Copper Efflux System Required in Murine Lung Infection by Haemophilus influenzae Composed of a Canonical ATPase Gene and Tandem Chaperone Gene Copies

ABSTRACT Copper is an essential micronutrient but is toxic at high concentrations. In Haemophilus influenzae mechanisms of copper resistance and its role in pathogenesis are unknown; however, our previous genetic screen by transposon insertion-site sequencing implicated a putative cation transporting ATPase (copA) in survival in a mouse lung infection model. Here, we demonstrate that H. influenzae copA (HI0290) is responsible for copper homeostasis involving the merR-type regulator, cueR, as well as six tandem copies of the metallochaperone gene, copZ. Deletion of the ATPase and metallochaperone genes resulted in increased sensitivity to copper but not to cobalt, zinc, or manganese. Nontypeable H. influenzae (NTHi) clinical isolate NT127 has the same locus organization but with three copies of copZ. We showed that expression of the NTHi copZA operon is activated by copper under the regulatory control of CueR. NTHi single copA and copZ mutants and, especially, the double deletion copZA mutant exhibited decreased copper tolerance, and the ΔcopZA mutant accumulated 97% more copper than the wild type when grown in the presence of 0.5 mM copper sulfate. Mutants of NT127 deleted of the ATPase (copA) alone and deleted of both the ATPase and chaperones (copZ1-3) were 4-fold and 20-fold underrepresented compared to the parent strain during mixed-infection lung challenge, respectively. Complementation of cop locus deletion mutations restored copper resistance and virulence properties. NTHi likely encounters copper as a host defense mechanism during lung infection, and our results indicate that the cop system encodes an important countermeasure to alleviate copper toxicity.

N ontypeable Haemophilus influenzae (NTHi) is a gram-negative bacterial pathogen that colonizes the nasopharynx of healthy humans. As a commensal microbe, NTHi is commonly isolated from the upper airways of the human host with a carriage frequency of approximately 20 to 80%. However, dissemination to other anatomical locations results in a range of infections, including otitis media, sinusitis, pneumonia, and exacerbations of chronic obstructive pulmonary disease (COPD). In COPD, NTHi is a leading cause of acute lung infections that accelerate disease progression, causing ;50% of bacteria-mediated exacerbations and ;30% of all COPD exacerbations (1)(2)(3)(4)(5).
In our previous studies we utilized genome-wide mutant fitness analyses to identify factors of H. influenzae required for growth and survival in the mouse lung model (6,7). Our approach, termed HITS (high-throughput insertion tracking by deep sequencing), in which transposon-chromosome junctions are identified and enumerated from transposon mutant libraries by high-throughput DNA sequencing, revealed a core set of genes required for survival in this site. These include known virulence determinants as well as genes in diverse processes, including metal transport (e.g., copper, zinc), for nutrient acquisition and efflux that were previously unrecognized to play a role in pathogenesis in this organism. In fact, validation of two genes with putative transport functions from our HITS screen led to the discovery of a novel zinc transport system crucial for NTHi survival in the mouse lung (8).
Copper is an essential micronutrient and serves as a biological cofactor yet is toxic in excess. All living organisms have developed strategies to handle the toxic effects of copper in the cell. Bacteria possess dedicated systems to maintain copper homeostasis and ensure the appropriate balance of intracellular levels (9)(10)(11). Copper is a necessary cofactor for certain respiratory complexes (redox enzymes) (9), but toxic doses can also cause degradation of Fe-S clusters in vital enzymes of key biosynthetic pathways leading to impaired cell function (12).
To prevent accumulation of toxic copper concentrations, excess amounts must be detoxified and exported out of the cell, and diverse canonical copper efflux mechanisms implicated in this function have been identified among bacteria. Copper homeostasis involves at least three main proteins, a copper-exporting ATPase (Cu ATPase) to remove copper from the cytosol, a copper chaperone that delivers copper to the Cu ATPase, and a metalloregulator that modulates activity of the Cu ATPase and chaperone (9)(10)(11). Many gram-negative bacteria (e.g., Escherichia coli, Pseudomonas spp.) possess additional systems for removing copper from the periplasm and out into the extracellular space (9,10,13). Studies of mechanisms of copper transport and homeostasis in these two prototypical model organisms indicate that they have multiple copper efflux systems with seemingly redundant functions. Two are chromosomally encoded: the primary copper efflux system, cue, which consists of a copper exporter (CopA), periplasmic oxidase (CueO), and copper-responsive metalloregulator (CueR), and the cus system (CusCFBA or CusCBA), which is involved in periplasmic copper efflux. The third is a plasmid-encoded system containing a six-to seven-gene operon (Pco or Cop system) also involved in periplasmic copper handling (10,(14)(15)(16). Most Proteobacteria have homologs or variations of the Cue, Cus, or Pco (Cop) systems.
Emerging evidence indicates that the innate immune response utilizes the toxic properties of copper to defend against infectious microorganisms (17,18). While the mechanism by which copper exerts antimicrobial activity is not well understood, copper tolerance has been recognized as an important virulence mechanism in pathogenic bacteria (19). The role of copA alone or in combination with other copper resistance genes has been evaluated in animal models of infection for numerous pathogenic gram-negative and gram-positive bacteria, and copA was found to be required for virulence in some bacteria (Acinetobacter baumannii, Klebsiella pneumoniae, Streptococcus pneumoniae, Pseudomonas aeruginosa, Streptococcus agalactiae and Salmonella enterica serovar Typhimurium) (20)(21)(22)(23)(24)(25)(26), but not in all cases. In some studies deletion of copA alone (Helicobacter pylori, Listeria monocytogenes, Streptococcus suis, Streptococcus pyogenes, Actinobacillus pleuropneumoniae) (27)(28)(29)(30)(31) or deletion in combination with a gene encoding copper amine oxidase in Vibrio cholerae (32) was inconsequential for in vivo virulence. Moreover, in some species, copA was differentially required for colonization in distinct tissue sites in mice (P. aeruginosa, S. agalactiae) (24,25). Regardless of whether copA was required for virulence in vivo, copA was needed for copper tolerance in vitro when tested in the studies cited above, and these studies suggest that the role of specific copper tolerance factors during pathogenesis may differ depending on the constellation of systems present in each species.
