Copper Intoxication in Group B Streptococcus Triggers Transcriptional Activation of the cop Operon That Contributes to Enhanced Virulence during Acute Infection

ABSTRACT

Bacteria can utilize copper (Cu) as a trace element to support cellular processes; however, excess Cu can intoxicate bacteria. Here, we characterize the cop operon in group B streptococcus (GBS) and establish its role in evasion of Cu intoxication and the response to Cu stress on virulence. Growth of a GBS mutant deficient in the copA Cu exporter was severely compromised under Cu stress conditions. GBS survival of Cu stress reflected a mechanism of CopY derepression of the CopA efflux system. However, neither mutant was attenuated for intracellular survival in macrophages. Analysis of global transcriptional responses to Cu by RNA sequencing (RNA-seq) revealed a stress signature encompassing homeostasis of multiple metals. Genes induced by Cu stress included putative metal transporters for manganese import, whereas a system for iron export was repressed. In addition, copA promoted the ability of GBS to colonize the blood, liver, and spleen of mice following disseminated infection. Together, these findings show that GBS copA mediates resistance to Cu intoxication via regulation by the Cu-sensing transcriptional repressor copY. Cu stress responses in GBS reflect a transcriptional signature that heightens virulence and represents an important part of the bacterium’s ability to survive in different environments.

IMPORTANCE Understanding how bacteria manage cellular levels of metal ions, such as copper, helps to explain how microbial cells can survive in different stressful environments. We show the opportunistic pathogen group B streptococcus (GBS) achieve homeostasis of intracellular copper through the activities of the genes that comprise the cop operon, and we describe how this helps GBS survive in stressful environments, including in the mammalian host during systemic disseminated infection.

INTRODUCTION

Copper (Cu) is the most reactive of the biologically relevant first-row d-block transition metals. It can readily displace other metals, including zinc (Zn), nickel (Ni), cobalt (Co), iron (Fe), and manganese (Mn) from metalloproteins in which these are bound (1, 2). In cells across all kingdoms of life, Cu-dependent enzymes are pivotal to many essential processes that underpin physiologic biochemical reactions, owing the ability of copper to redox cycle between Cu(I) and Cu(II) states (3). When present in excess, however, Cu is hazardous to cellular processes and macromolecules, due to the effects of localized free-radical damage (4). A key role for Cu in bacterial cell biology encompasses the host-pathogen interface, where a balance between Cu usage and avoidance of Cu intoxication resulting from host antimicrobial responses must be attained for bacteria to survive, as reviewed elsewhere (5). At this interface, human immune cells can mobilize Cu in a defense response to infection that culminates in exposure of intracellular bacteria to antimicrobial levels of Cu that kill the invading pathogen (6–9). Thus, bacterial responses to Cu stress and mechanisms of resistance to metal ion intoxication have emerged as important facets of bacterial disease pathogenesis (10).

There are several mechanisms that enable bacteria to tolerate excess extracellular Cu, and these have been characterized in several pathogenic species. These mechanisms, which represent survival strategies against Cu intoxication (10) typically involve Cu-transporting P-type ATPases, exemplified by CopA, which detoxify Cu by exporting it from the bacterial cytosol (11–13). This mechanism has been described in Gram-negative and Gram-positive bacteria, and among Streptococcaceae species (14–16), including pneumococci and Streptococcus pyogenes (17).

An opportunistic streptococcal pathogen of humans and animals for which resistance to excess extracellular Cu has not been described is Streptococcus agalactiae, also known as group B streptococcus (GBS). This organism, unlike other Streptococcaceae species, is associated with a comparatively broad host range that encompasses humans, cattle, and fish (18). In humans, GBS is a major cause of invasive infection in infants aged <3 months (19), and it causes a more diverse range of disease etiologies than other streptococci, including meningitis, pneumonia, skin and soft tissue infections, sepsis, arthritis, osteomyelitis, urinary tract infection, and endocarditis (19). GBS has several virulence factors that enable survival in cytotoxic environments, e.g., under acid stress, oxidative stress, and during host colonization, which are reviewed elsewhere (20). Cellular tolerance to Cu stress and the mechanisms of responding to and surviving Cu intoxication have not been defined in GBS.

In this study, we characterize a system that enables Cu efflux in GBS, encompassing a copA homologue, which we show mediates control of Cu efflux in the bacteria. We show that this system affects survival, growth, and virulence of GBS in the mammalian host.

RESULTS

Transcriptomic profiling of GBS responses to extracellular Cu.

To dissect the global response of GBS to extracellular Cu stress, we performed RNA sequencing (RNA-Seq) to define the complete primary transcriptional response to Cu. RNA-Seq analysis of mid-log-phase cells of GBS strain 874391 grown in the presence of 0.5 mM Cu (a subinhibitory level; see below) for 2.5 h (compared to control cultures that were grown without supplemental Cu) revealed a surprisingly modest Cu-responsive transcriptome. The response comprised 18 transcriptional responses, defined as significant based on −2 ≤ fold change (FC) ≤ 2 (false-discovery rate [FDR] < 0.05; n = 4 biological replicates), of 11 upregulated transcripts and 7 downregulated transcripts (Fig. 1 and Table 1). The most significantly upregulated were homologues of the copY-copA-copZ gene cluster (3.6- to 4.0-fold), which encode the putative CopY (Cu-binding transcriptional repressor), CopA (P-type ATPase Cu efflux system), and CopZ (Cu-chaperone) proteins. To validate the expression levels of selected targets identified by RNA-Seq, we targeted pcl1, hvgA, and copY in reverse transcription-quantitative PCR (qRT-PCR) assays, confirming nearly identical fold change values to those found by RNA-Seq analyses (Fig. 1B). The products of copY, copA, and copZ in GBS (Fig. 2A) are moderately conserved compared to those in other Lactobacillales taxa, including Streptococcus mutans, Streptococcus thermophilus, Streptococcus pyogenes, Streptococcus pneumoniae, and Enterococcus spp. (Fig. 2B, ranked by CopA identity), for which a homologous Cu management system is defined (9, 13–15). Together, these findings provide transcriptional and comparative insights that support a proposed model of Cu efflux in GBS (Fig. 2C).

