Differential roles for the Co2+/Ni2+ transporting ATPases, CtpD and CtpJ, in Mycobacterium tuberculosis virulence


  • Daniel Raimunda,

    1. Department of Chemistry and Biochemistry, Worcester Polytechnic Institute, Worcester, MA, USA
    Search for more papers by this author
    • Both authors contributed equally to this work.
  • Jarukit E. Long,

    1. Department of Microbiology and Physiological Systems, University of Massachusetts Medical School, Worcester, MA, USA
    Search for more papers by this author
    • Both authors contributed equally to this work.
  • Teresita Padilla-Benavides,

    1. Department of Chemistry and Biochemistry, Worcester Polytechnic Institute, Worcester, MA, USA
    Search for more papers by this author
  • Christopher M. Sassetti,

    1. Department of Microbiology and Physiological Systems, University of Massachusetts Medical School, Worcester, MA, USA
    2. Howard Hughes Medical Institute, Chevy Chase, MD, USA
    Search for more papers by this author
  • José M. Argüello

    Corresponding author
    1. Department of Chemistry and Biochemistry, Worcester Polytechnic Institute, Worcester, MA, USA
    • For correspondence. E-mail arguello@wpi.edu; Tel. (+1) 508 831 5326; Fax (+1) 508 831 4116.

    Search for more papers by this author


The genome of Mycobacterium tuberculosis encodes two paralogous P1B4-ATPases, CtpD (Rv1469) and CtpJ (Rv3743). Both proteins showed ATPase activation by Co2+ and Ni2+, and both appear to be required for metal efflux from the cell. However, using a combination of biochemical and genetic studies we found that these proteins play non-redundant roles in virulence and metal efflux. CtpJ expression is induced by Co2+ and this protein possesses a relatively high turnover rate. A ctpJ deletion mutant accumulated Co2+, indicating that this ATPase controls cytoplasmic metal levels. In contrast, CtpD expression is induced by redox stressors and this protein displays a relatively low turnover rate. A ctpD mutant failed to accumulate metal, suggesting an alternative cellular function. ctpD is cotranscribed with two thioredoxin genes trxA (Rv1470), trxB (Rv1471), and an enoyl-coA hydratase (Rv1472), indicating a possible role for CtpD in the metallation of these redox-active proteins. Supporting this, in vitro metal binding assays showed that TrxA binds Co2+ and Ni2+. Mutation of ctpD, but not ctpJ, reduced bacterial fitness in the mouse lung, suggesting that redox maintenance, but not Co2+ accumulation, is important for growth in vivo.


Upon infection, Mycobacterium tuberculosis replicates within a phagosome-like compartment of the host macrophage (Aderem and Underhill, 1999; Flynn and Chan, 2001; Vergne et al., 2004). Transition metal homeostasis (namely Cu+/2+, Zn2+, Fe2+/3+and Mn2+) can alter the outcome of this interaction by numerous mechanisms (Forbes and Gros, 2001; Argüello et al., 2011; Hood and Skaar, 2012; Rowland and Niederweis, 2012). Increased Cu+ and Zn2+ and decreased Fe2+ concentrations have been described in phagosomes of interferon-gamma-activated macrophages (Wagner et al., 2006). These alterations, in conjunction with the activation of the phagosomal inducible nitrogen oxide synthase (iNOS) and NADPH oxidase, lead to bacterial clearance (Flynn and Chan, 2001; Vergne et al., 2004). The pathogen adapts to this environment through the use of specific metal transporters. Transition metals are essential micronutrients, as they are cofactors in bacterial metalloenzymes necessary for various metabolic processes and coping with redox stress. As a result metal importers are required for bacterial virulence in a number of systems (Forbes and Gros, 2001; Argüello et al., 2011; Botella et al., 2011; Rowland and Niederweis, 2012; Padilla-Benavides et al., 2013). However, at high concentrations these metals can also be cytotoxic, and bacterial efflux systems are equally important for growth in the host environment. Consequently, M. tuberculosis infection is dependent on maintaining appropriate transition metal homeostasis.

The M. tuberculosis genome possesses an unusually high number of heavy metal transporting P1B-type ATPases.1 Members of this family of proteins are characterized by a highly conserved core protein structure and a common mechanism of transport (Argüello et al., 2007; 2011). These can be classified into distinct subgroups, clustered by conserved transmembrane metal binding sites (TM-MBS) and consequent metal specificity (Argüello, 2003). P1B-ATPases export cytosolic transition metals as part of metal excess responsive systems, and control cellular metal quotas (Raimunda et al., 2011). Moreover, emerging new data suggest that some P1B-ATPases are part of the redox tolerance machinery of intracellular pathogens, as they participate in the assembly of membrane and secreted redox metalloenzymes (González-Guerrero et al., 2010; Argüello et al., 2011; Osman et al., 2013; Padilla-Benavides et al., 2013). These represent a novel function for metal transporters; i.e., they are also part of metalloprotein assembly systems. More importantly, this has notable implications for the conceptual development of metal homeostasis models. This is, transition metals although not free are transported to specific target metalloproteins that from a system perspective operate as metal pool/compartments. The identification of additional examples, in particular with novel and unique metal specificities, validates the general importance of this mechanistic strategy.

