Total Synthesis of Natural Products

Carbene Complexes

Introduction

I. Benzannulation Reaction

II. Cyclohexadienone Annulation

III. Tautomer Arrested Annulation

IV. Aldol Reaction

V. Diels-Alder Reaction

VI. Cyclobutanone Formation

VII. Biaryl Synthesis

VIII. Macrocycles 

Asymmetric Catalysis

I.. Ligand Design and Synthesis

II. Asymmetric Diels-Alder Reaction

III.. Imino Aldol Reaction

IV. Asymmetric Aziridination

Studies on the Synthesis of (+)-Olivin.    Our approach to the total synthesis of (+)-olivin is outlined in Scheme XXVI and features the benzannulation of a Fischer carbene complex and an alkyne.  Olivin is the aglycone of olivomycin which is a member of the aureolic acid family of antitumor antibiotics.  Evidence suggests that a magnesium complex of these compounds binds in the minor grove of DNA in GC rich regions.  It has been shown that another member of the family, mithramycin, inhibits transcription of the c-myc protooncogene by binding to a GC rich promoter region and blocking gene transcription.   We have previously published the synthesis of the alkyne 197 in optically pure form which is to be the key substrate in the reaction with a 2,4-dioxygenated carbene complex of the type 196.[1]  The plan involves trapping the phenolic function formed in the benzannulation with triflic anhydride to give the naphthyl triflate 195.  Hydroboration and oxidation should give the carboxylic acid 194.  Friedel-Crafts cyclization and then reduction of the triflate should give the intermediate 193 which contains the intact carbon-skeleton of (+)-olivin.  Completion of the synthesis will involve the introduction of the hydroxyl group via the Rubottom reaction and deprotection.

It is well known that the benzannulation of Fischer carbene complexes with alkynes is sensitive to the electronic nature of the carbene complex with the outcome that increased electron density on the complex will give rise to increased amounts of side products, usually indenes and indanones, that result from the failure of a carbon monoxide ligand to become incorporated into the product (Scheme III).  As part of the effort in the total synthesis of (+)-olivin, we have recently published our results from the detailed examination of the effect of para and ortho substituents on the arene ring of an aryl carbene complex.[2]  As can be seen from the para-substituted complexes in Scheme XXVII, higher yields of the benzannulated product 199 are seen with more electron withdrawing substituents.  In this case, the phenol product was not isolated and instead an oxidative workup was employed to give a quinone product.  A para-methoxy group results in a drop in the yield of the quinone and the appearance of a 40 % yield of the five-membered ring annulated product 200.  Under the same conditions, the unsubstituted phenyl complex gives the quinone 199 in 78 % yield as the only detectable product that was mobile on silica gel.  The para- trifluoromethyl substituted complex gives a 95 % yield of the quinone.  Interestingly, no electronic effects were seen for groups in the ortho-position.  The ortho methoxy and ortho trifluoromethyl groups give essentially the same yield of quinone product.  Presumably, a substituent in the ortho position causes the aryl group to rotate out of conjugation with the carbene complex during the reaction and thus the electronic nature of the aryl group does not impact on the reaction.  This would lead to the expectation that a 2,4-dioxygenated aryl complex would give approximately the same yield as the ortho complex and that the electron donating nature of the para oxygen would not be felt.  As can be seen from the reactions of complex 198g and 198h, the actual result is somewhere in between.  These studies have resulted in the optimization of this reaction to the bis-benzyl complex 202 with alkyne 203 which gives after benzannulation and triflation, the naphthyl triflate 204 in 49 % yield for the two steps.  Studies on the Friedel-Crafts cyclization and triflate reduction are now in progress.

Studies Towards the Synthesis of Phomactin D.   A synthesis of the natural product Phomactin D 205 is being pursued in our group where the key step will be an asymmetric cyclohexadienone annulation of Fischer carbene complexes.  Phomactin D was identified in a screen for platelet activating factor (PAF) antagonist activity that Schering Plough and Sankyo have independently made over the last decade on the marine fungus Phoma sp..  Both groups have discovered active compounds which possess the same rare cleomane carbon skeleton.  The most active compound is phomactin D 205 which has an IC50 for platelet aggregation of 0.80 uM.  The only total synthesis of Phomactin D reported to date was accomplished in 36 steps from L-ascorbic acid.[3]

The key step in the synthesis of phomactin D is the cyclohexadienone annulation with a 1,4-asymmetric induction.  The original study of this process for 1,4-asymmetric induction in the construction of a cyclohexadienone were performed on the simple complex 60 and the chiral propargyl ether 40 as was discussed in Scheme VI.[4]  In this study it was found that the reaction was both stereoselective and stereospecific.   Our retrosynthesis of phomactin D is shown in Scheme XXVIII and involves the ring closing methathesis of intermediate 210 to give the [9.3.1]-bicylopentadecanone 209.  The stereoselective synthesis of the cyclohexadienone 210 was achieved with a 94 : 6 selectivity from the reaction of the carbene complex 211 with the chiral propargyl ether 212.  This selectivity could be improved to 98 : 2 if the temperature was reduced to 55 ^oC.  The chiral center of the propargyl ether does not appear in the natural product since oxidation to a ketone will need to occur at some point in the synthesis.  Nonetheless, the chiral propargyl ether chiral center will be used, via the cyclohexadienone annulation, to set the correct absolute configuration of the chiral center in the six-membered ring of the natural product.  Studies directed to the completion of the synthesis of phomactin D are now underway.     

