THE JOURNAL OF BIOLOGICAL CHEMISTRY VOL. 289, NO. 13, pp. 8799 –8809, March 28, 2014 © 2014 by The American Society for Biochemistry and Molecular Biology, Inc. Published in the U.S.A.

Identification of the Intracellular Gate for a Member of the Equilibrative Nucleoside Transporter (ENT) Family* Received for publication, January 7, 2014, and in revised form, January 31, 2014 Published, JBC Papers in Press, February 4, 2014, DOI 10.1074/jbc.M113.546960

Raquel Valdés‡, Johannes Elferich§, Ujwal Shinde§, and Scott M. Landfear‡1 From the Departments of ‡Molecular Microbiology and Immunology and §Biochemistry and Molecular Biology, Oregon Health and Science University, Portland, Oregon 97239 Background: Equilibrative nucleoside transporters (ENTs) play important roles in biology and disease. Results: The intracellular gate for an ENT, LdNT1.1, has been identified by computational and experimental methods. Conclusion: The intracellular gate consists of the ends of four transmembrane helices. Significance: An important functional component of ENTs has been identified. Equilibrative nucleoside transporters of the SLC29 family play important roles in many physiological and pharmacological processes, including import of drugs for treatment of cancer, AIDS, cardiovascular, and parasitic diseases. However, no crystal structure is available for any member of this family. In previous studies we generated a computational model of the Leishmania donovani nucleoside transporter 1.1 (LdNT1.1) that captured this permease in the outward-closed conformation, and we identified the extracellular gate. In the present study we have modeled the inward-closed conformation of LdNT1.1 using the crystal structure of the Escherichia coli fucose transporter FucP and have identified four transmembrane helices whose ends close to form a predicted intracellular gate. We have tested this prediction by site-directed mutagenesis of relevant helix residues and by cross-linking of introduced cysteine pairs. The results are consistent with the predictions of the computational model and suggest that a similarly constituted gate operates in other members of the equilibrative nucleoside transporter family.

Purine nucleosides and their derivatives play pivotal roles in a variety of cellular and metabolic processes, including energy production, cell signaling, and synthesis of nucleic acids. Clinically, nucleoside and nucleobase analogs are also widely used in anticancer, antiviral, and anti-protozoal therapies (1, 2). Because physiological nucleosides and the majority of their synthetic analogs are hydrophilic molecules, specialized nucleoside transporter proteins are required to facilitate their movement across biological membranes. Thus, nucleoside transporters play key roles in nucleoside physiology, pathophysiology, and in the therapeutic actions of many nucleoside drugs (1–5).

The solute carrier 29 (SLC29)2 (4) family, referred to as equilibrative nucleoside transporters (ENTs), is an important class of permeases that enable facilitated diffusion of nucleosides, nucleobases, and nucleoside analogs into a wide range of eukaryotes (2– 4), although some protozoan members are active proton-coupled symporters that harness energy from the electrochemical gradient across the plasma membrane to concentrate their substrates within the cell (6 – 8). Purine nucleoside and nucleobase transporters are of particular interest in parasitic protozoa such as Leishmania, Trypanosoma, and Plasmodium, as none of these organisms is able to synthesize purines de novo, and they are completely reliant upon uptake of preformed purines from their hosts via SLC29 permeases (9). Currently no ENT crystal structure is known. Additionally, ENT-like bacterial homologs have not been identified, and therefore, purification and crystallization of these permeases remains a challenge. Support for structural commonality between ENTs and the major facilitator superfamily (MFS) of transporters was recently provided by site-directed mutagenesis studies of mammalian and protozoa ENTs, suggesting a common evolutionary origin (10) and similar packing of transmembrane (TM) helices around a solvent-accessible permeant binding site (11). These observations allowed the construction, by computational approaches, of putative tertiary structures of several protozoan ENTs, including Plasmodium falciparum PfENT1 (12), Trypanosoma brucei TbNBT1 (13), and Leishmania donovani LdNT2 (14). and LdNT1.1 (15). The ab initio computational studies on the LdNT1.1 adenosine/pyrimidine nucleoside transporter of L. donovani (15) captured this protein in a conformation that is “closed to the outside, open to the inside,” similar in the disposition of TMs and conformation to that observed in the crystal structures of the Escherichia coli lactose permease (LacY) (16) and glycerol 3-phosphate transporter (GlpT) (17), both members of the MFS family. The ab

2

* This work was supported, in whole or in part, by National Institutes of Health Grant AI44138 (to S. M. L.). This work was also supported by National Science Foundation Grant 0746589 (to U. S.). 1 To whom correspondence should be addressed: Dept. of Molecular Microbiology and Immunology, Oregon Health & Science University, 3181 S.W. Sam Jackson Park Rd., Portland, OR 97239. Tel.: 503-494-2426; Fax: 503494-6862; E-mail: [email protected].

MARCH 28, 2014 • VOLUME 289 • NUMBER 13

The abbreviations used are: SLC, solute carrier family; ENT, equilibrative nucleoside transporter; hENT, human ENT; LdNT, L. donovani nucleoside transporter; TM, transmembrane helix; LacY, E. coli lactose permease; MFS, major facilitator superfamily; FucP, E. coli fucose transporter; MTS-3-MTS, 1,3-propanediyl bismethanethiosulfonate; MTS-17-O5-MTS, 3,6,9,12,15pentaoxaheptadecane-1,17-diylbismethanethiosulfonate; o-PDM, N,N⬘(1,2-phenylene) dimaleimide; p-PDM, N,N⬘-(1,4-phenylene) dimaleimide; Bis-Tris, 2-[bis(2-hydroxyethyl)amino]-2-(hydroxymethyl)propane-1,3-diol.

JOURNAL OF BIOLOGICAL CHEMISTRY

8799

Inner Gate of an Equilibrative Nucleoside Transporter initio model predicted that most of the LdNT1.1 mutants that we previously found to greatly impair transport activity or alter substrate specificity (15, 18) line the central permeation pore, as might be expected for mutations that induce strong transport phenotypes. In addition, analogous to the experimentally determined structures for the bacterial permeases LacY and GlpT, the LdNT1.1 ab initio model predicted that TM helices 1, 2, and 7 cluster together at the extracellular surface of the transporter to close off the pore or permeation pathway and form an “extracellular gate” (15, 19). Two lines of experimental evidence support the identification of this extracellular gate. First, the computational model predicts that Phe-48 in transmembrane domain 1 (TM1) and Trp-75 in TM2 interact to tether these two helices together when the extracellular gate is closed. Indeed, mutation of either residue abrogates transport activity. Second, paired cysteine residues introduced between TM1–2, TM1–7, and TM2–7 could be cross-linked to impair transport activity. The presence of such a gate is consistent with the commonly invoked “alternating access model” for membrane transport (20, 21) in which alternating opening and closure of an extracellular and an intracellular gate make substrate sequentially accessible to one side or the other of the membrane during the transport cycle (see Fig. 1C). The majority of structurally characterized transporters have only been experimentally captured in one conformational state. Therefore, insight into different conformations representing steps in the transport cycle often has to be inferred by analogy with transport proteins with the same overall fold crystallized in a distinct conformational state. Thus crystal structures of several Na⫹ symporters that are members of distinct sequence families with similar structural folds allowed capture of such permeases in different conformations thought to represent six sequential steps in a common transport cycle (22, 23). Similarly, three members of the MFS, LacY, GlpT, and a multidrug resistance transporter EmrD, have been crystallized and structurally resolved in the outward-closed conformation (16, 17, 24). Recently the crystal structure of another MFS member, the E. coli fucose/H⫹ symporter FucP (25), provided the first image of an MFS member in the inward-closed conformation. Given the success in defining the extracellular gate of LdNT1.1 (15, 19), the present work aimed to elucidate the structural components that constitute the intracellular gate located on the opposite or cytoplasmic face of the permease. The rationale for this work was that together these studies identify two critical structural components of this ENT family permease, the two gates that define the outward-closed/inward-open and inward-closed/outward-open conformations. Because the original ab initio structural model of LdNT1.1 suggested that it was conformationally related to members of the MFS family, we postulated that the FucP structure could provide a model for LdNT1.1 in the inward-closed conformation. This approach is analogous to the modeling of distinct transport intermediates for Na⫹ symporters (22, 23), which also employed conformationally related structures of permeases from different sequence families. A homology model of LdNT1.1 based on the FucP structure predicted that the cytosolic ends of TM4, -5, -10, and -11 constitute the intracellular gate, and it predicted specific regions of these helices that are

8800 JOURNAL OF BIOLOGICAL CHEMISTRY

likely to interact with each other. These predictions were subsequently tested by site-directed mutagenesis and chemical cross-linking. Notably, comparison of critical intracellular gate residues of LdNT1.1 with the sequence of other ENT permeases, including the human nucleoside transporters, suggests that components of the intracellular gate are shared between multiple members of this family of carriers. Hence these studies on LdNT1.1 illuminate the structure and function of the larger ENT family that is widely represented among eukaryotes.

