Structural basis for the reaction cycle and transport mechanism of human Na⁺-sulfate cotransporter NaS1 (SLC13A1)

Abstract

Sulfate (SO₄²⁻) is a pivotal inorganic anion with essential roles in mammalian physiology. NaS1, a member of solute carrier 13 family and divalent anion/sodium symporter family, functions as a Na⁺-sulfate cotransporter, facilitating sulfate (re)absorption across renal proximal tubule and small intestine epithelia. While previous studies have linked several human disorders to mutations in the NaS1 gene, its transport mechanism remains unclear. Here, we report the cryo–electron microscopy structures of five distinct conformations of the human NaS1 at resolutions of 2.7 to 3.3 angstroms, revealing the substrates recognition mechanism and the conformational change of NaS1 during the Na⁺-sulfate cotransport cycle. Our studies delineate the molecular basis of the detailed dynamic transport cycle of NaS1. These findings advance the current understanding of the Na⁺-sulfate cotransport mechanism, human sulfate (re)absorption, and the implications of disease-associated NaS1 mutations.

Five distinct conformations of human NaS1 provide insight into the mechanism of Na⁺-sulfate cotransport.

INTRODUCTION

As the fourth most abundant anion in human plasma (1, 2), sulfate plays essential roles in numerous physiological processes, particularly in human pregnancy, growth, and development (3–5). Sulfate deficiency inhibits detoxification, increases susceptibility to xenobiotics, and changes metabolism and activities of many endogenous compounds, revealing the critical functions of sulfate homeostasis in mammalian physiology (6–8). Most of the sulfate from dietary intake is absorbed in the gastrointestinal tract, and once in the bloodstream, sulfate is freely filtered by the kidneys and ~90% of sulfate in crude urine is reabsorbed in renal proximal tubule, thus maintaining the sulfate homeostasis (9–11). As a hydrophilic divalent anion, sulfate requires specific membrane sulfate carriers to mediate its influx/efflux cross cell membrane (7). Over the past half-century, a series of studies in various animals has identified two distinct sulfate transport systems in renal and intestinal epithelial cells: the Na⁺-sulfate cotransporter NaS1 (SLC13A1) on the luminal (apical) brush border membrane (BBM) mediates the lumen-to-cell sulfate transport (12–18), whereas the sulfate-anion transport (SAT)1 (SLC26A1) on the contra luminal basolateral membrane (BLM) mediates the cell-to-blood sulfate transport (19–28). These two transporters are responsible for the (re)absorption of sulfate to maintain the blood sulfate homeostasis, and their dysfunctions lead to pathophysiological conditions (6, 29, 30).

NaS1 belongs to the divalent anion/sodium symporter (DASS) family, which typically uses the preexisting Na⁺ gradient to transport Krebs cycle intermediates or sulfate across the cell membrane (31–33). Specifically, mammalian DASS proteins [from the solute carrier 13 (SLC13) family], such as NaS1 and NaS2, cotransport Na⁺ ions and sulfate, whereas NaCT, NaDC1, and NaDC3 facilitate the symport of Na⁺ ions along with either C6-tricarboxylate or C4-dicarboxylate (16, 34–37). NaS1 are mainly expressed in the kidney and intestine and functions as an electrogenic, pH-insensitive, and high-affinity Na⁺-sulfate cotransporter (16–18, 38–40). NaS1 has also been reported to transport sulfate analogs such as thiosulfate and selenate, and its transport activity can be inhibited by some sulfate analogs and carboxylic acid (17, 39, 40). NaS1 expression is regulated by various dietary and metabolic conditions, including the high/low sulfate diet, thyroid hormone and growth hormone level, vitamin D level, blood potassium level, and kidney function (41–47). Previous studies in murine model showed that knockout of NaS1 gene results in hyposulfatemia and hypersulfaturia, as well as variations in growth, fecundity, behavior, metabolism, intestinal physiology, and detoxification functions of the liver (6, 48–56). Further, pathological studies have reported that several human disorders such as hypersulfaturia, hyposulfatemia, dorsalgia, chronic pain, and intervertebral disc disorder are implicated in mutations of NAS1 gene (57–63), indicating the critical roles of NaS1 in human sulfate homeostasis.

Starting a decade ago, structural studies of the DASS family have been mainly focused on the NaDC family proteins, including the prokaryotic VcINDY, LaINDY, and the human NaCT (64–68), leading to an elevator transport model in which the two protomers of the protein transit from outward-open to inward-open conformations during transport. However, for another important class of DASS family proteins, the Na⁺-sulfate cotransporters, their molecular mechanism remain enigmatic. In this study, using single-particle cryo–electron microscopy (cryo-EM), we present the structures of full-length human NaS1 in five distinct conformations including three intermediate conformations that have not been reported in DASS family proteins. Connecting these structures provides structural visualization of the transport cycle for NaS1 from outward-open state to inward-open state in dynamic. Our findings not only provide insights into the substrates recognition and the transport mechanism of NaS1 and DASS family but also shed light on human sulfate (ab)sorption process.

