Biochemical Journal

Research article

SLC41A1 Mg2+ transport is regulated via Mg2+-dependent endosomal recycling through its N-terminal cytoplasmic domain

Tyler Mandt, Yumei Song, Andrew M. Scharenberg, Jaya Sahni

Abstract

SLC41A1 (solute carrier family 41, member A1) is a recently described vertebrate member of the MgtE family of Mg2+ transporters. Although MgtE transporters are found in both prokaryotic and eukaryotic organisms, and are highly conserved, little is known about the regulation of their Mg2+ transport function. In the present study, we have shown that endogenous SLC41A1 transporter expression is post-transcriptionally regulated by extracellular Mg2+ in TRPM7 (transient receptor potential cation channel, subfamily M, member 7)-deficient cells, suggesting that SLC41A1 transporters underlie a novel plasma membrane Mg2+ transport function. Consistent with this conclusion, structure–function analyses of heterologous SLC41A1 transporter expression demonstrate that SLC41A1 transporters exhibit the same plasma membrane orientation as homologous bacterial MgtE proteins, are capable of complementing growth of TRPM7-deficient cells only when the Mg2+ transporting pore is intact, and require an N-terminal cytoplasmic domain for Mg2+-dependent regulation of lysosomal degradation and surface expression. Taken together, our results indicate that SLC41A1 proteins are a central component of vertebrate Mg2+ transport systems, and that their Mg2+ transport function is regulated primarily through an endosomal recycling mechanism involving the SLC41A1 N-terminal cytoplasmic domain.

  • magnesium
  • membrane
  • MgtE
  • solute carrier 41
  • member A1 (SLC41A1)
  • transient receptor potential melastatin 7 (TRPM7)
  • transporter

INTRODUCTION

Targeted deletion of the TRPM7 (transient receptor potential cation channel, subfamily M, member 7) ion channel in tumour and stem cell contexts results in defects in cell growth that can be complemented through provision of supplemental Mg2+ to the growth medium [1,2], or through heterologous expression of a Mg2+ transporter or a phosphoinositide 3-kinase catalytic subunit [3,4]. These observations have suggested that Mg2+ uptake is a fundamental determinant of cellular anabolic metabolism, and that further study of the molecular mechanisms of Mg2+ transport may yield novel insights into the control of basic cellular anabolic processes.

Although M7-KO (TRPM7-deficient) cells require supplemental Mg2+ to grow, the concentrations required (∼5–10 mM) are modest in comparison with the ∼100 mM levels of Mg2+ supplementation required to support growth of bacteria deficient in all known bacterial Mg2+ transport proteins [5]. On the basis of this observation, we hypothesized that vertebrate cells normally express a plasma membrane Mg2+ transport system in addition to that mediated via TRPM7. To gain insight into the nature of the predicted alternative Mg2+ transport pathways, we subjected M7-KO DT40 B-cells to mutagenesis/selection in reduced Mg2+ conditions in order to identify cell populations that had recovered the capacity to grow without supplemental Mg2+. Analysis of the MgtE transporter homologue SLC41A1 (solute carrier 41, member A1) Mg2+ transporter expression in the selected clone in comparison with WT (wild-type) and M7-KO parental cells demonstrated up-regulated expression of SLC41A1. Characterization of native and heterologous SLC41A1 transporter expression under various conditions in both WT and M7-KO DT40 cells demonstrated that surface accumulation of SLC41A1 transporters is regulated through a Mg2+-dependent endosomal recycling mechanism. Structure–function analyses of SLC41A1 transporter endosomal/lysosomal trafficking demonstrated further that plasma membrane recycling of SLC41A1 is independent of Mg2+ transport function and requires an intact N-terminal cytoplasmic domain. Taken together, our results define SLC41A1 transporters as a central component of the vertebrate plasma membrane Mg2+-uptake system, and implicate endosomal recycling of SLC41A1 transporters mediated through the N-terminal cytoplasmic domain as a key Mg2+ transport regulatory mechanism.

MATERIALS AND METHODS

Cell lines

DT40 cells were grown in RPMI 1640 medium supplemented with 10% (v/v) FBS (fetal bovine serum), 1% chicken serum, 2 mM L-glutamine and 10 units/ml penicillin/streptomycin at 37 °C in a humidified 5% CO2 environment. HEK (human embryonic kidney)-293 cells were grown in Dulbecco's modified Eagle's medium supplemented with 10% FBS at 37 °C in a humidified 5% CO2 environment.

Electrophoresis and Western blotting

SDS/PAGE was performed using the Laemmli method with 10 or 12% (w/v) acrylamide gels. Cells were lysed in TNE buffer (150 mM NaCl, 50 mM Tris/HCl, pH 7.5, and 1% Nonidet P40) supplemented with protease inhibitor cocktail (cOmplete™ Mini; Roche) on ice and rotated at 4 °C for 30 min. Lysate was subsequently centrifuged at 20000 g for 15 min at 4 °C, and cell nuclei were removed. Cleared lysate was added to sample buffer and separated by SDS/PAGE. Protein was transferred on to a PVDF membrane (Millipore) and blocked for 1 h at room temperature (23 °C) in blocking buffer {5% (w/v) non-fat milk, TBS [Tris-buffered saline (140 mM NaCl and 20 mM Tris/HCl, pH 7.5)]/0.1% Tween-20 or Odyssey Blocking Buffer; LI-COR} and stained with rabbit anti-HA (haemagglutinin) tag mAb (monoclonal antibody) (Cell Signaling Technology, catalogue number 3724), mouse anti-Myc mAb (Cell Signaling Technology, catalogue number 2276) or anti-SLC41A1 polyclonal antibody raised against an N-terminal peptide of human SLC41A1 (Abcam, catalogue number ab83701). Bound antibody was detected using either the Amersham or the Pierce ECL (enhanced chemiluminescence) detection system or subsequently stained with either IRDye 800 CW-conjugated goat anti-rabbit mAb (LI-COR) or Alexa Fluor® 680-conjugated goat anti-mouse mAb (Invitrogen) IR-fluorescent dye-conjugated secondary antibody and detected using the LI-COR Odyssey® Imaging System.