Synthesis of Members of the Colchicine Family.  (-)-Colchicine 213 (Scheme XXIX), the major alkaloid from Colchicum autumnale, is one of the oldest known natural products. It binds to the cytoskeletal protein tubulin, disrupting the microtubule-dependent functions in the cell and thereby suppressing the cell division process.  (-)-Allocolchicine 222, is also a natural product and is an isomer of colchicine and has similar biological activity to colchicine.  Allocolchicine has the tropone ring replaced by an arene ring.  Only natural (-)-7S-colchicine 213, which adopts an aR biaryl configuration, but not its 7R enantiomer, binds effectively to tubulin.   We are presently pursuing a total synthesis of (-)-colchicine 213, and several of its analogs by a method which relies on the reaction of a carbene complex of the type 130 in Scheme XIII where there is a central to axial chirality transfer during the key benzannulation step.[5]  

We have also been interested in the synthesis of other natural products in the colchicine family and we have recently completed the first total synthesis of (-)-allocolchicine 222 which is outlined in Scheme XXIX.[6]  The key step is the Diels-Alder reaction of the diene 217 with methyl propriolate.  This reaction was found to go at 110 ^oC to give, after aromatization with DDQ, a single regioisomer of the allocolchicinoid 218.  The TBS protected alcohol function in 217 will later become the nitrogen substituent in allocolchicine and is also crucial to the regioselectivity of the Diels-Alder reaction.  In the absence of this substituent, the Diels-Alder reaction proceeds to give a nearly equal mixture of cycloadducts as indicated in the formation of 215.  Asymmetry is introduced during the reduction of the ketone 219.   This is accomplished by the recent method of Singram and coworkers who found that ketones could be reduced with high inductions with lithium borohydride in the presence of the chiral Lewis acid 216.[7]  Applied to 219, this method gives the alcohol 220 in 91 % ee.  The conversion of 220 to (-)-allocolchicine occurred with no substantial loss of optical purity in three steps.  A final crystallization gave the optically pure natural product.

Studies on the Synthesis of Fostriecin.   Fostriecin (CI-920) 223, is an interesting phosphate ester which has antitumor activity and has been brought to phase I clinical trials.  It has recently been shown by Roberge in British Columbia that it is an inhibitor of protein phosphatases 1 and 2A, a property not previously described for an antitumor drug.[8]  The antitumor activity has been shown to be the result of an interaction with topoisomerase II, but its protein phosphatase activity is involved in its ability to inhibit the mitotic entry checkpoint.  This means it induces cells to enter mitosis.  Thus it is suspected that it is a tumor promoter and this is apparently being checked by a number of research groups.  Fostriecin is somewhat unstable and when used in the clinic must be stored frozen in buffer.  This instability is presumably due to the trienediol portion of the molecule.  In addition to the total synthesis of fostriecin, our group is involved in the synthesis of analogs that would hopefully be more stable and yet still retain activity.  

Our synthetic strategy for the synthesis of fostriecin is shown in Scheme XXX and involves the introduction of the sensitive trans, cis, cis-triene portion along with introduction of the phosphate on the C-9 second alcohol at the end of the synthesis.  A second key disconnection is the Horner-Wadsworth-Emmons reaction that couples the lactone unit 227 with the central fragment 228 which in turn is to be derived from the b-hydroxy aldehyde 229.  The strategy outlined in Scheme XXX is a result of the ability of chiral imidazolidinone carbene complexes to stereoselectively provide asymmetric aldol reactions with 2-alkynals.

 In previous work we have shown that the chiral imidazolidinone carbene complex 86 is an excellent chiral acetate enolate equivalent (See section IV on aldol reactions).[9]  A minimum of 20 : 1 selectivity can be achieved with aryl aldehydes, branched aliphatic aldehydes and unbranched aliphatic aldehydes.  These carbene complexes can be prepared in good yields and in a single step.  The chemical yields of the aldol reactions for simple aldehydes are in the 80 - 90 % range, which includes removal of the metal in the workup by stirring the crude reaction mixture for a few minutes with aqueous cerium (IV).

  Given the small size of an acetylene group, it was quite surprising to find that the aldol reactions of carbene complex 86 with alkynals occur with very good selectivities as indicated by the data in Scheme XXX.  Even more surprising was the fact that they occur with the opposite selectivity that is observed for simple alkyl aldehydes.  This selectivity can be reversed by employing dicobalt complexed alkynals which give the aldol adducts 231.  The stereoselectivity of these reactions was confirmed by an X-ray structure of 231 (R = TBS) in collaboration with Arnold Rheingold.  The cobalt complexed alkynals thus give the "normal" selectivity.  While the source of this selectivity reversal for alkynals is unknown at this time, our current efforts are directed to the completion of the synthesis of fostriecin. 

