Total synthesis of spiroketal containing natural products: kinetic vs. thermodynamic approaches

This review details part of a lecture delivered at the 20 th ICHC in Palermo Italy, July 31-August 5 2005. The published total syntheses of reveromycin A ( 1 ) and the assigned structure for spirofungin B ( 26 ) are described. The approach to 1 utilized a hetero-Diels-Alder reaction to construct the spiroketal fragment in a kinetic stereocontrolled manner. The synthesis of the proposed spirofungin B ( 26 ) spiroketal fragment utilized a thermodynamically controlled cyclization of a protected dihydroxyketone precursor. This allowed for the reassignment of the structure of spirofungin B as epi-spirofungin A ( 48 ).

1. Total synthesis of (-)-Reveromycin A 2. Total synthesis of the proposed Structure for Spirofungin B: a reassignment of the stereochemistry

Total synthesis of (-)-Reveromycin A
The reveromycins A (1) and B (2) are members of a family of compounds isolated from the soil actinomycete Steptomyces sp. 1 Reveromycin A (1) is a potent inhibitor (IC 50 0.7 µgmL -1 ) of the mitogenic activity of epidermal growth factor (EGF) in a mouse keratinocyte.In addition, compound 1 exhibits antiproliferative activity against human tumor cell lines (IC 50 = 1.3-2.0µgmL -1 ) as well as antifungal activity (MIC = 2.0 µgmL -1 , pH 3).8][9] The total synthesis of (-)-reveromycin A (1) has been achieved 10 utilizing a hetero-Diels-Alder (HDA) strategy to construct the challenging spiroketal moiety of this molecule. 6R' It is known that the reveromycin A (1) type 6,6-spiroketal core I can undergo acid catalyzed isomerization to afford the spiroisomer II due to an unfavorable steric interaction involving the axial C19 side chain in I which is alleviated in II, albeit at the cost of an anomeric effect (Figure 1).
Thus, the energy difference between the 6,6-spiroketals I and II is small.In addition, when the C18 alcohol is free, acid induced isomerization provides the more stable reveromycin B (2) 5,6-spiroketal core III exclusively.This energy profile has been observed in the reported thermodynamic approaches (i.e.cyclization of a 'dihydroxyketone' precursor) to the 6,6spiroketal of 1 which provided mixtures of spiroisomers 3,4,5 while degradative 4,11 and synthetic studies 6,7,9b,12

Scheme 1
In light of the above results, we envisioned that a kinetic approach to the reveromycin A spiroketal may circumvent the thermodynamic dilemma cited above.As shown in Scheme 1, the indicated retrosynthetic disconnections of 1 lead to the intermediate 6,6-spiroketal 3.This could be obtained from unsaturated spiroketal 4 by regio-and stereoselective hydroboration from the direction shown followed by alkyne homologation.Key spiroketal 4 could be obtained by an inverse electron demand HDA reaction 13 between the diene 5 6 and the chiral methylene pyran 6. 9 This critical reaction should set the stereochemistry at the spiro center by an axial approach of the carbonyl oxygen in the HDA transition state. 9The stereochemistry at C18 and C19 is then set by the hydroboration/oxidation sequence which circumvents the thermodynamic lability of the reveromycin A type spiroketal I.In addition to being stereoselective and convergent, this route would also avoid the need for a large number of different protecting groups.

Scheme 2
Earlier, we had found that the HDA reaction of isomerizable methylene pyrans with simple hetero dienes can be achieved thermally in the presence of K 2 CO 3 , which effectively reduced the amount of the thermodynamically driven exo to endo isomerization of the dienophile. 6,9In the case of diene 5, using this additive gave low yields of spiroketal due to the base lability of the diene. 6Therefore, Lewis acid promotion of the cycloaddition was surveyed using the diene and model dienophile 7 (Scheme 2).After some experimentation, we found that the HDA between 7 and diene 5 was promoted by Eu(fod) 3 14 in hexane solvent at 0°C to give the spiroketal 8 in excellent yield.Unfortunately, in the case of the substituted methylene pyran 6 required for the synthesis of 1, no HDA reaction was observed in hexane solvent and only slow isomerization occurred to give the endo isomer of 6.Eventually, we found that the cycloaddition proceeded smoothly when a neat mixture 15 of the diene 5 and dienophile 6 was treated with 15 mol% Eu(fod) 3 at 0°C.Although the reaction gave the desired spiroketal 4 as one diastereoisomer in a higher yield than that obtained using K 2 CO 3 , 6 the byproduct 9, resulting from an ene reaction, was also isolated as a mixture of diastereoisomers.This was not observed during the studies using the model pyran 7. The HDA reaction between 5 and 6 was also promoted by ZnCl 2 in THF at 0°C to provide the adduct 4 in a comparable yield.

