Synthesis of antimicrofilament marine macrolides: Synthesis and configurational assignment of a C₅–C₁₆ degradation fragment of reidispongiolide A

Abstract

Reidispongiolide A is a representative member of the sphinxolide/reidispongiolide group of cytotoxic 26-membered macrolides of marine origin. By interacting with actin in the cell cytoskeleton, the reidispongiolides and sphinxolides are potent microfilament destabilizing agents that represent a promising mechanism of action for developing novel anticancer drugs. An aldol-based synthesis of a library of diastereomers of C₈–C₁₆ and C₅–C₁₆ fragments and detailed NMR comparison with a reported degradation fragment enabled a configurational assignment for a major part of the reidispongiolide macrocyclic core, thus setting a solid foundation for ongoing synthetic efforts.

Marine macrolides provide a fascinating range of structural and functional diversity, which may find application inter alia as specific molecular probes for the investigation of the cytoskeleton and cell cycle events. Unique to eukaryotic cells, the actin cytoskeleton plays a critical role in the determination of cell shape, and in a variety of cellular processes, including cell motility, division, adhesion, and intracellular transportation (1). In comparison with microtubule-interacting agents (2), such as the taxanes and vinca alkaloids, the development of anticancer drugs based on selective binding to actin and disruption of microfilament assembly has attracted relatively little attention. However, recent insight into the role played by microfilaments in cell division and metastasis has highlighted the potential of certain structural classes of actin-binding natural products as useful leads for cancer chemotherapy (3).

A growing family of actin-binding macrolides that disrupt microfilament organization includes the scytophycins (e.g., 1, Fig. 1), isolated from the blue-green algae Scytonema pseudohofmanni (4), aplyronines (e.g., 2), isolated from the sea hare Aplysia kurodai (5), and sphinxolides (e.g., 3, planar structure), first reported from an unidentified Pacific nudibranch (6) and later reisolated from the New Caledonian marine sponges Neosiphonia superstes and Reidispongia coreulea, along with the congeneric reidispongiolides (e.g., 4, planar structure) (7, 8). A related family of trisoxazole-containing marine macrolides, typified by mycalolide A (5) and jaspisamide A (6), are also antimicrofilament agents (9). Recently, an x-ray crystal structure of the latter compound bound to actin (10) highlighted the importance of the characteristic N-vinyl formamide-containing side chain, common to all these macrolides, in determining their biological activity. Moreover, all these actin-binding macrolides inhibit the proliferation of human cancer cell lines; for example, the scytophycins and sphinxolides exhibit potent cytotoxicity against L1210 and KB cells, with IC₅₀ values ranging from ng/ml to pg/ml (9). In addition, aplyronine A shows potent antitumor activity in vivo and has potential as a clinical candidate (9). Notably, both the scytophycins and sphinxolides circumvent multidrug resistance mediated by the overexpression of P-glycoprotein or multidrug resistance protein (11). Although these preliminary studies demonstrate the potential of these macrolides as leads for the development of novel anticancer drugs, the scarcity of material available from the natural source has generally hampered extensive preclinical development.

Fig. 1.

Structurally related actin-binding macrolides.

As part of ongoing studies into the stereochemical determination and synthesis of cytotoxic macrolides of marine origin and designed analogues (12–15), we initiated a project directed toward the total synthesis of the reidispongiolides and sphinxolides, linked intimately with the elucidation of their relative and absolute stereochemistry. Herein, we propose a configurational assignment of reidispongiolide A based on an accumulation of synthetic studies, performed in association with detailed NMR analysis.

Materials and Methods

Experimental details along with physical and spectral data for all new compounds are published as supporting information on the PNAS web site.

Previous Structural Studies on the Sphinxolides and Reidispongiolides

The planar structure 3 for sphinxolide A, a 26-membered macrocyclic lactone, was first reported in 1989 by Pietra and coworkers (6). However, only recently has a partial configurational assignment for this highly oxygenated and stereochemically elaborate macrolide (17 stereocenters and 6 alkenes) been proposed (16). Extensive analysis of sphinxolide by NMR methods, by Gomez-Paloma and coworkers (16), based largely on the interpretation of proton-carbon ^2,3J couplings, enabled the assignment of the relative configurations in the five isolated stereoclusters, as indicated in the boxed regions of structure 8 in Scheme 1. Additionally, controlled ozonolysis of the closely related macrolide reidispongiolide A (having the proposed stereostructure 7), followed by reductive workup, provided the three distinct polyol fragments 9 (mixture of C₅ epimers), 10 and 11 (mixture of C₃₁ epimers) (17).

