Dicarboxylate

Jul 21, 2017 - We have previously modeled mammalian members of the SLC13 family, including the Na+/dicarboxylate cotransporter NaDC1 (SLC13A2), based ...
25 downloads 14 Views 5MB Size
Download PDF
Article pubs.acs.org/biochemistry

Mapping Functionally Important Residues in the Na+/Dicarboxylate Cotransporter, NaDC1 Claire Colas,† Avner Schlessinger,*,† and Ana M. Pajor*,‡ †

Department of Pharmacological Sciences, Icahn School of Medicine at Mount Sinai, New York, New York 10029, United States Skaggs School of Pharmacy and Pharmaceutical Sciences, University of CaliforniaSan Diego, La Jolla, California 92130-0714, United States

‡

S Supporting Information *

ABSTRACT: Transporters from the SLC13 family couple the transport of two to four Na+ ions with a di- or tricarboxylate, such as succinate or citrate. We have previously modeled mammalian members of the SLC13 family, including the Na+/ dicarboxylate cotransporter NaDC1 (SLC13A2), based on a structure of the bacterial homologue VcINDY in an inwardfacing conformation with one sodium ion bound at the Na1 site. In the study presented here, we modeled the outwardfacing conformation of rabbit and human NaDC1 (rbNaDC1 and hNaDC1, respectively) using an outward-facing model of VcINDY as a template and identified residues in or near the putative Na2 and Na3 cation binding sites. Guided by the structural models in both conformations, we performed site-directed mutagenesis in rbNaDC1 for residues proposed to be in the Na+ or substrate binding sites. Cysteine substitution of T474 in the predicted Na2 binding site results in an inactive protein. The M539C mutant has a low apparent affinity for both sodium and lithium cations, suggesting that M539 may form part of the putative Na3 binding site. The Y432C and T86C mutants have increased Km values for succinate, supporting their proposed location in the outward-facing substrate binding site. In addition, cysteine labeling by MTSEA-biotin shows that Y432C is accessible from the outside of the cell, and the accessibility changes in the presence or absence of Na+. The results of this study improve our understanding of substrate and ion recognition in the mammalian members of the SLC13 family and provide a framework for developing conformationally specific inhibitors against these transporters.

T

he Na+/dicarboxylate cotransporter NaDC1 is a member of solute carrier family 13 (SLC13), which is part of the larger divalent anion sodium symporter superfamily (DASS).1 NaDC1 is important for the plasma membrane transport of citric acid cycle intermediates, such as succinate and citrate, in the small intestine and renal proximal tubule. The transport mechanism of NaDC1 involves ordered binding of three sodium ions followed by a divalent anion substrate, with the net transport of one positive charge across the membrane. In the presence of Na+, Li+ inhibits transport by competing at one of the three cation binding sites, providing a useful chemical tool for characterizing the Na+ binding sites in this protein.2 Currently, there is no known atomic-resolution structure of any mammalian SLC13 transporter. Recently, X-ray structures were determined for a DASS family member, VcINDY from Vibrio cholerae, exhibiting a unique structural fold.3,4 VcINDY is a sodium-dependent dicarboxylate transporter that also couples three sodium ions to the transport of succinate.5 VcINDY shares a sequence identity of 33−35% with the mammalian SLC13 family as well as a highly conserved binding site and similar substrate specificity, making it the most suitable modeling template for the SLC13 family. The VcINDY structures were determined in an inward-facing conformation, © 2017 American Chemical Society

bound to citrate and succinate and two sodium ions (Na1 and Na2).3,4 A structure of VcINDY with human specific mutations in the substrate binding site has also been determined.4 The VcINDY structure with citrate and one sodium bound facilitated the characterization of various mammalian SLC13 family members, including the Na+/dicarboxylate cotransporters NaDC1 and NaDC3.6−9 The models of human NaDC1 and NaDC3 (hNaDC1 and hNaDC3, respectively) contain conserved binding sites with two hairpin loops, HPin and HPout; each loop contains a serine-asparagine-threonine (SNT) motif that coordinates binding to the ligand and two of the three sodium ions. We have previously characterized residues involved in substrate and ion binding and transport of hNaDC3, using modeling and cell-based methods.6 Specifically, we demonstrated that residues belonging to the N-terminal SNT motif were important for coordinating interactions with the substrate and Na1. T253, which is in the proximity of this SNT motif, has also been shown to be essential for substrate transport, and T485, which belongs to the C-terminal SNT Received: May 24, 2017 Revised: July 20, 2017 Published: July 21, 2017 4432

