Novel Transporter Required for Biogenesis of cbb₃-Type Cytochrome c Oxidase in Rhodobacter capsulatus

ABSTRACT

The acquisition, delivery, and incorporation of metals into their respective metalloproteins are important cellular processes. These processes are tightly controlled in order to prevent exposure of cells to free-metal concentrations that could yield oxidative damage. Copper (Cu) is one such metal that is required as a cofactor in a variety of proteins. However, when present in excessive amounts, Cu is toxic due to its oxidative capability. Cytochrome c oxidases (Coxs) are among the metalloproteins whose assembly and activity require the presence of Cu in their catalytic subunits. In this study, we focused on the acquisition of Cu for incorporation into the heme-Cu binuclear center of the cbb₃-type Cox (cbb₃-Cox) in the facultative phototroph Rhodobacter capsulatus. Genetic screens identified a cbb₃-Cox defective mutant that requires Cu²⁺ supplementation to produce an active cbb₃-Cox. Complementation of this mutant using wild-type genomic libraries unveiled a novel gene (ccoA) required for cbb₃-Cox biogenesis. In the absence of CcoA, the cellular Cu content decreases and cbb₃-Cox assembly and activity become defective. CcoA shows homology to major facilitator superfamily (MFS)-type transporter proteins. Members of this family are known to transport small solutes or drugs, but so far, no MFS protein has been implicated in cbb₃-Cox biogenesis. These findings provide novel insights into the maturation and assembly of membrane-integral metalloproteins and on a hitherto-unknown function(s) of MFS-type transporters in bacterial Cu acquisition.

IMPORTANCE

Biogenesis of energy-transducing membrane-integral enzymes, like the heme copper-containing cytochrome c oxidases, and the acquisition of transition metals, like copper, as their catalytic cofactors are vital processes for all cells. These widespread and well-controlled processes are poorly understood in all organisms, including bacteria. Defects in these processes lead to severe mitochondrial diseases in humans and poor crop yields in plants. In this study, using the facultative phototroph Rhodobacter capsulatus as a model organism, we report on the discovery of a novel major facilitator superfamily (MFS)-type transporter (CcoA) that affects cellular copper content and cbb₃-type cytochrome c oxidase production in bacteria.

Introduction

Cells require copper (Cu) as a cofactor for many metalloproteins, including cytochrome c oxidases (Coxs) (1), superoxide dismutases (2), and multicopper oxidases (3). Among them, Cox enzymes terminate the electron transfer chains of aerobic organisms by catalyzing the four-electron reduction of dioxygen (O₂) to water during respiration. Cu is an essential catalytic and structural cofactor of this important energy transduction enzyme (1), but an excess amount of it is toxic, as it can activate O₂ via the Fenton reaction to generate dangerous reactive oxygen radicals (4, 5). Indeed, its shortage or surplus leads to severe human illnesses, like Menkes, Wilson’s, and Alzheimer’s diseases (6–9). Thus, cells need to precisely control Cu acquisition, trafficking, and incorporation into the target proteins.

Cox enzymes have a conserved catalytic subunit (i.e., subunit I) that contains a low-spin heme and a binuclear metal center, composed of a high-spin heme Fe and a Cu atom (i.e., Cu_B) (10). They also have additional subunits: the mammalian mitochondrial enzyme contains 13, whereas the bacterial Cox has 3 or 4 subunits in total (11). Subunit II is the primary electron acceptor and harbors extra cofactors, like another binuclear Cu center (Cu_A) in the mitochondrial aa₃-type Cox (10, 11). In the bacterial cbb₃-type Cox (cbb₃-Cox), which is important for aerobic respiration, the onset of photosynthesis, rhizobial symbiosis, and pathogenesis in various species (12–14), subunit II contains covalently bound hemes instead of a Cu_A center. The presence of various cofactors renders Cox biogenesis a complex process. Synthesis and maturation of individual subunits, their insertions into the membrane, and assembly of mature subunits into active enzymes are coordinated temporally and spatially (15). Regulated expression of the subunits and accurate insertion of the cofactors during biogenesis are required to control the concentration of potentially toxic, free redox-active cofactors or reactive assembly intermediates (5).

The purple nonsulfur facultative phototrophic bacterium Rhodobacter capsulatus contains only two terminal O₂ reductases, a cbb₃-Cox and a hydroquinone oxidase (Qox), that allow cells to grow under respiratory conditions even in the absence of an active Cox (16). Unlike in many other bacterial species, cbb₃-Cox is the only Cox in this species, making it an organism of choice for investigating cbb₃-Cox biogenesis. The R. capsulatus cbb₃-Cox has four subunits encoded by the ccoNOQP operon and, like all Cox enzymes, contains a heme Cu_B catalytic center located in subunit I (CcoN). It naturally lacks a Cu_A center; hence, the only Cu atom is that of the Cu_B center. Instead of a Cu_A center, cbb₃-Cox contains two membrane-bound c-type cytochrome subunits: the monoheme cytochrome c (CcoO or cytochrome c_o) and the diheme cytochrome c (CcoP or cytochrome c_p) as subunits II and III, respectively (17). Both CcoO and CcoP are required for electron transfer from the electron donor cytochromes c (i.e., cytochromes c₂ and c_y) to the heme Cu_B center at CcoN. The enzyme also has a fourth subunit (CcoQ) without any cofactor, but its absence hardly affects the enzymatic activity (18, 19).

Currently, a number of components involved in mitochondrial Cox biogenesis are known, and several of them are linked to human diseases (20–22). A smaller number of biogenesis components have been identified so far in the case of bacterial cbb₃-Cox. Earlier genetic screens for loss of Cox activity identified several genes affecting cbb₃-Cox biogenesis, including ccoGHIS (23), olsAB (24), and senC (25, 26). Recent proteomic approaches also associated dsbA and degP (27) with this process. However, many of the components and steps governing biogenesis of an active cbb₃-Cox still remain undefined. In particular, those involved in forming the universally conserved heme Cu_B binuclear center of Cox enzymes are unknown. In this work, we report on the molecular characterization of a cbb₃-Cox-defective mutant that requires Cu²⁺ supplementation to produce an active cbb₃-Cox. This mutant unveiled a novel transporter, CcoA, of the major facilitator superfamily (MFS), whose absence greatly diminishes the levels of intracellular Cu content and active cbb₃-Cox in R. capsulatus. To our knowledge, CcoA is the first example of an MFS-type transporter required for efficient Cu acquisition and cbb₃-Cox production in bacteria.

