A Critical Role for the cccA Gene Product, Cytochrome c₂, in Diverting Electrons from Aerobic Respiration to Denitrification in Neisseria gonorrhoeae

Abstract

Neisseria gonorrhoeae is a microaerophile that, when oxygen availability is limited, supplements aerobic respiration with a truncated denitrification pathway, nitrite reduction to nitrous oxide. We demonstrate that the cccA gene of Neisseria gonorrhoeae strain F62 (accession number NG0292) is expressed, but the product, cytochrome c₂, accumulates to only low levels. Nevertheless, a cccA mutant reduced nitrite at about half the rate of the parent strain. We previously reported that cytochromes c₄ and c₅ transfer electrons to cytochrome oxidase cbb₃ by two independent pathways and that the CcoP subunit of cytochrome oxidase cbb₃ transfers electrons to nitrite. We show that mutants defective in either cytochrome c₄ or c₅ also reduce nitrite more slowly than the parent. By combining mutations in cccA (Δc₂), cycA (Δc₄), cycB (Δc₅), and ccoP (ccoP-C368A), we demonstrate that cytochrome c₂ is required for electron transfer from cytochrome c₄ via the third heme group of CcoP to the nitrite reductase, AniA, and that cytochrome c₅ transfers electrons to nitrite reductase by an independent pathway. We propose that cytochrome c₂ forms a complex with cytochrome oxidase. If so, the redox state of cytochrome c₂ might regulate electron transfer to nitrite or oxygen. However, our data are more consistent with a mechanism in which cytochrome c₂ and the CcoQ subunit of cytochrome oxidase form alternative complexes that preferentially catalyze nitrite and oxygen reduction, respectively. Comparison with the much simpler electron transfer pathway for nitrite reduction in the meningococcus provides fascinating insights into niche adaptation within the pathogenic neisseriae.

INTRODUCTION

Two species of pathogenic bacteria, Neisseria gonorrhoeae and Neisseria meningitidis, occupy contrasting niches within the human body. Meningococci typically reside in well-aerated regions of the upper respiratory tract, but gonococci are, at least in the female host, surrounded by obligate anaerobes and lactobacilli that reduce the limited amounts of available oxygen to hydrogen peroxide. Despite a high respiratory capacity that is comparable to that of aerobic bacteria, gonococcal physiology is typical of a microaerophile. It has only a single cytochrome oxidase, cytochrome cbb₃, which in other bacteria has a very high affinity for oxygen (1–5). We previously suggested that its high respiratory capacity enables gonococci to minimize oxidative damage from reactive oxygen species generated both from its own respiratory metabolism and by other bacteria that share its environment (6). When the oxygen supply is growth limiting, it also catalyzes a truncated denitrification pathway in which nitrite is reduced to nitric oxide by AniA, a copper-containing nitrite reductase of the NirK family (7–12). Nitric oxide is then reduced to N₂O by a single subunit nitric oxide reductase, NorB, that receives electrons directly from ubiquinol (13).

Gonococci are a prolific source of six c-type cytochromes during aerobic growth. These are the c₁ component of the cytochrome bc₁ complex, the CcoP and CcoO subunits of the cytochrome oxidase, and two cytochromes, designated c₄ and c₅, that transfer electrons between the bc₁ complex and the terminal oxidase (6). These cytochromes are tightly associated with the cytoplasmic membrane: their purification and biochemical characterization in vitro are extremely challenging and present serious risks of generating nonphysiological artifacts.

The sixth c-type cytochrome readily detected during aerobic growth is an NO-binding protein, CycP or cytochrome c′, which provides some protection against nitrosative stress (14–16). During oxygen-limited growth, a seventh c-type cytochrome, the cytochrome c peroxidase (Ccp), is synthesized as part of the FNR (fumarate and nitrate reductase) regulon but is absent from N. meningitidis. Both cytochrome c peroxidase and cytochrome c′ are outer membrane lipoproteins that provide protection against reactive oxygen and reactive nitrogen species, respectively (16, 17). The nitrite reductase, AniA, and a lipid-associated azurin, Laz, are also outer membrane redox proteins (18–21). However, it is unknown how electrons are transferred across the periplasm from the cytochrome bc₁ complex in the inner membrane to these outer membrane proteins.

Analysis of the genomes of both pathogenic and commensal Neisseria strains (using the Gonococcal Genome Sequencing Project database at http://stdgen.northwestern.edu) revealed the presence of a gene, cccA, for an eighth c-type cytochrome that in N. gonorrhoeae has been designated cytochrome c₂. Although this cytochrome has never been visualized by SDS-PAGE, there is clear evidence that the gene is expressed, that a cccA mutation confers detectable phenotypic changes, and that the expected product with heme attached is readily detected when the cccA gene is expressed from a recombinant plasmid in the heterologous host Escherichia coli (6, 22). It was proposed that the corresponding meningococcal cytochrome, cytochrome c_x, is involved in electron transfer to oxygen (2). However, deletion of the gonococcal cccA gene resulted in a slight increase, not a decrease, in the rate of oxygen reduction, so a possible role for this cytochrome in gonococcal aerobic respiration was discounted (6). This highlighted the need to consider other roles for cytochrome c₂.

To identify its physiological role, we first compared the amino acid sequence of cytochrome c₂ with that of other known proteins using BLAST at NCBI (http://blast.ncbi.nlm.nih.gov). This analysis indicated a possible role in electron transfer to the copper-containing nitrite reductase AniA. We therefore systematically deleted genes singly and in various combinations for proteins implicated in electron transfer to the terminal nitrite reductase to determine pathways for electron transfer to nitrite in N. gonorrhoeae. This analysis revealed a key role for the enigmatic cytochrome c₂, solved several outstanding controversies, and revealed contrasting results with those of a previous study of nitrite reduction in the meningococcus (23). We propose a model to explain how cytochrome c₂ regulates the distribution of electrons between two alternative electron transfer complexes. In the context of niche adaptation, we discuss the contrasting simplicity of electron transfer processes in the closely related meningococcus with the more complex electron transfer chains of the gonococcus.

MATERIALS AND METHODS

Strains, media, and culture conditions.

Bacterial strains and plasmids are listed in Table 1, and oligonucleotide primers (synthesized by Alta Bioscience, Birmingham, United Kingdom) are listed in Table 2. Bacteria were stored as glycerol stocks at either −80°C (Escherichia coli and Neisseria gonorrhoeae strains) or in liquid nitrogen (Neisseria gonorrhoeae strains only). E. coli strains were grown as previously described (6, 24).

Table 1.