RESULTS

Structure determination of human NaS1

To obtain homogeneous proteins suitable for cryo-EM study, we overexpressed full-length human NaS1 in mammalian cells using a twin-strep affinity tag and subsequently reconstituted it in Lauryl Maltose Neopentyl Glycol/Cholesteryl Hemisuccinate (LMNG/CHS) micelle buffer. By adding different substrates during protein purification, we successfully obtained the NaS1/Na⁺/SO₄²⁻ complex (fig. S1). The NaS1/Na⁺/SO₄²⁻ complex samples were then subjected to cryo-EM analysis, resulting in a total of five distinct maps at resolutions of 2.7 to 3.3 Å (figs. S2 and S3 and table S1). All the NaS1 maps displayed well-resolved densities to allow confident model building for most of the protein regions (figs. S4 to S7). A total of five different conformations were identified in the NaS1/Na⁺/SO₄²⁻ complex from one dataset, including an inward-open conformation, an outward-open conformation, and three intermediate conformations (named intermediate states 1 to 3) that have not been previously reported in DASS family proteins (Fig. 1, A to E). Moreover, the number of particles for different conformations of NaS1 is not identical in our study. The three intermediate structures have larger proportion of particles compared to the inward and outward-open structures, indicating that most of the NaS1 proteins in our cryo-EM sample are in the intermediate conformations (fig. S2C). Intriguingly, in the two asymmetric states, intermediate states 1 and 2, the two transport domains of the NaS1 dimer are located at different positions, which is different from the classic “elevator” model wherein both transport domains always exhibit synchronous motion. (Fig. 1, B and C).

Fig. 1. Structure determination and overall structure of human NaS1/Na⁺/SO₄²⁻ complex.

(A to E) Cryo-EM map, ribbon model, and schematic model of human NaS1/Na⁺/SO₄²⁻ complex in the outward-open state [(A), green], intermediate state 1 [(B), blue], intermediate state 2 [(C), purple], intermediate state 3 [(D), salmon], and inward-open state [(E), orange]. The bound Na⁺ and SO₄²⁻ are shown as purple spheres and sticks, respectively. (F) Topology diagram of one protomer in human NaS1 dimer in the inward-open state. TM helices of NaS1 are schematically illustrated. The gray shaded regions indicate the inverted repeats of NaS1.

Overall structure of human NaS1

To better characterize the overall structure and the specific domains of NaS1, we will reference the inward-open NaS1/Na⁺/SO₄²⁻ complex structure because of the ease in structural comparison with other DASS family proteins. Human NaS1 forms a homodimer, with each NaS1 protomer containing a scaffold domain (residues 1 to 101, 346 to 451) and a transport domain (residues 120 to 306, 470 to 577) (Fig. 1, F and G). Specifically, the scaffold domain comprises transmembrane helix 1 (TM1) to TM4 and TM7 to TM9, whereas the transport domain is composed of TM5, TM6, TM10, and TM11 as well as the helical hairpin-like structures named HP_in and HP_out (Fig. 1F). The N-terminal (from TM2 to TM6) and the C-terminal halves (from TM6 to TM11) exhibit symmetrical repeating structures (Fig. 1F). Notably, within the transport domain, TM5 and TM10 exhibit discontinuous transmembrane span, with their segments connected by loops L5ab and L10ab, respectively.

Substantial interactions, both within and between protomers, contribute to the stabilization of the scaffold domain throughout the NaS1 transport cycle. These interprotomer interactions provide a robust interface for the scaffold domains (Fig. 2A), reinforced by interactions such as R12-Y418′ (Fig. 2B), Q428-M431′ (Fig. 2C), and π-π interactions between W433 and W433′ (Fig. 2D). Intraprotomer stability arises from interactions like R107 in TM4c with Y337 and G341 in H6b (Fig. 2E), anchoring the horizontal orientation of TM4c. Both horizontal TMs, H4c and H9c, serve as connectors between the scaffold and transport domains.

Fig. 2. Interactions within NaS1/Na⁺/SO₄²⁻ complex.

(A) Cylinder representation of the dimer interface of NaS1/Na⁺/SO₄²⁻ complex. The transport domains are colored orange, and the two scaffold domains from each protomer are colored pink and cyan, respectively. TMs of scaffold domain in one protomer are labeled. (B to D) Interactions in the NaS1 dimer interface. Potential hydrogen bonds are indicated with yellow dotted lines. (E) Interactions between H4c and H6b. Potential hydrogen bonds are indicated with yellow dotted lines. (F) The proline residues in the transport domain provide rigidity. The side chains of all proline residues in the transport domain are shown. (G and H) Interactions within the NaS1 transport domain. (I) Interactions between NaS1 scaffold domain and transport domain. Potential hydrogen bonds are indicated with yellow dotted lines.