FIG 1.

Global transcriptomic analysis of group B streptococci (GBS) in response to Cu stress. (A) Volcano plot showing data from RNA sequencing (RNA-Seq) analysis of wild-type (WT) GBS cultures exposed to 0.5 mM Cu compared to data from nonexposed controls. Transcripts detected as upregulated or downregulated in response to Cu (n = 4; −2 ≤ fold change (FC) ≤ 2; false-discovery rate [FDR] < 0.05) are highlighted in red or blue, respectively. Dotted lines show false-discovery rate (q value) and fold change cutoffs. Gray points indicate genes that were not significantly changed in expression, according to these analysis cutoffs. Selected genes are identified with black lines. (B) Validation of RNA-Seq data. Expression ratio (fold change) of pcl1, hvgA, and copY quantified by reverse transcription-quantitative PCR (qRT-PCR) in Todd-Hewitt broth (THB) medium containing 0.5 mM Cu compared to THB without Cu. Ratios for qRT-PCR were normalized using housekeeping dnaN, and RNA-Seq ratios were calculated using DESeq2. Bars show means and standard error of the mean (SEM) from 4 independent experiments.

TABLE 1.

Transcriptional signature of Cu intoxication in GBS

Locus taga	GenBank accession no.	Label	Annotation (NBCI)	FCb	FDRc
02565	ASZ00809.1	copY	CopY/TcrY family copper transport repressor	4.1	1.09E−12
02575	ASZ00811.1	copZ	Carbonate dehydratase	4.0	1.02E−37
02570	ASZ00810.1	copA	Copper-translocating P-type ATPase	3.7	2.25E−35
10970	ASZ02392.1		Peptidoglycan-binding protein LysM	3.4	6.13E−13
07965	ASZ01821.1	mtsA	Metal ABC transporter substrate-binding protein	2.9	2.15E−12
07960	ASZ01820.1	mtsB	Metal ABC transporter ATP-binding protein	2.6	1.02E−15
10190	ASZ02237.1	mntH2	Divalent metal cation transporter	2.4	6.34E−09
10965	ASZ02391.1		Transglycosylase	2.2	1.87E−06
09125	ASZ02032.1		CHAP domain-containing protein	2.1	1.58E−09
07955	ASZ01819.1	mtsC	Metal ABC transporter permease	2.1	2.62E−10
10305	ASZ02261.1		PAP2 family protein	2.0	6.13E−13
08045	ASZ01834.1		Glycosyltransferase family 2 protein	−2.0	9.51E−06
08200	ASZ01861.1		Branched-chain amino acid ABC transporter substrate-binding protein	−2.3	1.98E−04
10430	ASZ02284.1	hvgA	Pathogenicity protein	−2.3	2.18E−07
04720	ASZ01199.1	pcl1	Membrane protein	−4.3	1.38E−09
04185	ASZ01095.1	fetB	Iron export ABC transporter permease subunit, FetB	−4.9	1.39E−08
04180	ASZ01094.1	fetA	ABC transporter ATP-binding protein FetA	−4.9	4.58E−11

^aDenotes locus tag of S. agalactiae 874391; number preceded by CHF17_RS.

^bFC, fold change.

^cFDR, false discovery rate,

FIG 2.

Organization of the copY-copA-copZ locus in GBS. (A) copY-copA-copZ are adjacent in the GBS genome and are likely controlled by the promoter-proximal gene copY, encoding a putative Cu-sensing repressor. Locus tags from the GBS 874391 genome are indicated. (B) Distribution of homologous cop genes in other Lactobacillales species, arranged by percent identity of amino acid sequence to CopA of Streptococcus agalactiae. (C) Model of Cu efflux in GBS, highlighting CopA as a transmembrane Cu exporter transcriptionally repressed by CopY in the absence of Cu and likely inhibited by the Cu-binding chaperone protein CopZ. Figure based on previous studies in other Lactobacillales species (37, 62).

In addition to copYAZ, GBS upregulated putative metal transporters for Mn import (mtsABC and mntH2), and downregulated an iron (Fe) export system (fetAB) (Fig. 1 and Table 1). In addition, we detected significant downregulation of pcl1, which encodes a putative membrane protein, and of the gene for hypervirulence-associated factor, hvgA.

Roles of copA and copY in Cu resistance in GBS.

To functionally characterize two major elements of the Cu-responsive transcriptome in GBS, we targeted copA and copY by generating isogenic deletion mutants and comparing growth of these with that of the wild type (WT). In a nutrient-rich medium (Todd-Hewitt broth [THB]) supplemented with increasing amounts of Cu, we observed no growth-inhibitory effects of high Cu (up to 1.5 mM) for WT GBS; the lag phases, growth rates, and final biomass yields were equivalent for the WT at different levels of Cu (Fig. 3). In contrast, the growth of ΔcopA GBS was severely attenuated by Cu stress at concentrations of ≥1 mM; complementation of the copA mutation restored the phenotype to the WT (Fig. 3). Deletion of copY, which encodes a putative Cu-dependent repressor of copA, had no effect on the growth rate or lag phase in THB supplemented with Cu but significantly affected final biomass yield of cultures (attenuance [D] at 600 nm at 18 h) (see Fig. S1 in the supplemental material). Together, these results show that copA confers cellular resistance to Cu stress in GBS in a manner that supports bacterial growth under nutritive conditions.

