Functional analysis of the conserved hydrophobic gate region of the magnesium transporter CorA.

The Leu294 residue in the cytoplasmic neck of Thermotoga maritima CorA is considered to be the main gate for Mg2+ transport. We created three site-directed mutants at this position: in the Leu294Asp and Leu294Gly mutants we observed a defect in closing of the pore, while in the Leu294Arg mutant not only gating, but also the regulation of Mg2+ uptake was affected. Our results confirmed the importance of the Leu294 for gating of Mg2+ transport and in addition revealed the influence of the charge and structural features of the amino acid residues on the gating mechanism.

Introduction

Magnesium transporters of the CorA family are widely distributed amongst Eubacteria and Archaea. The CorA gene encodes a constitutively expressed integral membrane protein [1]. Previous investigations showed that it is essential for bacterial growth [2] but standard rich media provide sufficient magnesium concentrations for growth [3].

In the last years three crystal structures of Thermotoga maritima CorA have been published [4-6]. All showed that the transporter exists in a pentameric form, consisting of a cytosolic funnel shaped part, linked to the transmembrane region by long α7 helices (Fig. 1a, b). The pore is formed by the first trans-membrane helices and surrounded by the second trans-membrane helices, which anchor the complex in the membrane and end in a highly conserved positively charged motif (KKKKWL) called “basic sphincter”. In the cytoplasmic neck of the pore a hydrophobic ring is created by residues Leu294 and Met291, surrounded by the aforementioned basic sphincter (Fig. 1c). This concentration of positive charges and the significant conservation of the bulky hydrophobic residues at positions 291 and 294 in the CorA protein family is considered to be of high importance for gating of Mg2+ ions. Opening and closing of the gate is most probably regulated by interaction of the Mg2+ ion with a divalent cation sensing site (DCS), placed between Asp89 in the α3 helix in the N-terminal part of one monomer and Asp253 of the α7 helix of the adjacent monomer [4-6] (Fig. 1b). A second DCS site, involving residues Glu88 and Asp175, was identified by Eshaghi et al. [5] and Payandeh et al. [6] (Fig. 1b).

Fig. 1

Structure of the TmCorA Mg2+ transporter. (a) Single monomer: green — transmembrane domains TM1 and TM2, blue — α7 helix, red — N-terminal domain. (b) Side view of the homopentamer: cyan — DCS sites 1 and 2. (c) View from the top: blue — basic sphincter, yellow — aromatic ring, green — the gate forming residues Leu294 and Met291.

The exact gating mechanism of the TmCorA transporter could not been revealed yet. It has been proposed that binding of the Mg2+ ions to the DCS sites evokes a structural rearrangement of the cytosolic domain causing positively charged residues of the basic sphincter to close the pore by drawing the negative charges away from the middle of the pore, and thereby preventing the positively charged Mg2+ ion to pass. Removal of Mg2+ from the DCS sites would cause a movement of the N-terminal domain, resulting in drawing the basic sphincter away from the neck of the pore and allowing the Mg2+ ion to pass [4]. According to the recently performed 110-ns molecular-dynamics simulations, based on the CorA structures published by Eshaghi et al. [5] and Payandeh et al. [8], the binding or unbinding of Mg2+ ions to the DCS sites evokes structural rearrangements of the cytosolic domains and the α7 helices transmit these changes to the gate region causing closing or widening of the pore [7].

Leu294 in the hydrophobic ring is the critical residue for Mg2+ gating. It creates a strong energetic barrier for ion permeation and probably controls the movement of Mg2+ ions indirectly through the movement of water. According to Payandeh et al. [8], not only an energetic, but also a mechanic barrier can influence the uptake of Mg2+ and “opening sensitivity” of the transporter. To investigate this hypothesis in more detail, this mutational study was focused on Leu294 which was mutated to 15 different amino acids using random PCR mutagenesis. After a preliminary screen, three of these mutants representing different types of amino acids: positively charged hydrophilic arginine, negatively charged hydrophilic aspartic acid and small neutral glycine, were chosen for closer investigations.

