|Home | About | Journals | Submit | Contact Us | Français|
PIB-type ATPases transport heavy metals (Cu2+, Cu+, Ag+, Zn2+, Cd2+, Co2+) across biomembranes, playing a key role in homeostasis and in the mechanisms of biotolerance of these metals. Three genes coding for putative PIB-type ATPases are present in the genome of Thermus thermophilus (HB8 and HB27): the TTC1358, TTC1371, and TTC0354 genes; these genes are annotated, respectively, as two copper transporter (CopA and CopB) genes and a zinc-cadmium transporter (Zn2+/Cd2+-ATPase) gene. We cloned and expressed the three proteins with 8His tags using a T. thermophilus expression system. After purification, each of the proteins was shown to have phosphodiesterase activity at 65°C with ATP and p-nitrophenyl phosphate (pNPP) as substrates. CopA was found to have greater activity in the presence of Cu+, while CopB was found to have greater activity in the presence of Cu2+. The putative Zn2+/Cd2+-ATPase was truncated at the N terminus and was, surprisingly, activated in vitro by copper but not by zinc or cadmium. When expressed in Escherichia coli, however, the putative Zn2+/Cd2+-ATPase could be isolated as a full-length protein and the ATPase activity was increased by the addition of Zn2+ and Cd2+ as well as by Cu+. Mutant strains in which each of the three P-type ATPases was deleted singly were constructed. In each case, the deletion increased the sensitivity of the strain to growth in the presence of copper in the medium, indicating that each of the three can pump copper out of the cells and play a role in copper detoxification.
Thermus thermophilus is an aerobic facultative Gram-negative bacterium that grows rapidly in the temperature range from 60°C to 80°C and which yields sufficient biomass to be a practical source for thermally stable proteins to be characterized by biophysical and structural methods (13). In addition, T. thermophilus can be genetically manipulated and can be used as a host for expressing recombinant proteins (25, 43). As a source for thermally stable, recombinant membrane proteins, T. thermophilus can be particularly attractive. In this report, we explore this potential by the cloning, purification, and functional characterization, both in vitro and in vivo, of the three PIB-type ATPases present in T. thermophilus HB27. The PIB-type ATPases are ATP-driven heavy metal pumps (4, 5, 28, 32, 47), and the three encoded in the genome of T. thermophilus HB27 (32) are annotated as TTC1358 (CopA, a putative Cu+ pump), TTC1371 (CopB, a putative Cu2+ pump), and TTC0354 (a putative Zn2+/Cd2+-ATPase). All three were isolated as His-tagged proteins and demonstrated to have metal-stimulated ATPase activity.
The P-type ATPase superfamily includes many cation pumps, which establish and maintain steep electrochemical gradients of key cations across membranes at the expense of ATP hydrolysis (15, 47). They are grouped into five subfamilies, PI through PV, covering a wide range of cationic as well as lipid substrates (e.g., flippases) (6, 47, 52). Among the best-studied P-type ATPases are the Na+/K+ pump (44, 54), the plasma membrane H+ pump (48), and the Ca2+-ATPase (the PIIA group) (63, 64), commonly found in eukaryotic cells but also present in some prokaryotes (21). Structures of several P-type ATPases have been determined by X-ray crystallography (28, 44, 48, 54, 62).
PIB-type ATPases (3–5, 28, 32, 47) transport heavy metals (Cu2+, Cu+, Ag+, Zn2+, Cd2+, Co2+) across biomembranes, playing key roles in metal homeostasis and, by pumping toxic metals out of cells, increasing their biotolerance of these metals (42, 50, 58, 66). The PIB-ATPases include the two human Cu+-ATPases in which mutations are responsible for Menkes and Wilson diseases (17). Structures have been reported for the Cu+ pumps from Legionella pneumophila by X ray (28) and from Archaeoglobus fulgidus by cryoelectron microscopy and computer docking (2).