The involvement of M. tuberculosis P1B-ATPases CtpV (Rv0969) and CtpC (Rv3270) in Cu+ and Mn2+ homeostasis and virulence has been shown (Ward et al., 2010; Botella et al., 2011; Padilla-Benavides et al., 2013). In addition, a previous genomic analysis directed to identify genes essential for M. tuberculosis growth during infection showed the requirement of ctpD (Rv1469) for in vivo survival fitness (Sassetti and Rubin, 2003). CtpD is a member of the Co2+/Ni2+-transporting P1B4-ATPase sub-group (Rutherford et al., 1999; Argüello, 2003; Zielazinski et al., 2012; Raimunda et al., 2012a). Characterization of the M. smegmatis homologue (MSMEG_5403) showed that it transports Co2+ and Ni2+ and its transcription was induced by Co2+ (Raimunda et al., 2012a). In agreement with this, mutations in M. smegmatis Co2+-ATPase led to an increase in intracellular Co2+ and Ni2+ levels. Additionally, an increase in susceptibility to these metals was observed (Raimunda et al., 2012a). While maintaining cytoplasmic Co2+ and Ni2+ levels appears as a simple parsimonious role for this subtype of ATPases, this model is complicated by the presence of two homologous P1B4-ATPase coding genes, ctpD and ctpJ, in several Mycobacterium species including M. tuberculosis (Fig. 1A and B). The presence of two paralogous P1B4-ATPases was also observed in Gramella forsetii, Oligotropha carboxidovorans OM5 and Xanthobacter autotrophicus (Raimunda et al., 2012a). All these proteins present a large cytoplasmic ATP binding and hydrolysis domains and six transmembrane fragments (TM) containing metal binding residues S in TM4, and HEXXT in TM6 (Argüello, 2003; Zielazinski et al., 2012; Raimunda et al., 2012a).

Figure 1.

M. tuberculosis genome contains two P1B4-ATPases codifying genes.

A. Rooted tree of mycobacterial P1B4-ATPases obtained from genome-sequenced organisms.

B. Table list describing mycobacterial organisms having ctpD and ctpJ homologues (check marks indicate the presence of the paralogous gene in that organism).

C. Genetic environment of M. tuberculosis ctpJ and ctpD; arrows represent the DNA regulatory regions.

CtpD and CtpJ homologues from various mycobacteria are closely related at the primary sequence level (Fig. 1A). However, the genes flanking ctpD/J homologues are distinct and this genomic context can be used to assign orthology (Fig. 1B and C). The Co2+ sensing transcriptional regulator nmtR is always upstream of ctpJ orthologues. nmtR is a member of the ArsR-SmtB family of transcriptional repressors (Cavet et al., 2002). The ctpD orthologues are found together with two thioredoxins – trxA (Rv1470) and trxB (Rv1471) – and a putative enoyl-CoA hydratase (Rv1472) coding genes downstream (Fig. 1C). Unlike TrxB and TrxC, in vitro functional studies have shown that TrxA has an unusual low redox potential and is not functional in the presence of M. tuberculosis thioredoxin reductase (TrxR) (Akif et al., 2008). Although bioinformatics analysis suggests that ctpD transcription is monocistronic (http://www.tbdb.org/), it is notable that trxA is always present in the same position next to ctpD and is absent in all Mycobacterium species missing the ctpD homologue. Adding to these correlations, a proteomic study combining cellular fractionation and 2D-LC MS/MS showed the colocalization of CtpD, TrxA and the enoyl-CoA-hydratase in the membrane fraction (Mawuenyega et al., 2005).

Considering the different physiological functions observed in homologous P1B-ATPases (González-Guerrero et al., 2010; Argüello et al., 2011; Osman et al., 2013; Padilla-Benavides et al., 2013), the presence of CtpD and CtpJ in M. tuberculosis presents a unique opportunity to demonstrate the broad application of a strategy to employ pairs of metal transporters with similar specificity but different kinetics characteristics, to transport metal to different targets. This general model was tested in comparative in vivo and in vitro analyses. In vivo experiments confirmed the requirement of ctpD, but not ctpJ, for virulence. Biochemical analysis showed identical metal specificity and transport direction for both ATPases, although they work at different rates. These biochemical characteristics were in agreement with cellular metal accumulation assays and expression profiles under different stress conditions. The data suggest that while CtpJ is responsible for maintaining Co2+ cytoplasmic level; CtpD plays a unique role in redox stress response and adapting to the host environment. Furthermore, the cotranscription of CtpD with thioredoxins, their previously demonstrated colocalization, and the specific binding of Co2+ to TrxA, suggest that CtpD participates in the metallation of cobaloproteins.


CtpD and CtpJ transport Co2+ and Ni2+ at different rates

M. tuberculosis CtpD and CtpJ sequences contain conserved amino acids present in homologous proteins that are selective for Co2+ and Ni2+ transport (Zielazinski et al., 2012; Raimunda et al., 2012a). P1B-ATPases couple substrate transport across membranes to ATP hydrolysis following the Albers-Post E1/E2-like mechanism (Argüello et al., 2007). Both proteins were expressed in E. coli and affinity purified (Fig. 2A), to confirm their metal specificity and analyse their enzymatic characteristics. Protein preparations were incubated with TEV protease and then pretreated with chelating agents before metal dependent ATP hydrolysis was measured (Raimunda et al., 2012a). It is important to note that P1B4-ATPases lack amino- and carboxyl-terminus MBDs (N-, C-MBD) (Argüello, 2003) and that treatment of the protein with TEV protease removes the (His)6-tag used during in enzyme purification (Fig. 2A). Co2+ and Ni2+ activated CtpD and CtpJ ATPase activity in the μM range (Fig. 2B and C). Other transition metals such as Zn2+, Cu+/2+, Mn2+ and Fe2+ in concentrations ranging from nM to mM failed to activate the enzyme (data not shown). Importantly, CtpD showed a fourfold lower Vmax compared with CtpJ in presence of either substrate. This was also matched by slightly higher affinity of CtpD for the metal when compared with CtpJ. The Vmax value for both metals observed for CtpJ, resembles the biochemical behaviour of M. smegmatis Co2+-ATPase that we designate as its orthologue (Fig. 1). The different biochemical kinetics (low versus high transport rate of CtpD versus CtpJ), suggests that while CtpJ might control the cytosolic metal levels, CtpD could be required for additional physiological functions, as seen with other P1B1-ATPases (González-Guerrero et al., 2010; Raimunda et al., 2011; Osman et al., 2013).

Figure 2.