In the genome of H. influenzae Rd, HI0290, HI0291-2, and HI0293 are annotated as a putative copper-transporting ATPase, copper chaperones, and a copper efflux regulator, respectively, but no gene encoding the multicopper oxidase CueO is evident, nor are genes encoding the Cus or plasmid-encoded Pco systems. H. influenzae appears to have a simplified copper efflux system which could provide an opportunity to examine this potential virulence feature in a more streamlined context compared to that of organisms with multiple copper resistance pathways. The objective of this study was to characterize the system responsible for copper homeostasis in H. influenzae, for which no studies on copper tolerance have yet been reported, and apply this information to further examine its role in NTHi pathogenesis. The results of this study suggest that copper tolerance is an important virulence mechanism for NTHi in the lung and that the copper efflux system is required for optimal survival in this niche.

RESULTS
The genomic organization of the cop locus in H. influenzae shows variable copies of the metallochaperone gene. In our previous genetic screen using the HITS method (6, 7), we identified a putative H. influenzae copA (locus HI0290 when mapped to reference strain KW20) as likely to be essential for growth and survival in a mouse lung infection model. In this screen, abundant transposon insertions in HI0290 (called copA here) present in our input H. influenzae mutant library were underrepresented in the population recovered after infection with this mutant library in the mouse lung model (see Fig. S1 in the supplemental material) (6). Bioinformatic analysis suggests that copA encodes a copper-transporting P 1B -type ATPase that effluxes Cu 1 across the inner membrane and plays a role in copper homeostasis and tolerance (33,34). The predicted CopA Hi is a 722-amino acid protein with a single N-terminal metal binding domain (N-MBD) and 8 predicted transmembrane-spanning helices. The signature sequence motifs of CPCX 6 P in helix 6 (where X is any amino acid), YN in helix 7, and MX 2 SS in helix 8 are strictly conserved in P 1B-1 -ATPases, suggesting the presence of these residues as key determinants for copper specificity (34). For example, H. influenzae CopA has 38% identity/57% similarity to CopA1 of Pseudomonas aeruginosa (PA3920) and 37% identity/55% similarity to E. coli CopA (b0484). In H. influenzae genomes, examination of the DNA regions surrounding copA revealed genes that are likely involved in copper tolerance. Several metallochaperone genes (called copZ here) are located directly upstream of copA in an apparent operon, and this operon is likely controlled by the proximal MerR-like transcriptional regulator (called CueR here) (Fig.  S1). The putative H. influenzae CopZ metallochaperones have 41% identity and 65% similarity to the well-characterized Cu(I)-binding protein CopZ of Bacillus subtilis (BSU33510) and display the conserved metal-binding CXXC motif (35)(36)(37)(38). CopZ chaperones function to traffic copper ions to the ATPase and buffer against the toxic effects of copper in the cytoplasm (38)(39)(40). The amino acid sequence of the H. influenzae regulator CueR is highly similar to that of the copper-responsive CueR which controls copA expression in many gram-negative bacteria (41,42). For example, CueR Hi has 49% identity/67% similarity to CueR of P. aeruginosa (PA4778) and 40% identity/64% similarity to CueR of E. coli (b0487). The CueR Hi sequence possesses the signature motifs of CXGX 5 C for copper specificity in the metal binding loop and SX 2 V motif in the N-terminal region of the dimerization helix (43).
In the draft genome sequences of H. influenzae RdAW (at least 99.98% identical in nucleotide sequence to the reference strain Rd KW20 [6]) and the clinical NTHi isolate NT127, the cop locus is represented in several contigs. To determine the genetic organization of the locus in these strains, regions were amplified by PCR using primers that bound to sites in cueR and copA (Fig. S2). Sequencing of the NT127 and RdAW fragments revealed three copies of the chaperone, designated copZ1, copZ2, and copZ3, and six copies of the chaperone (copZ1 through copZ6), respectively. To examine whether tandem repeats of copZ also exist in other H. influenzae strains, we surveyed the length of the operon between cueR and copA by PCR (Fig. S3). We found variation of copZ copy number ranging from three to seven among NTHi isolates, with six copies of copZ in Rd KW20 and RdAW rather than the two annotated copies for KW20 in the NCBI database (GenBank accession no. L42023.1).
Comparison of the amino acid sequences of CopZ copies in both strains reveals three amino acid differences among the two strains that are either conservative or semiconservative substitutions as well as the conserved metal-binding CXXC motif (CGCC) at residues 13 to 16 (Fig. S4). CopZ1 and 3 are identical in amino acid sequence within NT127; CopZ3 through Z6 are identical in amino acid sequence within RdAW (the predicted CopZ proteins HI0292 and HI0291 of Rd KW20 are identical in amino acid sequence to CopZ1 and CopZ3 to Z6 of RdAW, respectively). The genetic organization of the putative cop locus in RdAW and NT127 is shown in Fig. 1. The occurrence of multiple nearly identical tandem copies of the copZ chaperone gene appears to be unique to H. influenzae and its close relative Haemophilus haemolyticus (see Discussion).
