OUP user menu

Nitrate reduction by Desulfovibrio desulfuricans: A periplasmic nitrate reductase system that lacks NapB, but includes a unique tetraheme c-type cytochrome, NapM

Angeliki Marietou, David Richardson, Jeff Cole, Sudesh Mohan
DOI: http://dx.doi.org/10.1016/j.femsle.2005.05.042 217-225 First published online: 1 July 2005

Abstract

Many sulphate reducing bacteria can also reduce nitrite, but relatively few isolates are known to reduce nitrate. Although nitrate reductase genes are absent from Desulfovibrio vulgaris strain Hildenborough, for which the complete genome sequence has been reported, a single subunit periplasmic nitrate reductase, NapA, was purified from Desulfovibrio desulfuricans strain 27774, and the structural gene was cloned and sequenced. Chromosome walking methods have now been used to determine the complete sequence of the nap gene cluster from this organism. The data confirm the absence of a napB homologue, but reveal a novel six-gene organisation, napC-napMnapAnapDnapGnapH. The NapC polypeptide is more similar to the NrfH subgroup of tetraheme cytochromes than to NapC from other bacteria. NapM is predicted to be a tetra-heme c-type cytochrome with similarity to the small tetraheme cytochromes from Shewanella oneidensis. The operon is located close to a gene encoding a lysyl-tRNA synthetase that is also found in D. vulgaris. We suggest that electrons might be transferred to NapA either from menaquinol via NapC, or from other electron donors such as formate or hydrogen via the small tetraheme cytochrome, NapM. We also suggest that, despite the absence of a twin-arginine targeting sequence, NapG might be located in the periplasm where it would provide an alternative direct electron donor to NapA.

Keywords
  • Periplasmic nitrate reductase
  • NapB
  • Desulfovibrio
  • Small tetraheme cytochrome c
  • NapM
  • Nitrate reductase evolution

1 Introduction

Analysis of the genomes of many Gram-negative bacteria has revealed the presence of operons encoding periplasmic nitrate reductases that differ in complexity, components, and organisation. Until recently four genes, napD, A, B and C (often in that order) were believed to be invariably present [[]. The molybdoprotein, NapA, is the catalytic subunit that accepts electrons from the quinol pool via the tetra-heme c-type cytochrome, NapC and the di-heme cytochrome, NapB [[]. NapD is apparently a pathway-specific chaperone that plays an uncharacterised role in NapA maturation [[,[]. Recently, several examples of nap operons that lack napC have been recognised. In Shewanella oneidensis, a NapC homologue, CymA, that is thought to play a more general role in anaerobic electron transfer, is encoded by a distant locus [[]. Wolinella succinogenes lacks NapC altogether, and there is direct evidence that neither of the two NapC-like tetra-heme cytochromes expressed by Wolinella succinogenes are required for periplasmic nitrate reduction [[]. However, the W. succinogenes nap complex includes genes for two iron–sulphur proteins, NapG and NapH, that are commonly found together in the nap gene clusters of obligate anaerobes, or bacteria like Escherichia coli that can adapt to an anaerobic lifestyle. Recently we showed that the membrane-associated E. coli NapH and periplasmic NapG form an alternative ubiquinol dehydrogenase that transfers electrons from ubiquinol via NapC to NapB and NapA [[].

With the sole exception of the nitrate reductase from the Gram-positive Symbiobacterium thermophilum (see Section 3), all of the nap gene clusters sequenced to date include a napB gene. However, there are considerable differences between the abilities of NapA from different bacteria to interact with NapB. The first periplasmic nitrate reductase to be purified was from Paracoccus denitrificans[[], in which NapA and NapB form a tight complex. At the opposite extreme, NapA from Desulfovibrio desulfuricans strain 27774 was purified as a single polypeptide devoid of NapB. This protein was crystallized, and its structure was determined [[,[]. No NapB subunit has ever been reported in this organism. We now report the cloning and sequencing of all of the components of the D. desulfuricans nap gene cluster, and confirm the absence of a napB gene. The data are entirely consistent with current biochemical evidence that NapA in D. desulfuricans is so far unique, and provide an explanation for the differences in the biochemical properties of the periplasmic nitrate reductases from different bacteria. They also provide direct experimental evidence for a proposal that periplasmic nitrate reductases might have evolved from a cytoplasmic, assimilatory nitrate reductase for which an iron–sulphur protein rather than a c-type cytochrome is the direct electron donor.

