The Total Synthesis of (±)-Friedelin Citation Walba, David Mark (1976) The Total Synthesis of (±)-Friedelin. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/YSVW-HF20. https://resolver.caltech.edu/CaltechTHESIS:12042017-160036925 Abstract The total synthesis of the unsymmetrical pentacyclic triterpene friedlein (10) was completed in 31 steps in an overall Refer to pdf for formulas. yield of 0.3%. The route, symbolized by the disconnections outlined in Chart 4, forms each of the 9 assymetric centers of friedelin stereoselectively. Chart 4 Refer to pdf for formulas. Item Type: Thesis (Dissertation (Ph.D.)) Subject Keywords: (Chemistry) Degree Grantor: California Institute of Technology Division: Chemistry and Chemical Engineering Major Option: Chemistry Thesis Availability: Public (worldwide access) Research Advisor(s): Ireland, Robert E. Thesis Committee: Unknown, Unknown Defense Date: 31 October 1975 Record Number: CaltechTHESIS:12042017-160036925 Persistent URL: https://resolver.caltech.edu/CaltechTHESIS:12042017-160036925 DOI: 10.7907/YSVW-HF20 Default Usage Policy: No commercial reproduction, distribution, display or performance rights in this work are provided. ID Code: 10581 Collection: CaltechTHESIS Deposited By: Benjamin Perez Deposited On: 05 Dec 2017 00:26 Last Modified: 23 Aug 2024 23:48 Thesis Files Preview PDF - Final Version See Usage Policy. 41MB Repository Staff Only: item control page

THE TOTAL SYNTHESIS OF (±)-FRIEDELIN Thesis by David M. Walba In Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy California Institute of Technology Pasadena, California 1976 (Submitted October 31, 1975) -ii- To my parents, my island brothers, and Holly. alone. Man does not live on chemistry -iiiAcknowledgements I wish to thank members of the Division of Chemistry a11d Chemical Engineering for making my graduate experience such an exciting and satisfying one. Special thanks go to Professor Robert E. Ireland for his guidance, encouragement, and friendship, and to members of The Group, past and present, for providing a most pleasant and stimulating environment. -ivAbstract The total synthesis of the unsymmet rical pentacyclic triterpene friedelin (J_Q_) was completed in 31 steps in an overall 10 yield of 0.3%. The route, symbolized by the disconnections out- 1 ined in Chart 4, fonns each of the 9 asymmetric centers of friedelin stereoselectively. -v- Chart 4 c D 0 0 friedelin OCH3 B 0 > CH 3CH o 2 e 0 25 00!3 A -8> -vi - Table of Contents INTRODUCTION DISCUSSION 9 EXPERIMENTAL SECTION 31 APP EN DIX 59 REFERENCES 71 PROPOSITIONS Abs tracts of Propositions 78 Proposition 1 79 Proposition 2 95 Proposition 3 11 0 Proposition 4 11 5 Proposition 5 127 _,_ INTRODUCTION The triterpenoids form a large class of structurally diverse naturally occurring compounds; at least 750 examples have been reported to date. 2 ' 3 A large majority of these molecules are polycyclic species which fall into one of about 30 basic structural types 2 differing in the number and array of fused rings, the structure of various appended side chains, and the array of methyl groups appearing on the skeleton . As a result of the rigid skeletal framework associated with many members of the class these structures fonn an ideal substrate upon which the synthetic chemist may test and revise methods of stereoselective directed total synthesis first put to use in the steroid field. 4 Indeed, many new techniques and much understandinghaveevolved from past efforts in triterpene total synthesis which have been ably reviewed by ApSimon. 5 Several highly imaginative and elegant total syntheses of triterpenes have been completed, 3 ' 5 and these may be divided into two basic types: (Chart 1) (1) syntheses in the onocerin (compounds .!_,£,1_,_1) and amyrin (compound i) structural groups which rely in essence upon attachment of an AB fragment to a DE fragment with formation of carbon-carbon bonds C-12a-C-13, C-13-C-14, or C-14-C-14a, followed by an acid catalyzed cyclization to form the C-6a-C-6b bond; and (2) the stereoselective 6 directed total syntheses of germanicol a (~_) and lupeol 6b(l) of the amyrin group, and alnusenone 7 (§_) and shionone 8 (2._) of the friedelin group. Though the syntheses of the first type were of a pioneering nature, every example suffers from the fact that the formation of the C-6a-C-6b bond lacks regiospecificity, is stereochemically ambiguous, and proceeds in very low yield to give a complex mixture of products from which the -2 - Chart 1: Triterpenes prepared by total synthes i s of type 1 °'- -onocer in 7( - onocerin 111/ 'h-1 tetrahyrnanol R R=H 2, olean-13(18)-ene R=~ -OH, o(., -H, ~ -amyrin hydroxyhopanone -3- Chart 1, continued: Triterpenes prepared by total synthesis of type 2 HO gennanicol lupeol alnusenone shionone -4des ired compound must be separated (excepting, of course, the synthesis of a-onocerin which has no C-6a-C-6b bond) . Those of the second type, however, were completed via elegant routes designed with maximum efficiency in mind, including regioselectivity in all C-C bond fanning steps and stereoselective introduction of each asymmetric center. Friedelin 9 (10) occupies place in the lore of terpene 0 chemistry 9b and its total synthesis has been a goal in these laboratories for almost a decade.?,B,lO,ll Friedelin is particularly inter- esting from the synthetic standpoint in that it contains, together with the difficultly accessible vicinal trans dimethyl grouping common to many pentacyclic triterpenes, other features unique to the friedelin group; specifically, the vicinal 8-dimethy l grouping at C- 4 and C-4a and the cis DE ring fusion with the synthetical ly challenging array of methyl groups on ring E. Because of its complexity friedelin was not chosen as the initial goal of total synthesis, rather two triterpenes in the same structural group, alnusenone (_~) and shionone (2) were attacked first. The initial phase of this work consisted of development of methods for the efficient synthesis of the pentacyclic diaromatic diether ll· 11 -5envisaged as a key intermediate for the synthesis of triterpenes in the friedelin group. Two routes to this compound were studied. The first 10 (Chart 2) was abandoned when the crucial alkylation of ketone 12 to give 11 with the vicinal-dimethyl grouping a C-6b and C-12b could not be performed in greater than 18% isolated yield . The second route (Chart 3) is based upon the observation in model systems that Friedel - Crafts cyclialkylation 11 of a tricyclic such as ]2_ should give predominantly the desired trans-anti-trans diether 11. Effort was then di rected toward an efficient synthesis of a tricyclic of this nature starting with the enone .J.E., and using as the method of introduction of the crucial C-6bB methyl group the Nagata procedure 12 for the stereosel ecti ve hydrocyanation of a,B-unsaturated ketones. The starting enone ]l was prepared from vinyl ketone l i and B-tetralone ]2 using methods developed previously . lOa Hydrocyanation of .J.E. using conditions known to give the thermodynamically more stable product (Et 2ApCN,¢H) produced the undesired cis - cyanoketone 20. However, when conditions of kinetic control were used (Et 3Al,HCN, THF) a dramatic increase in the amount of the desired trans cyanoketone 17 was observed (76% isolated yield), and by regenerating the enone by base catalyzed elimination of hydrogen cyanide from the cis isome r and recycling, a yield of 86% of lZ_ was realized. Thus the hydrocyanation route pro vides an elegant and efficient solution to the vicinal-transdimethyl problem. 13 Conversion of ..!l to the desired olefins 19 was accomplished by selective grignard addition to give cyano- alcohol ]_§_fol lowed by dehydration and reduction of the nitrile function. Finally, pTs OH catalyzed cyclization of }1 gave the diether 11 with the correct trans-anti-trans configuration as proven by single crystal X-ray analysis. Chart 2 - 6- ------4.t~ ------13 -7 - Chart f- 0CllzCI1 3 b - -- ···--""' 86% 16 rn cH 2 3 )Cil CH 2 3 . d c 98% CH 3 -CD> 98% Cllp 17 18 )CH CH 2 3 e 65% 19 ll aa, aqueous KOH, C8:~0H; b, (C 7H ) 3Al, HCN, THF (recycle cis-cyano ClQJ by dehyarocyanaticn 5with base); c, CH 3MgI; d , soc1 2 , pyr1d1ne; e, p-rn -C H so H, C H . ket?n~ 3 6 4 3 6 5 -8- The overall yield of this route to 11 from the vinyl ketone 14 is 25% , quite sufficient for the purposes of this work . With an efficient synthesis of the diether ll in hand, the total synthesis of alnusenone became fact with solution of the E ring problem (see discussion). This, coupled with methodology for construction of the friedelin A-ring obtained in connection with the synthesis of shionone (see Appendix) set the stage for completion of friedel in. Description of the conversion of a pentacyclic diaromatic diether similar to 11 to (±)-friedelin comprises the body of this thesis. -9- DI SCUSS ION In previous reports from these laboratories 7 •10 •11 an efficient route to the pentacyclic diether ll was described . A pentacyclic diaro- matic compound of this type, with the trans-anti-trans BCD ring fusion 8 aln usenone GI 3 21 10 friedelin and correct array of methyl groups, was envisaged as a useful key intermediate for synthesis of pentacyclic triterpenes in the friedel in structural group , and the successful completion of alnusenone 7 (~) proved t he vi ability of this approach . The present report describes the conversion of a simila r intermediate to the structurally sl ightly more complicated triterpene f riedelin (lQ) . The route used for the synthesis of friedelin closely parallels that used for alnusenone in many respects, and is outlined by the series of disconnections presented in Chart 4. The basic philosophical steps A-0 will be discussed in brief initially and will serve as a f r amework for discussion of the details of the synthesis: (A) the preparation of key intermediate _g]_ in a convergent manner from the vinyl ketone 24 an d the B- tetralone ~pa rallels exactly the synthesi s of ll from Ji and _!_i. Since these transformations are discussed in detail elsewh ere (see -10- Chart 4 c D > 0 0 f riedelin OCH 3 B 0 > CHfHzO OCH 3 A > -11- Introduction and Ref. 7) they are not included here; (8) methodology for the differentiation of the A and E rin~s of similar diaromatic precursors had been developed in connection with the alnusenone work. In this case, the A ring was functi.onalized initially to give the enone 23; (C) construction of the friedelin A-ring system from the enone 23 represented at the beginning of this work the principal synthetic challenge. The novel methods developed for accomplishment of this difficult vicinal dialkylation have been reported in connection with the synthesis of shionone (~). 8 Once the A-r ing problem was solved, simple Birch reductionlOa,l4 afforded the required enone 22; (D) the complet ion of friedelin involved first differentiation of the carbonyl groups of~ then completion of the E ring using methods reported for an analogous conversion in the alnusenone synthesis . Step A; synthesis of the pentacyclic diether 21 The diether _gJ_ was obtained in 25% overall yield from the vinyl ketone 24, or in 11 % overall yield from m-hydroxycinnamic acid from which 24 was prepared , in a manner analogous to that used for the prepa rat i on of 11 . Step B; synthesis of the pentacyclic enone _gl_ The crucial selective reduction of t he A ring of _gJ_ in the presence of an aromatic E ring is outlined in Chart 5, and is based upon the higher reduction potential of a phenoxide ion relative to that of a simple anisole. 7, 15 Thus the phenol _g§_ became the initial goal. Treatment of diether _g]_ with lithium diphenylphosphide in refluxing THF 16 ' 7 afforded the required phenol in high yield. However, selective Birch -12- Chart sa CH 3 -21 OH a b 96% CH 3CH 20 58% 26 c_,_Q__ 0 EtO a 27 )zP - + a,(C H Li, TI-IF; b , Li, NH , DME, EtOH; 3 6 5 HCl, ETOH, C H . 6 6 23 c, 01 1, DME; d, 5 N 3 -13reduction of phenol 26 was complicated by the fact that the phenoxide ion generated to protect the E ring with respect to reduction also caused the molecule to be only slightly soluble in the solven t used (33% DME / ammonia ) resulting in recovery of large amounts of sta rting ether 21 if normal concentrations and reaction times were used. 17 This same effect had been observed in the analogous reaction in the alnusenone series/ conversion of phenole 28 to enone 29a . In that system extensive 0 29a 28 experimentation with a variety of solvents and conditions had not improved the results obtained using simply high dilution, a large excess of lithium , and long reaction times . Prel iminary studies on reduction of phenol 26 in the presence of one equivalent of dicyclohexyl-18-Crown6 also indicated no improvement. Thus red uction of 26 as a 0. 002 molar solution in 33% DME/ammonia, with 125 molar equivalents of lithium for 9 hours, followed by remethylation of the intermediate phen ox ide '!:]_ af- forded enone ~i n 56% overall yield from diether 21 . Step C; synthesis of the pentacyclic enedione 22 An enone of type 23 appears an ideal intermediate for conversi on to a C-4S, C-4aS dimethylated ketone of type 22f4via conjugate addition 0 23 0 -14followed by alkylation. However, previous experience in these labora - toriesS, lS indicated that this transfonnation was quite difficult due to stereochemical problems which became apparent upon cloier inspection. A perspective drawing of enone 23 indicates that the double bond is quite hindered with respect to pseudo-axial (B) attack at C-4a as a result of the presence of the axial C- 14aB methyl group. Two approaches to solution of this problem were brought to fruition . 8 The first, based upon stereoselective Simmons-Smith methylenation 19 of the allylic alcohol 29 to give the cyclopropylcarbinyl alcohol 30 29 suffered from the difficulties observed in obtaining allylic alcohol _g2_ by reduction of an enone of type _g]_. The second app roach, outlined in Chart 6, derives from the proposition that acid catalyzed cyclization of an acetylenic alcohol of type 11 would proceed from the a-face (directed by the C-14aS methyl group) to yield the desired C-4aB methyl group configuration, 8 • 13 a analogous to the Friedel-Crafts cyclialkylation 11 of the tricyclic olefins 12_ to give the trans -antitrans configuration in pentacyclic diether 11. This route seemed Chart 6a -15- b 79% 0 CH 3 __d_ _______ 98% 00-13 TFAO 34 Li 37 -16- Chart 6, continueda 37 LiO \% "/47% OCH 3 CH 3 e HO 0 40 b 182% OCH 146%- 3 d,b,f 48% 0 d,b Y73% ,. 0 a a, CJ-1 3 1, THF; b , HCl, EtOH, C H ; 6 6 c, Zn-Ag, CH2I2 , 1HF, EtzO; d, Li, NH , THF, EtOH; e, KOtBu , Et 0; 3 2 Cr0 3-2pyr, CH 2Clz- f, -17- particularly attractive in view of the fact that recent work 20 i ndicated that an enol trifluoroacetate of type 34 would result under appropriate conditions, and 34 appears an ideal intermediate for the regioselective introduction of the required C-4 methyl group . Indeed, good precedent existed for this conversion in that an intermediate in the above - mentioned but discarded scheme, starting with the allylic alcohol, the enol acetate 12, was converted via enolate formation 21 followed by methylenation 22 and subsequent acid catalyzed ring opening to the dimethylated ketone l§_. Thus an efficient route to the alcohol 11 was required, and in fact, was available based upon the route of Eschenmos er et ~. 23 • 13 a involving fragmentation of an a,S epoxy hydrazine derivative. In execution, Eschenmoser cleavage of the derived epoxyketone 31 using p-toluene sulfonhydrazide in acetic acid-dichloromethane solution afforded, after careful choice of conditions, satisfactory yields of the acetylenic ketone 32. Cleavage using the recently reported modifications of Corey and Sachder 24 proceeded in comparable -18- yields, but gave~ contaminated with a highly colo red byproduct which could not be separated by chromatography or crystallization . sired tert . The de- alcohol }l was fanned in high yield upon treatment of 32 with methyllithium, and trifluoroacetic acid catalyzed ring closure indeed produced the expected enoltrifluoroacetate 34, also in high yield . It should be noted that the stereochemical as si gnment here · · usse d in · previous · res t s on f irm groun d , as disc reports. 8, l 3a With the enoltrifluoroacetate 34 in hand, fonnation of the desired dimethylated system seemed a trivial extension of work carried out on the enol acetate 35. However, treatment of 34 with methyl- lithium in a fashion analogous to that used with 35 afforded none of the desired enolate '}]__as evidenced by our inability to trap this species with Simmons -Smith reagent, te r t . butyl dimethyl chlorosilane, or acetic anhydride . Expe riments performed on model enoltrif luoro- acetates indic ated that this result was not an artifact of wet solvents or reagents, but apparently resulted from essentially quantitative protonation of the enolate by the initially formed trifluoromethyl methyl ketone . This problem was fina lly circumvented by use of lithium diisopropylamide to effect t he required cleavag e. This treatment pre- sumably results in initial fonnation of trifluoro-N ,N,dimethylacetamide which cannot protonate the enolate. The method of alkylation of enolate anion 37 received much attention, and a detailed discussion of this conversion seems justified . Though inspection of the perspect ive drawing of enolate 37 indicates - 19that attack of an electrophile (e . g., iodomethane) from the a-face (axial) should not be abnormally hindered, studies done on a model system indicated that direct methylation of this enolate gave a mixture LiO 'GCH 3 37 which always contained a substantial amount of unalkylated ketone, to gether with mono and polyalkylated products . Presumably alkylation of the enolate in this case is slow enough to allow exchange between sta rting enolate and mono-alkylated ketone to compete to an appreciable extent. In principle, if a dilute solution of the enolate 37 we re to be added very slowly to a large excess of iodomethane , then a maximum yield of mono-alkylated ketone 39 should result. In practice, even under care- fully controlled conditions, no better than a 50% yield of 12._ was ever obtained, and the yields in the reaction were quite variable, wi th different amounts of unalkylated ketone 38 (R 1= R2= H) , and polyalkylated ketone 38 (R 1= R2= H or CH 3) resulting. A possible explanation for this behavior becomes apparent when one observes that acid catalyzed epimeriz at ion of the C-4a methyl group of 4a-methyl ketone 12._ proceeds to give -20the C-4S epimer 28 with none of the C-4a epimer remaining at equilibrium. This effect has been explained by Elie1 25 as a consequence of puckering of the ring system due to the non-bonded interaction of the 3 methyl groups at C-4a, C-14a, and C-6b. This puckering causes the C-4a (axial) methyl group to interact with the C-14ba (axial) hydrogen to a much greater extent than is normal in conformationally rigid cyclohexanohe systems. This same puckering could be the cause of the anomalous behavior of enolate 37 with respect to direct al ky 1at i on. Though the desired ketone 28 was available via direct methylation, the only satisfactory yields obtained and the capricious nature of the reaction made the methylenation route used in the shionone work 8 seem the more attractive. Treatment of enolate 37 with the Simmons-Smith reagent 19 , 26 under carefully controlled conditions, followed by rapid workup and chromatography, afforded good yields of the cyclopropanol 27 40. Though treatment of 40 with potassium tert. butoxide in refluxing ether for long periods afforded mainly 4a-methyl ketone 12_ along with a small amount of 4S-methyl ketone 28,acid catalyzed rearrangement occurred with concomitant epimeri zati on of the C-4 methyl group to give ketone 28 directly identical to that obtained by epimerization of 4a-methyl ketone 39. It is interesting to note that in spite of the expectation that rearrangement of cyclopropanol 40 to 4S-methyl ketone ~should be nearly quantitative, only satisfactory yields (70%) were observed. Though no other products were isolated from this reaction, TLC analysis indicated the -21presence of small amounts of numerous other products in the crude reaction mixture. It seems possible that since a cation such as 41 is formed HO during the rearrangement, a series of 1,2 migrations up the backbone as indicated may account for the relatively low yield of this step. Rearrangements of this kind are well known in the friedelin group. 3 •9 With the ketone 28 in hand, completion of step C is straightforward. Birch reduction of 413-methylketone to give a 3Q-hydroxy enone followed by Collins oxidation afforded the yield . enedione 22 in acceptable However , since base catalyzed ring opening of cyclopropanols was known to require rather vigorous conditions (vide supra) direct Birch reduction of the cyclopropanol 40 was attempted. Indeed, treatment of 40 with lithium in DME-ammonia afforded a crude dihydroaromatic compound which upon treatment with acid gave enedione £g_ in good yield. Thus the key enedione £g_ was available from enone 23 in a respectable 19% overall yield in 6 steps. Step D; synthesis of (±)-friedelin (.lQ) Conversion of the E-ring enone 42 to the enone 43 in 14% yield in 5 steps has been reported in connection with the synthesis of alusenone 7 43 (_§). -22In order to fully understand the magnitude of this achievement one must consider the structure of friedelin from a conformational standpoint. Inspection of the perspective drawing of friedelin in the all chair conformati on (44) shows that the a-methyl group at C-11 is actually one C-C bond distance from the a-methyl group at C-12b if no distortion is allowed; clearly this is an unacceptable situation . Though the actual conformation of friedelin even in the crystal is not yet known, X-ray structures of the two derivatives 45 28 a and 46 28b have been solved. Though each of these derivatives contains features which could in principle distort its conformation relative to friedelin, both analyses gave the same gen er al result, that is, a conformation in which both the D and E rings are in the boat form. Taken together, these studies indicate that fried elin itself exists mainly as the conformer 47 . 47 -23Regardless of its exact conformation, it is clear that the stereochemi cal situation in the DE ring system of friedelin and alnusenone is quite crowded, and a challenging synthetic target. For conversion of the enedione ~to friedelin , it is first necessary to differentiate the carbonyl groups at C-3 and C-10. Initially it was thought that a selective ketalization would serve this purpose. How- ever, upon treatment of 22 with ethyleneglycol in refluxing benzene with 0 0 0 L + 49 p-toluenesulfonic acid catalysis, none of the desired monoketal 48 could be detected. tained. Rather, a mixture of monoketal 49 and bis ketal 50 was ob- Apparently the hindered nature of the carbonyl at C-3, at least with respect to axial attack (see structure 47), relative to the steri -· cally uncrowded environment at C-10 is enough to overcome the electronic preference for ketalization of a saturated carbonyl over an a,S-u nsaturated one. 29 In view of this unsatisfactory result another route was attempted which delayed the necessary differentiation. This route, based on a se- lective oxidation, is outlined in Chart 7. Treatment of enedione 22 with lithium tri-tert. butoxyaluminum hydride 30 in refluxing THF resulted in reduction of both carbonyl groups to give the 3-S, 10-S dihydroxy olefin -24- Chart 7a OH 0 a b ----ti. + 52 + 53 SS -25~ in high yield, analogous to the initial step used in the alnusenone series. It is interesting to note that none of the epimaric alcohol products could be isolated from this reaction. Hydride attack on C-3 occurs almost exclusively from the a (equatorial) face, directed by the C-4aS-methyl group, to give the S (axial) alcohol, while attack on C-10 occurs almost exclusively from the a (axial) face to give the S {equatorial) alcohol. Formation of the C-10 hydroxyl group here is critical since the next step, Simmons-Smith methylenation 17 of the diol ~' is directed by the C- 108 hydroxyl to put the carbon atom in on the B-face, affording~ with formation of the elements of the DE cis ring fusion. At this point, it was envisaged that conditions could be found to effect the necessary differentiation at C-3 and C-10 by selective oxidation of the cyclopropyl carbinyl hydroxyl group to give the hydroxy ketone 53. Abundant precedent for the selective oxidation of allylic alcohols in . t s 31 using . a variety . t he presence o f sec . hyd roxy l groups exis of reagents. However, treatment of the diol g with manganese dioxide, the most widely used reagent for such selective oxidations, under carefully controlled conditions, resulted in the formation of a mixture of starting diol g di one~' and both mono-oxidation products i'.! and 54. sequent studies using the cyclopropylcarbinyl Sub- alcohol 56 as a model for the E ring and natural friedelane-38-ol CiZ) as a model for the A ring indicated that neither use of different grades of manganese friedel inol -26di oxide, dichlorodicyanoquin ne in benzene, nor aluminum isopropoxide in acetone would distinguish between the hydroxyl groups of~ on electronic grounds. The dial 52 was also treated with Seloxette (Cr0 3 on graphite) in refluxing toluene in the hope that the hydroxyl groups could be distinguished on steric grounds, but only a low yield of 54 was observed. With the singular failure of the selective oxidation approach, a different tack was taken based upon the observation that treatment of the endione 22 with lithium tri-tert. butoxyaluminum hydride under more selective conditions (THF,0°,l hour) afforded the hydroxy enone 58 in excellent yield. This successful route is outlined in Chart 8. With the differentiation of the carbonyl groups at C-3 and C-10 accomplished by selective reduction, efforts were directed toward a method of protection of the C- 3 hydroxyl of 58. Work on the model alcohol 57 indicated that cyclopropylmethy lation 32 with cyclopropylcarbinyl bromide in THF using sodium hydride or n-butyl lithium as base in the presence or or absence of HMPA was not satisfactory, affording only sta rting material and decomposition products. Formation of the tetrahydropyranylet her (THP ether) of 58 occurred to us early on, but was not attempted initi ally because of supposed complications arising from the formation of diastereomers and the instability of THP ethers. However, upon treatment of§]_ in dihydropyran-dichloromethane solution with either p-toluene- sulfonic acid in benzene or a trace of phosphorylchloride, a good yield of the THP ether was isolated as a crystalline derivative stable to silica gel chromatography and comprising one spot on analytical TLC. -27- Chart 8a 0 a b 95% 73% 0 OH c .. 90% d 72% f e - - - -ll<P 91% TI-IPO a o a, Li(OtBu)3AlH , 1HF, C6H6' 0 ; b, DHP, POC1 3' CHzClz; c, Li(OtBu)3 AlH, 1HF, C6H6 , refl.; d , Zn-Ag, CHziz, EtzO; e, Cr03-2pyr, CttzClz; f, KOtBu , CH 3I, 1HF . -28- Chart 8, continueda OH g h .... ---- ~ 1HPO OPO(NMez)z ... 71% j l 1HPO TI-IP .£§ - f 100% ag, Li, NH3, 1HF, tBuOH; h, ClPO(NMez)z, DME, HMPA, nBuLi; l~ Li; EtNHz , tBuOH; j, MeOH, THF, pCH3C6H4S03H . -29- When attempted on the hydroxy enone 58 the THP formation reaction proved slightly capricious , but the protected hydroxy enone 59 could be obtained in fair yield . Reduction of .§2._ proceeded smoothly to give the allylic alcohol 60 . Stereoselective methylenation to give cyclopropylcarbinyl alcohol 61 followe d by Col°lins ox1dation gave the desired cyclopropyl ketone 62 in good yield. Regioselective alkylation of 62 afforded the gem. dialkylcyclopropylketone 63, the cyclopropyl ring serving to protect C-9 from alkylation. Treatment of the alkylated cyclopropylketone 63 with lithium in THF-ammonia in the presence of tert.-butyl alcohol gave the mixture of alcohol epimers 64 as judged by TLC and spectral properties. Presumably the cyclopro - pane ring is cleaved initially to form the enolate anion 68. This enolate is then protonated to give the ketone 69 which undergoes fur ther reduction to give 64. Normally dissolving metal reduction of a ketone would be expected to form the thermodynamically more favorable mixtures are reported to occur in the reduction of hindered or strained ketones. 33 In the present case, consider- epimeric alcohol, though ing the conformation situation in the E ring, formation of a mixture is not surprising. The epimeric alcohols were not separated, but were trapped with the ch l orophos pho rodi ami date reagent 34 by a s 1i gh t modi fi cation of the procedure used in the analogous reaction in the alnusenone - 30- series to fonn the mixture of TMPDA derivatives ~ in good yield. Re- ductive removal of the TMPDA grouping gave the THP ether 66 in an overall yield of 71% from 63. Cleavage of the tetrahydropyranyl grouping followed by Collins oxidation of the resulting alcohol 67 afforded (±) friedelin in high yield. -31EXPERIMENTAL SECTIONl,35 Since the experimental details for the preparation of diether 21 from the vinyl ketone 24 and the S-tetalone 25 are identical to those used for conversion of .!i and J2. to .ll (see Introduction), and the latter have been reported elsewhere, 7 they will not be included here. However, since the intermediates are new compounds, physical and spectral properties and analytical data are included. l-(2'-m-ethyoxyphenylethyl)-4,4a,9,10-tetrahydro-7-met ho xy-4aab.p . 