The transport domain of NaS1 is also stabilized by extensive interactions, especially by many π-π interactions. These interactions provide rigidity to the transport domain, allowing for rigid body-like motions during substrate translocation. Similar to the previous study in VcINDY (68), the rigidity of transport domain of NaS1 is also largely attributed to the proline residues (Fig. 2F) and further augmented by a series of interactions between side chains from the aromatic amino acids. Specifically, residues F281 and F285 from TM6 interact with W360 and F361 of TM7 (Fig. 2G). Likewise, residues F525 and F527 from L10ab interact with Y246 of TM5A (Fig. 2H). These aromatic clusters are essential for the stabilization of the hairpin-like structures (HP_in and HP_out) and the loop structures (L5ab and L10ab), which functions importantly in substrate binding. In contrast to the extensive interaction networks within the scaffold domain and transport domain, their interface mainly comprises a small number of hydrophobic interactions and only one hydrogen bonding interaction, N280-R364 (Fig. 2I). This arrangement creates a relatively smooth interface, hypothesized to lower the energetic barrier, thereby facilitating relative movement during the substrate transport cycle.

Sequences alignment of human NaS1 with other DASS family members reveals that NaS1 shares its highest similarity to NAS2 with 49% sequence identity. It also shows higher sequence identity with other SLC13 family proteins (NaDC1, 46%; NaDC3, 41%; NaCT, 44%) but notably less with two prokaryotic homologs (fig. S8). Our current human NaS1 exhibit an inward-open conformation with a pore in the middle of TMD open to the cytosol, similar to the previously reported prokaryotic VcINDY and human NaCT, with a root mean square deviation (RMSD) of 3.39 Å over 758 Cα atoms [Protein Data Bank (PDB): 7T9F; VcINDY] and 2.00 Å over 781 Cα atoms (PDB: 7JSK; NaCT) (fig. S9).

Substrate binding sites in human NaS1

To gain insights into the mechanism of substrate transport in NaS1, we embark on the investigation of ion binding sites in NaS1/Na⁺/SO₄²⁻ complex. The structure of NaS1/Na⁺/SO₄²⁻ complex in intermediate state 2 was first analyzed because of its highest resolution. Three prominent densities corresponding to two Na⁺ and one SO₄²⁻ were identified in the transport domain of each NaS1 protomer (Fig. 3, A to D). Structural comparison of the two transport domains indicates that despite some conformational differences, the binding sites for Na⁺ and SO₄²⁻ are almost identical (Fig. 3B). Consistent with the nomenclature of the Na⁺ binding sites in NaDC proteins, the two Na⁺ binding sites in NaS1 are also named as Na1 and Na2. Specifically, Na1 locates in a shell-like structure composed of the loop structure L5ab connecting TM5a and TM5b and the tip of the hairpin structure HP_in. Similarly, Na2 resides in a shell-like structure consisting of L10ab and HP_out (Fig. 3B). Further analysis of these two Na⁺ binding sites reveals that Na1 is coordinated by S135, L138, and N140 from HP_in and G258 from L5ab (Fig. 3, C and D), whereas, at the near-symmetric position, Na2 is stabilized by T485, A488, and N490 from HP_out and A532 from L10ab (Fig. 3, C and D).

Fig. 3. Substrate binding sites in human NaS1.

(A) Cylinder representation of the substrate binding sites of NaS1 in intermediate state 2. The dimer scaffold domain is colored white, and the two transport domains from each protomer are colored light green and light orange, respectively. The bound Na⁺ and SO₄²⁻ ions are shown as purple sphere and sticks, respectively. (B) Structure comparison of substrate binding sites between the two transport domains in (A). The TMs, hairpin-like structures (HP_in and HP_out) and the loops are labeled. The bound Na⁺ and SO₄²⁻ ions are shown as purple sphere and sticks, respectively. (C and D) Close-up view of the substrates binding sites in each of the two NaS1 protomers. The density of Na⁺, SO₄²⁻ and residues from the 2.7-Å map are shown as blue meshes and contoured at 7σ. (E) Transport activity measurement of human NaS1 (WT or mutants) in human embryonic kidney 293T cells. Transport activities of NaS1 variants are normalized to those of WT protein in the 10-min time course. The data represent mean ± SD of three biological independent experiments. Statistical significance was tested using the unpaired t test. Asterisks represent significance versus WT, for 95% probability (P < 0.05). (F) Structural comparison of substrate binding sites among human NaS1 (orange), VcINDY (gray, PDB accession code: 5UL7), and human NaCT (pink, PDB accession code: 7JSK). The Na⁺ ions are shown as spheres, and the sulfate, succinate, and citrate are shown as sticks.