Genes of the cop locus are required for copper tolerance in H. influenzae. The copA gene has been shown to be needed for copper tolerance in vitro in the model organism E. coli (44) as well as in many bacterial systems mentioned earlier, but this appears not to be the case for Bordetella pertussis (45). In some studies that have evaluated the role of copZ in copper resistance, for example in Bacillus subtilis and Streptococcus mutans, copZ was shown to contribute to copper tolerance in vitro (46,47), while in Listeria monocytogenes growth of a copZ mutant was not sensitive to copper (28). To initially evaluate the importance of copZA in copper homeostasis in H. influenzae, first a mutant that contains a precise deletion of the copper chaperones and the ATPase, RdDcopZA/v, was constructed by gene replacement in RdAW (Table 1) (Materials and Methods). Growth inhibition by metal cations commonly transported by P 1B ATPases was assayed in the parental strain Rd/v (48) and mutant RdDcopZA/v strains ( Fig. 2A). Culture densities after 5 h of growth in various concentrations of copper, cobalt, zinc, or manganese showed that Cu 21 at concentrations of 32 mM or greater inhibited the growth of the copZA mutant relative to the parental strain, Rd/v. Culture densities between both strains were equivalent in the presence of Co 21 , Zn 21 , or Mn 21 . Additionally, the growth kinetic profiles in Fig. 2B showed very little to no growth of the copZA mutant in 256 mM and 512 mM CuSO 4 , while complementation of the copZA mutant restored growth to parental levels. Therefore, these results implicate the copZA operon in copper homeostasis and tolerance in H. influenzae.
To address the role of copper tolerance genes in a clinical isolate of NTHi, a series of defined mutants was constructed in NT127 by deletion of genes encoding the regulator (NTDcueR/v), the chaperones (NTDcopZ/v), the Cu 1 -ATPase (NTDcopA/v), or both the chaperones and Cu 1 -ATPase (NTDcopZA/v) ( Table 1). Nonpolar deletions were constructed by replacing the coding regions with an antibiotic resistance marker, and mutants were complemented by expression of the intact genes ectopically at the xyl locus (Materials and Methods) (49). In growth assays comparing the effect of copper on NTHi strains, NTDcopZA/v showed the greatest sensitivity to increasing copper concentrations, and the growth profile of this mutant deviated from other strains at concentrations equal to or greater than 100 mM CuSO 4 (Fig. 3C). The strain NTDcopZ/v displayed inhibition at 250 mM CuSO 4 (Fig. 3E), while NTDcopA/v displayed inhibition only after the addition of 500 mM CuSO 4 (Fig. 3G). NTDcueR/v grew similarly to the parental strain at all concentrations tested. Complementation restored copper tolerance to parental levels in strains containing the DcopZA, DcopZ, or DcopA deletions (Fig. 3D, F, and H). However, NTDcopZ/c displayed partial complementation (Fig. 3H), likely due to the presence of only one copZ gene in the complementing construct relative to the three copies in the parental strain. Together, these data suggest that the copA and copZ Regulation of the copper locus and analysis of promoters. In the copA promoter of E. coli, the -35 and -10 motifs display extended 19-bp spacing and flank operator sites, a characteristic of MerR-type promoters. Metal-responsive regulators in this family can mediate repression and activation through modulation of DNA structure at the promoter site (50)(51)(52)(53). Transcriptional activation of CueR occurs following a conformational DNA change induced by the regulator in the presence of Cu 1 (41,50,54). The H. influenzae cop (i.e., copZ1) promoter contains features common to E. coli PcopA (Fig. 1). In H. influenzae, the sequence of the apparent CueR binding site (arrows in Fig. 1) is an imperfect palindrome interrupted by seven base pairs, ACCTTaA-N 7 -TaAAGGT. This motif is conserved with 100% nucleotide sequence identity in 101/102 complete genome sequences of H. influenzae strains (NCBI database https://www.ncbi.nlm.nih.gov/genome/) and one strain showing 86% nucleotide sequence identity. When compared to the PcopA In H. influenzae genomes, cueR is divergently transcribed from the copZA operon. Thus, two promoters (PcueR and Pcop) are likely located in this short intergenic region. A putative PcueR is predicted with 17-bp spacing between the -10 and -35 elements (bottom strand) (Fig. 1). In this proposed scheme, the -10 element of PcueR overlaps with the CueR box (right arrow) in the cop promoter, and this constraint may account for the differences between the CueR binding sequences of E. coli and H. influenzae.
To examine the activity of the cop promoter in H. influenzae in response to copper,  (Table 1). This strain carries the 76-bp region between the initiation codons of cueR and copZ1 ( Fig. 1) fused to the promoterless lacZ gene of E. coli. Growth in the absence of copper addition resulted in minimal detection of b-galactosidase activity, while in the presence of 250 mM copper, the activity increased approximately 12-fold (P , 0.0001) (Fig. 4). To determine if CueR regulates expression of the cop promoter in NTHi, the Pcop-lacZ fusion was integrated into a cueR-deficient strain, NTDcueRPcop-lacZ. In this strain background, deletion of cueR abolished the copper-mediated induction of the lacZ fusion (Fig. 4). These findings are consistent with a role for CueR as a copper-responsive activator that senses and responds to cytoplasmic copper status as seen in other organisms (41,(54)(55)(56). CueR HI also mediates repression in the absence of toxic levels of copper, as ;3.7fold higher levels of lacZ reporter expression were observed under this condition in the cueR-deficient strain compared to the parental cueR 1 strain, NTPcop-lacZ (Fig. 4). The role of CueR as a repressor is also consistent with reports in E. coli studies that show its ability to switch from repressor to activator mode via a DNA-distortion mechanism in response to metal binding (57), as well as by modulating RNA polymerase interactions to either repress or activate transcription at the copA promoter (58).
Deletion mutants of copZA accumulate copper. Copper sensitivity of the cop mutants in H. influenzae suggests excess copper is not adequately removed and could potentially accumulate to cause toxicity. To examine cellular copper accumulation, strains were grown in media containing sublethal concentrations of copper, and the total copper levels were measured using furnace atomic absorbance spectrometry. Figure 5 shows that the NTDcopZA/v mutant accumulated ;66% and ;97% more copper than parental strain NT/v when grown for 2 h in the presence of 250 mM (Fig. 5A) and 500 mM copper (Fig. 5B), respectively. Complementation of the mutation in NTDcopZA/c restored its ability to lower cellular copper levels despite slightly elevated levels of 27% and 30% above those of NT/v at 250 mM and 500 mM, respectively.
