The biosynthesis of the tetrapyrrole framework has been investigated in the sulfate-reducing bacterium Desulfovibrio vulgaris Hildenborough by characterization of the enzymes required for the transformation of aminolaevulinic acid into sirohydrochlorin. PBG (porphobilinogen) synthase (HemB) was found to be a zinc-dependent enzyme that exists in its native state as a homohexamer. PBG deaminase (HemC) was shown to contain the dipyrromethane cofactor. Uroporphyrinogen III synthase is found fused with a uroporphyrinogen III methyltransferase (HemD-CobA). Both activities could be demonstrated in this amalgamated protein and the individual enzyme activities were separated by dissecting the relevant gene to allow the production of two distinct proteins. A gene annotated in the genome as a bifunctional precorrin-2 dehydrogenase/sirohydrochlorin ferrochelatase was in fact shown to act only as a dehydrogenase and is simply capable of synthesizing sirohydrochlorin rather than sirohaem. Genome analysis also reveals a lack of any uroporphyrinogen III decarboxylase, an enzyme necessary for the classical route to haem synthesis. However, the genome does encode some predicted haem d1 biosynthetic enzymes even though the bacterium does not contain the cd1 nitrite reductase. We suggest that sirohydrochlorin acts as a substrate for haem synthesis using a novel pathway that involves homologues of the d1 biogenesis system. This explains why the uroporphyrinogen III synthase is found fused with the methyltransferase, bypassing the need for uroporphyrinogen III decarboxylase activity.

INTRODUCTION

SRB (sulfate-reducing bacteria) are environmentally important micro-organisms that are found in a wide range of habitats but are reliant on being able to reduce sulfate in a dissimilar manner to produce sulfide. To assist them in their metabolic challenge SRB employ a rich repertoire of redox groups including porphyrins and related modified tetrapyrroles. Genomic analysis suggests that SRB are able to make at least three types of modified tetrapyrroles, including haem, sirohaem and adenosylcobalamin, the coenzyme form of vitamin B12.

Modified tetrapyrroles have a similar structural architecture reflecting the fact that they are all derived from a common macromolecular intermediate, uroporphyrinogen III. Thus all modified tetrapyrroles are made along a branched biosynthetic pathway, as outlined in Figure 1. The pathway starts with the generation of ALA (aminolaevulinic acid), the first common intermediate of tetrapyrrole biosynthesis, which is synthesized either from succinyl-CoA and glycine (for some proteobacteria, fungi, yeast and animals) or from a tRNA-bound glutamate (in plants, algae, archaea and bacteria) [1,2]. The next step involves the asymmetric condensation of two ALA molecules to form the monopyrrole PBG (porphobilinogen), in a Knorr-type condensation reaction catalysed by the enzyme PBG synthase (also called ALA dehydratase), which is encoded by the hemB gene [3]. PBG synthases are distinguished according to their requirement in the active site for zinc or magnesium atoms and the need for magnesium in an allosteric binding site. The only known PBG synthases where these metal-binding sites are absent are from the bacterial genus Rhodobacter [4]. Rhodobacter capsulatus PBG synthase was shown not to require any metal ion for catalysis, having the added peculiarity of forming a hexameric structure [5], unlike most other studied PBG synthases that form octamers (e.g. [6]).

Four molecules of the pyrrole PBG are subsequently polymerized into the linear tetrapyrrole HMB (hydroxymethylbilane) by the action of PBG deaminase, which is encoded by hemC. The enzyme forms stable enzyme–intermediate complexes with one, two and three PBG-derived pyrrole molecules. A feature of this enzyme is the presence of a covalently bound dipyrromethane cofactor where the growing poly-pyrrole chain is assembled [7,8]. The HMB intermediate is further modified and cyclized by uroporphyrinogen III synthase (encoded by the hemD gene) to yield uroporphyrinogen III [9]. In the absence of uroporphyrinogen III synthase, HMB cyclizes spontaneously to yield the uroporphyrinogen I isomer [9]. The synthase is able to invert the terminal pyrrole unit of HMB, ring D, and cyclize the macromolecule to give the asymmetric and only biologically relevant type III isomer. In many organisms, uroporphyrinogen III represents the first branch point in the pathway, where the action of enzymes such as HemE, F (N) and G results in the formation of protoporphyrin IX, a precursor of modified tetrapyrroles such as haem and chlorophyll. Alternatively, uroporphyrinogen III can undergo a SAM (S-adenosylmethionine)-dependent-transmethylation at positions 2 and 7 (by CysG, Met1p or CobA depending on the organism) to generate precorrin-2, a highly unstable yellow dipyrrocorphin [10,11]. Precorrin-2 is further modified by a dehydrogenase (CysG, Met8p or SirC enzymes) to yield sirohydrochlorin. This isobacteriochlorin can be chelated with either ferrous iron to form sirohaem by the action of sirohydrochlorin ferrochelatase (CysG, Met8p or SirB enzymes) or with cobalt to generate cobalt–sirohydrochlorin, an intermediate that is directed along the cobalamin biosynthetic pathway. Thus, as with uroporphyrinogen, sirohydrochlorin also represents an important branch point in the biosynthesis of modified tetrapyrroles.

Tetrapyrrole metabolic pathway

Figure 1
Tetrapyrrole metabolic pathway

Adapted from [46] with permission. © (2000) Birkhäuser Basel.

Figure 1
Tetrapyrrole metabolic pathway

Adapted from [46] with permission. © (2000) Birkhäuser Basel.

In Escherichia coli and Salmonella enterica, the transformation of uroporphyrinogen III into sirohaem is performed by a single multifunctional enzyme, sirohaem synthase (CysG). The SAM-dependent methyltransferase activity is associated with the C-terminal domain of GysG (CysGA) and the NAD+-dependent dehydrogenation and ferrochelation activities are linked to the N-terminal domain (CysGB) [12,13]. In yeast, two independent enzymes, Met1p (SAM-dependent methyltransferase) and Met8p (NAD+-dependent dehydrogenase/ferrochelatase), are required for the sirohaem synthesis [14]. A third alternative is found in Bacillus megaterium, where the three reactions are catalysed independently, by SirA (SAM-dependent uroporphyrinogen III methyltransferase), SirC (NAD+-dependent precorrin-2 dehydrogenase) and SirB (sirohydrochlorin ferrochelatase) [15].

Fused enzyme systems are not just restricted to the sirohaem branch of the pathway. In some bacterial species, such as Selenomonas ruminantium [16], Lactobacillus reuteri [17] and Clostridium josui [18], the uroporphyrinogen III methyltransferase and uroporphyrinogen III synthase activities are performed by one single protein, since the cobA and hemD genes are linked together. The biochemical logic and metabolic importance of such fusions are not fully understood.

The sulfate reducers of the Desulfovibrio genus have many different haem proteins [1921] that play essential roles in the metabolism of the organism. Although tetrapyrrole biosynthesis has not been studied in any molecular detail in the SRB, earlier research on haem synthesis in Desulfovibrio vulgaris had uncovered one of the most interesting discoveries in tetrapyrrole biosynthesis of recent times as it demonstrated that haem is made of a novel route utilizing precorrin-2 or sirohydrochlorin as an intermediate [22]. This new pathway is also likely to be the major route for haem synthesis in the archaea. A follow-up study in D. vulgaris identified 12,18-didecarboxysirohydrochlorin (or its reduced form) as a possible intermediate but no specific enzymes/genes were linked with the metabolic activity. The authors also suggested that this novel pathway proceeded via coproporphyrinogen III, but this was not rigorously shown [23]. Despite the novelty of this intriguing pathway, no genomic or biochemical investigation into tetrapyrrole synthesis in D. vulgaris has been undertaken even though the genome has been fully sequenced. In the present study, we searched the D. vulgaris Hildenborough genome [24] and noted that it contains genes encoding enzymes for the transformation of ALA into sirohydrochlorin via uroporphyrinogen III. Consistent with the presence of a novel haem biosynthetic pathway, the organism appears to be missing most of the genes that encode the classical enzymes required for the transformation of uroporphyrinogen III into haem. In the present study, we report the characterization of the early enzymes of tetrapyrrole assembly and modification in D. vulgaris and suggest enzymes that may be involved in the novel haem pathway.