Scheme 3
With the desired spiroketal in hand, we next tackled the introduction of the C18 and C19 stereocenters and C5-C10 chain extension as shown in Scheme 3. Hydroboration of spiroketal 4 followed by oxidation proceeded in good yield to afford the tertiary alcohol 10 as the only isomer.

Scheme 4
The integrity of the 6,6-spiroketal 10 was evidenced by the characteristic 13 C NMR chemical shift (97.0 ppm) observed for the spiro carbon 9,10 while the stereochemistry was confirmed by its conversion into a known reveromycin A degradation product 4 as described previously. 6Protection of the tertiary alcohol followed by debenzylation afforded alcohol 11 ready for alkyne homologation.Oxidation 16 of 11 and alkyne formation following the Bestmann protocol 17 then gave compound 3. Selective desilylation of the primary TBS group (TBAF, rt) followed by oxidation and two step Wittig extension provided the diene 12 in excellent overall yield for the four steps.Reduction of 12 and subsequent oxidation provided the substrate 13 required for the reagent controlled aldol reaction (Scheme 4).Treatment of aldehyde 13 with the tin enolate derived from oxazolidine-2-thione 14 9,18 afforded the desired syn-propionate 15 which when exposed to NaBH 4 gave the diol 16 as a result of reductive auxiliary cleavage.
Removal of the C18 TBS ether required warming to 50°C to proceed at an appreciable rate and selective primary/secondary alcohol protection yielded bis-TBS ether 17 (Scheme 5).The stage was now set for the somewhat risky late stage introduction of the C18 hemisuccinate.Several methods for the acylation of hindered alcohols failed, 19 affording mostly the isomerized 5,6-spiroketal as found in 2, so we turned to the application of high pressure according to the procedure described by Shimizu and Nakata. 3,20Thus, treatment of a mixture of alcohol 17 and monoprotected succinic acid 18 9b with DCC and DMAP at 0.4 GPa in CH 2 Cl 2 solvent gave the desired ester 19 in very high yield.Interestingly, we found that the reaction proceeded efficiently at much lower pressures than reported previously (1.5 GPa) 20 and required the use of relatively simple high pressure equipment.

Scheme 5
Selective deprotection of the primary alcohol in 19 proceeded well using HF•pyridine complex buffered with pyridine 21 to afford the alcohol 20.Hydrostannylation of the alkyne in 20 proved somewhat difficult under Pd catalyzed conditions 22 owing to the steric hindrance at the axial alkyne functionality and excessive amounts of protodestannylation occurred during the reaction.We found that slow addition of the catalyst (up to 30 mol% over 5 h) was required to give the vinyl stannane 21 in acceptable yields.
The final steps to the target 1 are outlined in Scheme 6. Stille coupling 23 between stannane 21 and vinyl iodide 22 9b afforded the required tetraene 23 along with a small amount of what was tentatively identified as the 22E-isomer.This type of isomerization was observed by us during our studies on the total synthesis of reveromycin B (2). 9 Oxidation of the free primary alcohol and Wittig extension 24 afforded the fully protected reveromycin A precursor 24 which upon exposure to TBAF in DMF gave (-)-reveromycin A (1) in high yield.The synthetic material was purified by reverse phase chromatography (C18 SPE cartridge, 900 mg, 40-60% MeOH/H 2 O) and was identical (NMR, IR, UV, HRMS) to the natural product in all respects {[α]  In conclusion, we have completed the total synthesis of (-)-reveromycin A (1) which utilized a HDA reaction hydroboration sequence to construct the spiroketal core in a convergent and stereoselective manner.

