Heme Biosynthesis in Methanosarcina barkeri via a Pathway Involving Two Methylation Reactions

Abstract

The methanogenic archaeon Methanosarcina barkeri synthesizes protoheme via precorrin-2, which is formed from uroporphyrinogen III in two consecutive methylation reactions utilizing S-adenosyl-l-methionine. The existence of this pathway, previously exclusively found in the sulfate-reducing δ-proteobacterium Desulfovibrio vulgaris, was demonstrated for M. barkeri via the incorporation of two methyl groups from methionine into protoheme.

Many archaea, both euryarchaeota and crenarchaeota, contain heme proteins such as cytochromes a, b, and c and catalase. But until now, information on how heme is synthesized in these microorganisms has remained scarce. Biochemical and genomic information indicates that heme biosynthesis starts from δ-aminolevulinic acid, which is synthesized from glutamate, and proceeds via porphobilinogen, preuroporphyrinogen, and uroporphyrinogen III (Uro III) as intermediates (6, 9, 17) (Fig. 1). At this stage the pathway appears to deviate from that known to be operative in eucarya and most bacteria. Most archaeal genomes lack gene homologs of hemE, hemF or -N, hemG or -Y, and hemH encoding the enzymes catalyzing the successive conversion of Uro III via coproporphyrinogen III, protoporphyrinogen IX, and protoporphyrin IX to protoheme (18). One exception is the genome of Thermoplasma volcanium in which gene homologs of hemE, hemN, and hemH are present (14, 18).

FIG. 1.

Pathway of protoheme biosynthesis in eucarya and most bacteria (18) and, highlighted in gray, the new pathway found in Desulfovibrio vulgaris (1, 10). δ-ALA, δ-aminolevulinic acid; SAM, S-adenosyl-l-methionine; SAH, S-adenosyl-l-homocysteine; Copro III, coproporphyrinogen III; hemB, porphobilinogen synthase gene; hemC, porphobilinogen deaminase gene; hemD, uroporphyrinogen III synthase gene; hemE, uroporphyrinogen III decarboxylase gene; hemF, O₂-dependent coproporphyrinogen III oxidase gene; hemN, O₂-independent coproporphyrinogen III oxidase gene; hemG or Y, protoporphyrinogen IX oxidase gene; hemH, ferrochelatase gene.

The genes hemE, hemF or -N, hemG or -Y, and hemH are also absent from the genome of the sulfate-reducing Desulfovibrio vulgaris (8) despite the fact that this sulfate-reducing δ-proteobacterium is known to contain cytochromes b and c (16). The synthesis of protoheme in D. vulgaris has been shown to involve precorrin-2 (dihydrosirohydrochlorin) as an intermediate (1, 10). Precorrin-2, which is a 2,7-dimethyl derivative of Uro III, is formed from Uro III in two consecutive methylation reactions with S-adenosyl-l-methionine as methyl donor (21) (Fig. 1). Precorrin-2 is also a precursor for the biosynthesis of vitamin B₁₂, coenzyme F₄₃₀, siroheme and heme d₁ (2). The intermediates and proteins involved in precorrin-2 conversion into protoheme in D. vulgaris have not yet been identified (10). In cell extracts of this organism precorrin-2 was converted under strictly anoxic conditions into 12,18-didecarboxyprecorrin-2, indicating that the first step is a decarboxylation reaction (10). In other anaerobic bacteria (Chlorobium vibrioforme) protoheme biosynthesis was found not to require the incorporation of methyl groups from methionine (3).

The late heme genes are also missing from the genomes of Methanosarcina barkeri, Methanosarcina acetivorans (7), and Methanosarcina mazei (4). M. barkeri is a strictly anaerobic methanogenic archaeon that is able to grow on H₂ and CO₂, acetate, methylamines, or methanol as sole carbon and energy sources. It is a member of the order Methanosarcinales, the members of which are all known to contain cytochromes b and c and other proteins with protoheme-derived prosthetic groups (12, 15). The organism also contains the nickel tetrapyrrole coenzyme F₄₃₀ (5) and the cobalt tetrapyrrole hydroxybenzimidazolyl cobamide, which is a vitamin B₁₂ homolog (20). These tetrapyrroles are involved in different steps of methanogenesis. Here we demonstrate that the pathway of protoheme biosynthesis operative in D. vulgaris is most likely also operative in the methanogenic archaeon M. barkeri.