Scheme 1.

Proposed stereostructures of reidispongiolide A and sphinxolide A and chemical degradation fragments of reidispongiolide A.

As a prelude to a planned total synthesis of the sphinxolide/reidispongiolide macrolides, the preparation in our group of these three degradation fragments (or stereoisomers thereof) was required to simplify the stereochemical quandary and establish rigorously the interconnections between the isolated stereoclusters. To this end, we and the D'Auria group independently reported the synthesis of the degradation fragments 10 and 11 (17–19), and established unequivocally the relationship between the three corresponding stereoclusters in structure 7. We now focused our efforts on attaining a reliable configurational assignment for the remaining C₅–C₁₆ portion of reidispongiolide A, through the synthesis of the third degradation fragment 9, thus enabling the full stereostructure of these cytotoxic macrolides to be determined.

Initial Synthesis of the Postulated C₅–C₁₆ Degradation Fragment

Based on the uncertainty regarding the configurational relationship between the C₁₁–C₁₅ stereotetrad and the isolated methoxy-bearing stereocenter at C₇, we designed a flexible and modular route to the C₅–C₁₆ degradation fragment, having the stereochemistry proposed as shown in Scheme 2 based inter alia on the findings of Gomez-Paloma and coworkers (16) and D'Auria and coworkers (17). Initially, the (5R,7R)-configured stereoisomer 13 was targeted, incorporating the C₇ and C₁₃ stereochemistry of the related macrolide scytophycin C (4) (1, Fig. 1). It was anticipated that the construction of the dihydropyranone ring could be accomplished by using an intramolecular Horner–Wadsworth–Emmons reaction. Further, it was envisaged that a substrate-controlled 1,5-anti-boron aldol coupling would provide the necessary stereoinduction from C₇ to install the requisite C₁₁ center, and 14 and 15 could be accessed by means of boron (20) and Mukaiyama (21) aldol-coupling strategies, respectively.

Scheme 2.

Retrosynthesis for the postulated C₅–C₁₆ degradation fragment.

The convergent assembly of (5R,7R)-13 began with the aldol coupling of the methyl ketone 16 [prepared in 3 steps (51%) from methyl (S)-2-methyl-3-hydroxypropionate (22)] and crotonaldehyde (Scheme 3). Although reaction of the lithium enolate provided a 1:1 mixture of epimeric alcohols, which would find use in our later studies (see below), useful levels of diastereoselectivity could be realized for the 1,4-syn-adduct 18 when a boron aldol coupling was performed with c-Hex₂BCl/Et₃N (23, 24). The induction was further enhanced by using (-)-Ipc₂BCl (20) to provide 18 in 10:1 dr.† This aldol adduct proved to be a suitable substrate for hydroxyl-directed reduction, securing a 1,3-anti-relationship between C₁₃ and C₁₅. By using Me₄NBH(OAc)₃ (26), diol 20 was obtained cleanly (83%, >10:1 dr). The stereochemical relationship between the hydroxyl-bearing centers was determined by nuclear Overhauser effect analysis of the corresponding p-methoxybenzylidene acetal (see supporting information for details). Conversion into the bis-methyl ether (NaH and MeI), was followed by cleavage of the triisopropylsilyl (TIPS) ether [tetra-n-butylammonium fluoride (TBAF)] and Dess–Martin oxidation, providing the aldehyde 14 (84%).

Scheme 3.

Stereocontrolled synthesis of aldehyde 14. Conditions: a, Me₄NBH(OAc)₃, MeCN-HOAc (3:1), -20°C; b, NaH, MeI, THF; c, TBAF, THF; d, Dess–Martin periodinane, NaHCO₃, CH₂Cl₂.

