A Mutant of Paracoccus denitrificans with Disrupted Genes Coding for Cytochrome c₅₅₀ and Pseudoazurin Establishes These Two Proteins as the In Vivo Electron Donors to Cytochrome cd₁ Nitrite Reductase

Abstract

In Paracoccus denitrificans, electrons pass from the membrane-bound cytochrome bc₁ complex to the periplasmic nitrite reductase, cytochrome cd₁. The periplasmic protein cytochrome c₅₅₀ has often been implicated in this electron transfer, but its absence, as a consequence of mutation, has previously been shown to result in almost no attenuation in the ability of the nitrite reductase to function in intact cells. Here, the hypothesis that cytochrome c₅₅₀ and pseudoazurin are alternative electron carriers from the cytochrome bc₁ complex to the nitrite reductase was tested by construction of mutants of P. denitrificans that are deficient in either pseudoazurin or both pseudoazurin and cytochrome c₅₅₀. The latter organism, but not the former (which is almost indistinguishable in this respect from the wild type), grows poorly under anaerobic conditions with nitrate as an added electron acceptor and accumulates nitrite in the medium. Growth under aerobic conditions with either succinate or methanol as the carbon source is not significantly affected in mutants lacking either pseudoazurin or cytochrome c₅₅₀ or both these proteins. We concluded that pseudoazurin and cytochrome c₅₅₀ are the alternative electron mediator proteins between the cytochrome bc₁ complex and the cytochrome cd₁-type nitrite reductase. We also concluded that expression of pseudoazurin is mainly controlled by the transcriptional activator FnrP.

The electron transport system of the facultative organism Paracoccus denitrificans is one of the best-characterized electron transport systems of bacteria (6, 15). Many of its components are closely related to their counterparts in the mitochondrial respiratory chain. There is a good understanding of how nonmitochondrial features, such as oxidases that are alternatives to cytochrome aa₃, along with the apparatus for catalyzing oxidation of C₁ compounds or reduction of nitrate through to nitrogen, are connected to the underlying mitochondrion-type respiratory chain (6, 32). A key component in such connections is often considered to be cytochrome c₅₅₀ (analogous to cytochrome c₂ in photosynthetic bacteria). This protein has the same tertiary structure as mitochondrial cytochrome c and a similar reduction potential (32). Thus, it has been reasonable to propose that it is an electron acceptor from the cytochrome bc₁ complex and therefore the electron donor to periplasmic components of the electron transport system that receive electrons via this complex. One such component is the cytochrome cd₁-type nitrite reductase, with which cytochrome c₅₅₀ has been shown to interact in vitro (27, 39, 42). However, if cytochrome c₅₅₀ does indeed carry electrons from the cytochrome bc₁ complex to this enzyme, it cannot be the only component of the electron transfer system capable of this reaction. This is because a mutant of P. denitrificans that lacks the gene (cycA) encoding cytochrome c₅₅₀ is still able to reduce nitrite; indeed, no respiratory activity of P. denitrificans is lost as a result of interruption of the cycA gene (44). The role of the cytochrome bc₁ complex in electron delivery to cytochrome cd₁ has been established through studies with both inhibitors (2) and mutants specifically deficient in this complex (33). In the case of nitrite reduction, it was proposed by Moir and Ferguson (30) that the absence of cytochrome c₅₅₀ was compensated for by another protein that was able to act as an electron donor to cytochrome cd₁. On the basis of the finding that nitrite respiration was much more sensitive to the copper chelator diethyldithiocarbamate in the cytochrome c₅₅₀-deficient mutant than in the wild type (30) and because cytochrome cd₁ does not contain copper, it was suggested that the substitute protein might be pseudoazurin. The latter protein is a molecule with a single type I copper center per polypeptide chain and is a member of a family of Greek-key fold proteins that includes bacterial azurins and plant plastocyanins. However, another possible interpretation has been put forward on the basis of the observations made with diethyldithiocarbamate (24). A further indication of the importance of a copper protein in the absence of cytochrome c₅₅₀ came from the observation that the mutant grew anaerobically much more poorly in growth media with low copper contents than did the wild-type organism (33, 45). Furthermore, the nitrite reductase exhibits activity in vitro with pseudoazurin as an electron donor (29). A critical and definitive test of the proposal that cytochrome c₅₅₀ and pseudoazurin are alternative electron donors to cytochrome cd₁-type nitrite reductase would be provided by an examination of the properties of a mutant of P. denitrificans that has interruptions in both the cytochrome c₅₅₀ (cycA) and pseudoazurin (pazS) genes. This is the subject of the present paper.