CtpD and CtpJ are Co2+/Ni2+-ATPases with different transport kinetics. (A) Preparations of Ni-NTA purified CtpD and CtpJ analysed by Coomassie brilliant blue (CBB) and Western blot (WB) using anti-(His)6 tag antibody. (B) Co2+ and (C) Ni2+-dependent ATPase activity for CtpD (●) and CtpJ (■). Curves of ATPase activity versus metal concentrations were fit to v = Vmax[metal]/([metal] + K1/2). Kinetic parameters obtained with Co2+ for CtpD and CtpJ were Vmax = 0.35 ± 0.03 μmol mg−1 h−1, K1/2 = 4.83 μM ± 1.4 and Vmax = 1.38 ± 0.12 μmol mg−1 h−1, K1/2 = 6.3 ± 2.2 μM respectively. Kinetic parameters with Ni2+ for CtpD and CtpJ were Vmax = 0.36 ± 0.02 μmol mg−1 h−1, K1/2 = 4.90 ± 1.18 μM and Vmax = 0.94 ± 0.05 μmol mg−1 h−1, K1/2 = 13.23 ± 2.80 μM respectively. The reported standard errors for Vmax and K1/2 are asymptotic standard errors reported by the fitting software KaleidaGraph (Synergy). Data are mean ± SE (n = 3).

CtpD and CtpJ bind cytoplasmic metals

Previous reports on P1B4-ATPases have shown that these enzymes drive the efflux of cytoplasmic Co2+ (Rutherford et al., 1999; Raimunda et al., 2012a). The binding of the metal substrate to the TM-MBS facing the intracellular side is a well-known requisite for ATP hydrolysis by P1B-ATPases (Argüello et al., 2007; Raimunda et al., 2011). Nevertheless, it can be argued that CtpD or CtpJ might drive Co2+/Ni2+ influx and ATP hydrolysis is associated with the efflux of an alternative substrate. This would imply Co2+/Ni2+ binding to CtpD or CtpJ in the E2 conformation, i.e., exposing the TM-MBS to the extracellular space. Vanadate stabilizes P-ATPases in E2 conformation (Pick, 1982; Eren and Arguello, 2004). Thus, to evaluate the direction of transport the metal binding of Co2+ and Ni2+ to CtpD and CtpJ was measured in the absence or presence of vanadate. Incubation of each (His)6-less protein with the metals in a molar ratio 1:10 showed a TM-MBS binding stoichiometry of approximately 1:1 molar ratio in agreement with previous reports (Zielazinski et al., 2012; Raimunda et al., 2012a) (Table 1). This binding was largely abolished in the presence of 1.5 mM vanadate. The inhibition of metal binding observed in both proteins can be mechanistically explained considering the displacement of the E1/E2 equilibrium towards the E2 state (TM-MBS exposed to the extracellular side) preventing binding to the E1 form. Based on these data, we conclude that both proteins drive the efflux of the cytoplasmic substrate.

Table 1. Metal binding stoichiometry of His-less CtpD and CtpJ
ProteinCo2+ protein molar ratioaNi2+ protein molar ratioa
No vanadate1.5 mM vanadateNo vanadate1.5 mM vanadate
  1. aStoichiometry was estimated as moles metal : moles CtpD. Metal content in protein was determined by furnace AAS.
CtpD1.08 ± 0.140.35 ± 0.021.04 ± 0.050.38 ± 0.08
CtpJ1.06 ± 0.270.05 ± 0.010.92 ± 0.130.28 ± 0.03

Deletion of ctpD and ctpJ cause opposing alterations in cellular Co2+ levels

Previously described P1B-ATPases have been mostly associated with maintaining cytoplasmic transition metal levels (Argüello et al., 2007; 2011; Osman and Cavet, 2008). However, in the human opportunistic pathogen Pseudomonas aeruginosa, the two paralogous genes coding for Cu+ transporting ATPases show no redundancy in their functional roles. One gene is involved in Cu+ homeostasis (CopA1) while the other participates in metalloprotein biogenesis (CopA2) (González-Guerrero et al., 2010). Considering the possibility of similar alternative roles for M. tuberculosis P1B4-ATPases, their involvement in maintaining cellular metal quotas was tested. Wild-type, single ctpD::hyg and ctpJ::hyg mutants, and the double mutant ctpD-ctpJ::hyg strains were challenged by supplementing 7H9-OADC media with various metals and the resulting cellular metal levels were determined. A significant decrease of Co2+ content was observed in the ctpD::hyg mutant, whereas Co2+ accumulated in ctpJ::hyg and ctpD-ctpJ::hyg double mutant strains when compared with the wild-type (Fig. 3A). Complemented mutant strains showed metal contents similar to those of wild-type cells. No significant differences were observed in Ni2+ or Cu2+-challenged cells (Fig. 3B and C). These results, together with the biochemical kinetic parameters observed in the purified preparations of CtpJ (Fig. 2B), support the hypothesis that CtpJ is responsible for maintaining cytoplasmic Co2+ (but not Ni2+) levels. CtpD appears to have an alternative role since deletion of the coding gene does not lead to an increase but rather a decrease in cytoplasmic Co2+ level. As shown below this is likely due to a compensatory increase of ctpJ expression in the ctpD mutant. It is also notable that CtpD is not able to complement the ctpJ defect probably because of its slow turnover rate. Additional experiments exploring the tolerance of ctpD and ctpJ mutant cells to Co2+, Ni2+, Cu2+ and Zn2+ showed no differences between these and the wild-type strain (Fig. S1) .

Figure 3.

Metal accumulation in M. tuberculosis ctpD::hygctpD) and ctpJ::hygctpJ) mutant strains. Log-phase cell cultures grown in 7H9-OADC media were supplemented with (A) 100 nM Co2+, (B) 3 μM Cu2+ or (C) 1 μM Ni2+ for 2 h. Data represent mean ± SE (n = 3). *P < 0.01 and **P < 0.05 versus wild-type.