Next, we turned to the preparation of the methyl ketone 15. Methanolysis of polyhydroxybutyrate 22 (MeOH, H₂SO₄) was followed by t-butyldimethylsilyl (TBS) protection. Controlled reduction of 23 (diisobutylaluminium hydride, -90°C) gave the corresponding enantiopure aldehyde (76%). Treatment with 2-(trimethylsiloxy)propene and Me₂AlCl effected a 1,3-anti-selective (27) Mukaiyama aldol coupling with 5.2:1 dr.† After methylation of 24 with Meerwein's salt, we next explored the critical 1,5-anti-boron aldol coupling of the resulting ketone 15 and 14. By using our standard conditions (c-Hex₂BCl, Et₃N) (28), the desired adduct 25 was obtained in 70% yield (5:1 dr). With the six stereocenters of the postulated (5R,7R)-degradation fragment 13 secured, construction of the dihydropyranone moiety was sought. Yamaguchi esterification (29) of the β-hydroxyketone 25 with diethylphosphonoacetic acid provided the desired ester. Because of the inherent instability of this substance, it was treated directly with Ba(OH)₂ in wet tetrahydrofuran (THF) (30, 31) to effect the Horner–Wadsworth–Emmons cyclization, thus generating the dihydropyranone 26 (80%), along with minor amounts (< 10%) of the corresponding elimination product. Finally, acidic removal of the TBS ether was followed by brief exposure to ozone (5–10 s) and reductive workup (NaBH₄, MeOH) to selectively cleave the side chain alkene, generating the postulated C₅–C₁₆ degradation fragment 13 (44%).

Comparison of both the ¹H and ¹³C NMR spectra for 13 with the data reported by D'Auria and colleagues (17) indicated it was clearly a diastereomer of the C₅–C₁₆ degradation fragment obtained from reidispongiolide A. The three methoxy resonances, reported (17) at δ 3.37, 3.40, and 3.45 ppm appear at δ 3.42, 3.43, and 3.48 ppm, respectively, for (7R,5R)-13. Although these inconsistencies may reflect an incorrect relationship between the methoxy-bearing stereocenter at C₇ and the C₁₁–C₁₅ stereocluster, a number of more pronounced discrepancies (see comparison NMR data listed in Scheme 4) suggested the reported configurational assignment of the C₅–C₁₆ degradation fragment as 12 was erroneous. Understandably, the Naples group reported difficulties in assigning the relative configuration in the C₁₃–C₁₅ region of sphinxolide A because of the “rapid interconversion of different conformers with comparable populations” (16); in fact, eventual assignment of the relative stereochemistry at both C₁₃ and C₁₅ required reacquisition of some spectral data at lower temperatures. Based on these reported difficulties and the observed spectroscopic inconsistencies between 9 and 13, the stereostructure of this region of the parent sphinxolide/reidispongiolide macrolides, as depicted in the blue box of 8 (Scheme 1), now came under question.

Scheme 4.

Synthesis of the postulated C₅–C₁₆ degradation fragment 13. Conditions: a, MeOH, H₂SO₄; b, TBSCl, imidazole, dimethylformamide; c, diisobutylaluminium hydride, CH₂Cl₂, -90°C; d, 2-trimethylsiloxypropene, Me₂AlCl, CH₂Cl₂, -90°C; e,Me₃OBF₄,CH₂Cl₂; f,(i) c-Hex₂BCl, Et₃N, Et₂O, -78°C; (ii) 14,Et₂O; g, diethylphosphonoacetic acid, DMAP, Et₃N, 2,4,6-trichlorobenzoyl chloride, CH₂Cl₂, 0°C; h, Ba(OH)₂, THF, 0°C; i, p-TsOH, MeOH, 0°C; j, O₃, MeOH, -78°C; NaBH₄, -78°C to rt. †, 500 MHz, CD₃OD, δ (ppm).

Synthesis of Fragment Library for NMR Analysis

After this initial setback, we initiated a synthetic program to elucidate the relative stereochemistry within the C₁₁–C₁₅ region of the reidispongiolides/sphinxolides. As we expected, the remote C₅ and C₇ stereocenters would have little effect on the chemical shifts observed in the C₁₁–C₁₆ region of the degradation fragment, it was anticipated that access to a library of diastereomeric C₈–C₁₆ subunits would allow a confident configurational analysis. Although Kishi and coworkers (32, 33) have elegantly demonstrated the predictive ability of universal NMR databases for stereochemical assignment of acyclic compounds, this did not help the present situation. Our concerns regarding the assignment of C₁₃ and C₁₅ directed our initial efforts toward construction of the four C₈–C₁₆ diastereomers 27–30 detailed in Scheme 5. It was expected that the synthesis of this compound library would follow an route analogous to that disclosed above and, hence, discussion of this work is limited to the preparation of the (13R,15R)- and (13S,15R)-diastereomers 27 and 29, respectively.

Scheme 5.