Bacterial strains and plasmids used in this work

Strain	Genotype or characteristics	Source or reference
E. coli strains
JM109	recA1 endA1 gyrA96 thi hsdR17 (rK− mK+) relA1 supE44 Δ(lac-proAB) [F′ traD36 proAB lacIqZΔM15]	Promega
N. gonorrhoeae strains
F62	Parental strain	Laboratory stocks
AHGC101	ccoP-C368A kan	24
JCGC100	fnr::ermC	44
JCGC140	ΔaspC-lctP::plac-cccA+ 14 bp upstream, ermC	This work
JCGC142	cccA::cat ΔaspC-lctP::plac-cccA+ 14 bp upstream, ermC	This work
JCGC310	cycB Δ200–279::ermC	This work
JCGC320	ccoP-C368A kan cycB Δ200–279::ermC	This work
JCGC330	ccoP-C368A kan cccA::cat	This work
JCGC340	ccoP-C368A kan cycA::ermC	This work
JCGC800	cycA::ermC	6
JCGC805	cycA::3×FLAG kan	6
JCGC807	cccA::3×FLAG kan	This work
JCGC850	cycB::ermC	6
JCGC851	cccA::kan	6
JCGC852	cycB::ermC cccA::kan	6
JCGC853	cycA::ermC cccA::kan	6
JCGC861	cccA::cat	This work
Plasmids
pGEM T Easy	Commercially available vector for cloning PCR products	Promega
pLES940	Source of the cat gene	45
pAH107	cat gene ligated into the AgeI site of pYL12	This work
pYL12	430 bp of DNA upstream and 331 bp of DNA downstream of the cccA gene in a continuous sequence containing an AgeI restriction site at the end of the upstream DNA ligated into pGEM T-Easy	6
pAH130	502 bp of noncoding DNA upstream of cycB and 595 bp of cycB coding DNA linked by an AgeI site to 612 bp of DNA downstream of cycB in pGEM T-Easy	This work
pAH131	ermC gene ligated into the AgeI site of pAH130	This work
pAHC2FCD	336 bp of downstream DNA of the cccA gene ligated between the XhoI site and the NheI site of pDOC-F	This work
pAHC2F	435 bp of upstream and 465 bp of coding DNA of the cccA gene ligated between the EcoRI site and the KpnI site of pAHC2FCD	This work
pDOC-F	Source of 3×FLAG tag	46
pGCC4	Neisserial insertional complementation system; ermC kan lctP aspC lacIq lacP	47, 48
pGCC4_cccA	pGCC4 containing the cccA gene flanked by 14 bp of upstream DNA and 204 bp of downstream DNA inserted between the PacI and FseI sites of pGCC4	This work

Table 2.

Oligonucleotide primers used in this work

Primer	Sequence (5′–3′)	Restriction site
NTCYC4 FOR	GACCCAATGTGCGCGTACCG
NTCYC4 REV	GTGTGGGAGATATACGGGATTTACTC
NTERY FOR	AGAAGACCGGTTAAGAGTGTGTTGATAGTGC	AgeI
NTERY REV	CAAATTACCGGTAGGCGCTAGGGACC	AgeI
AHpGCC4PacIns	GCTGTGGTATGGCTGTG
AHCptC2Fb	CTACGTTTAATTAAGGAAACGCCAATGAACAC	PacI
AHCptC2R	CTACGTGGCCGGCCACGTTCGGGCGATTTGG	FseI
AHpGCC4_F	CCCGCATCAAACAGCTCGG
AHpGCC4_R	CCATTGTTCGGGCGTAGGG
AHCatAgeIF	ACTGACCGGTGGCAGGCCATGTCTGCC	AgeI
AHCatAgeIR	ACTGACCGGTGCTTACTCCCCATCCCC	AgeI
AHC5-2ndAUF	GGCAGGATGTGCGTCGCC
AHC5-2ndBURb	CTGCCCACCGGTACCGTTTATTAGTCAACGCCGACCGCAGG	AgeI
AHC5-2ndDDF	ACGGTACCGGTGGGCAGGTATCTGCTCCG	AgeI
AHC5-2ndCDR	ACAGGCATCCGGATGCCG
AHC5-2ndint_F	TCGCCGATGCGCTTGCC
AHC5-2ndint_R	CGCCCGACATATCGAGCC
AHC2AUFEcoRI	GTCTAGGAATTCACCGAGCACGCAGG
AHC2BURKpnI	GTCTAGGGTACCGAAAGGTTTGATTTGAATGCCG
AHC2CDRNheI	GTCTAGGCTAGCGGCGGAGGATTTGGCG
AHC2DDFXhoI	GTCTAGCTCGAGCGCGGTACTTTCAGCC
YLC552AUF	GGCAAAATGGGTTACACCGAGCACGCAGG
YLC552CDR	CGTGCTTCAGGCGGAGGATTTGGCG

N. gonorrhoeae inocula were cultured on GC agar, as previously described (6, 24). For N. gonorrhoeae strains, antibiotics were used at the following concentrations: 2 μg ml⁻¹ for erythromycin, 100 μg ml⁻¹ for kanamycin and 1 μg ml⁻¹ for chloramphenicol. Plates were stored at 4°C and used within 7 days. For transformation experiments, GC agar plates were supplemented with 3 mM isoleucine (25). Unless stated otherwise, liquid cultures were grown in gonococcal broth (GCB) following the protocol described previously (6, 24). When a standardized inoculum of more than 10 ml was required for an experiment, multiple 10-ml precultures were combined. Inocula were added to either 20 ml or 50 ml of GCB to generate 30-ml and 60-ml cultures, respectively, which were representative of oxygen-sufficient and oxygen-limited conditions, respectively. When the growth of different strains was to be compared, the amount of inoculum added was adjusted to ensure that the starting optical densities at 650 nm (OD₆₅₀) were almost identical. GCB was supplemented with sodium nitrite between 0.1 mM and 5 mM when appropriate.

Construction of a gonococcal cytochrome c₂ deletion mutation conferring chloramphenicol resistance.

The cat gene (conferring chloramphenicol resistance) was amplified from pLES940 using primers AHCatAgeIF and AHCatAgeIR, which introduced AgeI sites at each end of the fragment. Both the PCR fragment containing the cat gene and pYL12 were digested with AgeI and ligated to form pAH107. Linear DNA containing the cat gene flanked by DNA upstream and downstream of the cccA coding region was produced by PCR using primers YLC552AUF and YLC552CDR. Piliated Neisseria gonorrhoeae F62 and AHGC101 were transformed with these linear fragments to create strains JCGC861 and JCGC330, respectively, which were screened for chloramphenicol resistance. Candidate clones were confirmed by sequencing using the primer YLC552AUF.

Construction of a plasmid for overexpression of cytochrome c₂.

To construct the IPTG (isopropyl-β-d-thiogalactopyranoside)-inducible copy of the cccA gene, the gonococcal cccA coding region, and 14 bp of upstream DNA and 204 bp of downstream DNA were amplified by PCR from N. gonorrhoeae strain F62 genomic DNA using the primers AHCptC2Fb and AHCptC2R, which were flanked by PacI and FseI restriction sites, respectively. The plasmid pGCC4_cccAb was constructed by digesting the pGCC4 plasmid and the aforementioned cccA insert with PacI and FseI and then ligating the insert into the plasmid. The orientation of the insert was verified using primer pGCC4PacIns. In order to transfer the IPTG-inducible copy of cccA into the gonococcus, linear fragments were generated by PCR amplification using primers pGCC4_F and pGCC4_R and the plasmid pGCC4_cccAb as a template. Piliated colonies of Neisseria gonorrhoeae strains F62 and JCGC861 were transformed with these linear fragments to create strains JCGC140 and JCGC142, respectively. Candidates were screened by PCR and sequenced using the primer AHpGCC4PacIns.