MATERIALS AND METHODS

Cloning and expression of recombinant proteins

D. vulgaris Hildenborough genes encoding putative PBG synthase (hemB), PBG deaminase (hemC), uroporphyrinogen III methyltransferase/synthase (cobA/hemD) and sirohaem synthase (cysGB) were PCR amplified from genomic DNA and specific oligonucleotides (Supplementary Table S1 at http://www.BiochemJ.org/bj/420/bj4200317add.htm). Two additional oligonucleotides were designed for amplification of truncated forms of cobA/hemD to generate hemD or cobA encoding regions. The individual DNA fragments were cloned into pET vectors (Novagen) (Supplementary Table S1), which produced proteins with a His6 tag tail in the N- or C-terminus, depending on the vector utilized. The genes cobA/hemD, cobA and cysGB were also subcloned into pETac [25] for use in complementation studies. The integrity of the gene sequence for all cloned fragments was confirmed by DNA sequencing.

Plasmids pET-28a(+)cysGB and pET-22b(+)cobA/hemD were transformed into E. coli BL21Gold(DE3) cells (Stratagene) and plasmids pET-14b(+)hemB, pET-14b(+)hemC, pET-14b(+)-cobA and pET-14b(+)hemD into E. coli BL21Star(DE3)(pLysS) cells (Invitrogen). In order to overexpress the proteins, the E. coli cells were grown at 37 °C, in LB (Luria–Bertani) medium with appropriate antibiotics, until a D600 (attenuance at 600 nm) between 0.5 and 0.8 was reached. Induction of the genes was achieved by addition of 200–400 μM of IPTG (isopropyl-β-D-thiogalactoside) and by growing the cells for 4 h at 37 °C, for cobA/hemD and cysGB genes, or overnight at 19 °C for hemB, hemC, cobA and hemD genes. The truncated forms of cobA/hemD are designated cobAd and hemDd (where ‘d’ stands for domain).

Protein purification

Cells expressing HemB, HemC, CobA/HemD, CobAd (Met1–His246), HemDd (Lys247–Lys503) and CysGB proteins were harvested and resuspended in 20 mM Tris/HCl buffer (pH 8) containing 500 mM NaCl and 10 mM imidazole (buffer A); for the purification of CysGB all buffers contained 5% (v/v) glycerol. All cells were lysed by sonication and centrifuged for 20 min at ∼30000 g. The soluble fraction was loaded on to an NiCl2 charged chelating Sepharose resin (4 ml) (GE Healthcare) equilibrated with buffer A. After washing with 5–10 vol. of buffer A containing 50 and 100 mM imidazole, all proteins were eluted with 2–5 vol. of buffer A with 500 mM imidazole, except CysGB, which was eluted with 100 mM imidazole. Protein fractions were pooled together after detection with the Bio-Rad protein assay (Bio-Rad), concentrated with Millipore Ultra Centrifugal Filters with a 10 kDa membrane and buffer exchanged, by means of a PD10 column (GE Healthcare), into buffer B (50 mM Tris/HCl, pH 8), with the exception of CobAd protein, which was buffer exchanged into buffer B containing 500 mM NaCl and 200 mM imidazole. The purity of the proteins was assessed by SDS/PAGE [26] and the protein concentration was determined by the bicinchoninic acid method [27] using protein standards from Sigma. The native molecular mass of the various proteins was determined by gel-filtration chromatography on either a Superdex 200 or 300 HR 10/30 column (GE Healthcare), previously equilibrated with 20 mM Tris/HCl (pH 7.5).

The detection of the dipyrromethane cofactor of D. vulgaris PBG deaminase was determined by incubating 0.5 ml of the enzyme (0.2 mg) with an equivalent volume of Ehrlich's reagent [28] and by following the UV–visible spectral changes over 20 min.

Enzymatic assays

The various activity measurements were performed in a Hewlett–Packard 8452A photodiode array spectrophotometer, a Shimadzu UV-1700 spectrophotometer or a BMG Labtech Flurostar Optima plate reader. All activities were assayed in buffer B, with the exception of PBG synthase, which was assayed in 100 mM Tris/HCl (pH 8).

PBG synthase

HemB (0.5–5 μg) was pre-incubated for 5 min at 37 °C, and the reaction started by the addition of ALA (5 mM), in a final volume of 0.5 ml. The reaction mixture was incubated for a further 5 min at 37 °C. The reaction was stopped by the addition of 500 μl of 10% (w/v) trichloroacetic acid in 100 mM HgCl2. The mixture was centrifuged for 5 min at 9700 g and an equal volume of a modified Ehrlich's reagent [28] was added to the supernatant. After a 15 min incubation, at room temperature (25 °C), the absorbance at 555 nm was measured (ε555 6.02×104 M−1·cm−1) [28].

PBG deaminase

The HemC activity was determined essentially as described in [29], using 10–30 μg of enzyme and 200 nmol of PBG, and ε405=5.48×105 M−1·cm−1. One unit of enzyme is the amount of enzyme necessary to catalyse the utilization of 1 μmol of PBG in 1 h.

Uroporphyrinogen III methyltransferase

Uroporphyrinogen III methyltransferase activity was followed by coupling it to precorrin-2 dehydrogenase (SirC) activity, allowing the transformation of uroporphyrinogen III into sirohydrochlorin (ε376 2.4×105 M−1·cm−1) to be monitored. The substrate, uroporphyrinogen III, was generated anaerobically by incubating HemC and HemD from B. megaterium with 2 mg of PBG. The reaction mixture was prepared in an anaerobic chamber (Belle Technology, <2 p.p.m. O2), in a final volume of 250 μl with NAD (100 μM), uroporphyrinogen III (2 μM), B. megaterium SirC (20 μg) and with different amounts of D. vulgaris bifunctional CobA/HemD (0.5–10 μg); the reaction was started by the addition of SAM (100 μM).

Sirohaem synthase (CysGB)

Precorrin-2 dehydrogenase and sirohydrochlorin chelatase activities were assayed anaerobically following the formation of sirohydrochlorin or cobalt–sirohydrochlorin and sirohaem respectively (ε376=2.4×105 M−1·cm−1) [30]. The substrates, precorrin-2 and sirohydrochlorin, were generated anaerobically as described previously [30]. The dehydrogenase assay was performed in a 1 ml reaction, at 30 °C, with precorrin-2 (2–3 μM), NAD+ (800 μM) and different amounts of D. vulgaris CysGB (0.5–10 μg). The chelatase activity was measured with 4.2 μM of sirohydrochlorin, 20 μM of Co2+ or Fe2+ and varying the amount of enzyme (10–50 μg).

in vitro generation of precorrin-2 with D. vulgaris enzymes

HemB, HemC and CobA/HemD enzymes of D. vulgaris were tested simultaneously and individually in a linked assay for the in vitro generation of precorrin-2 [31]. The combination of enzymes and substrates used for each assay is described in Table 1. All the reaction mixtures were performed in 50 mM Tris/HCl (pH 8) in a final volume of 3 ml and contained between 0.05 and 5 mg of each enzyme, SAM (1 mg) and ALA (0.5 mg) or PBG (0.2 mg). The reaction was incubated for 2 h at room temperature, and after colour development, UV–visible spectra were recorded. B. megaterium SirC (0.1 mg) was added to the assays to generate sirohydrochlorin, as a further confirmation of precorrin-2 production (Table 1).