Expression of ctpD and ctpJ is induced by different stressors

M. tuberculosis manages to survive in the hostile phagosomal environment through to its ability to adapt to changes in pH, transition metal bioavailability, redox stress and nutritional starvation (Ehrt and Schnappinger, 2009; Rowland and Niederweis, 2012). Toward elucidating the mechanism of ctpD participation in the infection process, its expression under stress conditions present in the phagosome and upon exposure to various metals was studied. Sublethal concentrations for Co2+, Ni2+ and Zn2+ were chosen from metal sensitivity assays (Fig. S1). None of the tested metals induced ctpD expression (Fig. 4A). On the other hand, in a similar manner to the M. smegmatis orthologue, expression of ctpJ in M. tuberculosis was induced by the presence of Co2+ (Fig. 4B) (Cavet et al., 2002; Raimunda et al., 2012a). Addition of Ni2+ to the media did not induce the transcription of ctpD or ctpJ (Fig. 4A and B). Among a battery of tested redox stressors, nitroprusside and potassium cyanide (KCN) increased ctpD expression 10- and 7-fold respectively (Fig. 4A). Nitroprusside, triclosan and tert-butyl hydroperoxide (TBHP) also led to induction of ctpJ transcription, although these were smaller than the induction by Co2+ (Fig. 4B).

Figure 4.

Transcriptional analysis of ctpD and ctpJ in M. tuberculosis. Log-phase H37Rv M. tuberculosis cell cultures grown in 7H9-OADC (A, B, E) or Chelex-treated Sauton's media (C, D) were supplemented or not, with the indicated metals and redox stressors for 2 h. RNA was extracted and gene expression analysed for (A, C) ctpD or (B, D) ctpJ transcriptional induction. (E) ctpD and ctpJ fold-induction analysed in ctpJ::hyg (black bar) or ctpD::hyg (white bar) strains respectively. Data represent mean ± SE (n = 3).

The lack of ctpD induction by metals might be a consequence of the media used, as 7H9 contains significant amounts of Cu2+ and Zn2+. Similarly, lack of ctpJ Ni induction could be due to Ni2+ presence. Therefore, we tested the response to Cu2+, Zn2+, Co2+ and Ni2+ of ctpD and ctpJ transcription in cells grown in Sauton's media treated with Chelex. This treatment decreased metal levels to sub-nM concentrations as measured by AAS (not shown). None of the metals tested induced ctpD transcription (Fig. 4C). Unexpectedly, its transcription was decreased by Zn2+. Like in 7H9 media, ctpJ was induced only by Co2+ (Fig. 4D). Since CtpD and CtpJ are highly homologous and are both capable of binding and transporting Co2+, a compensatory induction of one ATPase in a mutant lacking the other might be hypothesized. No induction of ctpD was observed in the ctpJ::hyg mutant strain (relative expression was 0.55 ± 0.01 fold, relative to that observed in the wild-type cells). Alternatively, when ctpJ expression was analysed in the ctpD::hyg cells a 3.6 ± 0.6 fold induction was observed (Fig. 4E). This induction of ctpJ in the ctpD::hyg mutant strain likely explains the decrease in intracellular Co2+ observed in these cells (Fig. 3A).

ctpD is cotranscribed with thioredoxin A, B and an enoyl-CoA hydratase under reactive nitrogen species (RNS) stress

The described data support the role of CtpD as a Co2+-ATPase distinct from CtpJ, but did not reveal the role of this enzyme. ctpD's genetic environment was considered in the search for clues on its role (Fig. 1C). Genetic studies have shown that in M. tuberculosis ctpD and trxA are under regulation of the stress-responsive sigma factor SigF. M. tuberculosis sigF is induced under stress conditions such as nutrient starvation, macrophage infection, or stationary phase entry (Williams et al., 2007). Searching for a molecular/functional link between these genes, we evaluated whether or not ctpD is part of a contiguous operon cotranscribed with down-stream genes under stress conditions (Fig. 1C). RNA extracted from cells challenged for 2 h with nitroprusside was reverse-transcribed and the operon was mapped by PCR. Figure 5 shows that ctpD is cotranscribed in a single operon together with trxA, trxB and echA12, an enoyl-CoA hydratase, suggesting the participation of these genes in a co-ordinated response.

Figure 5.

ctpD is cotranscribed with trxA, trxB and Rv1472 under nitrosative stress. (A) Schematic illustration and (B) representative gel of the ctpD operon as determined by RT-PCR using the RNA obtained from M. tuberculosis wild-type cells grown under nitrosative stress conditions (1 mM nitroprusside in 7H9-OADC for 2 h). The location of primers (small arrows) and expected size of the amplification product are indicated. ‘–’ and ‘+’ indicate no retro transcribed RNA control and retro transcribed RNA (cDNA) respectively.

TrxA binds specifically Co2+ and Ni2+

The cotranscription of trxA and ctpD together with their colocalization (Mawuenyega et al., 2005), and the low turnover rate of CtpD suggest that TrxA could be a CtpD's client protein in the extracellular side of the membrane. In this way TrxA would be capable of accepting Co2+/Ni2+ as a metal chaperone or a redox buffer/operator. To test this, metal binding to heterologously expressed and purified TrxA was assayed. Prior to this, the (His)6-tag used for purification was removed from TrxA by TEV treatment (Fig. 6A). TrxA was capable of Co2+ and Ni2+ binding under reducing conditions with a stoichiometry of 2:1 (Fig. 6B). This binding appeared specific for these metals since the protein could not bind Zn2+.

Figure 6.

TrxA binds Co2+ and Ni2+.

A. Coomassie brilliant blue (CBB) and immune staining with anti-(His)6 antibody (WB-anti-(His)6) of (1) (His)6-TEV, (2) (His)6-Tev-TrxA, (3) (His)6-Tev-TrxA plus (His)6-TEV at time zero and (4) reverse purified (His)6-less TrxA after 3 hour incubation. Note that Tev refers to the cleavage site and TEV refers to the protease.