Construction of mutants defective in the second domain of cytochrome c₅.

Crossover PCR was used to construct a truncated version of cytochrome c₅. Two fragments of DNA were amplified from N. gonorrhoeae strain F62 chromosomal DNA. Primers AHC5-2ndAUF and AHC5-2ndBURb were used to amplify a region of DNA including 502 bp of noncoding DNA upstream of the cycB gene and 595 bp of coding DNA of the cycB gene. Primer AHC5-2ndBURb incorporated a stop codon, to allow the translation of the truncated version of the cytochrome c₅ protein. Primers AHC5-2ndDDF and AHC5-2ndCDR were used to amplify 612 bp of noncoding DNA downstream of the cycB gene. The fragments contained regions of sequence identity at the 3′ end of the upstream fragment and at the 5′ end of the downstream fragment that included an AgeI restriction site. The fragments were able to anneal to one another during a subsequent PCR using primers AHC5-2ndAUF and AHC5-2ndCDR to form a larger fragment. This contained the upstream noncoding and coding regions immediately followed by the downstream noncoding region with an AgeI restriction site located at the crossover point. To create a larger amount of this specific fragment, the fragment was gel extracted and the DNA amplified by PCR using primers AHC5-2int_F and AHC5-2int_R. The plasmid pAH130 was constructed by ligating this enlarged PCR fragment into the pGEM T-Easy vector. The ermC gene was amplified from JCGC100 using the primers NTERY FOR and NTERY REV, which introduced AgeI sites at each end. The ermC gene fragment and pAH130 were digested with AgeI and then ligated to create pAH131. Linear DNA fragments were amplified from pAH131 using primers AHC5-2int_F and AHC5-2int_R. Piliated Neisseria gonorrhoeae strains F62 and AHGC101 were transformed with these linear fragments to generate strains JCGC310 and JCGC320, respectively. Candidate clones were confirmed by sequencing using the primer AHC5-2int_F.

Construction of 3×FLAG-tagged cytochrome c₂.

To create a FLAG-tagged version of cytochrome c₂, linear fragments of DNA containing 336 bp of DNA located downstream of the cccA coding region were amplified from N. gonorrhoeae strain F62 genomic DNA by PCR using primers AHC2CDRNheI and AHC2DDFXhoI. The resulting fragments and the pDOC-F plasmid were digested with NheI and XhoI enzymes and then ligated to form pAHC2FCD. Primers AHC2AUFEcoRI and AHC2BURKpnI were used to amplify by PCR 435 bp of upstream and 465 bp of coding DNA of the cccA gene from F62 genomic DNA. The resulting fragments and the pAHC2FCD plasmid were digested with EcoRI and KpnI enzymes, and then the products were ligated to form pAHC2F. N. gonorrhoeae F62 was transformed with a linear DNA fragment amplified by PCR from pAHC2F using primers AHC2AUFEcoRI and AHC2CDRNheI. This DNA fragment contained the cccA upstream DNA and coding sequence, followed by an in-frame 3×FLAG tag, the kan gene, and then DNA downstream of the cccA. Purified colonies were screened by PCR for the FLAG-tagged version of the cccA gene and confirmed by sequencing using the primer AHC2DDFXhoI.

Construction of a mutant defective in cytochrome c₄ and the third heme group of CcoP.

PCR primers NTCYC4 FOR and NTCYC4 REV were used to amplify a linear fragment of DNA from strain JCGC800. This fragment contained the cytochrome c₄ gene insertionally inactivated with an erythromycin resistance cassette. Piliated N. gonorrhoeae strain AHGC101 was transformed with this linear DNA fragment to create strain JCGC340. Candidate clones were confirmed by sequencing using the primer NTCYC4 FOR.

Rates of oxygen reduction by gonococcal strains measured with an oxygen electrode.

To measure oxygen reduction rates, a standard protocol was used as described previously (6, 24). Rates of oxygen reduction by suspensions of whole, washed bacteria were measured in the presence of 50 mM lactate as a reductant and were obtained by using an S1/mini-Clark-type oxygen electrode (Hansatech Instruments) in conjunction with an Oxytherm control unit. The rate of oxygen reduction was plotted using the Oxygraph Plus program.

Oxygen-limited growth of bacteria for measurement of nitrite reduction during growth and rates of nitrite reduction by washed bacterial suspensions.

To measure the rate of nitrite reduction during growth, gonococcal strains were grown in 60 ml of GCB at 37°C and shaken at 100 rpm in an orbital shaker. The concentration of sodium nitrite added to cultures differed depending on whether rates of nitrite reduction were to be determined during growth or from washed bacterial suspensions. However, a standard protocol was used, unless otherwise stated. Where the measurement of the rate of nitrite reduction during growth by bacteria was required, 1 mM sodium nitrite was added to cultures after 1 h of growth, followed by a further 4 mM nitrite after 2 h of growth. For strains which are very sensitive to nitrite toxicity or for which the rate of nitrite reduction was very low, 0.1 mM and 1 mM nitrite was added to cultures after 1 h and 2 h of growth, respectively. Control experiments confirmed that rates of nitrite reduction by the parent strain were identical irrespective of whether the cultures had been supplemented with 0.1 + 1, 1 + 1, or 1 + 4 mM nitrite (data not shown). The concentration of nitrite in growth media was determined at regular intervals (26). For the measurement of rates of nitrite reduction by washed bacteria (where only the induction of AniA expression was required), 1 mM sodium nitrite was added after 1 h of growth, and a further 1 mM nitrite was added after 2 h of growth. The presence of nitrite was confirmed by mixing 100 μl of culture on a white tile with 100 μl of 1% sulfanilamide in 1 M HCl and 100 μl of 0.02% N-1 naphthylethylene diamine dihydrochloride in distilled water. The production of a pink color indicated the presence of nitrite. Immediately after all of the nitrite had been reduced (usually at an OD₆₅₀ of 0.6 to 0.7), cultures were harvested by sedimentation at 12,000 × g in an MSE model 18 centrifuge. The bacterial pellet was washed with 50 mM potassium phosphate buffer (pH 7.4), sedimented at 10,000 × g in an Eppendorf bench-top centrifuge, and resuspended in the same phosphate buffer. Rates of nitrite reduction by suspensions of whole, washed bacteria were measured as described previously (24, 26). When rates of nitrite reduction by particular mutants were very low, the sensitivity of the assay was augmented by increasing the volume of cells added to the assay tubes and decreasing the total volume of each assay tube.

SDS-PAGE and heme staining.

Proteins were resolved by Tris SDS-PAGE using a 15% (wt/vol) polyacrylamide gel. Proteins containing covalently attached heme were detected using heme-dependent peroxidase activity (27).

Analysis of AniA and 3×FLAG tag in bacteria by Western analysis.