216-225° (0 .002 mm); ir(CHC1 3) 1655 (C=O), 1610, 1580, and 1500 cm- 1(Ar); NMR(CDC1 3) ol.37(t , 3, J=7Hz9 Ar0CH C!!., ), l.47(s,3,C-4aaCH 3), 3. 77(s,3,ArOCH 3) and 4.00(q,2 , J=7Hz, 2 3 methyl-2(3H)-phenanthrone: Ar0C~CH 3 ). Anal. Calcd for c26 H30o3: C,79.96; H, 7.74 . H,7.63. Found : C, 79 . 98 ; lOas-eyano-l-(2'-m-ethoxyphenylethyl) - l ,4,4a,9 , 10,10a- hexahyd ro7-methoxy-4aa-methyl-2(3H)-phenanthrone : mp 138.5-140° (from dichloromethane-ether-hexane); ir(CHC1 3) 2240(C =N), 1715(C=O) , 1610 and 1580 cm- 1(Ar); NMR(CHC1 ) o l.26(s,3,C-4aaCH 3), l.33(t,3,J=7Hz,Ar0CH 2C.!:!_3), 3 3.66(s ,3,Ar0CH 3), and 3.97{q,2,J=7Hz, Ar0C~CH 3 ). Anal . Calcd for c27 H o N: C,77.66; H,7.48; N,3.35. 31 3 77.60; H,7.43; N,3.39. Found: C, -32 - CH 3CH 20 lOaS-Cyano-1-(2' -m-ethoxyphenylethyl )-1 ,2, 3, 4 ,4a, 9, 10, l Oa-octahydro-7-methoxy-2, 4aa-dimethyl-2-phenanthro 1: mp l 02-104 ° (from etherhexane); ir(CHC1 3) 3600(0H), 2240(C=N), 1610, 1580 and 1500 cm- 1 (Ar); NMR (CDC1 3) 6 1. 13(s,3,C4aaCH 3), l.20(s,3,C-2CH 3), l . 40(t,3,J=7Hz, ArOCH 2C!:i_3), 3. 76(s,3,ArOCH 3), and 4.06(q ,2,J =7Hz , ArOC~,CH 3 ). Anal. Calcd for c28 H35o3N: C, 77.56; H, 8. 14; N, 3.23. Found C, 77.51; H, 8.17; N, 3.27. 3 Cyano Olefin mixture: mp 105-107° (from ether-hexane); ir (CHC1 ) 3 1 2240(C=N), 1610, 1580, and 1500 cm- (Ar); NMR(CDC1 ) o 1.08, 1.20 3 (2s,rv2:1, C-4aaCH 3), l.40(t,3,J=7Hz , Ar0CH C!:i_3) , l.67(s,rv2, R C=CRCH ), 2 2 3 3. 73(s,3,ArOCH 3), 4.03(q,2,J=7Hz, Ar0C!:i_ CH 3), and 5.0-6.0(m , rv 1, 2 R2C=CHR). NMR indicative of a mixture of olefins . Anal. Calcd for c28 H33o2N: C, 80.92; H, 8. 00; N, 3.37. C, 81.00; H, 8.01; N, 3.39. Found: -333 rn3rnz Dimethyl Olefin mixture: mp 97-99° (from ether/hexane); ir (CHC1 3 ) 1610, 1580, and 1500 cm- 1 (Ar); NMR (CDC1 3) o 0.56, 0.70, 0.90, 1.06, 1. 10, l.20, 1.70 (7s, total 9H,C-2CH 3,C-4aCH 3 , C-lOaCH 3), l.40(t,3,J=7Hz, Ar0CH C~_-3)' 2 3.73(s,3,Ar0CH 3 ), 4.03(q,2,J~7Hz,Ar0C!!_2 CH 3}, and 4.6-6.4 (m, "'l, R2C=CHR). Anal. Calcd for c28 H36 o2: C,83. 12; H,8.97. Found: C, 83. 16; H, 8. 89. 3 rn3rnza 3-Ethoxy-10-methoxy-6bB,12ba,14aB-trimethyl-5,6,6aa,6b,7,8,l2b, 13,14,14a-decahydropicene(£]_): mp 152-154° (from dichloromethane-hexane); ir (CHC1 ) 1605, 1580, and 1495 cm - l ( Ar); NMR(CDC1 3 ) o 0. 60, 1.06, 1.16 3 (3s ,3 each, C-6bBCH , C-12baCH 3, C-14a8CH 3), l .36(t,3,J=7Hz, ArOCH 2C!!_3), 3 3.73(s,3,Ar0CH ), and 3.96(q,2,J=7Hz,Ar0C!!_2CH 3). 3 Anal. Calcd for c28H36 o2: C, 83.12; H, 8. 97. Found: C, 83.16; H, 8.92. -34- lO ~ Etho xy - 5,6 , 6a , 6 b a , 7 , 8 , l 2b,13,l4,14a- d ecahydro - 6a s ,12bS,l4aa. . trimethyl-2 - pi cenol (£§) . To a stirred and ice -coole d solution of 8.55 ml (49 .4 rrur1oles) of diphenylphosphine in 100 ml of dry TH F under argon were added dropwise 23.6 ml (51 . 9 mmoles) of a 2. 2 m solution of nBuli in hexane . After 10 min stirring without cool ing 10.0 gms (24.7 mmoles) of the pentacyclic diether _gJ_ was added vi a a powder funnel, and the resulting red solution was heated at reflux for 8-·l/2 hours. After cooling again with an ice bath an d addition of 4. 27 ml (24.7 mmoles) of diphenyl phosph i ne and 11 . 8 ml (25 . 9 mmoles ) of nBuli solution , reflux was continued fo r a f urther 12 hours . .The aqueous solution was cooled, poured into 350 ml of 5% HCl , and t he product isolated by ether extraction, 36 incl uding five 10% aqueous hy drogen peroxide washes . When the crude product from this and a seco nd identical run were combined and purified by silica gel ch r omatog raphy (1 :1 et her-petroleum ether) 18.6 gms (96%) of t he phenol 26 resulted as a colorless foam. Crystallization of a portion of this foam from ethe r -petro leum et he r afforde d ana lyt i cally pure mate rial whi ch melted at 202 . 5-204° (vacuum ): ir (CHC1 ) 3580(ArOH) , 3 1 and 1600 cm- (Ar) ; NMR (CDC1 3) o 0. 63, 1. 08, 1. 18 (3s, 3 each, C-6bS CH 3, :...3 5_ C-12ba CH 3, Cl4 aSCH 3) , l .38(t,3 ,J=7Hz,Ar0CH 2Ct!_3), 2.92 (broad t , 4,J=7Hz, ArC.!:!_2-),4.02(q , 2,J=7Hz,OC!:!_2CH 3), and 4.67 (broad s,l , ArO!:!_) . Anal. Ca led for c27 H34 o2 : C, 83 .03 ; H, 8. 77. Found: C, 82.81; H, 8. 80. 0 23 1 ,5,6,6aa,6b,7,8,12b ,13 , l4,l4a,l4ba- Dodecahydro-l0-methoxy-6bS, 12ba,14aS-trimethyl-3(2H)-picenone (23) . To a solution of 4.0 gms (10 . 24 mmoles) of the phenol 26 in 1800 ml of dry DME and 3600 ml of dry ammonia was added 7 cm ( 30 mg-atoms) of lithium wire with rapid stirring under argon. blue almost immediately. The solution went deep After 10 min, 46 cm (200 mg- atoms) of lithium wire was added in small pieces. The solution was stirred 10 min before 54 ml (922 mmoles) of dry 100% ethanol was added over a period of 10 min. During the next 8-1/2 hours one 46 cm port ion of lithium wire was added whenever the blue colo r fa ded. After addition of the fifth po rtion of 1 ithium another 54 ml portion of ethanol was added. A total of 280 cm ( 1260 mg-atoms) of l Hh i um wire and 108 ml (1850 mmo l es) of ethano 1 were used. After the solution had remained blue for 9 hours the excess lithium was quenched with 35 ml of meth anol , and the ammonia was removed under a stream of argon with the aid of a hot air gun. To the resulting white suspension were added 1200 ml of dry DME and 260 ml of dry methyl iodide (4.2 moles). The resulting mixture was stirred under argon for -3624 hours then filtered with suction through a celite pad. The cl ear yellow filtrate was concentrated at reduced pressure to a yellow paste which was diluted with one liter of 2% aqueous hydrochlori c acid. The crude dihydroaromatic product was isolated by ether ext raction , 36 ineluding a 10% aqueous sodium thiosulfate wash. The resul t ing yel low solid was hydrolyzed by heating under reflux with 120 ml of benzene, 600 ml of 95% ethanol and 320 ml of 5N aqueous hydrochloric acid for 40 min. After cooling to room temperature the mixture was poured in to 800 ml of 50% brine, and the crude enone fl was isolated by ether extraction. Silica gel chromatography of this material (1:1 ethe r- petroleum ether , then ether) gave 2,260 mg (58%) of fl as a slightly yellow foam . A portion of this material was crysta ll ized from cyc lo- hexane affording analytically pure material of mp 204-207° as very slightly yellow tiny crystals : ir (CHC1 3) 1660 (C=O), 1600 (C=C), and 1575 cm- 1(Ar); NMR (CDC1 3) c5 0. 60, 0.85, 1.15 (3s , 3 each, C-6b CH , 3 C-12b CH 3 , C-14a CH 3), 3. 77 (s ,3,ArOCH 3) , and 5. 90 (broad s,l~R 2 C=CHR). Anal. Calcd for c26 H34o2 : C, 82 .49; H, 9. 05. Found : C, 82.35; H, 9.17. 23 4,4a-Epoxy-l,5,6,6aa,6b , 7,8,l2b, l3,14,14a ,14ba-dodecahydro-10methoxy-6bS, l 2ba, l 4aS-tri methyl -3( 2H )-pi cenone (~_) . To a solution of 1.097 gms (2 . 9 mmoles) of the enone 23 i n 75 ml -37of di ch lov-omethane were added 100 ml of methano·1, 18 ml of 30% aqueous hydrogen peroxi de, and 9 ml of a 10% aqueous sodium hydroxide solution. Afte r stirring under argon for 18 hours the mixture, which contained a white precipitate , was po ured into 500 ml of brine, and the product was isolated by ether extraction. 36 Florisil chromatography of the crude product (CHC1 3) afforded 1.06 gms (92%)of the epoxy ketone IL as a colorless fo am . Crystallization of a porti on of this foam from cyclo- hexane gave analytically pure material as colorless needles: mp 184-186° (vacuum); ir (CHC1 3) 1700 (C=D) and 1610 cm- 1 (A r); NMR(CDC1 3) o 0. 55, 0.85 , 1.15 (3s, 3 each, C-6bCH 3 , C-1 2bCH 3, C-·1 4aCH 3), 3.20(s ,1 ,epoxy H), and 3. 78 (s ,3,ArOCH 3) . Anal. Cal cd for c26 H34o3: C, 79.15; H, 8.69. Found: C, 79.20; H, 8. 63 . 3 __..., 0 32 1S- (3'-Butynyl) - 3,4 ,4aa,4b,5,6,l0b,ll ,12,12a-decahydro-8-methoxy4bS- 10ba,12ae-trimethyl-2 (1H) - chrysenone (32) . To a dry mixture of 790 mg (2 . 0 mmoles) of the epo xy ketone I!_ and 410 mg (2.2 mmoles) of p-toluene -sulfonylhydrazid cooled with a - 20° bath and protected by an argon atmos phere was added, with stirr ing, 17 ml of a -20° solution of 1:2 glacial acetic acid -di chloromethane which had been previously degassed by alternate evacuation and ebulition with argon. After 23 hours at - 25 °, then 13 hours at room temperature, the -381ight orange mixture was poured into 500 ml of H20 and the product isolated by ether extraction, 36 including a base wash. Chromatography on Florisil (CHC1 3) afforded 595 mg (79%) of the acetylenic ketone 32 as a slightly yellow solid. One crystallization of a portion of this material from cyclohexane gave the analytical sample : mp 169-170° (vacuum); ir (CHC1 3) 3290 (C=C-H), 2110 (weak, C=C), and 1705 cm-l (C==O); NMR (CDC1 3) o 0.58, 0.75, 1.22 (3s, 3 each, C-4b CH 3 , ClOb CH 3 , C-12a CHs), and 3.75 s,3,Ar0CH 3). Anal. Calcd for c26 H34o2: C, 82.49; H, 9.05. Found: C, 82.37; H, 9.14. 32 l 13-(3 -Butynyl )-1 ,2, 3 ,4, 4aa,4b, 5 ,6, lOb, 11, 12, 12a-dodecahydro-81 methoxy-2a, 4b13, 10ba, 12as-tet ramethyl -2-ch ryseno l (33). To a stirred and ice-cooled mixture of 11 ml (20 mmoles) of 1.8M ethereal methyllithium and 24 ml of dry THF under argon was added a solution of 974 mg (2 . 57 mmoles) of the acetylenic ketone~ in 23 ml of dry THF over a period of 20 min . After 15 min stirring without cooling the excess methyllithium was carefully quenched with water . The resulting mixture was poured into 500 ml of brine, and the product was isolated by ether extraction. 36 Florisil chromatography of the crude product (CHC1 3) gave 1007 mg (99%) of the alcohol 33 as a colorless crystaline solid of mp 171-174° (vacuum). One recrystallization from n-hexane at -39- 20° gave analytically pure material as large milky prisms: mp 147-148° (vacuum); ir (CHC1 ) 3600 (OHL 3300 (C:::C - H), 1610, and 1575 cm·-l(Ar); 3 NMR (CDC1 ) 6 0.53, 0 .97, 1. 10, 1.18 (4s, 3 each, C- 2 CH , C- 4b CH , 3 3 3 C-lOb CH , C- 12a CH ), 3.75 (s,3,ArOCH ), and 6.7 - 7. 2 (m.3,ArH) . 3 3 3 Anal. Calcd for c H o : c, 82.18; H, 9.31. Found : C, 82.21; 27 38 2 H, 9 . 61. OCH 3 TFAO 3-Tri fl uoroacetoxy - 10- methoxy-4aB ,6bB, 12aa, 14a(3-tetramethyl- l ,2 ~ 4a ,5 ,6 ,6aa,6b, 7, 8, l 2a, 13, 14, l 4a, 14ba- t et radecahyd ropi cene ( 31:_). To 467 mg (1 . 184 mmoles) of the acetylenic alcohol ~1 cooled to -20° under an argon atmosphere was .added 80 ml of a -20° solution of 70:30 tri fluoroacetic acid-trifluoroacetic anhydride. After 40 min of stirring at -20° the solvents were removed at reduced pressure ("-' l mm). The res ·idue was taken up in 100 ml of ether, and after a base wash was isolated as usu ai. 36 The res ulting slightly yellow foam, which amounted to 510 mg (98%), was used in s ubsequent steps without purification: ir (CHC1 ) 1785 (C=O), 1690 (C=C), 1600 and 1575 cm- 1 (A r); NMR (CDC1 ) 3 3 6 0.53 , 0. 87, 1.03, 1. 10 (4s, 3 each, C- 4a CH , C-6b CH , C-12b CH , 3 3 3 and 5. 25 (s ,1, R C=CRH) . C- 14a CH 3 , 3.72 2 \ °" ,..-;. ---Gi> TFAO 34 40 CH3 -40- 1,2,4a,5,6,6aa,6b,7,8,l2b,l3,l4,14a,14ba-Tatradecahydro-3a, 4a-methano-10-methoxy-4aS,6bS,12aa,14aS-tetramethyl-3S-picenol (4~) To a stirred solution of lithium diisopropylamide prepared from 20.2 ml (52.6 mmoles) of a 2.60M heJcane solution of n-butyllithium and 9. 5 ml (63 mmol es) of di i sopropyl am·i ne in 85 ml of dry THF at -55° under an argon atmosphere was added a solution of 5.6 mg (1.05 mmoles) of the enoltrifluoroacetate 34 in 15 ml of dry THF over a period of 15 min. To the solution which resulted after 15 min of further stirring at -60° was added all at once 160 ml (100 mmoles) of the Sim~ons-Smith reagent prepared from the zinc- silver couple 26 [11.4 gms (175 mmoles ) of granular zinc, 88 mg of silver acetate, and 88 ml of glacial acetic acid] and 13.9 ml (172 mmoles) of diiodomethane in 158 ml of dry ether in the presence of a few strands of silver wool after the procedure of Conia. 26 After stirring at room temperature for 50 min, the· solution was poured into 300 ml of ice cold 20% aqueous ammonium sulfate sol ution containing 0.2% concentrated ammonium hydroxide . This rrrixture was di- luted with 300 ml of water, and the product was isolated by ether extraction36 including washes with the above ammoni um sulfate solution and 10% aqueous sodium thiosulfate solution . Immediate f-iltration of the oil which resulted after removal of most of the solvent with a rotary evaporator through 50 gms of basic 111 alumina (petroleum ether, then 75% ether-pet roleum ether) afforded the crude alcohol as ~n orange ' oil. Purification of this oil by chromatography on silica gel (l:l ether-petroleum ether) afforded 260 mg (60%) of the cyclopropanol 40 as a slightly yellow solid: mp-sintering at 92 ° , m 96° d(vacuum). Two -41recrystallizations of a portion of this material from n-hexane gave the ana.lytical sample as colo rless microcrystals: mp-shri nking and discoloration at 105°, m 111-112° d(vacuum); ir (CHC1 3) 3590 (OH), 1605, and 1575 cm- 1 (Ar); NMR (CDC 13) o 0. 50, 0.82, 1.09 , 1. 15 (4s , 3 each, C-4a CH 3 , C-6b CH 3 , C-12b CH 3 , C-14aCH 3) and 3.77 (s,3,ArOCH 3); Mass Spec. m/e 161(100%), 173(61), 187(47), 393(24) , 408( 15,M+ ), mass measured M+: Calcd for c28H40 o2: 408.302814. Found : 408.3029(±0.0004). Anal. Calcd for c28H40 o2 : C, 82.30; H, 9. 87 . Found: C, 82.24; H, 9. 81. 0 1 ,4,4a,5,6,6aa,6b,7,8,12a,13,14,14a,14ba-Tetradecahydro-10methoxy-4a , 4aS,6bB ,1 2aa,14aS-pentamethyl-3-2H)-picenone (39) . A. By Direct Methylation of the Enolate '}]_. To a stirred and ice-cooled solution of lithium diisopropylamide prepared from 1. 4 ml (3.l mmoles) of a 2.2M hexane solution of n-butyllithium and 0.54 ml (3 . 54 mmoles) of dry diisopropylamine in 5 ml of dry THF under argon was added 61 mg (0.123 mmoles) of the enoltrifluoroacetate 34 in 3 ml of dry THF in one min. After 15 min at room temperature the resulting dark red solution was added dropwise via syringe to a stirred and ice-cooled solution of 5 ml of dry methyl iodide in 5 ml of dry THF to which 0.5 ml of the above n-butyllithium solution had been added to insure aprotic conditions . -42After 30 min the reaction mixture was poured into 100 ml of water and the product isolated by ether extraction; 36 Silica gel chromatography (25% ether-petroleum ether) of the crude product gave 24 mg (47%) of the 4a-methyl ketone 39 as a colorless crystalline solid of mp 195-206° (vacuum): Rf on TLC (25% ether-petroleum ether) 0.23; ir (CHC1 3) 1700 cm- 1(C=D); NMR (CDC1 ) o 0.55, 0.90, 0.95, 1.17 (49,3 each, C-4a CH 3 , 3 C-6b CH , C-12b CH 3, C-14a CH 3 ), 1.14 (d,3,J=7Hz, C-4 CH 3), and 3.77 3 (s,3,AroCh 3); Mass measured molecular ion calcd. for c28H40 o2 ; 408.302814. Found: 408. 3029 ± 0. 0004. Also isolated was 7.6 mg of material homogeneous by TLC (Rf 0.31, 25% ether-petroleum-ether) but identified as a mixture of polyalkylated products by GLC analysis (215°, at least 3 peaks), NMR spectrum, and mass spectrum. B. By tert- Butoxide Catalyzed Ring Opening of the Cyclopropanol ~O. A stirred solution of the crude product (22 mg, 0.05 mmoles) from a preparation of the cyclopropanol 40 similar to that described above in 20 ml of dry ether was treated with 20 mg of potassium tert-butoxide at reflux under an argon atmosphere for 22 hours. The mixtu re was diluted with water and the product isolated by ether extraction. 36 NMR analysis of the crude material which resulted indicated the major component to be the 4a-rrethyl ketone 12_ with about 10%-30% of the 4S-epimer 28 present. Treatment of this material with ethanolic hydrochloric acid as described below resulted in complete epimerization to the 4S-methyl ketone 28. -43- 0 l,4,4a,5,6,6aa,6b,7,8,l2a,l3,l4,l4a,l4ba-Tetradecahydro-l0-methoxy4S,4aS,6bS,12aa,14aS-pentamethyl-3(2H)-picenone (28). A. By Epimerization of the 4a-Methyl Ketone 39. To a stirred solution of 43 mg (0. 104 mmoles) of the 4a-methyl ketone 39 in 1.2 ml of benzene and 6.1 ml of 95% ethanol under argon was added 3.2 ml of 5.ti_ hydrochloric acid and the mixture was heated to reflux. After 20 min the mixture was cooled, poured into 50 ml of 50% brine, and the product isolated by ether extraction. 36 The crude product thus obtained was purified by silica gel chromatography (25% ether-pet roleum ether) to afford 35 mg (82%) of the 48-methyl keto ne 28 as a colorless crystalline solid: ir (CHC1 3) 1700 (C=O), 1610, and 1580 cm- 1(Ar); NMR (CDC1 3) o 0.55, 0.75, 0. 92, 1. 17 (4s,3 each, C-4a CH 3 , C-6b CH 3 , C-12b CH 3 , C-14a CH 3), 0.91 (d,3,J~7Hz,C-48 CH ), and 3.77 (s,3,ArOCH 3) ; 3 + Mass measured M: Calcd for c28H40 o2 , 408.302814; Found, 408. 3029 ±0.0004. B. By Acid Treatment of the Cyclopropanol 40. To a stirred solution of 22 mg of the cyclupropanol 40 in 8 ml of 95% ethanol under argon was added 10 drops of concentrated hydrochloric acid. After 45 min at reflux the mixture was diluted with 40 ml of water, and the product isolated by ether extraction. 36 A slightly yellow oil resulted which amounted to 16 mg (73%). Ir and NMR spectra confirmed - 44the identity of this material as the 4S-methyl ketone 28, though some impurity peaks were present (notably in the RCH 2R region of the NMR). TLC analysis (25% ether-petroleum ether) showed one major component of Rf 0.25 with two minor components of Rf 0. 18 and 0.09. GLC analysis at 250° showed one majo r product with a 3.7 min retention time, approximately 90% pure . 1,2,3,4,4a,5,6,6aa,6b,7,8,10,11,12,12aS,12b,13 ,14,14a , 14baEicosahyd ro-4S,4aS,6bS ,12ba ,14aS-pentamethy1-10-oxo-3a-picenol . To a stirred solution of 105 mg (0.26 mmoles) of the ketone 28 in 112 ml of dry ammonia under an argon atmosphere were added 1.2 cm (5 mg atoms) of lith ium wire and 1.5 ml (25 mmoles) of dry ethanol . During the course of the next 160 min 8 additional 1.2 cm portions of lithium wire and one additional 1.5 ml portion of dry ethanol were added. Thus a tota l of 10.4 cm (45 mg atoms) of lithium wire and 3 ml (51 mmoles)of ethanol were used. To the resulting solution was added 2.5 gms (47 mmoles) of ammonium chloride and the ammonia was evaporated with a stre am of argon . The residue was diluted with 100 ml of water and the product isolated by ethe r extraction . 36 The resulting crude dihydro compound was hydrolyzed by refluxing with 3 ml of benzene, 15 ml of 95% ethanol , and 8 ml of 5!iaqueous hydrochloric acid . After 40 min the cooled mixtu re was poure d into 100 ml of 50% brine and the -45product was isolated by ether extraction, including a base wash. The resulting crude material was purified bysilicagel chromatography (40% ethylacetate-benzene) affording 59.7 mg (58%) of the hydroxy enone as a slightly yellow oil: ir (CHC1 3) 3600(0H), 1650 (C=O), and 1603 cm- 1 (C=C); NMR (CDC1 3) o 0.80, 0.98 (2s,3 each, angular CH ), 0. 88 (s,6, 3 angular CH 3), 0.93 (s,~1.5, part of C-48 CH 3), 3.3 (broad S,l,R 2C!:!_OH), and 5.8 (R 2c CRH) . This material was not further purified, but was used directly in subsequent steps. 1,4,4a,5,6,6aa,6b,7,8,l2,l2aB,l2b, 13,14,14a,14ba-Hexadecahydro4B,4aB,4bB,12aa,14aS-pentamethyl-3,10(2H,llH)-picendione (~). A. By Birch Reduction of the Cyclopropanol 40. To a stirred solution of 270 mg (66 mmoles) of the cyclopropanol 40 in 116 ml of dry DME and 230 ml of dry arrmonia protected by an argon atmosphere were added 2.96 cm (92 mg, 13.2 mg atoms) of lithium wire and 3. 85 ml (66 nmoles) of dry ethanol. During the next 160 min one 2. 96 cm portion of lithium wire was added whenever the solution decolorized. An additional 3.85 ml of ethanol was also added. A total · of 18 cm (79 mg atoms) of lithium wire and 7.7 ml (132 mmoles) of dry ethanol were used. The reaction was quenched by addition of 5 gms (94 mmoles) of arrmonium chloride and the ammonia was evaporated in a stream of argon with the aid of a heat gun. The resulting mixture was diluted -46wi th 250 ml of water, and the product i solated by ether extraction. 36 The crude dihydro compound which resulted was hydrolyzed by refluxing with 8 ml of benzene , 40 ml of 95% ethanol, and 21 ml of 5.r!_ aqueous hydrochloric acid for 40 min . Dilution of the cooled reaction mixture with 125 ml of 50% brine and isolation of the product by ether extraction, 36 including a base wash afforded crude product which was purified by medium pressure silica gel chromatography (20% ethylaceta te-benzene ). Pure endione (22) amou nting to 120 mg (46%) was obtained as a colorless crystalline solid of mp 24 7-250° (vacuum) . Also obtained were 19 mg of a mixture containing approximately 70% of the endione 22 along with overreduced material . Analytically pure material was obtained after one recrystallization of a po rtion of t he above pu re fracti on from cyclohexane: mp 247-250° (vacuum); ir (CHCls) 1700(C=O) , 1650 (C=O), and 1610 cm-l (C=C); NMR (CDC1 3) cS 0. 73 , 0.90, 0.93, 1.01 (45 , 3 each , C-4a CH , C-5b 3 CH 3, C- 8a CH 3, C-12b CH 3), 0.87 (d ,3, J ~6 Hz , C-48 CH 3) , 5.88 (b road s,l, R C=CHR) . 2 Anal. Calcd fo r c27 H40 o2: C, 81.77; H, 10.17. Found: C, 81.68; H, 10 . 23 . B. From Collins Oxidati on of the a-Hydroxy Enone To a sti rred solution of 60 mg (0 . 15 mmol es) of the a-hydroxy enone from above in l ml of dry dichloromethane under an argon atmosphere was added 7. 5 ml (1. 5 mmoles) of Co llins reagent. (This reagent was prepared by dropwise addition of 0.48 ml of dry pyridine to a stirred suspension of 300 mg of anhydrous chromium trio xide in 15 ml of dry dichloromethane under argon . The supernatant liquid was used -47after 20 min). After 10 min the dark solution was filtered through 10 gms of alumina, and the filter was washed with 75 ml of dichloromethane and 25 ml of ether. Removal of the solvents at reduced pres - sure afforded 49 mg (83%) of enedione 22 identical to that obtained above. 0 HO l,2,3,4,4a,5,6,6aa,6b , 7,8,l0, 11, 12,12aB,12b,13,l4,14a,14baEicosahydro-4B,4aB,6bB,12aa ,1 4aB -pentamethyl-10-oxo-3B-picenol (58) . To a stirred and ice-cooled solution of 135 mg (0.34 mmoles) of the enedione 22 in 26 ml of dry THF and 5. 3 ml of benzene protected by an argon atmosphere was added 430 mg (1.7 mmoles) of lithium tri-tertbutoxyaluminum hydride in 7 ml of ice - cold dry THF . After 2 hours at 0° the reaction was quenched by addition of 0.71 ml (l.79 IT111oles) of a 10% aqueous solution of sod ium hydroxide. After a further 15 min at 0° followed by 1 hour at ambient temperature the mi xture was filtered through 10 gms of silica gel, with suction. The filter was washed with 20 ml of dichloromethane and 100 ml of 1:1 ethylacetate-benzene. Re- moval of the solvents followed by medium pressure silica gel chromatography of the crude product (20% ethylacetate-benzene) afforded 130 mg (95%) of the 38-hydroxy enone 58 as colorless crystals of mp 249-252° (vacuum). Two recrystallizations of a portion of this material afforded -48the analytical sample: mp 252-255° (vacuum); ir (CHC1 3) 3600 (OH), 1655 (C=O) , and 1610 cm- 1(C=4:); NMR (CDC1 3) o 0.85, 0. 91 (2s,3 each, angular CH 3) , 0.98(s ,6,angular CH ), 3.74(s,l,w =8Hz, C-3a H), and 5.9l{s,l , 3 112 R2C=CHR). Anal. Calcd for c27 H41 o2: C, 81.35; H, 10.62. Found: C, 81.46; H, 10.50 . HO l,5,6,6a,6ba,7,8,8a,9,l0 ,l l ,12,12aa,12b,13,l4,l4a,l4bS-0ctahydro6aS, 8aS ,9S, l 2bS, l 4aa-pentamethyl - lOS( tet rahydropyran-2 -yl oxy )-3 (2H )1 pi cenone ( 59). To a stirred solution of 12 mg (0.03 mmoles) of the hydroxyenone 58 in 2 ml of dry dichloromethane and l ml of dihydropyran (freshly distil led from sodium) under argon was added 5µ1 of distilled phosphorous oxychloride. After 1. 25 ho urs another 5µ1 portion of phosphorous oxy- chloride was added. After 1.75 hours the reaction was quenched by ad- dition of 3 drops of distilled triethylamine. Remo val of the solvents at reduced pressure afforded 23 mg of the crude tetrahydropyranyl ether ~as an oily yellow solid. Silica gel chromatography (l:l ether- petroleum ether) gave 9.7 mg (67%) of the mixture of diastereomeric ethers 59 (a slightly broadened single spot by TLC, Rf 0. 19 with 1:1 ether-petroleum ether) as a slightly yellow solid of mp 214 ~ 216° (vacuum). Subsequently 86 mg of the hydroxy enone were converted into 76 mg (73%) -49of the ethers ~with mp 216-220° (vacuum) in two runs, the crude products of which were combined and purified together. One recrystallization of a portion of this material from methanol-dichloromethane afforded the analytical sample as small colorless prisms: mp 226-227° (vacuum); ir(CHC1 3) 1660 (C=O), and 1610 cm-l (C=C); NMR (CDC1 ) o 0.85, 0.91, 3 0.95, 0.98 (4s, complex, angular CH 3), 3.48, 3.69 (2s, 1/2 each, w = 112 6Hz, C-3a H), 3.50 (s,l,w =10Hz, C-6'H), 3.93 (broad m, l,C-6'H), 4.54, 112 4.70 (2s, 1/2 each, w = 6Hz, C-2'H), 5.90 (s,l,R 2C=CHR). This NMR is 112 indicative of the expected 50:50 mixture of diastereomers of the ether at C-2'. Anal. Calcd for c32 H50 o3: C, 79.62; H, 10.44. Found: C, 79.67; H, 10.43. OH TI-IPO l,2,3,5,6,6a,6ba,7,8,8a,9,10,ll,12,12aa, 12b,13,14,14a,14bBEicosahydro-6aS,8aS,9S,12bS,14aa-pentamethyl-10S(tetrahydro-2'-yloxy)3S-pi ceno 1 (60). To a stirred solution of 76.4 mg (0. 158 mmoles) of the tetrahydropyranyloxy enone ~ in 2. 5 ml of dry THF and 0. 57 ml of benzene under argon was added 400 mg (1.58 mrnoles) of lithium tri-tert-butoxyaluminum hydride in 2.5 ml of dry THF. The resulting homogeneous solution was heated under reflux for 1 hour then cooled and treated with 0.67 ml (1.66 mmoles) of aqueous 10% sodium hydroxide solution. After a further 1-1/2 hours the mixture was filtered through 5 gms of silica gel with -50suction, and the filter was washed with 100 ml of ether. Removal of solvents from the combined filtrates at reduced pressure followed by silica gel chromatography of the crude product (l :1 ether-petroleum ether) afforded 69 mg (90%) of the allylic alcohol 60 as a colorless solid of mp 239-244° (vacuum). Analytically pure material was obtaine~ after one recrystallization of a portion of the above alcohol from methanol-dichloromethane: mp 249-252° (vacuum); ir (CHC1 3) 3600 (OH), and 1650 cm- 1 (C=C); NMR (CDC1 3) 6 0.79 (s,3,angular CHS), 0.89 (m, angular CH 3), 0.93 (s,6,angular CH 3), 3.44, 3.66 (2s, 1/2 each, w = 112 7Hz, C-3a H), 3.50{s, l,w =10Hz, C-6'H), 3.94(m,l,C-6'H), 4.14 (s,l, 112 w =20Hz, C-loa 'H), 4.55, 4.67 (2s, 1/2 each, C-2 H), 5. 42 (s,l, 112 R2C=CHR). 1 Anal. Calcd for c32 H52 o3: C, 79.29; H, 10.81. Found : C, 79.22; H, 10. 75. 1HPO l,2,3,5,6,6a,6ba,7,8,8a,9,10,ll,12,12aa,l2b,13,14,14a,l4bSEicosahydro-4S,4aS-methano-6aS,8aS,9S, 12bS,14aa-pentamethyl-10S(tetrahydropyran-2 'yl oxy )-3S-pi ceno l {~.JJ . To a stirred solution of 38 mg (0.079 rnmoles) of the allylic alcohol 60 in 4 ml of dry THF under argon was added 4.3 ml of the Simroons-Smith reagent prepared from the zinc-silver couple 32 [850 mg (13 11VT1oles) of granular zinc, 6. 5 mg of silver acetate, and 6.5 ml of -51glacial acetic acid] and 0.81 ml (10 mmoles) of diiodomethane in 9.2 ml of dry ether in the presence of a few strands of silver wool afte r the procedure of Conia. 32 The colorless homogeneous mixture was heated at reflux for 1 hour then cooled and diluted with 50 ml of 20% aqueous ammonium sulfate solution containing 0.2% of ammonium hydroxide . Isolation of the product by ether extraction 36 including 20% buffered ammonium sulfate ana soa1um thiosulfate washes affo rded , after removal of most of the solvents at reduced pressu re, a colorles s oil which was immediately chromatographed on alumina (petroleum ether, then 1:1 etherpetroleum ether). The yield of cycl opropylcarbinyl alcohol §l was 28 mg (72%) as a slightly yellow solid of mp 252-257° (vacuum). One re- crystallization of a portion of this material from methanol -dichl oromethane gave analytically pure material as white microcrystals : mp 2532550 (vacuum); ir (CHC1 3 ) 3600 cm- 1(0H); NMR 0.42, 0. 59 , (2m,3,cyclopropyl H), 0.86 , 0. 94, 0.99, 1.09 (4s,complex,angular CH 3), 3.47, 3.65 (2s,l/2 each, C-3a H) . 3. 49 (s,l,C-6 H), 3.93 (m ,l,C-6' H) , 4. 17 (comp lex 1 m,l,C-lOa H), 4. 54, 4. 67 (2s,l/2 each ,C-2'H). Anal. Calcd for c33 H54o3: C, 79.46; H, 10 . 91. Found: C, 79 . 36; H, 11. 00. 11-IPO l,5,6,6a,6ba,7,8,8a ,9,l0,ll,12,l4aa,l2b,13,14,14a,14bB-0ctadecahydro-4S,4aS-methano-6aB,8aS,9S,12bS,14aa-pentamethyl-10S(tetrahydropyran-2' -yloxy)-3(2H)-picenone (62) . -52To a stirred solution of 37 mg (0.075 mmoles) of the cyclopropylcarbinyl alcohol §l in 1.3 ml of dry dichloromethane under argon was added 3.7 ml (0.746 mmoles) of Collins reagent. (This reagent was prepared by dropwise addition of 0.48 ml of dry pyridine to a stirred suspension of 300 mg of anhydrous chromium trioxide in 15 ml of dry dichloromethane under argon . The supernatant liquid was used after 20 min.) After 10 min at room temperature the dark brown mixture was fil- tered with the aid of suction through 10 gms of alumina, and the filter was rinsed with 100 ml of ether and 10 ml of dichloromethane . Removal of the solvents from the filtrate at reduced pressure afforded 33 . 7 mg (91%) of the ketone 62 as a slightly yellow solid of mp 265.5 -2 68° (vacuum) which was homogeneous by TLC (1:1 ether-petroleum ether, Rf 0.18). One recrystallization of a portion of this material from cy.clo- hexane afforded the analytical sample: mp 268.5-270° (vacuum); ir (CHC1 3 ) 1675 cm- 1(C=O) ; NMR (CDC1 3) 6 0.88, 0. 94, 1.01 (3s,complex,ang u1 ar CH 3) , 3.47, 3.66 (2s,l/2 each, C-3a H), 3. 