B. Co2+ (white bars), Ni2+ (grey bars) or Zn2+ (black bars) binding to (His)6-less TrxA was determined in the presence (+) and absence (−) of TCEP. Data represents the mean ± SE (n = 3).

ctpD is required for the growth of M. tuberculosis in mouse lung

Previous whole genome genetic screens suggested that ctpD but not ctpJ may be required for the growth or survival of M. tuberculosis during infection (Sassetti and Rubin, 2003). To verify this prediction, we directly determined the relative fitness of these mutants in a competitive infection assay. Mice were infected with a mixture of wild-type and mutant bacteria and the relative fitness of each strain was estimated by determining the change in ratio (wild-type/mutant) in mouse lungs over time (Fig. 7). Over a 28-day infection the representation of the ctpD::hyg mutant decreased by 10-fold. In contrast, the ctpJ::hyg strain and the complemented ctpD::hyg mutant grew indistinguishably from wild-type M. tuberculosis.

Figure 7.

CtpD is required for proper fitness of M. tuberculosis in a mix competition assay. Competitive index of each mutant relative to wild-type H37Rv was measured at 0 and 28 days post infection in mice lungs. White bars, ctpD::hyg mutant versus H37Rv; gray bars, ctpJ::hyg mutant versus H37Rv; black bars, ctpD::hyg mutant complemented with plasmid pJEB402-D versus H37Rv. Data are mean ± SE (n = 3). *P < 0.05 versus post-infection day 0.


The P1B-ATPase transition metal transporters are emerging as important determinants of virulence in a variety of intracellular pathogens (Osman et al., 2010; Argüello et al., 2011; Botella et al., 2011; Raimunda et al., 2011; McLaughlin et al., 2012; Padilla-Benavides et al., 2013). Genomes of pathogens frequently contain several P1B-ATPases, with identical metal specificity and direction of transport (Argüello, 2003; Argüello et al., 2011; Osman et al., 2013). Co2+ transporting P1B4-ATPases are widely distributed in nature (Rutherford et al., 1999; Argüello, 2003; Zielazinski et al., 2012; Raimunda et al., 2012a). However, only some plants and a few bacterial genera, M. tuberculosis among them, possess two P1B4-ATPases. Those present in mycobacteria have been referred as CtpD and CtpJ. Previous genetic screens suggested that these proteins serve non-redundant roles during infection (Sassetti and Rubin, 2003). On the other hand, evidence is emerging supporting distinct roles for Cu+- and Mn2+-ATPases not only in maintaining cytoplasmic metal quotas, but also participating in the assembly of metalloproteins. In this report, we present evidence of the alternative functions played by Co2+-ATPases in mechanisms of metal homeostasis and allocation. In particular, we characterize the functional role CtpD, and its singular importance for M. tuberculosis virulence.

Recent biochemical studies of bacterial P1B4-ATPases support their role in Co2+ and possibly Ni2+ homeostasis (Zielazinski et al., 2012; Raimunda et al., 2012a). M. tuberculosis CtpJ and CtpD are structurally homologous to those previously studied. Our results show that both Co2+ and Ni2+ trigger the CtpD and CtpJ ATPase activity. The apparent substrate affinities and the relative activation by Ni2+ and Co2+ are similar to those previously described (Zielazinski et al., 2012; Raimunda et al., 2012a). However, metal accumulation experiments performed in M. smegmatis (Raimunda et al., 2012a) and M. tuberculosis (this study), as well as induction of gene expression by Co2+ but not Ni2+ (Rutherford et al., 1999; Raimunda et al., 2012a), strongly indicate that these enzymes transport Co2+ rather than Ni2+ in vivo. The lack of CtpJ induction by Ni2+ appears to be at odds with previous LacZ reporter studies, which concluded that the operator-promoter region of ctpJ is driven by NmtR in mycobacterial cells grown with Ni2+ (Cavet et al., 2002). However, these studies were performed after much longer (20 h versus 2 h) exposure to similar metal levels. It is possible that this harsh condition might be eliciting pleiotropic effects leading to gene induction. Although unlikely, in our experimental conditions, a Ni2+ carryover in cells grown in 7H9 media could mask a transcriptional induction in the Chelex-treated Sauton's media.

M. tuberculosis CtpJ maximum activity is comparable to that of the M. smegmatis ortholog, although both are much lower (30-fold) than that reported for CoaT from Sulfitobacter sp. NAS-14.1 (Zielazinski et al., 2012). However, the high activity of CoaT might be exceptional, as it is much higher than any other reported for a transition metal transporter. The even slower turnover rate of CtpD suggests that this protein might not contribute to maintenance of the cytoplasmic levels of these metals. Metal accumulation experiments support this idea. Co2+ accumulated in the ctpJ::hyg mutant strain, while the ctpD::hyg showed a decrease in Co2+ content (the double ctpD-ctpJ::hyg mutant showed Co2+ levels comparable to the ctpJ::hyg mutant). This observation might raise the question of whether CtpD drives metal efflux or influx. Both CtpD and CtpJ, binds the substrate to be exported when intracellular-facing transport sites are available, strongly implying that both act as exporters. Instead, we speculate that the observed compensatory induction of ctpJ expression in the ctpD::hyg mutant is responsible for the decreased Co2+ levels. While ctpJ induction might be related to a RNS detoxification-like effect produced by the deletion of ctpD (ctpJ is induced by nitroprusside), it also results in a higher Co2+ exporting capability. This compensatory effect of ctpD deletion is reminiscent of the decrease in Cu+ content after deletion of Cu+-ATPases likely involved in metallation of cuproproteins rather than maintaining compartmental Cu+ levels (Tottey et al., 2001; Argüello et al., 2011; Raimunda et al., 2011).