RESULTS

Cu²⁺ supplement-dependent cbb₃-Cox mutants.

Previously, R. capsulatus mutants (e.g., GK1 [see Table S1 in the supplemental material]) that exhibited very low or no cbb₃-Cox activity in a growth medium-dependent manner were isolated following mutagenesis with ethyl methanesulfonate (28). These mutants had a NADI (α- naphthol plus dimethylphenylene diamine → indophenol blue plus H₂O)-negative (i.e., no cbb₃-Cox activity) phenotype when grown on enriched peptone yeast extract (MPYE) medium (23) but exhibited a NADI^slow (i.e., very low cbb₃-Cox activity) phenotype on minimal (MedA) medium (Fig. 1). Subsequently, additional similar mutants (e.g., HY70 [see Table S1 in the supplemental material]) were also obtained. Two of these mutants (GK1 and HY70) were retained for further studies.

FIG 1 .

Growth medium-dependent NADI phenotypes of R. capsulatus mutants. The wild-type (MT1131), the ΔccoNOQP (GK32), the ΔccoA or ccoA1 (SE8 or GK1, respectively), and the complemented ΔccoA/ccoA+ (SE8/pSE3) strains were grown at 35°C under respiratory conditions on MPYE or MedA media supplemented with 0, 5, or 10 µM Cu²⁺, as indicated. Colonies that contain an active cbb₃-Cox (NADI⁺) turn blue immediately (<0.5 min) when exposed to a NADI stain, and those that have no (NADI⁻) or low (NADI^slow) cbb₃-Cox activity remain green or become bluer upon longer (>10-min) exposure times. MedA with no Cu addition (MedA-Cu) was prepared by omitting 1.5 µM CuSO₄ from the regular MedA (indicated as 1.5 µM) and MedA with 6.5 µM by adding 5 µM CuSO₄ to the regular MedA (indicated as 0, 1.5, and 6.5, respectively).

Testing of the chemical constituents of different growth media indicated that the growth medium-dependent NADI phenotype of GK1 correlated with the Cu²⁺ content of the medium used (Fig. 1). Earlier inductively coupled plasma mass spectrometry (ICP-MS) analyses indicated that, in the absence of CuSO₄ supplementation or chemical chelation of Cu, both MPYE and Med A lacking Cu (MedA-Cu) media contained approximately 150 nM Cu²⁺ as a “contaminant,” but the level of bioavailable Cu was much lower in MPYE medium (23). When Cu²⁺ was omitted from MedA, which contains 1.5 µM CuSO₄, GK1 became NADI minus. Conversely, when MPYE that has no Cu²⁺ supplement contained 5 µM or more CuSO₄, GK1 regained a NADI^slow phenotype (Fig. 1). Membranes of GK1 exhibited ~2% or 15% of the wild type’s cbb₃-Cox activity (monitored as O₂ consumption activity in the presence of ascorbate and tetramethylphenylene diamine [TMPD]) when cells were grown on MPYE (i.e., lacking Cu) and MedA (i.e., containing Cu) media, respectively (28). Addition of metal ions other than Cu²⁺, including Fe³⁺, Zn²⁺, and Mn²⁺, or redox-active chemicals, such as cysteine/cystine or oxidized/reduced glutathione, did not affect the NADI phenotypes of GK1, indicating that cbb₃-Cox activity in GK1 responded specifically to increased exogenous Cu²⁺ availability. Moreover, R. capsulatus structural (ccoNOQP) or assembly (ccoGHIS) genes of cbb₃-Cox (23, 28) or other genes (dsbAB, senC, and olsAB) known to affect its biogenesis (23, 24, 27) were unable to complement GK1 or HY70 for a NADI⁺ phenotype in the absence of a Cu²⁺ supplement. We therefore surmised that these mutants might reveal a novel component(s) involved in cbb₃-Cox biogenesis.

New gene responsible for active cbb₃-Cox production.

The gene that was defective in GK1 was identified by complementation with chromosomal libraries that were constructed using either the EcoRI or HindIII restriction enzyme. These crosses yielded the plasmids pSE1 and pSE2, which complemented GK1 to a NADI⁺ phenotype without Cu²⁺ supplementation and carried 8.0-kb EcoRI and 4.8-kb HindIII fragments, respectively (see Fig. S1A in the supplemental material). DNA sequence determinations of the end portions of these fragments and their alignments with the R. capsulatus reference genome (http://www.ncbi.nlm.nih.gov) identified the chromosomal region that complemented GK1. The EcoRI fragment contained six intact open reading frames (ORFs), annotated as follows: a protein of unknown function (DUF88, GenBank accession no. RCAP_RCC02189), a heavy-metal-translocating P-type ATPase (RCAP_RCC02190), a transcriptional regulator of the MerR family (RCAP_RCC02191), a major facilitator superfamily member (RCAP_RCC02192), DNA-3-methlyadenine glycosylase II (RCAP_RCC02193), and a phospholipase/carboxylesterase family protein (RCAP_RCC02194). The HindIII fragment, which was contained within the EcoRI fragment, carried only the ORFs RCC02190, RCC02191, and RCC02192 (Fig. S1A). Three derivatives of pSE2 were constructed, namely, pSE201, pSE202, and pSE203, which contain deletions in RCC02190, in RCC02191, and in both, respectively. All three plasmids complemented GK1 to NADI⁺ in the absence of a Cu²⁺ supplement. Concurrently, the knockout mutants SE4, SE5, and SE6, with inactive copies of RCC02190, RCC02191, and both RCC02190 and RCC02191, respectively, were constructed by interposon mutagenesis (see Text S1 in the supplemental material). All three mutants showed a NADI⁺ phenotype (Fig. S1A). The plasmids pSE204 and pSE3 contained only RCC02192 (see Fig. S1A and -B in the supplemental material) and complemented both GK1 and HY70, whereas its derivatives with internal deletions in RCC02192, pSE5 and pSE6, were unable to do so (see Fig. S1B and Table S1 in the supplemental material). These data showed that RCC02192 was defective in GK1 and HY70.