DOI: 10.1021/acs.biochem.7b00503 Biochemistry 2017, 56, 4432−4441

Article

Biochemistry

lian SLC13, VcINDY is considered the most appropriate modeling template for mammalian NaDC1. For each protein, 500 initial models were generated with MODELLER-9v.14, which were assessed with the statistical potential Z-DOPE.17 The initial NaDC1 models were built with a sodium ion and a citrate, whose coordinates were derived from the template. The top-scoring model exhibited a Z-DOPE score of −0.3, suggesting that ∼50% of their Cα atoms are within 3.5 Å of their correct positions, while the template exhibited a score of −0.4. Finally, the side chains of S140 and T142, which constitute the N-terminal SNT motif, were refined with PyMOL (PyMOL Molecular Graphics System, version 1.8, Schrödinger, LLC). The movements of the transport domain between the inward and outward models were measured with Profit (www.bioinf.org.uk/software/profit/) and DynDom.18 Conservation Analysis. Evolutionary conservation was calculated with the ConSurf server (http://consurf.tau.ac.il/ 2016/) for the outward-open models of VcINDY and rbNaDC1, using the default parameters.19 The conservation scores were mapped onto both models with UCSF Chimera, using the scripts available on the ConSurf server.20 Additionally, the hydrophobicity profile was computed, using UCSF Chimera with the Kyte and Doolittle hydrophobicity scale, and mapped on the models (Figure S2).21 Mutagenesis and Cell Culture. Site-directed mutagenesis was performed using the QuikChange mutagenesis kit (Stratagene) as described previously.22 The mutants were generated in the C476S mutant of rabbit NaDC1 in the pcDNA3.1 vector. The endogenous C476 was replaced because it is sensitive to cysteine specific reagents, and the other endogenous cysteines were left to allow an increased level of protein expression.23,24 The model indicates that there are no cysteines in the proximity to form disulfide bonds, and our previous study showed no difference between C476S and mutants with fewer endogenous cysteines.24 The parental and mutant NaDC1 transporters were expressed in human embryonic kidney (HEK-293) cells (CRL-1573, American Type Culture Collection, Manassas, VA). The cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 25 mM HEPES, 2 mM Glutamax, 1 mM sodium pyruvate, 0.1 mM nonessential amino acids, 10% heatinactivated fetal calf serum, 100 units/mL penicillin, and 100 μg/mL streptomycin at 37 °C in 5% CO2.25 For most transport experiments, the cells were plated on poly-D-lysine-coated 96well plates (BD Biosciences, San Jose, CA) at a density of 0.6 × 105 cells/well and transfected with plasmids encoding SLC13 transporters using FuGene6 (Roche Applied Science, Indianapolis, IN) at a 3:1 ratio. For biotinylation experiments, the cells were plated in six wells, and for succinate kinetics, the cells were plated in 24 wells. Transport Assays. Transport activity was measured 48 h after transfection, as described previously.25 The sodium transport buffer contained 140 mM NaCl, 2 mM KCl, 1 mM MgCl2, 1 mM CaCl2, and 10 mM HEPES (pH adjusted to 7.4 with 1 M Tris). Lithium buffer contained 140 mM LiCl in place of NaCl, and choline buffer contained 140 mM cholineCl in place of NaCl. For most of the assays, each well was washed twice with choline buffer and then incubated with 50 μL of sodium buffer containing ∼10 μM 2,3-[14C]succinate (52 mCi/ mmol, Moravek, Brea, CA) for 30 min at 37 °C. The uptake assays were stopped, and surface radioactivity was removed with 4 × 1 mL washes of choline buffer. The cell monolayers were dissolved using Ultima Gold scintillation cocktail

motif, was shown to be important for substrate transport and coordination of Na2, in agreement with the crystal structure. Interestingly, hNaDC1 includes a substitution in the equivalent residues, where the threonine C-terminal SNT motif is mutated to valine (V477), thereby causing differences in substrate specificity and cation affinity compared with those of other SLC13 family members.26 Finally, mutagenesis of T527 in hNaDC3, which is located in the proximity of the C-terminal SNT motif, altered sodium transport, providing additional confirmatory evidence of the estimated location of Na2.6 Overall, our models suggested that a combination of substitutions determines substrate specificities within the SLC13 family, in agreement with the partially humanized structure of VcINDY.4 In the absence of atomic-resolution structures in different conformational states, the technique of repeat-swap modeling has been used to develop structural models of transporters in the opposite conformation.10−12 The VcINDY structure contains a topology pattern of two halves of the protein related to each other by inverse pseudosymmetry.3 The repeatswap modeling approach, combined with a variety of biochemical assays that include cross-linking, produced an outward-facing model of VcINDY.13 This model suggested large movements of the substrate and cation binding domains in an elevator mechanism of transport similar to the transport mechanism observed in GltPh.13,14 In the elevator mechanism, the domain containing the binding site, which is also termed the transport domain, moves along the axis perpendicular to the membrane, whereas a scaffold domain remains stationary. VcINDY also contains an oligomerization domain that connects the two monomers, which is also stationary.13 This outwardfacing conformation model of VcINDY provided a unique template for modeling the mammalian SLC13 family members, including NaDC1, in an outward-facing conformation. The purpose of this study was to model NaDC1 from rabbit and human in the outward-facing conformation on the basis of the new VcINDY model. We also characterized experimentally the role of residues in rbNaDC1 that were identified by the models as being functionally important. Furthermore, we predicted the location of the third cation binding site of NaDC1, called Na3, by comparing our models to a related model of the Na+/phosphate cotransporter.15 Residues predicted to be important in rbNaDC1 were then tested experimentally using site-directed mutagenesis followed by cellbased functional assays. Finally, we discuss potential alternative explanations for our experimental results and the relevance of our results in light of previous studies characterizing SLC13 transporters.