Library of synthetic C₈–C₁₆ diastereomers.

Having already prepared the aldehyde 14, the synthesis of the (13R,15R)-configured library member 27 was first investigated. Based on earlier work (27), we anticipated a useful level of induction for the Mukaiyama aldol coupling of 2-trimethylsiloxypropene and 14. In the event, only modest selectivity was observed for the BF₃·OEt₂-promoted reaction (3:2 dr, Scheme 6). Although other Lewis acids failed to offer any improvement, the diastereomeric aldol adducts were readily separable by flash chromatography and the major product, incorporating the (R)-configuration at C₁₁,^† was converted into the corresponding dihydropyranone 32 by using the Horner–Wadsworth–Emmons approach detailed above. Fortunately, the disappointing substrate control could be circumvented by relying instead on Brown's methodology (34) and ring-closing metathesis (35) to introduce the required dihydropyran moiety. Thus, treatment of 14 with the chiral methallylborane reagent derived from 2-methylpropene and (+)-Ipc₂BOMe gave homoallylic alcohol 33 exclusively (Scheme 6). Esterification with acrylic acid [4-dimethylaminopyridine (DMAP) and N,N′-dicyclohexylcarbodiimide] provided 34, which underwent ring-closing metathesis mediated by Grubbs' second-generation ruthenium catalyst to give the dihydropyranone 32. Finally, brief exposure to ozone, followed by reductive workup (NaBH₄), provided the (13R,15R)-fragment 27. Notably, its spectral data were in close agreement to those observed for the analogous region of the full synthetic C₅–C₁₆ degradation fragment 13 (Scheme 4), supporting our assertion that the isolated C₅–C₇ stereocluster would have minimal effect on the C₁₁–C₁₅ ¹H NMR chemical shifts.

Scheme 6.

Synthesis of the (13R,15R) fragment 27. Conditions: a, 2-(trimethylsiloxy)propene, BF₃·OEt₂,CH₂Cl₂, -90°C; b, diethylphosphonoacetic acid, DMAP, Et₃N, 2,4,6-trichlorobenzoyl chloride, CH₂Cl₂,0°C; c, Ba(OH)₂, THF, 0°C; d, (i) 2-methylpropene, n-BuLi, TMEDA, Et₂O, -78°C to rt; (ii) (+)-Ipc₂BOMe, Et₂O; (iii) 14, Et₂O, -78°C; e, acrylic acid, N,N′-dicyclohexyldicarbodiimide, DMAP, CH₂Cl₂; f, Grubbs' II, CH₂Cl₂,reflux; g,O₃, MeOH, -78°C; NaBH₄, -78°C to rt. †, 500 MHz, CD₃OD, δ (ppm).

We next turned to the synthesis of the (13S,15R)-diastereomer 29, which now used a 1,3-syn-selective reduction (24) of the β-hydroxy ketone 18 (c-Hex₂BCl and LiBH₄), providing the diol 35 in good yield (80%, >10:1 dr, Scheme 7). The relationship between the hydroxyl-bearing stereocenters was determined by ¹³C NMR analysis of the corresponding acetonide using the method reported by Rychnovsky and Skalitzky (36). After our disappointment in the Mukaiyama aldol coupling of the related aldehyde 14 (Scheme 6), synthesis of the dihydropyranone again relied on the Brown methallylboration–ring-closing metathesis protocol.^† Subsequent chemoselective ozonolysis, followed by reductive workup, generated the (13S,15R)-fragment 29. Synthesis of fragments 28 and 30 (Fig. 2), and hence completion of the library of C₈–C₁₆ diastereomers, involved treatment of the ketone 19, obtained previously from the lithium aldol reaction between methyl ketone 16 and crotonaldehyde (Scheme 3), in a manner identical with that presented in Schemes 6 and 7.

Scheme 7.

Synthesis of the (13S,15R) fragment 29. Conditions: a, (i) c-Hex₂BCl, Et₃N, -78°C; (ii) LiBH₄; b, NaH, MeI, THF; c, TBAF, THF; d, Dess–Martin periodinane, NaHCO₃,CH₂Cl₂; e,(i) 2-methylpropene, n-BuLi, TMEDA, Et₂O, -78°C to rt; (ii) (+)-Ipc₂BOMe, Et₂O; (iii) 37, -78°C; f, acrylic acid, N,N′-dicyclohexyldicarbodiimide, DMAP, CH₂Cl₂; g, Grubbs' II, CH₂Cl₂, reflux; h, O₃, MeOH, -78°C; NaBH₄, -78°C to rt.