EXPERIMENTAL PROCEDURES Chemicals, Materials, and Reagents—Restriction endonucleases and DNA-modifying enzymes were obtained from New England Biolabs, Inc., Roche Applied Science, or Invitrogen. Radiolabeled [2,8-3H]adenosine (50 Ci/mmol) was purchased from Moravek Biochemicals. Synthetic oligonucleotides were purchased from Invitrogen. 1,3-Propanediyl bismethanethiosulfonate (MTS-3-MTS) and 3,6,9,12,15-pentaoxaheptadecane-1,17-diyl bismethanethiosulfonate (MTS-17-O5-MTS) were purchased from Toronto Research Chemical Inc. N,N⬘(1,2-Phenylene) dimaleimide (o-PDM), N,N⬘-(1,4-Phenylene) dimaleimide (p-PDM), and dithiothreitol (DTT) were purchased from Sigma. All other chemicals, materials, and reagents were of the highest grade commercially available. Homology Modeling—The sequence of LdNT1.1 was first aligned to FucP using a Hidden Markov profile comparison, generated by HHsearch software, employing the HHPred web server and default parameters (26). This alignment produced a probability score of 98.4% and an E value 2.0 ⫻ 10⫺5. The profile for LdNT1.1 was built automatically by searching a profile database based on uniprot clustered at 20% sequence identity using HHblits (27). This profile for FucP was that catalogued in the HHPred database. A homology model of LdNT1.1 was then generated from this alignment and the FucP crystal structure employing MODELLER software (28) on this same server. The model was then visualized using PyMOL software (PyMOL Molecular Graphics System, Schrödinger LLC). Parasite Cell Cultures—L. donovani strains were propagated at 26 °C in RPMI medium (Invitrogen) containing 10% fetal bovine serum, 15 ␮g/ml hemin, and 100 ␮M xanthine. Null mutant ⌬ldnt1/⌬ldnt2 (29) was supplemented with drugs against the integrated resistance markers (50 ␮g/ml hygromycin (Roche Applied Science) plus 50 ␮g/ml phleomycin (Research Products International) as well as drugs that are cytotoxic to parasites expressing the wild-type LdNT1.1 or LdNT2 transporters (1 ␮M tubercidin (Sigma) or 1 ␮M formycin B (Berry & Associates), respectively). Parasites transfected with pX63NEORI constructs, described below, were selected and maintained in 100 ␮g/ml neomycin analog G418 (Invitrogen). Site-directed Mutagenesis and Plasmid Constructs—Mutagenesis was performed using the QuikChange威 II XL sitedirected mutagenesis kit, a polymerase chain reaction-based mutagenesis strategy (Stratagene). Mutagenic primers were designed to incorporate the desired mutation within the LdNT1.1 open reading frame (ORF) template that had been ligated into the EcoRI site of the episomal expression vector pX63NEORI (30). The codons used to introduce the point mutations are : T160A (ACG 3 GCG), Y161A (TAC 3 GCC), G162A VOLUME 289 • NUMBER 13 • MARCH 28, 2014

Inner Gate of an Equilibrative Nucleoside Transporter (GGC 3 GCC), M163A (ATG 3 GCG), F164A (TTC 3 GCC), F167A (TTC 3 GCC), T174A (ACC 3 GCC), M175A (ATG 3 GCG), M176A (ATG 3 GCG), L444A (CTG 3 GCG), V445A (GTG 3 GCG), L446A (CTT 3 GCT), M466A (ATG 3 GCG), G467A (GGC 3 GCC), I468A (ATC 3 GCC), S469A (TCC 3 GCC), I470A (ATC 3 GCC), and L471A (CTC 3 GCC). Single or paired double cysteines were also introduced within either the wild type or the cysteine-less (Cys-less) LdNT1.1 ORF templates that had been ligated into the EcoRI site of the episomal expression vector pX63NEORI. The codons used to introduce the point mutations were L99C (CTC 3 TGC), S158C (TCC 3 TGC), T172C (ACG 3 TGT), M176C (ATG 3 TGT), M442C (ATG 3 TGT), L446C (CTT 3 TGT), G447C (GGC 3 TGC), A462C (GCC 3 TGC), L465C (CTG 3 TGT), and G467C (GGC 3 TGC). For each mutant, two independent clones were isolated in parallel, and the presence of the mutations was verified by DNA sequencing at the Oregon Health & Science University Microbiology Research Core Facility using a model 377 Applied Biosystems automated fluorescence sequencer (PerkinElmer Life Sciences). Wild type, Cys-less, and mutant LdNT1.1 pX63NeoRI constructs were transfected into transport defective ⌬ldnt1/⌬ldnt2 L. donovani promastigotes using standard electroporation conditions (31, 32). Transfectants were selected and expanded in liquid medium containing 100 ␮g/ml G418 (Invitrogen). Transport Assays—L. donovani promastigotes expressing the wild type and the Cys-less LdNT1.1 transporters and the different mutant permeases were harvested in early-middle logarithmic phase (0.5–1.5 ⫻ 107 cells/ml), collected by centrifugation at 2000 rpm (835 ⫻ g) for 10 min at room temperature (⬃25 °C), washed 2 times in phosphate-buffered saline (PBS: 138 mM NaCl, 8.1 mM Na2HPO4䡠7H20, 2.7 mM KCl, 1.5 mM KH2PO4 (pH 7.4)), and resuspended in PBS to a final density of ⬃2 ⫻ 108 parasites/ml. The adenosine transport measurements (1 ␮M, 2.5 ␮Ci/ml, 40s) depicted in the figures were performed at least three times (n ⱖ 3), and values within a given experiment were the mean of triplicate determinations using 100-␮l aliquots by the previously described oil-stop method (33). For each mutant, two independent clones were isolated and tested in parallel to confirm each result (data not shown). In Vivo and in Vitro Disulfide Cross-linking—Both in vivo (transfected parasites) and in vitro (membrane preparations) disulfide cross-linking reactions were performed at room temperature (⬃25 °C). To induce the formation of disulfide bonds, transfected parasites, collected by centrifugation and resuspended in PBS (500 ␮l, ⬃2 ⫻ 108 cells/ml) and/or membrane preparations (10 –20 ␮g of protein; 1–2 mg/ml), were incubated with MTS-3-MTS (4 min, 0.05 mM), MTS-17-O5-MTS (10 min, 0.05 mM), o-PDM (4 min, 0.05 mM), and p-PDM (4 min, 0.05 mM) before transport measurements or SDS-polyacrylamide gel electrophoresis (SDS-PAGE), respectively. All stock solutions were freshly prepared for each experiment. For preparations of cells to be subsequently tested for reduction of the formed disulfide bonds, parasites were washed 3 times with PBS and resuspended in 500 ␮l of PBS followed by incubation with the reducing agent DTT at room temperature (⬃25 °C) before transport measurements. For DTT treatment, efficiency of reduction while preserving cell viability was found to be optimal at 5 mM/30 min. For monitoring disulfide formation by MARCH 28, 2014 • VOLUME 289 • NUMBER 13