2 Materials and methods

2.1 Strains, media and growth conditions

D. desulfuricans subsp. desulfuricans DSM 6949 (ATCC 27774) from the Deutsche Sammlung von Mikroorganismen und Zellkulturen GmbH, Braunschweig, Germany, was cultured in Postgate's C liquid medium that contained sulphate; and Postgate N, a modified sulphate-free Postgate's C medium containing 14 mM potassium nitrate [[0]. The cultures were grown in sealed serum bottles (pre-gassed with N2), inoculated with 10% (v/v) of a fresh culture, incubated at 30°C for 2 days and the bacteria were collected by centrifugation. Good growth was observed in both types of media, but the absence of a black deposit of iron sulphide in the Postgate N medium facilitated the purification of chromosomal DNA. The purity of the nitrate-grown culture was checked microscopically using the Gram stain, by checking for inability of the grown culture to form colonies on nutrient agar during aerobic growth, and for formation of a black sulphide deposit when sub-cultured in Postgate's C medium.

E. coli JM109 (Promega), used as host for recombinant plasmids, was grown at 37°C in Luria–Bertani (LB) broth with shaking or on LB agar. Where appropriate, ampicillin (100 μg ml−1; Sigma), IPTG (0.5 mM; Promega), and X-gal (40 μg ml−1; Sigma) were added to LB agar for selection.

2.2 DNA isolation

The DNA was extracted from bacteria using the GFX Genomic Blood DNA purification kit (Amersham Biosciences) according to the manufacturer's instructions.

2.3 Determination of DNA sequence upstream and downstream of the known napA sequence using the universal fast walking method

The published napA sequence [[] was initially confirmed by sequencing the product obtained by PCR using the forward primer DdV16 and the reverse primer DdR12 (Table 1). The resulting sequence coincided with base pairs 252–2300 of the known sequence. Using the universal fast walking method [[1], this sequence was extended by 1.2 kb at the 5′ end and by 2.1 kb at the 3′ end. Each extension by this method involves 5 stages and requires 5 forward primers based upon the known sequence. Essentially, a single strand is extended using the first primer (P1). This is next annealed with a second primer (P2) that has 18 specific bases followed by 10 random bases that bind by chance to specific sites leaving the 18 bp specific bases as a 5′ end. This 5′ end is next filled to provide a 3′ sequence that in the next stage of the procedure forms a lariat by intra-strand annealing to the P2 site, and extension of the fragment to the original P1 site. In the final stage forward primers P3 and P4, specific to sequences on either side of P2, are used to amplify the extended fragment. The sequence of the fragment is determined using primer P5. The primers used for the 5′ and 3′ extensions are listed in Table 1. For the 5′ extension, the primers corresponding to P1–P5 were DdV67, P21, DdR4, DdR5 and DdR1, respectively. The 3′ sequence extension was obtained in two stages. The five primers used for the first extension were P1, P2, DdF3, DdF4 and DdF5. For the second extension the corresponding primers were DdF4, B12-P2, B12-P3, B12-P4 and B12-P5. Each new sequence was confirmed using appropriate pairs of primers and chromosomal DNA for PCR amplification.