qPCR (quantitative PCR) analysis of endogenous SLC41A1 in DT40 WT, M7-KO and E1 cell lines

Total mRNA was extracted using the Qiagen RNeasy Mini™ and QIAshredder™ kits. Genomic DNA was removed during this step with an on-column DNase digestion using the Qiagen RNase-free DNAse kit. cDNA synthesis was subsequently reverse-transcribed on total RNA using the Invitrogen Superscript II® RT (reverse transcription)–PCR kit and poly(A)·oligo(dT)25 as a template primer. cSLC41A1 (chicken SLC41A1; GenBank® accession number XM_417968.2) cDNA was detected by PCR to produce a 1543 bp product using the forward primer 5′-CTTCCAAGCCAGAGCAGAAG-3′ and the reverse primer 5′-CTAATCCCCCACATCAGAGTC-3′. The amplicon was subsequently electrophoresed on a 1% (w/v) agarose gel and sequenced to confirm the identity of the PCR product as cSLC41A1. DT40 SLC41A1 transcript concentrations were determined by qPCR using the dsDNA (double-stranded DNA)-binding dye SYBR Green and a Bio-Rad Laboratories C1000 qPCR Detection Thermal Cycler. SLC41A1 forward primer, 3′-GCTGGGGCTTGTACAAAGAG-5′, and reverse primer, 3′-TTCTTGGCAATGATGATCCA-5′, were used for the PCR, producing a 103 bp amplicon. Actin primers (forward primer, 3′-CCGTGCTGTGTTCCCATCTAT-5′ and reverse primer, 3′-ATCCCAGTTGGTGACAATACCG-5′), producing a 160 bp amplicon, were used as an internal control. The resulting data were analysed by a modified version of the Pfaffl method [6].

Cloning and expression analysis of heterologous SLC41A1

Human SLC41A1 ORF (open reading frame) (GenBank® accession number NM_173854), a gift from Dr Hans-Christian Aasheim (Department of Medical Genetics, University Hospital Oslo, Oslo, Norway), was cloned into a modified T-Rex™ pcDNA5/TO or pCDNA4/TO doxycycline-regulated expression vector (Invitrogen) with an in-frame N-terminal HA tag and C-terminal Myc tag. The N-terminus of SLC41A1 was deleted by amplifying bases 275–1542 using the forward primer 3′-TTAGCGGCCGCAGGAGACCTCCTTTTCCATC-5′ and a reverse primer downstream of the SLC41A1 gene in the pcDNA5/TO-HA-SLC41A1-Myc expression vector. The N-terminal deletion mutant (ΔNT) was subsequently cloned back into the pcDNA5/TO expression vector containing the in-frame N-terminal HA tag and C-terminal Myc tag. Site-directed mutagenesis was performed on SLC41A1 using the Stratagene QuikChange® II XL site-directed mutagenesis kit. The pCDNA5/TO-HA-SLC41A1-Myc and SLC41A1 mutants or pCDNA4/TO-GFP constructs were linearized and electroporated into M7-KO DT40 cells and selected for either hygromycin or zeocin resistance. Inducible protein expression was analysed by flow cytometry and Western blotting.

Complementation of M7-KO cell lines

M7-KO DT40 cell lines transfected with the inducible construct were assayed for Mg2+-deficient complementation by measuring cell growth and proliferation in 0.5 mM Mg2+ (RPMI 1640 medium without supplemental Mg2+), 0.5 mM Mg2+ with 1 μg/ml doxycycline or 15 mM supplemental Mg2+. Cells were washed twice with PBS and 7.5×107–106 cells/ml were seeded into different medium preparations. Every 24 h, the populations were counted using Spherotech Accucount fluorescent particles and a BD LSR II™ flow cytometer. After determining cell concentrations, cells were split down to the seeding concentration using the appropriately treated medium.

Expression quantification and membrane topology analysis of heterologous SLC41A1

Intracellular staining for analysing total protein expression was performed by first treating 5×105 cells with the BD Cytofix/Cytoperm™ fixation/permeabilization kit followed by staining with either rabbit anti-HA or mouse anti-Myc primary antibody from Cell Signaling Technology. Next, samples were stained with either FITC-conjugated anti-rabbit or PE (phycoerythrin)-conjugated anti-mouse secondary antibodies. For surface protein detection, 5×106 cells were washed twice in PBS then resuspended and stained with either anti-HA or anti-Myc primary antibodies (Cell Signaling Technology), and 0.1% BSA in PBS. Cells were washed twice in PBS and subsequently stained with streptavidin-conjugated PE (BD Bioscience, catalogue number 554061), goat anti-rabbit IgG (Invitrogen, catalogue number A-11034) or PE-conjugated rat anti-mouse IgG2a (Southern Biotech, catalogue number 1155-09) and analysed for expression on a BD LSR II™ flow cytometer.

Computational analysis

SLC41A1 N-terminal and transmembrane domains were identified on the basis of primary sequence data produced by the TMHMM 2.0 (http://www.cbs.dtu.dk/services/TMHMM/) and TMpred (http://www.ch.embnet.org/software/TMPRED_form.html) transmembrane domain prediction programs. Secondary structures were identified using Jpred 3, a secondary structure prediction server (http://www.compbio.dundee.ac.uk/www-jpred/) and compared with the secondary structures of the crystal structure of MgtE on the basis of the alignment prediction between MgtE and SLC41A1 produced using the AlignX program (Invitrogen).

