Transition metal complexes of porphins, corrins and related compounds

The group of macrocyclic tetrapyrrolic ligands which includes the porphins, chlorins and corrins, occur in several biologically important molecules in the form of metal complexes. Intermediate in structure between the porphins and the corrins are the corroles and the 1-monosubstituted-, and I, 19-disubstituted tetradehydrocorrins. Syntheses and reactions of metal, particularly nickel, complexes of these novel ring systems are described. The vitamin B12 coenzyme, which is a cobalt-corrin derivative containing adenosine bound to the cobalt through a 5'-carbon-cobalt covalent bond, is considered in the light of this unique structural feature. The mode of biosynthesis of the coenzyme can be partially simulated in the laboratory and a brief survey of the biochemical reactions controlled by the coenzyme is given. The recagnitian af the structures af the vital transitian metal camplexes haemin(l) and chlaraphyll(ll) 1 gave an impartant bialagical impetus ta organametallic chemistry in its very early days, and at the sametime initiated a detailed study af the precise functian, in chemical terms, of these and other organametallic compaunds in Nature 2 . Bath haemin and chlarophyll are derivatives of the Iigand parphin(III), the addition af the twa extra ß-hydrogen atams in chlorophyll causing the red-brawn colaur af haemin to be modified to green in the dihydraparphin or chlorin. Parphin is a typically aromatic heteracyclic cantaining 18 n-electrans in its extended conjugated system as weil as having the four nitrogen atoms ideally arranged far square planar or octahedral coordination with metals. A wide range of these have been utilized in the formation of porphin complexes, rhodium3 , ruthenium4 , iridium5 , and gold6 being relatively new additians, all from Fleischer's laboratory, while germanium 7 and volatile silicon derivatives8 are examples of nan-transition camplexes recently described. Meanwhile a third natural product, the vitamin B 12 coenzyme(IV), has been isolated, and shown by x-ray crystallagraphy9 to cantain the carrin ring(V) as the chromophoric Iigand, as weil as a 5' -deoxyadenosyl group attached directly to cobalt through a cabalt-carbon bond. Our interest in the naturally occurring metallic porphins and corrins 10 has led us to examine a number af properties of these and related ring systems but at the autset af 375 P.A.C.-23/4 -B A. W. JOHNSON CH:CH 2 Me CH:CH 2 Me ~ ~ ~ CH:CH2 ~ ~ Me ~ Me ~ ~ Et N N-=- " J( \ Fe / " / N N Me ~ 1 "'- /."' # Me 3'\: 2 4 ~1 ~0 19 N 18 161 ~17 {I) /; HN 1 14 12 1~,~ (111) N N- " J( \ Mg Me ~ N N " '\.. H I h-. # Me { li) {IV) our work, no synthetic routes were available to macrocyclic tetrapyrrolic ring systems containing a direct link between two of the rings. The corrole ring(VI) is intermediate in structure between porphin and corrin and we were able to develop an easy synthesis of corroles by cyclization ofthe corresponding 1,19-dideoxybiladienes-ac(VII), themselves obtained from readily available pyrrolic intermediates 11. The cyclization of (VII) to corroles(VI) was effected by irradiation of alkaline solutions in presence of nickel or cobalt salts to yield the metal corroles in high yield. Alternatively in the absence of metal salts, one-electron oxidizing agents yield the free corroles 12. 376 TRANSITION MET AL COMPLEXES The corrole ring is aromatic as evidenced by the strong Soret band at ca. 400 nm and the n.m.r. signals of the three meso-protons at ca. -c 1. However, the formation of a divalent metal corrole(VIII) causes one hydrogen atom to migrate from nitrogen to a carbon of the chromophore (tentatively placed at clO in vnn and the resultant complex contains a non-aromatic chromophore. In this respect the metal corroles bear the same relation to porphin as does cyclopentadiene to benzene. Accordingly, IV) Me Me AcHO ~~H Me ~ ~ Et HN Me + <). NH CHO Me HN '\:: Et ~ Me Me ~ Et ~ Me Me lVIII) HBr ---+ (Vl) Me Me Me ~ '\: ~ ~ "\ Et NH HNNH HN_ Me t/ // ";:; + # Et Me Me !VII) Me /; / Et Me Me (VIllA) Formula VIllA. Structure of methylated copper corrole [R. Grigg, T. J. King and G. Shelton, Chem. Commun. 