Proteins separated by SDS-PAGE were transferred electrophoretically onto a polyvinylidene fluoride (PVDF) membrane (Millipore) and then detected as previously described (6, 28). The primary antibody used to detect AniA was anti-AniA (rabbit) antibody (gratefully received from J. Moir and M. Thomson) and was used at a dilution of 1:5,000. The primary antibody used to detect FLAG-tagged proteins was anti-FLAG monoclonal (mouse) antibody (Sigma-Aldrich) and was diluted 1:5,000. Where appropriate, either the anti-rabbit or the anti-mouse peroxidase-labeled secondary antibody (ECL Plus Western blotting detection reagents; Amersham) was used at a dilution of 1:5,000. To visualize proteins, an EZ-ECL kit (Biological Industries, Life Technologies) was used in conjunction with Hyperfilm ECL (Amersham). Films were exposed for 1 to 20 s and developed on a Xograph.

Reproducibility of the data.

Unless otherwise stated, a minimum of two but usually three or more biological replicates of each experiment were completed. In the figures, error bars show standard errors of the means, where appropriate. However, growth of mutants that are especially sensitive to inhibition by nitrite varied from one set of experiments to another, making it inappropriate to superimpose data. For these experiments (see Fig. 1, 2, and 6; also, see Fig. S3 in the supplemental material), data from a typical experiment are presented.

Fig 1.

Oxygen-limited growth and nitrite reduction during growth of strains F62 and JCGC851 (Δc₂) supplemented with nitrite or not. Solid lines indicate the optical density of cultures measured at a wavelength of 650 nm, and dotted lines indicate the concentration of nitrite present in the nitrite-supplemented cultures, which is plotted on the secondary y axis. Arrows indicate the addition of sodium nitrite to relevant cultures. Growth curves are representative of at least three independent experiments. (Inset) Western blot analysis of AniA protein accumulation by strains F62 and JCGC851 (Δc₂) during oxygen-limited growth in the presence of nitrite.

Fig 2.

Oxygen-limited growth of the Δc₂ Δc₄ and Δc₂ Δc₅ double mutants in the presence of nitrite. (A) Oxygen-limited cultures of strains F62, JCGC800 (Δc₄), and JCGC853 (Δc₂ Δc₄) grown in the presence of 1.1 mM nitrite. (B) Oxygen-limited cultures of strains F62, JCGC850 (Δc₅), and JCGC852 (Δc₂ Δc₅) grown in the presence of 2 mM nitrite. Arrows indicate the times at which nitrite was added. (C) Concentration of nitrite remaining in growth medium from cultures in panel A. (D) Concentration of nitrite remaining in growth medium from cultures in panel B.

Fig 6.

Oxygen-limited, nitrite-supplemented growth of CcoP third-heme-group mutants. (Top) Growth of strains F62, JCGC310 (c₅Δ200–279), AHGC101 (ccoP-C368A), and JCGC320 (ccoP-C368A c₅Δ200–279). Arrows indicate the addition of sodium nitrite to cultures. Growth curves shown are representative of at least three independent growth experiments. (Bottom) Concentrations of nitrite remaining in growth medium.

RESULTS

Bioinformatic analysis of the gonococcal cccA gene.

A BLAST search using the NG0292 sequence, excluding hits from the Neisseria genus, showed that cytochrome c₂ is similar to a domain of many copper-containing nitrite reductases. For example, cytochrome c₂ is 46% identical to the C-terminal domains of nitrite reductases of both Kangiella koreensis and Bdellovibrio bacteriovorus. More significantly, residues 30 to 132 of gonococcal cytochromes c₂ are 61% similar to residues 398 to 500 of the nitrite reductase of Moraxella catarrhalis strain 46P47B1, which due to its similar morphology was for many years classified as a member of the Neisseria genus (29). A sequence alignment of the M. catarrhalis and N. gonorrhoeae nitrite reductases shows that they share 63% sequence similarity. However, the M. catarrhalis protein is larger than the gonococcal AniA protein, with a C-terminal extension of 100 residues that include a binding site for covalently bound heme. It is this extra domain that is 61% similar to the gonococcal cytochrome c₂, suggesting that gonococcal cytochrome c₂ might play a role similar to that of the heme-binding domain in the M. catarrhalis AniA. This would be consistent with previous reports that electron donors to the NirK family of copper-containing nitrite reductases are either copper-containing azurins or c-type cytochromes (30). Although there is a lipid-associated azurin, Laz, in Neisseria, it has no role in nitrite reduction (21, 31).

M. catarrhalis AniA would be 72.1% similar and 60.6% identical to a hypothetical fusion protein containing all of the gonococcal AniA and residues 30 to 132 of cytochrome c₂ (see Fig. S1 in the supplemental material). There is also significant similarity between the hypothetical AniA-cytochrome c₂ hybrid and the copper-containing nitrite reductases of a small group of bacteria that include Pseudoalteromonas haloplanktis TAC125 (a member of the gammaproteobacteria), Bdellovibrio bacteriovorus HD100 (a member of the deltaproteobacteria), Psychrobacter sp. strain 1501 (a member of the gammaproteobacteria), and Kingella kingae ATCC 23330 (a member of the betaproteobacteria) (see Fig. S2 in the supplemental material). The crystal structures of AniA proteins from P. haloplanktis (PDB entry 2ZOO) and N. gonorrhoeae (PDB entry 1KBW) have been published (7). The similarity between the two structures is high, except for the additional heme-binding domain found in the P. haloplanktis AniA. Interaction between cytochrome c₂ and AniA might therefore facilitate efficient electron transfer either directly or indirectly from the cytochrome bc₁ complex in the inner membrane across the periplasm to AniA.

Cytochrome c₂ as a potential electron donor to AniA.

We previously reported that a mutant, strain JCGC851, with a deletion of the cccA coding sequence grew as rapidly as the parent strain under oxygen-sufficient conditions and reduced oxygen at a rate slightly higher than the parent strain, F62 (6). As cytochrome c₂ was shown not to be an electron donor to the terminal oxidase, the possibility that it is part of an electron transfer pathway to AniA was investigated. Both the parental strain and the deletion mutant were grown under oxygen-limited conditions in the presence and absence of nitrite. The growth of both strains was stimulated by the addition of nitrite, but the cytochrome c₂ deletion mutant grew slightly more slowly and reduced nitrite in the culture more slowly than the parent strain (Fig. 1). The rate of lactate-dependent nitrite reduction by washed suspensions of the cytochrome c₂ deletion mutant was only 52% that of the parent strain (Table 3). Western blotting revealed similar quantities of AniA in strains F62 and JCGC851 (Fig. 1, inset). This result was highly reproducible in many independent experiments, suggesting that the decreased rate of nitrite reduction by the mutant is due to a decreased rate of electron transfer to AniA, not to a lower abundance of AniA in cells lacking cytochrome c₂. Although these data are consistent with a role for cytochrome c₂ in electron transfer to the nitrite reductase, AniA, the rate of nitrite reduction by the mutant was still about half that of the parent strain, indicating the existence of multiple electron transfer pathways to AniA.

Table 3.