RESULTS

M7-KO cells adapted to grow without supplemental Mg2+ exhibit Mg2+-dependent up-regulation of SLC41A1

M7-KO B-cells and embryonic stem cells are able to grow only when provided with 10–15 mM supplemental Mg2+ in their cell culture medium [1,2]. In order to gain insight into the mechanisms of cellular Mg2+ homoeostasis, we subjected 108 cells from the M7-KO cell line to Acridine ICR 191-induced frameshift mutagenesis, and plated cells at limiting dilution in cell culture medium without supplemental Mg2+ in order to identify cell clones that had regained the capacity to grow at a physiological Mg2+ concentration. One clone (designated E1) was isolated that exhibited essentially normal growth properties in standard tissue culture medium (Figure 1A).

Figure 1 Functional analysis of endogenous SLC41A1

(A) M7-KO-derived E1 cells proliferate at WT levels in the absence of supplemental Mg2+. WT 0.5 mM, WT DT40 cells grown in standard cell culture medium; WT 15 mM, WT DT40 cells grown in 15 mM supplemental Mg2+; E1 0.5 mM, E1 cells grown in 0.5 mM Mg2+; M7-KO 0.5 mM, M7-KO cells grown in 0.5 mM Mg2+. Growth was analysed by cell counting as described in the Materials and methods section. (B) Native and heterologous SLC41A1 are detectable by anti-SLC41A1 staining. Left-hand panel: HA–cSLC41A1-expressing HEK-293T cells were induced and stained with anti-SLC41A1. Dox, doxycycline (1 μg/ml, where added). Right-hand panel: WT and M7-KO cell lysates were immunoblotted with anti-SLC41A1. (C) M7-KO and E1 cells demonstrate remarkably enhanced SLC41A1 expression that is sensitive to extracellular Mg2+. Anti-SLC41A1 stained lysates derived from DT40 WT, E1 and M7-KO cells cultured in the indicated Mg2+ concentrations. (D) Mg2+-sensitive SLC41A1 expression behaviour exhibited by E1 and M7-KO cells does not correlate with SLC41A1 transcript abundance. Values on the y-axis are means±S.E.M. SLC41A1 qPCR Ct values were expressed relative to β-actin using a modified version of the Pfaffl method [6] (n=3). The x-axis is first divided by cell type, subdivisions are mM Mg2+ in the culture medium. Because M7-KO cells cannot survive in 0.5 mM Mg2+ for 72 h, SLC41A1 mRNA was not analysed for M7-KO cells at 0.5 mM Mg2+.

Although up-regulation of several previously characterized Mg2+ transporters [7] could potentially allow E1 to proliferate in cell culture medium without supplemental Mg2+, we focused our study on SLC41A1, as we were able to detect it at an endogenous level using a commercially available antibody (Figure 1B). Using this antibody, quantitative analysis of endogenous SLC41A1 transporter expression in E1 cells, the parental M7-KO cells, and WT DT40 cells indicated that SLC41A1 transporters were up-regulated in M7-KO cells, and up-regulated further in the E1 clone (Figure 1C). Furthermore, analysis of SLC41A1 transporter expression in various concentrations of Mg2+ demonstrated Mg2+-dependent down-regulation of SLC41A1 transporters unique to M7-KO cells and the E1 clone, with increasing Mg2+ resulting in substantial suppression of SLC41A1 transporter expression. Taken together, these data suggest that TRPM7-independent growth of transitioned M7-KO and E1 cells is in part due to up-regulation of SLC41A1 transporter expression, and that SLC41A1 transporter expression is regulated in accordance with the status of cellular Mg2+ homoeostasis.

To gain insight into the mechanism through which SLC41A1 protein expression is up-regulated in the E1 clone, SLC41A1 mRNA abundance was assayed via qPCR in WT, M7-KO and E1 cells maintained over a range of supplemental Mg2+ concentrations (Figure 1D). We observed comparable mRNA abundance in all three cell types over the entire range of Mg2+ concentrations tested: in the E1 clone, the modest increase in SLC41A1 abundance in all Mg2+ conditions was not statistically different from either WT DT40 or M7-KO cells (see Supplementary Figure S1 at http://www.BiochemJ.org/bj/439/bj4390129add.htm). Furthermore, sequencing of the SLC41A1 transcript from WT, E1 and M7-KO cells showed no difference in SLC41A1 sequence between the various cell lines (see Supplementary Figure S2 at http://www.BiochemJ.org/bj/439/bj4390129add.htm). On the basis of these observations, we conclude that up-regulated SLC41A1 transporter expression in M7-KO cells is post-transcriptional, and that the enhanced growth capacity of E1 cells relative to M7-KO cells in 0.5 mM Mg2+ is in part due to up-regulation of SLC41A1 transporter expression via this mechanism.

Heterologous SLC41A1 complements the growth of M7-KO cells

To verify the functional capacity of SLC41A1 transporters, and to develop a model in which to further analyse potential post-transcriptional regulatory mechanisms of SLC41A1, we generated M7-KO DT40 cells with heterologous expression of WT SLC41A1. For this purpose, an SLC41A1 ORF with N-terminal HA and C-terminal Myc epitope tags was cloned 5′ to a doxycycline-regulated promoter, and stably introduced into DT40 M7-KO cells. Doxycycline-regulated expression of the encoded protein, designated HA–SLC41A1–Myc, was confirmed by immunoblotting with both anti-HA and anti-Myc antibodies. Consistent with previous observations [810], HA–SLC41A1–Myc exhibited an approximate molecular mass of 56 kDa (see Supplementary Figure S3, immunoblot panels, at http://www.BiochemJ.org/bj/439/bj4390129add.htm).