MATERIALS AND METHODS

Bacterial strains and growth.

The bacterial strains and plasmids used in this study are described in Table 1. Escherichia coli strains were grown aerobically in Luria-Bertani medium at 37°C. P. denitrificans strains were routinely grown at 37°C in minimal succinate medium (9). Media used for anaerobic growth contained KNO₃ at a concentration of 100 mM unless specified otherwise. Tests for aerobic growth of P. denitrificans strains on methanol or methylamine were performed by using the media described by Alefounder and Ferguson (1). Antibiotics were added as required at the following final concentrations: ampicillin, 100 μg ml⁻¹; kanamycin, 25 μg ml⁻¹; rifampin, 40 μg ml⁻¹; spectinomycin, 25 μg ml⁻¹; and streptomycin, 25 μg ml⁻¹.

TABLE 1.

Bacterial strains and plasmids used in this study

Organism or plasmid	Characteristics	Reference or source
Paracoccus denitrificans Pd1222	Spr Rifr, restriction deficient	13
Paracoccus denitrificans Pd21.31	Pd1222 cycA	44
Paracoccus denitrificans Pd77.71	Pd1222 nnr	47
Paracoccus denitrificans Pd29.31	Pd1222 fnrP::Kmr	47
Paracoccus denitrificans Pd92.30	Pd1222 fnrP::Kmrnnr	47
Paracoccus denitrificans DPΩNOSR	Pd1222 nosR::Smr	This study
Paracoccus denitrificans IP1013	Pd1222 pazS::Kmr	This study
Paracoccus denitrificans IP1121	Pd21.31 pazS::Kmr	This study
Escherichia coli XL1-Blue	supE44 hsdR17 recA1 endA1 gyrA46 thi relA1 lac F′ [proAB+lacIqlacZΔM15 Tn10(tetr)]	8
Escherichia coli GM2163	F′ dam-13::Tn9 dcm-6 hsdR2 leuB6 his-4 thi-1 ara-14 lacY1 galK2 galT22 xyl-5 mtl-1 rpsL136 tonA31 tsx-78 supE44 McrA− McrB−	41
Escherichia coli S17-1	pro thi hsdR− M+recA Strr with integrated RP4 2-(tet::Mu) (kan::Tn7)	41
Escherichia coli NM554	F−recA13 araD139 Δ(ara-leu) Δ(lac)X74 galU galK hsdR2 mcrB1 rpsL	38
pBluescript SK(+)	General cloning vector; Apr	Stratagene
pUC18	General cloning vector; Apr	Amersham Pharmacia Biotech
pUC4K	Kmr cassette vector; Apr Kmr	Amersham Pharmacia Biotech
pWE16	Cosmid cloning vector	Stratagene
pGRPd1	Mobilizable suicide vector; Apr Spr/Smr	44
pHP45Ω	Spr/Smr Ω cassette vector; Apr Spr/Smr	36
pAR0181	Mobilizable suicide vector; Kmr	35
pRK2013	Helper plasmid, carries tra genes; Kmr	16
pBK11.paz	1.9-kb EcoRI-SphI fragment containing the 5′ region of the P. denitrificans pseudoazurin gene, cloned in pBK11a	R. J. M. van Spanning, Vrije Universiteit, Amsterdam, The Netherlands
pIP0724	EcoRI-SphI fragment from pBK11.paz excised with EcoRI-HindIII and cloned in pBluescript SK(+)	This study
pDP0725	EcoRI-SphI fragment from pHP45Ω cloned into MunI-cut pIP0724	This study
pDP0726	nosR::Ω from pDP0725 ligated to EcoRI-cut pAR0181	This study
cIP0727A	Approximately 50 kb of genomic DNA from P. denitrificans DPΩNOS cloned in pWE16	This study
cIP0728	cIP027A cut with EcoRI and self-ligated	This study
pIP0729	XhoI-EcoRI fragment from cIP0728 cloned in pBluescript SK(+)	This study
pIP0730	pIP0729 cut with SmaI and self-ligated	This study
pIP0731	BamHI-Acc65I fragment from pIP0729 cloned BamHI-KpnI in pUC18	This study
pIP0515	pIP0731 cut with BclI and ligated to the BamHI-excised Kmr cassette from pUC4K	This study
pIP0516	BamHI-EcoRI fragment from pIP0515 cloned in pGRPd1	This study