Chromosomal knockout allele of RCC02192.

A chromosomal deletion-insertion allele of RCC02192 was obtained using the gene transfer agent with pSE5 (ΔRCC02192::spe) as a donor (Text S1), to yield the mutant SE8 (Fig. S1B). Like GK1 and HY70, SE8 was NADI⁻ on media lacking a Cu²⁺ supplement and NADI^slow on Cu-containing media. It was fully complemented to a NADI⁺ phenotype in the absence of Cu²⁺ supplementation by pSE3 but not by pRK-GK1 (Table S1), which carries an identical chromosomal DNA fragment derived from GK1. Thus, the defect in cbb₃-Cox activity seen in GK1 and HY70 was confined to RCC02192, which we subsequently named ccoA to recognize its role in cbb₃-Cox biogenesis. DNA sequencing of appropriate chromosomal regions encompassing ccoA defined the molecular bases of the mutation(s) in GK1 and HY70. A single base pair change (C to T) at nucleotide position 345 of ccoA, converting glycine 116 of CcoA to aspartate, and a 6-base-pair-long insertion at nucleotide position 257, resulting in an in-frame insertion of a threonine-alanine dipeptide between positions 86 and 87 of CcoA, were found in GK1 and HY70, respectively (Fig. 2).

FIG 2 .

Topological model of the CcoA (RCAP_RCC02192) protein. The CcoA protein is predicted to have 12 transmembrane helices, like the MFS-type transporters. The topology model was drawn by using the TmRPres2D program (64). The mutations present in GK1 and HY70, the motifs similar to the conserved motifs found in MFS transporters (rectangles), and the metal-binding Mets motifs (ellipses) are also indicated.

CcoA belongs to the major facilitator superfamily of transporters.

In the R. capsulatus reference genome, ccoA (RCC02192) is annotated as a protein of the major facilitator superfamily (MFS). Like the MFS-type transporters, CcoA is an integral membrane protein (predicted to be 405 amino acids), with 12 putative transmembrane helices split into two subdomains of six helices each, separated by a large cytoplasmic loop (Transporter Classification Database [TCDB], http://www.tcdb.org) (29, 30) (Fig. 2). These transporters contain two highly conserved DRXGRR motifs between transmembrane helices two-three and eight-nine (AVYGRR and ARFGRE in CcoA) (31). Remarkably, mutations that inactivated ccoA in GK1 and HY70 are located near the first motif (Fig. 2), suggesting that this portion of CcoA is important for its function. A striking feature of CcoA is the presence of several Mets motifs (M₃₀SM, M₇₀SSFM, M₂₂₃ICGM, and M₂₃₃NLVM) associated with Cu binding and transport in Ctr-type Cu importers (32, 33). Earlier, mutating a conserved and functionally important tyrosine residue located in the same transmembrane helix as a Mets motif (YFLMLIFMT) was shown to decrease Cu transport by a Ctr-type Cu importer (33). Remarkably, a similar sequence (Y₂₃₀ALMNLVMT) is also present on the seventh helix of CcoA (Fig. 2).

Properties of a mutant lacking CcoA.

In order to gain mechanistic insights into how the absence of CcoA decreases the activity of cbb₃-Cox in R. capsulatus, we compared the transcription levels of the ccoNOQP gene cluster between a wild type (MT1131) and its ΔccoA derivative (SE8) in the presence and absence of a Cu²⁺ supplement. Total cellular RNA isolated from appropriate strains grown in MPYE medium with or without a Cu²⁺ supplement were subjected to reverse transcription-PCR (RT-PCR) using ccoN-specific primers (see Materials and Methods). Analyses of the amplification products (Fig. 3A) indicated that the amount of ccoN mRNA was slightly lower (~74% of the wild-type amount) in a mutant lacking ccoA. In the presence of a Cu²⁺ supplement, these amounts further decreased slightly and comparably in the ccoA mutant and the wild-type strain. The data indicated that the lack of CcoA had no significant effect on ccoN transcription and thus most likely on the entire ccoNOQP cluster, which is initiated from a promoter located immediately upstream of ccoN (28).

FIG 3 .

Effects of ccoA mutation on the expression of ccoNOQP structural genes of cbb₃-Cox. (A) The top row shows the results of RT-PCR of total RNA from wild-type (MT1131) and ΔccoA mutant (SE8) strains grown in MPYE media supplemented with 0 or 5 µM Cu²⁺. The middle row shows 16S rRNA gene expression, which is not affected by the ccoA mutation, and the bottom row shows the ratios of ccoN to 16S rRNA gene expression for comparison. The intensities of the bands were determined by ImageJ software (NIH). A control PCR omitting reverse transcriptase enzyme was performed in each case to check for DNA contamination (data not shown). (B) β-Galactosidase activities measured in cell extracts derived from the wild type (MT1131) and ΔccoA mutant (SE8) strain carrying the ccoN::lacZ gene fusion, grown by respiration on MPYE media without Cu²⁺ (−Cu) or with (+Cu) 5 µM Cu²⁺ supplementation. Control reactions refer to the assays performed with cell extracts from the wild type and ΔccoA mutant carrying the pXCA601 plasmid, which contains a promoterless lacZ gene used to construct the ccoN::lacZ fusion. The assays were performed at least in duplicate, and the activities are presented in nanomoles of ONPG hydrolyzed per minute per microgram of total proteins.

Next, a transcriptional-translational ccoN::lacZ fusion construct (pXG1) that carried the 220 bp 5′ of the ATG start codon and the first 13 amino-terminal codons of ccoN (28) was conjugated to the wild-type strain MT1131 and its ΔccoA derivative SE8. The absence of CcoA decreased roughly 2-fold the amounts of β-galactosidase activity produced in these strains grown in MPYE medium without any Cu²⁺ supplement (Fig. 3B). Upon 5 µM Cu²⁺ supplementation, this activity increased slightly and comparably in both strains. Thus, neither the transcription nor the translation initiation of ccoN in R. capsulatus was abolished by the absence of CcoA, and Cu²⁺ supplementation enhanced it only marginally.