FIG 3.

Growth curve analyses of GBS cultured in nutrient-rich THB medium (A), or in THB supplemented with 0.5 mM Cu (B), 1.0 mM Cu (C), or 1.5 mM Cu (D), comparing WT, ΔcopA, or ΔcopY strains. Complementation of copA mutation (ΔcopA+C) restored growth of GBS to WT levels under conditions of Cu stress. Points and bars show mean and SEM from several independent experiments (5 for WT, ΔcopA, and ΔcopY; 2 for complemented strain) monitoring attenuance (D) at 600 nm.

Temporal and concentration-dependent bactericidal effects of Cu toward GBS.

We examined the effects of Cu on survival of GBS under nutrient-limited conditions by using a minimal chemically defined medium (CDM), in consideration of prior studies of antibacterial activity in minimal medium (21, 22), the influence of culture medium (23, 24), and possible cell-protective effects from buffering agents, e.g., glutathione (15). We observed that the level of Cu required to inhibit GBS growth in CDM was markedly lower than that in THB (see Fig. S2 in the supplemental material); for example, the growth of ΔcopA GBS was significantly attenuated in CDM in the presence of 0.5 mM supplemental Cu, a level that caused only slight inhibition of growth in THB (Fig. 3). To more precisely define the bactericidal effects of Cu toward GBS in CDM, we performed time-kill curves using 1 × 10⁶ to 5 × 10⁶ CFU/ml exposed to Cu concentrations ranging between 50 μM and 1 mM. We observed potent killing effects that depended on both Cu concentration and time, with ≥0.5 mM Cu significantly killing WT GBS after 3 h of exposure (Fig. 4, WT); after 24 h, there was an ∼230-fold reduction in CFU/ml at 0.5 mM Cu (0 mM Cu = 5.6 log₁₀ CFU/ml; 0.5 mM CFU/ml = 3.2 log₁₀ CFU/ml), and no viable GBS remained in cultures exposed to 1 mM Cu. The ΔcopA strain was significantly more susceptible than the WT to Cu-induced toxicity; Cu levels above ≥0.5 mM significantly reduced the number of viable ΔcopA GBS, beginning as early as 1 h postexposure (Fig. 4, ΔcopA); the degree of bacterial killing was significant at the 24-h time point for ≥0.5 mM Cu (Fig. 4). Interestingly, at the 6-h time point, Cu at ≥0.5 mM did not affect viability of ΔcopY GBS compared to that of the WT strain (Fig. 4, ΔcopY), a hyperresistance phenotype that was not apparent at the 24-h time point. Together, these findings are consistent with a role for CopA in resisting Cu stress and with CopY as a Cu-dependent repressor of copA.

FIG 4.

Bactericidal effect of Cu on GBS viability. Time-kill assays comparing WT, ΔcopA, or ΔcopY GBS (and the complemented strain ΔcopA+C) incubated in CDM or in CDM supplemented with 0.05, 0.1, 0.2, 0.5, or 1 mM Cu. Viable cells were quantified at 1 h, 3 h, 6 h, and 24 h postincubation. ^†, viable cell counts of 0 CFU/ml were assigned a value of 1 to plot on log₁₀ y axes. Points and bars show mean and SEM from 4 independent experiments. Data were analyzed by one-way analysis of variance (ANOVA) with Holm-Sidak multiple comparisons. Significance markers represent comparisons of the number of surviving bacteria under conditions of the defined Cu concentration (Cu) versus the number of bacteria in the nonexposed control (0 mM Cu) for the same time point and same strain (*, P < 0.05).

Regulation of Cu efflux by CopY and accumulation of metals during Cu stress.

We next sought to examine the control of Cu responses in GBS at the transcriptional level. The capacity of Cu stress to induce expression of copA for Cu export was examined by analyzing copA expression by qRT-PCR in GBS exposed to Cu concentrations ranging from 0.25 to 1.5 mM Cu in THB. GBS significantly upregulated copA in response to Cu (3.7-fold to 14.2-fold) with fold change values increasing as Cu concentration increased (Fig. 5A). To ascertain the role of CopY as a putative Cu-responsive repressor of copA expression in GBS, we quantified copA mRNA transcripts in the ΔcopY mutant exposed to 0.5 mM Cu. This level of Cu was carefully chosen as subinhibitory to enable comparisons independent of metabolic state and therefore limiting any potential bias from possible discordant Cu stress between the WT and mutants with varied resistance phenotypes. Deletion of copY resulted in severe deregulation of copA expression, causing an ∼207-fold ± 45-fold increase of copA transcripts in the copY background (Fig. 5B). Thus, these data establish that GBS copY represses copA in the absence of Cu and are consistent with a model of CopY-mediated regulation of copA that is responsive to intracellular accumulation of Cu.

FIG 5.

Expression analysis of copA and intracellular metal content in GBS strains. (A) Expression ratio (fold change) of copA quantified by qRT-PCR in THB medium containing 0.25, 0.5, 1.0, or 1.5 mM Cu, compared to THB without Cu. (B) Relative copA transcripts were quantified in WT and ΔcopY strains with and without Cu supplementation (0.5 mM) to demonstrate deregulation of copA expression in the ΔcopY background. (C to F) Intracellular accumulation of Cu (C), Mn (D), Fe (E), and Zn (F) was compared with and without Cu supplementation (0.5 mM) in WT, ΔcopY, and ΔcopA strains. Ratios in panel A were calculated as described previously (54) using threshold cycle (C_T) values, primer efficiencies, and housekeeping dnaN. Bars show means and SEM from 3 or 4 independent experiments, compared by one-way ANOVA with Holm-Sidak multiple comparisons (*, P < 0.05; ***, P < 0.001). Ctrl, control condition (i.e., no supplemental Cu in medium). The ANOVA P value from comparisons of between strains was 0.06 for cellular Fe and 0.01 for cellular Zn; subsequent pairwise Student’s t tests were used to compare each strain under the Ctrl and Cu conditions (*, P < 0.05).