To determine whether HA–SLC41A1–Myc was functional, we compared the Mg2+-dependent growth requirements of HA–SLC41A1–Myc-expressing cells (Figure 2A). Expression of HA–SLC41A1–Myc supported nearly WT levels of growth in the absence of supplemental Mg2+, whereas uninduced cells maintained in the absence of supplemental Mg2+ were unable to proliferate. Live-cell staining using antibodies targeted to the HA and Myc tags (Figure 2B, middle and right-hand panels respectively) revealed that the C-terminal Myc tag was readily detectable at the cell surface, demonstrating unequivocally that the tagged SLC41A1 was successfully trafficking to the cell surface, as would be expected for an extracellular Mg2+-uptake transporter. On the basis of these observations and our previous characterization of the Mg2+-dependent growth properties of this cell line, we conclude that heterologously expressed HA–SLC41A1–Myc is functional and able to mediate effective trans-plasma membrane Mg2+ transport. Furthermore, we were also able to detect the HA-tagged protein by intracellular staining in fixed and permeabilized cells (Figure 2B, left-hand panel). As a side note, our capacity to detect the C-terminal Myc tag, but not the N-terminal HA tag, on the cell surface strongly supports an 11-transmembrane span model for SLC41A1 in which the SLC41A1 MgtE domains are oriented as predicted by the prokaryotic MgtE crystal structure (Figure 2C). Notable topological characteristics include the predicted 11-transmembrane spans plus an approximately 92 residue intracellular N-terminal cytoplasmic domain, and a short, approximately nine residue, extracellular C-terminal tail.

Figure 2 Functional characterization and topological analysis of SLC41A1

(A) Growth and complementation of HA–SLC41A1–Myc-transfected TRPM7-KO cells. Cells were grown in 15 mM Mg2+ supplemented medium, Mg2+-unsupplemented medium (0.5 mM Mg2+) or Mg2+-unsupplemented medium with 1 μg/ml doxycycline (+Dox). (B) Surface and total cell staining indicate that SLC41A1 has an N-terminal-in C-terminal-out membrane orientation. Uninduced and induced HA–SLC41A1–Myc-expressing cells were fixed and permeabilized and anti-HA-stained (left-hand panel) or live-stained for either anti-HA (middle panel) or anti-Myc (right-hand panel) and analysed by flow cytometry. (C) Despite previous findings, SLC41A1 probably has 11 transmembrane spans and an N-terminal-in C-terminal-out membrane orientation. Lefthand panel: previously proposed ten-transmembrane span model of SLC41A1. Right-hand panel: an 11-transmembrane span N-terminal-in C-terminal-out SLC41A1 structure based on transmembrane prediction software, homology with MgtE, and surface detection of the C-terminal Myc tag.

Surface trafficking of SLC41A1 is regulated by Mg2+-dependent endosomal/lysosomal degradation

During the course of our initial experiments to characterize heterologous HA–SLC41A1–Myc expression, we observed that inducible expression of HA–SLC41A1–Myc was greater at 24 h than at 48 h (Figure 3A), a finding consistent across flow cytometry analyses of both total and surface protein expression as well as Western blot analysis. Previous studies have shown that there is a positive correlation between the length of induction and total protein expression in the T-Rex™ doxycycline-inducible expression system [11]. Thus, in the absence of post-translational regulation of expression, one would expect to observe a steady increase and eventual plateau in inducible SLC41A1 expression over time. As a control for doxycycline-induced expression in our system, a doxycycline-inducible GFP (green fluorescent protein) construct was transfected into the M7-KO DT40 background and analysed for expression via flow cytometry. Figure 3(B) shows that GFP expression did not decrease between 24 and 48 h when cultured in 15 mM Mg2+. Similarly, doxycycline-induced expression of the SLC41A1 homologue HA–SLC41A2 in the same expression system and under the same conditions did not lead to a similar time-dependent down-regulation of transporter accumulation (Figure 3B, right-hand panel).

Figure 3 Mg2+-dependent endosomal regulation of surface and total SLC41A1 expression

(A) SLC41A1 total and surface expression decrease with respect to length of induction in Mg2+-unsupplemented medium. HA–SLC41A1–Myc-expressing M7-KO cells were induced with 1 μg/ml doxycycline in standard medium for 0 (no induction), 24 or 48 h. From these cells, whole-cell lysate was immunoblotted and anti-HA-stained (left-hand panel). These cells were also fixed and permeabilized, anti-HA-stained and analysed by flow cytometry (middle panel). The same cells under the same conditions were live-stained with anti-Myc and analysed by flow cytometry (right-hand panel). (B) Doxycycline-inducible protein control expression does not decrease post-induction. M7-KO cells expressing either GFP (left-hand panel) or SLC41A2 (right-hand panel) were induced for 0 (uninduced), 24 or 48 h in medium supplemented with 1 μg/ml doxycycline and 15 mM Mg2+. GFP-expressing live cells were analysed by flow cytometry. SLC41A2 cells were fixed and permeabilized, anti-HA-stained and analysed by flow cytometry. (C) HA–SLC41A1–Myc expression is regulated in a Mg2+-dependent manner. HA–SLC41A1–Myc-expressing M7-KO cells were induced with 1 μg/ml doxycycline and grown in 50 μM, 0.5 mM or 15 mM Mg2+ for 24, 48 or 72 h. Cells were subsequently fixed and permeabilized, anti-HA-stained and analysed by flow cytometry. (D) NH4Cl treatment enhances endogenous SLC41A1 expression in WT DT40 cells. WT DT40 cells were treated with or without 40 mM NH4Cl. Whole-cell lysates were immunoblotted and probed with anti-SLC41A1; β-actin was used as a loading control. (E) NH4Cl effects dramatic increases in heterologous SLC41A1 surface expression. M7-KO and HA–SLC41A1–Myc-expressing cells were grown in 1 μg/ml doxycycline with either standard RPMI 1640 medium (left-hand panel) or medium containing 80 mM NH4Cl for 24 h (middle panel). Live cells were subsequently labelled with anti-Myc and analysed for expression by flow cytometry. Whole-cell lysate was obtained from cells grown in 1 μg/ml doxycycline (+/0), 80 mM NH4Cl (−/80), 1 μg/ml doxycycline and 40 mM NH4Cl (+/40), or 1 μg/ml doxycycline and 80 mM NH4Cl (+/80). Note that the +/0 lane (boxed) was taken from a different region of the same immunoblot at the same exposure and digitally pasted next to the other bands. Dox, doxycycline.