Rates of nitrite reduction by cytochrome mutants of N. gonorrhoeae^a

Strain	Genotype	Rate of nitrite reduction	% of parental rate	P value(s)
F62	Parental	222 ± 8	100
AHGC101	ccoP-C368A	115 ± 8	52	1.58 × 10−9 (F62)
JCGC850	Δc5	130 ± 7	59	1.37 × 10−9 (F62)
JCGC310	c5Δ200–279	87 ± 9	39	1.26 × 10−8 (F62), 0.003 (Δc5)
JCGC320	c5Δ200–279 ccoP-C368A	26 ± 1.3	12	0.001 (c5Δ200–279)
JCGC851	Δc2	115 ± 7	52	1.37 × 10−12 (F62)
JCGC852	Δc2 Δc5	86 ± 9	39	9.14 × 10−4 (Δc5)
JCGC330	Δc2 ccoP-C368A	115 ± 15	52	0.97 (ccoP-C368A)
JCGC800	Δc4	81 ± 6	36	1.66 × 10−13 (F62)
JCGC853	Δc4 Δc2	93 ± 5	42	0.17 (Δc4)
JCGC340	Δc4 ccoP-C368A	79 ± 7.7	36	0.81 (Δc4), 0.004 (ccoP-C368A)
JCGC142	ΔaspC-lctP::plac-cccA+ Δc2	119 ± 3	54	0.71 (Δc2)
JCGC142	ΔaspC-lctP::plac-cccA+ Δc2 + IPTG	128 ± 5	58	0.56 (ΔaspC-lctP::plac-cccA+ Δc2)

^aUnits for rates of nitrite reduction are nmol NO₂⁻ reduced min⁻¹ mg dry cell mass⁻¹; values are means ± standard error of the means. P values are relative to the strain shown in parentheses. The data are based upon results of between 5 and 29 independent experiments. Note that Δc₅ indicates that the complete coding sequence for cytochrome c₅ has been deleted, but c₅Δ200–279 lacks only the last 80 codons, which include the second heme-binding domain of cytochrome c₅.

Evidence that cytochrome c₂ is a component of the electron transfer pathway from the CcoP subunit of cytochrome oxidase to the outer membrane nitrite reductase.

The CcoP subunit of cytochrome oxidase in all of the Neisseria species except N. meningitidis is so far unique because it includes a third, C-terminal heme group that is absent from CcoP in other bacteria (23, 24). The effects of the cytochrome c₂ mutation on oxygen-limited growth and rates of nitrite reduction both during growth and incubation with lactate and nitrite were almost identical to those reported previously for mutants that lack the third heme group of CcoP (23). We therefore generated a ΔcccA ccoP-C368A double mutant (JCGC330) to determine whether cytochrome c₂ and CcoP are part of the same or different electron transfer pathways to nitrite. After oxygen-limited growth in the presence of nitrite, all three strains reduced nitrite at 52% of the rate of the parental strain (Table 3). As Western analysis showed that all of the mutants had accumulated similar quantities of the AniA protein (data not shown), the identical growth and nitrite reduction phenotypes of these three strains strongly implicate cytochrome c₂ as a component of a pathway in which the CcoP subunit of cytochrome oxidase mediates electron transfer between the cytochrome bc₁ complex and AniA in the outer membrane.

Electron transfer from cytochromes c₄ and c₅ to nitrite.

Unlike the gonococcal cytochrome oxidase, the CcoP subunit of the meningococcal cytochrome oxidase lacks the third heme-binding domain and therefore is not involved in electron transfer to nitrite (23, 24). Conversely, loss of the third heme group of gonococcal CcoP has no effect on oxygen-sufficient growth or rates of oxygen reduction (6, 24). There are two parallel pathways for electron transfer to cytochrome oxidase in both organisms, and in meningococci, cytochrome c₅ was reported to be essential for nitrite reduction during growth in the presence of a high concentration of nitrite (23). However, rates of nitrite reduction by meningococci were not determined. Furthermore, the cytochrome c₅ mutant was exposed to 5 mM nitrite, a concentration that is 100 to 1,000 times higher than that typically found in body fluids (32–34), and no attempt to adapt the cytochrome c₅ mutant to growth in gradually increasing concentrations of nitrite was reported. It is therefore not known whether the concentration of nitrite added to the cultures was so high that its toxicity would inhibit growth and nitrite reduction. Experiments were therefore designed to determine the effects of deletion of the gonococcal cycB gene encoding cytochrome c₅ on growth, nitrite toxicity, and rates of nitrite reduction. For comparison, parallel experiments were completed with the parent strain and a cycA mutant that lacks cytochrome c₄.

Consistent with the data for N. meningitidis (23), addition of nitrite to a final concentration of 5 mM either strongly inhibited or completely prevented growth of the cytochrome c₅ mutant. In contrast and as previously reported (6), growth of the parent strain was stimulated by nitrite (Fig. 1). Reproducible growth of the cytochrome c₅ mutant was obtained if the oxygen-limited cultures were first supplemented with only 0.1 mM nitrite and, once this nitrite had been reduced, with a further 1 mM or 2 mM nitrite. Under these conditions, growth of the mutant was still far less rapid than that of the parent, and although all of the nitrite had been reduced by the parent strain within 2 h of addition, none of the nitrite had been reduced by the cytochrome c₅ mutant until 5 h after addition (Fig. 2). Using this protocol, sufficient bacteria were obtained to determine that the rate of nitrite reduction by the cytochrome c₅ mutant was 59% that of the parent. Parallel experiments with the cycA mutant that lacks the alternative cytochrome c₄-dependent electron transfer pathway to cytochrome oxidase gave slightly different results. Although oxygen-limited growth in the presence of 5 mM nitrite was strongly inhibited and inhibited similarly to the cytochrome c₅ mutant in the presence of 1 mM nitrite, nitrite was reduced during growth slightly more rapidly by the cytochrome c₄ mutant than by the cytochrome c₅ mutant (Fig. 2). Surprisingly, the rate of nitrite reduction by bacterial suspensions was only 36% of the rate of the parental strain (Table 3).

Contrasting effects of loss of cytochrome c₂ on nitrite reduction by mutants defective in cytochrome c₄ or cytochrome c₅.

Previously constructed double mutants defective in cytochrome c₂ and either cytochrome c₄ or cytochrome c₅ were used to determine whether cytochromes c₄ and c₅ were components of the same electron transfer pathways to nitrite as cytochrome c₂. Deletion of cccA (encoding cytochrome c₂) restored oxygen-limited growth of the cytochrome c₄ mutant almost to that of the parental strain but had no detectable effect on growth of the cytochrome c₅ mutant. This was reflected in the rate of accumulation of the terminal nitrite reductase, AniA, which was markedly delayed in the cytochrome c₂ and c₅ double mutant compared with the parent or the cytochrome c₂ and c₄ double mutant (Fig. 3). The effects on nitrite reduction during growth were therefore in large part due to differences in the previously reported effects on oxygen reduction (6), which in turn affected the stage of growth at which the dissolved oxygen concentration was sufficiently low for FNR to be active and hence aniA expression to be induced. Nevertheless, the double mutants also differed in their ability to reduce nitrite. The rate of nitrite reduction by the cytochrome c₂ and c₄ double mutant was almost identical to those of the cytochrome c₂ or cytochrome c₄ single mutants. In contrast, the rate of nitrite reduction by the cytochrome c₂ and c₅ double mutant was only 39% that of the parent, significantly lower than that of either the cytochrome c₂ or cytochrome c₅ single mutant (Table 3). These data indicate that although in the absence of cytochrome c₄, cytochrome c₂ negatively affects electron transfer via cytochrome c₅ to oxygen (Table 4), it has no effect on cytochrome c₅-mediated nitrite reduction. Conversely, cytochrome c₂ stimulates, but is not completely essential for, cytochrome c₄-mediated nitrite reduction in a pathway that is independent of cytochrome c₅.