The steady-state amounts of cbb₃-Cox subunits were examined in membranes of appropriate strains grown with or without Cu²⁺ supplementation. Subunit I of cbb₃-Cox, CcoN, was monitored by immunoblot analyses using polyclonal anti-R. capsulatus CcoN antibodies (28). Subunits II and III, CcoO and CcoP, were visualized using SDS-PAGE with tetramethylbenzidine (TMBZ) staining, which reveals specifically membrane-bound c-type cytochromes (see Materials and Methods). The amount of CcoN in SE8 lacking CcoA was much lower than that seen in the wild-type strain, MT1131, and addition of 5 µM Cu²⁺ supplement increased this amount in both strains (Fig. 4A). Similarly, the amounts of CcoO and CcoP (cytochromes c_o and c_p, respectively) were lower in the absence of CcoA than in the wild-type strain. Cu²⁺ supplementation increased these amounts (Fig. 4B), even though no effect was seen with other cbb₃-Cox-unrelated membrane-bound c-type cytochromes (e.g., cytochromes c₁ and c_y). Thus, in the absence of CcoA, the steady-state amounts of the structural subunits of cbb₃-Cox in membranes were highly decreased, and Cu²⁺ supplementation palliated this defect(s) partially.

FIG 4 .

cbb₃-Cox subunit profiles and enzyme activity in membranes of various R. capsulatus strains. Chromatophore membranes derived from the wild-type (MT1131), ΔccoNO (GK32), ΔccoNO/ccoNOQP (GK32/pOX15), ΔccoA (SE8), and ΔccoA/ccoA (SE8/pSE3) strains grown at 35°C under respiratory conditions on MPYE medium supplemented (+Cu) or not supplemented with 5 µM Cu were prepared as described in Materials and Methods. (A) Immunoblot analysis to define the amounts of cbb₃-Cox subunit I CcoN. Approximately 50 µg of chromatophore membranes were separated by 12% SDS-PAGE and treated as described in Materials and Methods, using anti-CcoN polyclonal antibodies. (B) Comparison of membrane-associated cytochrome c profiles of various strains. Approximately 50 µg chromatophore membranes prepared from appropriate strains, grown as described above, were separated using 16.5% SDS-PAGE, and the c-type cytochromes were visualized using TMBZ staining as described in Materials and Methods. CcoO and CcoP refer to subunits II and III of cbb₃-Cox, and c₁ and c_y correspond to the cytochrome c₁ subunit of the cytochrome bc₁ complex and membrane-attached electron carrier cytochrome c_y, respectively. (C) cbb₃-Cox activities of various strains determined by monitoring the rate of oxidation of reduced horse heart cytochrome c, as described in Materials and Methods. The Cox activity exhibited by the wild-type strain, MT1131 (1.5 µmol of cytochrome c oxidized per min and per mg of total proteins), was taken as 100 to determine the relative amounts of Cox activities of appropriate strains. At least two independent duplicates were performed for each assay.

The cbb₃-Cox activity present in detergent-dispersed membranes from cells lacking CcoA was determined using reduced horse heart cytochrome c (see Materials and Methods). The total amount of cbb₃-Cox activity detected in SE8 lacking CcoA was ~5 to 10% of that seen in the wild-type strain, MT1131 (Fig. 4C). Upon addition of 5 µM Cu²⁺ supplement, the cbb₃-Cox activity of SE8 increased to ~20 to 30% of that of MT1131, which was unchanged under these conditions. As expected, upon complementation with a plasmid carrying ccoA (e.g., SE8/pSE3), the NADI phenotypes, steady-state amounts, and activities of cbb₃-Cox reached wild-type levels (Fig. 1 and 4). Therefore, the absence of CcoA affected a step(s) that was subsequent to transcription and translation initiation of ccoN during biogenesis of R. capsulatus cbb₃-Cox under respiratory growth conditions. This defect(s) decreased drastically the steady-state amount and activity of cbb₃-Cox, and Cu²⁺ supplementation alleviated it partially.

The absence of CcoA decreases the total Cu content of R. capsulatus cells.

In order to determine whether the absence of CcoA affected the intracellular Cu content, cells grown in MPYE medium with or without 5 µM Cu²⁺ supplement were analyzed by inductively coupled plasma dynamic-reaction cell mass spectrometry (ICP-DRC-MS) (see Materials and Methods). In the absence of a Cu²⁺ supplement, washed and lyophilized cells of SE8 lacking CcoA contained ~20% less Cu than its wild-type parent, MT1131, whereas the amounts of Fe, Mn, and Zn found in these strains were unchanged (Table 1). Both wild-type and CcoA-lacking cells grown in the presence of 5 µM Cu²⁺ contained larger amounts of Cu, but the amount found in the absence of CcoA was again ~60% smaller than that seen in the wild-type parent, MT1131. Thus, in the absence of CcoA, the intracellular Cu accumulation in R. capsulatus cells grown by respiration was significantly decreased but not completely abolished.

TABLE 1 .

Total Cu contents of the R. capsulatus wild type and of the ΔccoA strain and its revertants as determined by ICP-DRC-MS^b

Strain	5 µM Cu2+	Relative level of:
		Cu	Mn	Zn	Fe
Wild type (MT1131)	−	100a	100a	100a	100a
	+	363	111	ND	ND
ΔccoA mutant (SE8)	−	79	107	101	97
	+	146	118	ND	ND
ΔccoARev1 (SE8R1)	−	126	102	101	96
	+	490	137	ND	ND
ΔccoARev2 (SE8R2)	−	150	100	93	92
	+	433	129	ND	ND

^aA mean value of 12 µg Cu, 6.3 µg Mn, 34.5 µg Zn and 268 µg Fe per gr of lyophilized cells was determined and is referred to as 100% for the wild-type strain grown in the absence of Cu²⁺ in MPYE medium under respiratory growth conditions.

^bFor each strain, two sets of independently grown cells were analyzed; for each measurement, at least two repeats were done, and differences of ~10 to 20% were observed between the measurements. For a given strain, the absolute amounts of metals determined varied from culture to culture in MPYE medium, but the trend of metal contents of different strains remained unchanged between the cultures. In each case, the mean value of all measurements was presented as a percentage of the value obtained with the wild-type cells treated under the same conditions, as described in Materials and Methods. + or − refers to cells grown in the presence or absence of 5 µM Cu²⁺ added to MPYE medium.

The absence of CcoA does not affect the production of multicopper oxidase.

