Abstract
Synthesis of a novel class of C-10 halogenated and C-12 oxygenated prostaglandin-A2 derivatives 6(a–c) has been accomplished. (15S)-Prostaglandin-A2 (1), from the gorgonian Plexaura homomalla, served as the starting material for the synthesis. The absolute configuration was determined using NMR.
Δ12-Prostaglandin J2 (Δ12-PGJ2), a cross-conjugated enone, is known to inhibit ubiquitin specific isopeptidase activity causing apoptosis.1 Mechanistic studies have indicated that inhibition results from Michael addition of an isopeptidase cysteine residue to the endocyclic β-carbon of the cyclopentenone.2 Punaglandins (PNGs), C-10 chlorinated and C-12 oxygenated prostanoids, are more potent inhibitors of isopeptidase activity than Δ12-PGJ2 and the PGA series (Figure 1).3 On the basis of these results, increasing the electrophilicity of the endocyclic β-carbon (C-10) should increase reactivity. To test this hypothesis, a new series of halogenated PGA2 derivatives were synthesized (4a–4c) by substituting electron withdrawing groups (Cl, Br, I) at the α-position of PGA2 (Scheme 1).
Figure 1.
Electrophilic prostanoids
Scheme 1.
Synthesis of α-halogenated PGA2 analogues
Preliminary work with these α-halogenated PGA2 analogues (4a–4c) indicated that the potency of isopeptidase inhibition for the series is I≫Br≥Cl.4 Additionally, Iodo-PGA2 was a more potent inhibitor of ubiquitin specific isopeptidase activity compared to Δ12- PGJ2, but less potent than the PNGs. Thus it was recognized that halogenation and electophilicity do play a role in the ability of prostaglandins to inhibit ubiquitin isopeptidase activity, but are not the only factors. These results suggested C-12 hydroxylation similar to PNGs may be necessary to achieve optimum activity. To test this hypothesis, a series of C-10 halogenated and C-12 oxygenated PGA2 derivatives (6a–6c) have been synthesized (Scheme 3).4
Scheme 3.
Synthesis of C-12 oxygenated PGA2 derivatives
The starting material for the synthesis, (15S)-prostaglandin-A2 (PGA2) was isolated in abundance (5.7% recovery)5 from the gorgonian Plexaura homomalla.6 The side-chain functional groups of PGA2 (the terminal carboxylic acid at C-1 and the alcohol at C-15) were blocked as their methyl ester and acetate respectively, using standard derivatization conditions (Scheme 1).7
The 15-O-acetyl-PGA2 methyl ester (2) was treated with alkaline hydrogen peroxide at −15 °C to generate a mixture of C(10,11)- α,β-epoxides (3).8 The subsequent C-10 halide substitutions9 were achieved by regioselective ring opening of the C(10,11)-epoxide (3) with halide (Cl, Br, I) salts under mildly acidic conditions (Amberlyst 1510 or silica gel support11). Similar transformations of epoxy cyclopentenones into α-haloenones have been shown to occur via the halohydrin intermediates which are spontaneously dehydrated under acidic conditions to afford the vinylhalide products.10
Accordingly, the method of choice for the formation of vinylchloride (4a) was LiCl/Amberlyst/CH3CN system, where as the vinylbromide (4b) was formed by NaBr under otherwise identical conditions.12 However, attempts to form the vinyl iodide (4c) under similar conditions (LiI or NaI/Amberlyst/CH3CN) were unsatisfactory and resulted in the re-formation of the alkene (2). A subsequent literature survey of alternative methods of vinyliodide formation13 from epoxides, revealed that silica gel could be employed as an efficient acid catalyst in the nucleophilic ring-opening of epoxides under solvent-free conditions.11 Accordingly, the solvent-free iodination of epoxide (3) was carried out using LiI supported on silica gel. The reaction proceeded smoothly to give 54% of the vinyliodide (4c) along with 13% of the alkene by-product (2).