51 (s,l,C-6 1 H) , 3.92 (m,l , C-6 1 H), 4.55, 4.67 (2s,l/2 each, C-2'H) . Anal. Calcd for c33 H52 o3: C, 79.79; H, 10.55. Foun d: C, 79.75; H, 10.51. IBPO l,5,6,6a,6ba,7,8,8a,9,l0,ll,l2,l2aa,12b,l3,14,l4a , 14bB-0ctadecahydro-4S,4aS-methano-2a,2S,6aS,8aS,9B,12bB,14aa-heptamethyl-10B-(tetrahydropyran-2 1 -yl oxy )-3 ( 2H )-pi cenone (63) . -53- To a stirred solution of 13 mg (0.026 mmoles) of the ketone 62 and 56 mg (0.49 nmoles) of potassium tert-butoxide in 3 ml of dry THF under argon was added 0.03 ml (0 . 52 rrunoles) of dry iodomethane, resulting in the irnnediate formation of a white precipitate. After 19 hours an additional 0.01 ml (0 .16 mmoles) of dry iodomethane was added. After one hour the reaction mixture was poured into 50 ml of water, and the product isolated by ether extraction, 36 affording 13 .3 mg (97%) of the crude ketone 63 as a slightly yellow solid. Silica chromatography of this material (25% ether-petroleum ether) gave 12.4 mg (91%) of 63 as a slightly yellow solid which was homogeneous by TLC (same solvent, Rf 0.28) of mp 239-243° (vacuum). One recrystallization of a portion of this material from methanol-dichloromethane gave the analytical sample as small colorless prisms: mp 254.5-256° (vacuum); ir (CHC1 3) 1675 cm-l (C==O); NMR (CDC1 ) 6 0.88, 0.96, l. 11, 1. 15 (4s,3 each,angular CH 3), 3 1.06 (s,6,angular CH 3), 3.47, 3.66 (2s,l/2 each, C-3a H), 3.51 (s,l, C-6 1 H) , 3.95 (m,l,C-6 1 H), 4.57, and 4. 70 (2s,l/2 each,C-2 1 H). Anal . Calcd for c35H56 o3: C, 80. 10; H, 10.76. Found: C, 80. 15; H, 10.79. 0 THPO 63 Mixture of Alcohols 64. To a stirred solution of 16.8 mg (0.032 rrmoles) of the ketone 63 in 6.6 ml of dry THF and 13 ml of dry alllTlOnia under argon was added 1.4 nm (0 .062 mmoles) of lithium wire. The deep blue solution which resulted - 54was stirred for 20 min before 0.061 ml (0.66 mmoles) of dry tert-butyl alcohol and another 1.4 mm portion of lith ium wire were added. After 40 min the excess lithium was quenched by addition of 0.5 ml of anhydrous methanol, and the aITT11onia was evaporated in a stream of argon. The residue was diluted with 50 ml of water and the product isolated by ether extraction . 36 The resulting crude mixture of alcohols 64 amounted to 15.3 mg (91%) of a slightly yellow solid which by TLC (25% ether-petroleum ether) appeared to be a mixture of epimers, Rf 0.09 , and 0. 17. The ir spectrum of this crude mixture showed no carbonyl absorption, with the expected hydroxyl absorption at 3600 cm- 1. The NMR showed many signals in the angular methyl region, and some signals in the c-3-5 region besides those due to the THP ether. is indicative of a mixture of alcohol epimers . The spectrum Material from a similar reaction was purified by silica gel chromatography (25% ether- petroleum ether) to give a 94% yield of the mixture of alcohol s, homogeneous by TLC. However, normally the crude product was used without purification. Mixture of PhoSphoradiamidates 65 . To a stirred solution of 15 . 3 mg (0.029 mmoles) of the crude mixture of alcohols 64 from above in 3.5 ml of dry DME, 0.86 ml of dry HMPA,and 0.87 ml (4.34 ffflloles) of the chlorophosphoridiamidate reagent,34 and a trace of 1,10 phenanthroline indicator under argon was -55- added a 2.32M hexane solution of n-butyllithium dropwise until the dark red color persisted for 5 min (0. 12 ml total, 0.28 mmo1es). min the color faded to a light yellow. After 15 After a further 15 min the reaction was quenched by addition of a few drops of water, and the mixture was diluted with 50 ml of ether. Isolation of the product by ether extraction 36 including a 2% aqueous sodium hydroxide wash afforded 22 mg of the mixture of TMPDA derivatives §.2. as a yellow semisolid. The infrared spectrum of this crude mixture showed no alcohol absorption at 3600 cm-l. broad spot at Rf 0.38. TLC analysis (5% methanol-ether) showed one Silica gel chromatography of the crude mix- ture from a similar run starting with chromatographed alcohols 64 afforded a 76% yield of the mixture of diastereomers as a white solid . Normally, however, the crude product was used without further purification. JS-(tetrahydropyran-2 1 -yloxy)friedelane (66). To the deep blue solution resulting from addition of 12 mm (5.8 mmoles) of lithium wire to 30 ml of dry ethylamine,then vigorous stirring for 1 hour under argon,was added a solution of the crude mixture of esters from above in 3 ml of dry DME and 0.536 ml (5.78 mmoles) of dry tert-butanol . After 2 min the solution decolorized and an addi- tional 12 mm (5.8 mmoles) of lithium wire was added. The resulting deep blue solution was stirred for an additional 30 min before the -56excess lithium was destroyed by the addition of a few ml of concentrated aqueous ammonium chloride solution. The ethylamine was evaporated under a stream of argon, and the resulting residue was diluted with 50 ml of water. The crude THP ether 66 resulting from isolation by ether extrac- tion36 was purified by silica gel chromatography (5% ether-petroleum ether) to give 10.8 mg (73%) of a colorless solid of mp 235-238° (vacuum) which was homogeneous by TLC (same solvent, Rf 0.28) . The same reaction carried out on chromatographed phosphorodiamidates 65 gave a 90% yield of THP ether 66 of mp 237.5-239 . 5° (vacuum). The overall yield from the ketone 63 without purification of the intermediates 64 and 65 is 69% and with purification is 64%. A portion of the above THP ether was recrystal- lized from methanol-dichloromethane to give analytically pure material: mp 243.5-245.5° (vacuum); ir (CHC1 ) 2920 cm-l (broad, RH); NMR (CDC1 3) 3 o 0.86, 1.16 (2s, 3 each , angular CH 3), 0.94 (s,6,angular CH 3), 0.99 (s,9,angular CH 3), 3.48, 3.68 (2s , 1/2 each, C-3 2H)$ 3.52 (s,l,C -6 1 H), 3.95 (m,l,C-6 1 H), 4.57, 4.70 (2 s, 1/2 each , C-2 1 H). Anal. Calcd for c35 H602: C, 81 .97; H, 11 .79. Found: C, 81.96; H, 11.73. Mass measured M+: Calcd fo r c35 H60o2; 512.45930 . 512 4593 ± 0. 0005 . 0 Found: -57- TIIPO (±)-Friedelin (!Q). To a stirred solution of 10 . 8 mg (0.021 rrmoles) of the tetrahydropyranyloxy friedelane 66 in 6 ml of dry THF and 3 ml of methanol under argon was added 2 mg of p-toluenesulfonic acid enonohydrate. After 20 hours at ambient temperature the reaction mixture was poured into 50 ml of 1/2 saturated sodium bicarbonate solution and the product isolated by chloroform extraction . 36 The crude alcohol 67 which resulted was oxidized directly without further purification . To a stirred solution of the crude alcohol 67 in 5 ml of dry dichloromethane under argon was added l ml of Colli ns reagent . (Th i s reagent was prepared by dropwise addition of 0. 48 ml of dry pyridine to a stirred suspension of 300 mg of anhydrous chromium trioxide in 15 ml of dry dichloromethane under argon. after 20 min.) The supernatant liquid was used After 10 min at room temperature the brown suspension was filtered through 5 gms of alumina with suction, and the filter was washed with 50 ml of ether . Removal of the solvents at reduced pressure afforded 10.6 mg of a yellow solid . Silicagel chromatography of this material (10% E/PE) gave 9 mg (100%) of a slightly yellow solid of mp 236-240° (vacuum). Recrystallization of this material from methanol- dichloromethane afforded 7.1 mg (79%) of analytically pure (±)-friedelin (!Q) as fluffy tiny white rods: mp 246.5-248° (vacuum); ir (CHC1 3) -581700 (C=O) ; NMR (CDC13) o 0. 72, 0. 89 , 0,95 , 1 . 04 , 1.17 (5s , 3 each, angula r CH 3) , 1 . 00 (s ,6,angular CH 3) , 0.90 (d , 3,J= 6Hz , C- 4 CH 3) . Anal . Calcd for c30H50o: C, 84.44; H, 11 . 81 . Found : C, 84 . 48; H, 11 . 77 . Mass measu red M+: 426. 3860 ± o. 0005. Cal cd fo r c30 H50o; 426.386147 . Found: -59 - APPENDIX -60- ( R~pri n tc d from the Journa l of Organic Chemistry, 40, 990 (1975).J Copyright 1975 by tbc American Chemical Society and reprinted by pcrmis$ion of the copy'right owner. Experiments Directed toward the Total Synthesis of Terpenes. XX. Total Synthesis of (±)-Shionone, a Tet racyclic Triterpene 1 Robert E. Irela nd ,• Conrad J . Kowalski,2 Jefferson W. Tilley, 3 and David M. Walba 4 Contribulion No. 4973 from The Chemical /,aboratories, Cf!lifornia /11.stitute of T ech,;ology, Pa.adena, California 91125 R cc~ived October 18, 1974 The co nv~ rsi on of the tetracycl ic ketone I to the lriterpene shionone (23) is explored by two alternative sequences. Bot h approaches rely on t he introdurtion of the more or less completely formed side chain and thon .modi lic1tion of t he • romatic A ri ng. One unsuc·cc"ful approach ent,.ils incorporation of the intact side chain and t hen cleavage ann recycl izat inn of the enone 18. Acid-catalyzed recyclization of the A ring results in hydration of t he side-chai n dou ble bond. This problem wae overcome and the synthesis of (:1:)-shionone achieved through postp<>nement of th e introd uction of t he eide-chn in unsaturntion until the A r\ng sequence_ wae complete. In the preced ing paper5 in this sflri es t he development of a practical a nd effi ci ent synthesis of the tetracyclic ketone 1 is described. Th is m ateri a l, as well as some of the interm ediates used in i L~ synt hesis, were e nvisag-cd as key interm ediates for synth esis of both pent.a· and tetracyclic triterpenes. In t his report t he s uccessful CQnversion of the ke tone 1 to the tetracycli c t ri terpenc shionone (23) 6 is described .7 For th is synth es is it was necessary to d evise two mutu111ly compa tibl e sche mes for the rem a ini ng opc>rat.ions, namely, the introdu ct ion of the s ide cha in in rin g D and the modification of th e arom a tic A r in g t.o that of th e natura l product. Th ~ in vesti gati on of the lat t er problem was u n dertaken first (Cha rt 1). A convenient o;yst.em--the enone :! --with whi ch to <•xplore means for t.hc A ring conversion wns obtai n rd by fi rs t transformation of the tetracycli c ketone l to t h n enol phosphorodiamicl at.e ('J'MPD A) 2R nnd t hen Birch redu ct ion to remove the TMPDA ns well as reduce the aromatic ring. This two-st age transformation afforded the enone 3 in 70% overall yield; during the course of optimizing this yield, it was observed I.hat if a proton source, such· as alcohol, was omitted from t he B irch 1'eduction step , the TMPDA gro1; ping was still reduct ively removed in high yie ld , but t he a romatic ring remH in ed intact. Of course, t he correspond ing a romatic olefi n could be subsequently red uced to t he enone 3 under standard Birch reduction cond itions, and t his two step reduction sequence primarily serves to demonstrate the funct ional selectivity possible d uring the reductive removal8 of the TMP DA grouping. The a ,8-unsaturatcd ketone syste m of the enone 3 offers an ideal subst rn te fo r t ht> regi ose lective introduction of th e two remai n in g methyl groups at C -12a a nd C-1 through conjugate addition a nd t he n a-meth ylation. The etereochcmica l sit uation i ~, however, somew hat less satisfac tory. -61- Total SynthesiR of (± )- S hi o11one Chart I .Ring A Modification of Tel rncyclic Ketone 1• . ~ -;; ,- I ~MPOA Cli,O~ CIJ,O~ 2 o~ 6 11~~ 1 [7,X~O Bfrl'.> 8,X ~ /JQH,o·CHJ mt92% .~~~ 10 0 a, Li N[CH(CHalzJ.. THF; CIPO[NWH:ih).; b, Li , NH.,, THF, t-BuOH ; 5 N HCI, F.TOH : c, LiR:iBH. THF; d, Zn-Cu , CH21 7 , Et20: e, CrOa · 2Py. CH2Cl2: f, Li, NHJ, THF; Ac20: g, KOH. EtOII; b, CHaLi, Dl\-!E; Zn .-Cu, CH 2 1,, Et 1 0; i, HCI, H 7 0, EtOH; j, H202. aq NaOH , CH,Of-1; k. p-Ts~ llNH,. llOAc, CH 2 Cl 7 ; I. CH3Li, Et20; m, CF,CO,H -·(CF 3 COJ,O, Cll 3 COCHs. Cl l,OH, aq HCI. The C -4b/j angular methyl group spverel.v sh ic, lds t.hc C-1 ?.n carbon from at tack hy a reagent frmn.t lH, desired # fuce of the molecule. Th•.ts , whil e conj ugate tt rl d it inn of n met hyl group per se [LiCu(C'H: 1h ] to 1his cnotw system would lie ex pected to lend to a cis -fused pri.du c:l, even n reagent J . Org. Chem., Vol. 40, No. 8, 1975 991 known to produ,·e trans-fused ringM systems in other molecules9 (A!Et:i ·· HCN 10 ) gave no react.ion or predominantly a low yield of ci~-fused product herc. 11 To overcome this stereochemicnl sit uR.t1on a method was sought that relied on the intramo lecular orientntion of CR.rbon-carbon bond format ion at C-l 2a, and two such schemes were invcsti· gated . Oi1e method relies on the orie ntat.ion 12 of the SimmonsSmith methy lenation reaction 13 by the alcohol function in an allylic alcohol system, and here requires t he generation of a C-2(3 (axial) hydroxyl group. The formation of t he desired allylic alcohol 4 proved itse lf to be a thorny problem, for standard hydride reductions (LiA IH 4 , NaBH 4 ) produced little, if any, of the {J (axial) alcohol. The on ly satisfactory method for reduction of th is enone system was through the u~c of lithium perhydro-9b-boraphenalylhydride recently developed by Brown and Dickason14 and utilized effectively in an earlier stage 6 in the synthesis. Unfortunately, probably owing to t he flat, unhindered character of the enone system, the yield of the desired fj (axial) alcohol 4 was not as high as the yields experienced elsewhere when saturated ketones were reduccd. 14 It was possible, however, after a rnther tedioua and inefficient chromRtographic sequence, to realize a fair yield of the desired allylic alcohol 4 and pursue the sequence further, as shown in Chart I. These remaining stages resulted in quite satisfactory yields of the respective intermediate products. A useful consequence, of course, of the methylenation process for the formation of the C-12a bond in the {3 (axial) orientation is that lithium- ammonia reduction 15 of the cyclopropyl ketone 6 generates the enolate anion necessary for the introduction of I.he C-1 methyl group by methylation. In spite of the fact that this methylation wou ld be expected to take place through the ~nhindered , a (axial) approach to the tetracyclic enolate, direct methylation 16 of the enolate generated during reduction of the cyclopropyl ketone 6 or methylation 17 of the enolate regenerated in dimethoxyethane from the intermediate enol acetate 9 were singularly unsuccessful. The desired monomethylated ketone in low yield was always accompanied by unmethyloted material in much higher yield. While the reasons for t his behavior are unclear, a convenient solution to the prohlem was found in the Simmons-Sm i1h methylenation 1~of this same enolatea procedure suggested by the work of Whitlock and Over· man. 18 While it wao possible to achieve the desired end result by remova l of the ammonia and then addition of the Simmons-Smith ·reagent directly to the enolate formed from reduction of the cyclopropyl btone 6, a cleaner prod uct was obtained in more reproducible yields if this enolate was. regenerated in dimethoxyethanc with methyllithium from the initially trapped enol acetate 9. Contrary to the results reported by Whitlock and Overman, 18 there is no question but th at the expected cyclopropyl alcohol is the primary product of this process By rapid and careful chromatography of the <.·ude product, it is possible to rP•nove nil the iodide-iodine formed and isolate the cyclopropyl alcohol 11 in good yield. Th is materia l is quite labi le to traces of iodide ion in liydroxylic solvents and is rapid ly cleaved to the corresponding methylated ketone. This !ability and the fail ure Lo re move these by-products probably accou nts tor the fact that Wh itlock and Overman 18 d id not observe the format ion of u cyclopropyl alcohol in their investigations. For preparative purposes n more convenient means of cyclopropyl nlcohol cleavage is the use of mineral acid , which noL only provides for the cleavage but also isomer izes the in it ia ll y <Y (axi11 l) methy l gro u p. T he overall yield of the keto ne 12 fr om the enone 3 by this sequence is only 10.3%, end the route suffers primarily 992 J. Org. Chem., Vol. 40, No. 8, 1975 from the only fair yield of the ally lic alcohol 4 and t.hc tedious procedure necessary lo achieve even that result. As n consequence of this experience another sequence was investigated in which the stereochemical outcome of the formation of a carbon--carbon bond at C-12a is controlle>d in a desirable fashion by the C-4b,~ methyl group. For thi s re sult to pertain it is necessary to plan for the fornrntion of the C-l-C-l 2a ring bond which is a (equatoria l) t.o the 13 r ing. Such A plan implies the prior introduction of the potential C-12a methyl group, as well as the cleavage and reformation of the C-1-C-12a bond which already exists in the enone 3. A sequence which involvecl just such a process is outlined in Chart I, and in spite of what at first sight seems inefficiency owing to the 1wccssity of ring cleavage, this route is significantly more efficient than that just. described. Utilization of the sequence deve loped by E~chenmoser and coworkers1D provided an excellent means for cleavage of the A. ring of the enone 3 without the loss of any cRrbon atoms. Owing to the diversity of the functionality t-hat re sults from the Eschenmoser cleavage, it was now possib le to incorporate the potential C-12a angu lar methyl group through the direct addition of mcthyllit.hium to the acetylenic ketone 7 without the necessity of incorporating hlocking groupa in the sequence. With the Rcetylenic a lcohol 8 in hand the stage was set for the re-formation of the C-1-C12a ring bond through cyclization. The pioneering work of Peterson20 and the extensive work of .Johnson and Lansbury and .their coworkers 21 provided the basis for the se lection of the reaction cond iti ons. Co nfidence that the stcreochemistry of the molecule thBt would result from this cycli:zation would be that with the C-1'?.a methyl group in the desired fJ (axial) orientation stem med from the extensive previous experience2 2 in these lahorntories that demonstrated the stereochemical control pro\'ided by the axia l C-4b{J methyl group during similar cationic ring closures. It was nevertheless gratifyi ng to find that cyclizat.ion of the acetylenic alcohol 8 in trifluoroacetic acid led to an enol trifluoroacetate in excellent yield and that hydrolysis of this intermediate provided the same sRturated ketone I 0 that was obtained on saponification of the corresponding enol acetate 9 from t.he previously described route . The convergence of these twc routes at this puint serves to confirm the fJ (axial) orientution of the C-l 2a methyl group, fo r the /i (quasi-axial) assignment of the configuration of th t> C-2 hydroxyl group in the allylic alcohol 4-and hence t.he orientation of. the s ;.nmons-Smith met hyl enat ion reartion --rests on firm ground. In view of th e ease with which the enol trifluoroacetate could be isolated from this C)'Cli>.ation and the already proven utility of th e enol acetate 9 for the incorporation. of the remaining m<>thyl group al C-1, the ·present route seemed well suited to the construc tion of the shionone A ring, and attention wa' I.urned to the 111trod11ction of the side chain in ring D. Since the general plan for the total synthesis of sh ionrme (23) entailed the incorporation of the ring D side chain a" d then modification of the aromatic A ring by the proces~ discussed above, the tetracyclic kelonr' I ngain l..wcame the starting point. While the ketone func tionality in the ll ring of this material would ohviously serve to introduce the two required alkyl groups in the a<ljaccnt rr position, the efficiency and st.ereochemical out.<:onw of these alkylat ion reactions were circumspect. In 1tddit ion such a plan i11 c<>rporates the potential difficulties that would he associated with the ultimate necessary removal of what. would then he a very hindered ketone functiu11. In ntrler to circumvent these anticipHted chem ical pru!.kms, as well as hm·e A sound basis for the st.ereoclwmical results, rn eans were -62- Ireland, Kowalski, Tilley, and Walba sought to remove the existing ring D ketone and at the same time introduce activating functionality external to the ring system. This plan was effectively accomplished (Chart II) by a two-step process I.hat lc<l from the tetracyclic ketone I through the chloro aldehyde 13 from the Vilsmeier rear.tion23 and then by lithium-111nmonia reductionmeth¥lation16 to the aldehyde 14. The yields in this process were quite satisfactory, and the efficiency of the structural changes t hRt attend the red uction- methylation st.ep is noteworthy. The stereochemical outcome of the methylntion of the enolate from the lithium-ammonia reduction of the ch loro aldehyde 13 is well preceden ted 24 in simi lar systems, l1ut the preparative use of a,/j-unsaturated aldehydes in such recluctions to generate useful aldehyde enolates appears:!.~ to be novel. As· might be expected, the conditions for the reduction stage had to be carefully controlled (see Experimental Section) in order to prevent dimerization and overr~duction. With an eye toward the rapid completion of the side · chain, the aldehyd e 14 was quantitatively reduced to the corresponding prim1uy alcohol with lithium aluminum hydride and efforts were made to convert this alcohol to the iodide preparative to a coupling reaction 26 with 11'-(l,l-dimethylallyl)-nickel bromide complex. Unfortunately, both th e neopentyl character and the severe steric congestion about this axial hydroxymet.hyl grouping thwarted all attempts io prepare the iodide or other halides. Reactions of the alcohol with triphenyl phosphite-methyl iodide, 27 triphenylphosphine, carbon tetrabromide, and carbon tetrechloride,28 and thionyl chloride--quinoline led either to recovered alcohol after no reaction or a plethora of products that resulted from the intervention of cationic species that led to backbone rearrangements. ·To overcome these difficulties an alternate scheme was developed for the addition of the remainder of the side chain through the use of two successive Wittig-type condensations. The aldehyde 14 was first converted to the unsaturated aldehyde 15 by the efficient formylolefi nation procedure of Nagata and Hayase,29 and after reduction 30 of the unsaturated aldehyde 15 with triethylsilane in the presence of tris(triphenylphosphine)rhod iu m chloride, the process was completed in :!0% overall yield from the tetracyclic ketone I by the condensation of the saturated aldehyde with isopropylidenephosphorane. With the aromatic olefin 16 thus in hand, modification of the A ring hy the metl1od developed above was projected to complete the synthesis. Indeed this process proceeded well (Chart ! I) up to the stage of final reformation of the A ring from the acetylenic alcohol 20. This approach irreversibly broke clown at .this p:iint, for the acidic conditions necessary to effect the cycl izat ion invariably resulted in acid-catalyzed hydration of the side-chain trisubstituted double bond. When modifications were m11de in the reaction conditions in order to avoid this add ition reaction by reducing the acid ity, lower ing the tempemture , and/or changing the acid catalyst, it was fo.u nd that the sequence of event!' involved initial rapid addition to the side-chain double bond. In experiments where the conditions were vigorous enough, cyclizat.ion of the acetylen ic alcohol was a subsequent step. Indeed, it was possible to hydrate the side-chain double bond without affecting the acetylenic alcohol system. The !ability or this side-chain unsaturation was a surprise, particularly when the· model system used to explore this sequence--the acetylenic alcohol 8-was specifically chosen with this side cha in in mincl and itself contains an isolated (alheit <lisuhstitut.ed) double bond. The solution to this last problem dictated a change in methodology for either the A ring modifications or the -63- Total Synthesis of (±)-Shionnne .J !Jrg. Chem., Vol. 40, No. 8, 1975 993 Chart JI Conversion of Tetracyclic Ketone I to (:1:)-Shionone (23)a I '~ ~_,CHO d :; !lllf/> CH,O 14 13 _rrl~ Cli. ,O~i............- '*·~~ ~--~" 0 OH 18 2Q ·-·~ n.t ~:--~OH -- - 0 23 . • • -~OH k TI1' 6~1):. 0 19 HO Cii, 22 21 a a, POCI,. DMF; b, Li, NH, , '!'HF', 1-BuOI-l; c, Na02CCcHo. CH 3 1; d, NaH, (Rt0)2POCH2CH=NC 0 H 11 , THF, aq (C02 H J., CoHo; e, EtaSiH, l(CoHobP)sRhCl, C•flo. CH,COCHa. aq HCI; f, (CoHohPCH(CH 3 b• T-, C6 H0 Li, Tiff; g, (C0 2 H) 2 , aq EtOH, NoOH, aq EtOH; h, H20i. aq NaOH, CHaOH -CJ-1,Cl,; i, p-TsNHNI!,. HOAc - CH2Cl,; J. CHaLi, THF ; k, aq HCl, EtOH ; l, CF3 C0 2 H-(CF3CO)o0; m, LiN(CH(CH,J,J, THF, Zn - AR, CH,!,, Et 2 0; n, Cr0 3 • 2l'y, CH 2 Cl 2 . side-chain construction. RHther than tamper with the more intricate procedures in the former process, a reshuffling of the ~teps in the side-chain construction seemed ad visable. Since the offending functionality in the side chain was the double hond that resulted from the last stage in the process, this reaction was deferred until the <.:omplPtion of the A ring. Thrn;, complete reduction of the unsaturated aldehyde 15 led in good yield to the hydroxyenone 17, which could be carried through the A ring syn thesis with out major incident (Chart II). Th e result of the trifluoroacetic acid catalyzed cyclization of I he acetylenic: alcohol 19 wa s the cxpect~d hi s trifluoroncelal c, hut this posed no significant experimental problem in the subsequent stages that completed theshionone (2:l) synthesis. A noteworthy point did come to light when the bis trifl uoroacetate was used to generate the (onolate in ring A . Under the conditions used c.' u li,, r for the general ion and methylenation of the enolate from the enol acetat" 9, none of the expect.ed cyclnpropyl al..ohol was ohserved, 11nd the product was the C-1 dem1·thyl keto alcohol. Model studies showed that this was not the result of hydrolysi s or protonation of the enolate by traces of moistur~. nor was it the re sult of the failure of th<c nidhylenation re11ction. Heasoning that the enolate g1rnerated by methyllithiurn add i\i1,nl.7 was being rapidly protonated by the Initially formed 1,1,1-trifluoroaceton e, an aminolysis reaction was substituted for the Grignnrd reaction with salutory results. The enol trifluoro11cetate was readily cleaved by lithium diisopropylamide, and the resulting enolate behaved as expected in the methylenation ·reaction. T his observation should render enol trifluoroacetates generally useful for the formation of ketone enolates, and coupled with the addition of trifluoroacetic acid to acetylenes, the overall process is an interesting ketone synthesis. Experimental Sectioq3• 8-M ct hoxy-4n/1, I Ob/1, I 2aa- trimethyl-3,4,4a,4ba,6,G,l Ob,· 11, I 2, 12a-dccahydrochryscn-1-yl Tctramet.hylphosphorodl• 11midatc (2). To a solution of lithium diisopropylamide prepared from R ml (57 mmol) of diiwpropylamine and 13 ml of a 2.84 M hexane solut ion of n- butyllithium in 150 ml of dry ether under an nrgon ntmo~phere was added over a 10-rnin period a solution of ?.'!? g (7 mmoll of the kf"lone I in 20 ml of dry tetrahydrofuran a nd 8 ml of N,N,N',N'·tetramethylethylenediamine. The mixture wa• then cooled in an ice hath, and 15 ml (81 mmol) of tetramethyl<linmiclophosphorochloridate" was added dropwise. After the resulting yellow ,olution wus allowed to warm to room temperature and then stirred for 1.5 hr. the mixture was poured into ice and 400 ml of 10% aqueous hydrochloric acid, and the product was isolated hy ether extraction:i.1 including a base wash. On chromatography of 994 J. Org. Chem., Vol. 10, No. 8, l!J7.5 th<' crude product on 200 g ·o r si licn gel, ~-~1 g (bil%) of thr phos· phorodiamidatc 2, mp 108.-111° (v11c1111m l, wos r lut e d with 1800 ml of 10% acetone- Nhvl arclntt nft"r an in it inl wa'h with 800 ml of ethvl acetate and th~n GOU ml of 5% 11vetr111" ·etlw l ncetnt.,. 'l'hc analytical sample, obtni1wd a ft er eystnllirntion c>f n.portion of thi < mnter inl from ether- hrpt1111e, nlso melt1·d 111 11m- 11J 0 (vncuum l: ir (CHCbl 1670 (C=Cl. lliOf1 . 1500 (Ar), 1:in;, (l'.N), and 980 <·n1- 1 (P-0- C); NMR (CDCl:i) I> 1.02 is, 3, C -•111 CH:il. 1.2~ (<. 2 X a. c. lOh and C-12a CH,), 2.70 (<l, 12, J = 111 Hz. NCH,), 3.7'> (<, :1, OCHal, and 5.20 (m, 1, C="CHl. Anni. Calcd for C2r;H 41 0,,N,P: C, 67.80; II , 8.97; N, fi.08; P, 6.?:l. Found: C, 67,9fi; H, 8.8r; N, 6.16; P, 6.64. 8-Mclhoxy-4a/J, I Ob/l, I 2lia-t ri methyl-~ ,.f ,4 a ,4 ha,5,fi, I 0 b,l I, 12, 12a-decahyd roc h ryscn e . To n snl11tion of ;17 mg (5.~ mg· ntoms) of lithium in ;,u ml of dry ammonia nnd JO ml of dry letrahydrofuran under nn nrgon a tmosph~re wns tldciPci a solution of 210 mg (0.45 mmol) of the plrnsphoro<linrnirla te 2 in 6 ml of dry tetrnhydrofuran. After 1.5 hr the blue color fnded, ond an Hdditionul 37 mg (5.3 mg-atom<) of lit hi um wns added: A fl Pr •tirring for 3.fi hr longer, the reaction mixwre was trent<•d with .IQ() mg of sodium ben1.0ate and then 200 mg of solid ammonium chloride . After the ammonia was evaporated in a stream of argon, the residue was dissolved in 50 ml of wat e r, and the-product was isolated by ether extraction33 including an acid an<l base wash. On preparati\'e TLC (30% ether- petroleum ether) ,,f the crude prorl uct thNe was obtained 115 mg (82%) of the tetracyclic olefin (Rr 0.7) as a colorl<:ss oil. The analytical sample was obtained afLer further preparative TLC (30% ether--petrolcum ether) and then evaporative distill ation (120°, 0.01 mm) of a portion of this materia l: ir (CHCl3) 1605 and 1500 cm- 1 (Ar); NMH (CDCl3) 6 0.82 (s, :1. C-4a CH3l, 1.00 (s, 3, C-12a CH 3 ), 1.22 (s, 3, C· !Oli CH3}. 3.77 (s, 3, ArOCHa), 5 ..52 (m, 2, CH=CHl, and 6.50-7. 