Table 1
Enzymes and substrates utilized in the linked in vitro assays for anaerobic generation of precorrin-2

*Ps. denitrificans CobA and B. megaterium HemC and HemD were used in the assays where D. vulgaris enzymes were tested individually.

D. vulgaris enzyme Other enzymes and substrates present in the assay* Precorrin-2 generation 
HemB B. megaterium HemC and HemD, Ps. denitrificans CobA, ALA, SAM Yes 
HemC B. megaterium HemD, Ps. denitrificans CobA, PBG, SAM Yes 
HemDd B. megaterium HemC, Ps. denitrificans CobA, PBG, SAM Yes 
F446S-HemDd B. megaterium HemC, Ps. denitrificans CobA, PBG, SAM No 
CobAd B. megaterium HemC and HemD, PBG, SAM Yes 
CobA/HemD B. megaterium HemC, PBG, SAM Yes 
HemB, HemC,CobA/HemD ALA, SAM Yes 
None B. megaterium HemC and HemD, Ps. denitrificans CobA, PBG, SAM (positive control) Yes 
D. vulgaris enzyme Other enzymes and substrates present in the assay* Precorrin-2 generation 
HemB B. megaterium HemC and HemD, Ps. denitrificans CobA, ALA, SAM Yes 
HemC B. megaterium HemD, Ps. denitrificans CobA, PBG, SAM Yes 
HemDd B. megaterium HemC, Ps. denitrificans CobA, PBG, SAM Yes 
F446S-HemDd B. megaterium HemC, Ps. denitrificans CobA, PBG, SAM No 
CobAd B. megaterium HemC and HemD, PBG, SAM Yes 
CobA/HemD B. megaterium HemC, PBG, SAM Yes 
HemB, HemC,CobA/HemD ALA, SAM Yes 
None B. megaterium HemC and HemD, Ps. denitrificans CobA, PBG, SAM (positive control) Yes 

Complementation of E. coli cysG and E. coli hemD mutant strains

The pETac plasmids containing the genes cobA/hemD, cobA and cysGB were transformed into E. coli cysG mutant strain 302Δa (cysteine auxotrophic) (Table 2). E. coli cysG transformants were selected on LB plates, with 100 μg/ml ampicillin and 35 μg/ml chloramphenicol, and the mutant phenotype rescue was tested on minimal medium plates with the appropriate antibiotics, in the absence and presence of cysteine. The plasmids pET14b(+)cobA/hemD and pET14b(+)hemD were introduced in E. coli SASZ31 (a hemD gene mutant strain) and selected on LB plates, with 100 μg/ml ampicillin by incubation at 37 °C, for approx. 20 h. The mutant strains were re-streaked on LB plates and incubated at 37 °C for a shorter time period (∼12 h) to evaluate cell growth.

Table 2
Strains and plasmids used in complementation studies together with D. vulgaris enzymes

*CobA activity tested with CobA/HemD protein and CobAd. †HemD activity tested in the presence of CobA/HemD protein and HemDd. ‡CysGB protein was tested for precorrin-2 dehydrogenase and sirohydrochlorin ferrochelatase activities. §CysGB protein was tested for sirohydrochlorin ferrochelatase activity.

D. vulgaris enzyme Strain Plasmids Plasmid description References 
CobA/HemD* E. coli 302Δa pETac-cobA/hemD D. vulgaris cobA/hemD cloned in pETac The present study 
  pCIQ-met8p Sa. met8p cloned in pCIQ The present study 
CobAdE. coli 302Δa pETac-cobA D. vulgaris cobA cloned in pETac The present study 
  pCIQ-met8p Sa. cerevisiae met8p cloned in pCIQ The present study 
 E. coli 302Δa (positive control) pETac-PdcobA Ps. denitrificans cobA cloned in pETac The present study 
  pCIQ-met8p Sa. cerevisiae met8p cloned in pCIQ The present study 
CobA/HemD† E. coli SASZ31 pET14b(+)-cobA/hemD D. vulgaris cobA/hemD cloned in pET14b(+) The present study 
HemDd† E. coli SASZ31 pET14b(+)hemD D. vulgaris hemD cloned in pET14b(+) The present study 
 E. coli SASZ31 (positive control) pET14b(+)EchemD E. coli hemD cloned in pET14b(+) The present study 
CysGB‡ E. coli 302Δa pETac-cysGB D. vulgaris cysGB cloned in pETac The present study 
  pCIQ-PdcobA Ps. denitrificans cobA cloned in pCIQ [37
 E. coli 302Δa (positive control) pETac-met8p Sa. cerevisiae met8p cloned in pETac [30
  pCIQ-PdcobA Ps. denitrificans cobA cloned in pCIQ [37
CysGB§ E. coli 302Δa pETac-cysGB D. vulgaris cysGB cloned in pETac The present study 
  pCIQ-sirCcobA Methanothermobacter thermoautotrophicus sirC and Ps. denitrificans cobA cloned in pETac [37
 E. coli 302Δa (positive control) pETac-met8p Sa. cerevisiae met8p cloned in pETac [30,37
  pCIQ-sirCcobA M. thermoautotrophicus sirC and Ps. denitrificans cobA cloned in pCIQ  
D. vulgaris enzyme Strain Plasmids Plasmid description References 
CobA/HemD* E. coli 302Δa pETac-cobA/hemD D. vulgaris cobA/hemD cloned in pETac The present study 
  pCIQ-met8p Sa. met8p cloned in pCIQ The present study 
CobAdE. coli 302Δa pETac-cobA D. vulgaris cobA cloned in pETac The present study 
  pCIQ-met8p Sa. cerevisiae met8p cloned in pCIQ The present study 
 E. coli 302Δa (positive control) pETac-PdcobA Ps. denitrificans cobA cloned in pETac The present study 
  pCIQ-met8p Sa. cerevisiae met8p cloned in pCIQ The present study 
CobA/HemD† E. coli SASZ31 pET14b(+)-cobA/hemD D. vulgaris cobA/hemD cloned in pET14b(+) The present study 
HemDd† E. coli SASZ31 pET14b(+)hemD D. vulgaris hemD cloned in pET14b(+) The present study 
 E. coli SASZ31 (positive control) pET14b(+)EchemD E. coli hemD cloned in pET14b(+) The present study 
CysGB‡ E. coli 302Δa pETac-cysGB D. vulgaris cysGB cloned in pETac The present study 
  pCIQ-PdcobA Ps. denitrificans cobA cloned in pCIQ [37
 E. coli 302Δa (positive control) pETac-met8p Sa. cerevisiae met8p cloned in pETac [30
  pCIQ-PdcobA Ps. denitrificans cobA cloned in pCIQ [37
CysGB§ E. coli 302Δa pETac-cysGB D. vulgaris cysGB cloned in pETac The present study 
  pCIQ-sirCcobA Methanothermobacter thermoautotrophicus sirC and Ps. denitrificans cobA cloned in pETac [37
 E. coli 302Δa (positive control) pETac-met8p Sa. cerevisiae met8p cloned in pETac [30,37
  pCIQ-sirCcobA M. thermoautotrophicus sirC and Ps. denitrificans cobA cloned in pCIQ  

RESULTS AND DISCUSSION

The genome of D. vulgaris [25] was searched using BLAST of NCBI (National Center for Biotechnology Information; http://blast.ncbi.nlm.nih.gov/Blast.cgi) for homologues of known bacterial enzymes involved in the early steps of tetrapyrrole biosynthesis and modification, namely HemB, HemC, CobA, HemD and CysG. Four gene-derived amino acid sequences that have a high degree of sequence identity with these enzymes from different organisms, such as E. coli and B. subtilis, were identified (Table 3). To study the D. vulgaris proteins the corresponding genes were cloned, the recombinant proteins were produced and biochemically characterized.