The next task of the synthesis was the conversion of the C-10 halogenated PGA2 series (4a–4c) to the corresponding C-12 hydroxylated derivatives (6a–6c). This novel series of C-12 hydroxylated derivatives were believed to be accessible via allylic oxidation mediated by selenium dioxide,14 owing to the fact that the C-12 position of PGA2 is bis-allylic to C-10 and C-13 double bonds (Scheme 2). However, the likely formation of byproducts due to multiple potential oxidation sites, were foreseen as a possible drawback of this method.
Scheme 2.
SeO2 oxidation of 2
To test the feasibility of the allylic oxidation reaction, PGA2-acetate methyl ester (2) was employed. As speculated, compound 2 with SeO2 gave several oxidation by-products. One major by-product was identified as the conjugated diene (2a) possibly arising from allylic hydroxylation at C-7 followed by dehydration (Scheme 2). Shorter reaction times, lower reflux temperatures (95% EtOH) or reduced amounts of the oxidant did not facilitate formation of the desired tertiary alcohol product (6).
It was now evident that in order to selectively hydroxylate the bis-allylic position (C-12) of PGA2 and suppress by-product formation during SeO2 oxidation, the other allylic positions (C-4 and C-7) needed to be removed or masked in some manner. This was achieved by selective epoxidation of the C(5,6) double bond in 4 using m-CPBA (Scheme 3). The resulting C(5,6)- α,β-epoxide mixture (5) was treated as an intermediate and used without further purification for the SeO2 oxidation. As anticipated, allylic hydroxylation at C-12 position proceeded smoothly and was accompanied by C(5,6)-epoxide ring opening to yield the novel series of prostaglandin-A2 analogues (6a–6c)15 in moderate yields (Scheme 3).
The relative and absolute configurations of the five stereocenters (C-5/6/8/12/15) in 6(a–d) were deduced as follows. The absolute stereochemistry of C-8 was assumed (R), based on literature precedent for coral-derived PGA2.16 The assignment of the 15(S) configuration of 6 was based on the application of Mosher’s method17 to (S)- and (R)-MPA derivatives of PGA2-methyl ester.18 The trans (R,R) relationship between the two side chains (C-8/C-12) was established based on a strong NOESY cross peak observed between H-8 and H-13. Subsequent molecular modeling studies with energy-minimized conformations for each of the C-12 epimers supported this observation. The trans-(8R,12R) relationship was further corroborated by heteronuclear coupling constant analysis (2JH8/C12 = small)19, and dihedral angle measurements (−135.7°) for the 8(R),12(R) diastereomer of 6.20
The cis orientation of the C(5,6) epoxide mixture [(5S,6R) and (5R,6S) diastereomers] in 5 was supported by the homonuclear coupling constant value of 3JH5/H6 = 4.3 Hz.21 It was presumed that the C(5,6) diol in 6, generated during the allylic oxidation (SeO2/dioxane/H2O/reflux) of 5, occurs via nucleophilic opening of the C(5,6) epoxide by water. Accordingly, non-regioselective attack of water on the (5S,6R) and (5R,6S) cis-epoxides in 5 would generate two secondary alcohol diastereomers, (5R,6R) and (5S,6S). This was supported by the 1H-NMR spectra of 6(a–d) which showed doubling of resonances due to the formation of two diastereomers in a ~1:1 ratio. Accordingly, HPLC analysis of 6d showed two peaks eluting at 11.5 (7) and 13 min (8). The two peaks (7 and 8) were separated,22 and were employed for absolute stereochemical assignments of the C(5,6) centers by J-based analysis,19 Mosher’s method17 in conjunction with molecular modeling studies (Figure 2).
Figure 2.
C(5,6)-diol diastereomers of 6d
Initial attempts focused on utilizing J-based analysis19 to relate the configuration of the vicinal diol segment C(5,6) to that of the known absolute configuration at C-8. However, diastereomer 7 proved undesirable for J-analysis due to overlap of resonances for H-5 and H-6 (δ3.51–3.57, m, 2H) as well as for H-7h and H-7l (δ 1.98, m, 2H). Therefore, the viable alternative was to attempt the J-based analysis on the diastereomer 8.
For diastereomer 8, sufficient chemical shift dispersion of all proton resonances were observed in CDCl3 at 0 °C. While all of the couplings required for J-analysis were determined successfully (Table 1), the information failed to fit any of the typical rotamers with certainty.19
Table 1.