Fig 3.

Western analysis of AniA accumulation during oxygen-limited growth in the presence of nitrite of the Δc₂ Δc₄ and Δc₂ Δc₅ double mutants compared with the parent. N. gonorrhoeae JCGC853 (Δc₂ Δc₄) (A) and JCGC852 (Δc₂ Δc₅) (B) strains were grown under oxygen-limited conditions compared with the parent strain in the presence of 1.1 mM or 2 mM nitrite, respectively. Samples were taken during growth, separated by SDS-PAGE, and blotted onto PVDF membranes, and AniA protein was detected using anti-AniA antibodies and ECL-Plus chemiluminescent labeling.

Table 4.

Rates of oxygen and nitrite reduction by cytochrome mutants of N. gonorrhoeae

Strain or genotype	% of parental reduction ratea
	Oxygen	Nitrite
F62	100	100
Δc2	126	52
Δc5	80	59
Δc2 Δc5	57	39
Δc4	84	36
Δc2 Δc4	107	42

^aOxygen reduction rates are from reference 6. Nitrite reduction rates are from this work.

Detection of cytochrome c₂ synthesis.

Unlike the seven other gonococcal c-type cytochromes, cytochrome c₂ has not been visualized on an SDS-PAGE gel stained for the presence of covalently bound heme (6). This phenomenon was also observed for the Neisseria meningitidis homologue, cytochrome c_x (2). As the effects of mutations in the cycA and cycB genes (encoding cytochromes c₄ and c₅, respectively) were fully complemented by ectopically expressed copies of these genes integrated at the aspC-lctP locus on the chromosome, the same neisserial complementation system was used to construct strains with a functional cccA gene at this locus. The F62 derivative, strain JCGC140, and the parent strain were grown under oxygen-limited conditions supplemented with 5 mM sodium nitrite. In order to overexpress the ectopic copy of the cccA gene (encoding cytochrome c₂), 1 mM IPTG was added to relevant cultures immediately postinoculation. This revealed two additional heme-stained proteins in the IPTG-induced culture of strain JCGC140 compared with the parental F62 strain (Fig. 4). The upper band corresponds to cytochrome c₂ that has retained its leader peptide; the lower band is the fully processed cytochrome. This established that the absence of the cytochrome c₂ band on SDS-PAGE gels stained for covalently bound heme is not due to inability to translate the cccA mRNA into protein or to failure to incorporate the covalently bound heme group into the cytochrome c₂ apoprotein. Rather, it indicates that very little of the cytochrome accumulates in normal strains.

Fig 4.

SDS-PAGE gel stained for covalently bound heme in cultures of strains F62, JCGC851 (Δc₂), and JCGC140 (ΔaspC-lctP::plac-cccA⁺). Cultures were grown under oxygen-limited, nitrite-supplemented conditions. After 7 h of growth, samples of bacteria were harvested and lysed, and the proteins were separated by SDS-PAGE and stained for covalently bound heme. Lane 1, protein MW marker; lane 2, parent strain F62; lane 3, strain JCGC851 (Δc₂); lanes 4 and 5, strain JCGC140 (ΔaspC-lctP::plac-cccA⁺) unsupplemented, or supplemented with 1 mM IPTG, respectively. Note that the cytochrome c₂ band, although faint, is a doublet.

Complementation of the cccA deletion mutation.

In order to complement the cytochrome c₂ deletion, strain JCGC142, which lacks the cccA gene at its normal locus but has a functional copy of cccA regulated by the lac promoter inserted at the aspC-lctP locus, was constructed. Oxygen-limited cultures of strains F62, JCGC851 (lacking the cytochrome c₂-coding region), and JCGC142 (cytochrome c₂ deletion mutant complemented with the ectopically expressed copy of cytochrome c₂) were grown under oxygen-limited conditions in the presence or absence of 5 mM nitrite. Strain JCGC142 was also grown in the presence and absence of 1 mM IPTG. The growth rate of strain JCGC142 was improved on addition of IPTG, suggesting that the IPTG-inducible copy of cytochrome c₂ is partially able to complement the loss of cytochrome c₂ (see Fig. S3 in the supplemental material). However, the growth phenotype was not fully restored to the same level as the parent strain, even though the overexpression of cytochrome c₂ during these experiments was confirmed (data not shown). Furthermore, nitrite reduction during growth by strain JCGC142 was also only partially restored to that of the parental strain, and the addition of IPTG to the complemented strain failed to restore the rate of nitrite reduction to that of the parent (Table 3). Western analysis confirmed that the relative abundance of AniA was the same in all samples, suggesting that the reason why complementation was only partial was not a lower accumulation of AniA (see Fig. S3, inset).

Detection and regulation of cytochrome c₂ synthesis.

In order to study the regulation of cytochrome c₂ synthesis in the gonococcus, a 3×FLAG tag was cloned onto the 3′ end of the cccA gene to create strain JCGC807. This would allow detection of both cytochrome c₂ apoprotein and holoprotein by Western analysis. To confirm that cytochrome c₂ was expressed at a very low level compared to other gonococcal c-type cytochromes, the accumulation of 3×FLAG-tagged cytochrome c₂ was compared to the level of accumulation of 3×FLAG-tagged cytochrome c₄. Strains JCGC807, containing the 3×FLAG-tagged cytochrome c₂ protein, and JCGC805, containing the 3×FLAG-tagged cytochrome c₄ protein, were grown under oxygen-limited conditions supplemented with 5 mM nitrite. It was estimated by Western blotting that cytochrome c₂ had accumulated to a level only 10% of that of cytochrome c₄ protein (Fig. 5A). The absence of a band corresponding to cytochrome c₂ on an SDS-PAGE gel stained for covalently bound heme can therefore be explained by the limited sensitivity of the heme-staining procedure and not by the inability of heme to be incorporated into the cytochrome c₂ apoprotein. This reinforces the hypothesis that cytochrome c₂ is expressed at a very low level compared to the other electron transfer chain components. Parallel experiments with oxygen-limited cultures not supplemented with nitrite revealed that cytochrome synthesis is induced about 2-fold by nitrite (Fig. 5B).

Fig 5.

Accumulation of FLAG-tagged cytochrome c₂ compared with cytochrome c₄: regulation of cytochrome c₂ accumulation. (A) Accumulation of 3×FLAG-tagged proteins detected by Western blotting. Cultures were grown under oxygen-limited conditions supplemented with 1 mM sodium nitrite after 1 h of growth followed by 4 mM sodium nitrite after 2 h of growth. Samples were taken after 5 h of growth. In each lane, 20 μg of dry cell mass was loaded, but for strain JCGC805, different volumes were loaded (as shown below the gel) in order to titrate in the relative abundance of the 3×FLAG-tagged cytochrome c₄ control strain. (B) Accumulation of 3×FLAG-tagged cytochrome c₂ protein in strain JCGC807 (containing a 3×FLAG-tagged copy of cytochrome c₂ under the control of its natural promoter) detected by Western blotting. Cultures were grown under oxygen-limited conditions supplemented with 1 mM sodium nitrite after 1 h of growth followed by 4 mM sodium nitrite after 2 h of growth. Samples were taken after 1 h, 2 h, 4 h, and 6 h of growth from cultures in the absence and presence of nitrite (− and +, respectively). In each lane, 20 μg of dry cell mass was loaded.