immunoblotting, cross-linked membranes were diluted into 4⫻ NuPAGE威 LDS sample buffer (Invitrogen) supplemented with 5 mM DTT and analyzed by SDS/PAGE as described below. Integral Membrane Protein Preparations—Transfected L. donovani parasites in early-middle logarithmic phase (0.5– 1.5 ⫻ 107 cells/ml, 100 ml) were collected by centrifugation at 2000 rpm (835 ⫻ g) for 10 min at room temperature and washed 2 times with PBS. Plasma membranes were prepared as described previously (34), and the pellets were dissolved in 50 –100 ␮l of resuspension buffer (20 mM Tris-HCl, 100 mM NaCl, 2 mM CaCl2, and 2% SDS (pH 7.4)) supplemented with protease inhibitor mixture (Complete, Hoffmann-La Roche) and stored at ⫺80 °C as aliquots for further use. Protein quantification was determined using the Bio-Rad DC Protein Assay kit. Membrane preparations used for cross-linking were diluted in the aforementioned buffer at a protein concentration of 1–2 mg/ml. Immunoblot Analysis—10 –20 ␮g of membrane preparations were mixed with the corresponding volume of 4⫻ NuPAGE威LDS sample buffer in the presence of DTT (5 mM) and heated at 70 °C for 10 min. Samples were fractionated by SDS-PAGE on 4 –12% gradient NuPAGE威 Novex Bis-Tris gels (Invitrogen) in the presence of NuPAGE威 reducing agent using the XCell SureLockTM Mini-Cell system (Invitrogen). Subsequently, proteins were electro-transferred under denaturing conditions onto nitrocellulose membranes (Protran, Whatman, GmbH) using the XCell II™ Blot Module (Invitrogen). Nitrocellulose membranes were then blocked with 5% fat-free milk in PBS (pH 7.4) containing 0.2% Tween 20 (PBS-T) (overnight at 4 °C). After a wash with PBS-T (5 min, room temperature), membranes were incubated with rabbit polyclonal antiLeishmania mexicana NT1-loop VII (Asn-236 –Lys-331) antibody as previously detailed (35) (dilution 1:3000 in PBS-T with 5% fat-free milk) for 1 h at room temperature. The blots were then washed 3⫻ with PBS-T (10 min/wash) and incubated for 1 h at room temperature with anti-rabbit IgG antibody conjugated to horseradish peroxidase (dilution 1:10,000 in PBS-T with 5% fat-free milk). After one 10-min wash with PBS-T and two 10-min washes with PBS, proteins were visualized using ECL detection reagents (Thermo Scientific) and exposure to film (Kodak BioMax MR film). Biotinylation of Surface Protein—Transfected L. donovani parasites in early-middle logarithmic phase (0.5–1.5 ⫻ 107 cells/ml, 50 ml) were collected by centrifugation at 2000 rpm (835 ⫻ g) for 10 min, washed 2 times with ice-cold PBS (pH 8.0), resuspended in ice-cold PBS (pH 8.0) at 5 ⫻ 107 cells/ml, and incubated on ice for 30 min with freshly prepared 10 mM EZ-LinkTM Sulfo-NHS-SS-Biotin (Thermo Scientific) (80 ␮l of biotin solution per 1 ml of cell suspension). The biotinylation reaction was quenched by washing the cells three times with ice-cold PBS, 100 mM Tris (pH 8.0). Parasites were then incubated on ice for 15 min in 300 ␮l of lysis buffer (PBS (pH 8.0), 1% Nonidet P-40, and protease inhibitor mixture), and the cell lysate was pelleted at 4 °C in a microcentrifuge (15,000 ⫻ g) for 15 min, after which the supernatant was removed to a new microcentrifuge tube and incubated overnight at 4 °C with 100 ␮l of packed streptavidin beads (Invitrogen) prewashed in iceJOURNAL OF BIOLOGICAL CHEMISTRY

8801

Inner Gate of an Equilibrative Nucleoside Transporter cold PBS (pH 8.0). After the overnight incubation, the beads were washed 3 times with 100 mM Tris (pH 8.0), 8 M urea buffer and resuspended in the corresponding volume of 4⫻ NuPAGE威LDS sample buffer in the presence of DTT (5 mM). All samples were fractionated by SDS-PAGE on a 4 –12% slab gel, and biotinylated LdNT1.1 protein was detected by immunoblotting with the NT1-loop VII antibody (as described above). As a loading control, blots were stripped with RestoreTM Western blot Stripping Buffer (Thermo Scientific) and reprobed with total antiserum against the LdMIT transporter (36) diluted 1:1000 followed by anti-mouse IgG antibody conjugated to horseradish peroxidase (dilution 1:10,000). Multiple Sequence Alignments—Alignments of the LdNTs and human ENTs (hENTs) (Table 1) were performed using ClustalW Version 1.4 (37) with the Blosum similarity matrix, an open gap penalty of 10, an extend gap penalty of 0.05, a delay divergent of 40%, and a gap distance of 8.

RESULTS Prediction of Inward-Closed Conformation of LdNT1.1 by Homology Modeling—Because the previous computational model of LdNT1.1 captured the permease in an outwardclosed/inward-open conformation (15) and allowed definition of the extracellular gate, a model in the “alternate” conformation was required to define the structural components of the intracellular gate. Because ENT or SLC29 family members appear to fold similarly to MFS family permeases (12–15), we employed homology modeling of LdNT1.1 to the only member of the MFS whose structure has been solved in an inwardclosed conformation, the E. coli FucP transporter (Fig. 1). As a starting point for modeling, we aligned LdNT1.1 with FucP using the HHsearch algorithm (26), which was developed to optimize alignment between proteins of limited sequence homology by employing sequence profile comparisons and secondary structure predictions, and the three-dimensional model was then generated using MODELLER software (28). As anticipated from the FucP structure, this model predicts that the intracellular ends of TM helices 4, 5, 10, and 11 cluster together to close the aqueous channel in this conformation of LdNT1.1. Selection of Candidate Amino Acid Residues for Site-directed Mutagenesis and Functional Characterization of LdNT1.1 Mutants—To experimentally test the model for the intracellular gate and to determine which residues are likely important for gate function, we selected for mutagenesis a battery of residues near the intracellular end of each of the helices proposed to close the permeation pathway from the intracellular side (TM4: Thr-160, Tyr-161, Gly-162, Met-163, Phe-164, Phe-167; TM5: Thr-174, Met-175, Met-176; TM10: Leu-444, Val-445, Leu-446; TM11: Met-466, Gly-467, Ile-468, Ser-469, Ile-470, Leu-471), and we individually mutagenized each residue to alanine. The resultant mutant permeases were expressed in the transport-defective ⌬ldnt1⌬ldnt2 L. donovani double null mutant that is genetically deficient in the LdNT1.1, LdNT1.2, and LdNT2 genes and consequently provides a null background for transport of nucleosides (29). The ability of each transgenically expressed mutant transporter to mediate ligand translocation was evaluated by uptake assays using 1 ␮M [3H]adenosine, a natural substrate of LdNT1.1. As shown in Fig. 2A,

8802 JOURNAL OF BIOLOGICAL CHEMISTRY

FIGURE 1. Homology model of LdNT1.1 in the outward-open/inwardclosed conformation. TM helices are indicated as cylinders and are numbered I-XI, and the connecting hydrophilic loops are not shown. Images on the left represent views along an axis parallel to the membrane (inside of the cell on top, outside of the cell on bottom), whereas images on the right are views perpendicular to the membrane from the inside toward the outside. A, six critical residues whose mutation to alanine inhibits adenosine uptake by ⬎90% are indicated as space-filling structures in red. The yellow residue is Met175, discussed in the text, whose mutation to alanine inhibits uptake by 78%. B, residues at the inner ends of TM helices, whose mutation to alanine either inhibits adenosine uptake more modestly or activates uptake, are shown as space-filling structures in green. C, image of LdNT1.1 in the inward-closed and inward-open conformations showing the inner gate transitioning from closed to open state. The inward-closed image and the space-filling structures are as in A, whereas the inward-open image represents the ab initio model for LdNT1.1 that was reported previously (15).