Table 3
Amino acid sequence identities and similarities of D. vulgaris enzymes with orthologues

NCBI accession numbers in parentheses: HemB proteins: D. vulgaris (AAS95336), E. coli (NP_414903), Br. japonicum (P45622), Ps. aeruginosa (AAG08628), Pi. sativum (AAA33640); HemC proteins: E. coli (AAA67601), S. enterica (CAD09382), R. capsulatus (AAG50298), C. josui, (BAA05861); CobA/HemD proteins: Selenomonas rumantium (AAK00606), L. reuteri (AAX14527), C. josui (BAA05862), Listeria innocua (NP_470501); CobA proteins: Ps. denitrificans (AAA25773), B. megaterium (AAA22317), Paracoccus denitrificans (AAA93119), M. thermoautotrophicus (NP_275310); HemD proteins: E. coli (NP_418248), B. megaterium (CAD48147), Pa. denitrificans (YP_915785), S. enterica (AAL22782); E. coli CysG and CysG N-terminal (NP_417827 and residues 1–223 of NP_417827), Sa. cerevisiae MET8 (NP_009772), B. megaterium SirC (CAD48923). The amino acid sequence alignments were performed in Clustal W2 [45].

 Identity (%) Similarity (%) 
 D. vulgaris HemB  
E. coli 47 66 
Br. japonicum 42 60 
Ps. aeruginosa 44 62 
Pi. sativum 41 54 
 D. vulgaris HemC  
E. coli 52 63 
S. enterica 51 63 
R. capsulatus 40 54 
C. josui 33 50 
 D. vulgaris CobA/HemD  
Se. ruminantium 52 63 
L. reuteri 29 47 
C. josui 36 56 
Li. innocua 35 53 
 D. vulgaris CobAd  
Ps. denitrificans 40 56 
B. megaterium 48 67 
Pa. denitrificans 34 47 
M. thermoautotrophicus 42 62 
 D. vulgaris HemDd  
E. coli 23 33 
S. enterica 22 33 
B. megaterium 27 47 
Pa. denitrificans 16 27 
 D. vulgaris SirC  
E. coli CysG 12 20 
E. coli CysG N-terminal 26 42 
Sa. cerevisiae MET8 17 31 
B. megaterium SirC 21 41 
 Identity (%) Similarity (%) 
 D. vulgaris HemB  
E. coli 47 66 
Br. japonicum 42 60 
Ps. aeruginosa 44 62 
Pi. sativum 41 54 
 D. vulgaris HemC  
E. coli 52 63 
S. enterica 51 63 
R. capsulatus 40 54 
C. josui 33 50 
 D. vulgaris CobA/HemD  
Se. ruminantium 52 63 
L. reuteri 29 47 
C. josui 36 56 
Li. innocua 35 53 
 D. vulgaris CobAd  
Ps. denitrificans 40 56 
B. megaterium 48 67 
Pa. denitrificans 34 47 
M. thermoautotrophicus 42 62 
 D. vulgaris HemDd  
E. coli 23 33 
S. enterica 22 33 
B. megaterium 27 47 
Pa. denitrificans 16 27 
 D. vulgaris SirC  
E. coli CysG 12 20 
E. coli CysG N-terminal 26 42 
Sa. cerevisiae MET8 17 31 
B. megaterium SirC 21 41 

D. vulgaris PBG synthase is active as a hexamer

A comparison of the gene-deduced amino acid sequence of D. vulgaris PBG synthase with the PBG synthases of other organisms showed that D. vulgaris HemB contains the conserved cysteine-rich sequence for Zn2+ binding (Cys121, Cys123 and Cys131, D. vulgaris HemB residue numbering), the determinant residues for allosteric magnesium binding (Arg11 and Glu238, D. vulgaris HemB residue numbering) [4] and the amino acid residues known to be involved in the binding of the substrate and reaction mechanism [3234] (Figure 2). The recombinant D. vulgaris HemB migrates by SDS/PAGE analysis with an apparent molecular mass of ∼36 kDa, whereas gel-filtration studies indicated that the enzyme has a native mass of 226 kDa. This suggests that D. vulgaris exists as a homohexamer.

Amino acid sequence alignment of PBG synthases from D. vulgaris (AAS95336), E. coli (NP_414903), Bradyrhizobium japonicum (P45622), Pseudomonas aeruginosa (AAG08628) and Pisum sativum (AAA33640) (NCBI accession numbers in parentheses)

Figure 2
Amino acid sequence alignment of PBG synthases from D. vulgaris (AAS95336), E. coli (NP_414903), Bradyrhizobium japonicum (P45622), Pseudomonas aeruginosa (AAG08628) and Pisum sativum (AAA33640) (NCBI accession numbers in parentheses)

Strictly conserved residues are shaded. The symbols represent functionally relevant residues in the E. coli enzyme [32,33], which are also conserved in D. vulgaris PBG synthase: (#) cysteine residues that bind zinc; (*) residues involved in allosteric magnesium binding; (“) arginine residues involved in binding A-side ALA; (−) residues of the lid region that covers the active site; (&”) lysine residue required for correct function of the enzyme; (&) lysine residue that forms the Schiff base with P-side ALA; ($) residues involved in binding of P-side ALA.

Figure 2
Amino acid sequence alignment of PBG synthases from D. vulgaris (AAS95336), E. coli (NP_414903), Bradyrhizobium japonicum (P45622), Pseudomonas aeruginosa (AAG08628) and Pisum sativum (AAA33640) (NCBI accession numbers in parentheses)

Strictly conserved residues are shaded. The symbols represent functionally relevant residues in the E. coli enzyme [32,33], which are also conserved in D. vulgaris PBG synthase: (#) cysteine residues that bind zinc; (*) residues involved in allosteric magnesium binding; (“) arginine residues involved in binding A-side ALA; (−) residues of the lid region that covers the active site; (&”) lysine residue required for correct function of the enzyme; (&) lysine residue that forms the Schiff base with P-side ALA; ($) residues involved in binding of P-side ALA.

The specific activity of D. vulgaris HemB was determined in the presence of zinc to be 45 μmol of PBG·h−1·mg−1 with a KM value for ALA of 0.05 mM (Table 4). This activity is within the values measured for other bacterial PBG synthases such as E. coli [6], also assayed in the presence of zinc, and has a similar KM.