To examine the impact of copA and copY mutations on accumulation of Cu within the cell, we used the equivalent conditions to those of the RNA-Seq and quantitative PCR (qPCR) assays, exposing GBS to 0.5 mM Cu in THB and measuring the total Cu content of cells compared to that of nonexposed controls (Ctrl). Inductively coupled plasma optical emission spectroscopy (ICP-OES) demonstrated that standard THB contained 0.2 ± 0.08 μM Cu, reflecting trace amounts in the medium. In the absence of supplemental Cu, WT GBS limited intracellular Cu content such that only 0.8 ± 0.1 μg Cu · g dry weight⁻¹ were detected in cultures grown in THB. However, exposure of WT GBS to 0.5 mM Cu resulted in a significant increase in intracellular Cu to 31.9 ± 3.1 μg Cu · g dry weight⁻¹ (Fig. 5C). Strikingly, ΔcopY GBS exhibited significantly less cellular Cu upon exposure to Cu (6.7 ± 0.4 μg Cu · g dry weight⁻¹), consistent with the observation that transcription of the Cu exporter CopA is significantly elevated in this mutant. In addition, we noted significant accumulation of cellular Cu in the ΔcopA strain (109.6 ± 1.9 μg Cu · g dry weight⁻¹), confirming a Cu efflux role for copA. Interestingly, we noted modest but significant increases due to Cu stress in Mn and Fe in the ΔcopA strain and in Zn in the ΔcopY strain (Fig. 5D to F).

To extend our findings comparing mutant strains to the WT at a level of Cu that is subinhibitory to the ΔcopA strain, we next undertook experiments to investigate the impact of severe Cu stress on WT GBS by repeating metal accumulation analyses using cells grown at 1.5 mM Cu. We detected modest but statistically significant accumulation of Fe inside GBS cells exposed to 1.5 mM Cu, cooccurring with significantly higher Cu levels; however, there was no difference in Mn or Zn content (see Fig. S3 in the supplemental material). Fe status has been reported to influence bacterial resistance to peroxide stress (25), and therefore we also undertook experiments to investigate the consequence of Cu toxicity on resistance to oxidative stress using assays of susceptibility to hydrogen peroxide (H₂O₂). These experiments demonstrated a significant reduction in the viability of ΔcopA GBS exposed to H₂O₂, which was dependent on preexposure to Cu (see Fig. S4 in the supplemental material). Thus, GBS cells undergoing Cu stress are rendered significantly more susceptible to oxidative stress.

Role of CopA in macrophage killing of GBS.

To examine whether the CopA Cu efflux system in GBS supports survival of the bacteria in phagocytes, we performed antibiotic protection assays with human monocyte-derived macrophage-like cells. Macrophages were infected with the WT, ΔcopA GBS, or its complemented strain for 1 h, and antibiotics were added to kill extracellular bacteria. Viable intracellular GBS was quantified at 1 h post-antibiotic addition and at 24 h and 48 h to assess intracellular survival. We performed assays comparing the culture conditions of standard RPMI medium to those of medium that was supplemented with 20 μM Cu to ensure adequate availability of Cu for host cellular responses. These assays demonstrated significant reductions in the numbers of viable GBS bacteria over the time course (24 to 48 h); however, we did not detect any significant differences between the numbers of WT and ΔcopA GBS bacteria recovered from the host cells at any time point (Fig. 6). The numbers of bacteria recovered from WT and ΔcopA mutant were equivalent regardless of the presence of supplemental Cu in the medium. Thus, under these conditions, copA does not contribute to the intracellular survival of GBS in human macrophage-like cells.

FIG 6.

Interactions of GBS with human macrophage-like cells. Gentamicin protection assays with WT and ΔcopA strains in human (U937 monocyte-derived macrophage-like) macrophages in medium with and without 20 μM supplemental Cu. Surviving bacteria are expressed as CFU per milliliter, indicating the numbers of intracellular bacteria at 1 h after antibiotic treatment and at 24 h and 48 h postinfection (h.p.i.). Data are means and SEM of 3 or 4 independent experiments.

GBS CopA contributes to virulence in vivo.

To examine the contribution of Cu efflux to GBS virulence, we used a murine model of disseminated infection (26). In mice challenged with 10⁷ GBS, we detected significantly fewer ΔcopA mutant in the liver (median of 3.5 versus 4.2 log₁₀ CFU · g tissue⁻¹; P = 0.005), spleen (median of 3.8 versus 4.1 log₁₀ CFU · g tissue⁻¹; P = 0.013), and blood (median of 1.4 versus 1.9 log₁₀ CFU · g tissue⁻¹; P = 0.044) compared to the WT at 24 h postinoculation (Fig. 7). However, no significant differences were observed between the counts of the WT and ΔcopA mutant in several other tissues, including the brain, heart, lungs, and kidneys (see Fig. S5 in the supplemental material). These data support a modest but significant role for cellular management of Cu via copA in GBS in supporting disseminated infection in vivo.

FIG 7.

Virulence of WT or ΔcopA GBS in a mouse model of disseminated infection. C57BL/6 mice (6 to 8 weeks old) were intravenously injected with 10⁷ bacteria; bacteremia and disseminated spread to liver and spleen were monitored at 24 h postinfection. CFU were enumerated and counts were normalized using tissue mass in grams. Viable cell counts of 0 CFU/ml were assigned a value of 1 to enable visualization on log₁₀ y axes. Lines and bars show median and interquartile ranges, and data are pooled from 2 independent experiments, each containing n = 10 mice and compared using Mann-Whitney U tests (*, P < 0.05; **, P < 0.01).