Studies on the Synthesis of Taxol.  Taxol is a clinically important agent for the treatment of ovarian and other cancers and is the number one selling cancer drug in the world.  Our preliminary studies reveal that the reaction of Fischer carbene complexes with enynes has the potential to provide for rapid access to the taxol skeleton.  As described in Section VI on cyclobutanone formation, the reaction of Fischer carbene complexes with enynes can be controlled by the substitution pattern on the alkene to give either bicyclo-[3.2.0]-heptanones or bicyclo-[3.1.1]-heptanones.[10]   Subsequently, we have discovered that the reaction to give the latter can be utilized to provide a fundamentally new strategy for the synthesis of taxol and related molecules and this is indicated in Scheme XXXI.

We have published our initial studies on the viability of the reaction of Fischer carbene complexes with enynes as a key tactic in a strategy for the synthesis of taxol and taxol analogs.[11]  The first generation strategy involved the access of the tricyclic intermediate 235 from the reaction of the enyne 236 and the carbene complex 237.  This strategy was based on the reaction of complex 76 with enyne 112 which gave the bicyclo [3.1.1] heptanone 114 in 69 % overall yield (See Section VI).  The strategy was adapted to include the extra alkene in enyne 236 which, after reaction with complex 237 and epoxidation, would give intermediate 234.  The plan was to construct the eight membered ring and introduce the bridgehead double-bond by a Grob fragmentation of the epoxide.  This was not to be as it was found that the reaction of the enyne 236 with the complex 76 did not give a bicyclic species, but rather the highly unsaturated aldehyde 239.  A mechanistic interpretation of this outcome suggested a second generation strategy which involved the reaction of the exocyclic methylene substituted enyne 242 with complex 237 to give the bicyclic species 241.   This would then be followed by the isomerization of the exo-cyclic double-bond in 241 to give the intermediate 235 in the original strategy.  The possibility for this transformation is suggested by the isomerization of the bicycloheptanone 240a to 240b in quantitative yield.[12]  The reaction of complex 76 with enyne 242 was performed as a model study for the second-generation strategy.  Although this reaction gave high yield, it unfortunately gave a mixture of the desired product 243 along with the cyclobutenone product 244.  It was found that the ratio of these products could be controlled to some extent by electronic tuning of the carbene complex.[13]  Electron poor phenoxy complexes of the type 245 lead to increased proportions of the desired product 246 as indicated by the data in Scheme 1.  Further elaboration of the second-generation strategy for the synthesis of taxol is under investigation.

  Synthesis of Chloroamphenicol.       One of the oldest anti-bacterial agents is chloroamphenicol which was first isolated from Streptomyces Venezuelae in 1947.  This antibiotic is obtained commercially by chemical synthesis and is biological active only as its 2R,3R enantiomer.  It is used clinically as a broad spectrum antibiotic and is particularly useful for the treatment of salmonella, typhi, rickettsia and meningeal infections.  As a result of its link to bone marrow depression, its use is reserved for serious infections with organisms which have been demonstrated to be resistant to all other appropriate anti-microbial agents.  A number of chemical syntheses of racemic chloramphenicol have been reported as well as a few in the last decade that are selective for the formation of (-)-chloramphenicol.  We have recently published an asymmetric synthesis of optically pure (-)-chloramphenicol that is the shortest synthesis of the enantiomerically pure antibiotic to date.[14]

The synthesis of (-)-chlormaphenicol is presented in Scheme XXXII and features an asymmetric catalytic aziridination reaction that has recently been developed in our laboratories (See Section IV).[15]  The synthesis begins with commercially available para-nitrobenzaldehyde and its conversion to its corresponding benzydyl imine 249 which is obtained in 85 % yield after crystallization from ethanol.  The aziridination is carried out with 1 mole percent of the catalyst prepared from the VAPOL ligand in toluene at 0 ^oC to give the aziridine 250 in 80 % yield, 96 % ee and with a >100:1 cis/trans selectivity.  The enantiomeric purity of this aziridine could be improved to 99 % ee with a single crystallization from hexane / methylene chloride (84 % yield, 1st crop).  Treatment of the optically pure aziridine 250 with 10 equivalents of dichloroacetic acid in refluxing 1,2-dichloroethane for one hour gave the hydroxy acetamide 251 in 80 % yield as a single diastereomer.  Completion of the synthesis was accomplished by reduction of the ethyl ester with sodium borohydride which gave (-)-chloroamphenicol in 74 % yield and in greater than 99 % enantiomeric excess.  The synthesis of enantiomerically pure (-)-chloroamphenicol was thus achieved in four steps from commercially available starting materials in 38 % overall yield.

The conversion of 250 to 251 actually involves three steps in one as outlined in Scheme XXXIII.  In the first step, dichloroacetic acid protonates the aziridine nitrogen which is then opened by its counter ion with inversion to give the bis-ester 254.  A second molecule of dichloroacetic acid then protonates the amine in 254 which leads to cleavage of the benzhydryl group and the formation of the free amine 256.  The final step is an intramolecular acyl transfer which is thermodynamically driven to give the amide 251.