The copZA operon is required for NTHi infection of the mouse lung. Copper tolerance has been recognized as an important virulence mechanism in several pathogenic bacteria but not in all cases as mentioned earlier (20)(21)(22)(23)(24)(25)(26)(27)(28)(29)(30)(31)(32). To evaluate the fitness defect of the copA-deficient mutant predicted by HITS (Fig. S1) (6), NTHi strains were assessed for their ability to infect the lungs of 6-week-old C57BL/6 mice (Fig. 6). As duplicated genes such as copZ are functionally redundant and disruption of one copy does not result in any detectable phenotypic changes, this class of genes was excluded Copper Efflux System of H. influenzae Drives Lung Infection Infection and Immunity from our previous HITS analysis. To identify potential synergistic contributions of copA and copZ to survival in vivo, we also tested the copZA deletion mutant and its complemented derivative. The infections were conducted as competition experiments between an experimental strain and the lacZ-containing reference strain, NTlacZ (Table 1). At 24 h postinfection, competitive indices (CI) were calculated from CFU differentially enumerated on X-Gal (5-bromo-4-chloro-3-indolyl b-D-galactopyranoside)-containing media (Materials and Methods). Relative to the CI of the parent strain NT/v, mutations in copA and copZA (strains NTDcopA/v and NTDcopZA/v) were underrepresented during infection by ;4-fold (P , 0.05) and ;20-fold (P , 0.001) ( Fig. 6A and B), respectively. This suggests that the copZA mutation is more attenuating than the copA mutation alone. Complementation restored the ability of these mutants to survive in the lung. CI ratios between NTDcopA/c and NTDcopA/v differed by ;20-fold (P , 0.01), while NTDcopZA/ c and NTDcopZA/v showed an ;100-fold difference (P , 0.001) ( Fig. 6A and B, respectively), indicating that the attenuation in virulence was specific to the mutations of copA and copZA. In fact, the complemented strains of the respective mutants exhibited slight competitive advantages relative to the parent strain, as comparisons of CI ratios between each of the NTDcopA/c and NTDcopZA/c groups versus NT/v showed an ;5fold difference (P , 0.05 and P , 0.001, respectively). It is possible that in vivo expression of the complementing genes at the xyl locus exceeds that of the parental strain, leading to improved survival. Overall, the results from the lung infection data indicate that the copper efflux system is required for colonization and survival of pathogenic NTHI in the mouse lung.

DISCUSSION
To tolerate excess copper and prevent its toxic effects, bacteria use multiple pathways to maintain copper homeostasis in the cytoplasm and additionally in the periplasm of gram-negative bacteria. The mechanisms of copper transport and homeostasis have been best characterized in prototypical model species such as Escherichia coli and Pseudomonas syringae for gram-negative bacteria and Enterococcus hirae for gram-positive bacteria (9-11). One major copper transport system encoded by the cue (Cu efflux) locus (first identified in E. coli) encodes a copper-responsive regulator CueR, a periplasmic multicopper oxidase CueO, and a P-type ATPase transporter CopA responsible for controlling cytosolic copper levels (41,44,54,55). Subsequently, it was determined that the copA gene of E. coli also encoded a second protein, CopZ, a copper chaperone expressed as a result of programmed ribosomal frameshifting (59, 60). The cus system (cusCFBA), also first identified in E. coli (61), is proposed to remove excess levels of copper from the periplasmic to extracellular space (9,10). In addition to these chromosomally encoded copper resistance genes, the Pco systems are plasmid-borne copper resistance determinants first characterized in E. coli (pcoABCDRSE) and Pseudomonas syringae (copABCDRS) with cop homologs identified in other pseudomonads (9,10).
Most bacterial genomes encode at least one copper efflux system where the Cu-ATPase (copA) is usually present in an operon with a Cu-responsive transcriptional regulator and a copper metallochaperone. Typically, the metallochaperone (usually called copZ) is encoded by a separate gene (62), unlike the programmed frameshift seen in E. coli. In most gram-negative bacteria, copA is activated by a CueR homolog, whereas in most gram-positive bacteria and Mycobacterium tuberculosis the Cu ATPase is negatively regulated by CsoR (e.g., Bacillus subtilis) (42) or CopY (e.g., Enterococcus hirae, Streptococcus spp.) (11,23,47,63).
Gram-negative bacteria have homologs or variations of the Cue, Cus, or Pco (Cop) systems with homolog names differing between and sometimes within bacterial species. Some possess multiple copper resistance systems, such as E. coli and Pseudomonas spp. In this report, we showed that H. influenzae appears to rely primarily on a simplified copper defense system which comprises a copper-responsive regulator cueR, tandemly repeated copper chaperone genes, copZ(n), and P-type Cu-ATPase copA organized adjacently in the genome (Fig. 1). We showed that NTHi single copA, copZ, and double deletion copZA mutants are defective for copper tolerance in growth toxicity assays (Fig. 3). Complementation of these mutants restored growth in excess copper, indicating that these gene products function to detoxify copper in this organism. The copZA gene products are specific for detoxifying excess copper and not cobalt, zinc, or manganese (Fig. 2). Moreover, deletion of copZA resulted in cellular copper accumulation in the presence of increasing amounts of CuSO 4 during growth compared to the parent and complemented strains, consistent with the role of this system in copper transport and likely accounting for the growth inhibition (Fig. 5). Our collective results are consistent with observations that cop transport/chaperone mutants in diverse pathogenic bacteria are growth inhibited under excess copper conditions that lead to accumulated intracellular copper (20,23,25,28,30,31,47,64,65). For example, in Streptococcus mutans, copYAZ is required for copper resistance, and a copYAZ mutant accumulates a 2-fold increase in the amount of intracellular copper compared to the wild-type strain (47). We also showed with a cop promoter-driven lacZ reporter fusion that expression of the copZA operon is activated by copper under the regulatory control of CueR, consistent with its function as the transcriptional regulator that senses intracellular copper and activates synthesis of the copA efflux pump (Fig. 4). CueR Hi is within the MerR family of regulators first described in E. coli as a transcriptional activator of the Cu ATPase copA and copper efflux oxidase cueO genes (41,54,55). Promoter lacZ fusion assays in E. coli in these cited studies showed induction of the copA promoter by copper which was abolished in a mutant with a deletion of cueR (cueR is termed copR in the Petersen and Møller study [55]). In addition to its role as a transcriptional activator, CueR can also act as a repressor in the absence of copper (57), which we observed for NTHi in this report. It is not clear why negative regulation by CueR of cop promoter lacZ fusions in the absence of copper has been seen in some studies (56), including ours here, and not in others (41,54,55), but perhaps CueR repressors have alternative regulatory properties in different bacteria.