DISCUSSION

Transcriptional and cellular responses of bacterial pathogens to metal ions, including Cu, can influence host-pathogen interactions and thereby play a role in disease pathogenesis (4). The roles of Cu homeostasis and detoxification in the biology of GBS have not hitherto been characterized. The principal finding of this study is the establishment of a transcriptional and cellular connection between the response to Cu stress in GBS and survival of the bacteria under conditions of Cu toxicity; this connection is mediated through copA and controlled through copY, enabling the organism to resist Cu-mediated killing. Additionally, this study establishes a connection between Cu stress resistance in GBS and virulence in the host during systemic, disseminated infection. The new insights into gene function in GBS, viewed through the lens of the Cu stress transcriptome, combined with the findings of enhanced virulence, elucidate molecular mechanisms that underpin GBS survival of intoxicating conditions, including those likely to be encountered in the host.

The transcriptional remodeling in GBS that occurs in response to Cu stress, as defined in this study at a global level, comprises an intriguingly constrained subset of genes indicating a tightly controlled system of responses to Cu. Interestingly, equivalent low numbers of genes were identified in previous transcriptional analyses of other streptococci exposed to Cu stress (14–16). Our findings are consistent with these prior reports and support the notion that these transcriptional responses function in housekeeping or homeostasis to set a low limit of Cu availability in the cytoplasm (15). In GBS, copA is among the most strongly activated genes in the transcriptional response to Cu stress, and the mutational analysis performed in this study shows that copA is crucial for the bacterium to attain an essential Cu efflux response during Cu stress. However, copA is only one of an assembly of genes engaged by GBS during Cu stress, and it is likely that other genes in the transcriptome contribute to Cu detoxification via additional means. For example, we detected upregulation of putative metal transporters for Mn (mtsABC and mntH2), along with concurrent downregulation of a system that encodes Fe transport machinery (fetAB). These transcriptional insights are interesting because they hint at additional stress response mechanisms that occur during Cu stress in GBS, which extend beyond CopA and need elucidation. Metal content analysis by ICP-OES showed that high Cu stress disrupts additional pools of intracellular trace metals, including, for example, causing modest elevations in levels of Fe along with major increases in Cu content. The consequences of these alterations in metal content in GBS for cell activity need further examination.

Transcription of metal import and export genes, including those mentioned above, is typically controlled by metal-dependent regulatory proteins termed metalloregulators that sense metal ion bioavailability and work to maintain cellular metal homeostasis (27). In our study, we detected no major changes in the expression of genes for metalloregulators (25, 27), such as for Mn (mntR and mtsR), Fe (fur), and Zn (adcR and sczA) transport and for the sensing of peroxide (perR), in GBS undergoing Cu stress. It would be of interest to elucidate whether such regulators undergo mismetalation in GBS during Cu stress, since it is conceivable that Cu may displace Mn and Fe from proteins, as described in other studies (28).

In S. pyogenes, Mn uptake is facilitated by mtsABC, is protective against peroxide-induced stress (25), and is controlled by the MtsR regulator (25, 29). Mn and Fe transport are coordinated in order to control metalation of superoxide dismutase (25), which can use Mn or Fe at its catalytic site in streptococci (30). The consequences of Cu intoxication relating to oxidative stress were explored in this study using killing assays with hydrogen peroxide. These revealed reduced viability in copA-deficient GBS, but only in cells that had been preexposed to Cu. Taken with our observations that Cu accumulates significantly in this strain following Cu (and, to a lesser extent, Mn and Fe) intoxication, these data hint at a role for oxidative damage as a killing mechanism. Fe can produce hydroxyl radicals, and Cu(II) can be oxidized to Cu(I) in reactions with peroxide; both damage bacterial cells. In GBS, mntH encodes a dual system for the import of Mn and Fe and is regulated by pH (31). In some strains of GBS, including the hypervirulent sequence type 17 (ST17) lineage used in this study, two homologues of MntH exist, encoded by mntH and mntH2. This study demonstrates upregulation of mntH2, but not of mntH, in response to Cu; however, the role of mntH2 and that of its regulation are undefined in GBS. In Bacillus subtilis, MntR coordinates Mn import (and efflux) by control of genes homologous to GBS mntH and mntH2 that encode the import of Mn (32). This may provide context for elevated mntH2 expression in the presence of modest Cu stress in our study, since Cu-bound MntR has a much lower affinity for binding its target operator sequence (33). We also note that mtsABC and mntH2 are upregulated in concert with downregulation of Fe-transporting fetAB in response to Zn stress in GBS (34). Together, our data indicate that modulation of mtsABC, mntH2, and fetAB expression forms parts of a transcriptional signature of GBS exposed to Cu or Zn stress.