Materials and methods

Bacterial strains, growth media and genetic procedures

Salmonella enterica serovar Typhimurium strain LB5010 was used as wild-type reference.

Salmonella enterica serovar Typhimurium strain MM281 (DEL485 (leuBCD)mgtB::MudJ;mgtA21::MudJ;corA45::mudJ;zjh1628::Tn10(cam) CamR, KanR, Mg2+ dependent) was kindly provided by M.E. Maguire. It lacks all major magnesium transport systems CorA, MgtA and MgtB and requires Mg2+ concentrations in millimolar range for growth.

Strains were grown in LB medium (10 g tryptone, 5 g yeast extract, 10 g NaCl per liter) with ampicillin (100 μg/ml). MM281 required addition of 10 mM MgCl2.

LB plates contained 2% Difco™ Agar Noble, minimizing possible Mg2+ contamination.

Plasmid constructs

The Thermotoga maritima CorA coding sequence was kindly provided by S. Eshaghi and used as template for PCR. The sequence was amplified using the following primers: TmCorwoSfw 5′-CGCGGATCCGAGGAAAA-GAGGCTGTCTGC-3′ and TmCorrev 5′-TCCCCCGGGTCACAGCCACTTCTTT-TTCTTG-3′. The 1035 bp PCR product was cut with BamHI and SmaI restriction enzymes and cloned into the pQE80L vector with an IPTG-inducible promoter.

Random PCR mutagenesis

In order to introduce various amino acid substitutions in CorA, an overlap extension PCR according to Pogulis et al. [9] was used.

Random mutagenesis of the Leu294 amino acid with the mutagenic forward primer 5′-GCGGTCTTCTTGATGTGTACCTTTCGAGTGTAAGTAACAAAACAAACGAAGTGATGAAGGTGNNNACCATCATAGCG-3′ and the reverse primer 5′-CGCTATGATGGTNNNCACCTTCATCACTTCGTTTGTTTTGTTACTTACACTCGAAAGGTACACATCAAGAA-GACCGC-3′ was performed with mutagenic PCR according to standard protocols. PCR product was cut with BamHI and PstI restriction enzymes and cloned into a BamHI and PstI digested pQE CorA construct. Correctly ligated constructs were identified by deletion of the BsaBI restriction site in the CorA gene, resulting in a silent mutation from a thymine to a cytosine. No additional mutations were found by sequencing.

Complementation assays on solid media

The CorA, MgtA, MgtB deletion strain MM281 was transformed with pQE80L constructs harbouring TmCorA or mutated versions thereof. Single colonies were inoculated in LB medium containing 10 mM MgCl2 and grown over night. The cultures were washed twice with 0.7% saline, adjusted to an OD600 of 0.1, and diluted 1:10, 1:100 and 1:1000 and spotted on LB plates containing 100 mM MgCl2 or on standard LB plates, both supplemented with different IPTG concentrations (0; 0.01; 0.1 mM). The cultures were incubated for 24 h at 37 °C.

Growth curves

Overnight cultures of MM281 cells transformed with the aforementioned plasmid constructs were grown in LB medium with 10 mM MgCl2, washed twice with LB and inoculated to an OD600 of 0.1 into liquid LB medium containing different concentrations of MgCl2 and IPTG.

Measurement of Mg2+ uptake by spectrofluorometry

For these measurements cells were grown in low-phosphate LB medium in order to minimize complexation of Mg2+ by an excess of phosphate, which might cause variation in Mg2+ concentration. The measurements have been performed as described previously [10].