View this table:
Table 1

Primers used

PrimerSequencePosition and stranda
B12-P2CCCTGCTATCTCTGCATGNNNNNNNNNN+4919
B12-P3ATTTTTACGGCTGTGCGTCCGCTG+4798
B12-P4ATGTGCATGACCTGCTATGACCG+5039
B12-P5AATTACGCAGGGGGTGTCATGTGC+5057
DdV16TAAAGGGCGTCTGCCGTTATTGCGG+2264
DdV67AAAAAGCACGGGATGCGCCTCGG−2797
Dd-F1CAG AAA TGG ACA TGA GCG ATA TGA TG+93
DdF3AACGCCTGCGTAAGGAATCGGG+3860
DdF4TATCGACCACTGGCATACGGCAACC+4092
DdF5ATAGGAGGGGGTATGGCTATTGCCG+4407
DdR1TTTCCTTTGAATTGAGCACGGGG−2422
DdR4ATA TCG GCG TAG GTT CCC ATA GG−2729
DdR5AACCCGTGCCGCAATAACGGCAG−2296
DdR12AAAACGGCACCGCAATAAGACCGGG−4306
Dd-R13ATG CAG AGA TAG CAG GGT GTC TGG C−4935
Dd-RP3TGA CCT GGA CGT GGA AGG CAA GCT−6814
P1AACACCCTTGTGGAATTCGCCAGG+3724
P2CCCCGACCGTTTCTTTTTTNNNNNNNNNN+3969
P21GCCGAAGCCCCCCTTCAANNNNNNNNNN−2625
PnestF1GAC AGA TCT CGA CCG CGC GGC C+1282
ProbeF1AAT GTG GTG CGC ATG+1208
ProbeR1GGT CAG TGG GAA AAG−1654
ProbeR2ACG GCT GCG CGC AGC ATA CAC−5640
W12-P4TTC TGA ACT GGT GTA TGC TGC GCG+5610
W13-P4ATC GGC CGC ATG CGA TGA AAG GTG−1159
  • a+ indicates the position of the 5′ base on the coding strand; − indicates the 5′ base on the non-coding strand (reverse primer). For positions, see Fig. 1.

2.4 Use of Southern analysis to complete the sequencing of the 5′ and 3′ ends of the nap operon

Two probes were constructed based on the nap operon sequence obtained above. A 447 bp fragment (probe A) 129 bp downstream of the 5 ′ end and a 524 bp fragment (probe B) 60 bp upstream of the 3 ′ end of the available sequence were amplified using chromosomal DNA of D. desulfuricans DSM6949 as template. Probes A and B were generated by PCR using primers probeF1/probeR1 and B12-P4/probeR2, respectively (Table 1). The amplified probes were DIG-labelled using the Digoxigenin DNA labeling and detection kit (Roche Diagnostic) according to the manufacturer's recommendations. Within probe A sequence there is a single Nar I restriction site, and in probe B sequence there is a single Mfe I restriction site. Chromosomal DNA was digested overnight with Nar I or with Hin dIII, restriction fragments were size-fractionated by 0.8% agarose gel electrophoresis, transferred onto a nitrocellulose membrane and probed with probes A or B, respectively. Fragments of interest (1–2 kb for Nar I and 1.5–3 kb for the Hin dIII digests, respectively) were excised and DNA was extracted using the Qiaquick Gel Extraction kit (Qiagen). The resulting DNA fragments were next self-ligated overnight at 16°C. The ligation mix was then used as a template in 50 μl inverse PCR reactions in order to amplify the Nar I and Hin dIII fragments of interest.

Primers pnestF1 and W13-P4 were used for the amplification of the 1.5 kb Nar I fragment. The inverse PCR products were ligated into pGEM?-T easy vector (Promega) and transformed into E. coli JM109 competent cells (Promega). The transformants were selected by plating onto LB/Ampicillin/X-gal/IPTG selective plates. Plasmids were extracted from overnight cultures of transformants by using the QIAprep Spin Miniprep kit (Qiagen). The purified plasmids were screened for the correct insert using a series of restriction digests. The correct inserts were sequenced by the plasmid to profile sequencing service provided by the Functional Genomics Laboratory, School of Biosciences, University of Birmingham. Plasmid DNA and primer pnestF1 were mixed and the sequencing reaction set up using the Big Dye Terminator kit (applied Biosystems) in a 3700 sequence analyser (Applied Biosystems).

The above procedure was also used to obtain the 3′ sequence of the Hin dIII digest except that the primers used to amplify the fragments were DdR13 and W12P4 (Table 1) and the fragments of interest were sequenced using either primer W12P4 or DdRP3. The sequences obtained were confirmed by PCR using appropriate primers and the chromosomal DNA as template. The DNA sequence of the six-gene nap operon has been assigned the accession number AJ920046 in the EMBL database.