17 (m, 3, ArH). Anal. Calcd for C22H3.10: C, 85.11; H, 9.74. F'ound: C, 84.9G; H, 9.64. 4b,8,6aa,10a/j,Trimet hyl - 4,4Ra,4b,5,6,6n,9,l O, l Oa,l Oba, I I, 12dodccahydro-2(~11) - chryscnone (3). A. From tl1 e Phosphorodiamidate 2. A solution of 310 mg (53 mmol) of lithium wire in :,50 ml of dry ammonia and 140 ml of dry tetrahydrufuran was stirred for 30 min, and then A <olution of 1.53 g (3.:l 2 mmol) of thr phosphorodiamidate 2 in 30 ml of dry letrahydrofurnn wa< injected nil at once with a syringe. Aft.er 5 hr an additionnl 9tl0 mg (139 mmoll of lithium and 85 ml of dry lert-butyl a lcohol were udded . After the reaction had stirred for an additionnl 2 hr, the exce'5 lithi •.Hn wns decomposed with 20 ml of methanol, nn<l tlw ammonio wns nllowed to evapornte overnight. The gray re<ido c wns treated with 500 ml of water, and the product we< isolated hy eth er ex lrnction.3.1 A solution of the resulting resichw in 200 ml of eth ;rnol and 130 ml of 5 N aqueou< hydrochloric acid"'"' he"ted at 65--70° for 40 min ln an argon atmosphrrP. The coulf'cl n.·nction mixture wns t he n poured into 500 ml of water . and the prodoct wo< isolat<•d hy ether extraction,83 induding a base wash . On chromal.ogrnphy of the dnrk yellow, oily re<idue on JOO g of sil ica gel there wns obtained 782 mg (79%) of the enone 3, mp 88 -!12 ° , hy e luti on with 600 ml of ·50% ethl"- petroleum ethe r. Crys tullirntion (ethan<ol water) and then sub:.mation {120° , 0.01 111111) of n portion or this material ga\•e the ana lytical <ample: mp !J.1.. ~7°; ir (CHCl:il JGfF, (C=O) and 1620 cm- 1 (C'=C) ; NMH !CDCl:1l. ii 0. 87 (s, 2 x :l. c. 4b and C-lOa CH3), 1.07 (s, 3, C -6a C l!,), 5.47 (m. 2. Cll = <'H). and 5.92 (s, 1, O=C-CH=Cl. Anal. Calcd for C21H300: C, 84.M; H , 10.J:l. Fo11nd: C, 84.f>l; H . 10.22. B. From Tetracyclic Olefin. A stirred <nlu tir111 of 182 mi: (0/18 mmol) of the above olefin in 60 ml of dr y amnHrnin, 20 ml of drv lctrahy<lrofuran, and 10 ml of <lry terl· butyl nlrnh ,. I under an "~''°"n atmosphere was trent.cd with Ill mg (l fi mg etoms ) of lit hi11 m. After 2 hr the excess lithium was d£:(;ompos Pd with :i ml uf rn L- 1 ha nol, and the ammonia was l'Vapornted in :i s trc:nrn of argon. '1 'lw gray residue was dissolved in 150 ml of wntc·r , Rnd the prorluct wa~ isolnted by ether extrnr.tion. 3:' A ~olutiun of thr crude pr<Hlw: t in 30 ml of ethanol and 20 ml of;, N aquom" h;'drnchloric ncid wr" heated at 65- 70° for 40 111in undc>r an arg(lll rttmo.., ph£>re, ;111d nft1·r dilution of the cuolecl renc-lion mixture wilh JOO ml of wult.:r , t liP. product was isolated hy r1 her exlrnC'tion .: 1 1 inr.luciing a ha sc· w ;t': h. The crude product was chromatograplwd on 2'2 g: of sil icu g{·I. nnd elution with 175 ml of f10% ether- petrolt!u111 <:t her nffordN.J ·1;,:1 m~~ (78%) of the ~none 3, mp 91-9·1°, tha t v.:as idl·ntiral (rnixtur<.= 11l•'11 ing point, ir, NMR) with the mat e rial pre piirl'd " ho ve in p:trl A 2,8-Hydroxy-•I b/j,6na, I Oa/l-t rime th y l-2a .1,4 ,.I aa, ~ b,S ,t;,(i:i, 'I, . 10,IOa,toba,ll,12-tt-trndccahydroc hry' " " e U). Following tlw -64- Ireland, Kowalski, Tilley, and Wa lba ge neral procedure of Brown and Dickason, 14 nn ice-cold solution of 781 mg (2.6 2 mmol) of the enone 3 in If> ml of dry tctrahydrofuran und"r an argon ut.mo<phcre was treated with 6.0 ml (f>. I mmol) nf 11 O.Rfi M tetra hydrofu run solution of the trialkylh"rnhydridc. After 30 min the organoboronc wus decomp<» r <I by the seq ue ntia l addi· tion of J.0 ml "f :IN nq11rous sodium hvdroxidr <olution a nd 2.0 ml of 3ry*i hydrot{C'n prroxidc . The react ion mixt.ur P wa~ immedint-rly p1mred into 50 ml of snturated aqueou fi ~odium carbonate solution nnd tha'pro<luct isolated hy ether-ben1enc (4:1) extraction. The crude prod11ct was chromntographed 011 100 g of Flnris il which was eluted with ri9''· cth.,r--petroleum ether. The fir • l. 300 ml elute<l 75 mr. of a mixture of nonpolnr products thnt wa• di•cnrded. The n<·xt 200 ml affot·dc•<l 183 mg (2:l'!b) of the axial nlcohol 4, mp 130- 1~3°. r urther elution with 400 ml of the same solvent guve 316 mg of o m ixture of the· two alcr.1huls which on furth er separation by prepurnlive TLC (10% ethe r-chloroform) gave 188 mg (24%) of the equtt· toriol alcohol Ult 0.4) a nd 114 m g 0.5%) of the axia l alcohol 4 (fir 0.5). Finally, washing the rolumn with 500 ml of ether gave 206 mg (26%) of the eq un tor iol a lcohol, mp 119-120° (vacuum). The tot.al yie ld of nxial alcohol 4 was 296 mg (38%) and that of the equatorial a lcohol was 394 mg (fi0%). The analyticnl <ample of the axial alcohol 4, prepared by crystallization of a portion of similar material from a nother reduction ex periment from ethyl aceta te-heptane and then ethanol-water, melted at 13 1-13·1 ° (vacuum) : ir 3600, 3450 (OH), nnd 1650 cm- 1 (C=C); NMR (CDCl3) 6 0.85 (•, 2 X 3, C-4b an d C· lOa CH3), and 1.05 (s, 3, C-6a CH 3), 4.00-4 .20 (m, I, CHO), 5.48 (m, 2, CH= CH), and 5.50-5.77 (m, I, OCCH = C). Anal. Calcd for C2 1H 320: C, 83.94; H, 10.73. Foun d: C, 83.87; H, 10.78. · The analytical sample of the equatorial alcohol was prepared in the same fashiGn and melted at 114-116° (vncuum): ir (CHCla) 360fi, 3450 (OH), and 1655 cm- 1 (C=C); NMR (CDCl3) 6 0.77 , 0.82. and 1.05 (s, 3 each, C-4b, C-6a, and C -!Oa CHal. 3.95- 4.30 (m, 1, CHO), and 5.37-5.53 (m, 3, CH=C). Anal. Cnkd for C21H320: C, 83.94; H, 10.?:l. Found: C , 83 .93; H, 10.87. l,8,12u/j-Methano-4b/1,6aa, l Oa /j-trimcthyl-1 a,4,4oa,4b,5,6,fin,Y, I O,IOa,I Oba,1 1, 12, 12a-tet r adecah ydro-2(311)-chrysenone (6). To a sus prn<ion of 4.0 g (57 rnmol) of 1i11r- copper couple 3' in 4.6 ml (57 mmol) of diiodomethane and 60 ml of dry ether was ad <lcd a solution of 618 m~ (2 .21 mmol) of the ax ial alcohol 4 in 10 ml of dry.et.her, and the resulting mixture was heated at re nnx u nder an argon Atmosphe re for 4 hr. After C<X>ling, the rea ction mixture wns poured into 100 ml of satu rntcd aqueous sodium cn rbonate, and the product was isolated by ether-benzene (4 : I) ex· tract ion .S 3 On chr.>m ot<ography of the product on 250 g of grade I ll a lumina, elu t ion with 600 ml of 3% me t hanol-e t he r gave 512 mg (80%) of the corre'l)(lllding cyclopropyl nlcohol, mp 135- 139° (vnc 11um): ir (CHCl,i :!600, ~450 cm- 1 (OH); NMH (CDC'3) 6 0.85, 0 95 . 1.05 (3 s, a <·ach, C-4h, C-Ga, and C-IOn CH 3), 4.07-4.43 (m, 1, CHO), and 5.47 (s , 2- CH '=C H-) Afte r the procc•dure of :iadcliffe and Ho<lchorst, 36 a solution of 512 m g ( 1.69 rnmol) of the above cyclo propyl a lcohol in 8 ml of d ry dichloromethane was ad ded under an argon atmosphe re to a solution of 1.62 m l (20 mmol) of dry pyridine nnd 1.00 g ( 10 mmol) of a nh yd rous chromium triox id e in 50 ml of dry d ichloromethone, lmd t he reel sol11tiPp wns stirred for 10 min. Th e dnrl< mixture was then filtered through n pad of grade III aluminu wit h the a id of 200 ml of ether. Evaporation of the solvents from the filtrate at re · duc1,d prc.sme afforded 490 mg (77%, 61 % overall) of the ketone 6, mp 149- lf>2° (vacuum) . The analyticnl <ample, prepare<l from a portion nf thi< 11rnter ial h;· preparative TLC l50% ether-petroleum ethe r) a nd then nys tellizati on from et.l: ., r - h exa ne, melted at 1501;,:10 (vac uum): ir !C HCl:i' J670 cm- 1 (C= OJ; NMR (CDCl3) rl O.~O. 1.02, l 07 (;l s, 3 each, L-4h, C-6a, and C-IOa CH 3), and 5.47 1 ,, 2,CH=C Hl. Anni. Cnlcd fur C 2 2H:i20: C, 84 .56; H , 10.~2. Found: C, 84.49; H, IO.:J9. 2-Acetox)•-4 b/J,6aa, I O ~ /J. I 2a/j-letramethyl-3,4,4na,4 b,5,6,6a,~. IO, I On, I llha , l I, 12, 12a-tr: radecahydrochrysen e (9). To an urgu11 -proJ.<,cl.r-d <oluti«n uf .18 mg (2.6 mg-atom s ) of lithium in 60 ml of ciry Amrno ni n and 2 lt ml of dry tetrnhydrofura n was added a solution (Jf 1o:i mK (OJ;.s mmol) of the cyclopropyl ketone 6 in 5 ml of dry tetruliydrPfurnn. After stirring for I.~ hr, the blue colnr fudod; nn 1r<lditio1rnl IH m~ (2.6 mg-atoms) of lithium was then nddNI, and th<· mixturr was stirred for 2.h hr. Mo~t of the ammo· nin wns th e n rC'rno\'l'd by evnporation inn st re nm of nrgon throuRh n mercury huhhlc:r, and thC' re~ulting grny su,..pc nsion was treated Total Synthesis of (:1:)-Shionon~ -65- with 5 ml (53 mmol) of dry acfltic nnhydrido ot room tc•111p<•rulure. After s\irring for 6 hr, the re1wt.ion 111i>.:llll'tl wos poun•d into n mix· tul'lt ot ice und 70 ml of 1()% aqucm" µ.1tassi u m hyclwxide H<>i111 ic•n, and tho product was isolotod hy etht:r ··'1e111.eiw ( l:l) ''" rnrt ion."" On chromatu~rnphy of the n11du pnodnct (2115 m~) 011 :lo 1~ of si liea gel, elution wi th 150 ml of 20% •I her -·petroleum eth cH ):O\'l' 1,16 m~ (64%) of the ~ n ol acetote ~.mp 119- 121° (vacuum). '!'hr H1111l~·tirnl sample, prepared from R portion of this material hy pn•pnrntivc TLC (20% ether- petroleum ether) Hild cryHtollizH1.io11 from ether-· hexane, melted at 121-J2:l 0 (vncu11111): ir (CHChl 1·;;.c. tC=·O) , 1690 (C=C), and 1220 cm- 1 (C- 0 - ('J; NMR (CCI,),\ i\'i8, 0.88 (2 a, 3 each, C-4h, and C-lOb CH,) , l.O:l (s, 2 X 3, C-li11 '"'ci C-12u CHa), 2.00 (s, 3, CH 3CO), 4.U7 (s, l, C.J HJ, and &..10 (s, 2, CH~CH). Anal. Calcd for C2,Hoo02: C, BO.Bi'>; H, 10.18. Found: C, 81.05; H, 10.23. Further elution of the column with 25 ml of the same sol"ont mixture gave 37 mg of a mixture that consisted of approximate ly equal parts of the enol acetate 9 and the A 2 -enol ac~t.nle of the starting ketone G (2-acetoxy- lfJ, l 2o{J-mcthano-4 hfJ,6aa,l OafJ· &rll!lethyl- l ,4,4a1>,4b,5,6,6a,9,J 0, 1Ou, I Oha, 11, I 2,l 2a-te tra docahydrochrysenc) on the basis of the comparative integration of the acetyl methyl signala at~ 2.00 and 2.03 in the NMR spectrum. An analytically pure sample of the latter .l 2 -cnol acetate was obtained Crom another similar experime nt. after preparative TLC (20% ether-petroleum ether) and then cryR!allizntion of the mkterinl with R1 0.4 from hexane and melted at 110-112° ('·acuum): ir (CC4l 1755 (C-0), 1685 (C=C ), and 1220 cm- 1 (C-0-C); NMR (CC4l 6 0.85, 0.98, and 1.03 (3 s, 3 each, C-4b, C-6a, and C-JOa CHal. 2.03 (s, 3, CH3CO), 4.80-5.05 (m, 1, C-3 H), and S.40 (s, 2, CH-CH). Anal. Calcd for CuH:w02: C, 81.31; H, 9.67. Found: C, 81.46; H, 9.76. Saponification of this A2 -enol acetate in aqueous, alcoholic po· ~um hydroxide solution afforded a 70% yield of the cydopropyl he~e6, mp 123-126°, alone or in admixture with authentic material of the same melting range. . Finally, continued elution of the column with JOO ml of t.he same aolvent mixture a fford ed JO mg (5%) of the ketone 111, mp 1501580. Further purification of thi• material by preµarative TLC (40% ether-pet roleum ether) and then crystallization from hexanedichloromethane gave material that melted at 172-17G 0 , alone or in admixture with authentic ketone 10, mp 172-176°, prepared below by hydrolysis of the enol ncetnte 9. 4b,6,6aa,I OafJ,12a,6-Tetramcthyl-3,4,4aa,4 b,5,6,Ga,9, IO, I Oa,. IOba,l l,12,l 2a-tetradecahydro-2{ I /l)-"hrysenon" ( l 0). A. From Eno! Acetate 9. A solution of 90 mg (0.25 mmul) of the enol acetate 9 and ISO mg (2.7 mmol) of potAssium hydroxid e in 5 ml of ethanol was stirred at room tempernture for 14 hr under Argon atmosphere. The mi.xt.ure was then diluted with water, nnd the prod uct was isolated by ether-benzene (l:J) extraction. 3 ' Purification of the crude product by preparative TLC (40% ether--petroleum), then crystalli?.ation from hex•nc - clichloromethane, and fi nally aublimation ol 160-,170° and 0.025 mm gave 46 mg (61%) of analytically .pure .ketone "0: mp I 72-176° (vacuum); ir (CHCbl 1705 cm-• (C=O); NMR (CDCla) /, 0.82 (s, 3). 0.90 (s, 6), ond J .06 (s, 3) (C-4b, C-Ga, C-!Oa, and C-12a CH,), and 5.48 (s, 2, CH ~" CH). Anal. Calcd for C22H340: C, 84.0~; H, 10.90. Found: C, 83.90; H, 10.99. 8 . From Acetylenlc Alcohol R. A mixture of 2.0 ml of trifluoroacetic acid and 0.6 ml nf trinu .. roacetir anhydride wus cooled in il -1.8° both; 1.6 ml of this co ld solution was the n added lo 16.0 mg (0.051 mmol) of the acetylenie alcohol Kat -18°, a nd the mixture was stirred at this temperature for 20 min. The pressure in the system was then r~duced with u vacuum pump, and the cooling bath was then removed to facilitate evnpornt ion of the solvent s . Mu»t of the liquid was gone within 5 min, hut the residual oil was dried at room temperat ure for 20 min at en. O.OS mm pressurn. This oi l was handled so as to avoid prolonged cont.act with moist 11ir and AP· pear~d to be the desired enol tril'lunroacetate: ir (CHCl:i) 1790 (CF3 CO), 169.'> (-OC=C-l. 1.385 <CH:il. and J 220, I 170, l 140 cm- 1 (C-0-C and CF3); no remRiniug 331?0 cm- 1 (-C""'CH) ; NMR (CDCl;i) 6 0.81, 0.90, 1.04 (3 s, 3, :i, '"";Ii, r<»pectiv.,1_1', C-4n, C-6u, C-lOa, and C-12n CH 3), 5.28 (m, l. CF:1C01C=,CHl, and ~.44 (s, 2, HC=CH); analysis by GLC (2:i0° ) ,110wc·d only one J>CHk at. retention time J.2 min. A solution of this crude enol trifluoroocetate in 1 ml of ac•lune and 1 ml of methanol was trent.ecl "ith 5 drop• of water ond 5 drops of 10% nqueous hydrochloric acid and th•n stirred 11t room J. Or11. Ch;•m., Vol. 40, No. 8, 1975 1195 tr111prrnt uro fur 75 min. After neut.rn lizatlon of t hi• •nlut.lun with wlid sodium bicurb.unutu, lhc prucluct WHB iR11ln1ud hy l•thor ex· ll'1H:tio11."~ Purificnlion of th1i' crudo product (17.f> m~) hy prvpnro · tiv~ TLC (40% oth~r-putro l e um ether) nfforded 14.'I m1: (f)2%) of lho tctrn .. yclic ketone 10 us 11 whitu •olid, mp l(i() .. Jn° (vnntum); the ir and NMR sprctrn oft his matetinl wore ic!ontirul with those of purifi<!d ketone 10 prnpnn•d in pnrt A above. Cry•tAllizotion of this solid from dichloronwt hanc·- hnxune afforcfod white crysta ls, nip 170··17°t 0 (vacuum), ul111H.~ or in adm ixture with muteriel pre· porl'd nhove, mp l 72 .. l 7G 0 (n1cuu m), in pnrt A. 2fJ-l 1yd roxy- l a,2a-nwtha 110-4 bfJ,Gaa, I Oa/j, I 2a/j-letr11mcthyll /J,2,3,4 ,4110,4 b,5,6,fin,D, l 0, I Uu, IOha, 11, I 2, l2a hax11docahydro<'hry9onc ( 11 ). A solution of H>:l mg (0.43 mmol) of tho enol 11cotat e. 9 in :; ml of dry di mot hoxyothan• wati oddccl to nn mxon-protcctod Rolut ion of met.hyllithium (0.7 ml, 1.2 mmol), anci the mi xture wos stirred al room lt.· mpcro ture for 30 min. To thi• solution Wf19 added hy syringe tlrn supernatant •olution from the prepara tion of th• Simmons- Smith reagent from J.20 g (17 mnwl) ofzinccopper couple3 ' and 1.40 ml (l.20 mnwl) of diiodomcthane in 17 ml of dry ether. After stirr in g in an ice bath for I hr, tho reaction mixture was poured into f>O ml of saturated aqueous sodium carbonate solution, and the product was isolated by ether- benzene . (1:1) extrnction 33 including ii bnse nnd 10% aqueous sodium thiosulfate solution wash. On chromatography of the crude material on 70 g of grnde III alumina, •J ut ion with 200 ml of ~ther gave 100 mg (71%) of the cyclopropyl alcohol IL, mp 161-165° (vacuum). The analyticnl sample, obtained after cry>tallization of a portion of this material from dichlorcmethAne-hex81le, melted at 166·- 168° (vacuum): ir (CHCls) 3600, 3450 (OH), and 1180 cm- 1 (C- 0 -.CJ; NMR (CIJCl3) b 0.80, 0.87 , 1.02, nnd 1.15 (4 s , 3 each, C-4b. C-Ga, C-lOa, and C-l2u CH3), and 5.42 (s, 2, CH=CH). Anal. Coled for C2sHasO: C, 84.09; H, 11.04. Found: C, 83.85; H, 11.13. . . I (J,4 bfJ,6aa, I OafJ, I 2afJ-Pen tn methyl -3,4,4a<>,4 b,5 ,6,6a,9, l 0,IOa, l Oba, l I ,12,12a- tetradecahydro-2 ( I H)-chrysenone ( 12). A solution of 161 mg (0.49 mmol) of the cyclopropyl a lcohol 11 and 1 ml of concentrated hydrochloric ncid in 12 ml of ethanol was heater! und•r reflux in an a. gon atmosphere for I hr. After cooling, the solution was diluted with 50 ml of water, and the product was isolatt•d hy ether exlraction, 33 including a base wush. The resulting material amounted to 157 mg (98%) of the ketone 12, mp 169-176° (vacuum), from which the analytical •ample, mp 178-182° (vacuum), wns pm.pared by preparative TLC (40% ether-petroleum ether) ancl then sublimntion at 150-155° and 0.7 mm: ir (CHCl3) 1705 cm- 1 (C=OJ; NMH (CDCI,) b 0.73 (s, 3), 0.83 (s, 4.5), 0.95, (s, 4.li). and 1.05 (s, :l) (C-1, C-4b, C-6a, C-JOa, and C-12a CH3) , and 5.47 (s, 2, CH=CHJ. Anal. Coled for C23H 3s0: C, 84.09; H, I 1.04. Found: C, 84.16; H, I 1.16. la, 12aa-Epoxy-4 bfJ,Gaa, I OafJ-trimcthyl-3,4,4aa,4b,5,6,6.a,9,lO,IOa,l0ba, I I ,12, I 2a-tctradccahydro-2( I H)-chryscnonc (5). To A stirred solution of 125 mg (0.42 mmol)·of the enone 3 in JO ml of methanol at room teir pernlure was added I ml (ca. 300 mg, 16 mmol) of 30% aqueous hydrogen peroxide solution and 0.5 ml of 10% aqueous sodium hydroxide solution, and the mixture was stirred at room temperature for I hr. The solution was then diluted with ether and water, iind the product was isolntcd by ether extrnction."3 Purification of the res ulting semicrystnlline solid (126 mg) by preparative TLC (30% ether- petroleum ether) afforded 109 mg (8:1%) of the epoxy ketone 5 U/1 0.45), mp 98-· 100° (vacuum). The annlytical sample, mp 102.5-103.5° (vacuum), was obteined after two crystalliwtions of this material from mcthanol-dichloromc thane: ir (CH Cl3) 1700 (C="O), 1450 (CH2), and J:l85 cm- 1 (CH"); NMH (CDCla) b 0.82, 0.87, and '..05 (3 s, 3 each, C-4b, C-6a, uncl C-10u CHal. 3.16 (s, I C-1 H), and 5.49 (s 1 2, CH,~CHJ . Anni. Calcd for C21H•o' ,: C, 80.21; H, 9.6?.. Found: C, 80.44, H, 9.6\J. l fl-(:!• -B utynyl )- 4 h{J,Haa, I Oa(J-t rimethyl-3,4,4aa,4 b,5,6,Sa,9,l O, IOn-dccahydrophcnnnl h ren-2( I H)-one (7). A slight modification of the general procedure of F..schenmoser and coworkers 19 was •mployed. To a dry mixture of 77.0 mg (0.244 mmol) of the epoxy kct.vne 5 and 48.8 mg (0.261 mmol) of p-tolucne<ulfonylhydrnlinc nt -20° was added with stirring and swirli ng 1.5 ml of -20° acetic acid-dichloruml'thone (l:I). After stirring for f, min at -20°, th1> solution was stored at -20° for 15 hr. The mixture was I.hen st irrcd at room temperature for an additionul 4 hr (during which time it turned red) and then the product was isolntecl by ether extraction, 33 incl1iding A ho'" wash. Purification of th~ crude product (81 mg) by prcparutive TLC (30% ether-petroleum ether) 0 996 J. Org. Chem., Vol. 40, No . 8, 1!175 afforded 5-0.5 mg (69%) of the ncetylcnic ketone 7 a• n yellow o il <Rt 0.48) which was suitahlc for analysis: ir (C l lCl3) :i:ioo (-C<-"Cl-1), 2120 (-e'""C-), 1700 (C=O). nnd 13DO crn- 1 (CH3); NM R (CDC!,) h O.i7, 0.81>, 1.13 (3 s, 3 t•ac:h, C-·lb, C-8n, and C- llln CH3), Rnd f1.4!l (s, 2, HC=CH>. Anni: Calcd for C211-1,oO: C, 64 .51; H, 10.13. Found: C, 84.25; l!, 10.02. . l,6-(3'-Rutynyl)-2a,4 l>/j,811cx, l Oa/j-tet rn 111 et hy I- I ,2,3,4,4na,4 b,5,G,8a,9, I0,10a-dodecahydro-2/j-phena nth rol (8). To a stirred and ice-cooled mixture of o.:lG ml (0.68 1111110!) of 1.9 M ethereal methyllithium solution nnd 2.0 ml of dry ethnr wns nclded over a 2-min period a solution of 19.U mg (O.OG4 1111110!) of the acetylenic ketone 7 in 1.2 ml of dry ether. After stirriny, for 10 min lo11ger nt 0°, and for 5 min without cooling, the rcnction mixture was cnutlously quenrhed with 0.5 ml of water and t.hcn the product wns isolated by ether extraction . ~ 3 P u rificat.ion of the crude product (18.8 mg) by preparative TLC (50% ether-petroleum ether) afforded 16.0 mg (80%) of the alcohol 8 as a white solid, mp 84 -·88° (\'ncuum). The analytical sample, obtained aflcr two cry•talliz.ations of a portion-of this mater inI from ether-hexane, melted at 91.0-9 2.5° (vacuum): ir (CHCls) 3600 (OHl, 3300 (-C=CH ), 2115 (-C=oC-), and 1385, 1370 cm- 1 (CH 3 J; NMR (CDC!,) 6 0.81, 1.00, 1.03, l.17 (4 a, 3 each, C-2, C-4b, C-8a, and C-lOR Cl-ls) , a nd 5.45 (s, 2, HC=CH). Anal. Calcd for C22Hs4 0: C, 84.02; H, 10.90. Found: C, 84.01; 1-1, 11.04. l-Chloro-2-Corinyl-S-mcthoxy-4n/l, I Ob/l, J 2aa-t rimethyl-3,4,4a,4ba,5,6,10b,ll,12,12a-dccahydrochryse1ie ( 13). Following a modilication of t.he procedure of Moersch and Neuklis,23 icecooled phosphoryl chloride (12 ml, 20.1 g, 0.131 mmol) was Rti rred end treated over a I-min period with 13.6 ml (12.8 g, 0.176 mmol) of dimethylformamide. Af1er st irring for 30 min without cooling, the viscous solution of reagent was added at room temperature to a stirred solution of 1.158 g (3.54 mmol) of the tetracyclic ketone 1 in 24 ml of dimethylformomide. The stirred reaction mixture wn• then heated with a preheated, 60° oil bath for G hr so tha t the internal temperature rose to a canst.ant 55-56° . After cooling with an ice bath, the solution was poured onto 350 g of ire and 40 ml of 40% aqueous sodium hydroxide solution , and the product was isolated by dichloromethane extrnction. 33 The crude residue (1.310 g) was chromatographcd on 200 g of s ilica gel inn medium-pressure column with dichloromethane. After the first 400 ml of elue nt was discarded, evaporation of the next 600 ml of eluenl at red uced pressure provided 854 m~ (65'lo) of the chloroaldehyde 13 as a white solid, mp 196-198° (vac1rnm). The analytical sample, obtained after crystallization of a portion of thi• material from acetone-dichloromethane- wat e r, melted at 198.5-l!l9° (vacuum): ir (CHCb) 2750 (-CHO), 106.'i (C•=C-CHO) ](;05, 1575, 1500 (ArH), 1385 (CH 3 ). and 1150, 11)~0 «m- 1 (Ar0CH 3); NMR (CDC!a) b O.n2, 1.21 (2 s, 3 and C, C-4n, C-IOb, and C-12u CH3), 3.77 (s, a, Ar0CH 3 1, 6.i>-7.3 (m, 3. ArH), and 10.30 (s, I, CHO). Anal. Calcd for C2aH,.OiCI: C, 74.08; H, 7.84; Cl, 9.51. Found: C, 74.12; H, 7.96; Cl, 9.49. No material was elut 0 cl from the column by additional 350 ml of dichloromethane, but< -.·aporntion of the following f\00 ml of eluent afforded 154 mg (13%) of starling ket-0ne as a white solid; ir and NMH •Jl•Clra are the some as those of a purified sample of ketone I. Purther elution with 500 ml of 5% methanol-e ther gave 147 mg of a ·white solid, which on crystallization from acelone-dich loromethnne-water afforded 2,9-bisformyl-l -ch lo ro -8-mcthoxy4a/j,10h/j,1211a-trimcthyl-:l,4,4a,4ba,5,fi,1Oh,1l,12, 12a-decahydrochrysene: mp 26~ -- 2Gfi d<!c (vacuum); ir (CHCl3) 2770 (CHO), 1G70 (C=O), .1605, 1570, 1495 (ArH), and 11 50, 1055 cm - 1 (ArOCH3); NMR CCDCl,J b 0.92, 1.21 (2 s , :i and 6, C-4a, C-JOb, and C-12a CH3). 3.!10 (s , 3, ArOC:H3). 6.70 (s, l, (;.7 H) , 7.80 (",I . . C-10 H). 10.30 (s, 1, C-2 CHO), and 10.50 (s, I, C-9 CHO). Anal. Calcd for C,.Hw0 3Cl: C, 71.90; H, 7.2g; C:!, 8.84. Found: C. 71.82; H, 7.24; Cl, 8.90. The yield of the drsired ch lo ro aldehyde l 3 hn;<'d upon recovered starting material was 7b%. In a similAr experim e nt, in \\.·h ich the reaction mixtur~ was hea ted with a 6~ hath for 4.5 hr, the yi~ld of purified aldeh;Jc 13 (without recovc:1y of starting materi al) was 71%. 2a-Formyl-8-methoxy·2P,4a/j, l ObP, l 2aa-t I'! r11 methyl-1,2,:l,4 .4a,4 ba,5,6, I Ob, I l, 12, 12a-dodecu hydroch ryscno (14). To on argon-protected solution of 2GO mg (:!8 mi:-tttu•Q') of lith i11111 in 120 ml of dry ammonin and f1(1 ml of dry t etrnhyrlrof11rn11 wns s lowly added over a 50-min period with vigorous sl ir ring A solul ion of 253 mg (0.66 mmol) of thP ch loro Aldehyde l:l nnd 128 pl (100 1111:. 0 0 -66- ~:eland, Kowalelci, Tilley, and Walha 1.36 mmol) of dry tcrt- butyl alcohol in 60 ml of dry totrahydrofurun. After stirring for an add itional 15 min tho hlue color of tho rnaction mixture was . discharged by the portionwise addition of dry, powdered st1di u111 h«i11.0at.e, Rnd the nrnmonia was evaporated through a mercmy b11bhlN by heating t. he mixture with a hot air r,1111. After tlw addition of 40 ml of dry tctr9hydrofuran, the reaction mixture was stirred wit.h ·ice cooling end treated with 5 ml (11.4 g, 80 mmol) of iodomcthane. After stirring without cooling for 2 hr, rhe resultin1: white suspension wns diluted with 200 ml uf ether and the produd wtt s isolated by ether extraction,33 including both an acid and n bnse wash. Purification of the crude product (281 mg) by prcporntive TLC (15% ether-petroleum ether, double development) affordod 154 mg (64%) of the Aldehyde 14 (Rt 0.45): mp 111- 116° doc (vacuum); ir (CHClal 2805, 2705 (C HO), 1720 (C~O). 1605, 1575, 1500 (ArH), 1385 (Cl-lal. and 1245, 1040 cm- 1 (ArOCHs); NMH (CDC l:il 6 0.76 (s, 3, C -12A Cl-13). 0.93, 1.00 (2 s, 3 each , C-2 And C-4a CH3), 1.21 (s, ;l, C - IOh CH3). 3.76 (s, 3, ArOCHsl. 6.6-7 .3 (m , 3, Ari-I}. and 9.45 (s, 1, CHO). The same spectral properties were observed for the analytical sample which was prepared by cry•ta llization of a portion of this material from ether-hexane and also melted over the range 111-116° dee (vacuum). Anal. Calcd for Co.Ha.02: C, 81.31; H, 9.67. Found: C, 81.43; H, 9.67. Reduction of 143 mg (0.403 mmol) of ~his aldehyde Id in 10 ml of dry tetrahydrofuran with 110 mg (2.9 mmol) of lithium aluminum hydride afforded 139 mg (97%) of the corresponding primary alco hol as a white foam: ir (CHCla) 3625, :l470 (OH), 1606, 1575, 1495 (ArH), and 1385 cm- • (CHsl ; NM~ (CDCl 3 ) h 0.97, 1.00 (2 s, 6 a nd 3, C-2, C-4a, and C-12a CHs), 1.21 (s, 3, C-lOh CH3). 3.43, :1.M (2 s, 1 each, -CH20- ), 3.67 (s, 3, ArOCH,), and 6.6-7.3 (m, 3, ArH). The sa me spectral properties were observed for the analytical sample which was prepared by crystall ization of a portion of this material fro:n methanol and melted et 123.5- 125.5° (vacuum). Anal. Calcd for C2•HsG02: C, 80.85; H, 10. l 8. Found: C, 80. 73; H, 10.2(>. Attempts to converl this primary alcohol to the corresponding primary halid e with triphenyl phosphite-iodomethano,27 thionyl chloride-quinolino, and triphenylphosphine-ca rbon tetrachloride"11 led either to no observable reaction or a mi xture of numerous products, judged to he the result of deep-seated rea rrangements by their and Nl\IH spectra. 8-M ethoxy-2P,4a/j, l ObP, l 2aa-tetramethyl-2<>-( 3' -oxo-1 '-propenyl )-1 ,2,3,4 ,4a ,4 ba,5 ,fi, l Ob, I l, l 2, 12a-dodecahydrochrysenc (15) . Following An adoptnlion of the general procedure of Nagata arid Haynse,2 9 a stirred suspension of 880 mg (21 mmol) of 57% sodium hydride-mineral oil dispersion in 13.5 ml of dry tetrahydrofuran was conled with an ice bRth and treated over a 5-min period with u solution of 5.4G ~ (21 mmol) of cliethyl 2-(cyclohexylimino)E!thylµh osphonale in 25 ml of dry tetrahydrofuran. After 15 min, a solution of 1.100 1: (:l.10 mmol) of the aldehyde 14 in 20 ml of dry tetrahydrofuran was then added over " I-min period . This stirred mixture was heated with a preheated 60° oil both for 80 min , cooled with nn ice hath, and t hen poured onto 150 ml of ice nnd water. After isolation of the crude product by ether extraction"' there was ohlnined fl .3 g of a yellow-brown oil thnt contained the corresponding a ldirnine. Hydrolysis of the ~ ldimine was accomplished by t reatment of th i• crude product in 150 ml of benzene with 500 ml of 1% aqueous oxalic acid solution . This two-phase. system was stirred at room temperature for i!J hr. The organic layer was sepa rated and the aqueous laye r was then ex tracted with three 200-ml portions of ether The combined organic phases were washed with 2% aq ueous hyclroc hloric acid (200 ml), 2% aqueous •odium hydroxide solution (two 200-ml portions ), and r~,turated l>rine (200 ml ), and then dried (MgS0 4 ). After removal uf the drying agent and evaporation of the solvent nt rerlured pressure, 1.64 g of a yellow oil was obta ined. Purification of t hi s oil on 200 g of silica ge l in a mediumpr-e'5ure column wns accomplished by el ution with 40% ether-petroleum ether. When t. hc •er.ond 200 ml of eluent from the column wns evapornted nt r~duced pressure, there was obtained 1.066 g (fi9<:o) uf unsa turated aldehyde 15 as a white solid. Crystallization of A portion of th is mnlerio l from ether-petroleum ~ther afforded n1111lyticn lly pure mnt.eriul thut melted at 131 - 133° (vacuum): ir (CHCI,) 27:15 (CHO), 1675 (C=Ol, 1625 (C=C), 1605, 1575, 1495 (ArH) , 138.'i (Clla), und 10:15 cm- 1 (ArOCH3); NMR (CDCls) b ll.87, 1.04 (2 s, :l oncl fi, C -2, C-4a, and C-12n CHsl . 1.21 (s, 3, C-lOb CH:rl. :J.76 (s, 3, ArOCH,l , ti.ll5 (dd, I, J = 16 and 7.5,Hz, C-2' <> ~CH, 6.fi- 'i.:l (m, 4, ArH a ncl C-1' C=Cl-1). and 9.52 (d, 1, Jc 8 11,, CHO). Total Synthe.~is of (±)-S hionone -67- Anal. Cakd for C, 6 H3GOi: C, 82.0fl; H, 9.53. Found: C, 82 .07 ; H, 11.41. 8-Methoxy-2,3,•111,3, I Ob,3, 12aa- l <•I rn met h yl-2a-( 3' -ox up ro pyl)-l,2,3,4,4n,•l ba,:,,1;, I Ob . I I, 12, I 211-clodt•r·ahydroch1·ys"no. After the procedurt• of Nagni a nd C\l\\'orkPr~.:lt) a 8olution of I :\f) mg (0.355 mmo!} of the u11snturatcd aldc·hydc 15 and t.25 ml of triothylsilane in I ml of benzene was tr~atr«I with 4.[l mg (.1.9 1nnnl) of tris(triphenylphosphine) rhodium ch loride , nnd the mixturt• was heated ot 50° for 1.2~ hr. Whil" lwatin~ was ('(mt inued for nn Addi· iional 1.50 hr, two 2-mg (2.2 µmoll portions of the rhociium rntn· lyst were added al O.f,.hr intervals. Aflor dilution with 2;, rnl of ether and thrn fitrntinn , evaporal ion of the solven ts from l h1• filtrate at reduced pressu re afforded a yellow oil that cnnt:iincd the corrcspondin~ si lyl enol ether. A solution of this oil in 5 ml or acetone WHS treated with ll.f> ml of 5% aqueous hydrochloric acid, and tho mixture wao s tirred al room temperature for 20 min. Isolation of the product by ether exlrartion,33 including a base wash, afforded 200 mg of a y<•llow, sc mi crystalline solid which on purification l>y preparat,ivc TLC (35% ether-petroleum ether) gnve 113 mg (8:l%) of the saturated a id e· hyde <Rt 0.42) as a white, amorphous solid : ir (CHCl3) 27:15 (CHO), 1720 (C=O), !G05, 157f>, 1495 (ArHl, 1380 (CH 3 ), and 1035 rm-I (ArOCH3) ; NMR (COCl3) o 0.90, 1.UO, 1.05 (3 s, 3 each, C-2, C -4a, . and C-12a CH3l, 1.22 (s, 3, C-lOb CH,), 3.77 (s, 3, ArOCH,.l . 