Surprisingly, the growth of ctpD::hyg and ctpJ::hyg M. tuberculosis strains was not inhibited by Co2+. This could be explained on the basis of the heightened capacity of M. tuberculosis cells to sense and respond to increases in the bioavailability of Co2+. It is known that this function is shared by two transcriptional repressors (Cavet et al., 2002; Campbell et al., 2007). Responses mediated by NmtR lead to ctpJ transcriptional induction, while KmtR mediates the induction of Rv2025c, a putative cation diffusion facilitator (CDF) transporter (Campbell et al., 2007). NmtR responses are observed at high metal concentrations, whereas KmtR has been detected to act with an extremely high-affinity. In this scenario, and with a high metal content in the sensitivity assay, the Rv2025c gene product could be responsible for maintaining tolerable cytoplasmic Co2+ levels in the absence of CtpJ. While the unique roles of CtpJ and Rv2025c remain unclear, their differential regulation, substrate specificity, or energetic requirements are likely to underlie the retention of both in the genomes of many mycobacterial species.

Stringent and hostile conditions met by M. tuberculosis in phagosomes induce expression of genes required for fundamental metabolic tasks, such as detoxification of reactive oxygen species (ROS), RNS, and transition metal homeostasis. To this end, physiological stresses faced by M. tuberculosis in phagosomes were mimicked in vitro and ctpD and ctpJ expression levels in wild-type strain were analysed. Among the several stimuli, only nitroprusside and cyanide induced the ctpD transcription (Fig. 4A). Both of these compounds inhibit respiratory chain electron flow and might influence ctpD expression by altering cellular redox state.

CopA2, a Cu+-transporting ATPase in P. aeruginosa, was shown to be part of an operon containing the cbb3-1 and cbb3-2 involved in cytochrome c oxidase assembly by supplying Cu+ to the catalytic site (González-Guerrero et al., 2010; Hassani et al., 2010). The observed connections between ctpD and trxA in mycobacteria lead us to analyse the possibility that these genes are part of the same operon and the potential functional link between them. Using RNA obtained from cells grown under nitrosative stress, we determined the cotranscription of ctpD with trxA, trxB and echA12. To understand if M. tuberculosis TrxA is a metalloprotein that could be loaded by an ATPase such as CtpD, we assayed its metal binding capability. Under reducing conditions with TCEP, the protein was able to bind Co2+ and Ni2+, but not Zn2+, in a protein : metal ratio of 2:1 (Fig. 6B). Although differential competition for the metal between TCEP and TrxA cannot be discarded, its divalent metal adducts have similarly low stabilities. This, along with the identification of ctpD, trxA and echA12 gene products in membrane fractions, suggests a co-ordinated function, perhaps a role in lipid membrane metabolism (Mawuenyega et al., 2005). However, the participation of transition metals as cofactors of enoyl-CoA hydratases is unknown (Agnihotri and Liu, 2003) and trxA has been reported to be non-functional (Akif et al., 2008), although this could be due to inappropriate protein maturation. Interestingly, a recent report identifying virulence factors in P. aeruginosa proposed an enoyl-CoA hydratase to be relevant for infection of Caenorhabditis elegans. Its participation in a fatty acid synthesis is likely required for membrane signalling and quorum sensing (Feinbaum et al., 2012). A similar mechanism might be present in M. tuberculosis. The study of these hypothetical links is beyond the scope of the work presented here.

Finally, the requirement of CtpD and CtpJ for M. tuberculosis virulence was evaluated in a competition (mutant versus wild-type) infection model. Consistent with previous genomic screens using transposon libraries, our results highlight the importance of ctpD, but not ctpJ, during M. tuberculosis infection (Sassetti and Rubin, 2003). Mixed competition assays are useful in determining relative fitness and have been used to demonstrate indispensability of Zn2+ transport systems in Haemophilus influenza (Rosadini et al., 2011). The lack of CtpJ participation in virulence is not surprising considering that no apparent changes in Co2+ (or Ni2+) levels in activated phagosomes have been reported (Wagner et al., 2005). So why are ctpD mutants attenuated during infection? The decreased levels of Co2+ found in the ctpD mutant strongly argues against toxicity due to metal accumulation in the bacterial cell. Instead, the putative involvement of CtpD in the response to redox stress in the phagosome might explain why CtpD is required for virulence (Flynn and Chan, 2001; Vergne et al., 2004). CtpD's cotranscription with trxAB and echA12 may suggest that CtpD plays a role in activating these proteins through metallation, protecting M. tuberculosis from redox stress. CtpD may also play a role in activating other proteins required for growth.

In this study, we have characterized the functional roles of two M. tuberculosis P1B4-ATPases and established their requirement for the infection process. CtpJ is in charge of cytosolic Co2+ and Ni2+ levels and is dispensable for infection. CtpD appears necessary for metallation of secreted proteins and for overcoming redox stress, reflecting its requirement in M. tuberculosis to strive in the harsh environmental phagosome conditions. It remains to be elucidated the participation of CtpD in fatty acid synthesis, plasma membrane lipid remodelling, or signalling processes during lung colonization.

