Stereoselective synthesis and stereochemistry of seven isomeric spiroacetal structures based on the C17–C28 fragment (CD rings) of spongistatin 1

Brown allylation methodology has been employed to provide seven of the eight possible isomers of the ( ZE )( EZ ) spiroacetal system representing the CD ring fragment (C17–C28) of spongistatin 1. The stereochemistry of the isomers follows from high–field NMR examinations which furnish insights into the trends in the NMR data for these ketal structures


Introduction
In 1993-1994, the isolation of several series of marine-based macrolides was reported.2][3][4][5][6][7] There has been enormous interest in the pharmacology, structures and stereochemistry of these systems which are available in only meagre amounts from their sponge sources.A number of syntheses of key sub-structures, 8 and several total syntheses 9,10,11 have been reported.Synthesis of altohyrtin C (identical with spongistatin 2) and altohyrtin A (spongistatin 1) have defined the relative and absolute stereochemistry of the natural compounds 5 which is shown below in Figure 1, and considered to represent this spongipyran group.The members of these families of metabolites differ with respect to the substituents at C5, C15 and C50.
Sub-structure syntheses have focussed on the retrosynthetically attractive AB and CD spiroacetal assemblies and the EF tetrahydropyran moiety.These partial syntheses are noteworthy as it may be that the most efficient total synthesis will select from the list of syntheses of the key, readily linked, sub-structures.Much of this synthetic endeavour has been summarised. 6,8n this report, we describe synthetic approaches to the CD spiroacetal system (C17-C28) 12 and at the time this work commenced (1995), there was stereochemical uncertainty regarding the relationship between the spongistatins, altohyrtins and cinachyrolide groups. 5Our approach, at the outset, was to control the stereochemistry at C19 and C27 by asymmetric induction methods, and separate and examine the suite of spiroacetals that would result from lack of control at C21, C25 and the created spirocentre, C23.The structures of the resulting spiroacetals would provide insights into the thermodynamics 13 of these systems, whereas the fully assigned NMR spectra would clarify chemical shift trends and nOe's that would be valuable when additional stereocontrol was incorporated.The CD system is stabilised by a single anomeric effect (in contrast with the AB spiroacetal system) and hence there is a question whether the required ZE or EZ system would be favoured, and which isomer(s) within each of these systems.This feature has been considered and regulated in a number of syntheses, 9,12 and equilibration-separation-equilibration cycles have been utilised.Given the free-energy similarities, choice of protection groups on the various oxyfunctions could influence these equilibria.(In the assignment of the ZE, EZ descriptors, 14 the C ring has priority over the D ring, and the first descriptor always refers to the C ring, and the second to the D ring.In addition, if the ring substituents attached to the carbon atoms linked to the tetrahydropyranyl (ring) oxygen, are cis, then Z is used, and if trans, E is applied.)CD spiroacetal.