Total synthesis of the proposed structure for spirofungin B: a reassignment of the stereochemistry
The spirofungins A (25) and B (26) were isolated from a Streptomyces strain Tü 4113 collected in the Otway National Park, Australia. 25Compounds 25 and 26 possess a similar 6,6-spiroketal core to reveromycin A (1) 1 and have the same C1-C9 triene acid and C20-24 diene acid segments however there is no succinate half ester at C18 and the C18 substituent is a methyl group.The spirofungins show high inhibition activity against yeast and moderate activity against some fungi with a MIC of 15 µgmL -1 against Candida albicans.

ARKAT
Compounds ( 25) and ( 26) were isolated as a ~4:1 mixture respectively.Although they possessed slightly different retention times on reverse phase HPLC, they were not separated and the structural elucidation was conducted on this mixture.The structure of the major isomer spirofungin A (25) was assigned using various NMR techniques with the minor component spirofungin B ( 26) identified as a diastereoisomer of 25, epimeric at both C18 and C19.The absolute configurations of 25 and 26 are suggested by analogy to the reveromycins.
It was proposed 25 that the spiroketal core of 25 possessed the conformation A as depicted in Figure 2, with the oxygens (maximum anomeric stabilization) and the C19 diene acid sidechain in axial orientations while the C18 methyl group is equatorial.In compound 26, the C19 sidechain is now equatorial while the C18 methyl group is oriented axial as shown in proposed conformer B. This was supported by the vicinal coupling constant measured between H19 and H18, the C18 change in chemical shift, and the ROE observed between H19 and H11.

Figure 2
The natural occurrence of spirofungin B ( 26) is curious since it possesses a different absolute configuration at C19 to that from both spirofungin A (25) and the related reveromycins.Several groups have reported approaches 26,27 to the spiroketal systems of spirofungins A (25).Herein, we detail our total synthesis of the structure 26 proposed for the minor component spirofungin B which lead to the reassignment of the stereochemistry of this natural product.

Scheme 7
A brief retrosynthetic analysis of 26 is shown in Scheme 7. The key bond disconnections are similar to those utilized for our total synthesis of reveromycin B (2). 9 The C2-C3 and C8-C9 bonds could be formed via stabilized Wittig reactions while the the C4-C5 bond might be formed by a reagent controlled asymmetric syn-aldol reaction. 18Finally, the C21-C22 bond could be installed using a Stille cross coupling reaction. 23This leads to the key spiroketal intermediate 27 which could be constructed from a coupling 29 between the anion derived from alkyne 28 and Weinreb amide 29 followed by alkyne saturation and acid induced deprotection and spiroketal formation. 13O OTES

Scheme 8
The route to the proposed 25 structure for spirofungin B began with the known optically pure alcohol 30 9

Scheme 9
Silylation of 30 gave ether 31 which was smoothly converted into Weinreb amide 29 by oxidative alkene cleavage, Wittig extension and hydrogenation.Coupling fragment 28 was synthesized from known alcohol 32 30 by protection, alkene cleavage and Corey-Fuchs 31 alkyne formation.Treatment of the alkyne 32 with n-BuLi followed by addition of the amide 29 and subsequent complete saturation using H 2 and Lindlar catalyst gave the corresponding ketone 33 in excellent yield without any benzyl group removal.Compound 33 was subjected to acid deprotection and spiroketalization by treatment with PPTS in MeOH to give 27 as the only spiroisomer.Some of the primary TBS group was cleaved in this sequence but this was easily reinstalled by subjection of the crude product to resilylation.With the spiroketal fragment 27 in hand we next appended the C3-C10 sidechain as outlined in Scheme 9. ARKAT Debenzylation of 27 followed by oxidation 16 and homologation using the Bestmann protocol 17 gave alkyne 34.TBAF induced desilylation then afforded the alcohol 35 which fortunately was crystalline and an X-ray crystal structure 32 was obtained for this intermediate confirming the stereochemistry (Figure 3).