The above observation suggests that, like the native protein, heterologous SLC41A1 is subject to some form of posttranscriptional regulation. To determine whether the post-transcriptional regulation of HA–SLC41A1–Myc involved a Mg2+-dependent mechanism, we evaluated the influence of different concentrations of extracellular Mg2+ on HA–SLC41A1–Myc expression. M7-KO DT40 cells with inducible expression of HA–SLC41A1–Myc were treated with doxycycline and subsequently maintained in 15 mM, 0.5 mM or 50 μM MgCl2 for 24, 48 or 72 h before their expression was evaluated (Figure 3C). The 15 mM and 0.5 mM Mg2+ populations displayed the characteristic down-regulation from 24 to 48 h, and expression continued to decrease up to 72 h for these Mg2+ concentrations. Cells maintained in 50 μM Mg2+, however, did not down-regulate HA–SLC41A1–Myc expression over time; rather induction of HA–SLC41A1–Myc expression in 50 μM Mg2+ resulted in a consistent increase in expression up to the 72 h time point. Taken together, these observations demonstrate a strong dependence of heterologous protein expression on extracellular Mg2+ concentration. As no effect of high extracellular Mg2+ concentration was observed on the expression of other doxycycline-regulated proteins (e.g. see expression time course in Figure 3B), the above results indicate further that regulation of the accumulation of HA–SLC41A1–Myc protein is occurring at a post-translational step.

Lysosomal degradation is an important mechanism for post-translational regulation of expression of many surface proteins, and can be efficiently probed through the use of lysosomotropic agents, such as ammonium chloride (NH4Cl), that inhibit lysosomal proteases [1216]. To evaluate the role of endosomal/lysosomal degradative pathways in regulation of SLC41A1 transporter expression, we treated WT DT40 cells with NH4Cl, and evaluated endogenous SLC41A1 expression by immunoblotting. In the WT DT40 cell line, endogenous SLC41A1 expression was almost undetectable at baseline (e.g. Figures 1B and 1C). However, treatment of WT DT40 cells with 40 mM NH4Cl for 24 h caused a dramatic increase in expression level (Figure 3D, upper panel).

Inhibition of endosomal/lysosomal degradative pathways by NH4Cl would be expected to cause SLC41A1 transporter up-regulation by allowing SLC41A1 molecules previously fated for degradation to be recycled to the cell surface, thus it should result in parallel changes in total and surface expression of SLC41A1. To test this prediction, we treated cell lines expressing HA–SLC41A1–Myc with NH4Cl, and evaluated total and surface protein expression by Western blotting and live-cell flow cytometry respectively (Figure 3E). Evaluation of total expression by immunoblotting for the HA epitope tag demonstrated that treatment with NH4Cl resulted in markedly enhanced total protein expression, whereas live-cell-surface staining for the C-terminal epitope tag of this construct demonstrated large increases in HA–SLC41A1–Myc surface expression. Taken together, these results suggest a model in which the SLC41A1 expression is post-translationally regulated via internalization of SLC41A1 transporters from the cell surface; these transporters are then degraded in the endosomal/lysosomal pathway in a Mg2+-homoeostasis-regulated manner: if the cellular Mg2+ concentration is sufficient, the transporters are degraded, presumably in lysosomes, whereas if the cell requires a higher rate of Mg2+ uptake, the transporters are recycled from the endosomal compartment to the cell surface.

SLC41A1 pore mutants are unable to complement growth of M7-KO cells, but exhibit normal surface trafficking

One potential mechanism to account for Mg2+-dependent regulation of SLC41A1 degradation is that internalization and degradation are somehow linked directly to SLC41A1 Mg2+ transport function. To test this possibility, the effect of pore-region point mutations on Mg2+ permeation and growth complementation was investigated. As our surface staining and topological analyses suggested that SLC41A1 was located in the membrane similarly to prokaryotic MgtE proteins, predicted functionally important pore residues in SLC41A1 were identified and mutated on the basis of their homology with residues within the ion-conducting region lining the pore of the prokaryotic Thermus thermophilus MgtE protein crystal structure [17,18]. Since MgtE is a homodimer and SLC41A1 is a pseudodimer, each SLC41A1 molecule possesses two residues corresponding to highly conserved residues in the MgtE protein. The locations of the residues implicated in the MgtE Mg2+-conducting pore, and the location of the homologous residues in SLC41A1, are depicted in Figure 4(A). As expected, a D263A mutation, homologous with the MgtE D432A mutation that abolishes T. thermophilus MgtE Mg2+ transport [17,18], completely abolished the capacity of HA–SLC41A1–Myc to complement the growth of M7-KO DT40 cells (Figure 4B). Similarly, it was found that D487A mutant HA–SLC41A1–Myc channels were also unable to provide M7-KO cells the ability to proliferate without supplemental Mg2+.