modification of six residues (Y161A and G162A in TM4, V445A in TM10, and G467A, I468A, and I470A in TM11) strongly impaired transport capability (loss of ⬎90% activity versus the wild type). Operationally we refer to these six residues as “critical” for transport activity. Among the remaining mutants investigated, i.e. those that exhibited ⬎10% residual activity, the one with the lowest residual activity (22% activity versus the wild type) was M175A in TM5 (Figs. 1A and 2A), and we consider this mutant as well as being of potential functional significance. The remaining mutants exhibited a more modest reduction or an increase in adenosine uptake (between 30% residual activity for L444A to 229% for F164A) (Figs. 1B and 2A). Moreover, the loss of function in the mutant permeases that exhibited a pronounced loss of uptake was not due to decreased steady-state levels of the mutant transporters on the plasma membrane, because the intensity of the bands in the surface biotinylated mutants is similar or higher to that of wild type transporter, with the exception of I468A (Fig. 2B). NoneVOLUME 289 • NUMBER 13 • MARCH 28, 2014

Inner Gate of an Equilibrative Nucleoside Transporter

A

TM4

TM5

250 125 100 150

Ado Uptake (% of WT)

Ado Uptake (% of WT)

200

100 50 0

WT T160A Y161A G162A M163A F164A F167A

75 50 25 0

TM10

WT

T174A M175A M176A

TM11 125

125

100

Ado Uptake (% of WT)

Ado Uptake (% of WT)

100 75 50 25 0

75 50 25 0

WT

L444A V445A L446A

WT M466A G467A I468A S469A I470A L471A

B WT

Y161A G162A

M175A

WT

V445A G467A I468A

I470A

FIGURE 2. Effect on adenosine uptake of alanine mutations for amino acids predicted to be at the inner ends of helices TM4, TM5, TM10, and TM11, the predicted components of the inner gate. A, uptake of 1 ␮M [3H]adenosine (y axes) was quantified for the ⌬nt1/⌬nt2 double null mutant of L. donovani expressing either wild type (WT, filled bars) LdNT1.1 or this transporter with the designated alanine point mutations (open bars). The data are separated according to each TM helix. The level of uptake for WT was set at 100%, and numbers on the y axes represent percent uptake of each mutant relative to wild type LdNT1.1. Each uptake value in this and subsequent figures represents the mean and S.D. (error bars) for at least three independent uptake measurements. B, surface expression of WT and each mutant was quantified by surface biotinylation followed by purification of biotinylated proteins on streptavidin beads. The streptavidin bound fractions were separated by SDS-PAGE, blotted, and probed with antibody directed against the NT1-loop VII (top) or the myo-inositol transporter LdMIT (bottom). The numbers under each lane represent the relative surface expression of each alanine point mutant relative to that of WT LdNT1.1 after normalization to the LdMIT signal for each lane.

theless, the ⬎90% loss of function of I468A cannot be explained by its under expression in the plasma membrane (40% versus wild type). Furthermore, uptake assays performed at 10 ␮M adenosine (Ado), 10 times the Km value (33), showed very similar results for inhibition of transport (data not shown). This result indicates that mutations Y161A, G162A, M175A, V445A, G467A, I468A, and I470A reduce the Vmax of the permease and do not simply increase the Km for Ado. Overall, mutations that strongly impair transport activity (the six critical residues and Met-175) are predicted to cluster at the interfaces between the four gate helices and to fill the postulated gate region in Fig. 1A, whereas mutations that do not as strongly affect transport capability have a more peripheral location and do not project into the predicted gate region (Fig. 1B), with the possible exception of Leu-444 on TM10. It is notable however that Leu-444 is located farther from the tip of TM10 than Val-445 and thus would be predicted to have less of an effect on gate function, as observed. These results provide experimental support that helices 4, 5, 10, and 11 likely interact MARCH 28, 2014 • VOLUME 289 • NUMBER 13

to form an intracellular gate and that mutation of residues at the interface between the intracellular ends of these helices strongly impairs transport activity, presumably because the function of the inner gate is disrupted. Fig. 1C presents images of the permease in the inward-closed (present homology model) and inward-open (previously published ab initio model (15)) conformations. This image demonstrates that the residues depicted in red and yellow seal off the central cavity of the transporter in the inward-closed conformation but move away from each other and open the central cavity to the inside in the inward-open conformation, supporting the notion that they constitute components of the inner gate. It is notable that the seven residues whose mutation to Ala results in the strongest loss of activity and that are predicted to be positioned at the interfaces of TMs 4, 5, 10, and 11 have hydrophobic side chains (Table 1). This observation suggests that hydrophobic interactions between distinct TMs may mediate, to a large extent, the closing of the intracellular gate. JOURNAL OF BIOLOGICAL CHEMISTRY

8803

Inner Gate of an Equilibrative Nucleoside Transporter TABLE 1 Residues in other ENT family members that correspond to the 7 residues of functional interest of LdNT1.1 (top row in bold) located at the intracellular ends TMs 4, 5, 10, and 11 and discussed under “Sequence Alignment and Conservation of Hydrophobic Residues” The corresponding residues in the other three L. donovani ENTs and the four human ENTs are indicated in subsequent rows. Sequences were aligned using ClustalW Version 1.4. LdNT1.1

Tyr-161

Gly-162

Met-175

Val-445

Gly-467

Ile-468

Ile-470

LdNT2 LdNT3 LdNT4 hENT1 hENT2 hENT3 hENT4

Asn Tyr Tyr Phe Phe Tyr Tyr

Ala Ala Ala Tyr Gly Gly Gly

Ala Ile Met Ile Phe Leu Val

Ile Val Ile Cys Cys Leu Ile

Gly Gly Gly Ala Thr Ser Thr

Ile Val Ala Phe Phe Phe Val

Leu Leu Leu Leu Leu Val Tyr

Sequence Alignment and Conservation of Hydrophobic Residues—To determine whether the hypothesized hydrophobic interactions between residues critical for function of the LdNT1.1 inner gate could also apply for intracellular gate formation in other members of the ENT family, we performed a multialignment encompassing the sequences of the four L. donovani ENTs, LdNT1.1– 4 and the four human ENTs, hENT1– 4. Table 1 lists the residues in these other seven ENTs that correspond to the seven residues of interest in LdNT1.1 (Tyr-161 and Gly-162 in TM4, Met-175 in TM5, Val-445 in TM10, and Gly-467, Ile-468, and Ile-470 in TM11). Indeed with limited exceptions, (e.g. the Asn in LdNT2 that aligns with Tyr161 of LdNT1.1, and the Thr, Ser, and Thr of hENT2, hENT3, and hENT4, respectively, that align with Gly-467 of LdNT1.1), these residues are hydrophobic throughout this family of permeases. This observation suggests strong conservation of hydrophobicity between the proposed critical gate residues of LdNT1.1 and the corresponding residues in other ENTs and thus implies that a “hydrophobic ring” is responsible, at least in part, for closing the outer gate in this family of permeases. Effect of Thiol Cross-linking Agents of Different Lengths and Flexibility on Transport by LdNT1.1 Paired Double Cysteine Mutants—To further test the LdNT1.1 inward-closed structural model, paired double cysteine mutations at the interface between helices 4, 5, 10, and 11 (S158CTM4loop/L465CTM11, G162CTM4/S173CTM5, M176CTM5/M442CTM10, and G447CTM10/G467CTM11) (Fig. 3, A and B) were introduced into both the Cys-less LdNT1.1 construct (a fully functional LdNT1.1 mutant protein in which all five native cysteine residues were substituted with alanine) (30) and the wild type LdNT1.1 background. The later was done because significant transport activity was observed when three of these four pairs were introduced in the wild type construct (Fig. 4A), but introduction of these mutations into the Cys-less background was found to largely abrogate transport activity (Fig. 4B). According to the homology model, paired double cysteine mutants S158CTM4loop/L465CTM11 (1, Fig. 3A) and M176CTM5/M442CTM10 (2, Fig. 3A) were predicted to crosslink one domain of the transporter (N-terminal domain, TM1TM6) to the other (C-terminal domain, TM7-TM11), with calculated C␣ distances between the introduced cysteines of 5.8 and 7.8 Å, respectively. In addition, paired double cysteine mutants G162CTM4/S173CTM5 (3, Fig. 3B) and G447CTM10/ G467CTM11 (4, Fig. 3B) were selected to test the cross-linking of helices within either the N-terminal or the C-terminal domains of the permease, with predicted C␣ distances between the introduced cysteines of 12.4 and 13.5 Å, respectively.