2.5 Sequence alignments and comparison with the D. vulgaris genome sequence

Putative proteins were identified by similarity searches using blast analysis (http://www.ncbi.nlm.nih.gov/BLAST/). Gene sequences were analysed using either the ClustalW programme, or the GCG software (the Wisconsin Package Version 10.2, Genetics Computer Group, Madison, WI).

3 Results and discussion

3.1 DNA sequencing strategies

First the published sequence of napA was confirmed using primers designed from the published sequence of napA from D. desulfuricans strain 27774 [[]. This sequence was extended using the universal fast walking method [[1]. For unknown reasons and despite the replacement of many of the primers designed to complete the sequencing of the operon, success with this method was sporadic, terminating with sequence within napC at the 5′ end of the operon and within napH at the 3′ end (Fig. 1).

Figure 1

Organisation of the D. desulfuricans periplasmic nitrate reductase gene cluster and the location of adjacent. (A) Linear representation of the napCMADGH gene cluster, showing key restriction sites exploited in the Southern analysis. The ♦ symbol denotes the part of the sequence that was determined by the universal fast walking method [[1], while the symbol denotes the part of the sequence that was determined by Southern analysis. Map drawn to scale. (B) Corresponding location of adjacent genes on the D. vulgaris Hildenborough chromosome [[7].

To complete the sequencing, Southern blotting was used to identify restriction sites within 3 kb upstream or downstream of the completed sequence. This approach identified a 1.5 kb Nar I fragment that included napC and upstream DNA, and a 2.2 kb Hin dIII fragment that included napH and downstream DNA. This resulted in the complete assembly of the 5177 bp nap operon plus an additional 1 kb of upstream DNA and 0.5 kb of downstream DNA (Fig. 1A).

The DNA sequences either side of the D. desulfuricans nap operon are either very closely related or identical to contiguous sequences in the Desulfovibrio vulgaris genome that are transcribed in the same orientation (Fig. 1B). This would suggest either that a 6 kb nap region has been deleted in D. vulgaris, or more likely that there is a 6 kb insert including the nap operon in D. desulfuricans. Analysis of the DNA flanking the nap operon failed to reveal any other functional genes that are absent from D. vulgaris (Fig. 1B). The very short distance between the flanking genes in D. vulgaris suggests that they might be part of a single operon, and that the DNA on the 5′ side of the nap operon no longer encodes functional products. This would be consistent with an insertion of the nap region into a predecessor of D. desulfuricans, followed by the accumulation of random mutations or recombination events to generate what would then become non-functional genes.

3.2 Analysis of the sequence of the D. desulfuricans nap operon

Analysis of the nap cluster revealed the presence of six open reading frames that, on the basis of similarity to other nap genes, were designated napCnapMnapAnapDnapGnapH (Fig. 2). Translational coupling is predicted between NapC and NapM, NapD and NapG, and NapG and NapH, and the only gap between genes is 108 bp between napM and napA. The nap gene cluster is therefore most likely to be expressed as a single operon. Its organisation is so far unique, increasing further the various types of organisation of nap clusters in different bacteria (Fig. 2).

Figure 2

Organisation of nap gene clusters in various bacteria.

The first open reading frame encodes a 190 amino acid polypeptide with four Cys–X–X–Cys–His heme-binding motifs and a predicted N-terminal transmembrane α-helix typical of quinol dehydrogenase NapC. The predicted molecular mass of the apoprotein is 20.7 kDa, and the mature cytochrome with four covalently attached hemes would be about 22 kDa. Although it clearly falls within the NapC family of tetra-heme cytochromes, D. desulfuricans NapC is only 24% identical to the E. coli NapC. In contrast, the sequences of NapC from E. coli and Paracoccus pantotrophus are 48% identical. In fact multiple sequence analysis of ?80 members of the NapC family of tetra-heme cytochromes that includes the NrfH (electron donor to cytochrome c nitrite reductase) and TorC (electron donor to trimethylamine N-oxide reductase) subgroups places the D. desulfuricans NapC in the NrfH rather than the NapC clade. A key feature of the primary structure that gives confidence in this phylogenetic location for D. desulfuricans NapC is the absence of a histidine conserved in the NapC group that lies 3 amino acids after the second CXXCH heme binding motif (Fig. 3A). This His has been shown by site-specific mutagenesis of P. pantotrophus NapC to provide an axial ligand to one of the heme irons [[2]. The genes encoding NrfH tetra-heme cytochromes are normally found as part of a nrfHA gene cluster in δ and ε proteobacteria in which nrfA is the gene for penta-heme nitrite reductase and nrfH encodes the tetra-heme quinol-oxidising electron donor [[3]. Despite its location in the NrfH clade, we retain the NapC nomenclature for the D. desulfuricans cytochrome because of the genetic context of its gene in the nap cluster but note that this unusual nap cluster appears to be a hybrid of nap and nrf genes.

