Co2+ Selectivity of Thermotoga maritima CorA and Its Inability to Regulate Mg2+ Homeostasis Present a New Class of CorA Proteins*

CorA is a family of divalent cation transporters ubiquitously present in bacteria and archaea. Although CorA can transport both Mg2+ and Co2+ almost equally well, its main role has been suggested to be that of primary Mg2+ transporter of prokaryotes and hence the regulator of Mg2+ homeostasis. The reason is that the affinity of CorA for Co2+ is relatively low and thus considered non-physiological. Here, we show that Thermotoga maritima CorA (TmCorA) is incapable of regulating the Mg2+ homeostasis and therefore cannot be the primary Mg2+ transporter of T. maritima. Further, our in vivo experiments confirm that TmCorA is a highly selective Co2+ transporter, as it selects Co2+ over Mg2+ at >100 times lower concentrations. In addition, we present data that show TmCorA to be extremely thermostable in the presence of Co2+. Mg2+ could not stabilize the protein to the same extent, even at high concentrations. We also show that addition of Co2+, but not Mg2+, specifically induces structural changes to the protein. Altogether, these data show that TmCorA has the role of being the transporter of Co2+ but not Mg2+. The physiological relevance and requirements of Co2+ in T. maritima is discussed and highlighted. We suggest that CorA may have different roles in different organisms. Such functional diversity is presumably a reflection of minor, but important structural differences within the CorA family that regulate the gating, substrate selection, and transport.

Divalent metal ions are highly essential for cellular processes in all life forms. The supply of these ions to cells and organelles at appropriate levels is therefore critical. As a result, organisms have developed transport systems for maintaining adequate concentrations of intracellular divalent cations. Distortions in these systems cause various pathological conditions in higher eukaryotes and are lethal for microorganisms. CorA is a family of divalent cation transporters that is found in most bacteria and archaea (1,2). This family has been studied extensively at a functional level using homologues from Escherichia coli, Salmonella typhimurium, Methanococcus jannaschii, and Haemophilus influenza as model proteins (3)(4)(5)(6)(7)(8). These studies have shown that CorA is able to transport Mg 2ϩ , Co 2ϩ , and Ni 2ϩ with the binding affinities of 15-20, 20 -40, and 200 -400 M, respectively (4,8). However, because Co 2ϩ and Ni 2ϩ are trace elements, the required concentrations are considered nonphysiological for these organisms, thus leaving Mg 2ϩ as the main substrate of CorA. Nevertheless, the rather similar affinity of CorA for both Mg 2ϩ and Co 2ϩ per se indicates that it is not very selective, as each of these ions can readily inhibit the uptake of the other one via CorA.
Rather recently, crystal structures of CorA from the hyperthermophilic organism Thermotoga maritima were determined (9 -11). Although some metal binding sites were identified and important structural features of the protein were revealed, the molecular mechanism for the ion transport through CorA remained unsolved. This is despite further functional characterizations of the T. maritima CorA (TmCorA) 3 (12,13).