Nitrite reduction by a double mutant defective in both CcoP and cytochrome c₅.

As single mutants defective in the third heme group of CcoP, cytochrome c₂, or cytochrome c₅ all reduce nitrite at about half the rate of the parent strain and cytochrome c₅ was thought to be involved in a pathway to nitrite that is independent from CcoP and cytochrome c₂, it seemed likely that double mutants defective in cytochrome c₅ and either cytochrome c₂ or the third heme group of CcoP would be totally defective for nitrite reduction. Many attempts to construct a cytochrome c₅ ccoP-C368A double mutant failed, so instead the 3′ end of the cycB gene (encoding cytochrome c₅) was deleted to construct a mutant in which only the amino-terminal domain of cytochrome c₅ was synthesized. This mutant (JCGC310) was able to grow aerobically and reduce oxygen almost as rapidly as the parent (215 ± 7 nmol O₂ reduced min⁻¹ mg dry cell mass⁻¹ [mean ± standard error] for the parent strain versus 210 ± 3 nmol O₂ reduced min⁻¹ mg dry cell mass⁻¹ for strain JCGC310) but was as defective for oxygen-limited growth and nitrite reduction as the mutant in which all of the cycB gene had been deleted (Table 3). When this truncated cytochrome c₅ mutation was combined with the ccoP-C368A mutation, the resulting strain (JCGC320) grew much less rapidly than either of the single constituent mutants (Fig. 6). The rate of nitrite reduction by this double mutant was extremely low, just 12% of that of the parental strain (Table 3).

Nitrite reduction by a double mutant defective in both CcoP and cytochrome c₄.

The indication that loss of cytochrome c₂ function had no effect on cytochrome c₅-dependent nitrite reduction suggested that cytochrome c₂ primarily affects electron transfer from cytochrome c₄ via the third heme group of CcoP to AniA. If so, rates of nitrite reduction by a double mutant defective in both cytochrome c₄ and the third heme-binding domain of CcoP should be similar to those of the cytochrome c₄ single mutant. This prediction was confirmed. The rates of nitrite reduction by all three strains, the double mutant and the two single mutants, were in the range of 36 to 52% of that of the parent (Table 3). This result was highly significant because it suggests that cytochrome c₄ is the major electron donor to the third heme group of CcoP, but cytochrome c₅ is ineffective.

DISCUSSION

Cytochrome c₂ as a potential electron donor to the gonococcal nitrite reductase.

Analysis of neisserial genomes reveals the ubiquitous presence of a gene potentially encoding a 16.5-kDa c-type cytochrome. The gonococcal protein, which we have designated cytochrome c₂, is highly similar to the C-terminal domain of some other copper-containing nitrite reductases of the NirK family (see Fig. S1 and S2 in the supplemental material). In these soluble, periplasmic proteins, the heme domain transfers electrons from the membrane-associated respiratory chain to the active site copper of NirK (35, 36). This suggested that the gonococcal cytochrome c₂ might be the immediate electron donor to the copper-containing nitrite reductase AniA. Deletion of the cccA gene resulted in loss of about half the total rate of nitrite reduction, confirming a role for cytochrome c₂ in electron transfer to nitrite (Table 1). As the decreased rate of nitrite reduction was shown by Western blotting not to be due to the failure to accumulate sufficient AniA apoprotein, we conclude that it was due to an electron transfer defect. Furthermore, many SDS-PAGE gels of proteins from mutant strains stained for covalently bound heme revealed that synthesis of other electron transfer components was not disrupted in the mutant strains used in the present work. It is therefore unlikely that the mutant phenotypes are due to indirect effects on the synthesis of other electron transfer components.

Evidence that cytochromes c₂ and c₄ and the third heme of CcoP are components of the same electron transfer pathway to nitrite.

Data presented in Table 1 and in Fig. 1 established that cytochrome c₂ is not an essential electron donor to AniA. The cytochrome c₂ and c₅ double mutant reduced nitrite at a lower rate than either of the single mutants, consistent with predictions that there is an alternative, cytochrome c₅-dependent pathway to nitrite that is independent of cytochrome c₂ (23). In contrast, a second mutation in cytochrome c₂ in addition to deletion of cytochrome c₄ had no effect on cytochrome c₅-dependent nitrite reduction. This strongly indicates that cytochromes c₂ and c₄ are components of the same electron transfer pathway to nitrite. Furthermore, after oxygen-limited growth, a cytochrome c₂ and ccoP-C368A double mutant reduced nitrite at the same rate as single mutants defective in either cytochrome c₂ or the third heme group of CcoP: after oxygen-limited growth in the presence of nitrite, all three strains reduced nitrite at 52% of the rate of the parental strain (Table 3). We therefore conclude that cytochromes c₂ and c₄ and the third heme group of CcoP are components of the same electron transfer pathway to nitrite. As Western analysis showed that single and double mutants defective in these three components all had accumulated similar quantities of the AniA protein (data not shown), the identical growth and nitrite reduction phenotypes of these three strains strongly implicate cytochrome c₂ as a component of the electron transfer pathway in which the CcoP subunit of cytochrome oxidase mediates electron transfer between the cytochrome bc₁ complex and AniA in the outer membrane. The similarity between cytochrome c₂ and the C-terminal domain of other nitrite reductases of the NirK family strongly indicates that it is likely to be the direct electron donor to the active site copper of AniA. Conversely, the fact that deletion of the third heme-binding domain of CcoP does not disrupt either aerobic growth or the rate of oxygen reduction indicates that cytochrome c₄ precedes CcoP in the electron transfer chain to oxygen and presumably therefore also in electron transfer to nitrite. If these inferences are correct, electrons would be transferred from the cytochrome bc₁ complex via cytochrome c₄ to the first heme groups of CcoP and then via the third heme group of CcoP to cytochrome c₂ and AniA (Fig. 7A). This model assumes that cytochrome c₂ is a direct electron donor, but not the only electron donor, to AniA. We propose that the alternative pathway involves electron transfer across the periplasm directly to AniA by the second heme group of cytochrome c₅ (Fig. 7A).

Fig 7.

Proposed role of cytochrome c₂ in electron transfer to nitrite. (A) Proposed pathway for electron transfer from the cytochrome bc₁ complex via the CcoP subunit of cytochrome oxidase cbb₃ to the nitrite reductase AniA; (B) proposed alternative complexes of cytochrome c₂ and CcoQ with cytochrome oxidase cbb₃ that preferentially transfer electrons to nitrite and oxygen, respectively.

Are electron transfer complexes involved in neisserial electron transfer to nitrite?