PIB- and PII-type ATPases share common structural and functional features (8, 33, 47). Alignment of sequences and homology models suggest that transmembrane helices H6, H7, and H8 of the PIB-type ATPases are structurally equivalent to H4, H5, and H6, respectively, of the PII-type ATPases (3). These transmembrane helices provide signature sequences for metal selectivity (Fig. 1). A unique feature of these PIB-type proteins is the presence of a putative metal binding sequence (CPC, CPH, or SPC) in transmembrane helix 6 (69). However, the relationship between ion specificity and CPX sequences remains to be established. Another noticeable characteristic of the CPX ATPases is the presence of 1 to 6 metal binding domains in their amino termini (N-MBD). These cytoplasmic N-MBDs include either CXXC or His-rich sequences (Fig. 1). It has been shown that CXXC N-MBDs are not required for ion transport, and a regulatory role has been suggested for these domains (4, 40, 65, 67).
The DNA of the three genes encoding CopA, CopB, and Zn2+/Cd2+-ATPase from T. thermophilus HB27 (32) was amplified by PCR from genomic DNA (Table 1). The 8His tag was introduced at the C terminus of each protein using the QuikChange site-directed mutagenesis kit (Stratagene). DNA oligonucleotides were synthesized at either the University of Illinois at Urbana-Champaign (UIUC) Biotechnology Center or by Integrated DNA Technologies. The primers to amplify the three PIB-type ATPase genes were designed to introduce a 5′ NdeI site and a 3′ HindIII site at the end of each gene to clone the resulting DNA fragments directly in the shuttle vector pMKPbcbgaA (Table 1) (provided by J. Berenguer, Universidad Autónoma de Madrid, Spain). This plasmid contains the promoter from the cytochrome bc1 complex from T. thermophilus (41) and a kanamycin resistance gene (Table 1). The different gene sequences were confirmed by DNA sequencing (UIUC Biotechnology Center). All the genetic manipulations were done in Escherichia coli TOP 10 cells. Finally, T. thermophilus HB27 cells were transformed with these constructs and selected by the addition of 30 μg/ml kanamycin.
Heterologous expression in E. coli of the Zn2+/Cd2+-ATPase was done using the C43 (DE3) strain with the expression vector pET22b (Table 1). The Cys16Ala mutant of the Zn2+/Cd2+-ATPase was generated using QuikChange (Strategene) and also expressed in E. coli.
T. thermophilus HB27 was grown in 1 liter of culture medium in a 2-liter flask using rich medium (8 g/liter yeast extract, 6 g/liter tryptone, 3 g/liter NaCl, 3 ml glycerol, pH 7.4) at 60°C for 16 h in an incubator shaker (final optical density at 600 nm [OD600] ~ 1.4). E. coli C43 was grown in LB medium plus 100 μg/ml ampicillin at 37°C, and gene expression was induced by the addition of 1 mM IPTG (isopropyl-d-thiogalactoside) when the cell medium reached an A600 of ~0.7. All the purification procedures were carried out at 0 to 4°C. Cells were harvested and then resuspended in buffer A (25 mM Tris, pH 7.5, 100 mM NaCl, 10% glycerol, and 2 mM β-mercaptoethanol) with the addition of 3 mM MgSO4, DNase I, and a protease inhibitor cocktail (Sigma). The cells were then disrupted by twice passing the suspension through a Microfluidizer at a pressure of 80,000 lb/in2. Membranes were collected by centrifugation, resuspended in buffer A (plus the protease inhibitor cocktail), and then solubilized by the dropwise addition of dodecyl-β-d-maltoside (DDM) to a final concentration of 1%. After being stirred gently for 2 h at 4°C, the suspension was cleared by centrifugation to remove insoluble material. To this solution was added the Co2+ Talon resin (Clontech), preequilibrated with buffer A plus 0.05% DDM, and imidazole was added to a final concentration of 10 mM. The resin was placed in a glass chromatography column and washed with buffer A plus 0.05% DDM and 35 mM imidazole. The protein was eluted with buffer A plus 0.05% DDM and 250 mM imidazole. Fractions were concentrated using an Amicon spin filter with a molecular mass cutoff of 100 kDa, and imidazole was removed by dialysis with buffer A plus 0.05% DDM. The protein solution (approximately 4 mg/ml) could be flash frozen and stored at −80°C.