Based on the amino acid sequence and phylogenic analyses, the CorA family can be divided into at least two rather distinct subgroups (14). In this classification, CorAs from E. coli, S. typhimurium, and H. influenza belong to one group (subgroup B), whereas TmCorA belongs to the other group (subgroup A), which is distinguished largely by CorAs originating from extremophilic bacteria. The latter opens up possibilities to find new primary roles of CorA, as the physiological conditions for many of the organisms from subgroup A are significantly different from those of E. coli and S. typhimurium, for example. Together with the fact that CorA has the ability to bind multiple substrates, one could speculate that there are differences in the physiological roles of CorA between the two subgroups. In other words, is it possible that a CorA from subgroup A has other primary substrates than Mg 2ϩ ? If this would be the case, it could be a plausible explanation for why functional studies of TmCorA using Mg 2ϩ have failed to describe how CorA transports its substrate across the membrane (12,13). In the crystal structure of TmCorA, both Mg 2ϩ and Co 2ϩ were found at the metal-binding sites (9). Interestingly, in TmCorA crystals, Co 2ϩ could outcompete Mg 2ϩ for the metal binding sites at a four times lower concentration (9). This was the first indication that TmCorA might have a higher affinity for Co 2ϩ over Mg 2ϩ , hence suggesting Co 2ϩ as its primary substrate. In the present study, we have explored the substrate preference of TmCorA by performing in vitro and in vivo studies. Our data clearly show that TmCorA is specifically thermostabilized and undergoes structural changes through interactions with Co 2ϩ but not Mg 2ϩ . Moreover, we show through in vivo competition studies that TmCorA can select Co 2ϩ over Mg 2ϩ for transport at Ͼ100 times lower concentrations. We also show that in contrast to E. coli CorA (EcCorA) and S. typhimurium CorA (StCorA), TmCorA is unable to regulate the Mg 2ϩ uptake and intracellular Mg 2ϩ concentration. Altogether, our data show that TmCorA is a highly selective Co 2ϩ transporter and indicate that Co 2ϩ , but not Mg 2ϩ , is the primary substrate of TmCorA. Hence, this is the first study indicating functional diversity within the CorA family, which reflects important structural differences among the CorA proteins.

EXPERIMENTAL PROCEDURES
Cloning of CorA Homologues-The T. maritima corA gene was cloned into a pBAD vector as described previously (9). The E. coli corA gene was cloned into a pBAD vector (Invitrogen) according to the manufacturer's protocol. The gene was amplified by PCR using Phusion Hotstart II (Finnzyme) and the following primers: forward, 5Ј-AGCGCGCCTCGAGCTACC-AGCCAGTTCTTGCGCTTAA-3Ј; and reverse, 5Ј-GCGCGC-CATGGGTCACCATCATCATCATCA-3Ј. The PCR-amplified gene was inserted in the pBAD vector after 3 h of cleavage with NcoI and XhoI (New England Biolabs), followed by 16-h ligation with T4 DNA ligase (New England Biolabs). All reactions were performed at room temperature, except for the digestion, which was performed at 37°C. Both TmCorA and EcCorA contained the MHHHHHHSSGVDLGTENLYFQSM sequence at the N terminus.
Thermostability of TmCorA in the Presence of Different Metals-Protein was diluted to a concentration of about 0.5 mg/ml. The solution was divided in 50-l aliquots, and 1 l of stock solution of different metals was added. The samples were incubated at room temperature for about 30 min before heating in a thermocycler (Agilent Technologies). Samples were incubated for 10 min at each temperature point. To remove precipitated protein, the samples were applied on a 96-well filter-plate (0.65 m) (Millipore). The yield of the filtered protein was analyzed by SDS-PAGE. Protein-bands were Coomassie-stained using Simply Blue Safe Stain (Invitrogen).
Thermal Aggregation Shift Assay-Purified TmCorA at 0.2 mg/ml was mixed with various concentrations of Co 2ϩ or Mg 2ϩ and aliquoted in a final volume of 50 l into a clearbottom, 384-well plate (Nunc). The mixture was covered with 45 l of mineral oil (Sigma-Aldrich) to avoid evaporation. The mixtures were then gradually heated from 25°C to 80°C at 1°C/min using StarGazer-384 (Harbinger Biotech), and the data were analyzed with Bioactive software (Harbinger Biotech) (15).