The SO₄²⁻ locates at the tip of the hairpin structure HP_in, leaning more toward the Na1, and is stabilized by hydrogen bonding and electrostatic interactions between the residues S139, N140, and T141 in the conserved serine-asparagine-threonine (SNT) module and S260 in L5ab (Fig. 3, C and D). To further validate the substrate binding sites identified in our study, we performed whole-cell [³⁵S]-sulfate uptake assay by separately mutating the corresponding residues. First, the immunoblotting results showed that the NaS1 mutations do not notably affect the protein expression levels (fig. S10A). Second, in the time-dependent transport assay for wild-type (WT) NaS1, we observed the typical Michaelis-Menten saturation for SO₄²⁻ uptake, demonstrating both the protein’s activity and the effectiveness of this assay (fig. S10B). Last, we measured the transport activity between the WT NaS1 and the substrate binding–related mutants. The results showed that mutations in both the sodium-binding and sulfate-binding residues abolished the sulfate transport activity of NaS1 (Fig. 3E), signifying the essential roles of these residues in substrate transport cycle. By further analyzing the substrate binding sites in the other four conformations of the NaS1/Na⁺/SO₄²⁻ complex, we found that although the densities corresponding to the SO₄²⁻ could be observed in all conformations (fig. S7), the two Na⁺ densities could not be clearly determined in some conformations (fig. S11). This may be largely due to the density maps of these conformations containing fewer number of particles; hence, the densities of some Na⁺ densities are not well represented.

Substrate-recognition mechanism of human NaS1

To investigate the substrate-recognition mechanism of NaS1, we compared the structure of NaS1 with the currently available NaDC structures: the prokaryotic VcINDY and the human NaCT (Fig. 3F). For both VcINDY and NaCT, their substrate binding sites are situated midway between the two Na⁺ binding pockets Na1 and Na2. Each SNT motif contributes to stabilize one carboxylate group and one Na⁺ (Fig. 3F). However, unlike the NaDC family proteins, the SO₄²⁻ binding site in NaS1 is close to Na1 rather than the position midway between Na1 and Na2 (Fig. 3F). This might be attributed to the fact that the size of the sulfate group is smaller, unlike the binary or tricarboxylic acids that have multiple acidic groups for stabilization. Moreover, although spatially it seems that the Na2 could also accommodate another SO₄²⁻, we did not observe a second bound SO₄²⁻ in any of our obtained structures.

In both VcINDY and NaCT, each SNT motif can stabilize one Na⁺ and one carboxylate group. However, in NaS1, the second SNT motif is only involved in stabilizing Na⁺. Previously studies have suggested that the bound Na⁺ could enhance binding to the carboxylate group through electrostatic interactions in NaDC proteins (64, 65, 67). Therefore, we initially supposed that because the second binding pocket was further from the sulfate, it might be redundant for sulfate binding. Unexpectedly, our sulfate uptake assay revealed that both conserved SNT motifs are vital for the sulfate transport activity of NaS1. Mutations of any residues in the two SNT motifs resulted in a significant decrease in its transport activity (Fig. 3E), indicating that both Na⁺ are essential for sulfate transport in NaS1, albeit sulfate is farther away from Na2 than from Na1.

Sequence alignment between NaS1 and other DASS proteins show that the first SNT motif remains highly conserved from prokaryotes to humans. While the last amino acid in the second SNT motif has some variations, it remains pivotal for substrate transport. Furthermore, unlike VcINDY and NaCT, we found that besides the SNT motif, S260 on L5ab is also crucial for sulfate binding, and its mutation also results in a loss of NaS1 transport activity (Fig. 3E). Sequence alignment reveals that this site is nonconserved, identical only in NaS1 and NAS2, and different in VcINDY and NaCT. The involvement of residues of L5ab in substrate binding is another unique feature of NaS1 in its substrate-recognition mechanism.

Conformational changes during NaS1 transport cycle

Previous studies on the DASS family proteins have not identified different conformations within the same protein, e.g., the structures of prokaryotic VcINDY and human NaCT in the inward-facing conformation and the prokaryotic protein LaINDY in the outward-facing conformation. The currently resolved structures of NaS1 allow us to examine the reaction cycle of NaS1 transporter. In summary, we ultimately obtained five different conformations of NaS1/Na⁺/SO₄²⁻ complex (Fig. 4A), wherein the inward-facing conformation (Na1 and SO₄²⁻ ions identified) resembles those of VcINDY and NaCT (fig. S9), and the outward-facing conformation (SO₄²⁻ ions identified) is similar to LaINDY (fig. S9A). Within the three intermediate states, we have identified two protomer conformations (partial outward and partial inward protomers), which prompted the intermediate state conformations to exhibit as a chimeric structure that have never been reported previously. Specifically, intermediate state 1 consists of one outward protomer (Na1, Na2, and SO₄²⁻ ions identified) and one partial-outward protomer (Na2 and SO₄²⁻ ions identified). Intriguingly, in intermediate state 2, one protomer exhibits an outward-facing state that is similar to that in the outward conformation, whereas the other protomer displays a partial-inward state (Fig. 4B). This specific asymmetric conformation has two binding pockets with one facing outwardly and the other one partially accessible from intracellular (Fig. 4B). Intermediate state 3 exhibited a dimeric structure that includes two partial-inward protomers with Na2 and SO₄²⁻ ions identified. It seems that during the sulfate transport process of NaS1, the scaffold domain remains stationary. The alternate motion of the two transport domains completes the transition from the outward-facing conformation to the intermediate conformation and then to the inward-facing conformation, with a rigid-body 33° rotation of helix bundles that translocate the bound substrates by 11 Å across the membrane, thereby transporting substrates from crude urine into the renal epithelial cell (Fig. 4A).