6.6·7.8 (m, 3, ArOCH 3 ), and 9.80 (m, I, CHO). This material wos not further purified but used directly in the following experinwnt. 8-Mcthoxy-2,3,4a,3, 1Ob,6,l2a a-let ra melhyl-2a· ( 4' - met hyl-3'pentenyl )-1,2,~.4 ,4 o .~ ba,5,6, l Ob, 11 , l 2, 120 -dodecahyd roe.h rysenc (16). A s tirred suspei1sion of 4.62 g (10.7 mmol) of isopropyl triphenylphosphonium iodide in 30 ml of dry tetrnhydrofuran at room temperature was treated dropwise over a 3-min period with 4.03 ml (8.55 mmoll of a 2.12 M solution of phenyllithium in 30% ether-benzene. The red suspension wos stirred for 2.2f> hr, and then a solution or 819 mg (2.14 mmo l) oft.he a hove satura ted aldehyde in 13 ml of dry tet.rahydrofuran was nclded over 11 5-min period. After stirring at room tempernture for 50 min longer th<, product was isolated hy ether extraction,''' including a 10% h ydro~en peroxide wash and a 10% sodium thiosulfote wash. After removal of the desicC'anl and evaporation of the solvent at reduced pressure, a semisolid mixture wns ohteined. This material was filter ed through n glass wool plug with the aid or 100 ml of petroleum et her to remove most of the relatively insoluble triphenylphosphine oxide. Concentration of the filtrate •l. reduced pressure affordc•d 1.2 g of a y•llow oil which was purified by chroniatogrnphy on 120 g of silica gel with 4% et her-petrol eum e ther. After the fiT't. 1;,o ml of eluent was disc11rded, the next 25 ml cnntained IOI mg of a mixture which on preparative TLC (4% et her--petroleum et hc• r) affo rded 69 mg (8%) of the olefin l~ which w11' combined wilh l.111, bulk of the product obtained Int er. Con ce nlrntion of t.he follnwin~ 250 ml at reduced pressure afforded 6<il mg (7f>%) of the olefin Iii as a white solid: ir (CHCl3) 1605, 1575, 1·1% (ArH), 1385 (('JI;.), and 1035 cm- 1 (ArOCH:il; NMR (CDCI,) ,\ 0.92, 0.98, 1.05 (;1, s, :i each, C-2, C-4a, ancl C-12a CH,}, 1.21 (s, 3, C-!Ob CH,) , 1.59, 1.nG [2 s, C=C(CH3h ]. 3.7fi (s, 3, MOCH3l, 4.8:>- 5.35 (m, l, C ~-- Cll}, a nd 6.fi.-7.3 (m, :1, Ari:J . The coml>ineci yie ld of the olefin Iii"'"' thus 720 mg (8:J";,), Crystallization or H porlion of this nrntrr inl from methanol wilh o lrnce of methyl•nr chloride afforded thr orrnlyti· cal sample AS nuffy, White crysta ls, mp 71:1- 80° (V8CllU01), And with the same ir and NMH spectra as thuse recorded above. Anal. Calcd for CwH 44 0: C, 85.23; H, 10.85. Found: C, R'dlG; H, 10.66. 4b,3,6aa,8,3, I Oatl-Tetramethyl-8"-( 4' - me thy 1-3' -pen l ~nv I)4,4aa,4 b,5,G,6a, 7 ,8,9, I 0, 1Oa, IOba, 11, 12-tetradccahydro-2(:! f/)chysenonc (18) . To a solution of M mg (0.127 mmoll <•f tlic olefin 16 and 2.1 m l (J.66 g, 22 ..5 mmol) of <lry tcrt-butyl ulcohol in 5 ml of dry tetrtthyrlrofuran anrl lf> ml of d ry ammoniA uncicr fi ll :trgon atmosphere was added with stirr ing 2·1 mi; (3.5 r,-atoms) ol lithi· um, and thEl mbi::t.ure was allowed tu r ellux fnr 2 hr. The c>:<cc~c; li1 hium was thPn quenched with 0.6 ml of methanol, and after evupo· r1ttion of the Hmmoniu, the crude product (!)(} mg) was i1.;oln k d by ether extrRction .:ti To a stirrC"d soluti0n of t his cruel(· dihydroaromatic sy:-:h•m in l nll o:' dichlorometha11e at room tcuipt:raturc was addod s<~quf•nt.lal­ ly 2 ml of ethnnol, 0.!; ml of waler, u11rl fiO mg of oxalic aricl . After stirring for 2 hr, the reaction mixt11r<' wns dil11tcd with !llll inl of ether, and the procl11rt (f,0 mg) wns isolated hy ether exlrn('linu,"3 including A 1111 •0 wn•h. . Conjugation of th<· doublo lrnnd of this /l,"(·Unsaturntc·d kPlonc (ir (CHCI~) 1710 cm-I (C=O)] was eff<•cted by stirr ing nl roo m temperaturr· • •olution or the crud e mnlcrial with 1 ml nf ethanol, J . Org. ChPm., Vol. 40, No. 8, 1975 997 O.lfi ml o r w•ler, and 0.30 ml of 10% pqueous Rodium hydroxide in 5 ml of did1lurnmethane. Aftl'r isolation of the crudp product (50 lllG) hy ellH'r cxtraction 33 and purification of that mul.eriul hy prepArn1 i"" TLC (40'\o ether-petroleum t•ther), there was.ohtained .35 mg (72'.'<> l of th£> desired enone JM u:i ·a pnle yellow solid : ir (CHCl3) rn;,r, <t>,·,O J. 161:1 (C=Cl, nnd 1390, 1375 cm- 1 (CH,J; NMR (CDCJ ,. ) ,\ O.HH, 0.90. 0.9:!, I. J.'J (4 s, :l uach, C-4b, C-6a, C-R, and C10a CH,,), 1.61, 1.67 12 s, :! Mrh, C=C(CH3)2], 5.0H (m, l, HCll -· =C H"'), and 5.87 (hrs, I, O=C-CH~C); analysi• by GLC (~00°)" 1 in1lic>1lc·d the presenco of B single volatile component to tlw extPnl of >H9% wit h rel<•11lion I ime of 3.8 min. This material "'"" 11si·d diH•cl ly in subsequent experiment.• without further purificntion. · I, 12.n- Epoxy-~ b,3,6aa,8,3,10n/!·lE'tramcthyl-8a-( 4'-mcthyl3'-pen l enyl)-l ,4,411a,4b,5,6,lio ,7,8,9, I 0, I Oa,l0ba,11,12,12a-hexa· dccahydro-2(;l/f)-chryeenon~. To n stirred solution of 392 mg (1.0 mmol) of the enone IR in 9 ml of dichloromethane at room lempcruture wu> added sequcnt~ally 20 ml of methanol, 5'.:l ml of :im;. a411eou" hydrogen pernxidr •olutlon, nnd 1.2 ml of 10% aqueous sodi um h)·droxide solution. After stirring in a clo~ed OaRk for 2 1 hr, t.he rrude product wns isolated by ether extraction 3a and on crystu ll izution from methanol-dichloromethane afforded 310 mg (7n%) of the epoxy ketone: rnp 103.5-105.5° dee (vacuum); ir (CHCl:d 1700 (C= O) and 137fi cm- 1 (CH3); NMR (CDCl3) 6 0.90, 1.12 (~ s, 9 and :J, C-4b, C-6a,·C-8, and C-lOa CH3), 1.59, 1.67, [2 s, 3 each , C=•C(CH"l2], 3.11(s,1, epoxy H), a nd 5.07 (m, l, C=CH). The annl)·ti CHI sample, obtained nfter two further crystallizations of R port iull of this mot.erinJ from methanof,dichloromethnne, had the Mm• s pert.nl 1iropertias and mel ted at 107-108.5° (vacuum). AMA i. Calcd for C2AH.,02: C , 81.50; H, 10.75. Found: C, 81.49; H, 10.73. IP-( :1• - Huty ny I )-4 b/!, 7,3,8aa, I On,6-tetramethyl-7 a·( 4'-mothyl3'-pcn l eny I )-3,4 ,4 na,4b,5,6, 7,R,8a,9,10, I Oa-d odecahyd ro· 2(1//)-phcnanthrenone. According to the procedure described above for th e cleavuge 19 of the epoxy ketone 5, a solution of 51 mg (0.124 mmol) of the epoxy ketone above and 24.5 mg (0.131 mmol) of p-loluc nesulfonylhydrazine in 2.5 ml of 1:2 acetic acid-d ichlorumet.h•ne was stored for 30 hr at -20° and then stirred for 13 hr at room temperature. After isol•t ion of the product by ether extraction,"'' including n hnse wash, and then purification of the residue by prcparutivc TLC (15% ether- petroleum e ther), 31 mg (li3%) of t he A£:<•tylcnic ketone <Rt 0.32) was obtained as a white solid: ir (CHCl3I :J:JOO (C=CH ), 2120 (C""C ), 1700 (C=O), nnd 1390, 1380 cm- 1 (CH3); NMR (CDC'3) n 0.76 (s, 3), 0.91 (s, 6), 1.21 (s, 3) (C 4b, C-7, C-8a, and C-lOa CH 3 ), 1.62 and 1.68 12 l>< s, 3 each , C=C(Cll,,),], and 5.1 (m, I, HCH= C). T he unalytical sample, ob. tainccl nfler two c rystallizations of this material from methanol, harf the sa me spectral properties and melted at 90.5-91.5° (vacuum) . Anal. Calcd for C2al-14, 0: C , 84.79; H, 11.18. Found: C, 84.72; H, 11.21. l P-(3' - ll u lynyl )-2a,4 bp, 7{J,8aa, I Oa,6-pl!ntamethyl-7a·( 4' • methyl-3' -pentcnyl )- l ,2,3,4 ,4aa,4 b,5,6,7 ,8,8a,9,l O, I Oa- tc l radecnhyd ro-2,6-p hcna n throl (20) . To a stirred and ice-cooled solution· of 1.1 ml (1 .81 mmol) of 1.65 M ethereal me l.hyllithium in 2.2 m l of dry l.rl rahydrof11ran was add ed over a 7-m iu period a solution of 55 m~ (0 . 1 1~1 mmo l) of the acetylenic keto ne above in 4.5 ml of dry tetrnhydrofurnn, nnd the mixture wns allowed to stir at room tem p erature fc,r on 'add itional 21; min . After the reaction WRS quenched with I ml of \\'Ater, the product was isolated by ether extraction33 and amount!'cl to [16 mg (99%) of a white foam: ir (CHCl3) 3600 (0Hl, :1:100 (C=-,,CH ), 2115 (C.,.C), and 1:185, 1375 cm-I (CH 3); NMH (CDCl 3 ) o0.8n (s, 6), 0.97, 1.06, 1.13 (:ls, 3 each) (C-2, C-4b, C-7, C-Ha, nncl C.JOo CH 3 ), 1.6~ and 1.68 [2 br s, 3 each, C=C(Cl-1 3 ) 2 ], nnd fi.I (n', 1, HCH=C); analytical TLC (fi0% etherpetrolcurn rthrrl showed a one-component system with R1 0.50. This material was not further purified hut used directly in m1merou• •c:id-cnto lyzed cyclization react.ions, All of which resulted in hydrntio n of the terminal double bond with or without cycl ization of the uce tyleuic side chain. 8a-(:l' -1 l ydroxypropyl )-4,4a a,4 b,5,6,6n, 7 ,8,9, I 0, I Oa, I Oba, 11,· I 2-ktrnd"cahyd ro-4b,3,fiaa,8,3, l OnP·tetrnmothyl-2(3 H)-chryeenone (I 7). To a ool11tion of 19 ml of dry dimethoxycthane nnd 2.3:1 ml of Hhnnol in 65 ml or dry ammonia containing 56 mg (8 mg-Al.,rns) of lithium was add<·d dropwiRe with stirring a solution of 7() mg (0.2 mrnol) of thC! a ldehyde 15 in 19 ml of dry din\ethoxye lhane and 2.~ 1111 of ethm101. The hlue color of the solution was mainlnincrl ovr.r n 2-hr period hy the JJ-Orl.ionwise addition of 497 mg (71 mg -ntomsl of lithium, a nd then the excess lithium .was destroyed J,y t.he nddition of :1 ml of methanol. After evAporotion of 998 J. Org. Chem., Vol. 40, No. 8, lfl75 . the ammonin in a stream of 1n~,m nnci the n treutml•nt. of the residue with 150 ml of 5% oqu""'" hydrnch loric a cid, the product w11s isolated Ly ethcr.ext.raction."" incluc.Jini; n hasc wnsh. The residual, light. yellm-.i oil was di~solvecl inn mixture of H ml (1f Pl hn11c1l Rlld G ml of n N aqueous hydro<"hlc )rir 1wicl, EHH.J lhP n.•stilt.ing solution was r~fluxccl urirl er An argun ntmni-;plll'rc for l hr. Arter dilution of the Aolution with 50 ml of wut e r, tJw prnchict. wns isolu l.l'd hy r.t her extroction, 33 including a bast· wns h , and then purifi f'd b~· preparntive TLC (ethe r). The res ulting d!' tH, cnlorle'8 oi l ;HIHHll\ll'd tu :,1 mg (Gf!'lo) of the hyclroxycnone 17. C:rys tullization of n por tion of this oil from n· hexane-dichloromc t hnne afforded 011ul y1 i«a lly pu re material: mp 144-146.5° (vnrut11n); ir (C:HCl:i) :H>IO, 3°100 (OH), 1655 (C=O ). Rnd 1610 ctn- 1 (C= C:i; NMH (CllC l:i) h O.BO (• ,Ii) 0.95, I.I& (2 s, 3 each) (C-4b, C·Ga, C·R, nnd C. !Oo CH:i), and 3.G2 (t, 2, J = 6 Hz, -CH20H). Anal. Cak<l for C26H4002: C , 80.f>9; fl, 10.82. F ound: C, SO.Bil; H, 10.80. l,12n·Epoxy· I ,4,4aa,4b,5,6,fin,7,R,9,10, LOn, I Oba, 11,12, 12ahexadccnhyd ro-8a-(3' -hydroxy propyl )· ~hl)',6na,8/l', I Oa/3· tetra· methyl-2(Jfl)-chryscnone. Tu R <t irred solutinn of 52 mg (0.15 mmol) of the hydroxyenone 17 in 3.6 ml of dichlornmethanc were added &.4 ml of methanol, 0.9 ml of :lO% Rquenus hydrogr n perox· ide, and 0.44 ml of 10% aqueous sodium h ydroxide >olution. After stirring under argon for 8 hr, the react ion mixture was poured into 5Q mJ of urine, Rnd the product wa< isolated uy ether extraction .33 Purification hy medium-prrss ure chromatography (ether) afforded 34 mg {64%) of the epoxy ketone and I ;l mg (25%) of recovered hydroxyenone 17. Crystallization of a portion of the epoxy ketone from ether afforded analyticall)· pure material: mp lfi!i- JGG 0 (vacuum); ir (CHCl 3) 3620 (OH) nnd 1700 cm- 1 (C=O); NMH (CDCJ3) h 0.88 (s, 9), 1.10 (s, 3) (C-4b, C-611, C-8, and C-!Oa CH:iJ, :U3 (s, 1, epoxy H), and 3.62 (t, 2, J = 6 Hz, CH20H). Anal. Cnlcd for 0.isH4o()3: C, 17.27; H. 10.38. Founcl: C, 77.33; H, 10.44. 1P· (3' -Butynyl) -3,4,4na,4 b,5,G, 7 ,R,Rn,9, I 0, I Oa -dodeca hydro7a-(3' -hydroxypropyl )-4bl)', 71J,8aa, l Oa/l·tctramcthyl -2 ( I JI). phenanthronc. To a dry mi xture of I 12 mg (0.29 mmo)) of the epoxy ketone above ancl 59 mg (0.~2 mmol) of fJ · loluenes ulfon·ylhydrnzine cooled to -20° under on a rg on atmosphere was added 4 ml of R -20° solution of I:~ glacial ace tic acid.-cli.:hlorum ethane which had heen previous ly dL"gassed by a lternate t·v1u.: uation And ehullition with &rgon . After 2~ hr at - 15 to -25°., followed by 12 hr at room temperature, the rea ct ion mixture was poured into 100 m l of WRter, ancl the product was isolat ecl by ether '"lraction,33 in cluding n hnse wa~h . Purification by chromAtogruphy on F lorisi l {l:I ehloroform-·ether) gave 8~ mg (77 % ) of the Rcdylenic ketone as a slightly yellow oil that was n ~i ntd t- - component ~ystem hy tic (1:1 d1lornform-e1.her, Rt 0.1~): ir (CHCJ,) :lGlO (OHi, :1300 {C.,.C:ll), 2120 (C=oC), and 1700 rm- 1 (C= O); NMH (CDCl 3 ) o 0.77 {•, 3J, 0.90 (s, 6J, 1.18 (s, (C .4b, C .7, C -8a, and C-JOa CH3), e nd :l.fiO (I,?, J = 6 Hz, CH,OH). A port ion of this oil was crystal lized from etlurnol-wnter, mp 99.f>- lOI 0 (vacuum). However, the resultinl! solid did not give n s.11isfnct(1ry combustion an al ysi s. Satisfactory result~ \\•e rr qlJta it1erl from a sample prt-pared by flash distillatinn of n porlio1o of th t! original oil nt 10- 4 mm. Anni. C:nlccl for C20H 4o02: C, 80.:,n; H, 10.82. Found: C, 80.GO; H, 10.85. I fl· (:I'· But y nyl )· 7 a·(3'-hyd roxy pro pyl) - 2a,4 bfj,7 /3,8na, I Oafl· pcntnmdhyl-1,2,:l,4,4Ra,4b,5.G,7,8,Rn,9, l 0,1 Oa-ter radccahydro2/l·plicnant hrol (19) , To a stirn'd and ice-cooled mixture of 4.8 ml (8.5 mrnoll of 1.7 M ethcre11l n.othyllithium solution nnd 7.5 ml of dry tetruhydrofurnn L111d(> r on aq~on atmosphe re was added dropwise tt solutio!l of 82 mg- {0.'.22 rnmol) of the Acc:tyleuic ketone above in 8 rnl of tl'lrahydrofurnn. After 10 min of !->tin ing without cooling the exce!'~ m~thyllit hi um was d<!stroyed with wuter, a nd the fmlution was dilu ted with l 00 ml of IJrinc. Isolut ion of thf:' product hy ether c:xtrac-tion, 33 foll11wf1cl h~· purifica tion hy ('hromntography on Florisil ( t(\<l(, ether-c hl,,roform), afforded 75 mg (88%) of acetylenic Ricohol 19 os a whilt', cry,1.alline solid . Cry, tnllirntion of a portion of this nrnl crinl frc1m fl· hexane-ether afforded a nalytically pure rnat(·rial that 1111•lled al J~:l-·155° (vacuum): ir (CHCl 3) 3t>IO (OH), :i:mn (C="C Hl, and 2120 cm- 1 (C,=C ); NMR (CDCl:i) b 0.88 !<,GI, 1.0, 1.0~. 1.15 (3 s, :1 .,,.d>l (C.2, C·4h, C -7, C-8a, and C lOn CH:il. and ~.G2 (t, 2, J = u Hz, CH,OHJ. A nnl. C:Rkd for C 2 eH 44 0,: C, WUf>; H , 11.41. Found: C, 80.32; H, 11.41. Sa-(:l'-Hydroxypropyl)-1 a,2,,·melhRno-1,2,3,4 ,faa ,~ b,.5,6,6a,7,K,~. l 0, I Oa, I Oba, I I, 12, I 2a-<;ci:trl cca hydro-4 bfi,lin a ,R/9,1 On· (l,12afl·prntRmt•thyl-2/l-chry, enol (21). A. Preparntion of tho Eno! Bi• Trifluoroncetutr. To ;;;, mg (0.22 mmolJ of the nce tyle· :o -68- Ire land, Kowalski, Tillcy, 1111d Walba ni c ciiol 19 cooled to -2:,• under nn argon atmosphere was ndded 1·1.f> nil of u -25° •ol111i1111 of :10% trifluoroacetic anhydride in tri· fluoruncetic ncid. Afler ~fi mi.11 of stirring at -25°, th1• solvents were rcmovt•d Rt redtwed 1irc•sure (-1 mm), and the dark residue WRs tnken up in ether nncl wnsh ed with water nnd soturated aque· <Il l ~ ~u diurn bicnrhonatt' solution . 'T'he resulting oil, which nmounted lo 1:1r. mg. was ll<rd direr1ly in the next experiment: ir (CHCl1) 178tl cm- 1 (C=Ol; N:O.IH (CDCl 3) b 0.87 (s, 9), 1.0, 1.03 (2 s, 3 rnchl (C'.4h, C-6u, C-8, C.JOa, and c . 12., CHs), 4.3:l (t, 2, J = 6 Hz, CH,-OTFAl , nnd 5.30 (,, 1, c.1 H ). B. Cloavuge oft.he 'l'rifluoroauetato and Mcthylenation. To n s tirred ·solution of lithium diisopropy!amide prepared from 5.1 ml (10.9 mmoll of 2.13 M hexane solution of n-butyllithium and 1.82 rnl (12.0mmul) of diisopru1)ylamine in 12 ml of dry tetrahy· drofurnn Al 0° under 11n Ar~on Htmosphere Was added a solution of J;JG mg of the crude bis trilluoroacetatc from aho•·o in 5 ml of dry tclrnhydrofuran. To the clourly, red.brown solution which resulted a fter I 5 min of stirrin~ were added all at once 43.6 ml (43.6 mmol) of the Simmons-Smith reagent prepared from the zinc-silver couple'" [4.9 g (75 mmol) of granulnr zinc, 38 mg of si lver acetate, and 38 ml of glacial ace tic add] a nd 5.78 ml (72 mmol) of diiodo· methane in 66 ml of dry ether in the presence of a few strands of silver wool after the procedure of Conia.sa Afler stirring at room temperuture for 50 min, tho solution was diluted with WO ml of ice cold, saturated sodium carbonate solution Hild 30 ml of 40% aqueous ammonium sulfate solution. lsolnt.ion of the product by ether cxtractiona:I including saturated aqueous ~odi um carbonate and 1()% aq ueous sodium thiosulfate solution washes afforded a dnrk rrd oil which was immediately chro· malogrnphed on 130 g of Florisil. Elution with 300 ml of petroleum ether removed diiodomethane, and continued elution with. 250 ml of I: I elher- petroleum·ethcr and then 350 ml of ether afforded 59 rng (67%) o f the cyclopropanol 21 as a yellow, crystalline solid. C ryst a llization of this material from CHCb or hexnne-dichlorornethane ~n\·e colorless crystals, mp -200• dee (vacuum): ir (CHCI,.) 3.580 cm- 1 (Ol-1); NMH (DMSO-d 6 ) h 0.449 (m, 2, cyclo· propyl CH2l. 0.915 (s, 9) , 1.12, 1.17 (2 s, 3 each) (C -4b, C.Ga, C -8, C-JOn, mid (' . 120 CH3), ond UG (t, 2, J 6 Hz, CH20H); h igh· re<olution, mass measured molecular ion 402.3497 ± 0.0008 (calcd for C;; H 4 a0 2. 402.:l497G). I ,.l ,4oa,4h ,ii,G,Ga,7,8,9, I 0,1 Oa,l Oha,11,12,128-HexadecahydroI {J,4 b/3,Gaa ,8/J, I Oa(l, l 2e(l-h cx a methyl-8a-(3' -hydroxypropyl)· 2(3/l)-chryserrnne (22). To a Holut ion of 54 (0.13 mmo!) of the cydopropanol 21 in 25 ml of ethnnol was added 30 drops of conC<•nlrated hydrochloric •rirl solu t ion, a nd the mixture was refluxed in un nrgon atmosphere for 40 min. After dilution of the solution wi th 17~ ml of water, th" product wus isolated by ethe r e xtrac· t ion,"" inrlucling hose wash , nnd then purified by chromatography on 20 g of F'lorisi l. Elution with 120 ml of ether guve 41 mg (77%) of tetrncyclic hydro xy ketone 22. Crystallization of a portion of this material frnm n-hexane affo rded analytically pure moteriRI: mp J.'>8 - 160° (vacuum); ir (Cl!C 13 ) 3610 (OH) and 1700 cm- 1 (C=O); NMH (C:DCl3) b 0.72 (s, 3, < · l2a CH 3), 0.80, 0.90, 0.93, 1.13 (s , 15, C.J, C·4b, C·Ga , C-8, and C.JOo CH 3), and 3.62 (t, 2, J = 7 Hz, CH,OH). Anal. Cnlcd for C21H 4fi02: C, 80 ..'\4; H, 11.51. Found: C, 80.43; H, l J.4:!. I ,4,4aa,4b,5,6,6a, 7,,1!,9, IO, I On, I Oba,l l,12, 12a-Hcxadecahydrol fl.~ ti{J,tiaa,8/l, 1Oa(l,12n/J·hexamcthyl-8a-(3' -oxopropyl) 2(3/1)-chryscnonc. To a suspension of 300 mg (3 mmol) of chromic an hyclrid e in 15 ml of dry d ichloromethane under an argon atmosphe re was added dropwise 0.48 ml (6 mmol) of pyridine. After 20 m in of stirring et room temperature, 2.19 ml (0.48 mmol) of this deep bur~tmdy Holution wa< added to rn mg (0.048 mmol) of the hydrox y kelone 22, and the .nix ture was stirred for 10 min. The red anJ hlHek mixture wus then filtered with the aid of suction through a lumina (Ill), and lhe a lumina was washecl with 150 ml of dich lorometha ne. Removal of the solvent at reduced pressure afforded 16 mg (84%) of t he keto aldehyde us a slightly yellow cry,tnllinc solid . CrystRllizatiun of a portion of this material from n·h<·rnne ttffordecl Rnalytically pure material as colorless rry<tals: mp li7 -li9' (vacuum); ir (CHCIJ) 2775 (CHO), 1720 (C = O), and 1100 t·m · 1 ( Cr~ O); NMH tC llC l3 ) o 0 .71 (s, 3, c.120 C H 3 ), 0.88, 0.89, 0 . ~2. 1.11 (4 •. 3 e11eh. C·4h, C-60, C-8. and C· IOa CH 3 ), 0.88 (d , 3,,/ = 7 Hz, C-1CH,),1111cl 9.8 (s, 1, CHO). Anal. Coled for C21H4,02: C, 80.94; H, 11.07. Found: C, 80.93; H, 11.l l. (±)-8hiononc. To a solution of 62 mg (1 .3 mmol) of triphenyJ. isnpropylphosphonium ie>did e in 4.5 ml of dry tctrnhydrofuran under 011 11rgon atmo<phcre WRS added dropwiNe 0.50 ml (1 mmol) Q mr. Total Synthesis of (±) -Shionon r -69- of a 2 M 11-hexnne- phe nyll il hi um •olut ion . Afte r sl ir rinf: fo r 2 hr at room temperuture, 3.7 ml (0.7·l mrn ol) of thi • rcHgent wos added to a solution of 19.6 mg (0.049 mmol) of the a hove k<•t o oldehyde in 2 ml of' dry tetrnhydrofurnn . Afler ;lirring for W min nl room temperature the pro riucl wn~ i ~o lal f'd hy ether c xtraction,:l:t i11cludi11g a 10% aqul·ou~ h ydroge n pero>: id1: wu!-.h, und purifi C' rl by chroma· tography on 20 g of silica ge l. Elution with 100 ml of 10% etherpetroleum elher afforded 17 mg (82,\,) of (± ) -shionone. Annlylinil1y pure moteriul wus obt.ainerl u p11! 1 c: ryi:;t.all ization from. methnnol as small needles: m p J61.5-lt3:J 0 (vnruumi; ir (CHCJ3) 17!10 cm- 1 (C=O); N l'v!H (CDCl:i) o 0.71 (s , 3, C-120 CH3), O. S8, 0.90, 0.92, 1.13 (4 s, 3 cnch, C-lb, C-6n , C-8, end C - lOn CH3), 0. 8H (d, :!, J = 7 Hz, C-1 CH 0 ) , 1.6 1, 1.69 [:.! s, 3 ea ch , C=C(CHahJ. a nd fi.10 (l, I, J ~ 8 Hi, RCH 0 =Cn2l. Anal. Calcd for C 30 H 500: C, 8·1.44; H , 11.81. Found: C, 84.38; H, 11.90. The ir, NMR, GLC, anrl TLC of this m ate rial were irle nt ical with those found for an autheul ic sn mple of natural shiononc provided by Professor G. Ourisson. Registry No.-1, 53311-24-3; 2, 54 141-74-1; 3, 5403G-92-9; 4 axial alcohol, 54036-93-0; 4 l' Qll.1 lurial a lcohol, 5-lO:lH -94-1 ; 5, 64036-95-2; 6, 54036-96-3; 7, 540~6-97 -4 ; 8, 54036-98- 5; 9, 54036- . 99-6; 10, 54037-00-2; II, 54037-01-:l ; 12, 54037-02-4; I~. 54062·792; 14, 53311 -25-4; 1s, 54037-0:l-5; 16, 54054-05-6; 11, 5;n11-20-5; 18, 1>4054-0G-7; 19, 54037 -04-6; 20, 5·l054-07-8; 21, 5.1082-4 1-6; 22, &4037-05- 7; 23, 53402-15-6; tet.ra mcthyldiamidoplH>s phoruc hlo ridate, 1605-65-8; 8- methoxy-4a/l,10b/l, 12arc- t rimcthyl3,4,4a,4ba,5,6,10b,11,12, 12a-d ecohyd rochrysene , ii4 141-75-2; ltl,12a,8-mcthano-4 b/l,6an, I Oa{i-tri met.hy I- la ,2,3,4,.\ n(l .4 h,5,6,Ga ,· 9,10,10a, 10bn,l 1,12, 12a-hexndccahyd ro -21l -chrysenol , 54037-0G-8; 2-acetoxy-l /3,l 2aiJ-me lhano- 4bll .GJ«, J Oa,6-trimethyl-1,4 ,4H".1 b,5,6,6a,9,10, !Oa,!Obo, J l,l 2, 12n-telradeca hydrochry•e nP, 54062-80-5; 2-trifluoroacetoxy-4hll,Gna ,1 O!t /] ,l 2n IJ-lctrom e thyl-3,4,4 ll« ,4 b,5,6,6a,9,10, J Oa, l Oba ,! 1,12, J 211-lt!t rndccuhydrochrysPnc, 54037- 079; 2,9-bisformyl-1-chloro·S-met hoxy-4a;3,!0b/l, l ~aff- t rim othyl 3,4,4a,4hcr,5,6,10b, 1 l ,12,12o-dccah yd rnchrysene, 5'10;17-08-0; 8methoxy-2/l,4a/l, 1Obtl,12aa -tclra111 " th yl- l ,2,3 ,4 ,4a,4bn ,\6, 1Ob,11 , 12,12a-rlodcrahydrochryscn-2<>-mcthanol, 54037-09- 1; 8·meth oxy2{J,4a{J,10bll,12tta-lct ramethyl-2.v -(3' -oxop ropyl) -1,2,3,4 ,4 n,4 ba ,5,6, 10b, 11, !2, 12a -dod.,cahyd roch rysene, 54037-JQ ..1; I, 12aepoxy-4b/j ,6aa ,Sp , l Oa/l -te t rn methyl-&.-( 4 '-methyl -3' -pe nt cnyl)l,4,4aa;4 b,5,6,6o,7,8,9,10, 10n, 1Oh",11,12,12a-hexadrc11hydro2(3H) -chryscnonc , · 54054 -08-9; l/J-(3 -butynyl)-4bii,7/3,8u" !Oa/ltetramethyl- 7 n-( 4 '-methyl-3' -pen tenyl )-:J,4,4aa ,4 h,5, G, 7 ,R,8a,9,54054 -0Q- O; I ,12a10 ,1 0a -dorlecah yd ro-2( ll/) -ph e11on t.hro11e, epoxy-l,4,4aa,4 b,5,G,fla,7,8,9, !ll, I Oa, 1Obu ,l l,l 2,12a-h rxo<lccn hy.dro-8a-(3' -hydrox yp ropyl )-4 h/l,Gan ,8/i, l Oa/l-tetramt•I h vl -2(:\/ I) chry11eno11e, 54037-11-5; 1.8 -(3 ' -buly 11yl) -3,4,4nn,~h.:1,G,7 ,K,13n,9,10,10o-dodecahydro-7a- (3' -hyd rr,xypropyl)-4h/i,7/l,)laa ,1O;i/3- tetramethyl-2(1H)-phenanthrone, f•::l111-27-6; 2-trifluor";icet.oxy-8<r(3'-trifluoroacetox ypropyl) -4 li/3,6art ,8/l, l Oa/l, I 2a/l-pr· n In methyl 3,4,4aa,4 b,5,6,6a, 7,8,9,10, J .Jn, J OL<r. I l, 12, 12a-hexaclecH Ii .I'd ruch rysene, 54037-12-6; . 4,4acr ,4b,5,G,Go,7,8,9,IO,!Oa, !Oba, l l . 12, 12a- hexadecahydro-1/l,4 bp,6aa,8il, J Oa/l, 11.R;:!-hexRmcth yl-Sn- (3' -oxopro· pyl)-2(3H)-chrysenone, 54037 -13 · 7. Acknowledg me nt_ Grateful ack nowledgme nt is mnde for the support of this work hy gra nts from th e Nati onal Science ~'ound a ti o n nnd the Hc,ffmann-La Roche Fou?1dation. The Rut.hors thank Professor G. Ourisso (llnivcrsity of Strasbourg) for a generous silmp le of sh i onon~ used for comparison purposes. References and Notes (1) The structural formulas containing onu or more asymmoh ic carbon atoms d~pict one ona ntiomer lid refer to racomic com µou11d5 unluss otherwise designated. In tho text the( ± ) prefl )t will be on1itlt)d, and intermediates are to be assumed to b e raccmic. Tho tetrucyclic compour1ds will be des::;ribbd by the c~1rystJno nomc·r.c lature and ~;. c.h rncema le Is arbitradly ;e ~resentOO b~,. Iha: cn,:,n:.,,:,mer thal has lt.u C-Ga (C-12&j methyl f}'O\IP in the o con figu1a l1l>r 1. (2) National lnstltu:us 01 Heallh Tr ;iincc. 1969-- 1973. (3) National Sctence Foundation PrecfrJctma l Follow, 1960- 197::?. (4) National Oafonse Education Act Tt ~inP.o. 1971-1974. (5) R. E. Ireland, M. I. Dawson . C. J . Kowelskl, C. _A. Lipinski. D . R. Morshall. .J. W. Tilley, J. Bordner. and B. L. Tru, , J. Org. Gliem.. 40, 973 (rn75 ). (6) Y. Tarrahasr1i, Y . Moriy;,m~ . T. To\.c.:\ha!>hi, F. P~til. ;-ntJ G. Ourisson, Bun. Soc. Chim. Fr .. 237 4 ( lQ~ G) ; T. Ta hahashi. T . Tsuyuki, T. Ho:;himo, and M . Ito, Tetrahedron L<' lt .. 2997 (t 967 ). J. Org. Chem., Vol. 40, Nu. 8, 1975 999 (7) A preliminary report on somo of !his work has been published: R. E. lrelnnd. C. A. Llµinski, C. J. Kowalski, J. W. Tilley, nnd D. M. Walba, J. Am. Ch9m . Soc .. 96, 333~ (1974). (6) n. E. lrelHnd, D. C. Much moro, and U. Hangartner, J. Am. Chem. Soc., 94, ~09B (1972). (9) R. E. Ireland, M. I. Dawson, S. C. Welch, A. Haponbach, J. Bordner, end 13. Tr us. J Am. Chem. Soc .. 95, 7629 (1973). (10) W. Nogala •. M. Yoshioka, and S. Hirai, J. A m . Glie m . Soc., 94, 4635 (19 72 ). ( 11)· H. J. Czarny, M .S. Thcoio, California lnslltuto of Technology, 197 1. ( 12) W. G. Dauben nnd G. H. Bere,ln, J. Am. Chem . Soc., 85, 466 (1963); W. G. Dauben 811d A. C. Asl1cra fl , Ibid., 95, 3673 ( 1963 ). (13) H. E. Simmons and R. D. Smitl1, J. Am. Chem. Soc. , 81, 4256 (1959). (14) H. C. flrown and W . C. Dickason, J. Am. Chom. Soc., 92, 709 (1970). (15) W . G. Dnuben end E. J . lleviny, J. Org. Chem., 31, 3794 (1966). (16) G. Stork, P. Ro s~ n . and N. L. Goldman, J. Am. ChBm. Soc., 83, 2965 (196 11. . (17) A. 0. House and V. l( rarner, J. Org. Chem., 28, 3362 (1963); H. 0. Houso and B. M. Tr osl . Ibid., 30, 1341, 2502 (1 965). (1R) 11. W . Whitlock. Jr. , end L. E. Overman, J. Org. Gliem., 34, 1962 (1 969). (19) A. Eschonmoser, D. Felix, and G. Ohloft, HBlv. Cli/m. Acta, 50, 705 (1 967 ); 0. Felix. J. Sc l11 eitJar, G. Ohloff, and A. Eschonmoser, Ibid., 54, 2896 (1971); M. Tanabe, 0. F. Crowe, R. L. Dehn, and G. Detro, Tetra· hooron Lett. , 3739 (106 7); M. Tanaba, D. F. Crowe , and R. L. Dehn, l!Jlcf., 3943 (1 96 7). (20) P. E. Petorson and n. J. Ka mat, J. Am. Chem. Soc., 91, 5421 (1 969). (21) W . S. Johnson, M. B. Groneslock , R. J. Parry, R.R. F. Myers, T. A. Cryson, and D. H. Milos, J. Am. Chem. Soc.. 93, 4330 (197 1); P. T. Lansb1>1y and G. E. DuBois, Cl>.,m. Commun .. 1107 (1 97 1). (221 R. E. Irela nd, S. W. Ualdwln, and S. C. Welch, J. Am. Chem. Soc., 94, 2056 (1972). (23) G. W. Moersch and W. A. Neuklls: J. Chem. Soc., 788 (1965). (24) n . E. Ire land and L. N. Mander, J. Org. Chem .. 32, 6 69 (1967); H. 0. House and T. M. !lore, ltild ., 33, 943 (1966). (25) Seo. for oxamplo, H. H. lnhoffon, Angew. Chem ., 72, 875 (1960); 70, 576 (1958). (26) M. F. Sem1nelhack. Org. React., 19, 147 (1972). (27) J. P. H. Verheyden and J . G. Motfa tt. J. Amor. Chem. Soc .. 88, 5884 (1966). (26) J. G. Colzada and J. Hooz. Org. Synth. , 64, 0000 (1975). (29 ) W . Naga ta and Y. Hayese, J. Chem . Soc. C, 460 (19G9J. (30) I. Qj;ma, T. Kogure, and Y. N•gal, To trRliedron Lott., 5035 (1 972). (3 1) Molting points labtiled (vacuum} were taken In evacula ted cap illar\es on a Hoo ver capillary melting point apparatus, w hlle oil others were deter· mined on a Kofler mic1o hot st;:igo melting point apparatus. AH melting points and boiling points nro uncorrec ted. Infrared (Ir ) spectra were d&-term ;ncd on a Perki11-f. hner 2370 grating Infrared spectrometer, and n&.r clea r magnetic resonance (NMR ) spectra were recorded using either a Varian A GOA or T- 60 r.~ec trometer. Chemical shifts are reported as 6 values In parts per million ralalivo lo TMS (hn,5 0.0 ppm) as an Internal standard. Gas-iiquid pha se chrooiHlographlc (GLC) analyses we re delermlnad on either a Hewlc!t-Packard 5750 or F & M 810 resoa , ch chromatogruph using helium ca rr ier gas at a flow rato of 60 ml/min . Unless othe r~ wi•e noted, all analylicel GLC was conducted on a 6 ft X 0.125 In. column packed with 4% SF. -30 on 60-80 meshChromosorb W AW DMCS. PrcpMative thin lnyer chromntography (preparative TLC) was carried out on 20 X 20 X 0.2 cm gla:;s plates coated wilh silica gel PF 26 4+ 2eo (Brlnkmnn !nslrumonls Co). Ana lylica l thin layor ~h r o m a l ography (TLC) wus c onducted on 1 X 3 In. mk:roscopo slides coated with a 0 .5-mm layer of silica go\ G or PF254+2ee. Alumina used tor column chromatography refers to the grade I. neu-tral va riety manufact• .red by M. Woelm, Eschwege, Germany, and made Uf"J to grade II or Ill a~ indica ted by the addition o f 3% or 6% water prk>r • to use. Sfflcti gel columns used the 0.05 -0.2 mm silica gol manufactuied "for c olumn chrom&togra phy" by E. Merci..: & Co., Darmstadt, Germany. Prepara liva medium-pressure column chromatography was P"rformed using Yi X 20 In. or 2 X 20 ln. gl ass columns with fit11ngs suppliGd by Chrornatronlx, Inc., Berkeley, Calif .• and an Instrument mln~ pump suppl10d by Milt on Roy Co. , St Petersburg, Fla. (instrumentation designed by R H. Mueller. these laboratories, aod copiee, are availabk3 on rnqucst). The colu1n ~s wot e packed with &!Ilea gel H "for TLC acc. to Sl ahl" (10-40 µ) manufactured by E. Merck and Co .. Darmsladt, Gerniuny. Solvents woro degassed under water aspira tor vacuum prior to use. ' "Ory" solvents wero dried Immec'.,.