Edit · 2026 01 11 200723 Nature S41594 025 01568 8

Saving needs a token. Log in or register to get one. Edits are public and licensed CC-BY-SA-4.0.

# Substrate recognition and allosteric regulation of synaptic vesicle glutamate transporter VGLUT2 | Nature Structural & Molecular Biology

**URL:** https://www.nature.com/articles/s41594-025-01568-8

**Saved:** 2026-01-11T04:07:28.753Z

---

Article
Published: 03 June 2025
Substrate recognition and allosteric regulation of synaptic vesicle glutamate transporter VGLUT2
Fei Li, Jacob Eriksen, Juan A. Oses-Prieto, Yessica K. Gomez, Hongfei Xu, Surabhi Hareendranath, Poulomi Das, Janet Finer-Moore, Phuong Nguyen, Alisa Bowen, Andrew Nelson, Alma Burlingame, Michael Grabe, Robert M. Stroud & Robert H. Edwards

Nature Structural & Molecular Biology 
volume
 32, 
pages
1479–1487 (2025)Cite this article

3235 Accesses

3 Citations

3 Altmetric

Metrics
details

Abstract

The concentration of neurotransmitters inside synaptic vesicles (SVs) underlies the quantal nature of synaptic transmission. Uptake of glutamate, the principal excitatory neurotransmitter, is driven by membrane potential. To prevent nonquantal efflux of glutamate after SV exocytosis, the vesicular glutamate transporters (VGLUTs) are allosterically inhibited by the neutral pH of the synaptic cleft. To elucidate the mechanism, we determined high-resolution structures of rat VGLUT2 with a cyclic analog of glutamate. We propose a mechanism of substrate recognition in which a positively charged cytoplasmic vestibule electrostatically attracts the negatively charged substrate. We also identify modification of VGLUT2 by palmitoylation and find that this promotes retrieval of the transporter after exocytosis. The structure also reveals an extensive network of electrostatic interactions that forms the cytoplasmic gate. Functional analysis of a mutant that disrupts the network shows how this cytoplasmic gate confers the allosteric requirement for lumenal H+ required to restrict VGLUT activity to SVs.

You have full access to this article via your institution.

Download PDF
Similar content being viewed by others
Vesicular glutamate transporters are H+-anion exchangers that operate at variable stoichiometry
Article Open access
11 May 2023
Specialized astrocytes mediate glutamatergic gliotransmission in the CNS
Article Open access
06 September 2023
Opponent vesicular transporters regulate the strength of glutamatergic neurotransmission in a C. elegans sensory circuit
Article Open access
03 November 2021
Main

The storage of neurotransmitters inside SVs enables their release by regulated exocytosis, conferring the quantal mode of neurotransmission1. The amino acid glutamate serves as the principal excitatory neurotransmitter in most neural systems and is packaged into SVs by highly selective but low-affinity VGLUTs2,3,4,5. Mammals express three closely related VGLUT isoforms with similar transport activity but different expression patterns6,7,8. Loss of VGLUT expression eliminates glutamatergic transmission8,9.

The VGLUTs belong to the solute carrier 17 (SLC17) family of secondary active transporters10. Most secondary active transporters couple the concentration of substrate to the flux of another molecule, in many cases, an ion11. Indeed, the highly homologous lysosomal sialic acid transporter sialin (41% sequence identity to VGLUT2) and the Escherichia coli D-galactonate transporter DgoT (23% sequence identity) use H+ symport to drive organic anion transport into the cytoplasm12,13,14 (Extended Data Fig. 1). By contrast, the VGLUTs are not coupled to H+ but rather driven by membrane potential (ΔΨ), and transport in the opposite direction, moving glutamate from the cytoplasm into SVs, against a pH gradient3,15. Equilibrating the anionic substrate according to ΔΨ enables the VGLUTs to concentrate glutamate 10–20-fold. Indeed, SVs contain >100 mM glutamate16, a concentration similar to that predicted by coupling to ΔΨ (ref. 17). Despite the concentration of glutamate inside SVs, the VGLUTs do not leak glutamate even after dissipation of ΔΨ (refs. 18,19), and the mechanism responsible remains unknown.