Rhodobacter species contain, in addition to cbb₃-Cox, other Cu cofactor-containing enzymes, like Cu/Zn superoxide reductase (34) or multicopper oxidase (laccase or CutO) (35). Unlike Rhodobacter sphaeroides, R. capsulatus does not contain a Cu/Zn superoxide dismutase but has a twin-arginine translocation (TAT) signal sequence containing periplasmic CutO, which confers resistance to Cu²⁺. Mutants lacking CutO exhibit increased sensitivity to Cu²⁺ toxicity, and its protective effect is readily observed under anoxygenic photosynthetic growth conditions (35). In order to test whether CutO enzyme was defective in the absence of CcoA, the R. capsulatus cutO gene (RCC02110) was cloned and an insertion-deletion allele (cutO::kan) was constructed and introduced into both wild-type R. capsulatus strain MT1131 and its ΔccoA derivative SE8 (see Text S1 in the supplemental material). The sensitivity to Cu²⁺ of a ΔcutO single (SE15) and a ΔccoA ΔcutO double (SE16) mutant (Table S1) was determined by a plate growth inhibition assay (see Materials and Methods) and compared to the Cu²⁺ sensitivity of SE8. The sizes of growth inhibition zones surrounding filter disks soaked with various concentrations of Cu²⁺ were determined. As expected, a ΔcutO mutant (SE15) was sensitive to Cu²⁺ under photosynthetic growth conditions, unlike the wild-type strain (MT1131). Unlike the ΔcutO mutant, a ΔccoA mutant (SE8) was not sensitive to Cu²⁺, like the wild-type strain (MT1131) (Fig. 5A). Similarly, a ΔcutO ΔccoA double mutant was not more sensitive to Cu²⁺ than a ΔcutO mutant (not shown). Thus, the data inferred that CutO must still be functional to confer Cu²⁺ tolerance in the absence of CcoA in R. capsulatus.

FIG 5 .

Cu²⁺ response of R. capsulatus strains lacking CcoA and its derivatives. (A) Cu²⁺ sensitivity or resistance phenotypes of the R. capsulatus wild-type (MT1131), ΔccoA (SE8), and ΔcutO (SE15) strains under anaerobic photosynthetic growth conditions on MPYE medium were determined as described in Materials and Methods. Filter paper discs were soaked in 100 µM, 1 mM, and 10 mM Cu²⁺, as indicated, and plates were incubated for 2 days under photosynthetic growth conditions (Ps) to visualize the growth inhibition zones. (B) Cytochrome c profiles of appropriate R. capsulatus strains grown in enriched MPYE medium. See Fig. 4B for details. (C) The Cu²⁺ sensitivity or resistance phenotypes of the R. capsulatus wild-type (MT1131) and ΔccoA (SE8) strains and ΔccoA revertant 1 (ΔccoARev1; SE8R1) and 2 (ΔccoARev2; SE8R2) under respiratory growth conditions (Res) on MPYE medium were determined as described above for panel A, except that filter paper discs were soaked in 10 mM, 20 mM, and 100 mM Cu²⁺, as R. capsulatus is more tolerant to Cu²⁺ under aerobic growth conditions. Arrows indicate Cu-resistant derivatives of SE8R1 and SE8R2.

Bypass suppressors of ccoA are hypersensitive to Cu²⁺.

During the complementation experiments using genomic libraries, we noticed that GK1 reverted to the NADI⁺ phenotype at unusually high frequencies (~10⁻³ to 10⁻⁴) when grown by respiration without the Cu²⁺ supplement. DNA sequence analyses of the ccoA locus in several such revertants indicated that these revertants still retained the initial mutation (a C-to-T change at position 345 of ccoA) carried by GK1 and produced cbb₃-Cox enzyme (Fig. 5B). Similar high frequencies of reversion to the NADI⁺ phenotype were observed with SE8, which carried a deletion-insertion allele of ccoA, suggesting that the revertants restored the ability to produce cbb₃-Cox activity without any need for Cu²⁺ supplementation, thus bypassing the role of CcoA. Unexpectedly, when tested for their response to the Cu²⁺ supplement, the ccoA suppressors SE8R1 [Δ(ccoA::spe)rev1] and SE8R2 [Δ(ccoA::spe)rev2] (see Table S1 in the supplemental material) showed hypersensitivity to Cu²⁺ under both respiratory (Res) (Fig. 5B and C) and photosynthetic (Ps) (not shown) growth conditions. The R. capsulatus wild-type strain MT1131 and its derivative SE8, lacking CcoA, are tolerant up to a millimolar concentration of Cu²⁺, but the ccoA suppressors SE8R1 and SE8R2 were sensitive to micromolar amounts of a Cu²⁺ supplement for respiratory growth inhibition. Indeed, these mutants were partially growth inhibited in MedA medium containing 1.5 µM Cu²⁺ and completely inhibited by addition of ~25 µM Cu²⁺ supplement to MPYE medium. This hypersensitivity was specific to Cu²⁺ only, as no similar effect was seen with other metals, including Fe³⁺, Mn²⁺, Zn²⁺, and Ag⁺, and oxidants, such as cystine or glutathione. We therefore concluded that SE8R1 and SE8R2 regained the ability to produce cbb₃-Cox at the expense of decreased tolerance to Cu²⁺ toxicity.

Total intracellular Cu contents of SE8R1 and SE8R2 cells grown in MPYE medium with and without 5 µM Cu²⁺ supplement were also determined using ICP-DRC-MS analyses (see Materials and Methods). In the absence of a Cu²⁺ supplement, the total intracellular Cu contents of these mutants were greater than in their parent, SE8, which lacks CcoA, and similar or greater than in the wild-type strain, MT1131 (Table 1). In the presence of a Cu²⁺ supplement, all strains accumulated larger amounts of intracellular Cu. The levels found in SE8R1 and SE8R2 were much higher than those seen in the wild-type strain MT1131, although the intracellular amounts of Mn used as an internal control were unchanged in all cases (Table 1). Thus, the suppressor mutation(s) bypassed the absence of CcoA by increasing specifically intracellular Cu accumulation at the expense of compromised cellular tolerance to this toxic metal. The molecular basis of this suppression, which is beyond the scope of this work, remains to be identified.