We verified our previous HITS screening data indicating that a mutation in copA attenuates survival in the mouse lung (6,7). We also showed a contribution of the chaperone genes in the copZA mutant for in vivo survival after deletion of tandem copies of copZ compared with deletion of copA alone. The competitive indices suggest that both NTHi copZA and copA mutants were attenuated (;20-and 4-fold, respectively) compared to the parental strain, and complementation of these mutants restored survival above parental levels, indicating that the attenuation was specific to these mutations (Fig. 6). Deletion of copA and the duplicated copZ genes together caused greater attenuation in mouse lungs than deleting copA alone, establishing a contributory role for copZ and indicating that the copper chaperones and efflux ATPase are needed together for optimal survival in vivo. These results correlate with the more severe growth inhibition seen with the copZA double mutant versus the copA single mutant in the presence of increasing concentrations of CuSO 4 (Fig. 3), indicating synergism of the copZ and copA gene products for in vitro copper tolerance and in vivo pathogenicity. The additive effect of deleting copZ in the copA background suggests that copZ mediates a copA-independent function in copper resistance. Future studies will be required to determine whether CopZ interacts with an additional transporter or possesses a copper-sequestering capability.
The copA gene plays an important role in mouse pulmonary infection by several bacterial species; however, this is not always the case, and the role of copZ in pathogenesis has been underexamined. In animal models copA mutants have shown attenuation ranging from ;10-fold to 2-log differences compared to parent strains at 24 h of infection for the respiratory pathogen Streptococcus pneumoniae (22,23) and opportunistic pathogens such as Klebsiella pneumoniae (21) and Acinetobacter baumannii (20). However, a copA mutant was not attenuated for lung colonization compared to the parent in the porcine respiratory pathogen Actinobacillus pleuropneumoniae (31). In infection studies with M. tuberculosis, a mutant lacking the copA-like P-type ATPase exhibited a bacterial burden equivalent to that of the parental strain but less lung disease severity in mice (and guinea pigs) (66). In contrast, effective lung (and lymph node) colonization by M. tuberculosis required a separate putative copper transporter in the guinea pig model (67). Recently, in Bordetella pertussis, deletion of a three-gene operon encoding CopZ, glutathione-dependent peroxidase, and glutathione reductase did not affect colonization appreciably in mouse lungs, yet deletion of this operon did cause a reduction in persistence in the nasal cavity, suggesting that CopZ may contribute to nasopharyngeal colonization in this organism (45).
One interesting feature of the H. influenzae cop system is the tandem duplication of the metallochaperone gene, copZ. The adjacent genomic organization of cueR, copZ, and copA appears to be conserved in all H. influenzae strains and other members of the Pasteurellaceae family; however, we found that tandem variable duplication of copZ appears to be unique to H. influenzae and its close relative Haemophilus haemolyticus and to our knowledge has not been reported in other bacterial systems. Pseudomonas aeruginosa has been reported to have two copper chaperones, CopZ1 and CopZ2, but they have distinct functions, sharing only 37% amino acid sequence identity to each other, and are located in different genomic locations (68,69). Maintenance of gene duplication in bacteria is thought to be an adaptation to environmental stress to survive in diverse environments (70)(71)(72)(73). Perhaps this is the case for maintaining multiple copies of the copper chaperone gene in H. influenzae. Gene duplication of virulence determinants such as capsule and pilus have been documented in H. influenzae and appears to correlate with increased pathogenicity (74). The cap locus (11 genes encoding capsule production) has 10 of these 11 genes duplicated in a cluster in H. influenzae type b (75)(76)(77), whereas in H. influenzae biogroup aegyptius, the entire hifABCDE locus encoding pilus production is duplicated (78,79). During the course of genome sequencing and assembly of several H. influenzae biogroup aegyptius strains, Phillips et al. (80) noticed that the genome size of one of the strains sequenced was ;28 kb larger than the whole-genome shotgun contigs of the same strain deposited by another group. The authors concluded that the discovery of extra sequences was due to using a long-read sequencing platform instead of short-read sequencing. Interestingly, the extra sequences are duplications of several regions including known virulence determinants such as genes that encode for lipooligosaccharide (LOS) phosphorylcholine addition, pilin, and pilus export.
Gene duplication as an adaptive response to metal resistance has been documented in a variety of bacteria and fungi (70,81). For example, copper sequestration studies in the opportunistic pathogen Ralstonia pickettii showed that when isolated from lake sediment contaminated with high levels of copper, it had adapted by duplicating metal resistance and transporter operons, although tandem duplication of the copper chaperone gene was not reported (82). In the budding yeast Saccharomyces cerevisiae, adaptive gene amplification of the copper binding metallothionein gene CUP1 is associated with copper tolerance and detoxification (83). The mechanism by which copZ duplications have occurred in H. influenzae is unknown, but gene duplication as a defensive strategy likely favors survival against copper stress in host environments.