SV exocytosis exposes VGLUTs to the positive outside cell membrane potential, which has the same orientation as the SV membrane potential and should drive non-vesicular glutamate efflux from the neuron. To prevent the degradation of quantal neurotransmission, glutamate transport at the plasma membrane is prevented by the change in environment, in particular, the pH, which is neutral in the synaptic cleft rather than acidic as inside SVs20. Glutamate uptake by SVs also requires the efflux of Cl− trapped by endocytosis, and the VGLUTs exhibit an associated Cl− conductance as well as allosteric regulation by Cl− (refs. 2,3,4,20,21,22,23,24,25,26,27,28). These properties take advantage of the rapidly changing ionic conditions associated with the SV cycle to enhance and to control glutamate flux by the VGLUTs, but the mechanisms for allosteric regulation by H+ and Cl− remain poorly understood.

Results
Substrate recognition

The high rates of release observed at many synapses require mechanisms to recapture the vesicular transporters into recycled SVs and rapidly refill SVs with high concentrations of neurotransmitter. Direct measurement has shown that refilling with glutamate can occur within 15 s26. High presynaptic concentrations of glutamate facilitate this process and require an appropriately tuned transport affinity. At the same time, the VGLUTs must discriminate between glutamate and similar compounds that also accumulate in the nerve terminal, such as aspartate29. Consistent with these requirements, the VGLUTs transport glutamate with high specificity but with low apparent affinity (Km ~ 1–3 mM)30. Classical studies showed that synaptic vesicle glutamate uptake does not transport or recognize closely related aspartate2. More recent work has identified aspartate currents associated with the VGLUTs, but these were detected at 0.14 M aspartate31. To test recognition of aspartate, we used a recently developed assay for substrate efflux by plasma membrane-targeted (pm)VGLUT2 expressed in HEK293T cells32. In this assay, VGLUT-mediated efflux of preloaded 3H-glutamate depends on low external pH that mimics the lumenal pH of SVs. In contrast to the robust efflux of glutamate32, we cannot detect any aspartate efflux mediated by VGLUT2 (Extended Data Fig. 2). Despite expression at the plasma membrane, transport by VGLUT2 thus remains specific for glutamate.

To understand the basis for this substrate specificity, we used a cyclic analog of glutamate, L-trans-1-amino-1,3-dicarboxy cyclopentane (ACPD), which is a substrate of the VGLUTs with higher apparent affinity (Km ~ 0.23 mM) than glutamate33. To stabilize a substrate-bound conformation, we used a mutant (R181Q/E191Q) that impairs vesicular glutamate transport after reconstitution (Extended Data Fig. 3c) but allows glutamate efflux from transfected HEK293T cells (Extended Data Fig. 3d), suggesting transport at a lower rate that might be more amenable to a structure with substrate. Given that high concentrations of Cl− are also non-competitive inhibitors of glutamate transport25, we used low Cl− (1 mM). The structure of the R181Q/E191Q VGLUT2–ACPD complex was determined using a previously described antigen-binding fragment (Fab)34 to 3.0 Å resolution (Fig. 1a, Table 1 and Extended Data Fig. 4).

Fig. 1: Substrate binding site.

a, Overall structure of the R184Q/E191Q VGLUT2–Fab complex bound to ACPD. The N-domain is indicated in blue, the C-domain is in pink and Fab is in yellow. ACPD is shown as a space-filling model in green. EM density is shown as a white transparent surface. b, Structures of ACPD and glutamate. c,d, Binding site with the density of ACPD shown in green. e,f, Interactions of the labeled residues with the ligand are indicated with a yellow dashed line, with the distances (Å) in red. A cation-π interaction between the amine group of ACPD and F326 is also observed with a distance of 3.3 Å. g, Effect of mutations on glutamate efflux from transfected HEK cells by internalization-defective VGLUT2. 5M, D97A/E122A/E338A/E339A/E344A VGLUT2. Error bars, s.e.m. ****P = 9.55 × 10−12; n = 3 biological replicates for R322A, R322K, Y327S, Y327T and 5M; n = 4 for S323T, F326A and Y327F; n = 7 for EV; n = 8 for WT.