Figure 3

Multiple sequence primary structure alignments of fragments of the D. desulfuricans Nap proteins to illustrate some of the distinct primary structure features described in the text. (A) Alignment of segments of NapC and NrfH that bind hemes 1 and 2 of the four hemes. The His residue that is conserved in the NapC clade and provides a heme iron ligand is shown in blue and indicated by an arrow. (B) Alignment of D. desulfuricans NapM with the small tetra-heme cytochrome of Shewanella species. Histidine residues that lie outside of the heme-binding CXXCH motifs are highlighted in blue and in the case of structurally defined small tetra-heme cytochrome these have been shown to provide axial ligands to the four heme irons. (C) Alignment of two segments of periplasmic and assimilatory nitrate reductase polypeptides to illustrate the sequence insertions found in many of the proteobacterial periplasmic nitrate reductases. (D) Alignments of the N-terminal half of NapG illustrating the truncated N-termini of D. desulfuricans and S. thermophilum NapG, which lack a twin arginine signal sequence. TAT indicates the position of putative double arginine signal sequence; Cys1 and Cys2 indicate the position of the first two cysteine clusters. Key. Dd, D. desulfuricans; Ec, Escherichia coli; Pa, Pseudomonas aeruginosa; Pp, Paracoccus pantotrophus; Rs, Rhodobacter sphaeroides; Sc, Synechococcus elongatus; Sf, Shewanella frigidimarina; So, Shewanella oneidensis; St, Symbiobacterium thermophilum; Ws, Wolinella succinogenes. Nap, periplasmic nitrate reductase components as discussed in the text; NrfH, tetra-heme electron donor to cytochorme c nitrite reductase; NarB, assimilatory nitrate reductase of cyanobacteria; STC, small tetra-heme cytochromes. Numbers refer to the position of the first and last amino acid in the row on the polypeptide chain. The identical residues in the alignments are shown in red.

The second gene of the D. desulfuricans nap operon that we have designated NapM encodes a second tetra-heme c-type cytochrome. The protein has a putative Sec signal sequence consistent with the periplasmic location of c-type cytochromes and a predicted mass for the heme-free processed apoprotein of ?12.5 kDa. In terms of size this NapM is similar to the structurally defined small tetra-heme c-type cytochromes (STC) from Shewanella species [[4] and although overall levels of identity are low (?21%) there are similarities in the spacing of the CXXCH heme binding motifs, suggesting that NapM is more related to these cytochromes than to the structurally distinct tetra-heme cytochrome c3s of sulphate reducing bacteria. The positions of the four axial histidine heme ligands in STCs is not conserved in NapM, although there are still sufficient histidine residues for all four hemes of NapM to be bis-His ligated.

The third gene of the operon encodes the 755-residue NapA sub-unit, which is the molybdoprotein that catalyses nitrate reduction to nitrite. The predicted mass of apo-NapA, which includes a twin-arginine motif for targeting to the periplasm via the TAT export pathway, is 83.5 kDa. The predicted processed mass is ?80.3 kDa, which is consistent with the previous 80 kDa estimated by SDS PAGE analysis of the purified mature protein [[]. The size of this NapA is more similar to that of the cytoplasmic ferrodoxin-dependent assimilatory nitrate reductases with which the enzymes share ?34% identity. This compares to ?39% identity to E. coli NapA. For reference pairwise alignments of other proteobacterial NapA proteins from different phylogenetic groups gives around 65–70% identity (e.g., 68% between α-proteobacterial R. sphaeroides and γ proteobacterial E. coli NapA). Thus the D. desulfuricans NapA is at present the most divergent of the periplasmic nitrate reductase and most closely related to the cytoplasmic assimilatory enzymes. A primary structural feature that illustrates this is the presence of two insertions in most proteobacterial NapA polypeptides, which are absent in D. desulfuricans NapA (Fig 3C), that largely account for the difference in mass between periplasmic and assimilatory nitrate reductases.The napA gene is followed by the 109-codon napD gene, which encodes NapD of mass 11.9 kDa. The identity of NapD to E. coli NapD is low, 20%, but in contrast to NapC, there is only 16–20% similarity between other phylogenetically distinct NapD pairs (for example, 17% between NapD from E. coli and P. pantotrophus).