Does H. influenzae have multiple copper efflux systems? First, with respect to the E. coli Cue system, H. influenzae does not appear to encode the multicopper oxidase CueO in its genome. A BLASTP (84) homolog search with E. coli CueO identified a gene annotated as SufI (a member of the multicopper oxidase family) in H. influenzae. To our knowledge, no studies of SufI have been reported for this organism; however, the SufI (FtsP) homolog in E. coli was shown to not bind copper and functions specifically in cell division (85,86). H. influenzae does not appear to have a cus copper/silver efflux system for transporting excess copper from the periplasm to the extracellular milieu. BLASTP (84) searches of the H. influenzae NCBI taxonomy database (87) against E. coli cusCFBA genes did not identify significant sequence similarities to CusF (periplasmic chaperone that transports copper in the periplasm to the pump encoded by cusCBA) and CusB (periplasmic membrane fusion protein). However, sequence similarities to CusC were identified (outer membrane channel-forming protein) and CusA (cytoplasmic membrane substrate binding and transport protein belonging to the resistance-nodulation-division [RND] family of efflux pumps). CusC has ;25% amino acid sequence identity to the outer membrane protein TolC in H. influenzae (HI1462) (88,89), and CusA has ;23% amino acid sequence identity to multidrug efflux RND transporter permease subunit AcrB in H. influenzae (HI0895) (90)(91)(92)(93). AcrB and TolC are components of the tripartite AcrAB-TolC efflux pump that contribute to multidrug resistance found in gram-negative bacteria (94), including H. influenzae (89).
Studies of E. coli have shown that TolC is very similar structurally to CusC (95) but cannot functionally replace CusC (96). Interestingly, several recent reports have shown TolC to have a role in copper tolerance. In the nitrogen-fixing bacterium Bradyrhizobium liaoningense, a transposon mutant library grown under copper selection identified tolC among six other genes to be involved in copper resistance (97). The tolC gene in the cyanobacterium Synechocystis has a role in copper efflux, as growth of a tolC mutant is inhibited greater than that of the wild type in the presence of copper and also accumulated ;3-fold more intracellular copper than the wild type (98). Also, a chemical genetic screen of the Keio collection of 3,985 E. coli deletion mutants with prolonged exposure to copper identified tolC as required to tolerate copper stress (99). Together, these studies suggest the possibility that an unidentified TolC-dependent efflux system analogous to the cus system (which utilizes a TolC-like protein) may be involved in copper resistance, and it remains to be seen if such a system may represent the periplasmic copper efflux system of H. influenzae.
Copper plays an important role at the host pathogen interface, as it is used by cells of the innate immune system as an antimicrobial agent (19,100). Copper transport into macrophages has been shown to promote bactericidal activity, and bacterial systems such as E. coli, S. Typhimurium, and S. pneumoniae containing deletions in copper exporters (e.g., CopA, GolT) were more susceptible to macrophage-mediated killing (17,22,26). It would be interesting in future studies to examine the H. influenzae Cu 1 -ATPase and chaperone mutants for susceptibility to killing by macrophages. Another interesting attribute associated with the copper efflux system is its role in biofilm formation, as growth in a biofilm mode is used by many bacteria to survive in response to environmental stresses such as nutrient limitation, pH and temperature changes, and exposure to antimicrobial agents and, as such, may have a role in the pathogenesis of disease (101). For example, in S. pyogenes, which causes a wide range of clinical illnesses, biofilm formation of the wild-type isolate was inhibited by copper (63), while CopZ in the dental-associated pathogens Streptococcus mutans and Streptococcus gordonii was critical for biofilm formation and detachment, respectively, in vitro (102,103). It remains to be seen whether the cop system of H. influenzae might also have a role in biofilm formation/detachment.
In summary, our findings in this report increase our understanding of NTHi pathogenesis, as copper is a host immune defense used to intoxicate invading pathogens. As recent work has demonstrated that a copper efflux system can be targeted with antibacterial oligopeptides that enhance copper-mediated toxicity (104), understanding copper tolerance pathways in NTHi could help in designing potential antimicrobial strategies targeting copper homeostasis in this respiratory pathogen of significant clinical relevance.
Plasmid and strain construction. Standard molecular biology techniques (106) were used for PCR, cloning, and plasmid construction. Nonpolar deletions of genes or regions of interest were constructed via gene replacement with the aacCI Gm-resistance (Gm r ) cassette, which comprises 228 bp of sequence upstream of the initiation codon and the complete aminoglycoside-(3)-acetyltransferase coding sequence. The Gm r cassette was fused to flanking H. influenzae fragments by overlap extension PCR (OE-PCR) (107) via tails added to the amplification primers. PCR products used as templates in splicing reactions were gel purified (Qiagen). The primers used in the study are listed in Table S1. For complementation of H. influenzae mutants, DNA fragments were amplified by PCR and cloned between adjacent restriction sites of the chromosomal delivery vector pXT10, as previously described (49). Typically, pXT10 was digested with SapI or the isoschizomer BspQI, and inserts were digested with EarI or BspQI. Ligation reactions were dialyzed and electroporated into E. coli DH5a. Following purification, plasmids were linearized by digestion with ApaLI prior to transformation and selection for double crossover homologous recombination. Competent cell preparation and transformation was accomplished as previously described (108). The plasmids used in the study are listed in Table 1.