8804 JOURNAL OF BIOLOGICAL CHEMISTRY

FIGURE 3. Locations of introduced dual cysteine mutations in the computational model. For each substituted cysteine, amino acids located in the N-terminal helix bundle (TM1-TM6) are shown as space-filling structures in green, and those in the C-terminal helix bundle (TM7-TM11) are shown in red. The images show the LdNT1.1 inward-closed model seen from the inside toward the outside of the cell. Image A depicts the dual cysteine mutants between TM4 (IV) and TM11 (XI) (S158C/L465C) (labeled 1) and the cysteine pair between TM5 (V) and TM10 (X) (M176C/M442C) (labeled 2). Image B shows the dual cysteine pair between TM4 (IV) and TM5 (V) (G162C/S173C), labeled 3, and that between TM10 (X) and TM11 (XI) (G447C-G467C), labeled 4. The black lines show the interconnections between each cysteine pair.

The paired mutant transporters were then expressed in the ⌬ldnt1⌬ldnt2 L. donovani double null mutant, and the ability of each transgenic mutant permease to mediate ligand translocation was evaluated by uptake assays using 1 ␮M [3H]adenosine (Fig. 4A). G162CTM4-S173CTM5 led to an inactive transporter and was, therefore, excluded from further study. The rest of the paired double cysteine mutants exhibited measurable transport activity (⬃14 –35% of the adenosine accumulated by the wild type transporter), and therefore, the proximity relationships between the selected cysteine residues and by extension between the TM domains to which they belong were further analyzed by adenosine uptake after treatment with the homobifunctional thiol-reactive reagents MTS-3-MTS) MTS17-O5-MTS, o-PDM, and p-PDM. The use of thiol-reactive cross-linkers of different lengths and flexibility has proven very useful in the topographical mapping of proteins and can assist in the determination of the distance between two cysteine residues (38). For many of such studies that have been done previously on permeases (39 – 41), cross-linking is monitored by loss of uptake activity after application of the cross-linking reagent to cells, and we have used the same approach here. The results of pilot in vivo experiments involving time courses and concentration curves indicated that maximum cross-linking of dicysteine mutants preserving cell viability was achieved when parasites were incubated for 5 min in the presence of 0.05 mM MTS-3-MTS, 0.05 mM o-PDM, 0.05 mM p-PDM, and for 10 min in the presence of 0.05 mM MTS-17O5-MTS (data not shown). Cross-linking under such conditions ensures substantial reactivity without excessive cell toxicity, a problem that emerges when higher concentrations or longer times of cross-linking are applied. VOLUME 289 • NUMBER 13 • MARCH 28, 2014

Inner Gate of an Equilibrative Nucleoside Transporter

Ado Uptake (pmol/107 cells/min)

A

15

WT LdNT1.1 template 10

5

M

17

7C /G 46 G 44

6C /M

2C /S 1

44

7C

C 73

5C G 16

58

C/ L4 6

W T S1

B

2C

0

10

(pmol/10 cells/min)

7

Ado Uptake

Cys-less LdNT1.1 template

5

7C 46 7C 44

G

M

17

6C

2C 16 G

/G

/M

17 /S

L4 C/ 58 S1

44

3C

C 65

ss le sCy

2C

0

FIGURE 4. Uptake of [3H]adenosine (Ado Uptake) by each of the cysteine pair mutants (open bars) relative to that of wild type LdNT1.1 (WT, filled bars). A shows uptake of each mutant pair in the WT LdNT1.1 background, whereas B indicates uptake of each mutant in the cysteine-less LdNT1.1 (Cysless) background.

Notably, all three double cysteine mutant transporters completely lost the ability to catalyze active adenosine transport upon cross-linking with the relatively short reagents MTS-3MTS (⬃5 Å in length and flexible; range of S-S distances, ⬃3– 6 Å) (38) or o-PDM (⬃6 Å in length and rigid; range of S-S distances 7.67–10.47 Å) (42, 43) (data not shown). Transport sensitivity to inactivation may be interpreted as evidence for crosslinking the two sulfhydryl groups, presumably because the disulfide bond imposes restrictions on the protein conformational changes necessary for transport (19). Unfortunately, incubation with MTS-3-MTS and o-PDM also strongly inhibited single cysteine mutant transporters S158CTM4loop, M176CTM5, and G467CTM11, raising the possibility that partial cross-linking of these cysteine residues to an endogenous cysteine (Cys-36TM1, Cys-143TM4, Cys-144TM4, Cys260loopTM6 –7, or Cys-377TM7) might have occurred. In contrast, preincubation with the longer length cross-linking reagents p-PDM (⬃10 Å in length and rigid; range of S-S distances, 9.20 –12.29 Å) (42, 43) or MTS-17-O5-MTS (⬃22 Å in length and flexible; range of S-S distances, ⬃3–22 Å) (38) strongly inhibited adenosine influx in ⌬ldnt1⌬ldnt2 L. donovani parasites transfected with the three double cysteine MARCH 28, 2014 • VOLUME 289 • NUMBER 13

mutants, S158CTM4loop/L465CTM11 (p-PDM, ⬃6-fold inhibition; MTS-17-O5-MTS, ⬃5-fold inhibition), M176CTM5/ M442CTM10 (p-PDM, ⬃28-fold inhibition; MTS-17-O5-MTS, ⬃20-fold inhibition), and G447CTM10/G467CTM11 (p-PDM, ⬃8-fold inhibition; MTS-17-O5-MTS, ⬃8-fold inhibition) double mutant constructs (Fig. 5A). Treatment with these cross-linking reagents had almost no effect upon uptake by the wild type LdNT1.1, demonstrating that loss of activity upon cross-linking required introduction of the new cysteine pairs and was not due to modification of the wild type permease. Furthermore, incubation with p-PDM and MTS-17-O5-MTS only weakly inhibited single cysteine mutants S158CTM4 loop (MTS-17-O5-MTS, ⬃1.5-fold inhibition) and G467CTM11 (MTS-17-O5-MTS, ⬃2-fold inhibition) (Fig. 5B). The single cysteine mutant M176CTM5 was more strongly inhibited by the cross-linking treatment (p-PDM, ⬃3.5-fold inhibition; MTS17-O5-MTS, ⬃5-fold inhibition), but these levels of inhibition were still well below those observed (⬃28-fold) when this mutant was incorporated into the M176C/M442C pair. These results demonstrate that the pronounced loss of activity observed upon cross-linking the paired cysteine mutants was due to modification of these pairs and not primarily to reaction of newly introduced cysteines with endogenous cysteines. In addition, glucose transport was not affected under the experimental conditions reported (data not shown), thus confirming that the observed impairment in adenosine uptake is specific to the mutant LdNT1.1 proteins and not due to some unrelated modification that the mutant cell lines may have undergone upon treatment with the cross-linkers or to toxicity of the cross-linking reagents to the cells. Notably, inactivation of transport with MTS-17-O5-MTS was significantly restored after reduction with the sulfhydryl reducing compound DTT (5 mM, 30 min) (Fig. 6), further confirming that loss of transport activity is due to cross-linking of the relevant cysteine pairs. As expected, cross-linking with p-PDM, an irreversible maleimide-based cross-linker, was not reduced (data not shown). Additional evidence for cross-linking of the three double cysteine mutants was obtained by analyzing the p-PDM-treated mutants on SDS-PAGE followed by Western blotting (Fig. 7). The mobilities of the S158C/L465C and M176C/M442C mutants were shifted upward by modification with the crosslinking agent, indicative of covalent ligation of two distant residues (19). The migration of the G447C/G467C mutant was shifted upward only slightly, as expected for cross-linking of two cysteines that are only 20 residues apart. In contrast, treatment of each single cysteine mutant or of wild type LdNT1.1 with p-PDM did not alter the mobility of the protein. These results imply that the relevant cysteine pairs S158CTM4loop/L465CTM11, M176CTM5/M442CTM10, and G447CTM10/G467CTM11 and their associated helices 4 and 11, 5 and 10, and 10 and 11 are indeed in close proximity, as predicted by the homology model and further support the designation of these helices as components of the inner gate.