The final two genes, napG and napH, are predicted to encode non-heme iron–sulfur proteins that, in E. coli, have been implicated in electron transfer from ubiquinol to the terminal components of the Nap complex [[]. The predicted masses of NapG (174 amino acids; 18.0 kDa) and NapH (299 residues; 31.8 kDa) are similar to those of their homologues in E. coli, W. succinogenes, Campylobacter jejuni and S. oneidensis. However, although NapD and NapG appear from the sequence ATGA, which includes the initial methionine codon for NapG and a TGA stop codon for NapD, to be translationally coupled, there is neither a twin arginine motif for TAT-dependent transport into the periplasm, nor a Sec-dependent signal peptide (Fig 3D). Six independent clones from two independent experiments were sequenced to confirm this region of the operon. NapG is therefore either located in the cytoplasm, in which case it cannot be an electron donor to NapA, or it might be transported into the periplasm as part of a NapA–NapG complex. All four cysteine clusters predicted to bind four [4Fe–4 S] centres in the E. coli NapG are present in D. desulfuricans NapG, and with similar but not identical spacing. The first cluster is remarkably conserved: C–V–R–C–G–Q–C–V–*Q–A–C–P– in the E. coli protein with only Q43 (starred) replaced by alanine in the D. desulfuricans NapG (Fig 3D). The spacings between the cysteine residues in the subsequent three clusters in D. desulfuricans NapG are C–X2–C–X2–C–X3–CP (Fig. 3D); C–X8–C–X2–C–X3–C–P; and CX2CX2CX3C.The final gene encodes a NapH polypeptide with the expected conserved cysteine motifs. Although the N-terminal half is similar to the E. coli NapH, there is little sequence identity between the C-terminal half of NapH from D. desulfurican and other bacteria. The structure of the D. desulfuricans nap operon is especially interesting for at least five reasons. Perhaps the most significant point is that it lacks a napB gene. Prior to the analysis of nap gene clusters presented in this paper, NapB was previously considered to be a universal electron donor to the catalytic subunit, the molybdoprotein, NapA. Secondly, there is a gene for a second tetra-heme c-type cytochrome that we have designated napM. No napM counterpart has been reported so far in other nap gene clusters. Thirdly, although napB is absent, napC is present, indicating that the quinol dehydrogenase, NapC, is able to function independently of NapB. Also present, however, are both napG and napH, which we have proposed form an alternative quinol dehydrogenase to NapC in the complex electron transfer pathway to the periplasmic nitrate reductase in E. coli[[]. Finally, our data provide independent confirmation that the periplasmic nitrate reductase from sulphate reducing bacteria is significantly smaller than those from most other eubacteria, for example, E. coli or P. denitrificans. Possibly this might enable D. desulfuricans NapA to accept electrons from more than one electron donor.

Assuming that NapB is not encoded elsewhere on the chromosome, its absence inevitably raises the question which alternative Nap subunit might replace it in sulphate-reducing bacteria. One possibility, which we do not believe to be correct, is that electrons from the quinol pool are transferred directly to NapA by NapC. More likely is that NapM, mediates electron transfer between NapC and NapA. There are precedents for such arrangements in Shewanella species where STCs or STC-like domains of flavocytochrome c fumarate reductases draw electrons from CymA, a member of the NapC family [[5]. An alternative possibility, for which there is no experimental evidence, is that NapM might provide a direct link between the terminal components of the Nap electron transfer chain and one of the four periplasmic uptake hydrogenases revealed by the genome sequence of D. vulgaris Hildenborough [[6,[7]. However, no energy would be conserved as proton motive force in a redox process occurring totally in the periplasm. It is also possible that there are two electron transfer pathways to NapA, one via NapG and the other from a wider range of electron donors via NapM. Mutational analysis and physiological experiments will be required to test this speculation.