Scheme 10
Oxidation of alcohol 35 and sequential Wittig extensions gave the ester 36 in good yield.Adjustment of the oxidation level in 36 gave aldehyde 37 which underwent an asymmetric aldol reaction with the tin enolate derived from oxazolidine-2-thione 14. 9 This sequence of reactions was high yielding and easily conducted on large scale.Auxiliary removal by treatment of 38 with NaBH 4 in wet THF gave diol 39.
The final steps to compound 26 are outlined in Scheme 10.Conversion of alkyne 39 into vinyl stannane 40 using Pd catalysis 22 proceeded in good yield.The stannane was then cross coupled 23 with the known iodide 22 9 to cleanly provide ester 41 with the C20-C24 sidechain fully installed.
Protective group manipulation gave 42 which upon oxidation and Wittig reaction provided the protected precursor 43.Exposure of 43 to excess TBAF in DMF caused global deprotection to give the nominal structure for spirofungin B (26) which was purified by reverse phase HPLC.Unfortunately, the physical data obtained for synthetic 26 did not compare with that reported for spirofungin B. 25 Furthermore, the 1 H NMR spectrum of an authentic spirofungin mixture measured on the same instrument did not match our synthesized material.In particular, the chemical shift for H19 in the natural product (4.76 ppm) was vastly different to that found for compound 26 (4.29 ppm).In addition to other shift differences in the 1 H NMR spectrum, the ARKAT chemical shift for C15 in the 13 C NMR spectrum of 26 was also measurably different to the literature value (Natural 25   An investigation of the literature revealed a possible alternative for the structure of spirofungin B. In their synthesis of the spiroketal fragment of spirofungin A, Shimizu and coworkers 27 obtained a mixture of spiroketals 44 and 45 (ratio 1.53:1) isomeric at the spiro centre via a thermodynamic approach similar to that utilized above (Figure 4).This clearly demonstrates that isomerization of the spirofungin A spiroketal is possible under acidic conditions (due to the axial orientation of the C19 substituent) to afford a spiroketal with one less anomeric stabilization. 33The chemical shift for the H19 in the preferred isomer is upfield from the corresponding shift in the minor isomer where the axial H19 is deshielded by the axial oxygen 34 and the opposite change in chemical shift is evident for H11 (Figure 4).Comparable chemical shift differences are also observed for H11 and H19 in spirofungins A and B themselves (Figure 5).Another example can be found in the synthesis of the spiroketal fragment of reveromycin A (1). Theodorakis and co-workers 7 also obtained a mixture of spiroketal isomers 46 and 47 in a ratio of 1.5:1 respectively by acid induced spiroketalization.The chemical shift differences (Figure 4) for the H11 and H19 protons for each isomer are similar to those cited above.In addition, the C15 chemical shifts for each isomer differ in a similar manner to that for spirofungin A (95.7ppm) verses B (97.7ppm).This is consistent with the trend of a slight downfield shift of the spiro carbon in going from ax-ax to ax-eq oxygen orientations. 35An NOE was also observed between H19 and H11 which was also seen in spirofungin B (Figure 2).

Figure 5
The results cited above strongly suggest that the structure for spirofungin B is in fact 15epi-spirofungin A (48) (Figure 5).The absolute configurations at C18 (S) and C19 (R) in spirofungin B are identical to that for spirofungin A (26) while the spiro carbon is epimeric (15R).Spirofungin A (25) has the conformation A while pirofungin B has the conformation B' as shown in Figure 5 which is consistent with the NMR data presented.It is not unreasonable to suggest that biosynthetically, spirofungin A (26) could give rise to 48 by simple spiroketal isomerization in analogy to the spiroisomers of the pectenotoxins. 36 Figure 1 Scheme 6
have demonstrated the strong preference for the reveromycin B type 5,6-spiroketal III.
spirofungin B: 97.7 ppm.Synthetic 26: 96.2 ppm.)These differences lead to the obvious conclusion that the structure proposed for spirofungin B (26) is incorrect.