■

MATERIALS AND METHODS Homology Modeling. We modeled the rabbit and human NaDC1 (rbNaDC1 and hNaDC1, respectively) in an outwardfacing conformation based on a model of the VcINDY dimer in an outward-facing conformation,13 and a previously published NaDC1−VcINDY alignment.7 The sequence of VcINDY is ∼35% identical with those of rbNaDC1 and hNaDC1, and VcINDY exhibits conserved ligand and ion binding sites. Moreover, VcINDY cotransports three sodium ions with a dicarboxylate with the same stoichiometry exhibited by rbNaDC1 and hNaDC1.5,16 Previous homology models of mammalian SLC13 members based on VcINDY captured key specificity determinants among members of this family.6,7 Thus, despite sharing relatively low sequence identity with mamma4433

DOI: 10.1021/acs.biochem.7b00503 Biochemistry 2017, 56, 4432−4441

Article

Biochemistry

Figure 1. rbNaDC1 homology model. rbNaDC1 model in the (A and C) inward-facing7 and (B and D) outward-facing conformations. The top panels show the dimers from the membrane bilayer plane and the bottom panels those from the extracellular side. The rbNaDC1 models are represented in cartoon format, with the oligomerization domains colored wheat, the scaffold domains orange, and the transport domains blue. The bound citrate is represented as cyan sticks and the sodium ion bound to Na1 as a purple sphere. The citrate and sodium coordinates are derived from the VcINDY templates.

lated proteins were precipitated using streptavidin-agarose resin (Thermo Scientific), while the remaining supernatants were transferred to new tubes for detection of total intracellular protein. MTSEA-biotin (Toronto Research Chemicals) was used to directly label extracellularly accessible cysteines, also as described previously.22 Specific labeling of the MTSEA-biotin was assessed by a 10 min preincubation with another cysteine selective reagent, MTSET, in sodium or choline buffer. Controls were preincubated in sodium buffer without MTSET. MTSEA-biotin was used at a concentration of 0.1 mM and MTSET at 1 mM. Western Blots. The biotinylated or intracellular protein samples were separated by sodium dodecyl sulfate−polyacrylamide gel electrophoresis and transferred to nitrocellulose, as described previously.22,26 The blots were probed for 1.5 h with the anti-NaDC1 antibody27 at a 1:1000 dilution in PBS-TM (phosphate-buffered saline with Tween 20 and milk). The secondary antibody, peroxide-conjugated anti-rabbit IgG (Jackson Laboratories), was applied for 1 h at a 1:7500 dilution. Antibody binding was detected with a SuperSignal West Pico Chemiluminescent substrate (Thermo-Fisher Scientific). Images were acquired with an Image Station 4000R imager (Carestream Scientific), and the intensity of protein bands was analyzed using Image 1D analysis software. NaDC1 typically has multiple protein bands representing differently glycosylated forms of the transporter.26,27 Statistics. The experiments were repeated with two or three different batches of transfected cells from different passage numbers. Triplicate or quadruplicate measurements were taken for each data point. Significant differences between groups were identified by one-way analysis of variance followed by

(PerkinElmer), and then the plates were counted directly using a Wallac Microbeta plate scintillation counter. For all experiments, counts in vector-transfected cells were subtracted from counts in NaDC1 plasmid-transfected cells to correct for background. Cation specificity experiments were performed by measuring succinate transport in either sodium, lithium, or choline buffers. We also compared the transport of different substrates by measuring the transport of 100 μM [14C]substrate (combination of radioactive and nonradioactive) all in sodium buffer. The substrates were 2,3-[14C]succinate, 1,5-[14C]citrate (112 mCi/mmol, Moravek), 1,5-[14C]glutarate (50 mCi/mmol, ARC, St. Louis, MO), and 5-[14C]ketoglutarate (53 mCi/ mmol, Moravek). Sodium activation experiments were performed as described previously,6 by replacing sodium with choline buffer. The sodium activation curves were fitted to the Hill equation. Lithium inhibition experiments were similar to a previous experiment with NaDC1.2 The sodium concentration was kept fixed at 112 mM, and a combination of lithium and choline buffer was used to produce between 0 and 28 mM Li+. The inhibition curves were fitted to a three-part hyperbolic decay equation: y0 + (Imax[Li+])/(K0.5 + [Li+]), where Imax represents the maximum inhibition at a saturating Li + concentration and K0.5 is the Li+ concentration that produces half-maximal inhibition. Succinate kinetic experiments were performed with [3H]succinate and nonradioactive succinate between 5 and 1000 μM at 37 °C for 15 min. Twenty-four-well plates were used to increase the intensity of the signal. Biotinylations. Sulfo-NHS-LC biotin (Thermo Scientific), which labels extracellular lysine residues, was used to detect cell surface rbNaDC1 protein, as described previously.22 Biotiny4434

