Syntheses of porphyrins isolated from the Coral Sea demosponge Corallistes sp.

Methyl esters 2 , 3 , and 5 of corallistins B, C, and E, respectively, isolated from the Coral Sea demosponge Corallistes sp., were successfully synthesized by the MacDonald method (for corallistin B ester) and via a,c -biladiene cyclization (for corallistins C and E esters). In order to compare with the 1 H-NMR data reported in the literature, zinc(II) complexes of these corallistins were prepared, and their 1 H-NMR spectra were measured. Most chemical shifts were found to be within 0.1 ppm of the reported values, except in the case of corallistin B ester, where signal broadening due to porphyrin aggregation made the comparison with literature values difficult. Overall, the NMR data provided support for the structures proposed for the natural corallistins. Postulates on the biosynthetic origins of the corralistins are also presented


Introduction
In 1989, Pietra and coworkers 1 isolated a novel metal-free porphyrin (as its dimethyl ester) from the Coral Sea demosponge Corallistes sp., and named it corallistin A dimethyl ester (1).This became the second metal-free porphyrin isolated from a living organism (chlorophyll-c being the first).Corallistin A (the dicarboxylic acid) is present in huge amounts in the sponge, but the role of this porphyrin is still unknown.This stimulated Pietra and co-workers to further examine the demosponge, and eventually five more metal-free porphyrins were isolated. 2These included four new corallistins (as their methyl esters) -B (2), C (3), D (4), and E (5) -and the long-known deuteroporphyrin-IX dimethyl ester (6).Among the corallistins isolated so far, corallistin A was

Synthesis of Corallistin B dimethyl ester (2)
In principle, corallistin B dimethyl ester (2) can be synthesized by two different routes: the a,cbiladiene route or the MacDonald route. 7,8After consideration of symmetry issues and monopyrrole availability, we decided that the MacDonald route was the more straightforward approach.Scheme 1 shows the synthetic route which was used.Because of the symmetry in one of the required dipyrromethanes, only three different pyrroles were needed to accomplish this total synthesis.
The known 2-unsubstituted pyrrole 7 and acetoxymethylpyrrole 8 were condensed, in the presence of Montmorillonite K10 clay, to give dipyrromethane 9 which was catalytically debenzylated to give the dipyrromethane dicarboxylic acid 10.Using a variation of the MacDonald [2+2] cyclization procedure, 1,9-diformyldipyrromethane 11 and dipyrromethane-1,9-dicarboxylic acid 10 were condensed in the presence of p-toluenesulfonic acid.Air oxidation of the intermediate porphodimethene was facilitated by the addition of zinc(II) acetate with concomitant zinc insertion.Demetalation with TFA gave the desired corallistin B dimethyl ester (2) (Scheme 1).

Synthesis of Corallistin C methyl ester (3)
When one considers substituent symmetry aspects, Corallistin C methyl ester (3) can also, in principle, be synthesized by the MacDonald route. 7,8However, based on a retrosynthetic analysis of the ease of preparation of the required pyrroles, we chose instead to use a,c-biladiene cyclization 9 as the synthetic route for corallistin C methyl ester (3).Four different, but readily available, pyrroles were required for the synthesis (Scheme 2).Condensation of known pyrroles 12 and 13 in the presence of Montmorillonite K10 clay gave dipyrromethane 14.This was catalytically debenzylated to afford dipyrromethane-1,9carboxylic acid 15 before being reacted with formylpyrrole 16 in the presence of ptoluenesulfonic acid; the tripyrrin 17 was isolated as its hydrobromide salt (Scheme 2).After the removal of the tert-butyl ester group on tripyrrin 17 with TFA, and reaction with formylpyrrole 18, the a,c-biladiene was isolated as its dihydrobromide salt 19, and was macrocyclized in DMF in the presence of copper(II) acetate to form the copper(II) porphyrin 20.Corallistin C methyl ester (3) was obtained after the removal of copper from porphyrin 20 using 15% H 2 SO 4 /TFA.