Construction of H. influenzae Rd mutants. The Rd RdDcopZA mutant was constructed by replacement of the region containing the six metallochaperones (copZ 1-6 ) and HICG_01009 (copA) coding sequences with the Gm r cassette. The replacement construct was created from three fragments: a 59 flanking fragment (697 bp) from Rd, amplified using primers 5CuATPase1 and 3CuATPase1pG; the Gm r cassette (762 bp), using primers 5pGent1 and 3Gent2, and a 39 flanking region (796 bp) from Rd, using primers 5CuATPase2pG and 3CuATPase2. Products were joined by OE-PCR using primers 5CuATPase1 and 3CuATPase2. The resulting 2,207-bp amplicon was added to competent cells of strain Rd and plated on sBHI agar containing Gm, and Gm-resistant (Gm r ) transformants were isolated and verified by PCR. To generate the strain RdDcopZA/v, which contains an empty complementation vector, RdDcopZA was transformed with linearized pXT10, and Tc r transformants were isolated and verified by PCR. For complementation of the RdDcopZA mutant, a segment containing the upstream promoter element, the six metallochaperones, and the ATPase gene was amplified by PCR with primers CuATP_comp_F and CuATP_comp_R, which bear restriction sites at the 59 primer termini. The resulting 3,848-bp fragment was cloned into pXT10, creating plasmid pXRcopZA. This plasmid was linearized and added to competent cells of strain RdDcopZA, and Tc r recombinants were isolated and verified by PCR to yield strain RdDcopZA/c.
Construction of NTHi NT127 mutants. To create strain NTDcopZA, copA and the three copZ genes were deleted by transformation of the Gm-marked 2,207-bp stitched product from Rd (above) into strain NTX, a derivative of NT127 in which the xylose locus was modified for efficient recombination with pXT10-based plasmids (8). To generate NTDcopZA/v, which carries the empty complementation vector, NTDcopZA was transformed with linearized pXT10, and Tc r transformants were isolated and verified by PCR. To complement the NTDcopZA mutant, a segment containing the upstream promoter region, the three copper metallochaperones and the ATPase gene were amplified by PCR from NT127 with primers CuATP_comp_F and CuATP_comp_R. The resulting 3,006-bp fragment was cloned into pXT10 to create pXNTcopZA. This plasmid was linearized and added to competent cells of strain NTDcopZA, and Tc r recombinants were isolated and verified to yield strain NTDcopZA/c.
The NTDcopA deletion strain was constructed via gene replacement of copA (HIAG_01772) through the splicing of two fragments: the 59 flanking product, amplified using primers 5CuATPase1 and 3HI0290_1pG from NT127 cells as template, and a product containing the Gm r cassette and 39 flanking region, amplified using 5pGent1 and 3CuATPase2 from NTDcopZA cells as template. The intact replacement construct, amplified using primers 5CuATPase1 and 3CuATPase2, was transformed into NTX. Following selection on sBHI-Gm plates, isolates were validated by PCR to yield NTDcopA, which was then transformed with linearized pXT10 to create NTDcopA/v. To complement the copA deletion, two fragments were amplified from the pXNTcopZA plasmid template (a 1,168-bp fragment PCR amplified using primers x-xylF and HI0290_comp_R and a 2,180-bp fragment generated using primers HI0290_comp_F and CuATP_comp_R) and spliced by OE-PCR using primer x-xylF and CuATP_comp_R to generate a 3,327-bp product. From this intermediate, primers CuATP_comp_F and CuATP_comp_R were used to amplify a 2,239-bp fragment, which was cloned into pXT10 to create pXNTcopA. This plasmid was linearized and transformed into NTDcopA to generate strain NTDcopA/c.
To delete the three chaperone genes, a replacement construct was created from two fragments: a 1,459-bp fragment, amplified using primers 5CuATPase1 and 3gent2_290tail from NTDcopZA cells as template, and a 987-bp fragment, amplified with primers HI0290_comp_F and q3ATPase from NT127 cells as template. These fragments were joined by OE-PCR using primers 5CuATPase1 and q3ATPase, and the resulting 2,422-bp amplicon was transformed into competent NTX cells. Gm r transformants for strain NTDcopZ were verified by PCR. Strain NTDcopZ was transformed with pXT10 and selected for Tc r , and isolates were validated by PCR to yield NTDcopZ/v. To complement the DcopZ mutation, a single copZ gene was amplified using primers CuATP_comp_F and Chap_comp_R, as attempts to clone all three chaperones alone were unsuccessful. This 302-bp amplicon was cloned into pXT10 to create pXNTcopZ. Linearized pXNTcopZ was transformed into NTDcopZ to generate NTDcopZ/c. The NTDcueR mutant strain was constructed by replacement of the cueR (HIAG_01035) coding sequence with the Gm r cassette. Three segments were fused by OE-PCR: a 1,290-bp 59 fragment amplified using primers 5cueR1 and 3cueR1pG from NT127 cells as template, the 762-bp Gm r cassette amplified with 5pGent1 and 3Gent2, and the 672-bp 39 fragment amplified using 5cueR2pG and 3cueR2 from NT127 template. The resulting 2,683-bp product was transformed into NTX and NT/v to yield NTDcueR and NTDcueR/v, respectively. For complementation of the cueR mutation, a fragment consisting of the upstream elements and the cueR coding sequence was amplified using primers cueR_comp_F and cueR_comp_R. The resulting 486-bp product was digested and cloned into pXT10 to create pXNTcueR, which was transformed in NTDcueR to yield NTDcueR/c.
Construction of cop promoter LacZ reporter. To monitor cueR-dependent regulation, the copZ promoter/operator region (between cueR and copZ1) was fused to the translational start of lacZ and cloned into the xylose locus. The reporter was constructed from the following: an 867-bp fragment amplified with primers x-xylF and 3xylFcop from pXT10 template, a 3,150-bp fragment amplified using 5copLacZ and lacZtrcKan from pXKZ template, and a 2,415-bp fragment amplified using primers IFTrcF and x-xylB2 from pXTtrcKan template. The 76-bp intergenic region between the divergently oriented cueR and copZ1 genes was created via the overlapping primer tails 3xylFcop and 5copLacZ. The amplicons were spliced together via OE-PCR using primer XT10thyA-F and x-xylB2 to create a 6,404-bp final product. This product was transformed into strains NTDcueR/v and NT/v and selected for Km-resistant colonies to create strains NTDcueRPcopLacZ and NTPcop-LacZ, respectively. The copZ promoter region of the reporter fusion was verified in the resultant strains by sequencing in both directions using primers pXGPseqF and lac7.