Results and Discussion
Excision of the CD spiroacetal unit from spongistatin leads to spiroacetal 1 which may be modified to 1a and unravelled to ketohexol 2. The latter was planned to result from dithiane coupling of two ∃ oxy epoxides 3 and 4, as shown below.Arrangement 2 incorporates four stereogenic centres which increases to five on spiroacetalisation (C23).Our approach to open-chain precursor 2 incorporated control at C27 and C19, so that in principle, the final spiroacetal could be a mixture of eight diastereomers, with centers at C25, C23 and C21 uncontrolled.Construction of the chiral epoxides 3 and 4 was achieved with moderate to high asymmetric induction at (C19, C21) and (C25, C27) as outlined below.The route to epoxide 4 utilised the bis-benzyl ether 5 of 2-buten-1,4-diol and ozonolysis provided the benzyloxy aldehyde 6 15 Asymmetric allylation of this aldehyde, employing Brown's B-allybis(isocaranyl)borane (4-d Icr2Ball, 7), 16 provided 8 with an ee exceeding 95%, based on NMR analysis of its Moshers ester.b) Epoxidation (m-CPBA) of 8 provided the desired epoxide 9 as a mixture of diastereomers, which were protected as the TES ethers, 10.These were prone to decomposition and were stored at low temperature (-78° C). ( Scheme The coupling of 1,3-dithiane and the epoxide 17 at low temperature (-40 °C) (Scheme 3) proceeded efficiently (73%) to provide the monalkylated adduct 21 which was protected as the methyl ether.Suitable conditions for the opening of epoxide 10 by deprotonated 22 proved difficult to optimise.In our hands, t butyl lithium in THF containing 10% HMPA (freshly distilled) provided the best outcome and 23 was obtained as a diastereomeric mixture.Treatment of this with Stork's reagent, (bistrifluoroacetoxyiodobenzene) unmasked the carbonyl moiety, removed the p-methoxybenzyl acetal and the triethylsilyl group and effected cyclisation to a mixture of spiroacetals 24.The high resolution mass spectrum exhibited a molecular ion (M=366) and fragmentation ions consistent with the spiroacetal structure 24.For example, ions corresponding to the loss of OCH 3 (m/z 335), CH 3 OH (m/z, 334), -CH 2 CH 2 OH (m/z, 321), H 2 O and CH 3 OH (m/z, 316) and -CH 2 OBn (m/z, 245) are prominent, along with ions for oxygenated pyran fragments (eg.m/z, 237). 17,18cheme 3. Synthesis of spiroacetal 24 via dithiane coupling reactions.This spiroacetal system 24 contains five stereogenic centres, and as two of these (C19 and C27) were installed asymmetrically, there remains the possible formation of eight diastereoisomers.Reverse phase HPLC (acetonitrile/water) successfully separated the diastereomers and sufficient of each was obtained to enable high-field NMR studies, although in some cases, only very small amounts (sub-milligram) were available, and accurate optical rotations could not be determined.Benzene-d6 was employed as solvent to avoid possible acid catalysed equilibration (See Scheme 4) of the purified spiroacetals, a problem encountered previously with simpler systems in choroform. 19Scheme 4. Notional acid catalysed isomerization of spiroacetal 24.

NMR Analysis of the Stereoisomers
Seven isomers of spiroacetal system 24 (see Figure 4) were of sufficient purity to allow analysis of the proton and carbon NMR data utilising a combination of high field 1-D and 2-D NMR and nOe analysis (at 400, 500 and 750 MHz).These isomers are referred to as A-G for convenience and ease of comparisons, and are shown in Figure 4, together with the remaining ZE/EZ isomer, H, which was not formed in sufficient amount for separation and NMR identification.It is of interest that the relative amounts of the seven isomers A-G formed under our spiroacetalisation conditions, are comparable, and no isomer predominates over the others by a factor exceeding 4 or 5. Given the difficulties in calculating equilibrium positions in acetal isomerisations, meaningful reconciliation of the data with computed relative free energies is extremely difficult. 20Comparative analysis of the proton and carbon shifts (See Tables 1 and 2) reveals a number of characteristic resonances for certain centres in the Z,E spiroacetal system that mirror earlier findings on simpler systems.The 1 H and 13 C NMR chemical shifts for isomers A-G are presented in Tables 1 and 2, and a summary of 1 H shifts with coupling patterns and nOe data is given in Table 3. Note: The equatorial and axial protons on the carbon adjacent to the spiro centre in the E ring of the ZE/EZ system 24 are displaced markedly down-and up-field respectively.

Stereochemical Assignments
High field nOe studies were used to assign the configurations of the isomers A-G, with emphasis on the relative configuration of the CD ring and the axial or equatorial orientations of the methoxy and hydroxy groups at C21 and C25 respectively.None of the isomers A-G appeared to favour a conformation in which either of the benzyloxymethyl or 2Ν -hydroxyethyl groups was predominantly axially orientated.
Isomer A: Z,E-(19R,21S, 23S,25S,27R)-24.Examination of the spectra of isomer A confirmed its stereochemical correlation with the natural CD spiroacetal fragment. 21Specifically the CD ring configuration was assigned as Z,E with the C21 methoxy and C25 hydroxy groups concluded to be equatorially and axially disposed respectively, based on observed nOe's.(The Z OH or E nature of ring C (that bears the methoxy group) is always given first.)A strong nOe between the equatorial proton on C24 and the C19 proton together with the downfield shift of the C27 proton relative to the C19 proton (due to the deshielding 1,3-syn interaction with the axial C-O bond of ring C) 19 strongly supported the Z,E configuration for the CD ring system.