Our assays of bacterial growth in vitro in conditions of Cu stress revealed striking differences in growth phenotypes of GBS between nutritionally rich (THB) and limited (CDM) media, possibly reflecting relative quantities of compounds that confer a protective advantage for survival during Cu stress, such as glutathione (15) or other thiol-containing amino acids that may interact with free Cu ions in solution. Glutathione is not included in CDM as a separate chemical constituent, but the quantities of methionine, cystine, and cysteine in our preparations of CDM are 30, 62.6, and 50 mg · liter⁻¹, respectively. Importantly, Cu exposure assays in vitro are almost certainly influenced by the compounds present in the medium that likely affect levels of Cu that become inhibitory (15). For example, Cu-buffering effects likely occur in THB, which would influence the toxicity of Cu toward the bacteria. Our rationale for using THB to analyze Cu toxicity in GBS is that this medium is a standard rich growth medium for GBS and is widely used for studies of this organism. The challenges inherent in using such a rich medium, including, for example, Cu buffering effects, led us to also examine the impact of Cu toxicity on GBS in a defined, nutrient-limited medium (CDM), which revealed that lower ranges of concentrations were required to initiate Cu intoxication. Some of the Cu concentrations used in our study may be considered supraphysiological (1 to 1.5 mM), but the magnitude of these concentrations is likely much greater than the magnitude of “free” Cu ions (not bound by other molecules) that would be present and available to react biologically. The concentration ranges used were informed by available literature along with measures of expression of copA, since we and others have shown that the transcriptional activity of copA is influenced by cellular Cu. Thus, copA activity can be considered a bona fide reporter of free cellular Cu, or a surrogate marker for exposure to Cu stress. Consistent with this notion, we observed higher fold change in expression of copA (7.6-fold ± 2.3-fold upregulated; data not shown) in CDM with 100 μM Cu than in THB with 500 μM Cu (4.9-fold ± 0.5-fold upregulated) (Fig. 5). The Cu concentrations that are encountered by GBS at infection sites or in different intracellular compartments in host cells are unknown. It is also possible that the levels of Cu utilized in our assays in THB are distinct from those in host niches. Cu is elevated in the host at sites infected with S. pyogenes (15), including in the blood and in infected skin lesions. For mycobacterial phagosomes, investigators have reported Cu concentrations of 25 μM, and in Mycobacterium tuberculosis-infected macrophages, intravacuolar concentrations of 0.4 mM (35). Notwithstanding considerations of the possible physiological ranges of Cu, the Cu exposure assays used here are beneficial to establish bona fide gene function and bacterial responses to Cu intoxication.

Our finding that copY functions to repress copA in the absence of Cu is consistent with previous reports in other bacteria (36, 37). Disrupting the genetic systems for Cu efflux in GBS via mutation of the CopA exporter or the CopY regulator reveals divergent phenotypes that stem from loss of export or regulatory function, resulting in accumulation (ΔcopA) or reduction (ΔcopY) of cellular Cu. These phenotypes will be of interest to dissect in terms of the role of CopY in GBS biology in other models of infection and disease in the future. The results of the current study indicate that GBS might have a higher intrinsic level of resistance to Cu than other Streptococcus spp., including S. pyogenes. However, we did not directly compare different species, and differences in experimental approaches or background solution, as our study demonstrates, might affect these assays. Thus, further work more directly characterizing relative resistance between streptococci would be of interest to the field.

Studies have demonstrated increased Cu levels in the blood of humans during bacterial infection (38, 39); however, most of the insight into Cu-driven antimicrobial responses is derived from studies of mammalian cells infected in vitro. In macrophages, bioavailability of Cu correlates with antibacterial activities (40), and Cu “hot spot” formation mediates antimicrobial responses against intracellular bacteria (7). It would be of interest to examine if Cu release within phagocytic cells colocalizes with intracellular GBS. Our findings based on in vitro macrophage infection showed no attenuation of GBS devoid of CopA for survival in host cells, which was surprising, given the important role of Cu management in survival of other bacteria inside macrophages (6) and epithelial cells (12). Notably, however, intracellular survival of Salmonella deficient in cueO, which encodes an enzyme required for resistance to Cu ions, was not impaired in murine macrophages in a previous study (41), leading the authors to suggest multiple host factors are involved in clearance of the bacterium. In addition, copA-deficient S. pyogenes was not impaired for survival in assays with human neutrophils (15). Our results are consistent with these findings. Other researchers have described the limitations of in vitro tissue culture monolayer assays for determining intracellular survival of bacteria in the context of Cu homeostasis (42).

Systemic infection of mice exposed a connection between the ability of GBS to generate a Cu management response via CopA and bacterial virulence in vivo. Here, copA was essential for GBS to fully colonize and survive in the blood, as well as in other tissues. In demonstrating a significant attenuation of GBS deficient in CopA to be fully virulent in mice, we suggest that Cu toxicity may represent a form of stress experienced by the bacterium in vivo during systemic infection. In other bacteria, including Pseudomonas aeruginosa and Listeria monocytogenes, compromised Cu transport leads to attenuation for colonization in various infection models (42, 43). Attenuation of GBS for colonization of the blood, liver, and spleen indicates that Cu management in the bacterial cell is essential not only for efficient survival of the bacteria in the bloodstream but also for colonization of highly immunologically active tissues, i.e., Kupffer cells and splenic lymphocytes for innate and adaptive immune responses, respectively. It is plausible that Cu stress that might be encountered by GBS in vivo could influence virulence factor function, such as the hypervirulence-associated adhesin HvgA, which we have shown is downregulated under Cu stress. It would be of interest to analyze the effect of Cu transport deficiency in GBS in other relevant models of infection, including in vaginal colonization (44). In addition to defining the effects of metal homeostasis in GBS on the nature of infection and disease caused by this organism, small-molecules probes might hold promise for the identification of other molecular mechanisms of metal homeostasis in GBS, as reported previously for Gram-positive bacteria (45).

In summary, this study shows that management of Cu export in GBS is essential for the bacterium to survive in environments with Cu stress. Cu intoxication generates a transcriptional signature that includes activation of the cop operon to confer bacterial survival and virulence. The exact role for Cu ions as an antibacterial response against GBS warrants further investigation.

MATERIALS AND METHODS

Bacterial strains, plasmids, and growth conditions.