Figure 4 Pore mutation residue localization, expression and Mg2+-deficient complementation

(A) SLC41A1 half-domain has two regions of homology with MgtE. Depicted is a crucial residue in the ion-conducting channel of MgtE and the homologous locations in SLC41A1. The GenBank® accession numbers for the protein sequences in the alignment are: NP_776253 (Homo sapiens SLC41A1), XP_417968 (predicted Gallus gallus SLC41A1), XP_002933050 (predicted Xenopus tropicalis SLC41A1#1), XP_002938280 (predicted X. tropicalis SLC41A1#2), NP_001096676 (X. tropicalis SLC41A2) and YP_144326 (T. thermophilus HB8 MgtE). (B) SLC41A1 mutations D263A and D487A cause Mg2+-transport defects that do not affect expression and surface trafficking. M7-KO cells expressing SLC41A1-D263A or SLC41A1-D487A were either uninduced or induced with 1 μg/ml doxycycline, in 15 mM Mg2+ for 24 h. Induced and uninduced whole-cell lysate was extracted and immunoblotted for HA (left-hand panels). Cells under the same conditions were live-stained for Myc and analysed for surface expression by flow cytometry (middle panels). SLC41A1-D263A- and SLC41A1-D487A-expressing cells were grown in 0.5 mM Mg2+, 15 mM Mg2+ or 0.5 mM Mg2+ and 1 μg/ml doxycycline. Growth was analysed by cell counting and passaging as described in the Materials and methods section (right-hand panels).

To determine whether lack of complementation was the result of a loss of transporter trafficking to the cell surface (such as might occur if the transporter were unable to fold properly), cells were induced to express each pore mutant channel in 15 mM Mg2+, and then probed for the surface-exposed Myc-tag. Both the D263A and D487A mutant HA–SLC41A1–Myc channels were observed to traffic to the cell surface as effectively as WT HA–SLC41A1–Myc (Figure 4B, flow panels). In addition to providing additional support for the existence of functional conservation between bacterial MgtE and vertebrate SLC41A1 proteins, these data demonstrate that a functional Mg2+-conducting pore is not required for efficient folding or intracellular trafficking and surface expression of SLC41A1 transporters.

Mg2+-dependent regulation of SLC41A1 is compromised by deletion of its intracellular N-terminal domain

The 92-residue hydrophilic and intracellular N-terminus comprises approximately 20% of the total amino acid sequence of SLC41A1, and is a plausible region for regulation of SLC41A1 transporter intracellular trafficking due to its cytoplasmic accessibility [1922]. To evaluate the role of the N-terminal domain in SLC41A1 post-translational regulation, an SLC41A1 variant with truncation of the entire N-terminal domain was constructed and designated ΔNT. ΔNT was found to be expressed at the predicted molecular mass of ∼40 kDa, and appeared to function normally in that it supported nearly WT levels of growth in the absence of supplemental Mg2+ (Figure 5A, left-hand two panels). However, although we observed the presence of abundant ΔNT in fixed and permeabilized cells, surface expression of ΔNT was undetectable (Figure 5A, right-hand two panels). Given the capacity of ΔNT to complement the growth of M7-KO cells, we hypothesized that the ΔNT transporter was able to traffic efficiently to the surface, but exhibited a defect in regulation of its intracellular trafficking following internalization. As an initial test of this hypothesis, we evaluated Mg2+-dependent regulation of ΔNT expression. As before, total channel expression was measured via Western blot analysis and flow cytometry after 0, 24 and 48 h of induction in the presence of various concentrations of Mg2+. By comparing Figures 5(B) and 3(C), it can be seen that down-regulation of ΔNT from 24 to 48 h of induction is far less dramatic than for the WT HA–SLC41A1–Myc transporter. In addition, in contrast with the WT and pore mutant transporters, total expression of ΔNT did not up-regulate in the presence of 50 μM, 0.5 mM or 15 mM Mg2+, indicating that ΔNT is nearly entirely deficient in Mg2+-dependent regulation.

Figure 5 N-terminally deleted SLC41A1 Mg2+-deficient complementation and expression, and Mg2+-dependent regulation

(A) Induced ΔNT proliferates without supplemental Mg2+, but does not accumulate at the cell surface. ΔNT-expressing M7-KO cells were grown in RPMI 1640 medium supplemented with 0.5 mM Mg2+, 15 mM Mg2+ or 0.5 mM Mg2+ and 1 μg/ml doxycycline. Growth was analysed by cell counting and passaging as described in the Materials and methods section (left-hand panel). ΔNT-expressing M7-KO cells were induced in 0.5 mM Mg2+ for 0 (no induction), 24 and 48 h. From these cells, whole-cell lysate was immunoblotted and HA-stained (middle lefthand panel). These cells under the same conditions were fixed and permeabilized, anti-HA-stained and analysed by flow cytometry (middle right-hand panel). The uninduced and 24 h-induced cells were live-stained with anti-Myc and analysed by flow cytometry (right-hand panel). (B) In contrast with WT transporter, N-terminally deleted SLC41A1 does not exhibit Mg2+-dependent post-translational down-regulation. ΔNT-expressing M7-KO cells were induced with 1 μg/ml doxycycline and grown in 50 μM, 0.5 mM or 15 mM Mg2+ for 24, 48 or 72 h. Cells were subsequently fixed and permeabilized, anti-HA-stained and analysed by flow cytometry. Dox, doxycycline.