DOI: 10.1021/acs.biochem.7b00503 Biochemistry 2017, 56, 4432−4441

Article

Biochemistry Dunnett’s test, with P values of 1. The mean KNa in C476S was 57 mM, similar to our previous measurement of 36 mM using the Xenopus oocyte expression system.24 The Y432C and M539C mutants had KNa values that were significantly higher than those in C476S, 216 and 125 mM, respectively (Figure 6 and Table 1). NaDC1 has ordered binding where the three cations bind first and produce a conformational change that increases the affinity for the substrate, and then the substrate binds.2,31 NaDC1 can also transport succinate using lithium instead of sodium; however, the lithium presumably produces a less optimal conformational change, and the Km for succinate is approximately 5-fold larger in lithium than in sodium.2,31 In the study presented here, the C476S parental transporter had approximately 8% of the succinate transport activity in lithium compared with that in sodium (Figure 7A). T86C was not significantly different from C476S, but the other mutants had decreased transport activity in lithium ranging from 1.5 to 4.5% of that in sodium. NaDC1 is also sensitive to inhibition by lithium when both lithium and sodium are present.2 One of the three cation binding sites has an affinity for lithium higher that

Figure 5. Specific MTSEA biotinylation of Y432C compared with C476S and T482C. HEK-293 cells expressing NaDC1 mutants were preincubated with (+) or without (−) 1 mM MTSET in sodium (Na) or choline buffer (Ch). The cells were then washed and incubated with MTSEA-biotin, and labeled proteins were pulled down using streptavidin-agarose beads, as outlined in Materials and Methods. The blots were probed with anti-rbNaDC1 antibodies. The top panel shows proteins labeled from the outside of the cell with MTSEAbiotin. The bottom panel shows intracellular proteins from the same experiment, as an indication of gel loading. The position of size standards is shown at the left.

biotin to the C476S parental transporter, as we showed previously,22 but this binding was not specific because it was not blocked by MTSET, another cysteine selective reagent (Figure 5). The T482C mutant was used as a positive control. Our previous studies showed that this substituted cysteine is very sensitive to extracellular labeling by MTS reagents and this cysteine changes accessibility with different transport conditions.24 There was strong labeling of T482C by MTSEAbiotin (Figures 4 and 5), and this labeling could be blocked by preincubating the cells in MTSET and sodium buffer. However, the MTSEA-biotin labeling was not blocked by preincubating the cells in MTSET and choline buffer (Figure 5). This is consistent with our previous study showing that T482C is most 4437

DOI: 10.1021/acs.biochem.7b00503 Biochemistry 2017, 56, 4432−4441

Article

Biochemistry

Figure 7. Effects of lithium on succinate transport. (A) Transport activity in lithium. The parental C476S-rbNaDC1 and mutants were expressed in HEK-293 cells. Transport of [14C]succinate (10 μM) was measured in 140 mM lithium or 140 mM sodium buffers at 37 °C for 30 min. Transport activity in lithium is shown as a percentage of the activity measured in sodium. Bars show means ± the standard error of the mean (SEM) (n = 4 except for N525C, for which n = 3). An asterisk denotes a significant difference from the C476S control group (P < 0.05). (B) Lithium inhibition in the presence of sodium. Transport buffers contained 112 mM sodium and between 0 and 28 mM lithium (replaced by cholineCl). [14C]Succinate transport was measured at 37 °C for 30 min. The data were normalized to the transport activity in the absence of lithium. The data points are the means ± SEM (n = 3 replicates from a single experiment). The summary of data from multiple experiments is given in Table 2.

and mutant transporters (Table 1). The mean Km for succinate in C476S was 38 μM, similar to the value of 25 μM measured in our previous study.22 The succinate Km values in Y228C (178 μM) and Y432C (506 μM) were significantly higher than in C476S, indicating a possible decrease in substrate affinity in these mutants. The succinate Km values in the L83C, T86S, and M539C mutants were not significantly different from thta of C476S. The activity of N525C was too low for accurate kinetic measurement (the sodium activation curves could be measured because the specific activity of [14C]succinate was higher). Note that the Km values for transporters, although used to estimate substrate binding properties, also include rate constants for translocation steps.32 We also compared the transport of 14C-labeled substrates in cells expressing C476S and mutants to identify potential changes in substrate selectivity (Figure 8). The L83C mutant was similar to the parental C476S. In contrast, the T86C, Y228C, and Y432C mutants exhibited decreased rates of transport of citrate, α-ketoglutarate, and glutarate relative to succinate. The N525C and M539C mutants exhibited similar rates of transport of citrate to the C476S parent; however, the rate of transport of α-ketoglutarate was increased, and the rate of transport of glutarate was decreased.

that for sodium, called the high-affinity cation binding site, which results in transport inhibition in the presence of lithium and sodium together.2 We examined whether the rbNaDC1 mutants had changes in their sensitivity to inhibition by lithium as a way to assess the high-affinity cation binding site (Figure 7B and Table 2). In C476S, the K0.5 for lithium inhibition was Table 2. Li+ Inhibition of Succinate Transport in NaDC1 Mutantsa mutant

K0.5Li (mM)

maximum inhibition (%)

no. of experiments

C476S L83C T86C Y228C Y432C N525C M539C

3.4 ± 0.9 2.6 ± 1.1 0.7 ± 0.1 1.8 ± 0.8 0.4 ± 0.1b 2.6 ± 0.6 0.9, CF,c CFc

65 ± 3 76 ± 2 88 ± 1b 65 ± 4 94 ± 2b 77 ± 6 27, CF,c CFc

4 3 3 3 3 3 3

a

The transport mixtures contained 112 mM Na+ and between 0 and 28 mM Li+, with the difference made up by cholineCl. The data shown are means ± SEM of three or four experiments; a single experiment is shown in Figure 5. bSignificantly different from that of the C476S group (P < 0.05). cCannot fit.