·tely prior to use. Ether, benzene, tetrahydrofura n , and dimethoxyethane were di stilfod from lithium a luminum hydride: tort-bu!)'; ·.Jcohol, trlmethyl sulfoxide, pyridine, a n<' ' •examelhylphosphorarnirte (~lf\l\PA) were Distilled from ca lcium hydrkle; d.._ c hlornmethane. c arbon l1Jtrnchlorido, diodomAthan e, nnd m ethyl iodkHJ wern disUlled from phosphorus pentoxlde; ammonia was distilled from lh'o tc.nk and the n h (.lffi a blue lithium or sodium solution. "Petroleum P.lhnr" rofors to the "Analyi.ed Reage rt" graclH hydrocarbon fraction, bp 30-60°, which is suppliod by J. T. Baker Co .. Phillipsburg, N.J., and W:':IS not further purified. Rua ctions described as run under nttrogen or argon employed a mercur;• bubbler arrn r 1 ~jt:d so thn t tho system coutd he alh~rnately ovacuoled and lilled wi1h thi.l in~rt {PS Hnd let! under a posilivo pres sure. Mic roannt1ses wom pe rfCJrmod by Spang Microa nalytical Laboratory, Ann Aibor, Mich. (32) H. G. Gook, J. D . llell. B. C. Saunders. G. J. Stacey, H. G. Walson. I. G. E. W•ld;ng, nnd S. J. Woodcock, J. Chem. Soc. , 2021 (1949) . (3 3) In CS':iE'S where products were isolated "by solve nt axlraction", the procedure ge i'l er aU~· followed was lo exlract the aquoous layer with several pvrtivns tho indica te(! solvent; then the organic layers we ro comtJined Eind wa shed with water , followed by saturated brino. Tho org3nlc layer 1000 J. Or11. Clw111 .. Vu/ . .JO, No. 8 , l .'175 was dftod over anhyc1ruus sodium or m:1~111••si 11111 :"•lllRto, th r-n f1lh ·•1•1I, and the ~olvc nt Wet!' t!Vttp01Jlt-d fr on1 th.i 11111.1:<1 11nd tJr ro<1u1:1!11 111 1•··· SUfO (~'8 1or A5p iratnr) ~ 1 :-:in g 8 fOtmy l!V." ll'• 11: thl1 The USO o f th1 • l,.,1 ll S ''base wa ~: h" or .. acid w;i ~ ;h'' 111cl1cnl!• w.1 ·.hi1 i. 1 tilt~ combillthl 1'1 q .u.;, &dyers w it h snh" nh){l uquuous sodium l"ic. 1d 11. 111: 11t:' so lution 0 1 w1lli ;1,11 11t' -70- ln•l111icl <•I 111. nq11< 11•t.·. h )'C'1ro1' hln1 l1: .1H:ll1, rt •:,puc llvo ly , pr ll>r to the ufornn11111liont1.1 w, 1:-, l11r1q w ith wu lnr , ('.\4) r I P l ;n 11 . ,I l ,,, , (...'l t1• m' 29, :'04 11 ( Hl(i<1) ( :~!-.) 11 1:11o·ti 1t1 · 1.11\:I H f\od r 1ho1~ ; 1, .I O.o t:11c·111 . 35, •WOO ( Hl'IO ) p H) J . f.. I t lt• 111r., l i linnrd, u nd J .. fv\. C t•nln, 0y1Jtll,1~I.-., 5411 ( l!l /~ ) -71References l. (a) The structural formulas containing one or more asymmetric carbon atoms depict ohe enantiomer but refer to racemic compounds throughout. In the text the (±) prefix will be omitted and intermediates are to be assumed to be racemic. In this discussion, picene nomen- . lb w1·11 be use d t o descr1"be pen t acyc 1.ic comc1a t ure an d num ber1ng pounds, and each racemate is arbitrarily represented by that enantiomer that has the C-12b (C-14a) methyl group in the a-configuration. However . in the experimental section tricyclics will be described using the phenanthrene nomenclature and numberinglb while tetracycl ics will be described using the chrysene nomenclature and number;.. ing 1b; (b) IUPAC, "Nomenclature of Organic Chemistry", Butterworthsg London, 1969, 2. T. K. Devon and A. I. Scott, "Handbook of Naturally Occurring Compounds", Vol. 11, Academic Press, Inc., New York, 1972, pp. 281-384. 3. A. A. Newman, "Chemistry of Terpenes and Terpenoids", Academic Press, New York, 1972, pp. 207-287. 4. (a) L. F. Fieser and M. Fieser , "Steroids", The Waverley Press, Inc., Baltimore, Md., 1959; (b) A. A . Akhrem and Y. A. Titov, "Total Steroid Synthesis", Plenum Press, New York, 1970; (c) J. Fried and J. A. Edwards , "Organic Reactions in Steroid Chemistry", Van Nostrand Reinhold Co., New York, 1972. 5. "The Total Synthesis of Natural Products" , Vol . 11, Ed. J. ApSimon, Wiley-Interscience, New York, 1973, pp. 575-632. 6. (a) R. E. Ireland, S. W. Baldwin, D. J. Dawson , M. I. Dawson, J.E. Dolfini , J. Newbould, W. s. Johnson, M. Brown, R. J. Crawford, P. F. -72Hudrl ick, G. ·H. Rasmussen, and K. K. Schmiegal, J. Am. Chem. Soc., 92, 5743 (1970); (b) G. Stork, S. Uyeo, T. Wakamatzu ; P. Grieko , and J. Labovitz, J. Am . Chem . Soc., 93, 4945 (1971). 7. R. E. Ireland, M. I. Dawson, S. C. Welch, A. Hagenbach, J . Bordner, and B. Trus, J. Am. Chem. Soc., 95, 7829 (1973) . . 8. R. E. Ireland, C. J, Kowalski, J. W. Tilley, and D. M. Walba , ~ Org. Chem., 40, 990 (1975);Appendix of this thesis. 9. (a) E. J. Corey and J. J. Ursprung, J. Am. Chem. Soc ., 78, 5041 (1956); G. Brownlie, F. S. Sp r ing , R. Stevensen , and W. S. Strachen, J. Chem. Soc., 2419 (1956) ; (b) P. DeMayo, "The Higher Terpenoids" , Vol. III, Interscience, 1959, pp. 202-213; ref. (3), p. 220. 10. (a) R. E. Ireland, D. A. Evans, D. Glover, G. M. Rubottom, and H. Young, J . Org. Chem., 34, 3717 (1969) ; (b) R. E. Ireland, D. A. Evans and P. Loliger, J. Org. Chem ., ~~, 3729 (1969). 11. R. E. Ireland, S. W. Baldwin , and S. C. Welch, J. Arn . Chem. Soc ., 94, 20 56 ( 19 72) . 12. W. Nagata , M. Yoshioka, and S. Hirae , J. Am . Chem. Soc ., 94, 4635 ( 1972); W. Nagata , M. Yoshioka, and M. Murakami, ibid. , 94, 4654 (1972); W. Nagata, and M. Yoshioka, Org . Syn., 52, 90 (1972 ); ibid., 52' 100 ( 1972) . 13. For a detailed discussion of the hydrocyanation reaction used in the formation of trans-di axial methyl groupings and a short discussion of other routes see (a) C. J . Kowalski, Ph . D. thesis , C. I.T. (1974) . For other routes see (b) R. E. Ireland, D. R. Marshall, and J . W. Tilley, J. Am. Chem. Soc ., 92, 4754 (1970); R. E. Ireland, M. I . Dawson, C. J. Kowalski, C. A. Lipinski , D.R . Marshall, J. W. Tilley, J. Bordner, -73and B. Trus, J. Org Chem., ~Q· 973 (1975); R. E. Ireland, P. Bay, K. C. Cheng, R. J. Czarny, J. Moses, and R. I. Trust, ibid., 40, 1000 (1975); R. E. Ireland, T. C. McKenzie, and R. I. Trust, ibid., 40, 100 7 ( 19 7 5 ) . 14. A. J. Birch, Quart. Rev. (London), ~~' 17 (1958). 15. C. Djerassi, 11 Steroid Reactions 11 , Holden-Day, Inc., San Francisco, Calif., 1963, pp. 269-289 . 16. F. G. Mann and M. J. Pragnell, Chem. Ind. (London), 1386 (1964). 17. A. L. Wilds and N. A. Nelson, J. Am. Chem. Soc., 75, 5366 (1953) . 18. Alex Hagenback, unpublished results. 19. H. E. Simmons, T. L. Cairns, S. A. Vladuchick, and C. M. Hojness, 20. Organic Reactions, Vol. 20, J. Wiley and Sons, Inc., 1973, pp. 1- 131. 20. P. T. Lansbury and G. E. DuBois, Chem. Commun., 1107 (1971); for a review of this type of cyclization see ref. 13a, pp. 29-37. 21. H. 0. House and V. Kramer, J. Org. Chem., 28, 3362 (1963); H. 0. HOUSE and B. M. Trost, ibid., 30, 1341, 2502 (1965). 22. H. W. Whitlock, Jr., and L. E. Overman, J. Org. Chem., 34, 1962 ( 1969). 23. A. Eschenmoser, D. Felix, and G. Ohloff, Helv. Chim. Acta., 50, 705 (1967); D. Felix, J . Schreiber, G. Ohloff and A. Eschenmoser, ibid . , 54, 2896 (1971); M. Tanabe, D. F. Crowe, R. L. Dehn, and G. Detre, Tetrahedron Lett., 3739 (1967);M. Tanabe, D. F. Crowe, and R. L. Dehn, ibid., 3943 (1967). 24 . E. J. Corey and H. S. Sachder, J. Org. Chem., 40, 579 (1975) . 25. E. I. Eliel, N. L. Allinger, S. J. Angyal, and G. A. Morrison, 11 Conformational Analysis 11 , Interscience, 1965, pp. 345-346. -7426. J. M. Denis, G. Girard, and J.-M. Conia, Synthesis, 549 (1972). 27. D. H. Gibson and C.H. DePuy, Chem. Rev., 74, 605 (1974). 28. (a) T. Kikuchi, M. Niwa, and N. Masaki, Tetrahedron. Lett., 5249 (1972); (b) P. Rogers and D. J. Williams, ibid., 63 (1974) . 29. See S. W. Baldwin, Ph.D. thesis, C.I.T., 1969 for an example of such a selective ketalization. 30. H. C. Brown and R. F. McFarlin, J. Am. Chem. Soc., 80, 5370 (1958). 31. Ref. 4c, Vol. 1, pp. 222-264. 32. W. Nagata, private communication. 33. H. 0. House, 11 Mode m Synthetic Reactions 11 , Benjamin, Philippines , 1972, pp. 152-158. 34. R. E. Ireland, D. C. Muchmore, and U. Hengartner, J. Am. Chem. Soc.» 94' 5098 ( 1972). 35. Melting points labeled (vacuum) were taken in evacuated capillaries on a Hoover capillary melting point apparatus, while all others were determined on a Kofler micro hot stage melting point apparatus. melting points and boiling points are uncorrected. All Infrared (ir) spectra were determined on a Perkin-Elmer 237B gr ating infrared spectrometer, and nuclear magnetic resonance (NMR) spectra were recorded using either a Varian T-60 or HR-220 spectrometer. Chemical shifts are reported as 8 values in parts per million relative to TMS (8TMS 0.0 ppm) as an internal standard . Gas-liquid phase chromatographic (GLC) analyses were detennined on either a Hewlett-Packard 5750 or a Varian 940 research chromatograph using helium carrier gas at a flow rate of 60 ml/min. All analytical GLC was conducted on a 5 ft x 0. 125 in column packed with -754% SE-30 on 60-80 mesh Chromosorb WAW DMCS. Preparative layer chromatography (PLC) was carried out on pre-coated PLC plates with a 20x20x2mm layer of silica gel 60F-254 on glass plates manufactured by E. Merck, Darmstadt, Germany. Alumina used for column chromatography, unless otherwise stated, refers to the grade I, neutral variety manufactured by M. Woelm, Eschwege, Germany, and made up to grade 11 or 111 as indicated by addition of 3% or 6% water prior to use. Silica gel columns used the 0.05-0 . 2 mm silica gel manufactured "for column chromatography" by E. Merck and Co., Danns tadt, Germany. Prepara- tive medium-pressure column chromatography was perfonned using 1/2 x 20 in or 2 x 20 in glass columns with fittings supplied by Chromatronix, Inc., Berkeley, Calif., and an instrument minipump supplied by Milton Roy Co., St . Petersburg, Fla . The columns were packed with silica gel H "for TLC acc. to Stahl" (10-40µ) manufactured by E. Merck and Co., Dannstadt , Gennany. Ether and petroleum ether were degassed under water aspirator vacuum prior to use. "Dry" sol vents were dried immediately prior to use. Ether, benzene, tetrahydrofuran (THF), and dimethoxyethane (DME) were distilled from lithium aluminum hydride, tert.-butyl alcohol, pyridine , di isopropyl amine, hexamethyl phosphorami de (HMPA) , and n-hexane were distilled from calcium hydride; dichloromethane and iodomethane were distilled from phosphorous pentoxide; ammonia and ethylamine were poured from the tank and then distilled from a blue solution -76- of lithium. 11 Petroleum ether refers to the "Analyzed Reagent 11 11 grade hydrocarbon fraction, bp 35-60°, which is supplied by J. T. Baker Co., Phillipsburg, N.J ., and was not further purifi ed . · Reactions described as run under argon employed a mercury or silicon oil bubbler arranged so that the system cou ld be alternately evacuated and filled with inert gas and left under a posi tive pressure . Mass spectral analyses were performed by Ms. Beth Irwin, UCLA , Los Angeles, Calif . Micro -analys es were performed by Spang Micro- analytical Laboratory, Ann Arbor, Michigan . 36. In cases where products were isolated by solvent extraction 11 11 , the procedure generally followed was to extract the aqueous layer with several portions of the indicated solvent; then the organic layers were combined and washed with water, followed by saturated brine. The organic layer was dried over anhydrous magnesium sulfate, unless otherwise stated , and the sol vent was evaporated from the filtrate under reduced pressu re (water aspirater) using a rotary evaporator and the resulting residue was dried.:!..!:!. vacuo (~ 1 mm). the terms 11 The use of base wash or acid wash" indicate washing the combined 11 11 organic layers with satu rated sodium bicarbonate solution or with dilute aqueous hydrochloric acid, re spectively , prior t o the aforementioned washing with water. The term 50% brine refers to the solution resulting when a satu rated aqueous solution of sodium chloride is diluted with an equal volume of water. -77- PROPOSITIONS -78- Abstracts of rropositions Proposition 1: It is proposed that synthetic polyethers of the "crown ether" variety be studied as possible analogues of the natural ionopho.ric antibiotics. Proposition 2: Possible approaches to the total synthesis of carboxylic acid ionophores in the nigericin group are considered. Specifically, a hemi/acetal approach to the ABCD ring system of monensin, a diels-alder approach to the DE ring system of monensin, and a dihydropyrone approach to the DE ring system of nigericin are proposed. Proposition 3: A synthesis of the alkaloid loline is pro- posed. Proposition 4: A method is proposed for the generation of relatively stable solutions of electrons in solvents other than amines, utilizing crown ethers on a solid phase. Proposition 5: A method is proposed for the isolation of macromolecules responsible for long term memory using competitive hybridization techniques. -79Proposition 1 Synthetic Polyethers as Biologically Active Ionopho res: Analogues of X-537A (Lasalocid) futroduction The recent discovery that the ionophore monensin (_!_) 1 possesses useful antibiotic activity stimulated work on isolation 0 0 1 rnonensin I C0 2 H ~ _rvoH 0/\0 __/\ 7;/ 'I 2 X-537A -80and characterization of other ionophoric Streptomyces metabolites. This work led to the discovery of X-537A @) 2 as a new cardio- tonic and inotropic agent which in many respects is more ideally suited for therapeutic applications to pathological conditions of the heart than the two inotropic drug classe s in current use, the catecholamines and digitalis. 3 It is proposed that relatively easily synthesized polyether ionophores of the "crown ether" variety may be prepared which mimic the activity of X-537A. Furthermore, the synthetic origin of such compounds would allow access to a variety of structural analogues which could be tested and varied within a mechanistic framework allowing preparation of compounds for which desirable proper tie s are maximized. The Mode of Action of Ionophoric Antibiotics The naturally occurring ionophoric antibiotics may be divided into two basic classes, differentiated by their charge at physiological pH. The valinomycin class, including valinomycin, 4 the . . 5 an d the actins, 6 is composed of neutral macrocyclic enn1atins, polypeptides or lactones which form positively charged lipid soluble complexes with d 0 cations . nigericin, 7 monensin (1 ), l, The nigericin class, including 7 X -206, 8 lycosellin, 9 X-537 A (2 ), ~ - 2 and A-23187, lO is composed of monocarboxylic acids which are negatively charged at physiological pH and form neutral complexes with d 0 cations . A large variety of studies with compounds of this type involving ion complexation and transport through synthe tic and natural membranes have been carried out. The ionophores -81have been studied extensively as models for biological ion transport 11 because of their ability to distinguish between cations in a manner reminiscent of biological systems. Members of both classes have also proved useful as tools for the study of various biological processes involving ion transport. 12 13 14 ' ' In no case has it been shown that any biological activity of the compounds is a result of properties other than the ability to complex cations. The central thesis of this proposal is that the biological activity of the ionophores is indeed a result only of their ability to form lipid soluble complexes capable of transporting cations across membrane barriers and not a result of specific interactions with proteins or other structures in the cell. This unusual ability of the ionophores to form lipid soluble complexes with d 0 cations 15 may be readily expl ained by examina- tion of the structures of representative examples. X-ray struc- tural analysis of several complexes have been completed, . . . h e d : 17 many common f eatures may b e d 1stmgu1s 15 16 ' and 1) The complexes are cyclic in general, though the shape of the complex may be cylindrical, saddle shaped (as the seam of a baseball), or irregular. 2) The complexes contain between 5 and 8 oxygen atoms (hydroxyl, carbonyl, or ether) within hydrogen bonding dis- tance of the cation and there are few oxygen atoms on the outside surface of the molecule . 3) The oxygens may be regularly arrayed about the cation, or may form an irregular array. 4) The external surface of the complexes are mainly lipophilic, being -82composed of hydrocarb on groups or chains. not hydrated. 5) The metal cation is 6) The complexes may be positively charged or neutral. Properties of the Neutral "C rown " Polyether s In 1967 the synthesis of some examples of a novel class of compounds termed "crown ethers 11 was repor t ed. 18 Th ese hetero- macrocyclic polyethers possess the ability to complex d 0 cations in a manner similar to that exhibited by the natur a l ionophores. X-ray studies have shown that in a typical 1: 1 complex the rn.e tal is contained in a cavity surrounded by oxygen atoms in an orde re d array with the lipophilic hydrocarbon chain of the "crown " exposed to solvent. Measurements of a large number of physical properties, including the rmodynamic s and kinetics of complex formation 19 •ZO,Zl and the ability to render natural and synthetic membranes . . d icate . permea bl e to cations , 2 2' 2 3' 2 4 m t h at t h ese neutra 1 macrocyeles behave in a manner similar to the valinomycin class of antibiotics. There appears to be no reas on why the synthetic macrocycles should differ f rom the natural ionophores in more complex properties such as effects on respiring mitochondria and antibiotic activity. Howeve r , in an early study on the effect of some neutral polyethers on mitochondria though the macrocycles (~., 23 it was obse r ved that, dibenzo-18-crown-6) rendered mito- chondrial membranes permeable to K + as evidenced by increased A TPase activity, the important ability to uncouple oxidative -83phosphorylation, which is common to all members of the valinomycin class of antibiotics, was not observed . This puzzling result, which seemed to cast doubt upon the utility of synthetic macrocycles as biologically active species, appears now to be explicable in terms of the chemiosmotic theory of oxidative phosphoryla tion. 25 26 • It seems that energy obtained through electron trans - port may be stored by the formation of a pH and electrical potential gradient across the mitochondrial membrane. Valin o- mycin, forming a positively charged complex, moves K+ down the . potential gradient through the membrane, thereby 11 turning off the power" for oxidative phosphorylation by destroying the gradient. In order to do this the complex must move through the lipid barrier sans counter ion. Though the thermodynamic and kinetic parameters of complexation for valinocycin and some of the synthetic macrocycles are simil ar as measured for neutral ion pairs, in a single phase or diffusing through lipid barriers , it is not clear that these parameters adequately test for the ability to carry charge across a lipid bar r ier . that the 11 Indeed, it has been shown 24 crowns 11 are much less effective than valinomycin at reducing the electrical resistance of synthetic membranes, though the "crowns 11 are almost as effective at causing cations to "leak" through blood cell membranes in response to a concentration gradient. This difference in behavior may be rationalized on the basis of the structural differences between the crown complexes, where the M+ is close to the surface of the complex, and valinomycin complexes, where the M+ is surrounded more fully by the -84lipophilic skin. It should be stressed that the controversy over the mechanism of oxidative phosphorylation is not settled. However, the above arguments do serve to rationalize the apparently anomalous behavior of the neutral synthetic polyethers in the mitochondrial system. The Properties of X-537A The compounds X-537A ~) and A-23187 are anomalous mono- carboxylic acid ionophores in that A-23187 is specific for dications while X-537A binds mono- and dications efficiently. Examination of the structures of these molecules indicates that no cyclic con formation may be obtained with a cavity of oxygen atoms large enough to accommodate a centrally located c a tion. X-537A forms dimeric complexes with both monocations dications in the form of a sandwich. 28 and This behavior is alarmingly reminiscent of the behavior of synthetic neutral polyethers with small 11 holes, 11 ~·, the formation of a 2 : 1 sandwich complex of K + and benzo-15-crown-5. 29 As a result of its unique structure X-537A has many interesting properties not found in other ionophores of its class. Using the aromatic portion of the molecule as a fluorescent probe for complexation 3 a series of disa ssociation constants (kdiss> were obtained for X-537A for various cations in ethanol. It was found that X-537A binds K+ more effectively than Ca++. Most interesting, however, was the discovery that X-537A also binds ethanolamine hydrochloride, and othe r ammonium salts, presumably as a 1 : 1 -85complex. No other natural ionophore in either class has been reported to bind ammonium salts, though most of the synthetic polyethers do. Since kdiss in a single phase is not expected to be a reliable indicator for ability to transport cations across a lipid barrier, a simple vertically stacked bulk phase transport system (Fig. 1) was used to investigate rates of transport for various ions. 3 It flat bottom _ _ __..,_ tube - - - 1. 0 ml tricine + 10 mm cation - 1. 0 ml 39% CC1 : 61% c 4 H 10 22 + ionophore 1. 0 ml tricine + 10 mm mag . stir bar------------.. - labeled cation sucr ose + 30% Fig . 1 was shown that Ca++ and Cs+ a re transported fastest by X-537A. Also, X-537A shows a selectivity pattern for catecholamines pre dictable on the basis of steric factors, with mediated transport rates in the order norepinephrin > epinephrin > isopropylepinephrin. As a result of the observed affinity of X-537A for Ca++ its effect upon isolated sarcoplasmic reticulum (SR) vesicles was studied. 3 o, 3 l, 32 It was found that SR which had absorbed Ca++ -86by an ATP linked transport rn.echanism lost this C a + 1 upon addi tion of X-537A to the medium. It appears that this effect is due solely to the ability of X-537A to render SR membranes permeable to Ca++ so that simple passive transport down a concentration gradient occurs. This effect coupled with the observed ability of X-537A to complex catecholamines led investigators to test X - 537A .. on muse 1e tis sue. 3 The results of these studies are perhaps the most exciting of all natural ionophore studies to date . It was found that X-537A initiates contraction of the smooth muscle of the rabbit aorta and increases the str ength and rate of contrac t ion of perfused rabbit heart. Jn in vivo studies on live dogs a dramatic increase in contractile tension (IVST) of 81% was observed, while the heart rate and aortic pressure were increased by much smaller amounts (3% and 10 %/16%, respectivel y) . resistance actually decreased by 23%. The p e riphe r a l No condu c t i ve a nomali e s appeared on EKG or HBE t r aces and no othe r toxic manifestation s were observed. Also, preliminar y studies indicat e that t h e hemo - dynamic responses of dogs in surgically induced ca r diac shock are even more striking. The results indicate that X- 53 7A ma y b e c o me a useful inotropic or cardiotonic agent in human s . Synthetic Analogues of X-537A Jn order to prepare a series of synthetic analogues of X-537A designed to mimic its cardiotonic activity in vivo, it is necessary to determine the structural requirements for such activity. It would be possible to simply begin testing of analogues in vivo and -87in this way develop empirical correlations between cardiotonic activity and structure. However, it seems likely that better analogues would result if as many parameters as possible, including equilibrium association constants, kinetics of complexation-decomplexation, ability to transport ions across bulk phases, and ability to transport ions across natural membrane barriers, were determined and correlated to in vivo cardiotonic activity. In this way the con- tribution to the in vivo activity of various detailed processes may be determined, and eventually it is pas sible that a mechanistic model for the in vivo activity could be developed . This would allow intelligent design of new ionophoric molecules pas sessing desired activity. This process would be analogous to that used in determining the parameters necessary for activity in uncoupling oxidative phosphorylation. Unfortunately, very few correlations may be made with existing data. One is hindered by the diversity of methods us ed in the literature to chara cterize various ionophores and by the la r ge number of different parameters actually measured. The fi r st step in an investigation of this sort is the choice of a parameter to use as a screen for activity. It appears in the present ca se that measurement of the rate of ion transport through a bulk phase in response to a concentration gradient, using a system similar to that of Cram, et al. 27 (Fig. 2 ), provides such a parameter. The experiments are simple in execution and provide a crude model for the activity in vivo. After a number of synthetic analogues are tested and related to the behavior of X-537A by this criterion, -88Q' 13- side -side H 0 + tricine 2 + ion salts - -- H 0 + tricine 2 ~~,., ~ CHC1 3 + ionophore Fig. 2 testing of likely compounds by other methods (~., Ca++ transport in SR vesicles and rate data in single phase s) should lead to analogues active in vivo, and point the way to design of new compounds . The limited number of structure-function correlations already obtained for X-537A derive from the previously mentioned structural data for X-537A complexes and measurements made on some analogues of X-537A already prepared. 33 It has been shown 3 that Br-X-537A, a derivative brominated at C-6, has a partition coefficient in the 1-octanol/water system which is a factor of 2 larger (more soluble in the organic phase) than X-537A itself. Also, the transport rates through a bulk phase (Fig. 1) for Ca++ a nd Cs+ are greater for Br-X537A than for X-537A, though not by a factor of 2. The transport rates for Rb+, Sr++, Ba++ and ethanol ammonium+ are comparable for Br -X- 537A and X-537A. Thus there appears at least some correlation between the presence of lipophilizing -89groups on the molecule and transport rates through a bulk phase. However, the cardiotonic activity of Br-X-53 7A was not reported. Examination of all pertinent data indicates that a synthetic analogue of X-537A should conform to the following structural requirements. The synthetic analogues should be mono-carboxylic acids which possess the ability to form neutral lipid soluble cornplexes with dications in the form of a sandwich. The compounds should also form lipid soluble complexes with monocations and ammonium salts, and the rates of complexation-decomplexation should not be so slow as to render the ionophore incapable of transporting cations across lipid barriers at a reasonable rate. With these criteria in mind a large number of candidates for testing which have already been prepared spring to mind. acidic ionophores of Cram, et al., 27 These are the which have been designed as hosts for the binding of metal cations and ammonium salts in a very specific manner simulating the binding of substrates by enzyme systems. These compounds appear to offer an ideal starting point in the search for X -5 37A analogues . The macrocycles 5-8 represent only a small fraction of the total number of monoacid ionophores prepared to date but serve to exemplify structures available through synthetic procedures already worked out. The number and kind of heteroatoms in the macrocycle, the total number of atoms in the cycle, and the type of connecting and shaping units and appended functional groups have been varied to give nwnerous . 27 34 35 analogous compounds with widely differing properties. ' ' -90~C0 2 H 2, R = H / \0 .§_, R = CH -N 2 J_, R = \_/ 8 CH /O'-../C0 2 CH 3 2 Some of the properties of 2-~ with respect to binding of metal and a:mmonium ions have been reported and will be discussed briefly. Compound 5 has been shown to complex valine, and though no report of the properties of ~ - -M+} complexes ar~ available, thi s molecule would appear to fulfill the above structural requirements" More data is available for .§_, a nd this cycle serves to p oint up a possible problem with compounds of this type. been shown to form CHC1 alkalin earth cations. 3 Compound .§_ has soluble complexes with alkali and How e ver, the sandwich complex {.§_ - - Ba ++ } 2 has such a slow rate of decomplexation that a methanolic solution of (~ ++ } is stable for short periods to treatment with H So . - Ba 4 2 2 If Ca++ a l so exhibits this slow rate behavior in polar solvents, .§_ would not be expected to serve as a good model for X-537A. It is possible, however, that making the cycle of .