Fig. 2.

Additional synthetic C₈–C₁₆ diastereomers.

Examination of the ¹H and ¹³C NMR spectra of the fragment library in CD₃OD revealed distinctly different chemical shifts for each of the diastereomers 27–30. Comparative ¹H NMR data are presented in Fig. 3 as the difference in ppm between each individual proton resonance in the fragments 27–30 and those reported for the analogous protons of the natural C₅–C₁₆ degradation fragment 9 (17). This analysis convincingly defines the relative stereochemistry within this region of the degradation fragment, and consequently reidispongiolide A, as that represented by the (13S,15S)-diastereomer 30.

Fig. 3.

Comparative ¹H NMR analysis of the fragment library. Bars represent deviation in ppm between individual proton chemical shifts observed for the diastereomers 27–30 and those reported for the natural C₅–C₁₆ degradation fragments 9¹⁷ (500 MHz, CD₃OD).

Configurational Assignment of the C₅–C₁₆ Degradation Fragment of Reidispongiolide A

Confident in this reassignment of the relative configuration within the isolated C₁₁–C₁₅ stereocluster, the establishment of a relationship between this subunit and the isolated C₇ methoxy-bearing stereocenter in the parent macrolide was investigated next. We anticipated that a synthesis of potential C₅–C₁₆ degradation fragments, bearing the relative stereochemistry defined by 30 and either (R)- or (S)-configuration at C₇, would enable stereochemical determination of the C₅–C₁₆ degradation fragment 9. In particular, we expected that subtle differences in the chemical shift of the diastereotopic methylene protons at C₈ and C₁₀ would indicate the correct relative configuration. With this in mind, we designed a flexible aldol-based approach, using ketone 45 (or ent-45) and aldehyde 48 (Schemes 8, 9, 10). In addition, as the C₅–C₁₆ degradation fragment 9, obtained from the ozonolysis/reduction of reidispongiolide A, was isolated as a 2:1 mixture of C₅ epimers (17), our planned approach incorporated a chemoselective ozonolysis-reduction sequence of both a C₅ methylene and C₁₅ ethylidene in the presence of the dihydropyranone to access both C₅ epimers in each series.

Scheme 8.

Synthesis of the enantiomeric methyl ketones 45 and ent-45. Conditions: a, (i) 2-methylpropene, n-BuLi, TMEDA, Et₂O, -78°C to rt; (ii) (+)-Ipc₂BOMe, Et₂O; (iii) 41, -78°C; b, NaH, MeI, THF; c, TBAF, THF; d, Dess–Martin periodinane, NaHCO₃, CH₂Cl₂; e, MeLi, THF, -78°C.

Scheme 9.

Synthesis of the (7R)-epimer of the revised C₅–C₁₆ degradation fragment. Conditions: a, Me₄NBH(OAc)₃, MeCN-HOAc (3:1), -20°C; b, NaH, MeI, THF; c, TBAF, THF; d, Dess–Martin periodinane, NaHCO₃, CH₂Cl₂; e, (i) ent-45, c-Hex₂BCl, Et₃N, -78°C, Et₂O; (ii) 48; f, diethylphosphonoacetic acid, DMAP, Et₃N, 2,4,6-trichlorobenzoyl chloride, CH₂Cl₂,0°C; g, Ba(OH)₂, THF, 0°C; h, O₃, MeOH, -78°C; NaBH₄, -78°C to rt. †, 500 MHz, CD₃OD, δ (ppm). ‡, Isolated as 1:2 mixture of (5R)-51a/(5S)-51b.

Scheme 10.

Synthesis of the (7S)-epimer of the revised C₅–C₁₆ degradation fragment. Conditions: a, (i) 45, (-)-Ipc₂BCl, Et₃N, -78°C, Et₂O; (ii) 48; b, diethylphosphonoacetic acid, DMAP, Et₃N, 2,4,6-trichlorobenzoyl chloride, CH₂Cl₂, 0°C; c, Ba(OH)₂, THF, 0°C; d, O₃, MeOH, -78°C; NaBH₄, -78°C to rt. †, 500 MHz, CD₃OD, δ (ppm). ‡, Isolated as 1:2 mixture of (5S)-54a/(5R)-54b.