^aSee reference 21.

General DNA manipulations.

General DNA manipulations were performed as described by Sambrook et al. (40). DNA sequencing was performed by the PNACL facility at the University of Leicester, Leicester, United Kingdom. Southern blotting was performed by using the digoxigenin labeling system (Boehringer Mannheim) according to the manufacturer's instructions.

Cloning and disruption of the pazS gene.

The P. denitrificans DNA fragment cloned in pBK11.paz was excised with EcoRI plus HindIII and cloned in EcoRI-HindIII-digested pBluescript SK(+) to generate pIP0724 (Fig. 1). For construction of DPΩnosR, pIP0724 was digested with MunI and ligated to the Ω cassette isolated from EcoRI-cut pHP45Ω. The resulting plasmid, pDP0725, was digested with EcoRI, and the DNA fragment containing the Ω-disrupted nosR gene was cloned in pARO181. The resulting plasmid, pDP0726, was transferred to P. denitrificans by triparental conjugation by using the helper plasmid pRK2013 (14). Sm^r Km^s exconjugants were designated DPΩnosR. DNA contiguous with the introduced Ω cassette in DPΩnosR was isolated by cosmid cloning essentially as described by Viebrock and Zumft (48). DPΩnosR genomic DNA was digested with Sau3A to obtain a fragment length of approximately 50 kb and was ligated to BamHI-cut and dephosphorylated pWE16. Recombinant molecules were packaged into λ phage by using the Gigapack III XL system (Stratagene) and were transfected into E. coli NM554. Clones carrying the Ω cassette were selected as Ap^r Sp^r. One of the cosmids obtained, cIP0727A, was digested with EcoRI and self-ligated. Transformed E. coli cells were plated onto media containing spectinomycin in order to isolate clones carrying the Ω cassette. The resulting cosmid, cIP0728, consisted of a single EcoRI fragment of P. denitrificans DNA (approximately 30 kb) cloned in pWE16. Cosmid cIP0728 was digested with EcoRI plus XhoI, and the resulting restriction fragments were isolated and ligated to EcoRI-XhoI-digested pBluescript SK(+). Transformed E. coli cells were again plated onto media containing spectinomycin. This procedure yielded plasmid pIP0729, which contained the complete pazS gene together with nosC, the 5′ region of nosR, and the Ω cassette. pIP0729 was then digested with SmaI and self-ligated to obtain pIP0730, which contained pazS plus approximately 100 bp of upstream DNA and 1.4 kb of downstream DNA (Fig. 1).

FIG. 1.

Physical and restriction map of the P. denitrificans chromosomal DNA region containing the pazS gene. The map at the top is a restriction map. B, BamHI; Bc, BclI; E, EcoRI; M, MunI; Sm, SmaI; Sp, SphI; X, XhoI. The arrows indicate the positions of the pazS and nosC genes and the nosR 5′ region. The lines below the map are the P. denitrificans DNA fragments cloned in this study. Ω and Km^r indicate the positions of Ω (spectinomycin-streptomycin) and kanamycin resistance cassettes, respectively.

For pazS disruption, the pazS-containing BamHI-Acc65I fragment from pIP0729 was isolated and cloned in BamHI-KpnI-digested pUC18. The resulting plasmid, pIP0731, was digested with BclI and ligated to the Km^r cassette isolated from BamHI-cut pUC4K to produce pIP0515. This construct was digested with BamHI plus EcoRI, and the DNA fragment containing the disrupted pazS gene was ligated to BamHI-EcoRI-digested pGRPd1 to obtain pIP0516. This plasmid was transformed into E. coli S17-1 and transferred to P. denitrificans Pd1222 (wild type) and Pd2131 (cycA) by biparental mating (5).