Protein concentration was determined using the BCA kit (Thermo Scientific, Pierce Protein Research Products). SDS-PAGE was done using precast 4 to 20% gels (Nusep), and proteins bands were visualized with a His tag staining kit (Thermo Scientific, Pierce Protein Research Products) and with Coomassie brilliant blue (Thermo Scientific).
Phosphatase activity was measured colorimetrically using para-nitrophenyl phosphate (pNPP) as the substrate (12, 49). Reactions were performed at 65°C for 20 min. About 0.01 mg/ml of protein was incubated with 10 mM pNPP in reaction buffer (25 mM Tris [pH 6.5], 100 mM NaCl, 3 mM MgSO4, 0.05% DDM) and various metal salt concentrations. The reaction was started by addition of the substrate and stopped by adding 1 M NaOH. Absorbance at 410 nm was used to determine the amount of p-nitrophenol product using a ε410 value of 17.000 M−1 cm−1. Enzyme activity was determined as nmol of pNPP hydrolyzed mg−1 of protein min−1.
The ATPase activity was determined at 65°C using malachite green as previously described (36). The assay mixture contained 25 mM Tris, pH 6.5, 3 mM MgSO4, 100 mM NaCl, 0.05% DDM, 1 mM ATP (or as indicated in the legend for Fig. 3), 0.01 mg/ml purified enzyme, and different reagents (CuCl2, AgNO3, NiCl2, CoCl2, CdCl2, ZnCl2, CaCl2) as indicated in the figure legends and table footnotes. Dithiothreitol (DTT; 2.5 mM) was added to the CuCl2-containing mix to produce Cu+. The pH was measured at room temperature, and the pH value at 65°C was calculated using a pKa/°C conversion factor for Tris of −0.031 (27). Metal content of the buffer used for ATP and pNPP activity was determined by atomic absorption, and neither copper nor zinc was detected.
The concentration dependence of the initial velocity of ATPase activity on activating metal ions or ATP was fit to the Michaelis-Menten equation, ν = Vmax([L]/[L] + K0.5), where [L] is the concentration of variable ligand. Data analysis was done using Origin 8.0 software.
All deletions were obtained by double homologous recombination using suicide plasmids, each a derivative of pWUR112 (10). The structural genes of the four P-type ATPases were replaced by the bleomycin resistance gene under the control of the surface layer protein A gene promoter (slAp) (22). Upstream and downstream regions flanking the P-type ATPases (flank A and flank B) were amplified via PCR using T. thermophilus HB27 genomic DNA. For flank A the forward primers introduced an EcoRI site and the reverse primers introduced an XbaI site. For flank B the forward primers gave rise to a PstI site and the reverse primers introduced a HindIII site. The resulting products (~750 bp each for flank A and flank B) were digested and cloned in two steps of ligation into pWUR112 (Table 1). Each suicide plasmid was transformed into T. thermophilus HB27, and transformants were selected on plates at 60°C in the presence of 10 μg/ml bleomycin (Table 1). The deletion of interest was verified by PCR of the genomic DNA.
The cells were grown in liquid media (8 g/liter yeast extract, 6 g/liter tryptone, 3 g/liter NaCl, 3 ml glycerol, pH 7.4) supplemented with the desired metal concentration (CuCl2, ZnCl2, or CdCl2). Cell cultures were inoculated from overnight growths to an initial A600 of 0.1, and the A600 was obtained again after different times of growth at 60°C.