For this purpose, it was considered that in M. barkeri protoheme is synthesized from precorrin-2 as previously shown for D. vulgaris (10). Therefore, the methanogen was grown on methanol in the absence and presence of l-(methyl-d₃)methionine (Sigma, Munich, Germany). M. barkeri strain Fusaro (DSM 804) was obtained from the Deutsche Stammsammlung von Mikroorganismen und Zellkulturen (Braunschweig, Germany) and cultivated with 250 mM methanol as the sole carbon and energy source as described previously (13). The labeled amino acid l-(methyl-d₃)methionine was added to the medium to a final concentration of either 2 mM, 5 mM, or 10 mM. Growth was initiated with a 10% inoculum (grown in absence of methionine) and monitored photometrically at 578 nm. At the end of the exponential growth phase, the cells were harvested anaerobically by centrifugation at 10,000 × g for 30 min at 4°C. Noncovalently bound protoheme was extracted from M. barkeri by a modification of the method described previously (23). One gram of frozen cells was ground under liquid nitrogen, transferred to a test tube, and suspended in 10 ml of cold 90% aqueous acetone solution containing 600 mM HCl, followed by centrifugation at 10,000 × g for 30 min at 4°C. Four milliliters of peroxide-free diethyl ether was added to the supernatant. After mixing, 12.5 ml of cold water was added. The heme-containing upper ether phase was evaporated to dryness. For matrix-assisted laser desorption ionization-time of flight mass spectrometry analysis, the dried fraction was resuspended in 100 μl aqueous 20% (vol/vol) acetonitrile/1% (vol/vol) trifluoroacetic acid. Coenzyme F₄₃₀ is known to contain two methyl groups (11), and B₁₂ is known to contain seven methyl groups, all derived from S-adenosyl-l-methionine (22). These cofactors were analyzed as controls. The mass spectra were collected in the reflector positive-ion mode. Samples were measured after an appropriate dilution of the sample with saturated α-cyano-4-hydroxy cinnamic acid in aqueous 70% (vol/vol) acetonitrile/0.1% (vol/vol) trifluoroacetic acid. The spectra were determined with an Ultraflex instrument from Bruker (Bremen, Germany).

Mass spectra of commercially obtained protoheme (hemin from Fluka, Seelze, Germany), of heme extracted from unlabeled cells, and of heme extracted from cells grown in the presence of 2 mM, 5 mM, or 10 mM labeled l-(methyl-d₃)methionine are shown in Fig. 2.

FIG. 2.

Matrix-assisted laser desorption ionization-time of flight mass spectra of protoheme (hemin from Fluka) (A), heme extracted from M. barkeri cells grown in the absence of l-(methyl-d₃)methionine (B), and (C to E) heme extracted from M. barkeri cells grown in the presence of 2 mM (C), 5 mM (D) and 10 mM (E) l-(methyl-d₃)methionine.

The mass spectra of commercially obtained protoheme (Fig. 2A) and of heme extracted from the unlabeled cells (Fig. 2B) were found to be almost identical. The highest peak, representing a mass of 616 Da, corresponds to the mass of protoheme composed of only ¹²C, ¹⁴N, and ⁵⁶Fe. The other peaks, with masses of 614, 615, 617, 618, 619, and 620 Da, reflect the natural abundance of ¹³C (1.1%), ¹⁵N (0.37%), ⁵⁴Fe (5.82%), ⁵⁶Fe (91.72%), ⁵⁷Fe (2.1%), and ⁵⁸Fe (0.3%) in this molecule. The natural isotopic distribution in protoheme, F₄₃₀, and B₁₂ was calculated by use of the isotope pattern calculator provided by the University of Sheffield at the ChemPuter site (http://winter.group.shef.ac.uk/chemputer/isotopes.html). The fraction of protoheme, F₄₃₀, or B₁₂ labeled with methyl-d₃ (labeling efficiency) was determined by solving the binomial equation (a + b)ⁿ for an n value of 2 (protoheme or F₄₃₀) and an n value of 7 (B₁₂) using the Solver facilities provided by Excel 97 as described by Selmer et al. (19). The relative peak heights agreed well with those calculated for this isotope composition. The mass spectroscopic analysis thus yielded reliable results.

In Fig. 2C the mass spectrum of the heme extracted from M. barkeri cells grown in the presence of 2 mM l-(methyl-d₃)methionine is shown. The spectrum revealed the presence not only of 37% unlabeled protoheme with a mass of 616 Da but also of 44% protoheme with a mass of 619 Da and 19% protoheme with a mass of 622 Da. This observation is consistent with one and two methyl groups, respectively, of protoheme being derived from the methyl group of methionine. When the l-(methyl-d₃)methionine concentration in the growth medium was 5 mM, the percentage of protoheme with a mass of 622 Da (two methyl groups from methionine) was 33%. When the concentration was 10 mM, the percentage was 56% (Fig. 2D and E). A similar methionine concentration dependence has been reported for coenzyme F₄₃₀ labeling in Methanothermobacter marburgensis (11, 19).

As an internal control, the incorporation of the methyl group of l-(methyl-d₃)methionine into F₄₃₀ and B₁₂ in the same experiment was determined via mass spectrometry. F₄₃₀ contains two methyl groups, and vitamin B₁₂ contains seven methyl groups derived from methionine via S-adenosyl-l-methionine. Labeling efficiency followed the same concentration dependence as found for protoheme (Fig. 3). This clearly indicates that the mass increase of 6 Da observed in the case of protoheme is due to the incorporation of two intact methyl groups derived from l-(methyl-d₃)methionine during protoheme biosynthesis in M. barkeri.

FIG. 3.

Methyl group of methionine incorporated into heme (•), F₄₃₀ (▵), and B₁₂ (□) after growth of M. barkeri in the presence of l-(methyl-d₃)methionine at different concentrations. The labeling efficiency represents the methyl-d₃-labeled fraction of the total methyl content of the sample.