Analytical methods.

Preparation of total soluble protein fractions from P. denitrificans strains, determination of protein contents, sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), and immunoblotting were performed as described previously (34). Pseudoazurin was polymerized with glutaraldehyde as described by Hennig and Neupert (20), and antibodies were raised in sheep by the Scottish Antibody Production Unit, Law Hospital, Carluke, Scotland. Nitrite accumulation in the medium of growing cultures was determined as described by Nicholas and Nason (31).

Nucleotide sequence accession number.

The P. denitrificans pazS sequence, together with approximately 1,500 flanking base pairs, has been deposited in the GenBank database under accession no. AF334183.

RESULTS

Cloning and sequencing of the P. denitrificans pseudoazurin structural gene, pazS.

The starting material for the cloning of pazS from P. denitrificans was a DNA fragment (pBK11.paz) that contained (GenBank accession no. AJ010260) 135 bp of the 5′ end of pazS (Fig. 1). This fragment had been found unexpectedly, oriented in the opposite direction, adjacent to the nosC and nosR genes in a cluster of nos genes that are required for formation of nitrous oxide reductase (N. F. W. Saunders and R. J. M. van Spanning, unpublished observations). Because previous attempts to clone pazS had proved to be problematic, the strategy adopted in the present work was to generate a marked mutation in the chromosomal nosR gene and then to isolate DNA containing the marker and the flanking region containing the full-length pazS gene. The cosmid-based cloning procedure leading to plasmid pIP0730 is described in Materials and Methods.

The protein sequence translated from the P. denitrificans pazS structural gene sequence was in exact agreement with that obtained by direct protein sequencing (25). These gene and protein sequences are very similar to their counterparts in the closely related organism (37) Paracoccus pantotrophus (11, 25). However, immediately beyond the structural gene region the DNA sequences of P. pantotrophus and P. denitrificans had considerable differences. At the 5′ end, these differences not only were apparent in the noncoding region but also reflected much variation in the periplasmic targeting sequences, although this variation was conservative (Fig. 2). Nevertheless, in both organisms the pazS gene clearly had an anaerobox of the Fnr type centered at bp 80.5 upstream of the translational initiation point and most likely at a similar position, bp 41.5, relative to the transcriptional start point, if the latter was assumed to be identically positioned in P. denitrificans and P. pantotrophus (6, 25) (Fig. 2).

FIG. 2.

Nucleotide sequences of the 5′ ends of the pazS genes of P. denitrificans and P. pantotrophus together with the amino acid sequences of the signal peptides and the N-terminal regions of the mature proteins. The plus sign indicates the experimentally determined transcriptional initiation site for P. pantotrophus pazS. The shaded area is a presumed binding site for the transcriptional activator FnrP. Asterisks indicate DNA sequence identity.

Two transcription factors, NNR and FnrP, bind differentially to anaeroboxes in P. denitrificans (46, 47). The synthesis of pseudoazurin by mutants of P. denitrificans deficient in either FnrP or NNR was therefore examined. The former protein, along with anaerobic conditions, could be shown to be essential for significant expression of pseudoazurin, as judged by immunoblotting (Fig. 3), whereas the presence of NNR appeared to be less significant for pazS expression. The attenuated expression in the NNR mutant may be explained by the poor growth of the cells under anaerobic conditions, a consequence of their inability to reduce nitrite or nitric oxide (46).

FIG. 3.

Pseudoazurin expression in P. dentrificans nnr and fnrP mutant strains. Proteins were separated by SDS-PAGE, and pseudoazurin was detected by immunoblotting. Lane 1, Pd1222 (wild type); lane 2, Pd29.31 (fnrP::Km^r); lane 3, Pd77.71 (nnr); lane 4, Pd92.30 (fnrP::Km^r nnr). The lanes contained approximately 20 μg of total protein. The position of a 16.5-kDa molecular mass marker is indicated on the left.

Targeted disruption of pazS in P. denitificans Pd1222 (wild type) and Pd21.31 (cycA).