■

3.4 mM and the maximum inhibition was 65%. The Y432C mutant had a significant decrease in its K0.5 to 0.4 mM, and its maximal level of inhibition was 94%, indicating that this mutant is much more sensitive to inhibition by lithium than the parental C476S mutant is. The T86C mutant had an increased maximal level of inhibition compared with that of C476S, although the K0.5 was not significantly different. In contrast, the M539C mutant was insensitive to inhibition by lithium. In one experiment (Figure 7B), the K0.5 was 0.9 mM with a maximal inhibition of ∼27%, but in two other experiments, there was no inhibition and the curves could not be fitted accurately. Substrates. We next examined the succinate kinetics and substrate specificity of the cysteine mutants. First, the kinetics of succinate transport were compared in the C467S parental

DISCUSSION

Visualization of the interactions of transporters with substrates and ions in different transporter conformational states is needed to understand the mechanisms of secondary transport. Currently, atomic structures have been determined experimentally for only a small fraction of the mammalian SLC transporters. Recent improvements in computational modeling methods, coupled with newly determined structures of prokaryotic homologues, have provided a framework for describing specificity determinants in a variety of mammalian SLC transporters.10,33,34 In this study, we modeled the outward-facing conformation of the Na+/dicarboxylate cotrans4438

DOI: 10.1021/acs.biochem.7b00503 Biochemistry 2017, 56, 4432−4441

Article

Biochemistry

likely to be close to or part of the substrate binding site that is accessible from the outside of the cell. Because T86, L83, and Y432 are located in the proximity of the interface between the putative scaffold and mobile domains, it is plausible that the mutations of these residues disrupted helix packing or movement and had a minor direct effect on binding (Figure 2 and Table 1). It is also plausible that the sodium ion coordinates the binding of the two domains in a manner similar to that of a sodium ion in unrelated transporter families (e.g., LeuT). The sodium-coupled SLC transporters use the energy from the transmembrane electrochemical gradient for sodium to drive concentrative transport of substrates. In NaDC1, there is ordered binding with three Na+ ions binding first, resulting in an increased substrate affinity that allows the substrate to bind.37,38 The three Na+ ions show cooperative binding. The cation binding sites in rbNaDC1 will accept Li+ in place of Na+, but the Km for succinate in Li+ is more than 15 times greater than in Na+.2,31 Lithium is also an inhibitor of dicarboxylate transport when both Li+ and Na+ are present, mediated by binding at one site that preferentially binds Li+ over Na+.2 In most members of the SLC13 family, Li+ inhibits transport in the presence of sodium, indicating that it can compete with sodium.31,39 In the NaCT transporter (SLC13A5), which binds four Na+ ions, the Li+ can activate or inhibit transport depending on the substrate concentration.1,40 Despite the critical importance of the cations to the transport cycle in NaDC1, there is very little information about the structure of cation binding sites. The bacterial VcINDY structures show the location of two of the cation binding sites, Na1 and Na2,3,4 and the location of Na1 has been confirmed in NaDC3 (SLC13A3).6 The models in the study presented here predict the location in NaDC1 of the second and third cation binding sites, called Na2 and Na3, respectively. T373 in VcINDY is a conserved residue that forms part of the Na2 site, which corresponds to T474 in rbNaDC1 and T471 in hNaDC1 (Table S1). Indeed, mutation of this residue to cysteine resulted in an inactive protein, which highlights its critical role for transport. Our study has also proposed a possible location for the Na3 cation binding site that has a higher affinity for Li+ than for Na+. The M539C mutant in rbNaDC1 has impaired transport activity, although the protein is expressed on the plasma membrane. The main defect in rbNaDC1 M539C is in cation transport, with a decreased apparent sodium affinity and the loss of sensitivity to inhibition by lithium (Table S2). Therefore, the Na3 cation binding site in rbNaDC1 is likely to contain M539. The NaDC1 models visualize the outward-facing conformation, revealing functionally important residues that may constitute the substrate binding site in this conformation. For example, our experiments, guided by the structural model, indicate that Y432 in rbNaDC1 is a substrate binding residue that is accessible from the outside but not from the inside of the cell. The Y432C mutant has decreased transport activity despite no change in expression at the plasma membrane. This mutant also has a >10-fold increase in Km for succinate and a decreased transport activity of citrate, α-ketoglutarate, and glutarate relative to succinate, indicating that the affinity for other substrates may also have changed. Interestingly, in addition to changes in substrate binding, the Y432C mutation also affects cation binding, with decreased affinity for sodium and a weakened ability to transport in lithium. However, this mutant is more sensitive to inhibition by lithium. In Y432C, a lower