GBS, Escherichia coli strains, and plasmids used are listed in Table 2. GBS was routinely grown in Todd-Hewitt broth (THB) or on Todd-Hewitt agar (1.5% wt/vol). E. coli was grown in lysogeny broth (LB) or on LB agar. Routine retrospective colony counts were performed by plating dilutions of bacteria on tryptone soya agar containing 5% defibrinated horse blood (Thermo Fisher Scientific). Medium were supplemented with antibiotics (spectinomycin [Sp], 100 μg/ml, and chloramphenicol [Cm], 10 μg/ml), as indicated. Growth assays used 200-μl culture volumes in 96-well plates (Greiner) sealed using Breathe-Easy membranes (Sigma-Aldrich) and measured attenuance (D, at 600 nm) using a ClarioStar multimode plate reader (BMG Labtech) in well scan mode using a 3-mm 5 × 5 scan matrix with 5 flashes per scan point and path length correction of 5.88 mm with agitation at 300 rpm and recordings taken every 30 min. The media for growth assays were THB and a modified chemically defined medium (CDM) (46) (with 1 g/liter glucose, 0.11 g/liter pyruvate, and 50 mg/liter l-cysteine), supplemented with Cu (supplied as CuSO₄) as indicated. For attenuance baseline correction, control wells without bacteria were included for Cu in medium alone.

TABLE 2.

Bacterial strains and plasmids

Strain or plasmid	Descriptiona	Source or reference
Strains
E. coli DH5α	huA2 lac(Δ)U169 phoA glnV44 ϕ80' lacZ(Δ)M15 gyrA96 recA1 relA1 endA1 thi-1 hsdR17	Bethesda Research Labs
S. agalactiae 874391	Wild-type, sequence type 17, serotype III strain; vaginal isolate (Japan)	63
S. agalactiae GU2691	874391 ΔcopA (copA mutant); locus tag CHF17_RS02570	This work
S. agalactiae GU2857	874391 ΔcopY (copY mutant); locus tag CHF17_RS02565	This work
S. agalactiae GU3121	GU2691 (ΔcopA) containing copYAZ complement construct pGU3112; Spr	This work
Plasmids
pHY304aad9	ori(Ts); temperature-sensitive shuttle vector; Spr	48
pGU2650	pHY304aad9-derivative ΔcopA construct; Spr	This work
pGU2847	pHY304aad9-derivative ΔcopY construct; Spr	This work
pGU3112	pDL278 containing cloned copYAZ operon; Spr	This work

^aSp^r, spectinomycin resistant; Ts, temperature sensitive.

DNA extraction and genetic modification of GBS.

Plasmid DNA was isolated using miniprep kits (Qiagen), with modifications for GBS as described elsewhere (47). Deletions in copA (CHF17_00507 and CHF17_RS02570) and copY (CHF17_00506 and CHF17_RS02565) were constructed by markerless allelic exchange using pHY304aad9, as described previously (48). Plasmids and primers are listed in Table 2 and in Table S1 in the supplemental material, respectively. Mutants were validated by PCR using primers external to the mutation site and by DNA sequencing.

RNA extraction and qRT-PCR.

For Cu exposure experiments, 1 ml of overnight THB cultures were back diluted 1/100 in 100 ml of THB (prewarmed at 37°C in 250-ml Erlenmeyer flasks) supplemented with 0.25, 0.5, 1.0, or 1.5 mM Cu. Cultures were grown with shaking (200 rpm) at 37°C; after exactly 2.5 h, 10- to 50-ml volumes containing approximately 500 million mid-log-phase bacteria were harvested; RNA was preserved and isolated as described previously (49). RNA quality was analyzed by RNA LabChip using GX Touch (Perkin Elmer). RNA (1,000 ng) was reverse transcribed using SuperScript IV according to the manufacturer’s instructions (Life Technologies), and cDNA was diluted 1:50 in water prior to qPCR. Primers (Table S1) were designed using Primer3 Plus (50, 51) to quantify transcripts using Universal SYBR green Supermix (Bio-Rad) and a QuantStudio 6 Flex (Applied Biosystems) system in accordance with minimum information for publication of quantitative real-time PCR experiments (MIQE) guidelines (52). Standard curves were generated using 5-point serial dilutions of genomic DNA (5-fold) from WT GBS 874391 (53) and used to quantify relative transcript amounts. Expression ratios were calculated using threshold cycle (C_T) values and primer efficiencies described elsewhere (54) and using dnaN, encoding the DNA polymerase III β-subunit, as a housekeeper.

Whole-bacterial-cell metal content determination.

Metal content in cells was determined as described previously (30) with minor modifications. Cultures were prepared essentially as described above (“RNA extraction and qRT-PCR”) with the following modifications; THB medium was supplemented with 0.5 mM CuSO₄ or not supplemented (Ctrl), and following exposure for 2.5 h, bacteria were harvested by centrifugation at 4,122 × g at 4°C. Cell pellets were washed 3 times in phosphate-buffered saline (PBS) plus 5 mM EDTA to remove extracellular metals, followed by 3 washes in PBS. Pelleted cells were dried overnight at 80°C, resuspended in 1 ml of 32.5% nitric acid, and incubated at 95°C for 1 h. The metal ion-containing supernatant was collected by centrifugation (14,000 × g for 30 min) and diluted to a final concentration of 3.25% nitric acid for metal content determination using inductively coupled plasma optical emission spectroscopy (ICP-OES). ICP-OES was carried out on an Agilent 720 ICP-OES instrument with axial torch, OneNeb concentric nebulizer, and Agilent single-pass glass cyclone spray chamber. The power was 1.4 kW with 0.75 liter/min nebulizer gas, 15 liter/min plasma gas, and 1.5 liter/min auxiliary gas flow. Cu was analyzed at 324.75 nm, Zn at 213.85 nm, Fe at 259.94 nm, and Mn at 257.61 nm, with detection limits at <1.1 ppm. The final quantity of each metal was normalized using dry weight biomass of the cell pellet prior to nitric acid digestion, expressed as micrograms per gram weight. Baseline concentrations were determined from at least three independent assays to be 0.2 ± 0.08 μM Cu in THB medium and 40 ± 4 nM Cu in CDM.