Table 4
Kinetic parameters of D. vulgaris enzymes studied in the present study
D. vulgaris enzyme Specific activity KM (substrate) 
HemB 45 μmol·h−1·mg−1 0.05 mM (ALA) 
HemC 20 μmol·h−1·mg−1 214 μM (PBG) 
CobA/HemD 3 nmol·min−1·mg−1 0.4 μM (uroporphyrinogen III) 
SirC 700 nmol·min−1·mg−1 70 μM (NAD+
D. vulgaris enzyme Specific activity KM (substrate) 
HemB 45 μmol·h−1·mg−1 0.05 mM (ALA) 
HemC 20 μmol·h−1·mg−1 214 μM (PBG) 
CobA/HemD 3 nmol·min−1·mg−1 0.4 μM (uroporphyrinogen III) 
SirC 700 nmol·min−1·mg−1 70 μM (NAD+

PBG deaminase of D. vulgaris contains the dipyrromethane cofactor in its active site

The purified recombinant D. vulgaris HemC migrates by SDS/PAGE with an apparent molecular mass of approx. 38 kDa, which is in close agreement with the predicted mass of 34 kDa. The UV–visible spectrum of the oxidized protein, at pH 8, contains two weak bands at approx. 410 and 500 nm (results not shown), similar to the spectrum reported for the E. coli HemC enzyme [7]. To investigate the presence of the dipyrromethane cofactor, the enzyme was mixed with Ehrlich's reagent, and the reaction was followed by UV–visible spectroscopy over a 20 min period. A change in absorbance from 566 to 495 nm (Figure 3) was observed, consistent with the presence of the dipyrromethane cofactor at the catalytic site of D. vulgaris HemC [7]. The specific activity of the enzyme was measured as being 20 μmol·h−1·mg−1, which is similar to the activity of E. coli HemC enzyme (43 μmol·h−1·mg−1) [29]. D. vulgaris HemC has a KM of 214 μM for PBG, which is higher than the range of values measured for HemC of E. coli, Rhodopseudomonas sphaeroides, C. josui and Chlorella regularis (19–89 μM) [18,29,35,36].

D. vulgaris PBG deaminase spectral features

Figure 3
D. vulgaris PBG deaminase spectral features

Reaction with Ehrlich's reagent assessed by UV–visible spectroscopy. The spectra were recorded at the indicated times.

Figure 3
D. vulgaris PBG deaminase spectral features

Reaction with Ehrlich's reagent assessed by UV–visible spectroscopy. The spectra were recorded at the indicated times.

In D. vulgaris, the uroporphyrinogen III synthase and methyltransferase activities are performed by a single enzyme

A BLAST search with the locus DVU0734 revealed that the N-terminal domain of the protein (amino acids 1–246) exhibits significant amino acid sequence similarity to the uroporphyrinogen III methyltransferase enzymes (CobA), whereas the C-terminal region of DVU0734 (amino acids 247–503) has similarity to uroporphyrinogen III synthases (HemD). This suggests that in D. vulgaris the two activities are fused into a single multifunctional enzyme, which was named CobA/HemD. An amino acid sequence alignment of D. vulgaris DVU0734 with other separate bacterial CobA and HemD enzymes allowed us to predict accurately the point of fusion between the two functional domains. Consequently, three proteins were recombinantly produced: the whole CobA/HemD protein, the truncated form corresponding to the N-terminal region (named CobAd) and a second shorter protein corresponding to the C-terminal domain (named HemDd).

The initial characterization of DVU0734 was performed by complementation studies. In E. coli the multifunctional enzyme CysG synthesizes sirohaem from uroporphyrinogen III. A deletion of this gene produces a mutant strain that is not able to produce sirohaem and thus cannot synthesize cysteine. The mutant phenotype can be rescued by complementation with the expression of genes from other organisms, which is enzymes that catalyse the reactions performed by CysG (SAM-dependent methyltransferase of uroporphyrinogen III, NAD+-dependent dehydrogenation of precorrin-2 and ferrochelation of sirohydrochlorin) [25,31,37]. To test if the bifunctional CobA/HemD and CobAd of D. vulgaris have uroporphyrinogen III methyltransferase activity in vivo, they were used to rescue the E. coli cysG mutant strain by co-expression with the bifunctional MET8 of yeast, which is a protein with precorrin-2 NAD+-dependent dehydrogenation and sirohydrochlorin ferrochelation activities (Table 2). The E. coli cysG strain expressing MET8 of Saccharomyces cerevisiae and cobA/hemD or cobAd of D. vulgaris was able to grow in the absence of cysteine, demonstrating that the D. vulgaris enzyme variants are able to perform the in vivo transmethylation of uroporphyrinogen III to yield precorrin-2.

The in vivo uroporphyrinogen III synthase activity of the D. vulgaris CobA/HemD and HemDd variants was investigated by complementation of an E. coli hemD strain (SASZ31) (see [38]). This strain grows very poorly in LB medium, forming mini colonies, and thus complementation is observed only when the bacterium is transformed with a gene that allows the organism to form normal-sized colonies. Transformation of the mutant strain with plasmids harbouring D. vulgaris cobA/hemD or hemDd resulted in fast growing strains, demonstrating that the D. vulgaris cobA/hemD eliminates the cell growth deficiency and indicating that the encoded protein is functioning in vivo. Furthermore, E. coli cells producing the D. vulgaris CobA/HemD exhibited a reddish pigmentation, which is due to the accumulation of sirohydrochlorin and trimethylpyrrocorphin and is consistent with increased uroporphyrinogen III methyltransferase activity [13,39].

The in vitro activity of the recombinantly produced and purified enzyme variants was also studied. After purification, the CobA/HemD, which has a molecular mass of ∼55 kDa, was tested for uroporphyrinogen III synthase and methyltransferase activity. A specific activity of 3 nmol·min−1·mg−1 was determined for the uroporphyrinogen III methyltransferase activity of the fused enzyme, with a KM for uroporphyrinogen III of 0.4 μM (Table 4). This is within the range of values observed previously, e.g. the Pseudomonas denitrificans CobA (1 μM) [40]. In contrast to the B. megaterium and Ps. denitrificans CobAs [10,40], the D. vulgaris CobA/HemD did not display substrate inhibition with uroporphyrinogen III, a characteristic that had previously been observed with an archaeal uroporphyrinogen III methyltransferase [41]. The uroporphyrinogen III synthase activity of the CobA/HemD fusion protein was evaluated in a linked assay. Here, all the enzymes required to transform ALA into precorrin-2, except for the uroporphyrinogen III synthase, are mixed together. When ALA and SAM are added to this incubation the reaction stalls with the accumulation of HMB. If uroporphyrinogen III synthase is added, the reaction proceeds to the synthesis of precorrin-2 and the development of a characteristic yellow colour. When the CobA/HemD fusion was supplied to this assay (Table 1), the generation of precorrin-2 was confirmed with the appearance of a characteristic UV–visible spectrum and the development of the typical yellow chromophore of the compound. The further addition of the B. megaterium SirC enzyme, which oxidizes precorrin-2, led to the formation of sirohydrochlorin providing further evidence for the formation of precorrin-2 in the reaction mixture (results not shown).

The activities of the two truncated proteins were investigated separately in an in vitro linked assay. This revealed that the CobAd had uroporphyrinogen III methyltransferase activity but not uroporphyrinogen III synthase activity, while HemDd was found to possess uroporphyrinogen III synthase activity (Table 1). During the course of these studies a mutant variant, F446A, was serendipitously obtained in HemDd and this variant showed no measurable uroporphyrinogen III synthase activity. The function of Phe446 in HemD enzymes is unknown, and although is not strictly conserved this residue seems to play an essential role. However, more results will be required to confirm the catalytic role of Phe446.

The D. vulgaris CysGB has precorrin-2 dehydrogenase activity

The D. vulgaris CysGB has 17, 21 and 26% sequence identity with Sa. cerevisiae MET8, B. megaterium SirC and to the N-terminal part of the E. coli CysG respectively (Table 3). Although this is a relatively low value of amino acid identity, it is in the same range of identity observed between the first 223 amino acids of E. coli CysG and MET8 (16%). A common feature of all of these proteins is the presence of a consensus NAD+-binding motif at the start of the N-terminal region [42]. This motif on the D. vulgaris CysGB (GxGxxGx10G) is identical with the one observed in MET8 [14] and similar to the one found in the N-terminal domain of the B. megaterium SirC (GxGxxAx10G) and E. coli CysG (GxGxxAx3Ax6G).