DISCUSSION

The impetus behind this work was to understand how cells assemble catalytic metal cofactors into the heme Cu_B binuclear center of Cox, a process that is largely uncharacterized. Using R. capsulatus, we initiated a genetic approach to investigate bacterial Cox biogenesis and isolated various cbb₃-Cox-defective mutants. In this work, we focused on mutants that produced an active cbb₃-Cox only upon exogenous Cu²⁺ supplementation. Studies of these mutants uncovered a novel gene, ccoA, which was distinct from ccoGHIS (23), senC (26), olsAB (24), and dsbA and degP (27), known to affect this process. Mutants lacking CcoA were unable to produce normal amounts of cbb₃-Cox activity because the steady-state amounts of the subunits of this enzyme were drastically decreased in membranes. However, neither the transcription of the structural genes ccoNOQP nor the translation initiation of CcoN was abolished in the absence of CcoA, indicating that cbb₃-Cox assembly was defective. Whether ccoA affects only cbb₃-Cox is not known. Its absence does not inactivate the periplasmic Cu-containing multicopper oxidase CutO in R. capsulatus. The putative TAT signal sequence of CutO suggests that this enzyme might acquire its Cu cofactor in the cytoplasm prior to translocation, in a manner similar to that of its Escherichia coli homologue, CueO (3). Investigation of bacterial species that have CcoA homologues, like R. sphaeroides or Bradyrhizobium japonicum, and that also produce aa₃-Cox might further elucidate its role in the biogenesis of other Cox enzymes.

A major finding of this work was that the ccoA gene encodes a member of the MFS-type transporters, which have not been implicated hitherto into cbb₃-Cox biogenesis in bacteria. Remarkably, CcoA homologues with Mets motifs are present in most bacteria that contain cbb₃-Cox (except the epsilonproteobacteria and the Cytophaga-Flexibacter-Bacteroides [CFB] group) (see Table S2 in the supplemental material), suggesting that they are important for the production of this enzyme. How CcoA affects cbb₃-Cox assembly is intriguing. The MFS-type secondary transporters use the electrochemical potential difference generated by ion or solute gradients and transport a diverse range of substrates in and out of cytoplasm (36). Some MFS proteins have been implicated as importers or exporters of siderophores, including E. coli EntS (37), Vibrio parahaemolyticus PvsC (38), and Sinorhizobium meliloti RhtX (39), which secrete enterobactin, vibrioferrin, and rhizobactin, respectively. R. capsulatus mutants lacking CcoA produce various c-type cytochromes that rely on efficient siderophore trafficking and Fe supply for heme production, indicating that CcoA is not involved in Fe uptake. This is also supported by the unchanged intracellular Fe concentration in the CcoA knock out strain (Table 1). Instead, several findings suggest that CcoA is involved in cellular Cu acquisition for cbb₃-Cox assembly. First, mutants lacking CcoA exhibit enhanced cbb₃-Cox activity upon increased exogenous Cu²⁺ supplementation. This enhancement is specific to Cu²⁺ only, as Zn²⁺, Mn²⁺, or Fe³⁺ addition has no similar effects. Mass spectrometry measurements indicated that in cells lacking CcoA, only the total Cu (and not Zn, Mn, or Fe) content is lower than that of a wild-type cell under normal growth conditions. Upon Cu²⁺ supplementation, cellular Cu content increased in mutants lacking CcoA, although it never reached wild-type levels, whereas the cellular contents of metals other than Cu, like Zn, Mn, or Fe, remained unaffected. These findings are consistent with CcoA being involved in a Cu influx rather than a Cu efflux pathway. It is noteworthy that R. capsulatus mutants lacking CcoA still contain cellular Cu, indicating that unrelated and currently unknown Cu acquisition pathways also exist in this species.

Second, mutants lacking CcoA are not more sensitive to Cu²⁺ supplementation than a wild-type R. capsulatus strain under various growth conditions, suggesting that CcoA is not involved in Cu detoxification, unlike, for example, the P1B-type Cu exporters (40–42). Moreover, suppressor mutants that bypass the need for CcoA to recover cbb₃-Cox activity are extremely sensitive to very small amounts (~25 µM) of Cu²⁺ supplement in the medium. Again, this sensitivity is specific to Cu²⁺, as these revertants exhibit normal tolerance towards Zn²⁺, Mn²⁺, Fe³⁺, and even Ag⁺, known to mimic Cu⁺ (42, 43). Mass spectrometry measurements indicated that in these suppressor mutants, Cu content was similar to that seen in the wild-type cells, suggesting that the suppressors overcame the function of CcoA by enhancing Cu acquisition or retention. Indeed, upon Cu²⁺ supplementation, the suppressor mutants accumulated much larger amounts of intracellular Cu (and not Mn) than wild-type cells. Based on the overall findings, it is therefore compelling to rationalize that the absence of CcoA induces an intracellular Cu²⁺ shortage to decrease cbb₃-Cox production (Fig. 6). Consequently, the availability of an increased exogenous Cu²⁺ supply or the occurrence of an additional mutation(s) overcomes this shortage to yield normal amounts of cbb₃-Cox at the expense of a cell’s tolerance of Cu, which is compromised.

FIG 6 .

Hypothetical model of how CcoA affects cellular Cu content and the production of cbb₃-Cox in R. capsulatus. (A) A wild-type strain (MT1131) requires CcoA to produce normal amounts of active cbb₃-Cox and exhibits tolerance to Cu²⁺. (B) In the absence of CcoA (SE8), cbb₃-Cox formation is abolished and intracellular Cu amounts are decreased, but cells remain tolerant to Cu²⁺. (C) The absence of CcoA can be suppressed by a second mutation(s), referred to as ΔccoARevX (e.g., SE8R1 or SE8R2), that restores the production of active cbb₃-Cox at the expense of increased intracellular Cu amounts and renders cells sensitive to Cu²⁺.