The cloned pazS gene was disrupted by insertion of a kanamycin resistance cassette, and the disrupted gene was cloned in the mobilizable suicide vector pGRPd1. The resulting plasmid, pIP0516, was transferred by conjugation into P. denitrificans strains Pd1222 (wild type) and Pd21.31 (carrying an unmarked mutation in cycA). Kanamycin-resistant mutants of P. denitrificans, putatively carrying interruptions in either pazS alone or in pazS and cycA, were analyzed to identify mutants that had undergone the required double homologous recombination event leading to a genomic kanamycin-marked pazS gene. A single homologous recombination event would have resulted in both the marked pazS gene and the original intact pazS gene being present. There was no antibiotic resistance marker that could be used for this discrimination (the BamHI/EcoRI component of pGRPd1 present in pIP0516 did not contain the spectinomycin-streptomycin resistance locus), and so the presence of a large region of DNA originating from the vector pGRPd1 was used as a probe to detect the unwanted cells that had undergone only a single recombination event rather than the required double recombination event.

PvuII digests of genomic DNA from one in five putative pseudoazurin single mutants did not hybridize to the pGRPd1-specific probe, suggesting that each of these mutants was a product of a double recombination event. PvuII was used because there are no restriction sites for this enzyme within either the cloned pazS DNA or the kanamycin resistance cassette. The pazS gene was used as a second probe for the appearance in the P. denitrificans genome of a pazS gene interrupted by the kanamycin resistance cartridge. This proved to be problematic as only very weak probes could be generated in the present study when paz alone was labeled. Thus, we labeled pIP0731, which is a pUC18-derived construct containing a BamHI/Acc65I pazS fragment. Southern blotting with this probe showed that the putative mutants indeed had an interrupted pazS gene. One of these mutants, IP1013, was selected for further characterization. Immunoblotting (Fig. 4) confirmed the absence of the pseudoazurin protein from cells of this mutant after growth under anaerobic conditions.

FIG. 4.

Detection of pseudoazurin and cytochrome c₅₅₀ in total soluble extracts from wild-type and mutant strains of P. denitrificans. Proteins were separated by SDS-PAGE, and pseudoazurin (A) or cytochrome c₅₅₀ (B) was detected by immunoblotting. Lane 1, molecular weight markers; lane 2, Pd1222 (wild type); lane 3, Pd21.31 (cycA); lane 4, IP1013 (pazS::Km^r); lane 5, IP1121 (cycA pazS::Km^r). The lanes contained approximately 20 μg of total protein.

Coincidentally, one in five putative mutants with mutations in the pazS gene that resulted from a double recombination event were again isolated following conjugation of pIP0516 into P. denitrificans Pd21.31 (the cycA mutant). One of these mutants, designated IP1121, for which Southern blotting clearly demonstrated the absence of the vector pGRPd1 but the presence of the kanamycin insertion in pazS, was selected for further study.

The absence of pseudoazurin from the double cytochrome c₅₅₀ and pseudoazurin mutant, IP1121, was further confirmed by Western blotting. It is clear (Fig. 4) that whereas wild-type cells or the strain with the unmarked mutation for cytochrome c₅₅₀ contained pseudoazurin, both types of pazS mutants lacked this protein. It was noticed that the absence of cytochrome c₅₅₀ did not result in an increase in the amount of pseudoazurin in the P. denitrificans Pd21.31 (cycA) cells. As expected, cytochrome c₅₅₀ was still expressed in the single pazS mutant, but strikingly, it was expressed at a higher level than in the wild-type cells (Fig. 4), suggesting there was some compensatory effect of the absence of pseudoazurin on expression of this cytochrome.

The construction of the pazS cycA double mutant was designed to show whether in vivo nitrite reductase activity was compromised. In this context it was important to show that loss of both pseudoazurin and cytochrome c₅₅₀ from the cell had not caused a loss of cytochrome cd₁ from the cell. The Western blot in Fig. 5 shows that the latter cytochrome was still present after anaerobic growth, but at lower levels than those observed for the wild type or the pazS and cycA single mutants. The explanation for this is that maximal transcription of the nirS gene, coding for cytochrome cd₁-type nitrite reductase, requires production of the enzyme's reaction product, nitric oxide, in order to activate the transcription factor, NNR, that stimulates the expression of nirS. (46). Clearly, in a mutant unable to form nitric oxide the amount of cytochrome cd₁-type nitrite reductase synthesized is considerably attenuated. Thus, the decrease in the amount of cytochrome cd₁-type nitrite reductase as a consequence of the absence of both cytochrome c₅₅₀ and pseudodazurin is itself evidence that these molecules are the two physiological electron donors.