Mg2+-dependent endosomal recycling of SLC41A1 proteins requires an intact N-terminal domain

To define further the regulation of SLC41A1 intracellular trafficking, M7-KO cells with inducible expression of HA–SLC41A1–Myc or mutant transporters were induced and grown over several weeks in 50 μM Mg2+ (50 μM-adapted). Analysis of WT HA–SLC41A1–Myc expression after long-term induction in 50 μM Mg2+ revealed markedly increased total and surface transporter expression (Figure 6A), clearly demonstrating the robust influence of Mg2+ homoeostasis in the regulation of cell-surface SLC41A1 transporters. Similar attempts to grow the ΔNT mutant long-term in 50 μM Mg2+ failed, consistent with its deficiency in Mg2+-dependent regulation. Although pore mutants cannot support the growth of M7-KO cells in 0.5 mM Mg2+, evaluation of intermediate levels of supplemental Mg2+ demonstrated that the D263A cell line was able to grow long-term at reduced supplementation levels (1 mM supplemental Mg2+), and that complementation under these conditions was associated with highly up-regulated expression (Figure 6B, the D487A mutant was not able to grow under any conditions). These observations suggest that the D263A transporter has reduced, but not entirely absent, capacity to transport Mg2+, and that the capacity of cellular Mg2+ status to regulate its intracellular trafficking is intact. Taken together, these observations demonstrate that post-translational regulation of SLC41A1 expression is linked directly to the status of cellular Mg2+-transport; as plasma membrane Mg2+ transport is reduced, via either reduced extracellular concentration or restricted transporter activity, total and surface SLC41A1 expression is increased proportionately to achieve cellular Mg2+ homoeostasis. Furthermore, this mechanism is independent of the transporter's Mg2+-uptake capacity, and fully dependent on its N-terminal cytoplasmic domain.

Figure 6 Sensitivity of SLC41A1 and mutants to long-term Mg2+-dependent regulation and lysosomotropic compounds

(A) Long-term growth of SLC41A1-expressing M7-KO cells in 50 μM Mg2+ medium effects large increases in total and surface SLC41A1 expression. SLC41A1-expressing M7-KO cells were grown in approximately Mg2+-free (50 μM Mg2+) RPMI 1640 medium and 1 μg/ml doxycycline for several weeks. These cells along with uninduced SLC41A1-expressing cells were fixed and permeabilized, anti-HA-stained and analysed by flow cytometry (left-hand panel). The same cells were live anti-Myc stained, and analysed by flow cytometry (right-hand panel). (B) Enhanced transporter expression at reduced Mg2+ supplementation levels is only apparent in cells that express SLC41A1 containing a Mg2+-transport defect. M7-KO cells expressing WT SLC41A1, D263A or ΔNT were grown in 1 μg/ml and 0.5, 1, 5 and 15 mM Mg2+ for several weeks before analysis. Uninduced cells were grown in 15 mM Mg2+. Cells were subsequently fixed and permeabilized and HA-stained (left-hand panel) or live anti-Myc stained (right-hand panel), before analysis by flow cytometry. (C) Lysosomotropic agents dramatically increase surface expression of the 50 μM-adapted clone and SLC41A1 mutants including the otherwise undetectable ΔNT transporter. The 50 μM-adapted clone, ΔNT and D263A were grown in 80 mM NH4Cl with (bottom) or without (top) 1 μg/ml doxycycline for 24 h before live-cell anti-Myc staining and analysis by flow cytometry. M7-KO cells were treated with 80 mM NH4Cl for 24 h before anti-Myc staining and analysis by flow cytometry. (D) ΔNT and D263A-expressing M7-KO cells were grown in either 1 μg/ml doxycycline (+/−), 80 mM NH4Cl (−/80), 1 μg/ml doxycycline and 40 mM NH4Cl (+/40) or 1 μg/ml doxycycline and 80 mM NH4Cl (+/80), for 24 h before obtaining whole-cell lysate that was immunoblotted and anti-HA-stained. Dox, doxycycline.

The deficient Mg2+-dependent regulation of the ΔNT transporter could occur either as the result of reduced surface trafficking efficiency, or because the ΔNT transporter is able to efficiently traffic to the surface, but is subject to constitutive internalization and degradation. To differentiate these possibilities, we evaluated the influence of lysosomotropic agents on the trafficking and surface expression of the D263A and ΔNT transporters. Cell lines with inducible expression of D263A and N-terminally deleted HA–SLC41A1–Myc transporters were treated with doxycycline and, along with the 50 μM Mg2+-adapted HA–SLC41A1–Myc cells, treated subsequently with NH4Cl (Figure 6C). Treatment of the adapted 50 μM Mg2+ cells with NH4Cl further enhanced surface expression of the already high-expressing cell line. NH4Cl enhanced expression of the functionally compromised mutant D263A was comparable with the massive shift observed by the WT. Surface expression of the induced, otherwise untreated, ΔNT mutant was almost immeasurable (Figures 5A, right-hand panel, and 6B, right-hand panel), yet, remarkably, treatment with NH4Cl was able to completely reverse this phenotype to create high-level surface expression. Similar results were obtained in all cases tested for the alternative lysosomotropic agent monensin (see Supplementary Figure S4 at http://www.BiochemJ.org/bj/439/bj4390129add.htm).

We also evaluated total protein expression of HA–SLC41A1–Myc and its mutants by immunoblotting for the HA epitope tag on whole-cell lysates derived from NH4Cl-treated and untreated cells from each cell line (Figure 6D). In each case, treatment with NH4Cl resulted in enhanced total protein expression. Taken together, these data strongly support the hypothesis that the major step involved in post-translational regulation of SLC41A1 is lysosomal degradation. They also suggest that deletion of the N-terminal cytoplasmic domain does not compromise surface trafficking of SLC41A1; rather the loss of its N-terminal domain leads to a transporter that is constitutively internalized and degraded.

DISCUSSION

In the present study, we have shown that SLC41A1 proteins are central components of vertebrate plasma membrane Mg2+-uptake systems, functioning as plasma membrane Mg2+ transporters whose cell-surface expression is regulated in accordance with cellular Mg2+ homoeostasis. Using heterologous expression of WT and mutant proteins, we demonstrate further that Mg2+-dependent regulation of SLC41A1 transporters occurs through modulation of intracellular transport and lysosomal degradation mediated by the SLC41A1 N-terminal cytoplasmic domain.