Figure 8. Substrate transport in NaDC1 mutants. NaDC1 mutants and the parental NaDC1-C476S were expressed in HEK-293 cells. Transport of 100 μM 14C-labeled substrates was measured at 37 °C for 30 min. The transport activity with each substrate is expressed as a percentage of the transport of succinate in the same experiment. The data shown are means ± SEM (n = 3); in the case of L83C, n = 2 experiments and the range is shown. An asterisk denotes a significant difference from the C476S control group (P < 0.05).

porter NaDC1 (SLC13A2) from rabbit and human, using a model of VcINDY in this conformation as a template.13 The new outward-facing NaDC1 models were built on the basis of two major approximations. First, as a modeling template, we used a model of VcINDY in the outward-facing state.13 This VcINDY model was constructed using the repeatswap technique, which is based on the internal pseudosymmetry of the transporter.10 Although the repeat-swap method has been used to characterize transporters from diverse structural families such as the MFS and LeuT,11,35 this method provides low-resolution structures that may not always be sufficient for describing specific interactions involving amino acid side chains. Second, NaDC1 is a membrane protein that shares sequence identity of only 35% with VcINDY, making it a challenging target for homology modeling. Despite the conservation in fold among the DASS family members, models that rely on evolutionarily divergent template structures may be inaccurate, and most modeling methods are not optimized for membrane protein targets.33,34,36 Nevertheless, despite these limitations in our approach, our model of NaDC1 captures key specificity determinants that were in agreement with a range of cellular assays as well as with multiple previous studies by us and others characterizing the SLC13 family.6,7 The new NaDC1 models and previously published models of NaDC1 and NaPi-IIa were used to model the putative Na2 and Na3 cation binding sites in NaDC1. We tested the predictions of six residues from the model with site-directed mutagenesis, and four of these were found to be functionally important in rbNaDC1. Notably, membrane transporters, including the SLC13 family, are highly dynamic proteins, and many of the residues have more than one clear “role”. For example, a residue can directly bind an ion and a substrate while also participating in mediating domain−domain interaction and allosteric communications. Our experiments suggest that T474 can be part of the Na2 cation binding site and M539 might be a part of the Na3 cation binding site, whereas Y432 and T86 are 4439

DOI: 10.1021/acs.biochem.7b00503 Biochemistry 2017, 56, 4432−4441

Biochemistry

■

concentration of lithium is needed to inhibit transport in the presence of sodium, and the amount of inhibition was greater. At present, we do not have an explanation for the changed cation interactions in Y432C. Further evidence of the location of Y432C is shown by its accessibility to cysteine specific reagents from the outside of the cell. The substituted cysteine can be labeled directly by MTSEA-biotin. The level of labeling is higher in the presence of sodium and is decreased when sodium is replaced by choline. This result indicates that Y432 changes accessibility to the outside of the cell during the transport cycle, either by moving or by being occluded by another part of the protein. The rbNaDC1 model predicts that T86 is also a substrate binding residue located on the outside of the cell. Consistent with this, the T86C mutant shows impaired succinate transport with an increase in succinate Km and decrease in Vmax, while the protein is expressed well on the plasma membrane. The T86C mutant did not undergo changes in apparent sodium affinity but showed increased sensitivity to lithium inhibition. T86C was not strongly labeled by MTSEA-biotin (specific labeling inhibited by MTSET), indicating that it is not easily accessible to the outside of the cell. This residue is conserved in the Na+/ citrate transporter NaCT as T86, and mutations at this position decreased the Vmax for citrate and decreased IC50 for the dicarboxylate inhibitor PF-06649298.9 The third residue predicted to be in the outward-facing substrate binding site, L83, is not likely to be a substrate binding residue, although it may be important in the interface between monomers.

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected]. Phone: (858) 822-7806. Fax: (858) 822-6857. *E-mail: [email protected]. ORCID

Ana M. Pajor: 0000-0001-6696-1136 Funding

This work was supported in part by National Institutes of Health Grant R01 GM108911 to A.S. and C.C. and by a University of California−San Diego Academic Senate Bridging grant to A.M.P. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We thank P. M. Ung for technical assistance and maintenance of the computational resources required for this study. This work was supported in part through the computational resources and staff expertise provided by the Department of Scientific Computing at the Icahn School of Medicine at Mount Sinai. We thank Drs. Lucy Forrest and Joe Mindell for providing us the coordinates of the VcINDY model in the outward-facing conformation.

■

ABBREVIATIONS DASS, divalent anion sodium symporter family; SLC13, solute carrier family 13; MTS, methanethiosulfonate; MTSEA-biotin, N-biotinaminoethylmethanethiosulfonate; MTSET, 2(trimethylammonium)ethylmethanethiosulfonate; PBS/CM, phosphate-buffered saline (pH 7.5) containing 1 mM Ca2+ and Mg2+.

■

CONCLUSION This work describes new homology models of rbNaDC1 and hNaDC1 and experimental testing of rbNaDC1 using a range of cell-based assays. The model effectively predicted key residues relevant for substrate and cation binding. We have proposed for the first time in an SLC13 family member that M539 in rbNaDC1 may be located in the Na3 sodium and lithium binding site and have shown that T474 is a critical residue likely to be part of the Na2 cation binding site. Furthermore, Y432 is a key component of the substrate binding site accessible to the outside of the cell in the presence of sodium, but not in its absence. The Y432C mutant has a decreased substrate affinity but also decreased cation affinity, suggesting proximity to cation binding. The results in this study improve our understanding of substrate and ion recognition in the mammalian SLC13 members and provide a framework for developing novel inhibitors against these transporters.