PAGE and western blotting

Overnight cultures of MM281 cells transformed with the aforementioned plasmid constructs were washed twice with LB medium and diluted in fresh LB + Amp medium to an OD600 of 0.1. Expression was induced by adding 0.05 mM IPTG for 3 h. Equivalents of 3 ml culture with an OD600 of 0.5 were taken. The cell pellet was resuspended in 20 μl SDS-Laemmli buffer and 20 μl 8 M Urea and sonicated for 5 s. After 15 min centrifugation at 13,000 rpm, equal volumes of supernatant were loaded on 10% SDS/polyacrylamide gels, blotted and labelled with an antiserum against the 6xHis tag (Qiagen).

Results

There are several hydrophobic constrictions along the TmCorA ion conduction pathway. The narrowest one is formed by the highly conserved residues Leu294 and Met291 [4-6]. They are considered to be part of the potential “hydrophobic gate” [4,8]. To verify this hypothesis our mutations targeted the Leu294, we analysed mutants containing three different residues: positively charged arginine, negatively charged aspartic acid and small glycine.

Effect of positive charge at position 294 on gating of CorA: Leu294Arg mutant

In the growth complementation assay the Leu294Arg mutant showed similar growth as the wild-type TmCorA on plates without MgCl2 and with different IPTG concentrations, which means, that the change from hydrophobic leucine to the hydrophilic, positively charged arginine does not cause a dramatic change in the protein function (Fig. 2). While on plates containing 100 mM MgCl2 and 0.1 mM IPTG the wild-type TmCorA exhibited a growth defect, probably caused by Mg2+ overdose, the Leu294Arg mutant was still able to grow indicating decreased transport efficiency of this mutant (Fig. 2).

Fig. 2

Growth complementation assay of the gating mutants. S. typhimurium strain MM281 which lacks all three known magnesium transport systems was transformed with plasmids indicated. The pQE TmCorA and pQE80L empty constructs served as positive and negative control.

Mg2+ uptake measurements of the Leu294Arg mutant showed slower Mg2+ influx and lower steady-state values (~ 80% of wild-type values) (Fig. 3) suggesting that (i) the pore of the transporter is smaller than in wild-type TmCorA; (ii) the sensitivity of CorA to Mg2+ or the regulation of opening and closing of the pore was affected.

Fig. 3

Representative recordings of Mg2+ uptake. S. typhimurium strain MM281 which lacks all three known magnesium transport systems was transformed with plasmids indicated. The representative recordings show changes in fluorescence intensity of Mag-Fura-2 monitored over 10 min time after adding 10 mM MgCl2. The pQE TmCorA and pQE80L empty constructs served as positive and negative control.

The growth curves of cells transformed with TmCorA Leu294Arg mutant (Fig. 4a, b, c) support these hypotheses. In liquid medium supplemented with 0.05 mM IPTG the cells grew much slower than those with wild-type TmCorA (Fig. 4b), while the shapes of the curves remained similar, pointing to a lower Mg2+ transport capacity of this mutant. The same was true for cells growing in medium with 0.05 mM IPTG and 3 mM MgCl2 (Fig. 4c). We did not observe any indications of Mg2+ overdose, not even after 11 hour incubation (results not shown), suggesting that closing of the transporter remained tightly regulated. In medium without MgCl2 the cells with Leu294Arg mutant were unable to grow, exactly like those with wild-type TmCorA (Fig. 4a).

Fig. 4

Effect of gating mutations in different liquid medium. Growth curves of the S. typhimurium strain MM281 which lacks all three known magnesium transport systems, transformed with plasmids indicated. Cells were grown over night in LB amp medium containing 10 mM MgCl2. Cultures were diluted to an OD600 of 0.1 and grown over 5 h. The data were averaged from three independent experiments.

Effect of negative charge at position 294 on gating of CorA: Leu294Asp mutant

Cells transformed with TmCorA Leu294Asp mutant showed a growth defect on medium containing 0.1 mM IPTG as well as on media containing 0.1 mM IPTG and 100 mM MgCl2 in the growth complementation assay (Fig. 2). This suggests either a higher Mg2+ transport capacity, which might be caused by a defect in the regulation process of the opening and closing of the channel or in the closing itself, leading to Mg2+ overdose during the 24 hour incubation and consequently to a toxic effect of this protein at high expression levels.