Fig. 4. Conformational changes during NaS1 transport cycle.

(A) The movements of the transport domains in NaS1 leads to continuous conformational changes during NaS1 transport cycle. Cylinder representation of the transport domains and surface representation of the scaffold domains are shown. The protomers in the outward, partial-outward, partial-inward, and inward conformations are colored golden, blue, orange, and cyan, respectively. Red dotted circles indicate the position of the substrate binding sites. The bound Na⁺ and SO₄²⁻ are shown as spheres. (B) Heatmap of the surface electrostatics of NaS1 in intermediate state 2. Ribbon representation of the HP_in and HP_out in both protomers are shown. (C to E) Structural comparison among NaS1 protomers in the outward, partial-outward, partial-inward, and inward conformations viewed from outward (C), TMs (D), and inward (E). The red arrows indicate the movements of TMs and loop during conformational changes. (F) Structural comparison of substrate binding sites among NaS1 protomers in the outward, partial-outward, partial-inward, and inward conformations.

By comparing the four different NaS1 protomers within the five states of NaS1, we next set to delineate sequential movement of NaS1 helical bundles through the transport cycle. Viewing from the extracellular side (Fig. 4C), once the transport cycle initiates, the conformational transition from the outward to partial-outward/inward and to inward state is brought about by multiple adjustments of the TM helices, including the large anticlockwise rotation of TM10a, TM5b, and HP_out, thus moving the transporter domain toward the scaffold domain and sealing the outward facing substrate binding cavity. This conformational change is also accompanied by the anticlockwise rotations of other TM helices, like TM11, TM6, and H9C, which jointly help accomplish the configuration transition (Fig. 4D). Intriguingly, the connecting segment between TM3 and TM4 is a small helix both in the outward and occluded protomer but transforms into an irregular loop in the inward protomer, largely because of the steric contact introduced by the close movement of TM5b. On the other hand, viewing from the cytoplasmic side (Fig. 4E), such a configuration transition results in clockwise rotation of transport domain, especially for the movement of HP_in, TM10b, and TM5a, that further open up the substrate binding cavity and translocate the binding substrates to intracellular. Collectively, among the four different protomers, the sulfate ion is spatially closer to the Na1 site than that to Na2 site, albeit the conformation transition through the transport cycle leads to the subtle configuration changes within the substrate binding pocket (Fig. 4F).

DISCUSSION

The absorption and reabsorption of sulfate mediated by the Na⁺-sulfate cotransporter NaS1 and the anion exchanger SAT1 are crucial for the maintenance of sulfate homeostasis in mammalian physiology. In this process, sulfate in the lumen is first transported into the renal tubular or small intestinal epithelial cells by NaS1 located at the BBM, followed by transportation across the BLM by SAT1, and lastly entering the blood. Through our study of the transport mechanism and conformational changes of NaS1, we propose a model of the NaS1 transport cycle in human sulfate (re)absorption (Fig. 5). Previous studies on the DASS family of proteins have identified two conformations: inward-open and outward-open, proposing an elevator transport model, in which the two protomers of the protein transit from outward-open to inward-open conformations during transport. In this study, our results suggest a more complex mechanism during NaS1 transport cycle. First, the outward-open to inward-open conformational change of NaS1 contains many intermediate conformations. Second, the identification of two asymmetric intermediate conformations—intermediate states 1 and 2—indicates that the two protomers of NaS1 do not always move synchronously during transport. Third, the existence of symmetrical intermediate conformation—intermediate state 3—suggests that the conformational change in NaS1 is not initiated by one protomer completely transitioning from outward-open to inward-open before the other. In addition, in this study, we analyzed the conformational changes and transport mechanisms of NaS1 during the substrate translocation from the lumen to the cell. In this process, NaS1 is in a substrate-bound state. The complete transport cycle of NaS1 also includes the conformational change from inward-open to outward-open after substrate release, initiating the next round of substrate transport cycle. We hypothesize that NaS1 may undergo similar intermediate states and conformational changes during this process as it does during substrate binding.

Fig. 5. Proposed model for NaS1 transport cycle in sulfate (re)absorption process.