Recombinant, His-tagged forms of each of the three PIB-type ATPases were successfully expressed homologously in T. thermophilus HB27 using the pMKPbcbgaA expression plasmid. Between 5 and 10 mg of each PIB-type ATPase was obtained from 12 liters of culture medium. Densitometry of Coomassie brilliant blue-stained SDS-PAGE gels shows that this procedure routinely yielded a >95% pure protein (Fig. 2, lanes 1 to 3*), and His tag staining confirms the presence of the expressed protein (Fig. 2, lanes 4 to 7*). The SDS-PAGE pattern of the Zn2+/Cd2+-ATPase indicates heterogeneity which must result from proteolysis near the N terminus, since each band is also visualized by the His tag stain and the tag is located at the C terminus (Fig. 2, lines 3* and 7*).
Figure 3 shows the influence of the presence of 1 μM and 10 μM metal cations on the ATPase activities of the three different PIB-type ATPases. CopA is stimulated to the greatest extent (about 3.5-fold) by Cu+ but also shows some stimulation by Cu2+, Ag+, and Ca2+ (Fig. 3A). CopB is stimulated more by Cu2+ (more than 4-fold) than by Cu+, as expected (Fig. 3B). Surprisingly, the putative Zn2+/Cd2+-ATPase (Fig. 3C) was stimulated in vitro by Cu2+, Cu+, Ag+, and Ni2+ but not by Zn2+ and Cd2+.
For comparison, heterologous expression of the Zn2+/Cd2+-ATPase was successfully accomplished in E. coli. Interestingly, the purified recombinant protein from E. coli does not exhibit any evidence of proteolytic fragmenting at the N terminus (Fig. 2, lanes 3* and 7*). The ATPase activity of the Zn2+/Cd2+-ATPase expressed in E. coli was increased by the addition of Zn2+ and Cd2+ but also by Cu+ (Fig. 3D). Since the sequence of the TTC0354 gene encoding the Zn2+/Cd2+-ATPase indicates a single CXXC metal binding motif (starting at Cys16) at the N terminus, the C16A mutation was made to test if perturbing this motif altered the metal specificity of the enzyme. No significant difference in the metal selectivity was observed compared to the wild type (Fig. 3D), and 75% of ATPase activity was maintained.
Cu+, Cu2+, and Zn2+ appear to have the largest effects on the ATPase activities of CopA, CopB, and the Zn2+/Cd2+-ATPase (expressed in E. coli), respectively. The concentration dependence of ATPase activity by each enzyme with its activating metal was determined, showing K0.5 values in the submicromolar range for each (Table 2). In all cases, the addition of the activating metal beyond 5 μM is inhibitory, possibly due to metal binding to the E2 form of the enzyme, as previous shown for the sodium pump (9).
The enzymatic hydrolase assay of P-type ATPases can be readily tested with the chromogenic substrate para-nitrophenyl phosphate (pNPP) instead of ATP (12, 30, 49). The assay was performed at 65°C. At higher temperatures the background hydrolysis rate of pNPP was too high. Relative activities of metal stimulation of pNPPase activity (not shown) are similar to those obtained using ATP as the substrate (Fig. 3). Inhibition by vanadate is a signature feature of P-type ATPases (7, 34), and the pNPPase activity of each of the three T. thermophilus enzymes is inhibited by vanadate (Table 2).
The effects of different metals on the growth of T. thermophilus were compared in strains in which each of the putative metal-pumping ATPases was overproduced or in which the chromosomal gene was knocked out (Table 1). In each case, cells were grown in the presence of the indicated concentration of the metals (Fig. 4) and the OD was checked after different times of growth at 60°C. The OD600 at 24 h is shown in Fig. 4 because it illustrates the biggest difference in metal sensitivity between wild type and mutants. It is important to note that the cellular growth rates in rich medium without metal addition were not modified in cells either overexpressing or lacking the P-type ATPases separately (not shown).