In considering the overall relatedness of nap gene clusters we have analysed the nap genes of the recently published genome sequence of the Gram-positive bacterium S. thermophilum[[8]. There are four compelling parallels between this cluster and that of D. desulfuricans. First, the gene order of this cluster is napGADC (Fig. 2). S. thermophilum NapC, like that of D. desulfuricans, falls into the NrfH clade and lacks the key histidine residue that defines the NapC clade (Fig 3A). Second, S. thermophilum NapA is similar in length to that of D. desulfuricans (?.750 aa rather than the ca. 830 aa characteristic of other NapAs). The D. desulfuricans NapA and S. thermophilum NapA also share the highest level of identity with each other (?50%). The length of these NapA polypeptides after signal sequence cleavage is predicted to be ?725 amino acids, which is very similar to that of the cytoplasmic cyanobacterial assimilatory nitrate reductases (e.g., 729 aa for S. elongatus nitrate reductase) and this is largely due to the absence of two insertions in the polypeptides that are present in all other proteobacterial NapA as discussed above (Fig 3C). Third, both D. deulfuricans and S. thermophilum NapG do not have a putative twin arginine sequence motif that in E. coli is known to locate NapG into the periplasmic compartment [[4]. Finally, and most notably, neither cluster includes a napB gene (Fig. 2).

The current consensus of opinion is that NapA evolved from cytoplasmic eubacterial assimilatory nitrate reductases, for which the electron donors are ferredoxin or flavodoxin [[9,[0]. It is suggested that, during evolution, these cytoplasmic enzymes picked up a signal peptide and were translocated to the periplasmic compartment. We propose that, once in the periplasmic compartment, the nitrate reductase would have evolved a new electron donor, NapB, able to exploit electrons from the respiratory electron transport chain. The monomeric NapAs of D. desulfuricans and S. thermophilum are the closest known relatives of, and a similar size to, the ferredoxin-dependent assimilatory nitrate reductase (NarB) of Synechococcus sp. It is reasonable to speculate that the non-heme, iron–sulfur protein, NapG, can be a direct electron donor to D. desulfuricans and S. thermophilum NapA but that to do this it must pass through the TAT system as a mature NapAG complex taking advantage of the NapA TAT signal sequence. In S. thermophilum the NapC protein most likely serves to reduce NapG but in D. desulfuricans the presence of a membrane-associated NapH suggests a route of electron transfer from quinol to NapG via NapH that parallels that proposed for other NapGH containing bacteria [[4]. The D. desulfuricans and S. thermophilum Nap systems might therefore be evolutionary links between the ferredoxin-dependent assimilatory nitrate reductases and the more mature periplasmic nitrate reductase systems of α, β and γ proteobacteria for which a di-heme c-type cytochrome, NapB, is the electron donor.

Acknowledgements

The authors are grateful to Dr. Ligia Saraiva and Dr. Ines Pereira, ITQB, Oeiras, Portugal for the initial D. desulfuricans 27774 DNA sample; to Dr. Jörg Simon for access to his NapC/NrfH data base; to Dr. Douglas Samyahumbi for helpful advice about growing sulphate-reducing bacteria; and to Professor Lynne Macaskie for allowing us to use her anaerobic cabinet facilities. This research was funded by Project grants to DJR and JAC from the UK Biotechnology and Biological Sciences Research Council, and EC Programme Framework V contract EVKI-CT2000–00054.

References

  1. [1].
  2. [2].
  3. [3].
  4. [4].
  5. [5].
  6. [6].
  7. [7].
  8. [8].
  9. [9].
  10. [10].
  11. [11].
  12. [12].
  13. [13].
  14. [14].
  15. [15].
  16. [16].
  17. [17].
  18. [18].
  19. [19].
  20. [20].
View Abstract