Schematic model of the NaS1 transport cycle in sulfate (re)absorption process. Sulfate is absorbed in small intestine and reabsorbed in the renal tubule, which are dependent on two sulfate transporters, NaS1 and SAT1. Sulfate is first transported into the renal epithelial cell across the BBM by NaS1 and then transported into the capillary across the BLM by SAT1, thus completing the sulfate (re)absorption process. During NaS1 transport cycle, apo NaS1 adopts an outward-open conformation (C_out), exposing its substrate binding sites to the lumen. The binding of substrates induces the movement of transport domains, resulting in a series of sequential conformational changes (C_inter1–3-S) from the outward-open state (C_out-S) to the inward-open state (C_in-S), allowing the substrates to be transported into the intestinal/renal epithelial cell. After the release of the substrate, NaS1 undergoes the conformational change from the inward-open state (C_in) to the outward-open state (C_out), ready to initiate another transport cycle. During this step, NaS1 may also exhibit a series of intermediate conformations (C_inter). The scaffold domains of NaS1 are colored wheat, and the two transport domains of NaS1 are colored bluish violet and light green, respectively. The Na⁺ and SO₄²⁻ ions are shown as spheres. The blue triangle indicates the substrates of SAT1 such as Cl⁻, bicarbonate and oxalate. C, conformation; in, inward-open state; out, outward-open state; inter, intermediate state; S, substrates.

During our cryo-EM data collection and analysis, the uneven distribution of protein particles in different conformations on the one hand allowed us to identify three intermediate states but one the other hand limited our ability to identify more. We supposed that more intermediate states may exist during NaS1 transport cycle. Future studies will focus on elucidating more intermediate conformations of NaS1 to further clarify its transport mechanism. Sequence comparisons indicate that NaS1 is conserved across various organisms (fig. S12). This suggests that the transport mode of human NaS1 may also be present in other mammals. In addition, we hypothesize that the transport mechanism of NaS1 could be widespread in other DASS family proteins; however, further studies will be required.

Previously studies reported that several human diseases are implicated in NaS1 mutations. In addition to many nonsense and frameshift mutations that lead to complete loss of protein function, two mutations, N174S and R237C, occurring in the transport domain are noteworthy. Particularly, N174 is in the loop region connecting HPin and TM5a, while R237 is in TM5a near L5ab. These two mutations may affect the stability of the HPin and L5ab structures, which play crucial roles in substrate binding, thereby affecting substrate transport.

In conclusion, our research sheds light on the substrate-recognition mechanism and the transport mechanism of NaS1. These findings not only broaden our comprehension of the transport mechanisms inherent to DASS family proteins but also shed light on human sulfate (re)absorption.

MATERIALS AND METHODS

Materials

The following reagents were purchased from Sigma-Aldrich: NaCl (S9888), KCl (P9541), MgCl₂ (M8266), CaCl₂ (C3306), Na₂SO₄ (239313), Hepes (H3375). In addition, protease Inhibitor Cocktail Tablets (04693116001) were obtained from Roche. LMNG/CHS Pre-Made Solution (NG310-CH210) (10:1) was obtained from Anatrace. Strep-Tactin resin (2-1208-500) and d-desthiobiotin (2-1000-005) were ordered from IBA. For cell culture, human embryonic kidney (HEK) 293F cell line (R79007) and HEK293T cell line (CRL-3216) were obtained from Thermo Fisher Scientific and American Type Culture Collection, respectively. The HEK293F cell culture medium (SMM 293-TII) was obtained from Sino Biological, Dulbecco’s modified Eagle’s medium (C11995599BT) and fetal bovine serum (10091148) were ordered from Gibco, and the Penicillin-Streptomycin Solution (SV30010) was ordered from Hyclone. We obtained the transfect reagent polyethylenimine (PEI, Linear, molecular weight 25,000; 23966) from Polysciences and the VigoFect (T001) from Vigorous Biotechnology. For immunoblotting, mouse anti-Strep monoclonal antibody (catalog no. BE2076, RRID:AB_2801290) and mouse anti-actin monoclonal antibody (catalog no. BE0021) were ordered from Bioeasytech, and the Goat Anti Mouse IgG, HRP Conjugated Antibody (catalog no. CW0102, RRID:AB_2814710) was ordered from CWbiotech.

Cell culture and transfection

For protein expression and purification, the optimized coding cDNA for Homo sapiens NAS1 (Uniprot: Q9BZW2-1) including a C-terminal tandem twin Strep-tag was cloned into the pcDNA3.1(−) vector. HEK293F cells were cultured in medium supplemented with 1× penicillin/streptomycin in a Multitron-Pro shaker (Infors, 120 rpm) at 37°C with 5% CO₂. To produce intact human NaS1 proteins, 1 mg of plasmids was preincubated with 2.6 mg of PEI in 50 ml of fresh medium for 25 min before adding the mixture to 1-liter cells when cell density reached 1.8 × 10⁶ per milliliter. The transfected cells were cultured for 50 hours before harvesting.

For whole-cell transport assay, HEK293T cells were seeded in 24-well plates (PerkinElmer, 1450), with around 125,000 cells per well. After 24 hours, the cells were transfected with NaS1 (WT or mutants) in pcDNA3.1 vector. Transfection was performed using VigoFect according to manufacturer’s protocol, with 500 ng of DNA and 1 μl of VigoFect per well.