FIG. 5.

Detection of cytochrome cd₁ in total soluble extracts from wild-type and mutant strains of P. denitrificans. Proteins were separated by SDS-PAGE, and cytochrome cd₁ was detected by immunoblotting. Lane 1, molecular weight markers; lane 2, Pd1222 (wild type); lane 3, Pd21.31 (cycA); lane 4, IP1013 (pazS::Km^r); lane 5, IP1121 (cycA pazS::Km^r). The lanes contained approximately 20 μg of total protein.

Growth characteristics.

It was not thought that aerobic growth on a heterotrophic carbon and free energy source should require either cytochrome c₅₅₀ or pseudoazurin, because cytochrome c₅₅₂ is a protein that accepts electrons from the cytochrome bc₁ complex and delivers them to the cytochrome aa₃ oxidase (33), while cytochrome ba₃, which is a quinol oxidase, functions independent of either pseudoazurin or cytochrome c₅₅₀ (15, 33). This expectation was confirmed by the finding that aerobic growth with succinate as a carbon and free energy source was not affected in any of the mutants. Thus, the stationary phase was reached after approximately 14 h with a turbidity (optical density at 650 nm [OD₆₅₀]) between 2.5 and 3 for both wild-type cells and cells having mutations in the gene(s) encoding either cytochrome c₅₅₀ or pseudoazurin or both.

Aerobic growth on methanol or methylamine is more dependent on the cytochrome bc₁-dependent oxidase, cytochrome aa₃, than growth on succinate is. In Methylotrophus methylophilus a pseudoazurin has been implicated in electron transfer to oxidases from the methylamine dehydrogenase (4). However, growth of P. denitrificans on either of these two one-carbon sources was also not affected by the absence of both pseudoazurin and cytochrome c₅₅₀, implying that neither of these molecules has an obligatory role in this case, a conclusion for pseudoazurin that is consistent with the presence of the upstream anaerobox (Fig. 2).

Growth under anaerobic conditions with nitrate as an added electron acceptor and succinate as a carbon and free energy source gave a different result. The double mutant (cycA pazS) grew to a lower final turbidity than the other organisms; nitrite accumulated and remained in the medium (Fig. 6). When both the cell density and the nitrite concentration in the medium were measured under these conditions, it was clear that the nitrite concentration in wild-type and single mutant cultures rose and then declined to zero (Fig. 6); gas production resulting from denitrification proceeding beyond nitrite could clearly be observed in the medium. In contrast, the double mutant cells converted nitrate to nitrite almost quantitatively (Fig. 6), and production of gas did not occur. The formation of 25 mM nitrite correlated with cessation of cell growth, which occurred at an OD₆₅₀ of 0.4, compared with an OD₆₅₀ of 1.2 for the wild-type and single mutants (Fig. 6). This was expected, because once all the added nitrate had been reduced to nitrite, there was no useable electron acceptor available in such anaerobic cultures (i.e., the nitrite could not be used owing to the double mutation, which in turn meant that nitric and nitrous oxides could not be generated for respiration). All types of cells took approximately 12 h, after any lag phases, to reach the final OD_650, showing that the double mutant had a similar time course of growth but produced less biomass. These observations clearly imply that the double mutant was unable to reduce nitrite, in contrast to the wild-type strain and strains having a single mutation in either cycA or pazS.

FIG. 6.

Nitrite accumulation (A) and growth curves (B) for cultures of the P. denitrificans wild type and cycA, pazS, and cycA pazS mutants. The data are typical of data for three replicates for each strain; in each case variable lag phases following inoculation were omitted to permit comparisons between strains. For these experiments the concentration of nitrate initially present was lowered to 25 mM in order to facilitate measurement of the stoichiometry of nitrite accumulation; concentrations of accumulated nitrite higher than 25 mM are toxic. Symbols: ▪, wild type; ▴, pazS mutant; ○, cycA mutant; ▿, cycA pazS mutant.