As outlined in Scheme 8, we first prepared the methyl ketones 45 and ent-45, to access both possible relationships between the C₇ stereocenter and the C₁₁–C₁₅ stereotetrad. The aldehyde 41 was treated with the methallylborane reagent derived from 2-methylpropene and (+)-Ipc₂BOMe (34) to afford the homoallylic alcohol 42 in 85% yield (96% enantiomeric excess).‡ After methyl ether formation, the TBS group was removed to give alcohol 44 in good yield. Oxidation to the aldehyde, addition of MeLi, and subsequent reoxidation provided the volatile methyl ketone 45. By using (-)-Ipc₂BOMe, ent-45 was prepared from 41 in an analogous manner. The specific rotation of the secondary alcohols 42 and ent-42 in CHCl₃ were +2.9 and -2.8, respectively.

We next examined the coupling of the ketone ent-45 and the aldehyde 48, which had been previously used for the construction of both 28 and 30 (Fig. 2), as would be required for the C₅–C₁₆ fragment library. The synthesis of 48 commenced with a 1,3-anti-selective reduction (Me₄NBH(OAc)₃) of ketone 19 (Scheme 3). Conversion into the bis-methyl ether was then followed by removal of the TIPS group and Dess–Martin oxidation (89%) to afford the desired aldehyde 48. By following our standard protocol (27), the 1,5-anti-boron aldol coupling between 48 and the methyl ketone ent-45 proceeded smoothly, affording the desired aldol adduct in high yield and diastereoselectivity (90%, 10:1 dr).§ Completion of the synthesis of the (7R)-degradation fragments 51 involved formation of the diethylphosphonoacetate, which was directly treated with Ba(OH)₂ to provide the Horner–Wadsworth–Emmons product 50 in 67% overall yield. Finally, brief subjection of 50 to ozone and addition of NaBH₄ yielded the alcohols 51 as a 2:1 mixture of epimers at C₅, which were separated by HPLC. Spectroscopic analysis of alcohols 51 (500 MHz, CD₃OD) indicated that they were clearly diastereomers of the C₅–C₁₆ degradation fragment 9 (17). Most notably, in the ¹H NMR spectra of the purified alcohols, the resonances corresponding to the diastereotopic methylene protons at C₈ and C₁₀ were significantly different in chemical shift from those reported for the naturally derived material (Scheme 9) (17). Although this result suggested an incorrect relative configuration at C₇ in the alcohols 51, it remained to synthesize the corresponding fragment with the opposite or (S)-configuration at C₇ to corroborate this proposal. The synthesis of the (7S) C₅–C₁₆ fragment is detailed in Scheme 10. By following our standard conditions (20), the aldol coupling between 45 and 48 with (-)-Ipc₂BCl/Et₃N afforded the desired β-hydroxyketone 52 as the major adduct.§ In this case, the inherent 1,5-anti-selectivity observed in the coupling of ent-45 and 48 (Scheme 9) was overturned by appropriate choice of Ipc ligand chirality. Completion of the synthesis of (7S)-degradation fragment involved an analogous sequence of reactions to those discussed above, affording the alcohols 54 as a 2:1 mixture of epimers at C₅, which were separated by HPLC. Gratifyingly, the spectral data obtained for these alcohols were in good agreement to that reported by D'Auria and coworkers for the C₅–C₁₆ fragment obtained in their degradation work (17), thus allowing our assignment of the relationship between the remote stereocenter at C₇ and the C₁₁–C₁₅ stereocluster in the parent macrolide reidispongiolide A (7).¶

Conclusion

Based on the proposed relative configuration within the C₁₁–C₁₅ subunit of the reidispongiolide/sphinxolide family of cytotoxic marine macrolides, an efficient aldol-based synthesis of the postulated C₅–C₁₆ degradation fragment of reidispongiolide A was accomplished. Subsequently, several inconsistencies observed between the spectral data recorded for the synthetic and naturally derived fragments led us to construct a library of C₈–C₁₆ and C₅–C₁₆ stereoisomers and confidently reassign the relative stereochemistry within this region of reidispongiolide A. At present, the full macrolide may be represented by either stereostructure 55 or 56 (Fig. 4), where the final assignment will rely on a stereocontrolled synthesis of the macrocyclic core and reidispongiolide A itself.

Fig. 4.

The two potential stereostructures of reidispongiolide A, incorporating the revised C₅–C₁₆ stereochemistry determined in this work.