Protein expression and purification for cryo-EM analysis

For one batch of NaS1 purification, about 6 liters of transfected cells were harvested by centrifugation at 3000g. All procedures below were carried out at 4°C or on ice.

For NaS1/Na⁺/SO₄²⁻ complex purification, harvested cells were resuspended in lysis buffer containing 25 mM Hepes (pH 7.4), 75 mM Na₂SO₄, 2 mM dithiothreitol (DTT), and protease inhibitor and then lysed by a high-pressure homogenizer. After removal of cell debris by centrifugation at 10,000g for 45 min, cell membrane fraction was pelleted by ultracentrifugation at 150,000g for 1 hour. The membrane fraction was resuspended and solubilized in lysis buffer plus 1% (w/v) LMNG and 0.1% (w/v) CHS for 2 hours with gentle rotation. After ultracentrifugation at 150,000g for 30 min, the supernatant was passed through a column filled with Strep-Tactin Sepharose resin. The resin was washed 50 CV (column volume) with wash buffer containing 25 mM Hepes (pH 7.4), 75 mM Na₂SO₄, 2 mM DTT, 0.005% (w/v) LMNG, and 0.0005% (w/v) CHS. The target NaS1 protein was eluted with wash buffer plus 10 mM desthiobiotin. The eluted NaS1 protein was concentrated to a final volume of ~100 μl by a 100-kDa cutoff centrifugal filter (Millipore) and further purified by size exclusion chromatography (SEC) (Superose 6 5/150, GE Healthcare) in SEC buffer containing 25 mM Hepes (pH 7.4), 75 mM Na₂SO₄, 0.003% (w/v) LMNG, and 0.0003% (w/v) CHS. The SEC fractions corresponding to NaS1/Na⁺/SO₄²⁻ complex were collected and verified by SDS–polyacrylamide gel electrophoresis (SDS-PAGE) for cyro-EM sample preparation (fig. S1). The peak fractions were concentrated to 10 to 12 mg/ml for grid preparation.

Electron microscopy sample preparation and imaging

The cryo-EM grids of NaS1/Na⁺/SO₄²⁻ complex were prepared using Vitrobot Mark IV (FEI) at 8°C and 100% humidity. Aliquots of samples (3 to 4 μl) at a concentration of 10 to 12 mg/ml were applied onto glow-discharged holey carbon grids (Quantifoil Au R1.2/1.3 300 mesh). After a waiting time of 10 s, the grids were blotted for 2 to 5 s and plunged into liquid ethane for quick freezing. The grids were screened on a Tecnai Arctica microscope (FEI) operated at 200 kV using a Falcon II direct electron detector (FEI). The qualified grids were transferred into a Titan Krios microscope (FEI) operated at 300 kV equipped with an energy filter (slit width, 20 eV; GIF Quantum LS, Gatan) for data acquisition.

Images of NaS1/Na⁺/SO₄²⁻ complex were recorded using a K3 submit direct electron detector (Gatan) in a super mode at a nominal magnification of ×105,000, corresponding to a calibrated pixel size of 0.4187 Å. Date acquisition was performed automatically using AutoEMation2.0 (69) in a movie mode, with a frame exposure time of 0.08 s and a total exposure time of 2.56 s, resulting in a total of 32 frames per stack, and the total dose rate for each stack was ~50 e⁻ Å⁻². All 32 frames in each stack were aligned and summed using the whole-image motion correction program MotionCor2 (70) and binned to a pixel size of 0.8374 Å.

Image processing and three-dimensional reconstruction

For the dataset of NaS1/Na⁺/SO₄²⁻ complex, a total of 14,234 micrographs (movie stacks) were collected, respectively. The image processing steps were carried out using cryoSPARC (71).

Micrographs were imported and the contrast transfer function (CTF) corrections were performed using Patch CTF estimation. After the deletion of bad micrographs, 300 micrographs were used for automatic picking by blob picker, and these particles were subjected to two-dimensional (2D) classification. The class averages representing projections of the NaS1/Na⁺/SO₄²⁻ complex in different orientations were chosen as templates for template picking from the whole dataset. A total of 10,754,130 particles were picked from 13,969 micrographs. These particles were extracted and binned three times and subjected to 2D classification. After three rounds of 2D classification, ~1896 K particles in good 2D averages were chosen. These particles were subjected to the ab initio reconstruction and the following heterogeneous refinement for six classes. The map of each class was measured in Chimera (72), the particles from the best class were selected and re-extracted to the original pixel size 0.8374 Å. After another round of ab initio reconstruction and heterogeneous refinement, three different states of NaS1/Na⁺/SO₄²⁻ complex were identified. The particles for each state were respectively set to do another one or two rounds of ab initio reconstruction and heterogeneous refinement to identify more different states and delete bad particles. Then, the best particles for each of the five different NaS1 states were used for 3D reconstruction by nonuniform refinement (73) to further improve data quality. Last, for intermediate states 1 and 2, a 3.04-Å resolution map and a 2.73-Å resolution map with no symmetry imposed were obtained. For the outward-open state, the inward-open state, and intermediate state 3, a 3.30-Å resolution map, a 3.35-Å resolution map, and a 3.25-Å resolution map with C2 symmetry imposed were obtained, respectively. The local resolution maps of the five states of NaS1/Na⁺/SO₄²⁻ complex were calculated using local resolution estimation in cryoSPARC and displayed in ChimeraX (74). Please refer to figs. S2 and S3 for the workflow of image processing.