■

Article

■

REFERENCES

(1) Pajor, A. M. (2014) Sodium-coupled dicarboxylate and citrate transporters from the SLC13 family. Pfluegers Arch. 466, 119−130. (2) Pajor, A. M., Hirayama, B. A., and Loo, D. D. F. (1998) Sodium and lithium interactions with the Na+/dicarboxylate cotransporter. J. Biol. Chem. 273, 18923−18929. (3) Mancusso, R., Gregorio, G. G., Liu, Q., and Wang, D. N. (2012) Structure and mechanism of a bacterial sodium-dependent dicarboxylate transporter. Nature 491, 622−627. (4) Nie, R., Stark, S., Symersky, J., Kaplan, R. S., and Lu, M. (2017) Structure and function of the divalent anion/Na+ symporter from Vibrio cholerae and a humanized variant. Nat. Commun. 8, 15009. (5) Fitzgerald, G. A., Mulligan, C., and Mindell, J. A. (2017) A general method for determining secondary active transporter substrate stoichiometry. eLife 6, e21016. (6) Schlessinger, A., Sun, N. N., Colas, C., and Pajor, A. M. (2014) Determinants of substrate and cation transport in the human Na+/ dicarboxylate cotransporter, NaDC3. J. Biol. Chem. 289, 16998−17008. (7) Colas, C., Pajor, A. M., and Schlessinger, A. (2015) StructureBased Identification of Inhibitors for the SLC13 Family of Na(+)/ Dicarboxylate Cotransporters. Biochemistry 54, 4900−4908. (8) Klotz, J., Porter, B. E., Colas, C., Schlessinger, A., and Pajor, A. M. (2016) Mutations in the Na+/citrate cotransporter NaCT (SLC13A5) in pediatric patients with epilepsy and developmental delay. Mol. Med. 22, 310−321. (9) Pajor, A. M., de Oliveira, C. A., Song, K., Huard, K., Shanmugasundaram, V., and Erion, D. M. (2016) Molecular basis for inhibition of the Na+/citrate transporter NaCT (SLC13A5) by dicarboxylate inhibitors. Mol. Pharmacol. 90, 755−765. (10) Vergara-Jaque, A., Fenollar-Ferrer, C., Kaufmann, D., and Forrest, L. R. (2015) Repeat-swap homology modeling of secondary

ASSOCIATED CONTENT

* Supporting Information S

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.biochem.7b00503. A description of the hNaDC1 models in inward- and outward-facing conformations (Figure S1), an analysis of the conservation and hydrophobicity profiles of rbNaDC1 and vcINDY in outward-facing conformations (Figure S2), a comparison of the outward-facing models of NaPi-IIa and rbNaDC1, a comparison of amino acid substitutions within the SLC13 members for the residues selected for mutations (Table S1), and the functional effects observed for each mutated residue in rbNaDC1 (Table S2) (PDF) 4440