The transient accumulation of nitrite (Fig. 6) in cultures of anaerobically grown wild-type P. denitrificans has been observed previously (7; Ferguson, unpublished observations) and is thought to reflect the later onset of nitrite reductase synthesis compared to the onset of nitrate reductase synthesis. The presence of nitrite at millimolar levels disrupts the pattern of growth away from the ideal behavior (33). In the wild-type and single mutant cells nitrite formation correlated with the early stage of growth following any lag phase, but why different strain-dependent nitrite concentrations were reached (Fig. 6) before the concentrations declined to zero is not clear. Thus, lower nitrite concentrations were attained in the media of the cycA mutant cells than in the media of either the wild type or the pazS mutant. This observation suggests that there may be subtly different consequences of the absence of either cytochrome c₅₅₀ or pseudoazurin. The reason for this behavior is not clear but may be related to the requirement for nitric oxide formation to in turn stimulate the production of more nitrite reductase (see above). Thus, there may be a relationship between the rate of nitrite reduction in the early stages of growth and the subsequent formation of more nitrite reductase.

DISCUSSION

Whereas the absence of either pseudoazurin or cytochrome c₅₅₀ has no major effect on anaerobic respiration of nitrite in P. denitrificans, this process is clearly severely compromised in the P. denitrificans mutant that lacks both these periplasmic electron transfer proteins. Thus, either cytochrome c₅₅₀ or pseudoazurin is required for one, or most probably the only, step of the electron transfer pathway from the cytochrome bc₁ complex to cytochrome cd₁. This conclusion is in agreement with the recent report that addition of pseudoazurin stimulated the rate of electron transfer from succinate to nitrite in right-side-out vesicles, which must have retained externally bound cytochrome cd₁-type nitrite reductase, from cytochrome c₅₅₀-deficient P. denitrificans (23). A previous proposal (28) that one or more additional c-type cytochromes act as alternatives to cytochrome c₅₅₀ in electron transfer to nitrite reductase is not supported by the present work. This conclusion is enhanced by the demonstration (33) that another c-type cytochrome, cytochrome c₅₅₂, functions only in electron transfer from the cytochrome bc₁ complex (or from amicyanin during growth on methylamine) to cytochrome aa₃ oxidase. Cells of a P. denitrificans mutant deficient in cytochrome c₅₅₂ denitrify as well as cells carrying a disrupted cytochrome c₅₅₀ gene, but the former, unlike the latter, do not require copper in the growth medium for denitrification to occur. The implication is again that only a copper-containing protein can substitute for cytochrome c₅₅₀ in nitrite reduction during denitrification. As might be predicted on this basis, a double deletion mutant with deletion of both cytochrome c₅₅₀ and cytochrome c₅₅₂ also requires copper in the medium for anaerobic growth on succinate plus nitrate (33). Cytochrome c₅₅₂ is expressed constitutively, and its globular heme-containing domain has a structure very similar to that of cytochrome c_550, although it possesses an acidic patch and overall has a less positive charge (18). Nevertheless, although it shares a capability with cytochrome c₅₅₀ and (as implied by the present work) pseudoazurin to interact with the cytochrome bc₁ complex, it seems to be specific as an electron donor to cytochrome aa₃ (33). The acidic patch or the transmembrane anchor of cytochrome c₅₅₂ may prevent its productive interaction with cytochrome cd₁.

In an early in vitro study it was concluded that cytochrome c₅₅₀, but not pseudoazurin, is the electron donor to cytochrome cd₁-type nitrite reductase (26, 42). However, recent work with P. pantotrophus proteins has shown that pseudoazurin is an effective electron donor to Paracoccus cytochrome cd₁, provided that the latter molecule is first activated by reduction (3, 39). Presumably, this explains why Martinkus et al. (26) observed little activity with pseudoazurin, in contrast to Moir et al. (29), who in retrospect fortuitously activated cytochrome cd₁ before their assay. Since the presence of at least cytochrome c₅₅₀ or pseudoazurin is required for in vivo nitrite respiration, any activator for this cytochrome cd₁ must be a different protein for which the midpoint potential is less than 60 mV and thus able to change the inactive oxidized protein into the active reduced state (3, 22, 39).