Isomer A
The C21 methoxy group was equatorially oriented on the basis of the cross ring nOe's of the C21 proton and the equatorial proton on C24 and lack of any nOe's to the adjacent ring protons at C20 and C22.In constrast, the hydroxyl substituent on C25 was disposed axially because nOe's are detected between the C25 proton and adjacent axial and equatorial protons on C24 and C26, but not detected to either C19 or C27 (ie.no diaxial interaction or cross ring proximity).
These stereochemical assignments for A were consistent with the patterns of proton-proton coupling, interpreted with the aid of 2D COSY and HSQC experiments.The analyses that follow for the isomers B through G provided cohesive and consistent sets of assignments and are based on the types of interpretations outlined above.
Isomer B: Z,E-(19R,21S,23S,25R,27R)-24.Analyses of spectra of isomer B identified the same strong nOe between a proton on C24 and C19 as for Isomer A. When considered with the downfield shift of the proton on C27 relative to the C19 proton, the assignment of the CD ring system as Z,E configured is strongly indicated.
The C21 methoxy was assigned as equatorial, from the cross ring nOe of its geminal proton to the equatorial C24 proton, and absence of nOe's to adjacent protons.Similarly, the C25 hydroxy substituent was also assigned as equatorial, based on observed diaxial and cross ring nOe's of H25 with protons on C27 (not indicated on accompanying structure) and C19, respectively.

Isomer B
Isomer C: Z,E-(19R,21R,23S,25R,27R)-24.The spectra of isomer C exhibited the characteristic nOe between the protons on C24 and C19, and the downfield shift of the proton on C27 relative to the C19 proton, as in isomers A and B. Consequently the CD ring was designated as Z,E.The C21 methoxy group was deemed to be axially oriented from the nOe's of its geminal proton to adjacent axial and equatorial protons on C20 and C22 and absence of nOe's to either C19 or C24 (ie.no diaxial interaction or cross ring proximity).The observed diaxial and cross ring nOe's of the C25 proton to the C27 (not shown) and C19 protons respectively, supported assignment of the C25 hydroxyl substituent as equatorial.

Isomer C
Isomer D: Z,E-(19R,21R,23S,25S,27R)-24.The Z,E nature of the CD ring system of isomer D was also based upon the presence of a strong nOe between the equatorial proton on C24 and the C19 proton, together with the downfield shift of the proton on C27 relative to the C19 proton, as described above for isomers A, B and C. As with Isomer C, the C21 methoxy was concluded to be axial because of the nOe's between its geminal proton (H21) and adjacent axial and equatorial protons on C20 and C22 rather than with either C19 or C24 (ie.no diaxial interaction or cross ring proximity).The C25 Isomer D hydroxyl was assigned as axial from the observed nOe's between the C25 proton and adjacent axial and equatorial protons on C24 and C26.
Isomers E, F and G are located in the alternative E,Z system (for the C and D rings respectively), and again considerations of nOe's, chemical shifts and coupling patterns lead to the stereochemical conclusions for these isomers.These are not discussed in detail, but the important nOe's are shown on the structures below, and full listings of NMR data are presented in the Tables.

Conclusions
Seven of the eight possible isomers of the (Z,E) (E,Z) spiroacetal system representing the CD fragment (C17-C28) of spongistatin 1 incuding the one whose stereochemistry correlates with the natural fragment, have been characterised by high-field NMR measurements.These provide useful sets of comparison data for this system and similar oxygenated spiroacetal sub-structures.Stereocontrol at C19 and C27 was incorporated using Brown's allylation methodology.The Sharpless asymmetric dihydroxylation procedure has been explored as a vehicle for the introduction of further stereocontrol.For example, incorporation of epoxide 20 into the general scheme would control the configuration at C21, and similarly, control at C25 could be enforced.

Experimental Section
General Procedures.All operations involving air-sensitive reagents were performed under an inert atmosphere using syringe and cannula techniques.Glassware was assembled hot, evacuated and purged with nitrogen.THF and ether were distilled under nitrogen from sodium benzophenone ketyl.DCM was distilled from calcium hydride. 1 H and 13 C NMR spectra were recorded at the frequencies stated on a Bruker DMX750, AMX500, AMX400 or AC200F NMR spectrometer.Unless otherwise stated all 1 H NMR spectra were referenced to residual CHCl 3 (δ 7.24) and all 13 C NMR spectra were referenced to the central component of the CDCl 3 triplet at δ 77.0.750 MHz 1 H and 187 MHz 13 C spectra were obtained on sub-milligram samples using NMR microtubes purchase from the Aldrich chemical company.Reverse phase HPLC was performed on a Dynamax 60A C18 column.GC-MS analyses were carried out on a Hewlett Packard HP5890 GC using a 30mx0.25mmBP5 column and a HP5970 mass selective detector.High resolution mass spectra were obtained from a Kratos MS25RFA instrument.Optical rotations were measured on a Perkin-Elmer 241 MC polarimeter using the sodium D line (589nm).