Co 2ϩ Transport Assay-The E. coli MG1655 with knockedout corA (⌬E. coli) (National Institute of Genetics, Japan) was transformed with corA from T. maritima in pBAD. The wildtype E. coli MG1655 and the ⌬E. coli, both containing empty pBAD, were also used as positive and negative controls, respectively. Cultures were then grown overnight at 37°C in Luria Broth (LB) medium (ForMedium, UK) supplemented with 100 g/ml ampicillin. The overnight cultures were then diluted 50 times with LB medium supplemented with 0.02% (w/v) L-arabinose and 100 g/ml ampicillin, followed by incubation at 37°C for 3 h. Cells were harvested by centrifugation at 3000 ϫ g for 5 min at room temperature. Pellets were washed twice with 10 ml N buffer (5 mM KCl, 7.5 mM (NH 4 ) 2 SO 4 , 0.5 mM K 2 SO 4 , 1 mM KH 2 PO 4 , and 0.1 M Tris (pH 7.4)). The washed pellets were then resuspended in N buffer containing Co 2ϩ at various concentrations to a final A 600 of 0.15. The mixture was incubated 10 min at 37°C. Three parts LB medium containing the same concentrations of Co 2ϩ was added into each reaction and incubated at 37°C for another 3 h. The final A 600 was recorded and analyzed in comparison with the starting A 600 . The competition assays were done similarly by including Mg 2ϩ or Mn 2ϩ .
Mg 2ϩ Transport Assay-The S. typhimurium strain MM281, depleted from all Mg 2ϩ transporters, was transformed with corA from T. maritima and E. coli, respectively. MM1278, a special strain of MM281 that encodes StCorA, as well as MM281 transformed with empty pBAD, were also used as positive and negative controls, respectively. Cultures were then grown overnight at 37°C in LB medium supplemented with 100 g/ml ampicillin, 50 g/ml kanamycin, 34 g/ml chloramphenicol, and 100 mM MgSO 4 . Cells were harvested by centrifugation at 3000 ϫ g for 5 min at room temperature and then washed three times with LB medium. The washed pellets were then resuspended in LB medium supplemented with 0.02% Larabinose and antibiotics as well as Mg 2ϩ at various concentrations to a final A 600 of 0.1. The cultures were grown at 37°C for 10 h and the A 600 was recorded every 2 h. The competition assays were done similarly by including Co 2ϩ .
Fluorescence Measurements-The protein was diluted to 6.4 M in a desalting buffer, and the tryptophan fluorescence was measured by excitation at 283 nm and scanning the emission between 295 and 425 nm (Cary Eclipse). The tryptophan fluorescence-quenching effects of Mg 2ϩ , Ni 2ϩ , and Co 2ϩ were determined by titration of each of these metal ions.

SDS-PAGE and Western Blot Analysis of the Whole Cells-
The volume of cell cultures were adjusted to reach the same A 600 for all the samples. Then, 10 l of total cells was prepared and loaded on 4 -12% gels according to the manufacturer's recommendations (Invitrogen). The protein bands were then transferred into nitro cellulose membranes using I-Blot according to the manufacturer's protocol (Invitrogen). The bands were then detected using HRP-conjugated His probe and West-Pico according to the manufacturer's protocol (Pierce).

Thermostability of TmCorA Is Improved with the Substrate-
To test the thermostability of TmCorA, the isolated protein was gradually heated from 25°C to 95°C, and samples were taken for analysis at every 10°C increment. The precipitated (i.e. destabilized) protein was removed by filtration. The stability of the protein at each temperature was monitored through analyzing the level of retained protein with SDS-PAGE. In the absence of divalent metals, TmCorA was totally stable up to 65°C (Fig. 1). At 75°C, almost 50% of the protein had precipitated, and complete denaturation was observed at 85°C. When 1 mM MgCl 2 was added to the buffer, TmCorA thermostability was improved, and the protein was almost totally stable at 85°C. However, in the presence of 1 mM CoCl 2 or NiCl 2 , TmCorA was almost fully stable at 95°C. These results indicate that Co 2ϩ and Ni 2ϩ are more competent than Mg 2ϩ to keep TmCorA stable at temperatures that are physiologically relevant. In the presence of cobalt hexamine, the traditional inhib-itor of CorA, no significant changes in the protein stability in comparison with the control were observed (Fig. 1).
To verify whether the stabilizing effect of the metal ions were relevant and specific, we also tested thermostability of CorA in the presence of non-substrate metal ions. Ca 2ϩ showed the same stabilizing effect as Mg 2ϩ , whereas Fe 2ϩ and Zn 2ϩ showed a negative effect on the stability of TmCorA, already causing protein destabilization at 65°C or even lower temperatures. Thus, the thermostability of CorA was presumably specifically improved with the metal substrates and not generally by any divalent metal ions.