Synthesis of Corallistin E methyl ester (5)
The substituent pattern in corallistin E methyl ester ( 5) is totally unsymmetrical, so this porphyrin is best synthesized by the a,c-biladiene route. 9A simple inspection leads to the fact that the structure of corallistin E methyl ester (5) is very similar to that of deuteroporphyrin-IX dimethyl ester (6).It has been known that deuteroporphyrin-IX dimethyl ester (6) can be obtained from hemin (21) by double protio-devinylation followed by the removal of iron.In planning the synthesis for corallistin E methyl ester (5), we decided to take the same approach to fashion the vacant positions at the 3-and 8-positions, as in the deuteroporphyrin-IX dimethyl ester (6) synthesis.Therefore the key intermediate would be a divinylporphyrin 22, which, being totally unsymmetrical, would be synthesized through the a,c-biladiene route.
The known bis(2-chloroethyl)dipyrromethane 23 (Scheme 3) was treated with TFA to remove the tert-butoxycarbonyl group, and the reacted with formylpyrrole 18 to form tripyrrin hydrobromide salt 24.To cleave the benzyl ester, tripyrrin hydrobromide salt 24 was stirred at room temperature in TFA/30% HBr/acetic acid, and the product was then reacted with formylpyrrole 16 to form the required a,c-biladiene, which was isolated as its dihydrobromide salt 25. a,c-Biladiene 25 was then cyclized with copper(II) acetate/DMF, and removal of copper gave bis (2-chloroethyl) Double dehydrohalogenation with 3% aqueous potassium hydroxide/pyridine and subsequent esterification with 5% H 2 SO 4 /MeOH afforded the divinylporphyrin methyl ester 22.To obtain the desired corallistin E methyl ester (5), porphyrin 22 was first converted into its iron(III) complex, and this was then heated in a resorcinol melt to accomplish protiodevinylation. 10inally, removal of iron with HCl/methanol/FeSO 4 afforded corallistin E methyl ester (5).   1, 2, and 3, respectively, along with the data from the corresponding natural zinc(II) corallistins reported by Pietra and coworkers. 2t should be mentioned that zinc(II) corallistin B dimethyl ester (27) (in CDCl 3 ) gave a 1 H-NMR spectrum having broadened signals, particularly those from the meso protons (broad singlets at 9.06 ppm and 9.30 ppm) (Table 1).The broadening of the signals is probably a result of the aggregation of the porphyrin molecules.Porphyrin NMR spectroscopy is known to be solvent and concentration dependent since the porphyrin molecules can form aggregates in solution.11 Various degrees of aggregation often affect the NMR spectrum.To alleviate this problem, a donor molecule is often added to break up the aggregates.Therefore, in our case, the 1 H-NMR spectrum of zinc(II) corallistin B dimethyl ester (27) was measured again with the compound in CDCl 3 containing a small amount of pyridine-d 5 .All the signals were sharpened, but as a result of axial ligation with pyridine, the chemical shifts of the compound differed from those observed using CDCl 3 alone.For example, the meso proton signals appeared as four singlets at 9.98, 9.99, 10.00, and 10.07 ppm, more downfield than the reported values (Table 1).

II) complexes of Corallistin B dimethyl ester, Corallistin C methyl ester, and Corallistin E methyl ester
with zinc(II) acetate to give the corresponding zinc(II) complexes from which 1 H-NMR data were obtained.The 1 H-NMR data of synthetic zinc(II) corallistins B (27), C (28), and E (29) are listed in Tables

Table 1 .
1H-NMR data from natural and synthetic zinc(II) corallistin B dimethyl ester (27) Chemical shifts (ppm) in CDCl 3 of zinc(II) complex of: Surprisingly, aggregation of the porphyrin molecules did not seem to occur in the cases of zinc(II) corallistins C methyl ester (28) and E methyl ester (29).Their 1 H-NMR (CDCl 3 ) spectra appeared to be normal. 13

Table 3 .
1 H-NMR data from natural and synthetic zinc(II) corallistin E methyl ester (29) Chemical shifts (ppm) in CDCl 3 of zinc(II) complex of:

Honor of Prof. Charles Rees ARKIVOC 2002 (vi) 264-278 ISSN 1424-6376 Page 272
Melting points were measured on a Thomas/Bristoline microscopic hotstage apparatus and are uncorrected.Electronic absorption spectra were measured on a Hewlett-Packard 8450A spectrophotometer using solutions in dichloromethane.Proton NMR spectra were recorded at 300 MHz using a QE-300 spectrometer, with chemical shifts in parts per million (ppm).Elemental analyses were performed by the Microanalytical Laboratory at the University of California, Berkeley.All reactions were monitored by thin layer chromatography and were performed on cut strips (Merck silica gel 60 F254 precoated (0.25 mm thickness) plastic-backed sheets.For column chromatography two types of packing media were generally employed; pyrroles and dipyrromethanes were usually chromatographed on Merck silica gel 60, whereas porphyrins were purified over Merck natural alumina [70-230 mesh; Brockmann Grade III (6% water)]. 2 for 24 h.The clay was filtered and the solvent of the filtrate was removed.The brown oil was chromatographed on a silica gel flash column eluting with ethyl acetate/cyclohexane (30/70) to afford the title compound as a light brown oil (194 mg, 55%).

17-Diethyl-13-(2-methoxycarbonylethyl)-2,7,12,18-tetramethylporphyrin (Corallistin C methyl
The mixture was stirred at 140 ˚C for 5 min before being allowed to cool to room temperature.It was then diluted with CH 2 Cl 2 , washed with water (2x), dried (anhy.Na 2 SO 4 ), and the solvent was removed to give the crude copper(II) porphyrin 20.This porphyrin was stirred in 15% H 2 SO 4 /TFA (15 mL) for 1 h at room temperature before being poured into iced water.After the ice had melted, the solution was extracted with CH 2 Cl 2 several times.The combined organic layer was washed with saturated sodium bicarbonate solution, brine, dried (anhy.Na 2 SO 4 ), and the solvent was removed to give a dark brown residue.The residue was chromatographed on silica gel preparative TLC plates eluting with 1% methanol/ CH 2 Cl 2 to afford the title porphyrin 3 (30 mg), mp 192-194 ˚C.UV-Vis (CH 2 Cl 2 ): λ