Mg2+ uptake of Leu294Asp mutant increased slowly and did not reach a steady-state level within a time period of 600 s (Fig. 3), suggesting that the pore is neither completely open nor can it close properly, causing slow, constant Mg2+ uptake, leading finally to a Mg2+ overdose. This mutation apparently causes a defect in the regulation of opening and closing or/and in the closing process itself, which was corroborated by the growth curves (Fig. 4a, b, c): neither cells transformed with the TmCorA nor cells transformed with the Leu294Asp mutant were able to grow in medium without MgCl2 (Fig. 4a). In medium containing 0.05 mM IPTG we observed rapid growth but after 4 h a steady-state phase was reached, probably due to a Mg2+ overdose (Fig. 4b). In medium containing 0.05 mM IPTG and 3 mM MgCl2 the shape of the Leu294Asp mutant growth curve differed from that of the wild-type TmCorA: the mutant cells grew significantly slower, reaching only an OD600 of 0.6 (~ 50% of the wild-type CorA) (Fig. 4c), which indicates a Mg2+ overdose caused by the constant Mg2+ uptake depicted in Fig. 3.

Effect of removal of side chain at position 294 on gating of CorA: Leu294Gly mutant

As expected, the Leu294Gly mutant was even more Mg2+ sensitive than the Leu294Asp variant. The cells did not grow on plates containing 0.1 mM IPTG suggesting a defect in the closing and/or in the regulation of the transporter and the growth on plates containing 0.01 mM IPTG points to a higher Mg2+ uptake capacity compared to the wild-type TmCorA (Fig. 2). In case of the Leu294Gly mutant the cells also grew on plates without additional MgCl2 or IPTG (Fig. 2) which might be due to not tight enough regulation of the used plasmid, leading to leaky expression of the CorA gene in the absence of IPTG. This was tested by determination of expressed CorA levels using western blots (Fig. 6). The low expression level of the wild-type TmCorA protein was not sufficient to provide cells with enough Mg2+ in a low Mg2+ medium, but in the case of the Leu294Gly mutant cells are able to take up enough Mg2+ to survive. This observation supports the suggested increased Mg2+ transport ability of the Leu294Gly mutant, which was also observed in Mg2+ uptake measurements (Fig. 3). The Mg2+ influx was as fast as in cells transformed with wild-type TmCorA, but it did not reach a steady-state, which indicates a defect in the closing of the transporter.

Fig. 6

Western blot analysis of whole cell samples. S. typhimurium strain MM281 which lacks all three known magnesium transport systems was transformed with plasmids indicated. The synthesis of CorA protein and variants thereof was induced by addition 0.05 mM IPTG for 3 h.

In liquid medium containing 0.05 mM IPTG we observed fast growth of the Leu294Gly cells, followed by a steady-state after 4 h (Fig. 4b). In medium supplemented with 0.05 mM IPTG and 3 mM MgCl2 the cells stopped to grow already after 3 h, reaching only ~ 40% of the wild-type level, and contrary to the wild-type cells, they began to die (Fig. 4c). This effect can be related not only to an increased Mg2+ transport ability but also to a closing defect of the transporter. Increased Mg2+ transport ability of the Leu294Gly mutant (Fig. 3) allowed growth of the cells also in medium without any supplements (Fig. 4a), suggesting eminent structural changes caused by this mutation.