56 (1970)]. treatment of nickel corrole with bases readily formed a stable anion which showed a strong Soret band in its visible spectrum indicating a reversion to an aromatic structure. The anion reacted readily with alkyl halides to give crystalline mono-alkylated nickel corroles, the methyl derivative, for example, being obtained in 56 per celit yield 13. The alkyl derivatives are diamagnetic and the position of alkylation was deduced mainly from n.m.r. studies. The protons of the new methyl group were associated with a singlet at r 12·65, the marked upfield shift being the effect of the corrole ring current and providing good evidence that the methyl group was attached directly 377 A. W. JOHNSON to the nickel atom(IX)*. The three meso-protons of (IX) were represented by a singletat -r 0·6, suggesting an appreciable amount of aromatic character in the molecule, also indicated by a strong Soret band in the visible spectrum. Analysis of the product supported the assigned structure and also suggested, as did the infra-red spectrum, the Iack of any water or alcohol in the molecule as a sixth Iigand In order to accommodate an aromatic 18 n-electron Iigand, the molecule can be regarded as a derivative of nickel(IV). Analogous ethyl and n-propyl derivatives were prepared by similar methods. Possibly the most interesting of the reactions of the alkyl-nickel corroles is the rearrangement caused by heat When the methyl-nickel derivative(IX) was heated at 130° (reflux) for 90 min in chlorobenzene solution in an atmosphere of nitrogen, it was converted mainly (41·5 per cent) to an isomeric compound Likewise, the ethyl-nickel analogue was found to rearrange at 130° after 30 min, or, as with the n-propyl-nickel analogue, at 80° after 6 h. The methyl-nickel derivative, however, was unchanged after being heated at 80° for 8 h, showing that the thermal rearrangement of the I arger alkyl groups was more facile than that of methyl. The structure (X) of the rearrangement products was deduced from the n.m.r. spectra of the rearrangement products of three methyl-nickel derivatives of isomeric diethyl-hexamethylcorroles using the ethyl groups as markers. In the 8,12-diethyl series, the rearranged product (X; R = Me) contained a gem-dimethyl group and two almost identical ethyl groups, suggesting that the methyl-nickel substituent had not migrated to rings Bor C. In the 2,18-diethyl series, the rearranged product(XI) contained a gem-dimethyl group, indicating that the rearranged methyl group was not at C2 ; finally in the 3,17-diethyl series, the rearranged product(XII) was shown to contain a gem-ethyl-methyl group and thus the migration of methyl was from the metal to C3. In related experiments we have shown that the C2 position is not involved in the rearrangement process and that subsequent rearrangements of alkyl groups from C3 to C2 do not occur. The intramolecular nature of the alkyl migrations has been established as follows. First, thermolyses of alkyl-nickel 2,18-diethyl-3,7,8,12,13,17-hexamethylcorroles at 130°/90 min in presence of radical acceptors, e.g. hydroquinone or cumene, caused no decrease in the normal yield of rearranged product. Secondly, thermolysis of a mixture of ethyl-nickel 3,17-diethyl-2,7,8,12,13,18- hexamethylcorrole and methyl-nicke} 2,18-diethyl-3,7,8,12,13,17-hexamethylcorrole gave a product which showed no 'crossing' of the migrating alkyl group, i.e. the product contained a mixture of a 3,3-diethyl- and a 3,3- dimethyl-corrole derivative and none of the 3-ethyl-3-methyl analogue was observed. In another series, the nicke! corrole (VIII) was converted to its anion which was treated with allyl bromide. In this case none of the allyl-nickel derivative was obtained, but instead the main product (35 per cent) proved to be the 3-allyl-3-methyl nickel corrole, (X; R = allyl), i.e. the expected product from the thermolysis of the allyl-nickel compound. In addition neutral diallyl and triallylderivatives were obtained which appear to contain meso gem-diallyl groupings, e.g. (XIII). The position of the allyl group in the main products * Footnote added April 1970. X-Ray studies have revealed that the methylated copper and nickel corroles contain the additional methyl group as a substituent on the nitrogen atom of ring A of VIII, see previous page. 378 TRANSITION METAL COMPLEXES Me Me Me R Me Me Et ~ Et Me '\ Et Et ~ h Me Me Me Me (IX) (X) Me Et Me Me Me (XI) (XII) (e.g. X; R = allyl) of the reactions was again established by the technique of using ethyl groups as markers, e.g. nickel 2,18-diethyl-3,7,8,12,13,17-hexamethylcorrole gave a product containing a gem-allyl-methyl substituent, but the isomeric nickel 3,17-diethyl-2,7,8,12,13,18-hexamethylcorrole gave a product(XIV) containing a gem-allyl-ethyl grouping. Me (XIII) Me lXIV) The research on nickel corroles has been extended to other metals. We have shown, for example, that palladium(II) corrole readily forms a crystalline pyridinium salt which on methylation yields a mixture of the methyl-palladium derivative and its rearranged product, the palladium corrole containing a 3,3-dimethyl group. Rearrangement of the former to the latter required 379 A. W. JOHNSON heating at 180°/90 min, more stringent experimental conditions than are necessary in the nicke I series. The anion of copper(n) corrole on methylation gave a methyl-copper complex the structure of which is at present under investigation by x-ray crystallography. Howeve