Hydrogen peroxide assay.

Peroxide survival assays were based on prior studies (55, 56) with minor modifications to encompass using GBS cells that had been pregrown under conditions of Cu stress. Overnight cultures of WT and mutants were grown in THB, and cultures were back diluted 1/100 into either fresh THB or THB supplemented with 0.5 mM Cu and grown for exactly 2.5 h at 37°C with 200 rpm agitation. Bacteria were harvested by centrifugation at 4,122 × g, washed twice in PBS, and resuspended in assay buffer (0.1 M sodium phosphate buffer [pH 7.5]). This was prepared by combining 41 ml of 0.2 M dibasic sodium phosphate and 9 ml of 0.2 M monobasic sodium phosphate in a volume of 200 ml. GBS cells were diluted in assay buffer alone or in assay buffer containing 5 mM H₂O₂ (catalog no. H1009; Sigma-Aldrich) to a final concentration of ∼5 × 10⁶ CFU/ml. Cell suspensions ± H₂O₂ were then incubated for 1 h at 37°C, and survival was monitored by serial dilution and plating for CFU/ml counts.

RNA sequencing and bioinformatics.

Cultures were prepared as described above (“RNA extraction and qRT-PCR”) to compare mid-log-phase cells grown in THB plus 0.5 mM Cu to those grown in THB without added Cu. RNase-free DNase-treated RNA that passed Bioanalyzer 2100 (Agilent) analysis was used for RNA sequencing (RNA-seq) using the Illumina NextSeq 500 platform. We used a bacterial rRNA depletion kit (Invitrogen) prior to library construction and TruSeq library generation kits (Illumina, San Diego, CA). Library construction consisted of random fragmentation of the RNA and cDNA production using random primers. The ends of the cDNA were repaired and A-tailed, and adaptors were ligated for indexing (with up to 12 different barcodes per lane) during the sequencing runs. The cDNA libraries were quantitated using qPCR in a LightCycler 480 (Roche) with a Kapa Biosystems kit (Woburn, MA) prior to cluster generation. Clusters were generated to yield approximately 725,000 to 825,000 clusters/mm². Cluster density and quality were determined during the run after the first base addition parameters were assessed. We ran single-end 75-bp sequencing runs to align the cDNA sequences to the reference genome. For data preprocessing and bioinformatics, STAR (v2.7.3a) was used (parameters used: “-outReadsUnmapped Fastx -outSAMtype BAM SortedByCoordinate -outSAMattributes All”) to align the raw RNA sequencing fastq reads to the WT S. agalactiae 874391 reference genome (53). HTSeq-count v0.11.1 (parameters used: “-r pos -t exon -i gene_id -a 10 -s no -f bam”), was used to estimate transcript abundances (57). DESeq2 was then used to normalize and test for differential expression and regulation following their vignette. Genes that met certain criteria (i.e., an absolute fold change of ≥2 and a q value [false-discovery rate] of <0.05) were accepted as significantly altered (58).

Mammalian cell culture.

U937 monocytes were differentiated into human monocyte-derived macrophages (MDMs) as follows. Briefly, monocytes were seeded (5 × 10⁵ per well, such that 1 × 10⁵ adhered) into the wells of a 96-well tissue culture-treated plate (Falcon), essentially as described elsewhere (59, 60), except that U937 monocytes were differentiated by exposure to 30 ng/ml phorbol 12-myristate 13-acetate (PMA) for 48 h and cells subsequently rested in medium without PMA for 72 h to enhance the morphological and phenotypic markers of MDMs (61). A multiplicity of infection (MOI) of 100 bacteria:macrophages for 1 h was used in RPMI without antibiotics. Nonadherent bacteria were removed by five 200-μl washes of PBS using a Well Wash Versa (Thermo Scientific). RPMI containing 250 U/ml penicillin-streptomycin (Gibco), and 50 μg/ml gentamicin (Sigma-Aldrich) were used for antibiotic protection assays to kill extracellular bacteria, as described previously, by incubation for 1 h at 37°C in 5% CO₂ (60). Samples were processed after 1 h (time zero), 24 h, or 48 h after infection; monolayers were washed five times with 200 μl PBS and lysed by brief exposure to 50 μl of 0.25% trypsin EDTA (Gibco) and 0.1% Triton X-100 (10 min) prior to dilution with 150 μl PBS and estimation of CFU per milliliter by serial dilution and plate counts on agar. Additional assays that ran in parallel were identical, except that 20 μM Cu was added to RPMI culture medium at all stages post-PMA treatment of U937 cells.

Animals and ethics statement.

Virulence was tested using a mouse model of disseminated infection based on intravenous challenge with 10⁷ GBS (WT or ΔcopA) as described elsewhere (26). This study was carried out in accordance with the guidelines of the Australian National Health and Medical Research Council. The Griffith University Animal Ethics Committee reviewed and approved all experimental protocols for animal usage according to the guidelines of the National Health and Medical Research Council (approval no. MSC/01/18/AEC).

Statistical methods.

All statistical analyses used GraphPad Prism v8 and are indicated in the respective figure legends. Statistical significance was accepted at P values of ≤0.05.

Data availability.

Raw and processed data were deposited in the Gene Expression Omnibus (accession no. GSE167895 for S. agalactiae 874391 under Cu conditions and GSE167894 for S. agalactiae 874391 under control conditions).