2,3J values of 8 for the C-6/C-7 segment measured in CDCl3
| 2,3J valuesa | J (Hz) | classificationb |
| 3J (H-6, H-7h) | + 6.3 | medium |
| 3J (H-6, H-7l) | + 6.3 | medium |
| 3J (H-6, C-8) | + 2.2 | small |
| 3J (H-5, H-7h) | +4.4 | medium |
| 3J (H-5, H-7l) | +2.2 | small |
| 2J (C-6, H-7h) | −4.3 | medium |
| 2J (C-6, H-7l) | −1.1 | small |
H-7h and H-7l represent the high- and low-field H-7 protons respectively.
Classification for magnetude of the coupling constants: values between large and small are regarded as medium19
The reason for the occurrence of atypical coupling values in 8 could be explained based on a molecular modeling study, which revealed that the cyclopentenone carbonyl (C-9) is involved in an intra-molecular H-bonding with C-6(OH) adopting a seven-membered cyclic conformation (Figure 3). The measured J-values and NOESY data were in accordance with such a seven-membered conformation.
Figure 3.
The MM2/AM1-optimized model for 8
Having failed to unequivocally determine the C(5,6) stereochemistry by J-based analysis, Mosher’s method was attempted on 8. Accordingly, reaction of 8 with (R)- and (S)-MPA acids23 was expected to yield the corresponding C(5,6)-bis-MPA esters. However, 1H-NMR data confirmed that the C-5 MPA derivative was the sole product. Comparison of the chemical shift differences (Δδ = δR−δS) in the 1H-NMR spectra of the C-5 (R)- and (S)-MPA derivatives of 8 at 0 °C indicated the absolute configuration was (R) (Figure 4).
Figure 4.
Δδ values for 8 (Δδ = δR−δS, 500 MHz, CDCl3, 0 °C)
Thus the absolute configuration of all five stereo-centers (C-5/6/8/12/15) in 8 was established as (5R,6R,8R,12R) and 15(S). Consequently, the absolute configuration of the diol diastereomer 7 was assigned as (5S,6S,8R,12R) and 15(S). The absolute configuration assignments of 7 and 8 were in good agreement with their 1H NMR data, coupling constant analysis and molecular modeling studies employing MM2 and AM1.24
Supplementary Material
Experimental details and spectroscopic data for synthetic compounds 2-8. This material is available free of charge via the Internet at http://pubs.acs.org.
Acknowledgments
The authors wish to acknowledge Dr. Joseph R. Pawlik (University of North Carolina at Wilmington) for collection of the gorgonian. Dr. P. Krishna at the University of Utah Mass Spectrometry and Proteomics Core Facility is acknowledged for performing the LRESIMS experiments. The HRESIMS data were acquired at the University of Iowa High Resolution Mass Spectrometry Facility. This project was funded by NIH grant CA 36622.
References
- 1.(a) Mullally JE, Moos PJ, Edes K, Fitzpatrick FA. J Biol Chem. 2001;276:30366. doi: 10.1074/jbc.M102198200. [DOI] [PubMed] [Google Scholar]; (b) Mullally JE, Fitzpatrick FA. Mol Pharmacol. 2002;62:351. doi: 10.1124/mol.62.2.351. [DOI] [PubMed] [Google Scholar]
- 2.Masaaki S, Mori M, Niwa T, Hirata R, Furuta K, Ishikawa T, Noyori R. J Am Chem Soc. 1997;119:2376. [Google Scholar]
- 3.Verbitski SM, Mullally JE, Fitzpatrick FA, Ireland CM. J Med Chem. 2004;47:2062. doi: 10.1021/jm030448l. [DOI] [PubMed] [Google Scholar]
- 4.Manuscript thoroughly describing the pharmacology (outcome of the SAR studies and details of the biological assays data) will be published elsewhere.
- 5.See Experimental section for details on the isolation of PGA2 from P. homomalla.