Experimental procedures

Recombineering, mutant and complemented strains preparation

Deletion mutants were prepared in the background of the M. tuberculosis H37Rv wild-type strain using primers listed in Table 2 according to standard protocols (van Kessel and Hatfull, 2007). For ctpD mutation, a 1000 bp fragment corresponding to the 5′ first 500 bp and 3′ last 500 bp of the ctpD gene was designed using tuberculist (http://tuberculist.epfl.ch). These regions included 30 bp that flanked upstream and downstream of ctpD gene. Additionally, an insertion cassette containing the restriction sites for SpeI-HpaI-AscI was added between the 5′ and 3′ 500 bp regions. The resulting synthesized fragment was then inserted in a HindIII site into a pUC57 expression vector (GenScript), resulting in pEL2a. Vector pKM342 contains a hygromycin resistance (hygR) cassette flanked by SpeI-AscI sites. To insert the hygR cassette in pEL2a, both pEL2a and pKM342 were digested with SpeI-AscI. The 1.2 kbp hygromycin fragment was then ligated into pEL2a, resulting in pEL2b. To generate the deletion mutant of ctpJ two 500 bp amplicons corresponding to the 5′ first 400 bp plus 100 bp upstream, and the 3′ last 400 bp plus 100 bp downstream of the ctpJ gene were obtained by PCR. These were ligated in pJM1 flanking a hygR cassette resulting in pJM1-J. To generate ctpD and ctpJ mutants, the resulting 2.2 kbp ctpD-hygR-ctpD fragment from digestion of pEL2b with HindIII and the 2.8 kbp ctpJ-hygR-ctpJ generated by PCR using as template pJM1-1 plasmid, were both transformed into M. tuberculosis H37Rv recombineering strain. Briefly, the M. tuberculosis H37Rv recombineering strain bearing plasmid pJV53 was grown till OD600 = 0.7 and incubated for 18 h with 1 μM isovaleronitrile. The culture was treated with 0.2 M glycine for 8 h before making electrocompetent cells. The recombineering strain was electroporated and selected by hygromycin resistance on 7H10 plates. After 18 days, hygR colonies were isolated and transferred to 2 ml inkwells containing 7H9 media supplemented with 50 μg ml−1 hygromycin. To verify the presence of the ctpD and ctpJ deletion mutation, PCR amplification of the hygR cassette flanked by the N- and C- terminals was performed. The double mutant was obtained as described above for ctpJ, although using a cleaned-up hygromycin sensitive ctpD mutant as parental strain.

Table 2. List of primers used in this work
F-ctpJ-EcoRIAGCTGAATTCGTGGCTGTTCGTGAACTCTCTCClone ctpJ in pJEB402 under mop promoter regulation
R-ctpJ-HpaIGACTGTTAACTCACCTGTGCGCGGTGCCAGCClone ctpJ in pJEB402 under mop promoter regulation
prEL4-ctpD forwardATTCTAGACGATGATTAGCGCGGCCAACClone ctpD in pJEB402 under mop promoter regulation
prEL6-ctpD reverseATGGTACCAGCGTGGGCAACAACTTGACClone ctpD in pJEB402 under mop promoter regulation
F-5UTR-ctpJACTGGCGGCCGCGACCAAGGTGCAAGTGCTCGCTGClone 400 bp upstream ctpJ including first 100 bp of the gene
R-5UTR-ctpJCAGTACTAGTCCCAACGCATCTCCGACAACGCClone 400 bp upstream ctpJ including first 100 bp of the gene
F-3UTR-ctpJACTGCTCGAGCGCCCGACACGAAGGTTCCACCClone 400 bp downstream ctpJ including last 100 bp of the gene
R-3UTR-ctpJCAGTGGGCCCAGCGAGCGTCCTCGGTGGACCClone 400 bp downstream ctpJ including last 100 bp of the gene
For-1469ACGATCATCGGGTTGGCACGctpD genetic environment
Rev-1470GTCAAAGTGTTTCCGCGACGctpD genetic environment
For-1470CGAGAAAGATCTGGCCTCGctpD genetic environment
Rev-1471CGCTGCAAGCTCTCGTTCctpD genetic environment
For-1471GAACGAGAGCTTGCAGCGctpD genetic environment
Rev-1472CGCTAAGGCCTCTTTGAGCctpD genetic environment
For-1472GATGAACAGCTGCTAGATGCctpD genetic environment
Rev-1473CTTCTCCAGATCAGTGAGCctpD genetic environment

The complementation assay constructs were made by amplifying M. tuberculosis ctpD and ctpJ from genomic DNA. The resulting PCR fragments were digested and ligated into pJEB402 which confers kanamycin resistance (kanR) resulting in pJEB402-D and pJEB402-J. The ligation reactions were transformed into DH5α cells and the presence of the insert was verified by colony PCR and restriction digests. The plasmids were then purified and transformed into the ctpJ and ctpD M. tuberculosis mutant strains. Transformants showing kanR were analysed for the presence of ctpD and ctpJ by PCR.

Mice infection

The relative growth rates of M. tuberculosis H37Rv wild-type and ctpD::hyg and ctpJ::hyg mutant strains, were examined in vivo competition experiments. Briefly, the two strains were mixed with wild-type H37Rv in approximately 3:1 (wild-type:ctpD::hyg or wild-type:ctpJ::hyg mutants) ratio (final volume 200 μl), containing 6 × 105 cfu and were inoculated into the tail vein of female C57BL/6J mice. Groups of three mice were sacrificed at indicated time points and the bacterial burden in the lung homogenates were obtained by plating on 7H10 agar medium with or without 100 μg ml−1 hygromycin for mutants cfu and total cfu counting respectively. A competitive index was calculated as (cfu WT/cfu mutant)output/(cfu WT/cfu mutant)input. Mice were housed under specific pathogen-free conditions and in accordance with the University of Massachusetts Medical School, IACUC guidelines.