§_ larger or smaller would increase the rate of decomplexation and provide a better -91analogue. The cycle '!.._ and the hydroxy-acid resulting from selec- tive reduction of the ester function are two more members of the class of ionophores using the binaphthyl connecting and shaping unit. The unique class of ionophores with the carboxyl group placed in the center of a macrocycle is represented by 8. When complexed, the phenyl ring is out of the plane of the oxygen atoms in the macrocycle with the carboxylate sitting above the metal atom. This compound has been shown to complex tert. -butyl ammonium thiocyanate, Ca++, K+, and Na+. Presumabl y the Ca++ complex is a sandwich structure . With information gained by testing compounds of the general type 5-8 it should be possible to design new ionophores with desirable properties. It is hard to predict what sort of structures would be most interesting to try, but some possibilities are : 1) macrocyclic ionophores with more lipophilic hydrocarbon units appended to the cha in, ~·, 2_; 2) macrocycles with different connecting and shaping units, e.g., ~ · Jl:.. Jl• and ]2; open chain de rivatives using connecting and shaping uni ts more closely resembling those in the natural ionophores, e . g . ,~.; 4) and dl X - 537A itself. Though preparation of compounds of the type 2:i and total synthesis of X-537A will require the development of new synthetic methodology and understanding of acyclic conformational analy si s, it is clear that molecules of this type will be available in the near future. - 92- -Io 12 13 14 -93References 1) L. K. Steinrauf, J. Mol. Biol., ......,... 49, 533 (1971 ) • ~· 2) J. W. Westley, et al., J.C. S. Chem. Comm., 71. 3) B. C. Pressman and N. T . deGurman, Ann. New York Acad. Sci., 227, 380 (1974). -"""" 4) M. M. Shemyahin, et al., Tet. Lett., 1963, 1921. ~ 5) P. A. Plattner, et al. , H elv. Chim. Acta, 46, 927 (1963). """" 6) J. Bech, et al., Helv. Chim. Acta, 11• 620 (1962). 7) L . K. Steinrauf, et al. , Biochem. Biophys. Res. Comm ., _il, 1279 (1971). 8) J. F . Blount, J . c. s. Chem. Comm., 1975, 533 . """"'"" 9) N. Otake, J. c. s. Chem. Comm., 1975, 92 . ~ 10) M. o. Chaney, J. Am. Chem. Soc ., 96, 1932 (1974). 11) B. c. Pressman, Fed. P roc., 27, 1283 (1968). ""'-"' "'"'"'" 12) H. A. Lardy, et al., Arch. Biochem. Biophys. , 78, 587 (1958). """" 13) P. J. F. Hende rson and H. A. Lardy , Antimicrob. Agents Chemother., 1969, 18. ~ 14) P. W. Reed and H. A . Lardy, "The Role of Membranes in Metabolic Regulation, 11 (Eds. M. A . Mehlman and R . W. Hanson), Academic Press , N.Y., pp. 111-131 (1972). 15) M. R . Truter, Chemistry in Britain, 7, 203 (1971 ). "' 16) M. Dobler, Biochem. Soc. Trans., 1, 828 (1973). "' 17) W. Simon and W. E. Morf, "Lipid Bilayers and Antibiotics, 11 (Ed. G. Eisenman), Vol. 2, Dekker, N.Y., pp. 329-375 (1973). -9418) C. J. Pedersen and H. K . Frensdorff, Angew . Chem. Internat. Edit., 11, 16 (1972). - """'"' 19) J. J. Christensen, et al., Chem. Rev., 20) B. C. Pressman, etal., Biochem., J1, 351 (1974). .l.Q. 852 (197 1 ). 21) P. B. Chock, Proc. Nat . Acad . Sci., _,..,,,... 69, 1939 (1972 ). 22) B. C. Pressman, Fed. Proc ., -""" 27, 1283 (1968). 23) H. Lardy, Fed. Proc., 27, 1278 (1968) . """" 24) D. C. Tosteson, Fed. Proc ., 27, 1269 (1968 ). /'-" 25) H. A. Lardy and S. M. Ferguson, Ann. Rev. Biochem., ll• 991 (1969). 26) P. Mitchell, ' 1Chemiosmotic Coupling and Energy Tr an s d uc t ion, " (Glynn Research Ltd., Bodmin, 1968). 27) D. J. Cram, 11 Synthetic Host-Guest Chemist r y , 1 1 Chapte r in "Application of Biochemical Systems in Organic Chemist r y , 11 in press. 28) P. G . Schmidt, et al., J . Am. Chem. Soc . , 96, 6189 (1974 ). - """ 29) J. J. Christensen, et al. , Sc ien ce , 4 59 (197 1 ). 30) B. C. Pressman, et ~l. , Bioc h em. Biophys . Res . Comm., 48, '"'-"' 847 (1972). 31) B. C. Pressman and A. H. Caswell, Biochem. Biophys. Res. Comm., 49 , 292 (1972). """" 32) A. Scarpa and G. Ines i , Febs. Let t., 22, 275 (1972). -"-'""'- 33) J. W. Westley, et al., J. Med . Chem., 16, 397 (1973 ). """ 34) D. J. Cram, et al., J . Am . Chem. Soc . , 95, 3023 (1973). ""'"' 35) D. J. Cram, et al., J. Am . Chem. Soc., 97, 1257 (1975). ""'" -95Proposition 2 Total Synthesis of the Carboxylic Acid Ionophores: The Nigericin Group INTRODUCTION Since 1967 the structures of a number of novel mold metabolites possessing interesting biological properties have been determined . Among these are the structurally unique ionophoric antibiotics. This class . . . , 3 sep t amyc1n, . 4 A- 204A , 5 . 1udes monens1n, . l niger1c1n, . . . . 2 gr1sor1x1n inc dianemycin, 6 X-206, 7 alborixin, 8 salinomycin, 9 lysocelli n, 10 X-537A, 11 and A-23187 12 (see chart). Inspection of the structures of these examples indicates that the monocarboxylic acid ionophores may be di vided into four structural groups : 1) those molecules containing the nigericin ABCDE pentacyclic ring system ; 2) those containing the X-206 pentacyclic skeleton; 3) those contain ing the ketonic 5-carbon chain present in X-537A; and 4) the group represented by A-23187, which so far contains but one member. This proposal consists of the outline of a possible strategy for the synthesis of membe rs of group 1, the nigericin type ionophores. THE PENTACYCLIC BACKBONE Inspection of the structures of members of the nigericin group shows that the pentacyclic skeleton may be generalized as in structure 1. Rz R R20µtR~" R211" '" ~ R ,..z ... O side chain R1 = H, OH , or OCH 3 1 -96Ionophoric Antibiotics ,,,, 111 I rnonensin R=OH , nigericin R=H, grisorixin septamycin A-204A ' ''~CO H 2 -97- dianernycene X-206 alborixin n '-() salinomycin j\_.J)H ~-'''',,,'IJ1 1 -98- lysocellin X-537A A-23187 -99- Al though a variety of structures are represented, a number of features common to all members of the group are evident: 1) the stereochemistry along the chain is identical for each member of the group; 2) the A-ring is very similar, differing only in R = H or OH ; 3) the oxygen substi1 tuent at C-3 and the ketal oxygen atom of ring Dare in a 1,3 diaxial orientation. These similarities suggest that a basic unified approach applicable to the synthesis of all these compounds may be devised. How- ever, the structural variety apparent within the group also demands that each molecule be considered as an individual with routes designed for each structural type requiring their own specialized methodology. MONENSIN: THE ABCD RING SYSTEM During the course of the structural determination of monensin some chemical degradative work was carried out. la It was found that monensin was stable in basic solution but upon treatment with acid followed by oxidation with KI0 4-KMn0 4 a crystalline dilactone with the formula c23 H36o6 was formed as part of a mixture. The properties re- ported for thi s dilactone are consistent with structure 2. 2 This molecule appears to offer an ideal initial goal for total synthesis since it contains many of the most difficult synthetic problems present in the molecule while still much less difficult than monensin itself . Also, at least in princ iple , routes may be designed -100- Scheme 1 Route to the Dilactone 2 from an AB Fragment ""'"' 1) H+ EtMgBr R 2) Jones R 6 ........ Anhyd. HCl, ----~ <d)i----- ......9 8 - -1 01 - Scheme 1, Continued 1) KOH 12 13 ,,...,. 15 "'"""- dilactone l -102- which elaborate 2 back to the parent structure. Scheme 1 outlines a· tentative route from an AB fragment l to the dilactone ~by consecutive formation of the C and D rings. Although the proposed route is long and non-convergent, it serves to point out some of the problems associated with synthesis of polyethers of this type via a hemiketal approach. Inspection of models of the AB fragment l indicates that because of the methyl substituent at C-14 the ketone carbonyl should adopt a configuration which by application of Cram's rule should lead to predominate formation of the correct stereochemistry at C-12 upon nucleophil ic attach by ethyl grignard giving the alcohol .1_. If cond ition s can be found which afford a large excess of the correct diastereomer, this technique could prove of general utility in these systems . of alcohol 4 to the ketoacetate §.is straightforward . Conversion Formation of a substituted tetrahydrofuran derivative such as~ from a hydroxy ketone derivative such as §.appears to offer some advantages . The central feature of this route involves reduction of a hemiketal deri va t i ve such as the chloroether 10 to the saturated ether 12 in a stereoselective manner. Since the anomeric effect is operative in 5-membered ri ngs 13 the chlorine atom will prefe r a pseudo-axi al orientation. Thus, in order to minimize an undesirable 1 ,3 interaction the chlorine should prefer to be cis to the smaller subs tituent at C-12. In the present case it is likely that a nearly 50:50 mixture will result , though separation and reequilibration could lead to isolation of a good yield of the desired epimaric chloroether 10. The crucial reduction of 10 must then proceed with retention. It is reported in the literature 14 that Ag+ catalyzed nucleophilic dis- placement on an a-acetoXY chloroether by an amine proceeds presumably -103- with participation via a cation such as .l!. to the product corresponding to displacement with retention. In the present case, displacement by hydride with retention would give the desired acetate g. tions of this general reaction type may be envisaged. Many varia- For example, re- duction of a hemiketal corresponding to chloroether ..!.Q. with the mixed hydride reagent, 15 or displacement of a hemH.etal hydroxyl by CN in the presence of Et 3Al-HCN. A variety of mechanisms, and stereochemical con- sequences, are possible for these transformations. From the acetate .Jl., conversion to the ketone .!l is straightforward and adds the possibility of epimerization to other methods of control of the C-14 stereochemistry. The transformation of the ketone .ll to the alcohol .!i points out another general problem in construction of 5-membered rings by this route. An anion of the type .!Z. is required to place the necessary alcohol and carbonyl functions appropriately . A possible solution to this prob l em is the -17 14 ketene thioacetal anion 14. 16 Conversi on of the alcohol 15 to the desired dilactone ~via the protected lactone l§_ 17 is straightforward. Thus, while a possible route to the CD system of monensin is outlined, several problems and many unanswered questions have become apparent. Although a more convergent approach is desirable, the elaboration of the dilactone 2 to monensin is a reasonable goal. This entails the -104- - Scheme 2 Conversion of 2 Into Monensin lr~R 18 Q .n~ r~ ----- Acc/J'..~~rv--j:-R l9 ACO 21 ........ · 19 1) Elab of A Ring 2) Li~RNH 2 3) H Elab of side chain 0 -105- addition at C-5 of the elements of the E-ring and side chain, followed by formation of the spiroketal DE ring system and elaboration of the side chain . A route designed to accomplish this task is outlined in scheme 2. Although no stereochemical control in the attachment of the two pieces is attained, the route serves to point out some interesting features of the DE system. Inspection of models of the monesin DE ring system indicates that once the stereochemical relationship between the side chain at C-1, the methyl group at C-2, and the hydroxyl at C-3 are estab1ished, formation of the thermodynamically more stable spiro ketal should produce the correct orientation as indicated below . This observation suggests the use isomer of a dihydropyran derivative as precursor to the spiro-ketal 22. Also9 the cis relati onship of the alkyl grou ps at C-1 and C-2 suggests the use of a Diels-Alder reaction for formation of such a dihydropyran. Thus, addition of acetaldehyde enalate to a non-protected lactone lZ.' followed by treatment with acetic anhydride un de r appropriate conditions should allow formation of the diene 19. Condensation with a large excess of the cyclic dienophile 20 would hopefully lead to formation of two diastereomers of the dihydropyran £]_(starting with resolved lactone l?)· The carbonyl group of the dienophile 20 serves to direct the sense of addition in the Diels-Alder . 18 Cleavage of the C-S bonds and acetate -106- grouping of the adduct£!_, followed by acid catalyzed ring closure could afford spiro ketal f£, which offers the possibility of elabora tion to monensin. THE NIGERICIN DE RING SYSTEM During the course of structural work on grisorixin 3 the DE fragment 23 was isolated. Much as the dilactone £represents an ideal 24 initial goal for development of methods for synthes i s of the ABCD rings of monensin, the spiro ketal ~represents such a goal for the DE rings and side chain of: nigericin and grisorixin. Inspection of models indi- cates that, with the exception of the axial methoxyl at C-3, this system represents the most stable conformation at C-3, C-4, C-5, and C-6. Also, since the oxygen at C-5 is axi al , one wou ld expect that reduction of a C-3 ketone grouping with a hydride reagent should result in formation of the proper orientation of the oxygen atom at C-3. In order to test these ideas , a synthesis of the spiro ketal 24 is proposed as outl ined in scheme 3. Preparation of the ketone ~is straightforward. Trapping of the derived enclate with acid chloride'?]_ followed by hydrolysis should afford the unsaturated dione 28. The cyclization of intermediates of - 107- Scheme 3 The Ni gericin DE Rings ~ 1) LDA 2) o::J . 28 ....,_... -108this type to dihydropyrones such as 29 is wel ~ precedented, and proceeds in high yield with simple alkyl side chains. 19 Cyclization to the spiro ketal 30 should also represent a facile process. Since positions C-1 and C-4 are epimerizable and the cyclization is reversible, one would expect the most stable pyrone to result. As indicated in the per- spective drawing, the carbomethoxy group at C-1 is equatorial, as is the methyl group at C-4. The methyl group at C-6 is not epimerizable. However, depending upon the mode of cyclization of the side chain either the a or 8 methyl configuration may result. Although a mixture is expected, it is likely that the correct orientation, in which the methyl group eclipses a C-0 bond instead of a C-C bond 9 will predominate at equilibrium . Hydride reduction by equatorial attack would then afford the desired alcohol 24. -109References 1. a) A. Agtarap, et 21·, J . Am. Chem. Soc., 89, 5737 (1967) . b) w. K. Lutz, et 21· , 2. Helv. Chim. Acta, 54, 1103 (1971) . L. K. Stei nrauf, et .!!_. , Biochem. Bioehys. Res. Commun .,~ ~ , 29 ( 1968). 3. P. Gachon, et .!!_., Chem. Comm., 1970, 1421. 4. T. J. Petcher and H.P. Weber, Chem. Comm., 1974, 697 . 5. N, D. Jones, et.!!._., J. Am. Chem. Soc., ~~' 3399 (1973). 6, E. W. Czerwinski and L. N. Steinrauf, Biochem. Biophys. ~·· 45, 1284 (1971}. Comm.,!~?~· 7. J. F. Blount and J. W. Westley, Chem. 533. 8. M. All€aume, et al., Chem. Comm., 1975, 411. ......,......,......,......, 9. H. Kinashi, et.!!._., Tet.Lett. , 1973, 4955. 10. N. U"take, et 2l_., Chem. Comm., 1975, 92 . 11. J. W. Westley , et al., Chem. Comm., 1970, 71. ......,......,......,...., 12. M. O. Chaney , J . Am. Chem . Soc. , 96, 1932 (1974). 13. J. F. Stoddart, "Stereochemistry of Carbohydrates," Wiley, New York , -- \ 1971, pp. 97 - 102. 14. J. Davell , et 2l_., J . C.S~, 1948 , 967. 15. a) E. L. Eliel, et 2l_., J. Am. Chem. Soc., 84, 2371 (1962). b) G. R. Pettit, et.!}_., J. Org . Chem., 38, 2197 (1973). 16. D. Seebach , et.!!._., Angen . Chem. I.E., 12 , 69 (1973). 17. E. J. Corey, J. Am. Chem. Soc., 95, 18. J. Colonge and G, Descotes, in "1,4-Cycloaddition Reactions", (ed. -- J. Hammer), Academic Press, New York, 1967, pp. 217-253. 19. S. Gelm, et tl·, Bull . Soc. Chim. Fr., 1968, 288. -110- Proposition 3 The Total Synthesis of Loline A total synthesis of the tricyclic alkaloid loline (!) 1 is proposed as outlined in scheme 1. The key step in this route is cyclization of 2 8 the tritosytate ~to give tosylate ~· Condensation wi th methylamine would then afford loline stereoselectively. Examination of the proposed intermediate ~suggests that it may be derived from ribose. The previously prepared chloroacetate l. 2 has been shown to undergo nucleophil i c displacement with retention with amine nucleophiles in the presence of Ag+ . 2 Thus, for synthesis of.§. one requires a carbon nucleophile of the type 4a. 4a possible candidate. Enamine i_ appears a 4 Generation of an electrophilic species such as cation 39 in the presence of! may afford the immonium salt 5 after tautomeriza- -111- Scheme 1 OAc iA.c HCl, EtzO ~O~ 1 ~Ac 3 1rC> MeOH, AgN0 NaOMe 3 ....... OTs Ts Cl ~2 --~ [N\.P_rr ~ \._S)'bMe H H 6 ..... - - - - -e9 ........ - loline 10 ~ -112- 3a tion. Trapping of the immonium salt 5 with sodium methoxide with con- comitant cleavage of the acetate functions gives the protected amine £_. Tosylation and careful removal of the ketal protecting group could afford the desired tritosylate ~, which may cyclize to the tricyclic 2_. An alternative route from the immonium salt 5 is indicated in scheme 2. This route does not require an acid catalyzed hydrolysis in the presence of the sensitive tosyl esters, but utilizes the photolabile m-nitrobenzyloxycarbonyl protecting group for the amine func t ion. 3 -113- Scheme 2 OAc NBOCCl, Pyr N+ H 0+ 3 r;~~ 2- 5 OH 11 """'"" 1) TsCl 2) hcl (NBOC)._.._ __.., {NBOC) KOH iA.c IA.c ~ IA.c ... 8 -114References J. Chem., 22, 2021 (196 9). 1. A. J. Aasen and C.C.J. Culvener, Aust. 2. Lythgoe, 3. A. Patchornick, et tl., J. Am. Chem. Soc., 92, 6333 ( 1970). et~., I.C.S. 1948, 967. - 115- Proposition 4 A method is proposed for the generation of relatively stable solutions of electrons in solvents other than amines, utilizing crown ethers on a solid phase. One of the principal limitations of the dissolving metal red uction is the relatively narrow spectrum of solvents available . There are several advantages to running solvated electron reactions in solvents other than amines. Aside from the advantages derived from use of sol - vents with different solvent properties, there is evidence that some reductions, specifically that of an enone, give different stereochemical results depending upon whether they are run in trimesi tyl borane-ether , or ammonia. 1 Also, in some cases it is desirable to be able to con t rol st arti ng material Li, liq.NH 3, tBuOH Et 2o, Na, tBuOH , trimesitylborane 13% 87% 7% 42% 58% 34% the proton source in a reaction to avoid side reactions resulting from the strongly basic amide anions resulting from reductions in ammonia. Also, since primary amines and arrmonia are quite nucleophilic compared to THF or Et 2o the advantages of these solvents with molecules sensitive to nucleophiles are apparent. Several investigators have been working -116to devise ways of producing relatively concentrated solutions of solvated electrons in solvents other than amines. Recently several advances have been made, for example, use of HMPA 2 or trimesityl borane (c.f. above), both of which involve a radical anion as electron transfer agent. One of the most exciting is the discovery of Dye et~· 3 , that blue solutions of potassium or cesium in THF or diethyl ether may be prepared by addition of a crown ether , specifically dicyclohexyl-18-crown-6 (I) to the reaction mixture. These solutions were prepared by addition of DCC to r1 a: 0 OX) ~~J dicyclohexyl-18-crown-6 (DCC) to THF or Et 2o in the presence of a mirror of potassium or ces ium, and "'Jere stable for seve ral hours at room temperature. The concentrations of metal obtained (1x10- 4M in the presence of 5 x l0- 3M (I)) we re not close to that obtainable in , for example , a sat urated solut ion of sodi um in liquid ammonia at -33.8°C (4.8 M) 4 . However, the res ult seems t o be worthy of further investigation as a possible source of solvated electrons in an aprotic medium. In fact, the paper of Dye is not the first indication in the literature that crown ether type compounds may be valuable as sources of stable solvated electron solutions. In 1959, G. Wilkinson, et ~. 5 reported a study of stabilities of the blue solutions formed when potassium and sodium-potassium alloy were added to a number -117- of acyclic and cyclic ethers . Results were rather qua litative, the ex- tent of dissolution for mos t species being estimated by the intensity of the color produced . In a few cases , i ncluding DME and THF, a mo re quantitative approach was tried , that of measuring the amount of hyd rogen formed upon addition of water. The general concl usio n was that the metal solutions in ethers were meta stable, decolorizing at room temperature at varying rates depending upon the ether. Wilkinson set the upper limit of solubility at l0- 4M. He did , however, report one anomalous case. He observe d t hat "the cyclic tetramer of propylene oxide (I I) gives a deep blue sol ution of appreciable stability". However, though (II) realizing that the chelati ng abilities of II are obviously important, he felt that some of the stability of this so luti on was due to the viscosity of the material, noting that the stability of the filtered blue solutions of other ethers seemed to be higher for more viscous ethers. Today we recognize.!...!_ as a 12-crown-4 derivative, and the stab ility of alk ali metal solutions in this compound are not surprising . It would indeed be i n- teresting to study a solution of.!...!_ in THF with lithium added, since th e lithium complex would be predicted to be the most stable by correlating ring size to ionic diameter . 6 (The most stable lithium-crown ether complex is obtained with DCC. However, di(t-butylperhydrobenzo)-14-crown-4 - 118- i s the most selective for lithium. 7) In a more recent paper relating to the work of Wilkinson, Catterall et ~. 8 , they studied the potassium-THF solution using optical spectra and esr methods. The solutions appear similar to metal-prima ry amine solutions. The concentration of metal, estimated from optical spectra 9 was 6 xlO - 6M. The concentration of unpaired spin was 2-5% in concentrated solutions and up to 25% in dilute solutions . The solutions are closely related to ethylamins - potassium solutions, except that the total concentration of unpaired spin is lower, and the electrons are closer to the potassium nucleus in the monomer species present in solution (larger coup ling constant in the esr.) It is clear that the simplest procedure for running a reduction using crown ethers in THF or Et 2o would be to dissolve the metal in a solution of crown ether and THF, then add the substrate. After the re- action is judged complete, a standa rd workup could be performed. However , further steps would be necessary to separate crown ethe r from product, since crown ethers are soluble in most organic solvents. In fact 3 DCC (one of the most likely candidates for the reaction be cause of its affinity for cations) is soluble even i n petroleum ether, and yet displays appreciable water solubility . 7 Of course a liquid phase column chromatography could be used to get pu re product , but the procedure also has the disadvantage that at least molar quantitie s of crown ether (with respect to electrons transferred) must be used, and reductions involve the transfer of at least two electrons . Thus two moles of crown ether would be necessary to run the reaction for every mole of substrate in the favorable case of reduction of an enone to a ketone. This involves -119- quite large quantities of crown ether. It is possible that the crown ether used in the reaction could be recovered and regenerated by conventional methods. However , this would be a tedious process. A solution to this problem , and one which has many other applications as well, is to attach the crown ether molecules to a polymeric support. In fact, Christensen , et ~. 9 in a review of crown ether chemistry , suggested the use of crown ethers in ion chelating resins, and Pedersen and Frensdorf , 7 in their review, mention that dibenzo-18-crown-6 has in fact been condensed wi th formaldehyde to give resins containing polyether rings, though no mention is made of the prope rties of these resin s. If crown ether rings could be attached to a suitable polymeric support, a simple filtration of a que nche d reduction reaction mixture would be all that is necessa ry to separate product from crown ether. Also , the polymer bound crown ethers could be regen erated by simply washing the polymer with water. One more important advantage gained by use of crown ethe rs on a soli d support could be enhanced solubility of the metal. As previously stated, Dye only obtained l0- 4M solutions of potassium in THF . It is not clear that this is t he maximum concentration of metal obtainable by using free crown ethers, but it is probably true that solutions comparable to those obtainable with ammonia and amines would be very difficult to produce using free crown ethers . Howeve r, with the crown ether attached to a polymer, the sol ubility of the metal may be increased . If only a small percentage of possible polymeric re- active sites were crown ether (i.e., the molecules were isolated f rom one another in space), the properties of a very dilute solution could be obtained in a relatively 11 concentrated 11 one. This is perhaps the most -120fascinating consequence of polymer bound crown ethers. The general principle has already been demonstrated by Potchornick and Kraus. 10 In order to obtain the monoacylation of an ester enolate with an acyl halide without formation of products derived from self-condensation of the ester, or diacylation, the ester was attached to a polymeric support during the reaction. Thus, the properties of a very dilute solution were achieved, in that only monoacylation product resulted after cleavage of the ester. from the support. SYNTHESIS OF THE POLYMER There are numerous cases in which polymers containinq function ality attached covalently to the polymer backbone have been used to carry out reactions on small substrate molecules, atoms or ions. Three of most well known cases are the vulcanization of natural and synthe tic rubbers, the many ion exchange resins currently ava ilab le (both the synthesis of these resins, and their function), and the solid phase peptide synthesis developed initia1ly by R. B. Merrifield . 11 Recently, much work has been done to expand the method , including the monoacylation previously described, a polymer bound Wittig reagent developed at Dow Chemical Co . 12 , and a unidirectional Dieckmann cycl i zation achieved by H. Rapoport. 13 By far the most common polymeric support used for these reactions is a suspension polymerized copolymer of styrene and divinyl benzene (DVB). Suspension polymerization involves free radical polymerization of the monomers in small globules in a vigorously aqitated aqueous solution containing various stabilizers to keep the partially polymerized -121beads from sticking together. Using this technique beads as large as 'l cm and as small as 1 µha ve been obtained. 14 The most common beads for synthetic work have been about 30µ in diameter, containing about 2% DVB. The resulting bead is insoluble in all solvents , but appreciably swelled by most. Its structure is indicated below . Essentially every benzene ring in the entire polymer may be func ti ona l ized by reaction of the beads with chloromethyl methyl ether in the presence of stannic chloride. This chloromethylation may be controlled to yield polymers with a very wide range of functionalization by varying the temperature and time of reaction. For example, the polymer used by Rapoport 13 contained 1 .5% chlorine, and was prepared by r eaction of the beads in ClCH 20CH 3 in the presence of SnC1 4 at 0°C for 8 min. A great variety of functional groups may be attached to the polymer via the chloromethyl groups. For example , reaction of the chloromethyl polymer -122with the sodium salt of a half ester at 150° for 10 hours gives the ester linkage. 13 Al ternatively, the resin may be converted to the acetoxymethyl derivative with KOAc in hot benzyl alcohol, then saponified to give the hydroxymethyl derivative, which may be esterified under mild conditions with dicyclohexylcarbodiimide. Anion exchange resins may be prepared by reaction of the chloromethyl polymer with a tertiary amine, to give the quaternary arrmonium salt. 15 Another method used to functionalize the polymer is reaction of substituted styrenes which yield substituted poly mer upon which vari ous reactions may be carried out . For the preparation of the Wittig resin, 12 p-bromostyrene was suspension copolymerized with styrene and 2% DVB . The resulting p-bromostyrene polymer (about 25% brominated) was treated with .!!_-butyl lithium in benzene, then with chlorodiphenylphosphine to give the poly-p-styryldiphenylphosphine. A few other types of polymers have been used in solid phase peptide synthesis. Inukai, et ~. 