DISCUSSION Studies on the high resolution structures of membrane transport proteins have proven remarkably revealing regarding the molecular mechanisms whereby they mediate ligand transport JOURNAL OF BIOLOGICAL CHEMISTRY

8805

Inner Gate of an Equilibrative Nucleoside Transporter

A

Control

p-PDM

WT

S158C/L465C 125

Ado Uptake (% of Control)

Ado Uptake (% of Control)

125 100 75 50 25

100 75 50 25 0

0

M176C/M442C

G447C/G467C 125

Ado Uptake (% of Control)

Ado Uptake (% of Control)

125 100

100

75 50 25

75 50 25

0

0

S158C 100

100

100

75 50 25 0

50

0

M442C 200

150

100

50 25 0

G467C 200

100

100

0

75

G447C 250

Ado Uptake (% of Control)

200

Ado Uptake (% of Control)

M176C 125

Ado Uptake (% of Control)

Ado Uptake (% of Control)

Ado Uptake (% of Control)

L465C 150

125

Ado Uptake (% of Control)

B

MTS-17-O5-MTS

50 0

0

FIGURE 5. Effect of cross-linking on uptake of [3H]adenosine by paired or single cysteine mutants of LdNT1.1. All mutants were generated in the wild type LdNT1.1 background. Parasites expressing either wild type LdNT1.1 (WT) or each of cysteine mutants (indicated above each box) were either not treated (control, filled bars) or treated with the cysteine cross-linking reagents p-PDM (open bars) or MTS-17-O5-MTS (horizontally striped bars) as described in the text under Effect of Thiol Cross-linking Agents of Different Lengths and Flexibility on Transport by LdNT1.1 Paired Double Cysteine Mutants. A shows data for the paired cysteine mutants, and B shows the corresponding data for single cysteine mutants.

across membranes (see for example Refs. 22 and 23). In particular, distinct conformational states in the transport cycles have been captured, and structural components of the permeases along with individual amino acids that play central roles in the transport mechanism have been identified and their contributions to ligand permeation defined. However, for membrane proteins for which no crystal structure exists, an alternate approach is to develop computational models of these carriers (44) and to validate or optimize these models using targeted

8806 JOURNAL OF BIOLOGICAL CHEMISTRY

experimental approaches. For LdNT1.1, an arrangement of TM helices similar to that of bacterial MFS members such as LacY and GlpT (16, 17) was predicted from ab initio computational studies (15), and helical components and individual amino acids that constitute the extracellular gate were identified (15, 19). To obtain a model of LdNT1.1 in the previously undefined inward-closed conformation that would capture the closed intracellular gate, another critical component for understanding the molecular mechanism of action, we performed homolVOLUME 289 • NUMBER 13 • MARCH 28, 2014

Inner Gate of an Equilibrative Nucleoside Transporter Control

MTS-17-O5-MTS

MTS-17-O5-MTS/DTT S158C/L465C

WT 125

Ado Uptake (% of Control)

Ado Uptake (% of Control)

125 100 75 50 25 0

100 75 50 25 0

M176C/M442C

G447C/G467C 125

Ado Uptake (% of Control)

Ado Uptake (% of Control)

125 100 75 50 25

100 75 50 25 0

0

FIGURE 6. Reversal of cysteine pair cross-linking by DTT. Parasites expressing either WT LdNT1.1 or each of the three paired cysteine mutants (S158C/L465C, M176C/M442C, G447C/G467C) were first treated with the reversible cysteine cross-linking reagent MTS-17-O5-MTS (open bars) and then with DTT (horizontal bars), and uptake of 1 ␮M [3H]adenosine was measured. Uptake for untreated controls (filled bars) was set at 100%, and uptake from treated samples was graphed relative to that control value.

p-PDM:

_

+

WT p-PDM:

_

+

S158C/L465C p-PDM:

_

+

M176C/M442C p-PDM:

_

+

G447C/G467C

_

+

_

+

S158C

L465C

_

_

+

+

M176C

M442C

_

_

+

G447C

+

G467C

FIGURE 7. Cross-linking of cysteine pair mutants by p-PDM. Membranes were isolated from parasites expressing WT or double cysteine mutant LdNT1.1 cross-linked (⫹) or not (⫺) with p-PDM, separated on SDS-PAGE, blotted onto a membrane, and probed with the NT1-loop VII antibody (left panels). Single cysteine mutants were treated identically (right panels).

MARCH 28, 2014 • VOLUME 289 • NUMBER 13

ogy modeling using the E. coli fucose transporter FucP, the only MFS member crystallized in the inward-closed conformation. Although the sequence identity between LdNT1.1 and FucP is low (13%), homology models of other proteins with similar or lower degrees of identity, such as the HIV protease modeled on distantly related aspartyl proteases (45, 46), have provided valuable structural models to interrogate the function of those proteins. In the case of the HIV protease model the computational structure was subsequently validated by the crystal structure of a retroviral protease (47). Indeed, modeling from remote sequence similarities has been justified by “the general observation that protein structures are better conserved through evolution than are the sequences which overlay them” (46). Furthermore, we have performed site-directed mutagenesis and chemical cross-linking that support this model and define specific residues likely to contribute to the intracellular gate. The homology model of LdNT1.1 (Fig. 1) predicts that the inner tips of TM helices 4, 5, 10, and 11 come into close apposition to close the pore and thus constitute the inner gate, as has been postulated previously for the MFS family member LacY (16). Mutagenesis of seven hydrophobic residues (TM4: Tyr161, Gly-162; TM5: Met-175; TM10: Val-445; TM11: Gly-467, Ile-468, Ile-170) at the tips of these helices strongly impairs transport function, consistent with the prediction of the model that the tips of these helices interact to close the pore. Furthermore, within each helix mutation of residues on one face greatly reduces activity, whereas mutation of residues adjacent to these positions has considerably less inhibitory effect (Fig. 2). Moreover, amino acids whose mutation strongly impairs activity project toward the interface between the N-terminal (TM1– 6) and C-terminal (TM7–11) helix bundles in the model, where the helices are predicted to close off the pore (Fig. 1A). These JOURNAL OF BIOLOGICAL CHEMISTRY