Co 2ϩ and Ni 2ϩ Specifically Stabilize TmCorA at Low Concentrations-A metal ion concentration of 1 mM is a relatively high substrate concentration and may not reflect the physiologically relevant concentrations for T. maritima. Therefore, the stabilizing effects obtained by the metals at this concentration could be consequences of nonspecific binding to the protein. To explore how specific each substrate thermostabilizes the protein, a titration of metal ions from 10 to 1000 M was performed at 85°C because at this temperature, all three known substrates (Mg 2ϩ , Co 2ϩ , and Ni 2ϩ ) could stabilize the protein. Our analysis showed that below 1 mM Mg 2ϩ (same as the non-substrate ion Ca 2ϩ ) was no longer able to keep the protein stable at 85°C (Fig. 2). On the other hand, thermostabilization could be detected already in the presence of 80 M of Co 2ϩ or Ni 2ϩ , and a major fraction of the protein remained stable in the presence of Ն200 M CoCl 2 or NiCl 2 (Fig. 2). To further investigate the specificity of the stabilizing effect of Co 2ϩ and Ni 2ϩ on TmCorA, we performed a competition study in which titration of Co 2ϩ and Ni 2ϩ was done in the presence of 1 mM MgCl 2 . Interestingly, Mg 2ϩ showed no negative effect, even at 10 times the concentration of Co 2ϩ or Ni 2ϩ (supplemental Fig. S1). These data indicate that in an environment containing a mixture of Mg 2ϩ , Ni 2ϩ , and Co 2ϩ , either of the last two can compete out Mg 2ϩ for binding to TmCorA, even at much lower concentrations.
The thermostability effect of Co 2ϩ as well as Mg 2ϩ on TmCorA was further explored by thermal aggregation shift  assay. This assay is more sensitive and can detect protein destabilization and aggregation in real time. However, the upper temperature limit for this assay is 80°C, which makes it less suitable for such a hyperthermostable protein. Nevertheless, at temperatures below 80°C, this assay clearly shows that at ϳ60 M Co 2ϩ , already the aggregation temperature (T agg ) is shifted by ϩ1°C and that the T agg continues to increase dramatically by increasing Co 2ϩ concentration until complete stabilization is reached (Fig. 3). On the other hand, Mg 2ϩ could only moderately cause a shift in T agg without reaching complete stabilization. The results obtained with the differential thermal shift assay were thus in full agreement with the other results shown in Figs. 1 and 2.
TmCorA Undergoes Structural Changes upon Specific Binding of Co 2ϩ -At this point, our results suggest dominant and specific binding of Co 2ϩ and Ni 2ϩ to TmCorA, whereas Mg 2ϩ appeared to bind much weaker. Nevertheless, these experiments do not provide any information on whether the three substrates bind to and stabilize the protein in a similar manner. Therefore, we also performed a ligand dose tryptophan fluorescence quenching experiment (16). Surprisingly, Mg 2ϩ did not induce any quenching of the fluorescence signal from TmCorA, even at a concentration of 5 mM, whereas Co 2ϩ quenched the signal already at ϳ10 M (Fig. 4), indicating structural changes in the protein induced upon Co 2ϩ binding. The titration with Ni 2ϩ shows that this metal ion has some effect on the structure of TmCorA, but this effect is much smaller than for Co 2ϩ . Thus, the ligand dose tryptophan fluorescence quenching clearly indicated that Co 2ϩ could specifically interact with TmCorA, resulting in structural changes already at low concentrations.
