v31934a The adduct from MeCH=PPh3 and PhCHO(presumably mostly cis oxaphosphetane be-cause of the salt-free conditions)was reported to have a half-life at-8 C of ca.30 min,and at 20 C of ca.1.5 min;oxaphosphetanes from CH2=PPh3 were found to be much more transient.20 Decomposition of oxa-phosphetanes derived from cyclobutanone and 2-nor-bornanone was the most retarded,requiring tempera-tures in excess of 0 0C.20 We have performed rate studies on the decomposition and interconversion of cis and trans oxaphophetanes,details of which will be described in section II.A.2.b.In the case of semistabilized ylides,oxaphosphetanes have generally not been detected even at temperatures as low as-100 to-80 0C,31934c and there probably can be little hope for oxaphosphetanes from stabilized yl-ides.This,of course,presents a problem for mecha-nistic studies on these ylides,which is discussed further in section II.A.3.An exceptional case,in which oxa-phosphetanes from semistabilized ylides have been observed,is mentioned there as well.34C Some oxaphosphetanes are stable enough to be iso-lated;examples of these are presented in section 11.A.1.c.The mechanism of collapse of an oxaphosphetane to ylide and aldehyde is presumably the opposite of direct condensation,given microscopic reversibility.The mechanism for decomposition of an oxaphosphetane to alkene and phosphine oxide is a separate issue of con-siderable interest.10J2p35,36.4t42 Is this process concerted or stepwise,syn or anti,in nature?Bestmann has+WX-+He+MX+HB Since the course of the Wittig reaction virtually de-mands an oxaphosphetane stage,the question arises:Does a betaine precede the oxaphosphetane stage(Wittig reaction of three distinct steps:ylide+alde-hyde-betaine-oxaphosphetane-alkene)or is the oxaphosphetane formed directly from ylide and alde-hyde?From the body of experimental data,Vedejs20g34b has argued that a four-centered transition state leading dire