One conclusion of the present work is that not only must pseudoazurin and cytochrome c₅₅₀ be alternative electron donors to cytochrome cd₁, but they must also be alternative acceptors from the cytochrome bc₁ complex. This conclusion leads to the question of how two such structurally distinct proteins as pseudoazurin and cytochrome c₅₅₀ can interact with the same reductant and oxidant. In the case of interaction with cytochrome cd₁-type nitrite reductase, it has been proposed that pseudoazurin and cytochrome c₅₅₀ each interact with a hydrophobic patch on the surface of the enzyme, with the reaction promoted by patches of positive charges on the cytochrome c₅₅₀ or pseudoazurin and negative charges on cytochrome cd₁ (50). There is at least one precedent for the cytochrome bc₁ complex interacting with a copper protein. In the green sulfur bacterium Chloroflexus aurantiacus there is no evidence that there are water-soluble c-type cytochromes analogous to cytochrome c₅₅₀, and the cytochrome bc₁ complex is thought to use auracyanin, a copper protein related to pseudoazurin (43), as its electron acceptor. In thylakoids of both cyanobacteria and Arabidopsis, it is known that either cytochrome c₆ or plastocyanin, which is related to pseudoazurin, is an alternative molecule for transfer of electrons from the cytochrome bf complex to photosystem I (12, 17). Negatively charged patches on cytochrome c₆ or plastocyanin are thought to present a recognition feature to complementary charged regions on partner proteins. Thus, there are several good examples of an organism that is able to use a type I copper protein instead of a c-type cytochrome in an electron transfer step from membrane-bound electron transfer complexes that occur in both respiratory and photosynthetic systems.

It has been deduced that only cytochrome c₅₅₁, and not azurin, could act as an electron donor to the cytochrome cd₁-type nitrite reductase from Pseudomonas aeruginosa (49); this contrasts with the present work. There is much evidence which shows that azurin can function as an electron donor to the P. aeruginosa nitrite reductase in vitro, and thus the data are slightly surprising. The apparent failure of azurin to sustain substantial nitrite respiration in this organism in the absence of cytochrome c₅₅₁ might be attributed to several factors. Among these factors could be a failure of azurin to accept electrons from the cytochrome bc₁ complex. An explanation based on the possibility of a failure to express cytochrome cd₁-type nitrite reductase in the absence of azurin was eliminated by Vijgenboom et al. (49). In fact, in P. aeruginosa it appears that denitrifying conditions do not promote the expression of azurin (49), and thus in this organism only one electron-carrying protein, cytochrome c₅₅₁, can act between the cytochrome bc₁ complex and the nitrite reductase. The situation in P. aeruginosa has been further complicated by the subsequent finding that a second c-type cytochrome, NirC, can also act in vitro as a donor to cytochrome cd₁-type nitrite reductase (19), a result that contrasts with the data of Vijgenboom et al. (49).

The characteristic anaerobox sequence that is found upstream of the pazS gene in P. denitrificans could be a binding site for FnrP or NNR, two molecules which bind to similar sequences, although the context of these sequences must confer specificity (e.g., for NNR binding rather than FnrP binding upstream of nirS [47]) for one of the two transcription factors. The results in the present paper suggest that it is FnrP that is the more important of these transcription factors for activating transcription of pazS. Stimulation of pseudoazurin production by anaerobiosis signaled via FnrP rather than by NNR signaled by nitric oxide would be consistent with the finding (N. F. W. Saunders and S. J. Ferguson, unpublished data) that microanaerobic growth conditions stimulate production of not only the cytochrome c peroxidase, which is known to be under FnrP control (47), but also pseudoazurin. The −10 regions of the pazS genes of the two Paracoccus species have a CCTA sequence in common. The TA pair, in conjunction with the presence of an anaerobox at position −41.5, is diagnostic of the class II-type CAP family of promoters, in which the −10 region is likely to be dependent upon a σ⁷⁰ factor (10).