Model building

The model of NaS1/Na⁺/SO₄²⁻ complex in intermediate state 2 was first built from de novo in COOT (75) because of its highest resolution (fig. S6). Before model building, models of full-length NaS1 were predicted on AlphaFold (76, 77). Sequence alignment and secondary structure prediction of NaS1 were used to aid the model building. The predicted model of human NaS1 was docked into the cryo-EM map with a resolution of 2.73 Å in Chimera and manually adjusted in Coot to acquire the atomic model of NaS1 (72, 78). Model refinement was performed on the main chain of the two atomic models using the real_space_refine module of PHENIX (79) with secondary structure and geometry restraints to avoid overfitting. Then, the model of NaS1/Na⁺/SO₄²⁻ complex in intermediate state 2 was used as initial model for the model building of the other four NaS1 structures. After manual adjustment in COOT, all five NaS1 models were subjected to real-space refinement in PHENIX, respectively. Cryo-EM data collection and refinement statistics are shown in table S1.

Immunoblotting

To examine whether WT NaS1 and its mutants have the same expression level, NaS1-FLAG WT or mutants were expressed by transfect 10-ml HEK293F cells with a ratio of 10-μg plasmid: 25-μg PEI. After 48 hours, transfected cells were collected and lysed in 1.5 ml of lysis buffer containing 25 mM Hepes (pH 7.4), 150 mM NaCl, 1% (w/v) Triton X-100, and protease inhibitors. The cell lysate was incubated for 2 hours at 4°C with gentle rotation. After centrifugation at 15,000g for 10 min, the cell debris were removed, and the supernatant was collected. For each protein sample, 20 μl of supernatant was incubated with peptide N-glycosidase F at 37°C for 30 min and then mixed with SDS loading buffer. Then, the sample (10 μl of each protein) was resolved by 4 to 20% SDS-PAGE (GenScript) and transferred to a polyvinylidene difluoride membrane (Millipore). The same SDS-PAGE and the following transfer procedure were repeated to obtain two membranes containing the same amount of proteins. The two membrane were blocked with 5% nonfat dry milk (Bio-Rad) in phosphate-buffered saline with 0.1% Tween 20 (PBS-T) for 1 hour. Then, one membrane was incubated with a mouse anti-Strep monoclonal antibody (Bioeasytech, catalog no. BE2076) at a 1:3000 dilution for 1 hour at room temperature (RT), whereas the other membrane was incubated with a mouse anti-actin monoclonal antibody (Bioeasytech, catalog no. BE0021) at a 1:3000 dilution for 1 hour at RT. The two membranes were washed three times with PBS-T for 5 min each. Goat Anti Mouse IgG, HRP Conjugated Antibody (CWbiotech, catalog no. CW0102) was then added to the membranes at a 1:5000 dilution for 1 hour at RT. The membranes were then washed three more times with PBS-T, and the proteins were detected with enhanced chemiluminescent substrate (Pierce) by Amersham Imager 600 (GE Healthcare).

Whole-cell transport assay

HEK293T cells were maintained in 24-well plates (PerkinElmer, 1450) and transfected with NaS1 (WT or mutants). Forty-eight hours after transfection, the culture medium was removed, and the cells were washed with wash buffer containing 150 mM NaCl, 2 mM KCl, 1 mM MgCl₂, 1 mM CaCl₂, and 10 mM Hepes (pH 7.4). Then, the buffer was removed, and cells were incubated with wash buffer containing 500 μM total sodium sulfate, including 50 μM [³⁵S] sodium sulfate (PerkinElmer, NEX041005MC, 5 mCi/ml) at 37°C. After a certain amount of time, the [³⁵S] sodium sulfate buffer was removed, and cells were washed three times with 0.5 ml of ice-cold wash buffer to terminate the reaction and remove residual isotopes completely. Next, 0.5 ml of OptiPhase HiSafe 3 liquid scintillation cocktail (PerkinElmer, 1200.437) was added per well and mixed on a shaker for 10 min. Counts were measured using a MicroBeta2 Microplate counter (PerkinElmer, 2450). All procedures were carried out at RT except the incubation step. Statistical significance was tested using the unpaired t test, with P < 0.05 considered significant.

Supplementary Materials

This PDF file includes:

Figs. S1 to S12

Table S1