Bis-(2-isocarenyl)-allylborane (7).
To a cooled (-10° C) and stirred solution of boranedimethylsulfide complex (10 mL, 100 mmol) in dry THF (200 mL) under nitrogen, 2-carene (30g, 200 mmol) was added dropwise with the solution maintained at -10°C throughout the addition.The reaction mixture was then stored at 0° C for 24 h, and the product appeared as white needles.The THF was decanted using a cannula and the solid was washed with cold ether (0° C, 3 x 100 mL) with the exclusion of air, and then dried under vacuum (1 mbar, 1 h).This bis-(2-isocarenyl)borane hydride (12.39 g, 43.3 mmol) was then transferred (glove-bag) into a dry, round bottomed flask and ether (17 mL) was added and the mixture stirred at 0° C. Dry methanol (spectrometric grade, 3 ∆ sieves) (3.6 mL) was introduced dropwise over 10 minutes, after which the reaction mixture was stirred for a further 3 h at 0° C. The solvent and excess methanol were then evaporated (1 mmbar, 3 hours).After diluting with dry ether (17 mL), the reaction mixture was cooled to -78° C and allyl magnesium bromide (41.5 mL, 1.0 M, 41.5 mmol) was added dropwise from a pressure equalising dropping funnel.The reaction mixture was stirred for a further 15 minutes at -78° C and then for 1 h at room temperature.The product was used immediately.(R)-5-benzyloxy-4-hydroxypentene (8).Aldehyde 6 (1.62 g, 10.77 mmol) was dissolved in precooled (-78° C) dry ether, and added via cannula to freshly prepared allyl borane 7 cooled to -78° C.After 2 hours at -78° C, the reaction was quenched with 3 N NaOH (5 mL) followed by 10 mL of 30% aq.H 2 O 2 and left in the freezer overnight.After a further 2-3 h of reflux, the organic layer was separated, and the aqueous layer extracted after treatment with brine.The organic layer was dried (MgSO 4 ) then purified by flash chromatography (silica; 20% EtOAc in hexane) to provide the desired product 8 (

1-Benzyloxy-4,5-epoxypentan-2-ol (9).
To the alkene 8 (1.0 g, 5.2 mmol), stirred in DCM (10 mL) at room temperature, was added mCPBA (2.2 g of a 60% mixture, 1.5 equivs), in one portion.The reaction was monitored by TLC (silica, 50% EtOAc in hexane; anisaldehyde) and after 5 h a further 0.5 g of mCPBA was added.Stirring was continued for an hour after which time the reaction was quenched by the addition of saturated NaHCO 3 solution (10 mL).The aqueous phase was separated and extracted with DCM (3 x 20 mL).The combined organic layers were combined, dried (MgSO 4 ) and concentrated under reduced pressure.The residue was purified by flash column chromatography (silica, 50% EtOAc in hexane) to yield the desired epoxide 9, (0.80 g, 74%).GC/MS: (M + , O) 139 (6%), 107 (13), 105 (8), 92 (26), 91 (100), 87 (12), 69 (11). 1   (5 mL).The mixture was concentrated to remove MeOH and then extracted with ether (3 x 10 mL) and EtOAc (1 x 20 mL).The combined organic layers were dried (MgSO 4 ) and concentrated under reduced pressure.Initial purification of the residue by flash column chromatography (silica, 5% MeOH in DCM) was followed by gradient elution on semi-preparative reverse phase HPLC (75% acetonitrile/water through to 95% acetonitrile/water) to provide the diastereomers A -G whose spectra are tabulated and discussed in the text.Representative HPLC times for some of the spiroacetal isomers under these conditions were 16.6mins for isomer C and 11.6 mins for isomer F. Isomers B and E were obtained as a mixture which eluted at13.1mins.The peak area ratios of (B and E):C:F was 2.4:1.2:1.Isomers B and E were separated by further reverse phase HPLC elutions to afford pure B and E in a ratio of 1.6:1.Isomers A, D and G were separated under similar chromatographic conditions and were obtained in relatively similar quantities to B,E,C and F with the exception of isomer A of which only a small amount was obtained.HRMS: C 20 H 30 O 6 requires 366.2042.Measured, 366.2038.

Table 1 1
H Chemical shift assignments

Table 3
Isomer A