CysGB of D. vulgaris was assayed for NAD+-dependent precorrin-2 dehydrogenase activity by following the appearance of sirohydrochlorin, and was found to have a specific activity of ∼700 nmol·min−1·mg−1 and a KM for NAD+ of 70 μM (Table 4). The specific activity of D. vulgaris CysGB is significantly higher than the NAD+-dependent precorrin-2 dehydrogenase activity reported for the B. megaterium SirC (60 nmol·min−1·mg−1) [15]. However, the D. vulgaris enzyme is not able to perform the insertion of iron or cobalt into sirohydrochlorin, a result that is consistent with complementation studies, where the corresponding gene was unable to rescue the E. coli cysG mutant strain phenotype (Table 2). Cumulatively, these results show that the putative CysGB of D. vulgaris, although annotated in the D. vulgaris genome as an orthologue of the yeast bifunctional enzyme Met8p [31], is in fact a functional SirC enzyme, a single NAD+-dependent precorrin-2 dehydrogenase that does not possess any chelatase activity [15].

D. vulgaris enzymes generate precorrin-2 from ALA in vitro

An in vitro incubation with purified recombinant D. vulgaris HemB, HemC and CobA/HemD together with ALA and SAM resulted in the generation of precorrin-2 (Figure 4 and Table 1). The presence of the yellow dipyrrocorphin was confirmed by the loading of purified B. megaterium SirC and NAD+ to the incubation, an addition that generated the isobacteriochlorin, sirohydrochlorin (Figure 4).

UV–visible spectra of precorrin-2 (A) and sirohydrochlorin (B) generated in the linked assay performed with D. vulgaris recombinant enzymes

Conclusion

D. vulgaris makes sirohydrochlorin

In the work reported in the present study, the biochemical and functional characterization of several enzymes that catalyse the early steps of the modified tetrapyrrole biosynthetic pathway in D. vulgaris was undertaken. Thus the D. vulgaris PBG synthase (HemB) was produced recombinantly and shown to be a zinc-dependent enzyme that appears to have a native hexameric structure, in contrast with most of the bacterial HemB enzymes that are octamers. For PBG deaminase (HemC), the presence of the dipyrromethane cofactor, which covalently binds the growing poly-pyrrole chain, further supports the universal nature of this essential prosthetic group.

The recombinant D. vulgaris CobA/HemD was shown to be active both in vivo and in vitro, and represents the first characterization of a bifunctional uroporphyrinogen III synthase/methyltransferase. This suggests that, in this organism, uroporphyrinogen III is not released as a free intermediate. Rather, uroporphyrinogen III seems to be directed towards the formation of precorrin-2. This highly unstable intermediate is acted upon by a SirC-like protein, which, in the presence of NAD+, transforms it into sirohydrochlorin. It is known that sirohydrochlorin acts as an intermediate for B12 biosynthesis, since it is chelated with cobalt to generate cobalt-sirohydrochlorin by CbiK [37]. It also plays a role in sirohaem synthesis where ferrous chelation generates this prosthetic group, which is required by sulfite reductase and which is also found in D. vulgaris. However, it is also likely that sirohydrochlorin acts as an intermediate in the biosynthesis of haem.

D. vulgaris does not contain HemE, N, G or H

One of the interesting features of the D. vulgaris genome is the absence of the genes that encode enzymes for the classical synthesis of haem from uroporphyrinogen III. Thus there are no homologues to hemE, hemN, hemG or hemH, which encode the enzymes uroporphyrinogen III decarboxylase, coproporphyrinogen III oxidase, protoporphyrinogen IX oxidase and protoporphyrin IX ferrochelatase respectively. The absence of these genes is consistent with the presence of a novel pathway for the transformation of precorrin-2 or sirohydrochlorin into haem. Previous work by Sano and co-workers had shown that haem must be made via precorrin-2, since the two methyl groups found on rings A and B of the final porphyrin product are derived from methionine and not from the C2 position of ALA [22]. However, in their follow-up work, it was reported that an incubation of crude cell extracts of D. vulgaris with uroporphyrinogen III resulted in the appearance of sirohydrochlorin, 12,18-didecarboxysirohydrochlorin, coproporphyrin III and protoporphyrin IX. However, in this study, only the structure of the novel 12,18-didecarboxysirohydrochlorin was fully characterized [23].

If coproporphyrinogen III is an intermediate, then it could be acted upon by an anaerobic coproporphyrinogen oxidase such as HemN. Interestingly, the D. vulgaris genome does encode an orthologue of the radical SAM-reliant coproporphyrinogen III oxidase. The HemN proteins are characterized by the presence of an Fe-S redox group, which is constituted by four cysteine residues within the conserved motif (CxxxCxxCxC). However, in the D. vulgaris homologue, one of the cysteine residues is missing, suggesting that this D. vulgaris protein is not a HemN. Proteins missing this conserved motif have been shown not to be involved in haem biosynthesis [43]. Thus, if coproporphyrinogen III is an intermediate, it is not clear how it is converted into protoporphyrin IX as there are no recognized enzymes present to undertake the required reactions.

Possible proteins involved in the novel haem synthesis pathway

A scan of the genome reveals a number of genes encoding enzymes that are designated for the synthesis of haem d1 (nirD and nirJ), the cofactor of the nitrite reductase cd1. However, D. vulgaris does not seem to have cd1. These haem d1 synthetic homologues are also found in some genomes of methanogens and other bacteria that do not have cd1. In the methanogens, it has also been shown that haem is also made via a dimethylated uroporphyrinogen III derivative such as precorrin-2 or sirohydrochlorin [44]. Significantly, there are some mechanistic similarities between the pathways for haem d1 synthesis and haem construction via precorrin-2. In haem d1 synthesis, the two propionic acid side chains attached at C3 and C8 of the macrocycle are lost and are replaced by oxygen. To achieve haem synthesis from precorrin-2/sirohydrochlorin, the acetic acid side chains at C2 and C7 have to be removed. Both processes may involve radical chemistry and thus could be mediated via a radical SAM enzyme system. It is therefore of interest to note that NirJ is proposed to be a radical SAM protein. We thus propose that the D. vulgaris nirJ and nirD genes are involved in the transformation of either sirohydrochlorin or sirohaem into haem.

In summary, the research described in the present study outlines how the basic tetrapyrrole framework is synthesized up to the first branch point, which we believe to be sirohydrochlorin. Genomic analysis thus suggests that cobalamin and sirohaem are made by classical previously described routes but that haem is synthesized along a novel route using homologues of the haem d1 biosynthetic apparatus. The presence of a bifunctional CobA/HemD is consistent with uroporphyrinogen III not being a branch point in this bifurcated pathway and may represent a useful marker for an alternative haem biosynthetic route. More research is required to elucidate this surrogate synthesis. It is not clear why some of the Desulfovibrio or archaea have an alternative haem biosynthetic pathway. Possibly, the pathway evolved as a method of producing the more oxidized porphyrin ring system under anaerobic conditions at a time when the major tetrapyrroles were sirohaem and cobalamin. Then, as molecular oxygen became more abundant, the haem biosynthesis pathway involving several oxidative steps in the synthesis of protoporphyrin evolved and became the more common route.

Abbreviations

     
  • ALA

    aminolaevulinic acid

  •  
  • HMB

    hydroxymethylbilane

  •  
  • LB

    Luria–Bertani

  •  
  • PBG

    porphobilinogen

  •  
  • SAM

    S-adenosylmethionine

  •  
  • SRB

    sulfate-reducing bacteria

FUNDING

This work was supported by the FCT (Fundação para a Ciência e a Tecnologia) project PTDC/BIA-PRO/61107/2006 [grant number SFRH/BD/19813/2004 to S. A. L. L.]. The BBSRC (Biotechnology and Biological Sciences Research Council) is acknowledged for financial support.