TmCorA Is a Highly Potent Co 2ϩ Transporter-All the results discussed above have shown specific and stabilizing binding of Co 2ϩ to TmCorA. To investigate whether this specific binding is related to the transport activity, the ability of TmCorA to transport Co 2ϩ was examined. For these experiments, we used a corA-less mutant of the E. coli MG1655 strain (⌬E. coli), which is highly resistant to Co 2ϩ . Thus, ⌬E. coli was incubated with Co 2ϩ at various concentrations for 10 min, followed by 3 h of growth. Using this assay, we could monitor the Co 2ϩ uptake of the strain as a direct measure of the growth activity, i.e. the more Co 2ϩ uptake, the lower the growth activity. In the absence of any corA, ⌬E. coli remained resistant toward Co 2ϩ and retained ϳ80% of its growth activity at concentrations up to 500 M (data not shown). Once the ⌬E. coli was transformed with the corA from T. maritima, it became highly sensitive toward Co 2ϩ and already lost ϳ50% of its growth activity in the presence of 1 M Co 2ϩ (Fig. 5). Thus, TmCorA could readily transport Co 2ϩ at submicromolar concentrations and thereby reduce the ⌬E. coli growth activity. The growth activity was completely abolished at 100 M Co 2ϩ , indicating total toxification of ⌬E. coli cells (data not shown).
EcCorA and StCorA can transport both Mg 2ϩ and Co 2ϩ , and each of these ions can inhibit the uptake of the other one via CorA. We attempted to reproduce this inhibitory effect on the EcCorA using our Co 2ϩ transport assay, where the wild-type  E. coli strain MG1655 with the endogenous CorA was used. As expected, the wild-type E. coli showed Co 2ϩ sensitivity through its endogenous CorA by showing reduction in growth activity. However, the growth activity was restored by the addition of Mg 2ϩ at concentrations similar to Co 2ϩ or slightly higher (supplemental Fig. S2). In fact, Mg 2ϩ was able to completely restore the E. coli growth activity already at concentrations three to four times higher than that of Co 2ϩ (data now shown). This result was in good agreement with the results from a previous study (8). We then examined the ability of TmCorA to select Co 2ϩ in the presence of Mg 2ϩ or Mn 2ϩ in ⌬E. coli. The data show that TmCorA was extremely selective in Co 2ϩ uptake because neither Mg 2ϩ nor Mn 2ϩ could restore the ⌬E. coli growth activity, even at concentrations Ͼ100 times higher than those of Co 2ϩ (Fig. 6). It is noteworthy that Mn 2ϩ was able to show some inhibitory effect, whereas Mg 2ϩ was almost completely incapable of outcompeting Co 2ϩ . In a previous study, Mn 2ϩ showed to be more competent than Co 2ϩ to inhibit the Mg 2ϩ uptake by StCorA (8).
Mg 2ϩ Is a Poor Substrate for TmCorA-Finally, we performed a Mg 2ϩ transport assay using the S. typhimurium MM281 strain, which is depleted of all Mg 2ϩ transporter genes. This strain can grow only in the presence of Ն10 mM Mg 2ϩ , i.e. the Mg 2ϩ content of the LB medium is not sufficient for the growth. MM281, when expressing TmCorA (MM281 TmCorA ), could slowly grow in LB medium without supplementary Mg 2ϩ and seemed to reach a growth steady state (at A 600 of ϳ0.6) after 10 h (Fig. 7A). Addition of 1 mM Mg 2ϩ to LB medium increased the growth activity of MM281 TmCorA ; however, after 6 h the cell density declined and reached an A 600 of ϳ0.6 after 10 h (Fig.  7B). Interestingly, when 10 mM Mg 2ϩ was added to the LB medium, the MM281 TmCorA reached an A 600 of ϳ0.6 after only 4 h, and no further growth was observed after 4 h (Fig. 7C). These results indicate a slow transport of Mg 2ϩ through TmCorA, as increasing supplementary Mg 2ϩ increased the growth of MM281. However, the Mg 2ϩ uptake by TmCorA seems to be uncontrolled and not regulated because it could stop the growth and even cause cell death, indicating toxicity to the cell. On the other hand, MM281 expressing EcCorA (MM281 EcCorA ), although showing a slow start in growth, had a very stable growth rate regardless of Mg 2ϩ concentration in the medium (Fig. 7, A and B). Even with an additional 10 mM Mg 2ϩ in the medium, MM281 EcCorA continued a stable growth, indicating a tight regulation of Mg 2ϩ uptake through EcCorA as expected (Fig. 7C). Notably, the initial slow growth rate of MM281 EcCorA was significantly elevated after 8 h. An expression analysis of both CorA proteins showed a strong expression of TmCorA, already reaching saturation after 4 h, whereas the expression of EcCorA could first be detected after 10 h (Fig.  7D). The analysis of the expression levels could clearly show that EcCorA is significantly more competent than TmCorA in transporting Mg 2ϩ and, especially, in regulating the Mg 2ϩ import and homeostasis. Hence, these results underscore that TmCorA, unlike EcCorA, has a very slow Mg 2ϩ uptake activity and is incapable of controlling the intracellular Mg 2ϩ concentration.