Protein expression and purification

cDNA encoding M. tuberculosis ctpD, ctpJ and trxA were amplified using genomic DNA as template and reverse primers that introduced a Tobacco etch virus (TEV) protease site coding sequence at the amplicon 3′ ends. The PCR products were cloned into pBAD-TOPO/His (ctpD and ctpJ) or pEXP5-CT-TOPO (trxA) vector (Invitrogen) that introduce a (His)6-tag at the carboxyl end of the protein. cDNA sequences were confirmed by automated sequence analysis. For CtpD and CtpJ expression the constructs were introduced into E. coli LMG194 ΔcopA cells (Rensing et al., 2000). For protein expression, cells were grown at 37°C in ZYP-505 media supplemented with 0.05% arabinose, 100 μg ml−1 ampicillin, and 50 μg ml−1 kanamycin (Studier, 2005). Affinity purification of membrane proteins and removal of the (His)6-tag was performed as previously described (Mandal et al., 2002; Raimunda et al., 2012b). Solubilized lipid/detergent micellar forms of CtpD and CtpJ proteins were stored at −20°C in buffer containing 25 mM Tris, pH 8.0, 50 mM NaCl, 0.01% n-dodecyl-β-D-maltopyranoside (DDM) (Calbiochem), and 0.01% asolectin until use. TrxA was expressed in BL21 (DE3)pLysS cells transformed with pEXP5-CT-TrxA following the protocol described previously (Akif et al., 2008). Protein determinations were performed in accordance to Bradford (Bradford, 1976). Protein purity was assessed by Coomassie brilliant blue (CBB) staining of overloaded SDS-PAGE gels and by immunostaining Western blots with rabbit anti-(His)6 polyclonal primary antibody (GenScript) and goat anti-rabbit IgG secondary antibody coupled to horseradish peroxidase (GenScript). Previous to ATPase activity determinations, proteins (1 mg ml−1) were treated with 0.5 mM EDTA and 0.5 mM tetrathiomolybdate (TTM) for 45 min at room temperature. Chelators were removed using Ultra-30 Centricon (Millipore) filtration devices.

ATPase assays

These were performed at 37°C in a medium containing 50 mM Tris (pH 7.4 at RT), 3 mM MgCl2, 3 mM ATP, 0.01% asolectin, 0.01% DDM, 20 mM cysteine, 50 mM NaCl, 2.5 mM DTT, and 10–40 mg ml−1 purified protein plus the indicated metal concentrations. DTT was replaced by TCEP to prevent interference with the colorimetric reaction when Co2+ activation was measured. ATPase activity was measured after 10 min incubation and released Pi determined according to Lanzetta et al. (1979). ATPase activity measured in the absence of metal was subtracted from plotted values. Curves of ATPase activity versus metal concentrations were fit to v = Vmax[metal]/([metal] + K1/2). The reported standard errors for Vmax and K1/2 are asymptotic standard errors reported by the fitting software KaleidaGraph (Synergy).

Metal binding determinations

Maximum metal binding to isolated CtpD and CtpJ in the presence (1.5 mM) or absence of vanadate was measured as previously described (Raimunda et al., 2012a). Ten-micromolar CtpD was incubated for 1 min at 4°C in 50 mM HEPES-NaOH, pH 7.5, 200 mM NaCl, 1 mM TCEP, and 50 μM of CoCl2 or NiCl2. After incubation, excess metal was removed by size exclusion using Sephadex G-25 columns (GE Healthcare). Eluted protein was acid digested with 1.25 ml HNO3 (trace metal grade) for 1 h at 80°C and then overnight at 20°C. Digestion was concluded by addition of 0.25 ml of 30% H2O2 and dilution to 3 ml with water. Metal binding to TrxA was similarly determined except that Sephadex G-10 columns were used to remove metal excess. Metal content in digested samples was measured by furnace atomic absorption spectroscopy (AAS; Varian SpectrAA 880/GTA 100). Background metal level in control samples lacking protein was < 10% of those observed in protein containing samples.

Metal content analysis

M. tuberculosis H37Rv wild-type, the single ctpD::hyg and ctpJ::hyg, and the double ctpD-ctpJ::hyg deletion mutants and complemented strains were grown to the late exponential phase and incubated in the presence or absence of 1 μM CoCl2, NiCl2 or CuCl2 for 1 h. After the incubation, cells were washed with 25 mM Tris pH 7.0 and 100 mM KCl and protein levels were determined. Samples were acid digested as described above and metal concentrations measured using furnace AAS.

Gene expression analysis

M. tuberculosis H37Rv wild-type cells in exponential phase in 7H9-OADC were cultured for 2 h with 100 μM Cu2+, Zn2+, Co2+, Ni2+, Mn2+ or Fe2+ as chloride salts, except for Zn2+ that was added in the sulphate form. Alternatively cells were exposed to 1 mM triclosan, a lipid metabolism inhibitor, nitroprusside, a nitrosative stressor, or the oxidative stressors TBHP, Paraquat, KCN. In alternative experiments, cells grown in Sauton's pretreated with Chelex (Sigma) (1 g/100 ml) were exposed to 10, 50 and 100 μM of Cu2+, Zn2+, Co2+ or Ni2+. Cells were harvested, resuspended in 1 ml of TRIzol reagent (Invitrogen), and disrupted using lysing matrix B (MP Biomedicals) in a cell disrupter (FastPrep FP120, Qbiogene). RNA pellets were air dried and redissolved in 50 μl of diethyl pyrocarbonate-treated ultrapure water. Remaining DNA was removed with RNeasy minikit and an on-column DNase I kit (Qiagen). The RNA samples (1 μg) were used as templates for cDNA synthesis with random primers and SuperScript III reverse transcriptase (Invitrogen). Quantitative reverse transcription-PCR (qRT-PCR) was performed with iQ SYBR green supermix (Bio-Rad Laboratories), using primers listed in Table 2 and cycler conditions previously described (González-Guerrero et al., 2010). The RNA polymerase sigma factor (sigA) was used as an internal reference. Determinations were carried out with RNA extracted from three independent biological samples, with the threshold cycle (Ct) determined in triplicate. The relative levels of transcription were calculated by using the 2-ΔΔCt method (Livak and Schmittgen, 2001). The mock reverse transcription reactions, containing RNA and all reagents except reverse transcriptase, confirmed that the results obtained were not due to contaminating genomic DNAs (data not shown). ctpD expression in the ctpJ mutant, and vice versa ctpJ expression in ctpD mutant cells, were measured in cells treated in similar manner.


This work was supported by NIH awards F32A1093049 (J.E.L.), 1R21AI082484 (J.M.A.), AI064282 (C.M.S.) and the Howard Hughes Medical Institute (C.M.S.).


  1. 1

    For simplicity P-type ATPases will be referred as P-ATPases, P1B-ATPases, etc.