There is direct biochemical evidence that the four cytochrome oxidase cbb₃ subunits in other bacteria form a multiprotein complex with cytochromes c₄ and c₅ and possibly also with the cytochrome bc₁ complex (4, 37–39). Blue native PAGE allows intact supercomplexes of proteins to be separated by gel electrophoresis, after which the supercomplexes can be resolved on an SDS-PAGE gel and the individual protein components within them can be identified (40). Attempts to isolate similar complexes from the gonococcus and to identify the components by blue native PAGE have been unsuccessful, but until such complexes have been isolated and characterized physically, the proposals that follow will remain unproven. However, five lines of evidence suggest that similar complexes might be formed in the neisserial electron transfer chains, and the limited biochemical data currently available suggest a more complex role for cytochrome c₂ in gonococcal electron transfer to nitrite.

First, the cytochrome c₂ single mutant reduces oxygen more rapidly than the parent strain. Second, the cytochrome c₂ c₄ double mutant also reduces oxygen more rapidly than a single mutant that is deficient only in cytochrome c₄. These two observations imply that cytochrome c₂ in some way restricts electron transfer via cytochrome c₅ to oxygen, possibly by competing for complex formation with other components of the cytochrome oxidase complexes. Third, until the present work, cytochrome c₂ had never been seen as a hemoprotein on SDS-PAGE gels of gonococcal proteins stained for covalently bound heme. Data presented in this paper establish that this is due to very low levels of its accumulation during growth, even following modest induction during oxygen-limited growth in the presence of nitrite. Fourth, the implication that cytochrome c₂ is needed in only very small quantities must be reconciled with the loss of about half of the nitrite reduction ability of the cytochrome c₂ mutant, which indicates that cytochrome c₂ plays an important role in the electron transfer chain to nitrite. Finally, an ectopically expressed copy of cccA only partially complements the above phenotypes, even when it is overexpressed.

Construction of gonococcal strains with multiple mutations is far from a trivial process, and multiple attempts to construct strains with several combinations of mutations were unsuccessful. In some cases, this is easily explained. For example, single mutants defective in cytochrome c₄ or c₅ were easily obtained, but it is impossible to make a cycA cycB double mutant because cytochromes c₄ and c₅ provide alternative pathways of electron transfer to oxygen, either of which can fulfill an essential function (6). Less obvious is why we were unable to make a double mutant with deletions in both cytochrome c₅ and the domain encoding the third heme group of CcoP. The most likely reason is that cytochromes c₄ and c₅ form a single complex or two parallel complexes with cytochrome oxidase cbb₃ and that disruption of these complexes results in loss of electron transfer efficiency. This problem was solved by deleting the coding region for the second heme-binding domain of cytochrome c₅, which we suggest resulted in a complex that could still transfer electrons to oxygen but was partially defective in nitrite reduction. However, we were also unable to construct a double mutant combining the truncated cycB gene (encoding truncated cytochrome c₅) with a cytochrome c₄ deletion, despite the fact that the truncated cytochrome c₅ was still competent for electron transfer to oxygen. As all of the c-type cytochromes apart from cytochrome c₂ are firmly membrane attached, these results would be entirely consistent with the inability to assemble functional cytochrome oxidase complexes in the complete absence of cytochrome c₅ or CcoP but the ability to assemble active complexes when only their C-terminal domains are absent.

The cccA deletion construct used in the present study truncated the 3′ end of the convergent downstream gene NG0291 that is predicted to encode an Na⁺/H⁺ antiporter. However, multiple lines of evidence indicate that only partial complementation by an ectopic copy of cccA was not due to failure to compensate for loss of expression of this downstream gene, the most persuasive being that the cytochrome c₂ mutation had no effect on aerobic growth and even slightly enhanced the rate of oxygen reduction compared with the parent strain. Furthermore, deletion of cytochrome c₂ and also loss of the final codons of the convergent downstream antiporter gene fully restored the rate of respiration, resistance to oxygen toxicity, and the growth rate of a cytochrome c₄ single mutant. As many of the phenotypes of the cytochrome c₂ mutation were reversed by deletion of cytochrome c₄, explanations other than possible loss of function of the downstream antiporter gene must be suggested for the incomplete complementation of the cccA mutation by an ectopically expressed copy of the cccA gene.

Alternative possible mechanisms for partitioning electron transfer to oxygen or nitrite.

The fourth subunit of cytochrome oxidase cbb₃, CcoQ, has been suggested to be an assembly factor required for the formation of a fully active cytochrome oxidase complex (38, 39). However, unlike most other bacteria, the gonococcal cytochrome oxidase also transfers electrons to nitrite, and the nitrite and oxygen reduction functions are essentially independent (compare data in reference 24 with those in Fig. 1). We therefore propose that cytochrome c₂ forms a complex with cytochrome oxidase cbb₃. It is possible that the reduced and oxidized cytochromes regulate the distribution of electrons to nitrite and oxygen, respectively. This might explain why a second mutation in cytochrome c₂ in addition to deletion of cytochrome c₅ adversely affects cytochrome c₄-dependent respiration, as detected by decreased rates of oxygen reduction and a parallel increase in oxygen sensitivity during growth (6). A more likely explanation is that cytochrome c₂ is an alternative assembly factor to CcoQ that facilitates the formation of a complex that transfers electrons from cytochrome c₄ to AniA via the third heme group of CcoP (Fig. 7B). If so, we suggest that a low concentration of cytochrome c₂ compared with cytochromes c₄ or c₅ is necessary to avoid compromising the essential cytochrome oxidase function of cytochrome cbb₃. Overexpression of cccA from the ectopic, IPTG-inducible copy of cccA might adversely change the balance between electron transfer to nitrite or oxygen, compromising the ability of the gonococcus to protect itself against oxygen toxicity (6). This remains speculation until the proposed complexes have been purified and characterized in vitro.

Contrasting electron transfer pathways to nitrite in gonococci and meningococci as an example of niche adaptation.

The complexity of electron transfer pathways to nitrite in the gonococcus are in marked contrast to the single cytochrome c₅-dependent pathway in meningococci (23). However, while most gonococcal isolates have retained the ability to reduce nitrite during oxygen-limited growth, some meningococci have lost this capacity, presumably because it does not confer a sufficiently strong selective advantage (41). The simplest explanation for these observations is that in contrast to meningococci that inhabit sites at which oxygen is abundant, gonococci are found where oxygen is scarce. In such environments, it is clearly advantageous to retain the ability to supplement oxygen reduction with reduction of alternative electron acceptors such as nitrite (9) and possibly also hydrogen peroxide and nitric oxide generated both by the host and by neighboring bacteria. Furthermore, meningococcal nitrite reduction is regulated by an FNR protein that appears to be less sensitive to inactivation by oxygen than the gonococcal FNR (42), and meningococci lack the cytochrome c peroxidase, Ccp, for protection against hydrogen peroxide generated by neighboring lactobacilli. The systematic deletion of relevant genes has enabled the close links between prevention of toxicity and exploitation of such molecules as oxygen, hydrogen peroxide, nitrite, and nitric oxide to be revealed. We conclude that processes that enable a pathogen to survive host defenses by detoxifying their environment are linked to redox processes that support growth, with no clear distinctions between them (43).