Our results suggest a model for regulation of SLC41A1 Mg2+ transport in which, in order to meet the metabolic demands of a growing cell, SLC41A1 transporters traffic to the cell surface, thereby allowing Mg2+ flow through their Mg2+-selective channel pores via a combination of chemical and electrical gradients. This transporter-specific trafficking process is similar to the surface regulation of the glucose transporter GLUT3, which is required for energy homoeostasis in synaptic cells [23]. When cells are Mg2+-replete, transporters are internalized and predominantly traffic to lysosomes for degradation (Figure 7, left-hand panel). However, when cells are Mg2+-deficient, SLC41A1 transporters either avoid internalization or are internalized as usual, but avoid lysosomal degradation and are recycled to the cell surface (Figure 7, right-hand panel). The constitutive internalization and degradation observed for the ΔNT mutant is most consistent with the latter model.

Figure 7 SLC41A1 Mg2+-dependent endocytic down-regulatory mechanism

We propose that, independent of extracellular Mg2+ concentration, SLC41A1 is constitutively endocytically recycled from the cell surface. When the cellular Mg2+ concentration is low (left-hand panel), recycling of endocytosed SLC41A1 to the cell surface is favoured over degradation at the lysosome. When the cellular Mg2+ concentration is high (right-hand panel), lysosomal degradation is favoured.

On the basis of the trafficking phenotype of the ΔNT mutant, we speculate that the SLC41A1 N-terminal cytosolic domain contains subcellular trafficking motif(s) that provide for regulated intracellular trafficking. It is very common for transporters and other surface protein cytosolic domains to possess intracellular sorting motifs [15,19,21,2426]. In the case of SLC41A1, cellular Mg2+ would influence the recognition of one or more such motifs by endosomal or lysosomal protein sorting machinery, such that a low cellular Mg2+ concentration would result in recycling of the transporter to the cell surface from sorting endosomes/ERC (endocytic recycling compartment), rather than being targeted to the lysosomes. An important implication of this model is that the SLC41A1 N-terminal domain has the capacity to sense or receive information regarding cellular Mg2+ homoeostasis; if confirmed, this would represent a significant insight into the molecular mechanisms through which vertebrate cells are able to monitor and respond to their internal Mg2+ status. Furthermore, the capacity of the ΔNT mutant to complement cell growth despite our inability to detect its accumulation on the cell surface suggests that ΔNT transporter gating may be dysregulated, such that ΔNT transporters are able to support sufficient Mg2+ permeation to maintain cell growth and proliferation despite being rapidly cycled from the cell surface, as has been observed previously to occur with bacterial MgtE transporters [18]. SLC41A1 Mg2+ transport function and regulation are reminiscent of the function and regulation of yeast Alr family Mg2+ transporters demonstrated by Schweyen and colleagues [27]. Alr family proteins provide a plasma membrane Mg2+ transport function that is required for yeast cell division, and are observed to undergo Mg2+-dependent regulation of surface expression via endosomal internalization and vacuolar degradation, yeast processes directly analogous to the internalization and lysosomal degradation that we have demonstrated to regulate SLC41A1 function in vertebrate cells. As SLC41A1 transporters and the yeast Alr family transporters are members of distinct protein families with disparate structures, the conserved nature of their regulatory mechanisms suggest that Mg2+ homoeostasis may be regulated in accordance with conserved principles related to the unique role of Mg2+ in eukaryotic cellular energy metabolism, anabolic processes and growth.

Finally, our results highlight a central question in the field of Mg2+ transport: what are the relative roles and relationships of the multiple reported vertebrate Mg2+ transport or regulatory proteins in the context of overall cellular Mg2+ homoeostasis? In addition to SLC41A1, various other proteins, namely SLC41A2, MagT1 (magnesium transporter 1) and TUSC3 (tumour suppressor candidate 3), TRPM6 and TRPM7 and ACDP2 (ancient domain protein 2), have all been reported to have Mg2+ transport function [4,2831], and/or have been implicated in vertebrate cellular or organism level Mg2+ homoeostasis [29,3137]. The definition of a Mg2+-dependent regulatory mechanism for SLC41A1 opens the door to an in-depth understanding of how Mg2+ transport mediated by the various reported cellular Mg2+ transporter proteins are co-ordinated and contribute to overall cellular Mg2+ homoeostasis, as well as how these pathways interface with cellular metabolism, growth regulation and signalling.

AUTHOR CONTRIBUTION

Tyler Mandt built constructs, made cell lines, designed and performed experiments, collected and/or assembled data, analysed and interpreted data, and wrote the paper. Yumei Song designed and performed experiments, and analysed data. Andrew Scharenberg gave conceptual advice, financial support, supervised the project and wrote the paper. Jaya Sahni built the initial construct, made cell lines, designed and performed experiments, analysed data, edited the paper and supervised the project.

FUNDING

This work was supported by the National Institutes of Health [grant numbers PO1GM078195 and GM64316 (to A.M.S.)].

Abbreviations: FBS, fetal bovine serum; GFP, green fluorescent protein; HA, haemagglutinin; HEK, human embryonic kidney; mAb, monoclonal antibody; ΔNT, N-terminal deletion mutant; ORF, open reading frame; PE, phycoerythrin; qPCR, quantitative PCR; SLC41A1, solute carrier 41, member A1; cSLC41A1, chicken SLC41A1; TRPM7, transient receptor potential cation channel, subfamily M, member 7; M7-KO, TRPM7-deficient; WT, wild-type

References

View Abstract