8807

Inner Gate of an Equilibrative Nucleoside Transporter experimental results support the model and suggest that these residues cluster together, likely at least in part by hydrophobic interactions, to mediate closing of the inner gate helices. Indeed these residues can be thought of as a hydrophobic ring that provides favorable interhelix interactions that allow the inner gate to close. Although the activity disrupting mutation introduced in these studies, alanine, is also a relatively hydrophobic amino acid, the alteration in side chain size and physical properties must be sufficient to impair the necessary helix interactions. Indeed, several more conservative mutations in these critical residues still reduced substantially transport function (e.g. Y161F, 13.4%; Y161W, 6.1%; and V445I, 24.3% residual activity, respectively, data not shown graphically), although not as markedly as the alanine substitutions (Y161A, 1.4% and V445A, 7.4% residual activity, respectively, Fig. 2A), suggesting that the interhelix interactions are sensitive to the structural intercalation of residues with the interacting interface on other helices. In contrast, those amino acids whose mutation affects transport less strongly are predicted to be oriented away from the interface between the N- and C-terminal bundles (Fig. 1B). The one exception among the latter group of residues is Leu444 in TM10 that does project into the predicted interface, albeit at a more distal position from the helix terminus, but mutation of this reside to alanine still inhibits uptake relatively strongly, by ⬃70%. Clearly, any cutoff for distinguishing between “strong” versus “modest” inhibition is somewhat arbitrary. However, the overall conclusion from these studies is that mutagenesis of residues located at the predicted interface between helix bundles inhibits transport function more strongly than mutagenesis of those predicted to be at peripheral positions with regard to the helix bundle interface. We emphasize that the mutations introduced in this study were chosen explicitly because they were predicted to be located at the interface between helices that close off the central aqueous cavity in the LdNT1.1 model (Fig. 1A). Indeed the reductions in transport activity after mutagenesis closely accord with the predictions of the model. For this reason we postulate that the residues and helix termini identified are molecular components of the inner gate per se and not simply residues required for other reasons for translocation of substrate or gate opening-closing. The apposition of these residues in the inward-closed but not in the outward-closed model (Fig. 1C) underscores the role of these residues as components of the inner gate. The cysteine cross-linking studies also support the identification of the intracellular gate, as residues within predicted gate helices are close enough to each other in the inward-closed conformation (Fig. 3) to be cross-linked to each other. Because p-PDM cross-links all three cysteine-cysteine pairs, they must be within 9.20 –12.29 Å of each other, not far from the predictions of the model. The shorter predicted distances between introduced Cys-158 –Cys-465 (5.8 Å) and Cys-176 –Cys-442 (7.8 Å) could either indicate that these residues are somewhat farther apart than predicted by the model or that they are able to react with p-PDM as they breathe apart during opening of the gate. It is noteworthy that the inward-closed and outward-closed conformations are probably not the only two structural states

8808 JOURNAL OF BIOLOGICAL CHEMISTRY

important for function of LdNT1.1. Crystallographic studies have identified “occluded states” of multiple MFS members (24, 48 –50), intermediate structures in which the substrate is not directly accessible to solute on either the extracellular or intracellular face of the membrane. Such states are likely to exist for LdNT1.1 as well. The conservation of the proposed hydrophobic ring among all four ENTs in L. donovani and among all four human ENTs, which are highly divergent in sequence from the parasite orthologs (e.g. sequence identity between LdNT1.1 and hENT1 is 33%), argues for a common mechanism for inner gate formation between these diverse members of the SLC29 family. This observation underscores that modeling and confirmatory biochemical studies on one member of the family, the parasite encoded LdNT1.1, can illuminate structure-function relationships that are broadly applicable to many family members including those expressed in evolutionarily distant humans. In summary, the results reported here (i) provide a computational model for LdNT1.1 in the inward-closed conformation, (ii) identify four helices proposed to form the inner gate that closes the permeation pore in that conformation, (iii) implicate by mutagenesis seven residues at the inner ends of those helices that are likely to interact by largely hydrophobic forces to close the inner gate, (iv) demonstrate that the predicted interface between N- and C-terminal helix bundles is critical for transport function, thus supporting the accuracy of the computational model, (v) test by cysteine cross-linking the proximity of three pairs of helices predicted by this model to be in close apposition, and (vi) demonstrate, for the seven functionally important residues in the inner gate, relative conservation of hydrophobicity among ENT family members, suggesting that this gate and its mechanism of closing is likely broadly conserved in the SLC29 transporters. REFERENCES 1. Zhang, J., Visser, F., King, K. M., Baldwin, S. A., Young, J. D., and Cass, C. E. (2007) The role of nucleoside transporters in cancer chemotherapy with nucleoside drugs. Cancer Metastasis Rev. 26, 85–110 2. Young, J. D., Yao, S. Y., Sun, L., Cass, C. E., and Baldwin, S. A. (2008) Human equilibrative nucleoside transporter (ENT) family of nucleoside and nucleobase transporter proteins. Xenobiotica 38, 995–1021 3. Kong, W., Engel, K., and Wang, J. (2004) Mammalian nucleoside transporters. Curr. Drug Metab. 5, 63– 84 4. Baldwin, S. A., Beal, P. R., Yao, S. Y., King, A. E., Cass, C. E., and Young, J. D. (2004) The equilibrative nucleoside transporter family, SLC29. Pflugers Arch. 447, 735–743 5. King, A. E., Ackley, M. A., Cass, C. E., Young, J. D., and Baldwin, S. A. (2006) Nucleoside transporters. From scavengers to novel therapeutic targets. Trends Pharmacol. Sci. 27, 416 – 425 6. de Koning, H. P., Watson, C. J., and Jarvis, S. M. (1998) Characterization of a nucleoside/proton symporter in procyclic Trypanosoma brucei brucei. J. Biol. Chem. 273, 9486 –9494 7. Stein, A., Vaseduvan, G., Carter, N. S., Ullman, B., Landfear, S. M., and Kavanaugh, M. P. (2003) Equilibrative nucleoside transporter family members from Leishmania donovani are electrogenic proton symporters. J. Biol. Chem. 278, 35127–35134 8. Ortiz, D., Sanchez, M. A., Koch, H. P., Larsson, H. P., and Landfear, S. M. (2009) An acid-activated nucleobase transporter from Leishmania major. J. Biol. Chem. 284, 16164 –16169 9. Hammond, D. J., and Gutteridge, W. E. (1984) Purine and pyrimidine metabolism in the Trypanosomatidae. Mol. Biochem. Parasitol. 13, 243–261