DISCUSSION
Despite the ability of various CorA homologues to transport Co 2ϩ , CorA has generally been considered as the primary Mg 2ϩ transporter of all organisms in which it is encoded. This is a result of extrapolating the functional characterization of a few organisms, such as S. typhimurium and E. coli. The reason for this generalization is that the micromolar concentration of Co 2ϩ that is required for a transport to occur is not physiologically relevant for these organisms, which led to excluding Co 2ϩ as the physiological substrate of CorA. Nevertheless, when per- forming functional analyses of CorA from T. maritima using Mg 2ϩ as a substrate, the results were not easily explainable and are rather inconclusive. For example, the function of TmCorA was investigated by focusing on its Mg 2ϩ transport activity using numerous mutants of the protein that were designed based on the crystal structures (13,17). In addition, unexpected behavior such as gain of function was observed (13), which means that some mutants were more capable in transporting Mg 2ϩ than the wild-type TmCorA and became as active as the wild-type EcCorA. In a recent in vivo study, when monitoring the Mg 2ϩ -transport activity of TmCorA, the host strain showed sensitivity to Mg 2ϩ as if the metal ion was toxic (17). Thus, despite the available crystal structures as well as well performed functional studies, it is yet unclear how TmCorA can transport its substrate across the membranes and how its function is regulated. The reason could simply be that the right substrate was not used in these studies.
The capacity of Co 2ϩ to replace Mg 2ϩ at the metal-binding sites at four times lower concentration (9) was the first indication of TmCorA to be more selective for Co 2ϩ rather than for Mg 2ϩ . However, this conclusion could have been biased because Ca 2ϩ , which is not a substrate of CorA, could also bind to the same site (11). Also, the actual role of these metal-binding sites is not yet known. Nevertheless, our further exploration of this possibility has clearly shown that the TmCorA-Co 2ϩ interactions are highly specific. Among the many divalent cations tested, Co 2ϩ and Ni 2ϩ were the only ions that could increase the thermostability of TmCorA at relatively low concentrations ( Fig. 1-3), but only Co 2ϩ showed specific interactions with TmCorA (Fig. 4). Various studies have already shown that thermostability is a good measure for assaying the substrate or ligand binding (18 -20). Our thermostability data here were complementary to the in vivo transport studies, which showed that not only is TmCorA capable of transporting Co 2ϩ at submicromolar concentrations, but it also can efficiently select Co 2ϩ over other divalent ions such as Mg 2ϩ (Fig. 5 and 6). The poor ability of Mg 2ϩ to stabilize TmCorA was reflected by the poor capability of TmCorA to take up Mg 2ϩ and its inability to regulate the uptake (Fig. 7), which can explain the unusual toxicity observed in (17). Altogether, it means that in an environment with mixed metal concentrations, Co 2ϩ is the preferred substrate of TmCorA as it is taken up much faster and at much lower concentrations than Mg 2ϩ . This leads to the question regarding the physiological relevance of Co 2ϩ uptake and the need for Co 2ϩ .