16 used a phenol resin prepared by reac- tion of phenol with s-trioxane in bis-( 2- ethoxyethyl)ether containing a catalytic amount of PTSA at 150° in a sealed tube . The resin obtained was amorphous, but was swel l ed by the usual solvents used in peptide synthesis . The first N-protected amino acyl group was attached to the resin via an ester bond with the phenolic hydroxyl group. Several workers have used styrene-DVB popcorn polymers in peptide synthesis . The popcorn polymer is fo rmed in a neat mixture of the monomers in which the polymer is somewhat insoluble and forms as hard nodules which are not swelled by the solvent, but are evidently quite freely permeable . A study of the rates of aminolysis of p-n itro phenyl ester groups on a -123swel lab le s-DVB bead polymer, a popcorn polymer, and in homogeneous solution of p-nitrophenyl benzoate 17 showed that both polymers reacted at the same rate, which was about 1/5 the homogeneous reaction rate. Perhaps the largest departure from the norm is the work of Shemyakin, et~· ; 18 in which peptide synthesis was carried out on a styrene polymer chain which was soluble in the reaction solvents (no cross linking) . After reaction, the polymer-peptide could be precipi- tated with water and washed. There are numerous possible routes to a suitable polymer with attached crown ether molecules. The requirements are that the polymer be insoluble (therefore probably crosslinked) , fusable, and stable to the reaction conditions . The latter requirement, that the polymer be stable to the vigorous conditions of the Birch reduction, puts severe limitations on the possible polymeric supports. Obviously, an ester linkage of crown ether molecule to polymeric support i s unsuitable, but the most limiting factor of all is the requirement that no aromatic rings be present . The reason styrene polymers are so simple to prepare in useful forms is that styrene monomer may be polymerized by free radical processes initiated by simple organic (benzoyl peroxide) or inorganic (potassium persulfate) radical sources . Unfortunate ly, unac- tivated olefins do not undergo this reaction, as the necessary radical intermediate is not stabilized by the aromatic ring . Olefins such as ethylene may be polymerized at high pressures and temperatures or by stepwise mechanisms, but these polymers are long, straight or branched chain molecules, which do not have properties ideal for the present purpose. Thus, for this application, the property of the monomer which -124makes it convenient for polymerization and functio nalizat i on is al so not acceptable in the final product . There are a number of ways around this problem. The most obvious is the attempted reduction of an unsaturated , crosslinked polymer. Syn- thesis of s uch a polymer, containing crown ethers, would be quite simple and strai ghtforward. In fac t, it is possible that under the right conditions a polymer obtained by reaction of H2co with dibenzo-18- crown -6 7 could be useful. This type of polymer would be an alogous to Inuka i s 15 1 phenol, s-trioxane polymer, and would probably have a cross linked structure as shown. If crosslinking could be controlled such that the final polymer were insoluble but fu sable , a potentially useful material would result. It would also be relatively stra ightfo rward to make a bead type polymer by simple s uspension copolymerization of sty rene, DVB, and a crown ether monomer such as I I I. Such a monome r could be prepared by r 1 0 ©(° av a (I II) reaction of 3-4 dihydroxy styrene, with catechol and bis(2-chloromethyl) ether as shown in the diagram. 7 The desired product should make up half the total product. is 45%. The yield of this reaction to make dibenzo-18-crown-6 Thus the total yield of desired product should be about 22% -125- separate mixture A perhaps more elegant and very fascinating way of preparing a suitable monomer would be to prepare the crown ether from a catechol derivative attached to a solid phase, taking advantage of the immobility of the reactant to increase the yield of desired product. For example, a protected 3,4 dihydroxy benzoic acid could be esterified to a hydroxymethylated styrene-DVB bead using dicyclohexylcarbodiimide, followed by deprotection . Reaction with triethylene glycol ditosy ·late in the presence of base, followed by cleavage using the standard procedures developed in peptide synthesis (e . g. , TFA 10 ) would give benzo-18crown-6 with a carboxyl group at the 4 position. Esteri fication , reduc - tion with DIBAH, and Wittig would give the requi red vinyl monomer . ......-:--. ~u~ 6 C 0 f-ol / o)QI o o.~ o4f-c · o ~o.J ~ 0 ~~ .. H ~12 6- The next problem, reduction of the polymer, is not so amenable to simple solution. There are few ways of reducing an aromatic ring. The most common is catalytic hydrogenation . Unfortunately, an inhomo - geneous hydrogenation of an insoluble polymer would not be likely to succeed. However , there i s a report of a Ziegler type nic kel catalyst which converts benzene to cyclohexane under 70 atm of H2 at 150- 210° in essentially quantitative yield.l 9a,b If this catalyst could be made to convert an aromatic polymeric material into the saturated analogue in high yield, the problem would be solved . available. If not, other routes are Although Birch reduction normally does not lead to saturated compounds, the 1-4 dihydro product out of a Birch could possibly be reduced using tris (triphenyl phosphine) rhodium chloride to yield the saturated polymer . There are drawbacks to this pro ce dure , however, since a possible side reaction in the Birch is the reductive cleavage of the ether20 which would res ult in some loss of acti vity of the polymer. Although a polystyrene type backbo ne wou ld be very conveni ent and could possibly lead to useful product, a method of obtaining saturated polymer from mono-olefinic monomers would be very usefu l . There are many methods of polymerizing non-activated ole fi ns such as ethylene. The two most common are high pressure free radical polyme ri zation and use of the Ziegler-Natta catal ysts (R 3Al + TiC14). The convenience and applicability to lab scale operations make the use of comp l ex catalysts the prefer red method for the purposes of this discussion. As an example of the efficiency of the complex catalysts, linear ethylene may be polymerized at 190-210°, with 0.03% 02 at 1500 atm to give highly branched polyethylene (low density). Using Ziegler-Natta catalysts -127( e.g., TiC~ 4 and Et 2A1Cl in xylene) an all linear, high molecular weight polyethylene may be prepared in about 30 min at room temperature. 14 Poly (vinycyclohexane) has been prepared by reaction of vinylcyclohexane with triisobutyl aluminum-TiC1 4 . 21 It should be possible to prepare an insoluble, loosely crosslinked, fusable, saturated polymer by reaction of a mixture of vinylcyclohexane - divinylcyclohexane with mono-olefinic monomers containing crown ethers, using Ziegler catalysts. could be of type_!:£. or'}__. Monomers Evidence indicates that a polyether of type (V) (IV) V would be at least as good as a complexing agent for cation s as rv. 22 The parent compound of '}__has been prepared in 93%yield as shown . The reason for the high yield of'}__ is postulated to be the action of the K+ ion as a template for ring closure. Stability constants indicate that 18-crown-6 comp1exes Na+ and K+ in methanol even better than DCC . The synthesis of monomers of this type should be possible using reactions already well worked out and straightforward. DME, KOtBu -128- References 1. G. Frainkel, et tl·• J. Am. Chem. Soc., 87, 1406 (1965) . 2. 3. - - - - -J. L. Dye, et tl·, J. Am. Chem. Soc., 92(17), 5226 (1970) . 4. Solvated Electron, Adv. in Chem. Ser. , 50, A. C. S., p. 2 (1965) . 5. G. Wilkinson, et tl·, J. Chem. Soc., 1959, 3763 (1959) . 6. H.K . Frensdorff, J. Am. Chem. Soc., 93, 600 (1971). 7. C. J. Pedersen and H.K. Frensdorff, A New. Chem. Int. Ed. , 11(1), S. D. Darling, et al., ibid., 92(3), 696 (1970). 16 (1972). 8. Catterall, et tl·, J. Chem. Phys., 52 1003 (1970). 9. Christensen, Izatt, and Hill, Science, 174 (1971) . 10. A. Patchornik and M. A. Kraus, J. Am. Chem. Soc . , 92(26), 7587 ( 19 70). 11. R. B. Merrifield, Adv. in Enzym., 32, 221 (1969) . 12 . McKinley and Rakshys, Chem . Commun., 1972, 134. 13. H. Rapoport and J . I. Crowley, J. Am. Chem . Soc ., 92(21) , 6362 (1970). 14. A. Ravve, Q!:ganic Chem. of Macromolecules, Marcel Dekker Inc ., New York, N. Y. (1967). 15. F. W. Billmeyer, Jr., Textbook of Polymer Science , I nterscience, New York, N. Y. (1962). 16. Inukae, et tl·• Bull. Chem . Soc. Jap., 41, 182 (1968) . 17. Leetsinger and Jerin a, J . Pol ym . Sci . A-1, 5, 1977 (1967) . 18. Shemyakin, ~tl·• Tet. Lett., 1965, 2323. 19. (a) M. F. Sloan, J. Am. Chem. Soc., 85, 4014 (1963); (b) R. E. Harmon, et tl·, Chem. Rev., 73(1), 49 (1973). -129- 20. H. 0. House, Modern Synthetic Reactions, 2nd ed., Benjamin, Inc., Menlo Park, 201 (1972). 21. Chem. Abstr., 54, 7215 (1960). 22. R. N. Green, Tet. Lett., 1972, 1793. -130- Proposition 5 A method is proposed for the isolation of macromolecules responsible for long term memory using competitive hybridization techniques. In the last two decades a considerable amount of effort has been expended by many \'lorkers trying to understand the mechanisms of memory and learning on a molecular scale. Advances have been made, but very little is known about the details of the biochemistry of memory storage and even less is known about retrieval. The literature has been full of reports of results which later proved irreproducible and workers in the field are in agreement on only a very few points. Some scientists feel that the problem is either not amenable to solution at all, or that technology is in much too primitive a state to even begin background work. These arguments, however, sound very much like arguments against the study of the molecular mechanisms of heredity before the double helix was suggested, and though higher nervous function is perhaps the most complex biological phenomenon of ~11, it seems clear that work in the area today can yield useful and interesting data. ing 11. Before continuing, it wi 11 be useful to define 11 memory 11 and 11 learnIt has been proposed 1 that the term 11 learning 11 not be used in scientific language since different researchers have differing opinions concerning a definition. tions are of value. However, for our purpose the following defini- Memory is the storage of some type of information which may, under certain conditions, be observed as a behavioral ch ange . Evidence (see below) strongly indicates that there are two kinds of memo ry: -131- short term and long term. Short term memory may be detectable only on the order of hours, whereas long term memory is detectable under certain circumstances after very long periods of time (the lifetime of the subject?). Learning will be defined as the aquisition of long term memory. There have been three relatively successful approaches toward understanding the mechanisms of memory. First is the study of inhibi- tion of memory by application of various chemical or physical agents. It has been shown 2 that generalized convulsions induced either by electric shock or chemical means shortly after training produce a memory deficit in animals. Perhaps the most interesting studies in this general area involve the administration of various antimetabolites which block either the synthesis of RNA (actinomycin D) or protein (cycloheximid) in the brain. This work has been revie\·1ed by Barondes. 3 The general results of the inhibitor vwrk are as follows. An animal (e.g., goldfish or rat) may learn a specific response by methods of operant conditioning. The learned response is retained for many days. If an inhibitor (protein 6r RNA) is injected into the brain of a subject during or just after training, the response is retained for up to 6 hours (short term memory), but is then lost. If inhibitors are injected about 1/2 hour after training, there is no memory inhibition. This evidence indicates that there are two memory mechanisms and that long term memory is dependent upon protein synthesis. It must be remembered that these metabolic inhibitors exhibit very complicated activity in organisms besides their primary functions. Even so, it looks as though the conclu- sions based on these experiments are correct. -132- A second general approach towards elucidation of the mechanisms of memory is the transfer effect. This is certainly the most controver- sial of the techniques, and probably one of the most controversial subjects in the life sciences today. The supposition that macromolecule synthesis is necessary for memory is consistent with two basically different mechanisms. One possibility is that the synthetic activity is merely a quantitative increase in the synthesis of molecules and tissues already present in the brain (e.g., new synaptic connections). possibility The other is that the new molecules are qualitatively different from substances already present in the brain and that somehow the structures of the new molecules are related to the learned response. mechanism is necessary to explain transfer of memory. The latter The philosophy behind the transfer experiments is that if the macromolecules responsible for long-term memory are unique, then introduction of these molecules into the brain of a naive subject could cause the subject to 11 remer.:ber 11 training which he never received. The experiments in general are very similar, differing only in detail. At least tvw groups of donors and two groups of recipients are used. One donor- group is given some sort of training (a conditioned response) and the other donor group is handled in very close to the same manner, but not trained (the control group). Soon after the last training period, the brains of the trained and control donors are withdrawn, and either a whole brain homo genate, or some fraction of the brain matter is introduced into the recipient's brain. This is accomplished in one of four ways--either by feeding, or injection intracranially (IC), intravenously (IV), or introparitoneally (P). The tvm recipient groups (trained and control) are then tested for altered -133- behavior. Statistical analysis is used to show whether the trained and control animals behave differently to a certain probability. (In any test of behavior, differences between two groups of animals may be caused by random chance. Psychologists have worked out tests to deter- mine, with any two groups, what the odds are that differences in behavior are random fluctuations. Usually two groups are deemed statistically different if the odds are about l in 100 or less that differences are due to random chance.) Since the first report of this type of transfer experiment in 1962 4 at least 53 laboratories have performed transfer experiments. James A. Dyal 5 has made what he claims to be an impartial survey of these experimenters, and finds that out of 263 experiments reported (in this field negative results are often published; however, Dyal cormiunicated privately with many of the workers in the field to find out if significant negative results were not published, and found this not to be the case) 115 showed no transfer, 133 showed statistically significant transfer, and 15 were equivocal. Considering the tremendous dif- ficulty of the experiment it is not surprising that reproduction of results is very difficult. The surprising fact is that so many labora- tories have reported positive results. It would be very hard indeed to rationalize a theory of memory without taking the transfer phenomenon into account. effect is real. One is almost forced, against his will, to believe the Still, the argument for transfer is of the "overv;helm- ing evidence" type, and no single paradigm has proved to give a robust reproduceable effect in many different laboratories. -134- Research in the transfer field, once the effect is considered real, is _directed towa~ds two goals. Is the transfer specific and what is the chemical agent of the transfer? Much work has gone into eluci- dation of these points and still no general agreement has been reached. (Indeed, many reputable scientists still doubt the reality of the effect.) The most ambitious work has come from the lab of Georges Ungar, at Baylor College of Medicine, in Houston. In fact, Ungar, using the trans- fer effect as a biological assay, claims to have succeeded in isolating the substance, a pentadeca peptide, responsible for transfer of the dark . ra t s. 6 In fact, he has deduced the amino acid seavo1"d ance response in quence of the material, called scotophobin, and a total synthesis using the Merrifield technique has been performed. The experiments are run as follows. Donor rats are trained to fear the dark by conditioning using electrical shock in a dark cage. Rats, when given a choice, will spend more time in a dark part of their enclosure than in a lighted part. However, rats conditioned as above spend much more time in the light. Brains of conditioned and control rats are excized and various fractions injected IP into recipient mice. The recipients are then put in a cage with a dark part and a light part. The amount of time spent in each is recorded automatically. The recipients of trained brain extract spent significantly more time (p < 0.001) in the light part of the cage than the recipients of control brain extract. Synthetic scotophobin is as active as the natural scotophobin in causing fear of the dark in rats. Ungar 1 s experiments seem quite straightforward and conclusive. However, replication of the dark avoidance transfer has been difficult -135- ( though some labs have done it) and it is interesting to note that, while Ungar's article in Nature is 4-1/2 pages long, the critical comments of the referee, which were published as a paper directly following Ungar's paper, are 7-1/2 pages long. It is easy to see that even assuming the transfer phenomenon is real, use of the effect as a tool to elucidate the mechanism of memory is very difficult. This brings us to the third general approach which has been of considerable value in studying memory. This involves direct studies of chemical changes which occur in the brain ~uring learning. Most of the work in this area has been of the following type. Either an increase in the uptake of radioactively labeled RNA precursors is shown to take place during learning, or differences in base-pair ratios of brain RNA are observed between control animals and recently trained· animals. While these experiments elegantly show th~t indeed synthesis of some RNA species is increased during learning, they do not give a clue as to whether the RNA is unique. Clearly, if transfer is real, and whether RNA, as some claim, or protein as Ungar claims, is the transfer agent, ~~A species unique to certain training conditions must be formed in the brain during training. At first thought, direct proof of the existence of such unique RNA's would seem almost impossible, especially when it is remembered that, assuming the scotophobin work is correct, the unique m-RNA species coding for scotophobin would be a mere 45 nucleotides long. However, recent advances in biochemical genetics make proof of the existence of these unique RNA's possible, and I will propose a method of showing their existence and isolating the RNA's in question. -136- The technique that makes this type of detection and fractionation possible is based upon the following simple argument. Any RNA (and therefore any protein) produced in a cell must be coded for in the genome of the cell, at least according to one prevalent theory. If an RNA molecule with a certain nucleotide sequence is produced in the cell, a complementary sequence must be present in the DNA of the cell. Therefore, if one isolates the DNA of a cell and separates it into single strands, then incubates this DNA with purified RNA from that cell, every RNA species will base-pair with its complementary DNA sequences. This process is called hybridization, and the DNA-RNA double strand is referred to as a hybrid. Incubation of this mixture with radioactively labeled RNA from a similar cell should show no incorporation of radioactivity into the hybrid molecules, since all the complementary DNA sequences corresponding to the labeled RNA are complexed with cold RNA. However, if labeled RNA from a cell containing any different RNA species is incubated with the DNA mixture, the new RNA species will bind to their (u no ccupied) complementary sites on the DNA, causing radioactive label to be incorporated into the hybrid. This is, in principle, a very sensitive and specific test for unique RNA molecules (see Fig. 1). In 1965 a method for the quantitative assay of DNA-RNA hybrids was described 7 based on the fact that single stranded DNA will bind and be immobilized by a nitrocellulose filter, whereas RNA will not. One merely passes a solution of single stranded DNA through the filter. The DNA sticks and is, in effect, held rigid on a solid phase. The entire filter is incubated in a solution of labeled RNA which is to be tested, counted in a scintillation counter. The filter is then washed and -137- DfVA LJLJ in cu barf' ,·., <:: <-< ba.fe w;f/, f ii beled RN /1 fron-> 5 1 'nu lu ('e J/ ~~,/"'V;lf-1 ~L>f ;,.,cuh,df' w/f/., l;ibeleJ l p, µA Fig. 1 7 u.(',,., re ~"'" t'>e -138- The hybridization reaction obeys second order kinetics and some of the RNA species may be very dilute. reaction is critical. For ;his reason the time of It is important to incubate the mixture until saturation of the DNA occurs, that is, until every sequence of DNA which is represented by an RNA molecule in the solution has been bound. In 1968 Machlus and Gaito published a series of three papersBa,b,c describing the use of this technique to prove the existence of unique RNA species formed during learning. They followed the procedure of Fig. l, called successive competition hybridization. three groups. Rats ·were divided into One group was given shock avoidance training; the other two were given shocks but could not escape from the shock chamber, and therefore no avoidance training resulted. Unlabeled orotic acid was injected IC into one control group, and labeled orotic acid was injected IC into the other control group and the trained group. After the train- ing the rats were killed, their brains were removed, and DNA extracted and purified. This DNA was put on nitrocellulose (50 µg). RNA was ex- tracted and purified and the competition performed (50 µg of RNA- fo solution was used and incubation was performed on the order--of hours). Results indicated that a sizeable proportion (about 37%) of the brain RNA in shock avoidance trained rats was unique. The above described results were received with enthusiasm by workers in the field. However, in 1971 Kern von Hµngen, from Berkeley, reported that his own data and other reports in the literature indicated that Machlus 1 values were "unrealistic and incorrect 11 • 9 Von Hungen repeated Machlus 1 work and showed that under those conditions the DNA was far from saturated. (In fact, the latest evidence indicates that long -13 9- DNA, bound to a solid phase and hybridized with equal amounts of RNA would probably not reach saturation in a year.) Later in 1971 Gaito t 1 ~L· 10 . he d a re t rac t.ion of th e1r . previous . pu bl 1s war k their findings were possibly due to artifacts. . tat h a dm1. t ting Recently it has been shown 11 that with IC injections of labeled precursors into rat brain an unequal distribution of label is obtained. This unequal distribution is undesirable because in the cells of higher organisms there is a type of RNA synthesized in the cell which turns over very rapidly and never leaves the nucleus.~ Slight changes in precursor pools in the cell could cause different amounts of labeling to occur giving rise to quantitatively false values in a Machlus type experiment. This is probably the reason for the false conclusions reached by Machlus and Gaito. Another general fault of the competi tfve hybridization experiment pointed out by van Jungen 9 is the fact that since the genome in the cells of higher organisms contains a large fraction of repetitive sequences 12 (10% of mouse DNA is present as a segment of about 300 nucleotides repeated about one million times in a single cell), saturation may be difficult if not jmpos_sible. Techniques are now available which should lead to extremely sensitive tests for the presence of unique RNA's using competitive hybridization. First, it \I/as shown by Kottler 11 that IV injection of labeled precursors leads to equal distribution in the brain. Second, and most im- portant, it has become possible to separate the repetitive sequences of the genome from the single copy DNA. 12 This is accomplished by taking advantage of two very powerful tools. First, a special type of calcium phosphate (hydroxyapatite, or HAP) when used as column packing, binds double stranded DNA but not single strands. Second, the rate of -14 0- renaturation of repetitive sequences is much faster than the rate of renaturati on of single copy DNA. In fact, the number C0 t , where t 112 112 is the time to reach 1/2 renaturation and C0 is the initial concentration of DNA, is a constant characteristic of different types of DNA, dependent upon the length of non-repetitive segments. Thus in order to separate repetitive and non-repetitive DNA s one first shears the DNA to 1 about 400 nucleotides in length. allowed to cool to Tm-25°. The DNA is then heated to above Tm and At a certain . time essentially all of the repetitive sequences will be double stranded and all the unique sequences will still be single stranded. HAP chromatography then separates the double stranded, repetitive DNA from the single stranded, unique DNA. This unique DNA could be used in a competitive hybridization experiment similar to Machlus', using pulse-labeled RNA. However, Grouse 13 has reported a much improved procedure for detection of the amount of hybridization which takes place in experiments of this nature. DNA is highly labeled, sheared to a M.W. of about l.7xl0 5 (about 500 nucleotides long) and the unique sequences isolated. RNA is then added in great excess (about 1000-fold) and after satu r ation (about 4 days) HAP chromatography separates any DNA which hybridized (i.e., which contained sequences corresponding to the RNA present) from that which remained single stranded. Both fractions are counted, and the percent DNA hybridized is calculated. Grouse used pulse labelling techniques to label his DNA with H3 • A procedure which is an improveme nt over that of Grouse is to label the DNA with 1125 . Iodine reacts with the double bonds of cytosine. Reaction of 1125 (T = 57 days) with only 1% of the cytosine in a sample can re112 sult in DNA with a specific activity of 2xl0 7CPM/µg. Also, when the -141- . . t ac t . 14 in I 25 does decay l•t l eaves th e cy t os1ne Combining these procedures we have an extremely sensitive method for determining whether a unique RNA is produced during learning. experiment would be as follows. The Rat DNA from the strain chosen for the experiment is purified, sheared, and the unique sequences isolated. This DNA is treated with I 125 to give DNA of very high specific activity. The test rats are divided into two groups, learning (L) and nonlearning (NL). The rats are raised in as close to identical conditions as possible, then the L group is subjected to dark avoidance training . . The NL group is shocked, but not trained. Immediately after training both groups are killed, the brains excised, and RNA removed. Note that no injections of labeled precursors are necessary. The RNANL is then hybridized with the hot DNA using a 100-1000-fold excess of RNA. Aliquots are removed and HAP chromatography is used to determine the percent of double strands. _After saturation is reached (about 4 days) the entire reaction mixture is chromatographed on HAP. As a test, the hybridization procedure may be repeated using RNANL-. No activity should remain on the column after HAP chromatography. If this is true, then the single stranded DNA coming off the column contains no sequences complementary to any RNANL. the RNAL. This DNA is then hybridized to If any activity remains on the column after HAP chromatography this will be due to unique RNA's which were present in the trained rat brain. Note also that we have effectively separated this unique RNA from all other RNA's that were present. This procedure could be performed with many different training tasks, levels of training, etc. in a relatively rapid, efficient manner. -142- If it \'ms found that unique RNA 1 s were indeed formed during various learning experiences, it would be very. interesting to be able to compare these RNA 1 s. Since nucleotide sequencing is extremely tedious I feel that the following alternative might be tried first. The double stranded DNA-RNA hybrids containing the unique RNA may be treated with a DNase specific for single stranded DNA. 15 This would leave a small, double stranded DNA-RNA hybrid which could be melted, and the DNA separated by filtration on nitrocellulose. This DNA vJOuld contain the complementary sequences of the specific learning RNA to which it was originally bound. Other learning RNA s could then be tested for ability to hybridize with this DNA. 1 If the other m!A s 1 are close to the first, considerable hybridization should take place. A semi-quantitative measure of the extent of hybridization would be obtained by study of the melting curves of the various hybrids. -14 3- References l. G. Ungar, Molecular Mechanisms in Memory and Learning, Plenum Press, New York, N. Y. (1970), pp. v-x. 2. E. Fjerdingstad, Chemical Transfer of Learned Information, NorthHolland Publ . Co., Amsterdam (1971), p. XIV. 3. S. H. Barondes, Int. Rev. Neurobiol., 12, 177-201 (1970). 4. J. V. McConnell, J. Neuroesychiat., 3 (Suppl . 1), 42 (1962). 5. J. A. Dyal, Chapt. 13 in Ref. 2. 6. G. Ungar, et _tl., Nature, 238, 198 (1972) 7. D. Gillespie and S. Spiegelman, J. Mol. Biol., 12, 829 (1965). 8. (a) Machlus and Gaito, Psychon. Sci., 10(7), 253 (1968); (b) ibid., 12(3), 111 (1968); (c) Machlus and Gaito, Nature, 222, 573 (1969). --- 9. K. von Hungen, Nature, 229, 114 (1971). 10. Gaito, et~-, Nature-Biol., 234, 90 (1971). 11. P. D. Kottler, et tl· , Physiology and Behavior, 8, 291 (1972). 12. R. J. Britten and D. E. Kohne, Science, 161, 529 (1968). 13. L. Grouse, et tl·, Biochemistry, 11(5) (1972). 14. L. Grouse, private communication. 15. R. Britten,at Caltech, is working on such an enzyme, an acid nucl ea se. Private communication.