Bacteria utilize multiple transporters to achieve the movement of metals across the membrane and maintain homeostasis with metals in nontoxic concentrations (44). In the case of Cu, homeostatic pathways are multiple and complex (45, 46). For example, energy-dependent primary transporters of the P1B-type ATPases are involved in Cu⁺ efflux from the cytoplasm to the periplasm. Of these, CopA1-type transporters have high efflux rates and are involved in Cu detoxification. Their expressions are induced by an excess of Cu, and their absence induces Cu sensitivity (40, 42, 47, 48). In contrast, CopA2-type transporters, like R. capsulatus CcoI of the ccoGHIS gene cluster, have low efflux rates and no role in Cu toxicity but are involved in cbb₃-Cox biogenesis (23, 40). RND (resistance-nodulation-cell division protein family)-type transporters are also involved in the efflux of Cu ions from both the cytoplasm and the periplasm to the extracellular milieu to detoxify cells (49). Clearly, while Cu efflux and detoxification pathways are elaborate, how Cu is imported into the bacterial cytoplasm is less well known (41, 46). Only a few proteins, including Enterococcus hirae CopA (48), Pseudomonas aeruginosa HmtA (50), Bacillus subtilis YcnJ (51), Pseudomonas syringae CopCD (52), the cyanobacterial P1B-type ATPases (CtaA and PacS located in the cytoplasmic and thylakoid membranes, respectively) (53), and two plant chloroplast P1B-type ATPases (PAA1/HMA6 and PAA2/HMA8 of the inner and thylakoid membranes, respectively) (54), have been implicated in Cu import. In eukaryotic microbes like Saccharomyces cerevisiae, mainly the plasma membrane-located Ctr-type transporters are known to import Cu with very high affinity into the cytoplasm (45, 46). So far, no bacterial Ctr homologue has been reported, and until very recently, no MFS-type Cu transporter was known. Very recently, Beaudoin et al. (56) reported a novel forespore membrane Cu transporter, Mfc1, which is distinct from the Ctr-type transporters (55) and involved in meiotic and sporulating cells of Schizosaccharomyces pombe. Mfc1 is a member of the MFS-type transporters, and it functions as a specific Cu importer during meiotic differentiation under Cu-limiting conditions in S. pombe (56). Excitingly, S. pombe Mfc1 is homologous to R. capsulatus CcoA, has a topology similar to CcoA’s (with 12 transmembrane helices with cytoplasmic amino- and carboxyl-terminal ends), and contains other landmarks of MFS-type transporters. Moreover, both Mfc1 and CcoA contain the Mets motifs (MXM, MXXM, and MXCXM) involved in Cu transport. The Cu²⁺ importer function of Mfc1 was established by direct-transport assay using radioactive ⁶⁴Cu and by observation of its ability to complement appropriate S. cerevisiae mutants lacking the known Cu-importing Ctr-type transporters. The similarities found between CcoA and Mfc1 suggest that CcoA also acts as a Cu importer, although direct proof of this is lacking. Finally, as in some methane-oxidizing bacteria (methanotrophs), Cu is imported into the cytoplasm, which is associated with siderophore-like molecules (called chalkophores or methanobactins) (57, 58), and whether CcoA-mediated Cu acquisition involves additional compounds (59) needs investigation.

In summary, our findings establish for the first time that the MFS-type transporter CcoA is required for maintaining normal amounts of intracellular Cu and cbb₃-Cox in R. capsulatus and possibly in other bacterial species. Future work will hopefully elucidate the links between CcoA, cellular Cu acquisition, and cbb₃-Cox biogenesis.

MATERIALS AND METHODS

Strains, culture conditions, phenotypes, and molecular genetic techniques.

Detailed descriptions of strains, culture conditions, phenotypes, and molecular genetic techniques used in this study are presented in the supplemental material (Text S1).

Cu²⁺ sensitivity assays.

Strains to be tested for Cu²⁺ sensitivity were grown to the exponential phase (optical density at 630 nm [OD₆₃₀] of ~0.5) in MPYE medium under appropriate growth conditions. Cells for respiratory (1.7 × 10⁷ cells) and photosynthetic (2.6 × 10⁷ cells) growth conditions (estimated with 1.0 OD₆₃₀ unit equal to 7.5 × 10⁸ R. capsulatus cells per ml) were added to 4 ml of the same medium containing 0.7% top agar and poured on top of regular plates containing 10 ml medium. Whatman 3 MM paper discs (3 mm in diameter), soaked with 8 µl per disc of desired concentrations of CuSO₄ solution, were placed on the surfaces of the plates after solidification of the top agar. Plates were incubated under the desired growth conditions and scanned at the end of incubation period, and the sizes of the growth incubation zones exhibited by different mutants were measured to estimate their responses to CuSO₄ toxicity. Each assay was repeated at least three times.

RNA isolation and RT-PCR.

R. capsulatus cultures were grown semiaerobically in MPYE medium until mid-log phase (OD₆₃₀ of approximately 0.5), and total RNA was isolated from about 2 × 10⁸ cells using the Qiagen RNeasy minikit, digested with DNase I for 25 min at room temperature, and ethanol precipitated. Fifty nanograms of total RNA was used per RT-PCR with the Qiagen OneStep RT-PCR kit, and the ccoN BHK20B (5′ CCAGTCGGGCAGCGCGGTAT 3′) and N3-RT (5′ CGGCAACGGGATGCTGAACTTC 3′) primers amplified a 652-bp-long region internal to ccoN that corresponds to positions 299 and 972 of ccoN. The primers 16SrRNA-F (5′ ATATTCGGAGGAACACCAGTGGC 3′) and 16SrRNA-R (5′ CAGAGTGCCCAACTGAATGATGG 3′) were used as a control for amplification of a 450-bp region of the 16S rRNA gene. In each case, RT-PCR controls were prepared by omitting the reverse transcriptase enzyme from the reaction mixture, the amplification products were separated using 1% agarose gels, and their intensities were compared by ImageJ (NIH).

Cell extract preparation.

Cells were grown in 10 ml of MPYE medium by respiration and harvested at 4,000 pm for 10 min. Pellets were resuspended in 200 µl of CellLytic B 2× cell lysis solution (Sigma Inc.) supplemented with 1 mM phenylmethylsulfonyl fluoride (PMSF), 10 mM EDTA, 50 µg lysozyme, 20 µg DNase, and 10 mM MgCl₂, incubated at room temperature for 15 min, and centrifuged at 14,000 rpm for 10 min. Supernatants thus obtained were taken as whole-cell extracts. Intracytoplasmic membrane vesicles (chromatophore membranes) were prepared in 50 mM MOPS (pH 7.0) containing 100 mM KCl and 1 mM PMSF as described earlier (17). Protein concentrations were determined using the bicinchoninic acid assay according to the supplier’s recommendations (Sigma Inc.; procedure TPRO-562).