Figure 4A, ,B,B, and andDD show that overexpression of CopA, CopB, or the putative Zn2+/Cd2+-ATPase, respectively, increases the Cu2+ tolerance for growth compared to that of the wild type. Consistent with these results, the growth of the corresponding mutant strains was significantly more sensitive to copper in the growth medium. As a negative control the TTC0776 gene, encoding the Ca2+-ATPase in T. thermophilus, was deleted. That deficiency has no effect in copper tolerance in vivo, as is shown in Fig. 4C. These experiments demonstrate that the putative Zn2+/Cd2+-ATPase plays a role in copper detoxification, along with CopA and CopB, and are consistent with the observed copper stimulation of the ATPase activity of the isolated Zn2+/Cd2+-ATPase (Fig. 3). The dosage effect of the Zn2+/Cd2+-ATPase on growth tolerance of either Zn2+ or Cd2+ is less clear (Fig. 4E and and4F).4F). There does appear to be an increased tolerance when the Zn2+/Cd2+-ATPase is overproduced, but gene knockout does not have the opposite effect.
Each of the three membrane-bound PIB-type ATPases encoded in the genome of Thermus thermophilus has been expressed and purified in an active, His-tagged form using a T. thermophilus expression system. Two of the enzymes are homologous to copper pumps CopA and CopB, while the third is homologous to Zn2+/Cd2+-ATPases. The yield obtained for these membrane proteins is 5 to 10 mg from 12 liters of culture. All three enzymes exhibited both ATPase and pNPPase activities at 65°C and were inhibited by vanadate.
The copper ATPases studied in this work (T. thermophilus CopA [TtCopA] and TtCopB) are among the small number of copper ATPases that have been purified in an active form (31), and only a few CopB enzymes have been purified and characterized (16, 38, 57). Based on metal stimulation of both the ATPase and pNPPase activities (Fig. 3), the current work shows that in vitro TtCopA favors Cu+ whereas TtCopB favors Cu2+, similar to the pattern observed with CopA and CopB from Archaeoglobus fulgidus (38, 39, 68), with the exception that the ATPase activity of TtCopA is not further stimulated by cysteine (5 to 20 mM) in the presence of 1 μM Cu+. By contrast, the CopB from Enterococcus hirae has been shown to act as an ATP-driven pump for Cu+ and Ag+ in inverted membrane vesicles (57). In vivo studies strongly suggest that both TtCopA and TtCopB play a role in detoxification of copper and their presence increases the tolerance of T. thermophilus for Cu2+ in the growth medium (Fig. 4). Hence, both of these pumps are likely to pump copper out of the cell. The copper affinity in vitro is higher for CopB than for CopA. This is consistent with the higher copper sensitivity of the growth of the CopB knockout strain than of that of the CopA knockout. Therefore, CopB in T. thermophilus appears to be more functionally relevant at lower copper concentrations in the medium.
One would expect that any copper present in the bacterial cytoplasm would be in the reduced Cu+ form, which is potentially very deleterious (19, 37). Indeed, there appear to be no copper-requiring proteins within the cytoplasm, other than chaperones or DNA binding proteins which are parts of the systems for actively ridding the cytoplasm of copper. Although CopA or CopB are stimulated to different extents by Cu2+ in vitro, it is not clear that in vivo they would ever encounter Cu2+ but, rather, would utilize chaperone-bound Cu+ in the bacterial cytoplasm as a substrate.