References

References
1
Gibson
K. D.
Laver
W. G.
Neuberger
A.
Initial stages in the biosynthesis of porphyrins. 2. The formation of delta-aminolevulinic acid from glycine and succinyl-coenzyme A by particles from chicken erythrocytes
Biochem. J.
1958
, vol. 
70
 (pg. 
71
-
81
)
2
Jahn
D.
Verkamp
E.
Soll
D.
Glutamyl-transfer RNA: a precursor of heme and chlorophyll biosynthesis
Trends Biochem. Sci.
1992
, vol. 
17
 (pg. 
215
-
218
)
3
Nandi
D. L.
Shemin
D.
Delta-aminolevulinic acid dehydratase of Rhodopseudomonas spheroides. 3. Mechanism of porphobilinogen synthesis
J. Biol. Chem.
1968
, vol. 
243
 (pg. 
1236
-
1242
)
4
Jaffe
E. K.
An unusual phylogenetic variation in the metal ion binding sites of porphobilinogen synthase
Chem. Biol.
2003
, vol. 
10
 (pg. 
25
-
34
)
5
Bollivar
D. W.
Clauson
C.
Lighthall
R.
Forbes
S.
Kokona
B.
Fairman
R.
Kundrat
L.
Jaffe
E. K.
Rhodobacter capsulatus porphobilinogen synthase, a high activity metal ion independent hexamer
BMC Biochem.
2004
, vol. 
5
 pg. 
17
 