Enzyme activity measurements.

Cytochrome c oxidase activity was measured spectrophotometrically using reduced horse heart cytochrome c (Sigma, St. Louis, MO) in a stirred cuvette at 25°C. Horse heart cytochrome c was reduced by incubation for 15 min at room temperature with a 1 mM final concentration of fresh sodium dithionite (100 mM stock solution), which was then removed using a PD10 desalting column (GE Healthcare Life Sciences). R. capsulatus chromatophore membranes were detergent solubilized with 1 mg (wt/wt) dodecyl β-d-maltoside per mg of membrane proteins added to the assay buffer (10 mM Tris-HCl, pH 7.0, 120 mM KCl, and 25 µM reduced cytochrome c). The enzymatic reaction was started and stopped by the addition of solubilized membranes and 100 µM KCN, respectively. The linear range of the assay was controlled by using different amounts of solubilized membranes, and KCN-sensitive Cox activity was calculated as micromoles of cytochrome c oxidized per milligram of membrane protein per minute using an absorption coefficient (ε₅₅₀) of 20, as described earlier (17, 60). The β-galactosidase activities of whole-cell extracts prepared using 10-ml cultures of appropriate strains were measured spectrophotometrically at 420 nm using o-nitrophenyl galactoside (ONPG), as described earlier (23), and specific activity in nanomoles of ONPG hydrolyzed per minute per milligram of protein was determined using an absorption coefficient (ε₄₂₀) of 21,300 M⁻¹ cm⁻¹.

SDS-PAGE, immunoblotting, and heme staining.

For CcoN immunodetection, chromatophore membrane proteins (50 µg) in 62.5 mM Tris-HCl (pH 6.8), 2% (wt/vol) SDS, 25% (vol/vol) glycerol, 0.01% (wt/vol) bromophenol blue, and 5% β-mercaptoethanol were incubated at room temperature for 15 min prior to being loaded and were separated by 12% SDS-PAGE (61). The gels were electroblotted onto Immobilon-P polyvinylidene difluoride (PVDF) membranes (Millipore, Billerica, MA) and probed with R. capsulatus CcoN rabbit polyclonal antibodies (28). Alkaline phosphatase-conjugated monoclonal anti-rabbit IgG (clone RG-16) was used as the secondary antibody (Sigma-Aldrich, Saint Louis, MO), with BCIP (5-bromo-4-chloro-3-indolylphosphate)-nitroblue tetrazolium (NBT) as a substrate (Sigma-Aldrich, Saint Louis, MO) for the detection. For detection of the c-type cytochromes, ~50 µg of total membrane proteins was separated by 16.5% SDS-PAGE (62), and the gels were stained for endogenous peroxidase activity of the c-type cytochromes by using 3,3′,5,5′-tetramethylbenzidine (TMBZ) and H₂O₂ (63).

Determination of cellular Cu content by ICP-DRC-MS.

The cellular Cu contents of various strains were determined using inductively coupled plasma-dynamic reaction cell-mass spectrometry (ICP-DRC-MS). In this technique, aliquots of sample digestions are introduced into a radio frequency plasma where energy transfer processes cause desolvation, atomization, and ionization. The ions thus formed are extracted from the plasma via a differentially pumped vacuum interface and travel through a pressurized chamber (DRC) containing a specific reactive gas that preferentially reacts with interfering ions of the same target mass-to-charge ratios (m/z). A solid-state detector detects ions transmitted through the mass analyzer on the basis of their mass-to-charge ratios, and the resulting current is processed by a data handling system. For sample preparation, at least 1 h prior to use, all containers, glassware, and tubes were washed with 2% nitric acid and rinsed with metal-free Milli-Q water to prevent metal contamination. Metal-free water and buffers were prepared by stirring them for 1 h at room temperature with 5 g Chelex 100 per liter. For each strain, a 1-liter culture was grown by respiration in MPYE medium to an OD₆₃₀ of 0.8 to 0.9, and cells were harvested by centrifugation and washed three times with a metal-free buffer of 20 mM Tris-HCl, pH 8.0, and once with metal-free Milli-Q water. Cell pellets were lyophilized until complete dryness and shipped to Applied Speciation and Consulting, LLC (WA), for determination of total Cu, Zn, Mn, and Fe contents. Fifty milligrams of lyophilized cells was digested completely with aliquots of concentrated HNO₃ and H₂O₂ at 95°C. The digests were diluted to a known final volume (50 ml) with metal-free reagent water and analyzed via ICP-DRC-MS according to a standard procedure of this company. The data were provided in µg of metal of interest per g of cells (ppm).

Chemicals.

All chemicals were of reagent grade and were obtained from commercial sources.

SUPPLEMENTAL MATERIAL

Text S1

Supplemental methods. Download Text S1, DOCX file, 0 MB.

Figure S1

Restriction map of plasmids pSE1 and pSE2 and their derivatives used to complement GK1 and to construct chromosomal knockout mutants. Various plasmids, isolated or constructed as described in Materials and Methods, are shown on the left, and their ability to complement the NADI⁺ phenotype of GK1 is shown on the right. When appropriate, R. capsulatus strains carrying the related chromosomal knockout alleles are also indicated under the plasmids. (A) Plasmids derived from pSE1 and pSE2. See the text for the annotations of the ORFs RCAP_RCC02189, RCC02190, RCC02191, RCC02192, RCC02193, and RCC02194, located on pSE1 and pSE2. (B) The RCAP_RCC02192 gene corresponds to ccoA, which complements GK1, HY70, and SE8 and encodes a major facilitator superfamily (MFS)-type transporter responsible for a Cu²⁺ supplement-independent NADI phenotype and cbb₃-Cox production in these mutants. E, H, Ns, C, B, Bl, As, Ba, Bg, XB, and K correspond to the restriction endonuclease sites for the EcoRI, HindIII, NsiI, ClaI, BstBI, BlpI, AsiSI, BamHI, BglII, XbaI, and KpnI enzymes, respectively. Superscript commas indicate partial genes. Download Figure S1, DOCX file, 0.1 MB.

Table S1

Strains and plasmids used in this work.

Table S2

Phylogenetic distribution of CcoA among various bacterial genomes.