The copA gene in T. thermophilus is part of an operon containing the genes encoding the cytoplasmic Cu+ chaperone CopZ and the Cu-dependent transcriptional regulator CopY (CsoR), fully consistent with a role of CopA in Cu+ detoxification of the bacterial cytoplasm (4, 46, 56). The copB gene in T. thermophilus is adjacent to the gene encoding the periplasmic multicopper oxidase (53), suggesting a functional relationship between these two proteins. One of the reactions catalyzed by the multicopper oxidases is the oxidation of Cu+ to Cu2+ in the periplasm (18), and a functional link between multicopper oxidases and either CopA or CopB has been previously noted (29, 42). The genes encoding two periplasmic copper chaperones, Sco1 and PCuAC, have been identified in the genome of T. thermophilus. PCuAC appears to deliver Cu+ into subunit II of the T. thermophilus cytochrome ba3 respiratory oxygen reductase (1), and a role for the equivalent protein in Rhodobacter sphaeroides in the assembly of copper-containing respiratory oxidases has also been reported (60). Similarly, the Sco1/PrrC/SenC family of proteins has been implicated in copper acquisition by bacteria and in the assembly of copper-containing proteins, including respiratory oxygen reductases (11, 23, 24, 59, 60). Since PIB-type copper pumps have been implicated in the assembly of the copper-containing cbb3-type respiratory oxygen reductases (26, 35), it is reasonable that the periplasmic copper chaperones present in T. thermophilus may function together with CopA and/or CopB in the assembly of the ba3-type and caa3-type copper-containing respiratory oxygen reductases present in T. thermophilus (55, 61).
The stimulation by metal cations of ATPase activity (Fig. 3) depends on whether the recombinant protein is obtained from the T. thermophilus or E. coli expression system. The expected activation by Zn2+ and Cd2+ is observed only for the enzyme expressed heterologously in E. coli. This difference in metal selectivity is likely due to proteolytic cleavage for the enzyme isolated from T. thermophilus but does not appear to be due to an alteration of the single CXXC metal-binding sequence motif near the N terminus. Mutagenesis of the first Cys of the motif to Ala has no influence on metal selectivity and only a minor influence on the ATPase activity (80% of the wild-type activity). Similar results were obtained upon mutagenesis of the Zn-ATPases from E. coli (45) and from Arabidopsis thaliana (20).
The recombinant TtZn2+/Cd2+-ATPase protein expressed either in T. thermophilus or in E. coli is significantly activated by Cu+ (Fig. 3). This is consistent with the observed increase in sensitivity of T. thermophilus growth to copper when the gene encoding this enzyme is knocked out and the increased resistance to copper toxicity upon overexpressing the putative Zn2+/Cd2+-ATPase (Fig. 4). The data on whether the putative Zn2+/Cd2+-ATPase may have a role in zinc or cadmium tolerance are equivocal and do not support a clear conclusion. Interestingly, a PIB-ATPase regulator (CsoR) in T. thermophilus has been shown to respond to either Cu+ or Zn2+ (51), suggesting an interaction in the detoxification responses of the organism to these two cations. Further work will be needed to clarify these interactions.
One caveat about the test used to evaluate the physiological role of the metal pumps is that the influence of either the overexpression or deletion of the genes expressing any individual pump might influence the expression levels of other pumps or transport systems that influence metal homeostasis. This is emphasized by the response of the CsoR regulator protein to either Cu+ or Zn2+ (51). In short, the measured sensitivity to growth could be due to an indirect effect of deleting or overexpressing the gene of interest. Along the same lines, it is important to mention that there are several different systems encoded in the genome of T. thermophilus which may be involved on zinc or copper homeostasis, including both permeases and ABC-type transporters. The activity of these systems could also be influenced by genetic manipulations of the P1B-type transporters. Nevertheless, taken at face value, the data strongly suggest that the Zn2+/Cd2+-ATPase does assist in pumping copper out of the cytoplasm of T. thermophilus.
We thank J. Berenguer and the late J. Fee for providing plasmid pMKPbcbgaA and the HB27 strain, respectively. We are also indebted to James Hemp, Young Ahn, Ranjani Murali, and G. Gokce Yildiz for their help and for useful discussions.
This research was supported by grant GM095600 from the National Institutes of Health.
Published ahead of print 25 May 2012