6
Senior
N. M.
Brocklehurst
K.
Cooper
J. B.
Wood
S. P.
Erskine
P.
Shoolingin-Jordan
P. M.
Thomas
P. G.
Warren
M. J.
Comparative studies on the 5-aminolaevulinic acid dehydratases from Pisum sativum, Escherichia coli and Saccharomyces cerevisiae
Biochem. J.
1996
, vol. 
320
 (pg. 
401
-
412
)
7
Jordan
P. M.
Warren
M. J.
Evidence for a dipyrromethane cofactor at the catalytic site of E. coli porphobilinogen deaminase
FEBS Lett.
1987
, vol. 
225
 (pg. 
87
-
92
)
8
Jordan
P. M.
Berry
A.
Mechanism of action of porphobilinogen deaminase. The participation of stable enzyme substrate covalent intermediates between porphobilinogen and the porphobilinogen deaminase from Rhodopseudomonas spheroides
Biochem. J.
1981
, vol. 
195
 (pg. 
177
-
181
)
9
Burton
G.
Fargerness
P. E.
Hosozawa
S.
Jordan
P. M.
Scott
A. I.
13C-NMR evidence for a new intermediate preuroporphyrinogen in the enzymic transformation of porphobilinogen into uroporphyrinogens
J. Chem. Soc. Chem. Commun.
1979
(pg. 
202
-
204
)
10
Robin
C.
Blanche
F.
Cauchois
L.
Cameron
B.
Couder
M.
Crouzet
J.
Primary structure, expression in Escherichia coli, and properties of S-adenosyl-L-methionine:uroporphyrinogen III methyltransferase from Bacillus megaterium
J. Bacteriol.
1991
, vol. 
173
 (pg. 
4893
-
4896
)
11
Sattler
I.
Roessner
C. A.
Stolowich
N. J.
Hardin
S. H.
Harris-Haller
L. W.
Yokubaitis
N. T.
Murooka
Y.
Hashimoto
Y.
Scott
A. I.
Cloning, sequencing, and expression of the uroporphyrinogen III methyltransferase cobA gene of Propionibacterium freudenreichii (shermanii)
J. Bacteriol.
1995
, vol. 
177
 (pg. 
1564
-
1569
)
12
Stroupe
M. E.
Leech
H. K.
Daniels
D. S.
Warren
M. J.
Getzoff
E. D.
CysG structure reveals tetrapyrrole-binding features and novel regulation of siroheme biosynthesis
Nat. Struct. Biol.
2003
, vol. 
10
 (pg. 
1064
-
1073
)
13
Warren
M. J.
Bolt
E. L.
Roessner
C. A.
Scott
A. I.
Spencer
J. B.
Woodcock
S. C.
Gene dissection demonstrates that the Escherichia coli cysG gene encodes a multifunctional protein
Biochem. J.
1994
, vol. 
302
 (pg. 
837
-
844
)
14
Hansen
J.
Muldbjerg
M.
Cherest
H.
Surdin-Kerjan
Y.
Siroheme biosynthesis in Saccharomyces cerevisiae requires the products of both the MET1 and MET8 genes
FEBS Lett.
1997
, vol. 
401
 (pg. 
20
-
24
)
15
Raux
E.
Leech
H. K.
Beck
R.
Schubert
H. L.
Santander
P. J.
Roessner
C. A.
Scott
A. I.
Martens
J. H.
Jahn
D.
Thermes
C.
, et al. 
Identification and functional analysis of enzymes required for precorrin-2 dehydrogenation and metal ion insertion in the biosynthesis of sirohaem and cobalamin in Bacillus megaterium
Biochem. J.
2003
, vol. 
370
 (pg. 
505
-
516
)
16
Anderson
P. J.
Entsch
B.
McKay
D. B.
A gene, cobA+hemD, from Selenomonas ruminantium encodes a bifunctional enzyme involved in the synthesis of vitamin B12
Gene
2001
, vol. 
281
 (pg. 
63
-
70
)
17
Taranto
M. P.
Vera
J. L.
Hugenholtz
J.
De Valdez
G. F.
Sesma
F.
Lactobacillus reuteri CRL1098 produces cobalamin
J. Bacteriol.
2003
, vol. 
185
 (pg. 
5643
-
5647
)
18
Fujino
E.
Fujino
T.
Karita
S.
Sakka
K.
Ohmiya
K.
Cloning and sequencing of some genes responsible for porphyrin biosynthesis from the anaerobic bacterium Clostridium josui
J. Bacteriol.
1995
, vol. 
177
 (pg. 
5169
-
5175
)
19
Timkovich
R.
Burkhalter
R. S.
Xavier
A. V.
Chen
L.
LeGall
J.
Iron uroporphyrin I and a heme c-derivative are prosthetic groups in Desulfovibrio gigas rubredoxin oxidase
Bioorg. Chem.
1994
, vol. 
22
 (pg. 
284
-
293
)
20
Romao
C. V.
Louro
R.
Timkovich
R.
Lubben
M.
Liu
M. Y.
LeGall
J.
Xavier
A. V.
Teixeira
M.
Iron-coproporphyrin III is a natural cofactor in bacterioferritin from the anaerobic bacterium Desulfovibrio desulfuricans
FEBS Lett.
2000
, vol. 
480
 (pg. 
213
-
216
)
21
Murphy
M. J.
Siegel
L. M.
Siroheme and sirohydrochlorin. The basis for a new type of porphyrin-related prosthetic group common to both assimilatory and dissimilatory sulfite reductases
J. Biol. Chem.
1973
, vol. 
248
 (pg. 
6911
-
6919
)
22
Akutsu
H.
Park
J. S.
Sano
S.
L-methionine methyl is specifically incorporated into the C-2 and C-7 positions of the porphyrin of cytochrome c3 in a strictly anaerobic bacterium, Desulfovibrio vulgaris
J. Am. Chem. Soc.
1993
, vol. 
115
 (pg. 
12185
-
12186
)
23
Ishida
T.
Yu
L.
Akutsu
H.
Ozawa
K.
Kawanishi
S.
Seto
A.
Inubushi
T.
Sano
S.
A primitive pathway of porphyrin biosynthesis and enzymology in Desulfovibrio vulgaris
Proc. Natl. Acad. Sci. U.S.A.
1998
, vol. 
95
 (pg. 
4853
-
4858
)
24
Heidelberg
J. F.
Seshadri
R.
Haveman
S. A.
Hemme
C. L.
Paulsen
I. T.
Kolonay
J. F.
Eisen
J. A.
Ward
N.
Methe
B.
Brinkac
L. M.
, et al. 
The genome sequence of the anaerobic, sulfate-reducing bacterium Desulfovibrio vulgaris Hildenborough
Nat. Biotechnol.
2004
, vol. 
22
 (pg. 
554
-
559
)
25
Raux
E.
Thermes
C.
Heathcote
P.
Rambach
A.
Warren
M. J.
A role for Salmonella typhimurium cbiK in cobalamin (vitamin B12) and siroheme biosynthesis
J. Bacteriol.
1997
, vol. 
179
 (pg. 
3202
-
3212
)
26
Laemmli
U. K.
Cleavage of structural proteins during the assembly of the head of bacteriophage T4
Nature
1970
, vol. 
227
 (pg. 
680
-
685
)
27
Smith
P. K.
Krohn
R. I.
Hermanson
G. T.
Mallia
A. K.
Gartner
F. H.
Provenzano
M. D.
Fujimoto
E. K.
Goeke
N. M.
Olson
B. J.
Klenk
D. C.
Measurement of protein using bicinchoninic acid
Anal. Biochem.
1985
, vol. 
150
 (pg. 
76
-
85
)
28
Mauzerall
D.
Granick
S.
The occurrence and determination of delta-amino-levulinic acid and porphobilinogen in urine
J. Biol. Chem.
1956
, vol. 
219
 (pg. 
435
-
446
)
29
Jordan
P. M.
Thomas
S. D.
Warren
M. J.
Purification, crystallization and properties of porphobilinogen deaminase from a recombinant strain of Escherichia coli K12
Biochem. J.
1988
, vol. 
254
 (pg. 
427
-
435
)
30
Schubert
H. L.
Raux
E.
Brindley
A. A.
Leech
H. K.
Wilson
K. S.
Hill
C. P.
Warren
M. J.
The structure of Saccharomyces cerevisiae Met8p, a bifunctional dehydrogenase and ferrochelatase
EMBO J.
2002
, vol. 
21
 (pg. 
2068
-
2075
)
31
Raux
E.
McVeigh
T.
Peters
S. E.
Leustek
T.
Warren
M. J.
The role of Saccharomyces cerevisiae Met1p and Met8p in sirohaem and cobalamin biosynthesis
Biochem. J.
1999
, vol. 
338
 (pg. 
701
-
708
)
32
Kervinen
J.
Jaffe
E. K.
Stauffer
F.
Neier
R.
Wlodawer
A.
Zdanov
A.
Mechanistic basis for suicide inactivation of porphobilinogen synthase by 4,7-dioxosebacic acid, an inhibitor that shows dramatic species selectivity
Biochemistry
2001
, vol. 
40
 (pg. 
8227
-
8236
)
33
Erskine
P. T.
Norton
E.
Cooper
J. B.
Lambert
R.
Coker
A.
Lewis
G.
Spencer
P.
Sarwar
M.
Wood
S. P.
Warren
M. J.
Shoolingin-Jordan
P. M.
X-ray structure of 5-aminolevulinic acid dehydratase from Escherichia coli complexed with the inhibitor levulinic acid at 2.0 Å resolution
Biochemistry
1999
, vol. 
38
 (pg. 
4266
-
4276
)
34
Frankenberg
N.
Erskine
P. T.
Cooper
J. B.
Shoolingin-Jordan
P. M.
Jahn
D.
Heinz
D. W.
High resolution crystal structure of a Mg2+-dependent sporphobilinogen synthase
J. Mol. Biol.
1999
, vol. 
289
 (pg. 
591
-
602
)
35
Jordan
P. M.
Shemin
D.
Purification and properties of uroporphyrinogen I synthetase from Rhodopseudomonas spheroides
J. Biol. Chem.
1973
, vol. 
248
 (pg. 
1019
-
1024
)
36
Shioi
Y.
Nagamine
M.
Kuroki
M.
Sasa
T.
Purification by affinity chromatography and properties of uroporphyrinogen I synthetase from Chlorella regularis
Biochim. Biophys. Acta
1980
, vol. 
616
 (pg. 
300
-
309
)
37
Lobo
S. A.
Brindley
A. A.
Romao
C. V.
Leech
H. K.
Warren
M. J.
Saraiva
L. M.
Two distinct roles for two functional cobaltochelatases (CbiK) in Desulfovibrio vulgaris Hildenborough
Biochemistry
2008
, vol. 
47
 (pg. 
5851
-
5857
)
38
Chartrand
P.
Tardif
D.
Sasarman
A.
Uroporphyrin- and coproporphyrin I-accumulating mutant of Escherichia coli K12
J. Gen. Microbiol.
1979
, vol. 
110
 (pg. 
61
-
66
)
39
Roessner
C. A.
Scott
A. I.
Fluorescence-based method for selection of recombinant plasmids
BioTechniques
1995
, vol. 
19
 (pg. 
760
-
764
)
40
Blanche
F.
Debussche
L.
Thibaut
D.
Crouzet
J.
Cameron
B.
Purification and characterization of S-adenosyl-L-methionine: uroporphyrinogen III methyltransferase from Pseudomonas denitrificans
J. Bacteriol.
1989
, vol. 
171
 (pg. 
4222
-
4231
)
41
Blanche
F.
Robin
C.
Couder
M.
Faucher
D.
Cauchois
L.
Cameron
B.
Crouzet
J.
Purification, characterization, and molecular cloning of S-adenosyl-L-methionine: uroporphyrinogen III methyltransferase from Methanobacterium ivanovii
J. Bacteriol.
1991
, vol. 
173
 (pg. 
4637
-
4645
)
42
Bellamacina
C. R.
The nicotinamide dinucleotide binding motif: a comparison of nucleotide binding proteins
FASEB J.
1996
, vol. 
10
 (pg. 
1257
-
1269
)
43
Sofia
H. J.
Chen
G.
Hetzler
B. G.
Reyes-Spindola
J. F.
Miller
N. E.
Radical SAM, a novel protein superfamily linking unresolved steps in familiar biosynthetic pathways with radical mechanisms: functional characterization using new analysis and information visualization methods
Nucleic Acids Res.
2001
, vol. 
29
 (pg. 
1097
-
1106
)
44
Buchenau
B.
Kahnt
J.
Heinemann
I. U.
Jahn
D.
Thauer
R. K.
Heme biosynthesis in Methanosarcina barkeri via a pathway involving two methylation reactions
J. Bacteriol.
2006
, vol. 
188
 (pg. 
8666
-
8668
)
45
Larkin
M. A.
Blackshields
G.
Brown
N. P.
Chenna
R.
McGettigan
P. A.
McWilliam
H.
Valentin
F.
Wallace
I. M.
Wilm
A.
Lopez
R.
, et al. 
Clustal W and Clustal X version 2.0
Bioinformatics
2007
, vol. 
23
 (pg. 
2947
-
2948
)
46
Raux
E.
Schubert
H. L.
Warren
M. J.
Biosynthesis of cobalamin (vitamin B12): a bacterial conundrum
Cell. Mol. Life Sci.
2000
, vol. 
57
 (pg. 
1880
-
1893
)

Supplementary data