The hyperthermophilic T. maritima is a rod-shaped, Gramnegative bacterium found in geothermal heated marine sediment, and it can live at temperatures between 55-90°C, with an optimum growth activity at 80°C (21). In the natural habitat of T. maritima, the Mg 2ϩ concentration is relatively low, and it decreases with increasing temperatures (22). In contrast, in the same environment the level of Co 2ϩ is increased to even micromolar concentrations (22). Thus, T. maritima has a very different habitat than e.g. E. coli, even based on the concentration and availability of metal ions. In addition, there are reports of cytosolic enzymes from this organism that prefer Co 2ϩ as the cofactor, rather than Mg 2ϩ or Zn 2ϩ , which are the most common divalent cations in E. coli, for example (23)(24)(25)(26). Co 2ϩ has also been shown to thermostabilize proteins from T. maritima (Refs. 23, 24 and Figs. 1-3). Therefore, T. maritima most likely has a higher intracellular Co 2ϩ concentration than E. coli, for example, as it has a higher requirement for Co 2ϩ . The intracellular Co 2ϩ concentration of T. maritima is not known. However, it must have a very tightly regulated Co 2ϩ -uptake system and/or high intracellular Co 2ϩ content because in laboratories it is usually grown in a medium containing nearly 1 mM Co 2ϩ (27,28), and it can also tolerate Ͼ1 mM Co 2ϩ (29). Hence, the Co 2ϩ requirement for T. maritima and its ability to grow in environments containing high Co 2ϩ concentrations points to the necessity of a selective and tightly regulated uptake system for Co 2ϩ . Whether or not CorA is the main system responsible for regulating the Co 2ϩ uptake in T. maritima cannot be established at this moment.
It is also unknown how well the Co 2ϩ uptake is regulated by TmCorA. In our in vivo studies, the ⌬E. coli already becomes sensitive toward Co 2ϩ at very low concentrations (Fig. 5). It would be reasonable to ask why TmCorA does not regulate the Co 2ϩ transport by closing and thus rescuing the cell. However, this could be explained by the differences between E. coli and T. maritima in Co 2ϩ sensitivity/tolerance. The above-mentioned necessity of Co 2ϩ for the stability and activity of the T. maritima cellular enzymes would most certainly result in a higher intracellular Co 2ϩ concentration compared with E. coli. For example, the intracellular concentration of Cu 2ϩ in T. maritima has been measured to be 15 M (29), which is unusually high and for E. coli is considered toxic. Another important parameter to consider when studying the activity of proteins from thermophilic organisms is the temperature. Isolated proteins from thermophilic organisms typically display optimal activities at elevated temperatures (30 -33). Therefore, to better understand the biological role of TmCorA, one should perform the studies using corA-deleted T. maritima strains or alternatively create assays mimicking conditions similar to the environment of T. maritima.
The high selectivity of TmCorA for Co 2ϩ with its poor ability to transport and regulate Mg 2ϩ uptake is in contrast to the function of EcCorA as well as StCorA. This is most likely a direct reflection of the structural differences between TmCorA and the other two CorAs, and perhaps more generally also between subgroups A and B (14).
Although the results from two previous studies have indicated EcCorA to be a tetramer (34,35), a pentameric structure for all CorAs is most probable, considering that even the crystal structure of a distant homologue of CorA, ZntB, was determined as a pentamer (36). However, only a minor difference, such as the nature or position of the side chains in the active or substrate-binding site, can make a major difference, as has been reported in other cases (37,38). Therefore, it is important to avoid unreasonable extrapolations and excessive generalization of the CorA function before more functional as well as structural data from various subgroups is obtained. Understanding the functional differences between various CorA proteins at the molecular level is important for understanding ion selectivity, gating, and transport mechanisms of the different CorA subgroups. This information should also be valuable for understanding the evolutionary aspects in the differences between the CorA proteins.
This study is the first report of a CorA protein that is highly selective for Co 2ϩ and incapable of Mg 2ϩ transport and regulation, suggesting new roles for this transporter family.