DOI: 10.1021/acs.biochem.7b00503 Biochemistry 2017, 56, 4432−4441

Article

Biochemistry active transporters: updated protocol and prediction of elevator-type mechanisms. Front. Pharmacol. 6, 183. (11) Forrest, L. R., Zhang, Y. W., Jacobs, M. T., Gesmonde, J., Xie, L., Honig, B. H., and Rudnick, G. (2008) Mechanism for alternating access in neurotransmitter transporters. Proc. Natl. Acad. Sci. U. S. A. 105, 10338−10343. (12) Forrest, L. R., Kramer, R., and Ziegler, C. (2011) The structural basis of secondary active transport mechanisms. Biochim. Biophys. Acta, Bioenerg. 1807, 167−188. (13) Mulligan, C., Fenollar-Ferrer, C., Fitzgerald, G. A., VergaraJaque, A., Kaufmann, D., Li, Y., Forrest, L. R., and Mindell, J. A. (2016) The bacterial dicarboxylate transporter VcINDY uses a two-domain elevator-type mechanism. Nat. Struct. Mol. Biol. 23, 256−263. (14) Reyes, N., Ginter, C., and Boudker, O. (2009) Transport mechanism of a bacterial homologue of glutamate transporters. Nature 462, 880−885. (15) Fenollar-Ferrer, C., Patti, M., Knopfel, T., Werner, A., Forster, I. C., and Forrest, L. R. (2014) Structural fold and binding sites of the human Na(+)-phosphate cotransporter NaPi-II. Biophys. J. 106, 1268− 1279. (16) Mulligan, C., Fitzgerald, G. A., Wang, D. N., and Mindell, J. A. (2014) Functional characterization of a Na+-dependent dicarboxylate transporter from Vibrio cholerae. J. Gen. Physiol. 143, 745−759. (17) Shen, M. Y., and Sali, A. (2006) Statistical potential for assessment and prediction of protein structures. Protein Sci. 15, 2507− 2524. (18) Poornam, G. P., Matsumoto, A., Ishida, H., and Hayward, S. (2009) A method for the analysis of domain movements in large biomolecular complexes. Proteins: Struct., Funct., Genet. 76, 201−212. (19) Ashkenazy, H., Erez, E., Martz, E., Pupko, T., and Ben-Tal, N. (2010) ConSurf 2010: calculating evolutionary conservation in sequence and structure of proteins and nucleic acids. Nucleic Acids Res. 38, W529−W533. (20) Pettersen, E. F., Goddard, T. D., Huang, C. C., Couch, G. S., Greenblatt, D. M., Meng, E. C., and Ferrin, T. E. (2004) UCSF Chimera–a visualization system for exploratory research and analysis. J. Comput. Chem. 25, 1605−1612. (21) Kyte, J., and Doolittle, R. F. (1982) A simple method for displaying the hydropathic character of a protein. J. Mol. Biol. 157, 105−132. (22) Weerachayaphorn, J., and Pajor, A. M. (2007) Sodiumdependent extracellular accessibility of Lys-84 in the sodium/ dicarboxylate cotransporter. J. Biol. Chem. 282, 20213−20220. (23) Pajor, A. M., Krajewski, S. J., Sun, N., and Gangula, R. (1999) Cysteine residues in the Na+/dicarboxylate cotransporter, NaDC-1. Biochem. J. 344, 205−209. (24) Pajor, A. M. (2001) Conformationally-sensitive residues in transmembrane domain 9 of the Na+/dicarboxylate cotransporter. J. Biol. Chem. 276, 29961−29968. (25) Pajor, A. M., and Sun, N. N. (2013) Non-steroidal antiinflammatory drugs and other anthranilic acids inhibit the Na+/ dicarboxylate symporter from Staphylococcus aureus. Biochemistry 52, 2924−2932. (26) Pajor, A. M., and Sun, N. N. (2010) Single nucleotide polymorphisms in the human Na+-dicarboxylate cotransporter affect transport activity and protein expression. Am. J. Physiol Renal Physiol 299, F704−F711. (27) Pajor, A. M., Sun, N., and Valmonte, H. G. (1998) Mutational analysis of histidines in the Na+/dicarboxylate cotransporter, NaDC-1. Biochem. J. 331, 257−264. (28) Drew, D., and Boudker, O. (2016) Shared Molecular Mechanisms of Membrane Transporters. Annu. Rev. Biochem. 85, 543−572. (29) Vergara-Jaque, A., Fenollar-Ferrer, C., Mulligan, C., Mindell, J. A., and Forrest, L. R. (2015) Family resemblances: A common fold for some dimeric ion-coupled secondary transporters. J. Gen. Physiol. 146, 423−434. (30) Fenollar-Ferrer, C., Forster, I. C., Patti, M., Knoepfel, T., Werner, A., and Forrest, L. R. (2015) Identification of the first sodium

binding site of the phosphate cotransporter NaPi-IIa (SLC34A1). Biophys. J. 108, 2465−2480. (31) Wright, E. M., Wright, S. H., Hirayama, B. A., and Kippen, I. (1982) Interactions between lithium and renal transport of Krebs cycle intermediates. Proc. Natl. Acad. Sci. U. S. A. 79, 7514−7517. (32) Parent, L., Supplisson, S., Loo, D. D. F., and Wright, E. M. (1992) Electrogenic properties of the cloned Na+/glucose cotransporter: II. A transport model under nonrapid equilibrium conditions. J. Membr. Biol. 125, 63−79. (33) Schlessinger, A., Yee, S. W., Sali, A., and Giacomini, K. M. (2013) SLC classification: an update. Clin. Pharmacol. Ther. 94, 19− 23. (34) Colas, C., Ung, P. M., and Schlessinger, A. (2016) SLC Transporters: Structure, Function, and Drug Discovery. MedChemComm 7, 1069−1081. (35) Radestock, S., and Forrest, L. R. (2011) The alternating-access mechanism of MFS transporters arises from inverted-topology repeats. J. Mol. Biol. 407, 698−715. (36) Baker, D., and Sali, A. (2001) Protein structure prediction and structural genomics. Science 294, 93−96. (37) Meinild, A. K., Loo, D. D., Pajor, A. M., Zeuthen, T., and Wright, E. M. (2000) Water transport by the renal Na+-dicarboxylate cotransporter. Am. J. Physiol. 278, F777−F783. (38) Yao, X., and Pajor, A. M. (2000) The transport properties of the human renal Na+/dicarboxylate cotransporter under voltage clamp conditions. Am. J. Physiol. 279, F54−F64. (39) Pajor, A. M., Hirayama, B. A., and Wright, E. M. (1992) Molecular evidence for two renal Na+/glucose cotransporters. Biochim. Biophys. Acta, Biomembr. 1106, 216−220. (40) Inoue, K., Zhuang, L., Maddox, D. M., Smith, S. B., and Ganapathy, V. (2003) Human sodium-coupled citrate transporter, the orthologue of Drosophila Indy, as a novel target for lithium action. Biochem. J. 374, 21−26.

4441

DOI: 10.1021/acs.biochem.7b00503 Biochemistry 2017, 56, 4432−4441