VOLUME 289 • NUMBER 13 • MARCH 28, 2014

Inner Gate of an Equilibrative Nucleoside Transporter 10. Young, J. D., Yao, S. Y., Cass, C. E., and Baldwin, S. A. (2003) Equilibrative nucleoside transport proteins in Red Cell Membrane Transport in Health and Disease (Bernhardt, I., and Ellory, J. C., eds.) pp. 321–337, Springer, Berlin 11. Parkinson, F. E., Damaraju, V. L., Graham, K., Yao, S. Y., Baldwin, S. A., Cass, C. E., and Young, J. D. (2011) Molecular biology of nucleoside transporters and their distributions and functions in the brain. Curr. Top. Med. Chem. 11, 948 –972 12. Baldwin, S. A., McConkey, G. A., Cass, C. E., and Young, J. D. (2007) Nucleoside transport as a potential target for chemotherapy in malaria. Curr. Pharm. Des. 13, 569 –580 13. Papageorgiou, I., De Koning, H. P., Soteriadou, K., and Diallinas, G. (2008) Kinetic and mutational analysis of the Trypanosoma brucei NBT1 nucleobase transporter expressed in Saccharomyces cerevisiae reveals structural similarities between ENT and MFS transporters. Int. J. Parasitol 38, 641– 653 14. Arastu-Kapur, S., Arendt, C. S., Purnat, T., Carter, N. S., and Ullman, B. (2005) Second-site suppression of a nonfunctional mutation within the Leishmania donovani inosine-guanosine transporter. J. Biol. Chem. 280, 2213–2219 15. Valdés, R., Arastu-Kapur, S., Landfear, S. M., and Shinde, U. (2009) An ab initio structural model of a nucleoside permease predicts functionally important residues. J. Biol. Chem. 284, 19067–19076 16. Abramson, J., Smirnova, I., Kasho, V., Verner, G., Kaback, H. R., and Iwata, S. (2003) Structure and mechanism of the lactose permease of Escherichia coli. Science 301, 610 – 615 17. Huang, Y., Lemieux, M. J., Song, J., Auer, M., and Wang, D. N. (2003) Structure and mechanism of the glycerol-3-phosphate transporter from Escherichia coli. Science 301, 616 – 620 18. Valdés, R., Liu, W., Ullman, B., and Landfear, S. M. (2006) Comprehensive examination of charged intramembrane residues in a nucleoside transporter. J. Biol. Chem. 281, 22647–22655 19. Valdés, R., Shinde, U., and Landfear, S. (2012) Cysteine cross-linking defines the extracellular gate for the Leishmania donovani nucleoside transporter 1.1 (LdNT1.1). J. Biol. Chem. 287, 44036 – 44045 20. Jardetzky, O. (1966) Simple allosteric model for membrane pumps. Nature 211, 969 –970 21. Kavanaugh, M. P. (1998) Neurotransmitter transport. Models in flux. Proc. Natl. Acad. Sci. U.S.A. 95, 12737–12738 22. Krishnamurthy, H., Piscitelli, C. L., and Gouaux, E. (2009) Unlocking the molecular secrets of sodium-coupled transporters. Nature 459, 347–355 23. Abramson, J., and Wright, E. M. (2009) Structure and function of Na⫹symporters with inverted repeats. Curr. Opin. Struct. Biol. 19, 425– 432 24. Yin, Y., He, X., Szewczyk, P., Nguyen, T., and Chang, G. (2006) Structure of the multidrug transporter EmrD from Escherichia coli. Science 312, 741–744 25. Dang, S., Sun, L., Huang, Y., Lu, F., Liu, Y., Gong, H., Wang, J., and Yan, N. (2010) Structure of a fucose transporter in an outward-open conformation. Nature 467, 734 –738 26. Söding, J., Biegert, A., and Lupas, A. N. (2005) The HHpred interactive server for protein homology detection and structure prediction. Nucleic Acids Res. 33, W244 –W248 27. Remmert, M., Biegert, A., Hauser, A., and Söding, J. (2012) HHblits: lightning-fast iterative protein sequence searching by HMM-HMM alignment. Nat. Methods 9, 173–175 28. Sali, A., and Blundell, T. L. (1993) Comparative protein modelling by satisfaction of spatial restraints. J. Mol. Biol. 234, 779 – 815 29. Liu, W., Boitz, J. M., Galazka, J., Arendt, C. S., Carter, N. S., and Ullman, B. (2006) Functional characterization of nucleoside transporter gene replacements in Leishmania donovani. Mol. Biochem. Parasitol. 150, 300 –307 30. Valdés, R., Vasudevan, G., Conklin, D., and Landfear, S. M. (2004) Transmembrane domain 5 of the LdNT1.1 nucleoside transporter is an amphipathic helix that forms part of the nucleoside translocation pathway. Biochemistry 43, 6793– 6802 31. LeBowitz, J. H., Coburn, C. M., McMahon-Pratt, D., and Beverley, S. M.

MARCH 28, 2014 • VOLUME 289 • NUMBER 13

32. 33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

44.

45. 46. 47.

48.

49.

50.

(1990) Development of a stable Leishmania expression vector and application to the study of parasite surface antigen genes. Proc. Natl. Acad. Sci. U.S.A. 87, 9736 –9740 LeBowitz, J. H. (1994) Transfection experiments with Leishmania. Methods Cell Biol. 45, 65–78 Vasudevan, G., Carter, N. S., Drew, M. E., Beverley, S. M., Sanchez, M. A., Seyfang, A., Ullman, B., and Landfear, S. M. (1998) Cloning of Leishmania nucleoside transporter genes by rescue of a transport-deficient mutant. Proc. Natl. Acad. Sci. U.S.A. 95, 9873–9878 Ortiz, D., Valdés, R., Sanchez, M. A., Hayenga, J., Elya, C., Detke, S., and Landfear, S. M. (2010) Purine restriction induces pronounced translational upregulation of the NT1 adenosine/pyrimidine nucleoside transporter in Leishmania major. Mol. Microbiol. 78, 108 –118 Detke, S. (2007) TOR-induced resistance to toxic adenosine analogs in Leishmania brought about by the internalization and degradation of the adenosine permease. Exp. Cell Res. 313, 1963–1978 Drew, M. E., Langford, C. K., Klamo, E. M., Russell, D. G., Kavanaugh, M. P., and Landfear, S. M. (1995) Functional expression of a myo-inositol/H⫹ symporter from Leishmania donovani. Mol. Cell Biol. 15, 5508 –5515 Thompson, J. D., Higgins, D. G., and Gibson, T. J. (1994) ClustalW. Improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res. 22, 4673– 4680 Zhou, Y., Guan, L., Freites, J. A., and Kaback, H. R. (2008) Opening and closing of the periplasmic gate in lactose permease. Proc. Natl. Acad. Sci. U.S.A. 105, 3774 –3778 Brocke, L., Bendahan, A., Grunewald, M., and Kanner, B. I. (2002) Proximity of two oppositely oriented reentrant loops in the glutamate transporter GLT-1 identified by paired cysteine mutagenesis. J. Biol. Chem. 277, 3985–3992 Crisman, T. J., Qu, S., Kanner, B. I., and Forrest, L. R. (2009) Inward-facing conformation of glutamate transporters as revealed by their inverted-topology structural repeats. Proc. Natl. Acad. Sci. U.S.A. 106, 20752–20757 Zhang, X., and Qu, S. (2011) Proximity of transmembrane segments 5 and 8 of the glutamate transporter GLT-1 inferred from paired cysteine mutagenesis. PLoS ONE 6, e21288 Wu, J., and Kaback, H. R. (1997) Helix proximity and ligand-induced conformational changes in the lactose permease of Escherichia coli determined by site-directed chemical cross-linking. J. Mol. Biol. 270, 285–293 Green, N. S., Reisler, E., and Houk, K. N. (2001) Quantitative evaluation of the lengths of homobifunctional protein cross-linking reagents used as molecular rulers. Protein Sci. 10, 1293–1304 Etchebest, C., and Debret, G. (2010) Critical review of general guidelines for membrane proteins model building and analysis. Methods Mol. Biol. 654, 363–385 Pearl, L. H., and Taylor, W. R. (1987) A structural model for the retroviral proteases. Nature 329, 351–354 Taylor, W. R. (1994) Protein structure modelling from remote sequence similarity. J. Biotechnol. 35, 281–291 Miller, M., Jaskólski, M., Rao, J. K., Leis, J., and Wlodawer, A. (1989) Crystal structure of a retroviral protease proves relationship to aspartic protease family. Nature 337, 576 –579 Hirai, T., Heymann, J. A., Shi, D., Sarker, R., Maloney, P. C., and Subramaniam, S. (2002) Three-dimensional structure of a bacterial oxalate transporter. Nat. Struct. Biol. 9, 597– 600 Newstead, S., Drew, D., Cameron, A. D., Postis, V. L., Xia, X., Fowler, P. W., Ingram, J. C., Carpenter, E. P., Sansom, M. S., McPherson, M. J., Baldwin, S. A., and Iwata, S. (2011) Crystal structure of a prokaryotic homologue of the mammalian oligopeptide-proton symporters, PepT1 and PepT2. EMBO J. 30, 417– 426 Pedersen, B. P., Kumar, H., Waight, A. B., Risenmay, A. J., Roe-Zurz, Z., Chau, B. H., Schlessinger, A., Bonomi, M., Harries, W., Sali, A., Johri, A. K., and Stroud, R. M. (2013) Crystal structure of a eukaryotic phosphate transporter. Nature 496, 533–536

JOURNAL OF BIOLOGICAL CHEMISTRY

8809