Sequencing of chaperone repeats. H. influenzae strains RdAW and NT127 were previously sequenced as whole-genome shotgun sequencing projects with GenBank accession numbers NZ_ACSM00000000 and NZ_ACSL00000000, respectively. For gap closure of shotgun contigs containing the copper efflux operon, PCR with primers CuChapSeq_F and CuChapSeq_R was used in amplification and sequencing (Fig. S2). The length of the operon between cueR and copA in additional NTHi strains was surveyed by PCR. Cells from single colonies or genomic DNA were used as template in PCRs using primers q5MerR and 5merR1 (Fig. S3). Samples were analyzed by agarose gel electrophoresis and imaged using a Gel Logic 200 system (Kodak) followed by staining with ethidium bromide.
Metal sensitivity and growth curve analysis. Standing overnight cultures were inoculated by a 1/ 20 transfer into 5 mL of sBHI in culture tubes and incubated with shaking at 250 rpm and 35°C until mid-log phase. These cultures were adjusted to 0.09 optical density (OD)/mL, and 25 mL was inoculated into 200 mL sBHI medium in wells of a 96-well microtiter plate (Corning) to a final OD/mL of 0.01. For growth assays, sBHI was supplemented with metals (CuSO 4 , MnCl 2 , CoCl 2 , and ZnSO 4 ) at final concentrations of 0.032, 0.064, 0.128, 0.256, 0.512, and 1.024 mM or not supplemented. Microtiter plates were incubated at 35°C for 16 h in a VersaMax microplate reader (Molecular Devices, Sunnyvale, CA), and absorbance was recorded at 600 nm every 6 min. Growth curves were analyzed in Microsoft Excel and GraphPad Prism software.
Reporter expression analysis. Strains were grown in duplicate or triplicate biological replicates in 5 mL sBHI with or without the CuSO 4 (250 mM final concentration) in culture tubes following a 1/10 transfer of standing overnight culture into sBHI at 35°C with shaking at 250 rpm. Absorbances of 200-mL culture samples were read at 600 nm in a microtiter plate reader. The b-galactosidase assay was performed as reported (109) with slight modification. To wells of a polypropylene 96-well DeepWell plate (Nunc 260251) containing 750 mL Z buffer, 15 mL 0.1% SDS, and 30 mL chloroform, 200 mL culture was added, mixed vigorously, and allowed to phase separate. Next, 100 mL of the permeabilized cells was added to a microtiter plate containing 20 mL o-nitrophenyl-b-D-galactopyranoside (ONPG; 4 mg/mL) and mixed. Absorbance was read at 420 nm in a kinetic assay every min for 30 min.
Murine lung infection model. Standing overnight cultures were used to inoculate 25 mL of sBHI in a 50-mL flask to a final optical density at 600 nm (OD 600 ) of 0.01. The resulting cultures were incubated with shaking at 250 rpm and 35°C to mid-log phase. Experimental strains were mixed with the NTlacZ reference strain at a 1:1 ratio, washed, and diluted in Hanks' balanced salt solution to a final concentration of 5 Â 10 8 CFU/mL. Then, 40 mL of bacteria (2 Â 10 7 CFU total) was inoculated into the nares of female 6-week-old C57BL/6 mice (Charles River Laboratories, Boston, MA) anesthetized with ketamine (65 mg/kg of body weight) and xylazine (6.5 mg/kg) by intraperitoneal injection. At 24 h of infection, lungs were harvested, homogenized, and plated on sBHI agar plates with 1 mM D-xylose and X-Gal (5-bromo-4chloro-3-indolyl b-D-galactopyranoside; Sigma-Aldrich, St. Louis, MO) for CFU enumeration. Ratios of CFU of the experimental strains (white colonies, LacZ-) to competitor strain (blue colonies, LacZ1) were reported as the competitive index. All animal procedures were conducted in accordance with NIH guidelines and with prior approval by the University of Massachusetts Medical School Institutional Animal Care and Use Committee.
Copper accumulation. Overnight cultures were used to seed fresh 10-mL cultures and were grown to mid-log phase. Cells were harvested and adjusted to 1 OD/mL, and 250 mL was used to inoculate 10 mL of sBHI in 50-mL DeLong flasks (0.025 OD/mL final). Cultures were grown for 45 min and then supplemented with 10.5 mM or 21 mM CuSO 4 (diluted in sBHI) for a final concentration of 250 mM or 500 mM CuSO 4 , respectively. Cultures were grown at 35°C with shaking at 250 rpm for 45 min or 2 h. Cells were harvested by centrifugation and washed twice with 5 mL of 150 mM NaCl and 10 mM HEPES, pH 7.5. Cell pellets were digested with 0.5 mL of NO3H (trace metal grade) for 1 h at 80°C and further incubated overnight at room temperature. Digestions were terminated by the addition of 0.1 mL of 30% H 2 O 2 . Samples were diluted 1:5 with water, and copper content was measured by furnace atomic absorption spectroscopy (AAS) (Varian SpectrAA 880/GTA 100, Santa Clara, CA).

SUPPLEMENTAL MATERIAL
Supplemental material is available online only. SUPPLEMENTAL FILE 1, PDF file, 1.2 MB.
ACKNOWLEDGMENT This work was supported by the National Institutes of Health/National Institute of Allergy and Infectious Diseases (NIH/NIAID) R01AI095740 (to B.J.A.).