In order to confirm that the higher Mg2+ uptake capacity of this mutant is due to the Leu294Gly mutation, we used Co(III)hexamine, a known blocker of the CorA transporter [8,11,12]. We compared the wild-type Salmonella typhimurium strain LB5010, which has all major Mg2+ uptake systems (CorA, MgtA, MgtB) with strain MM281 deficient for these genes, transformed with wild-type TmCorA and with TmCorA Leu294Gly mutant. In liquid LB medium (Fig. 5) strain LB5010 grew fast, whereas strain MM281 containing the Leu294Gly mutation showed reduced growth, and strain MM281 with wild-type TmCorA did not grow at all. After addition of 3 mM Co(III)hexamine, growth of LB5010 cells and MM281 cells containing wild-type TmCorA remained unchanged, whereas growth of MM281 cells with the Leu294Gly mutation was completely inhibited proving that growth of cells under low Mg2+ conditions is the result of this mutation.

Fig. 5

Effect of Co(III)hexamine, inhibitor of the CorA-mediated magnesium transport. Growth curves of the S. typhimurium strains LB5010 (wild-type strain) transformed with pQE80L parental vector and MM281 (corAΔ, mgtAΔ, mgtBΔ) transformed with pQE80 TmCorA and pQE80 TmCorA L294G in the presence or absence of cobalt hexamine. The data were averaged from three independent experiments.

Discussion

Gating of the TmCorA magnesium transporter is a complex process, involving several events resulting in structural rearrangements of the pentamer: binding of Mg2+ ions to the regulatory DCS site(s) with consequent rearrangements of α-helices in the N-terminal domain, the pore-forming α7 helices and the basic sphincter are considered to lead collectively to conformational changes in the gating region. Leu294 lies in the 15 Å long region termed MM stretch (spanning residues Met291 to Met301) proposed to play a central role in controlling the ionic conduction profile by representing both a steric as well as an electrostatic bottleneck for Mg2+ translocation [7]. This region involves a series of hydrophobic residues among which Leu at position 294 plays a critical role in the gating process. The effects of the Leu294 mutation to Asp and Gly may be mechanistically and structurally explained by local structural and electrostatic changes due to introduction of a negatively charged or a small amino acid residue with no side chain.

In case of glycine, the absence of the side chain could on one hand cause a wider opening of the pore and on the other impair its complete closing. This is in agreement with results of Payandeh et al. [8], who showed that exchange of Leu294 to a hydrophobic amino acid with a smaller side chain (Ile, Val or Ala) results in an increased ability of cells to grow on media supplemented with low MgCl2 concentrations.

In case of introduction of an aspartic acid at position 294, the negatively charged side chains coming from five protomers repulse each other and thereby maximise the distance between them, leading to the incapacity of the transporter to properly close and to maintain a stable magnesium concentration in the cell. Furthermore, slow magnesium uptake of this mutant indicates that Mg2+ ions are probably trapped in the ring of negative charges.

The effects observed with the Leu294Arg mutant cannot only be explained by local changes. Concentration of positively charged arginine residues in this region might have two opposite effects: repulsion of the positively charged residues causing a local structural distortion, and/or repulsion of the Mg2+ ion and hindering its passing through the pore. Both effects can explain the slow-down of Mg2+ transport, but not the low steady-state values observed in Mg2+ uptake measurements, which indicate a change in the regulation of Mg2+ influx. According to Payandeh et al. a valine mutation at position 294 can alter Mg2+-binding properties of the transporter [8]. Due to the distance of ~ 65 Å between the Leu294 and the DCS sites a direct interaction seems impossible. Since the loss of Mg2+ ions from DCS sites can induce conformational rearrangements of the magnesium-binding domains, transmitted to the gate region by the α7 helices and leading to changes allowing ions to pass through [7,8], also a reverse process might be possible. Certain mutations in the gate region may evoke structural rearrangements of the α7 helices leading to conformational changes in the DCS sites, resulting in their changed affinity to Mg2+ ions.

Concluding remarks

In summary, we confirmed the importance of the Leu294 as a mechanical barrier to Mg2+ permeation. We furthermore propose an additional function of this residue in the regulation of Mg2+ uptake. Since Mg2+ transport involves multiple coordinated structural movements, further structural and functional analyses will be required to fully understand the molecular basis of this process.