- 6.Schneider WP, Hamilton RD, Rhuland LE. J Am Chem Soc. 1972;94:2122. doi: 10.1021/ja00761a060. [DOI] [PubMed] [Google Scholar]
- 7.Schneider WP, Bundy GL, Lincoln FH, Daniels EG, Pike JE. J Am Chem Soc. 1977;99:1222. doi: 10.1021/ja00446a039. [DOI] [PubMed] [Google Scholar]
- 8.1H NMR and HPLC analysis of crude mixture (3) showed no evidence of epimerization at C-8 during the epoxidation reaction.
- 9.(a) Dominguez JN, Taddei A, Cordero M, Blanca IJ. J Pharm Sci. 1994;83:472. doi: 10.1002/jps.2600830405. [DOI] [PubMed] [Google Scholar]; (b) Iguchi K, Kaneta S, Tsune C, Yamada Y. Chem Pharm Bull. 1989;37:1173. [Google Scholar]
- 10.Righi G, Bovicelli P, Sperandio A. Tetrahedron Lett. 1999;40:5889. [Google Scholar]
- 11.Kotsuki H, Shimanouchi T, Ohshima R, Fujieara S. Tetrahedron. 1998;54:2709. [Google Scholar]
- 12.In a few instances during scale up, the intermediate chloro and bromohydrins were isolated. These halohydrins can be readily dehydrated to form the vinyhalides, by heating with glacial acetic acid on a steam bath for 2–5 h.
- 13.(a) Johnson CR, Adams JP, Braun MP, Senanayake CBW. Tetrahedron Lett. 1992;33:917. [Google Scholar]; (b) Konaklieva MI, Dahl ML, Turos E. Tetrahedron Lett. 1992;33:7093. [Google Scholar]; (c) Palumbo G, Ferreri C, Caputo R. Tetrahedron Lett. 1983;24:1307. [Google Scholar]
- 14.(a) Sharpless KB, Lauer RF. J Am Chem Soc. 1972;94:7154. [Google Scholar]; (b) Groszek G, Tyrlik KA, Wicha J. Tetrahedron. 1989;45:2223. [Google Scholar]
- 15.NMR data were recorded at 0 °C. Compounds stored at r.t. tend to decompose slowly over 3–5 days.
- 16.(a) Baker BJ, Okuda RY, Yu PTK, Scheuer PJ. J Am Chem Soc. 1985;107:2976. [Google Scholar]; (b) Weinheimer AJ, Spraggins RL. Tetrahedron Lett. 1969;10:5185. doi: 10.1016/s0040-4039(01)88918-8. [DOI] [PubMed] [Google Scholar]
- 17.(a) Ohtani I, Kusumi T, Kashman Y, Kakisawa H. J Am Chem Soc. 1991;113:4092. [Google Scholar]; (b) Seco JM, Quiñoà E, Riguera R. Tetrahedron: Asymmetry. 2001;12:2915. [Google Scholar]
- 18.An attempt to determine the absolute stereochemistry by a variable temperature 1H NMR experiment with a single MPA derivative, 15(R)-MPA, in CDCl3 at −1 °C and 25 °C was not successful: Latypov SK, Seco JM, Quiñoà E, Riguera R. J Am Chem Soc. 1998;120:877.
- 19.Matsumori N, Kaneno D, Murata M, Nakamura H, Tachibaba K. J Org Chem. 1999;64:866. doi: 10.1021/jo981810k. [DOI] [PubMed] [Google Scholar]
- 20.Contreras RH, Peralta JE. Prog NMR Spectrosc. 2000;37:321. [Google Scholar]
- 21.Due to the overlap of H-5 and H-6, a combination of 1H-NMR, HOMODEC and 1D-TOCSY experiments were employed to extract the coupling constants with certainty.
- 22.See Experimental section for details of the separation procedure.
- 23.2 eqvi. of MPA per OH was employed. Reaction was carried out at room temperature for ~16 h, until complete convertion to the product, as indicated by TLC and NMR.
- 24.CS Chem3D Pro; ChembridgeSoft Corporation: Cambridge, MA.
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Experimental details and spectroscopic data for synthetic compounds 2-8. This material is available free of charge via the Internet at http://pubs.acs.org.