Source data

Full size image
Table 1 Cryo-EM data collection, refinement and validation statistics
Full size table

The VGLUT2–ACPD complex adopts a lumen-open conformation, with ACPD bound at the center of the protein between the amino-domain and carboxy-domain helical bundles (Fig. 1a–f). The two carboxyl groups of ACPD are coordinated in almost symmetrical fashion, each by one arginine and two tyrosine residues. The 3′-carboxyl is coordinated by R88 and Y135 from the N-domain and Y327 from the C-domain, whereas the 1′-carboxyl is coordinated by R322 from the C-domain as well as Y195 and Y223 from the N-domain. Both domains thus coordinate glutamate to make multiple hydrogen bonds as well as electrostatic interactions with each carboxyl (Fig. 1e,f). Interestingly, the critical residue R322 (ref. 25) uses its ε-NH in a folded-back configuration, rather than its terminal η-NH2, to form a hydrogen-bonded salt bridge with the 1′-carboxyl (Fig. 1e,f). This configuration can rationalize the inability of VGLUTs to recognize aspartate2: with ~3.4 Å between the 1′-carboxyl oxygen of ACPD and the ε-nitrogen of R322 (Fig. 1e,f), the shorter aspartate would not bridge the distance between R88 and R322. Previous studies showed that the amine group of glutamate is critical for recognition by the VGLUTs2,33. The amine of ACPD lies 3.3 Å away from F326, forming a cation–π interaction with the phenyl ring (Fig. 1f). In addition, an outer layer of polar residues including N431, N450 and S323 surrounds the recognition site and may also interact with the amino group of ACPD and glutamate in the fully occluded conformation.

We tested the importance of the glutamate and ACPD interacting residues using the efflux assay (Fig. 1g) and confirmed the important roles for these residues in transport. The Y327F mutation greatly impairs but does not eliminate transport, supporting a role for the sidechain hydroxyl in forming a hydrogen bond with the substrate carboxyl. Although this mutant shows reduced expression, the surface fraction is not significantly reduced (Extended Data Fig. 2). Consistent with a role for this residue in hydrogen bond formation with substrate, Y327T impairs release to a lesser extent. The more severe effect of Y327S further indicates that the position of the sidechain hydroxyl is important. Other than Y327F, all of these mutants are normally expressed (Extended Data Fig. 2). They are also highly conserved among the VGLUTs (Extended Data Fig. 1). Mutation of the immediately adjacent F326 (F326A), which forms a cation–π interaction with the amine group of ACPD, also eliminates activity. The conservative mutation of the nearby S323 (S323T) reduces glutamate efflux.

As validation, ACPD was docked into the R184Q/E191Q structure using the Schrödinger software. This resulted in a similar pose that preserves the interaction landscape with nearly identical ring and functional group placement (Extended Data Fig. 5a). We also docked glutamate into the wild-type (WT) apo structure (PDB 6V4D)34. The resulting top-ranked glutamate pose was again similar to the reported ACPD structure, with good alignment between the carboxyl groups but a different orientation of the amine group (Extended Data Fig. 5b). To probe the stability of these interactions, molecular dynamics simulations of glutamate with WT VGLUT2 and of ACPD with R184Q/E191Q VGLUT2 were carried out. The ACPD simulation was initiated from the structure, and the glutamate simulation was initiated from the docked pose. Both molecular dynamics simulations indicated that R88 and R322 form stable bidentate interactions with the ligand carboxyl groups (Extended Data Fig. 5e,f); as observed in the structure, the ε-NH and one η-NH2 of R322 form a salt bridge with the substrate. The amine group of ACPD engages F326 for the majority of the simulation but sometimes flips and reorients towards N450 with a distance ~4 Å (Extended Data Fig. 5d). Similarly, the amine of glutamate persistently contacts the sidechain oxygen of N450 and occasionally rotates to interact with F326 (Extended Data Fig. 5d), consistent with the importance of the amine group for recognition by VGLUT2,33.

To understand how the VGLUTs discriminate against aspartate, we also performed unbiased molecular dynamics. The simulations showed that, in contrast to glutamate, aspartate is loose in the pocket and tumbles frequently (Extended Data Fig. 5c,g,h). The sidechain of R322 is flexible, with rotamers that could extend to accommodate the shorter L-aspartate ligand, but the engagement is less stable, with increased tumbling frequency and calculated root mean squared deviation values relative to the starting pose. Water molecules could potentially bridge the distance to aspartate, but quantitative alchemical free energy perturbation calculations show a 4 kcal mol−1 greater selectivity for L-glutamate (see ‘Alchemical free energy calculations’ in the Methods).

To test the role of the two arginines in substrate recognition, we examined the effect of mutations. We previously found that neutralization of the transmembrane helix 1 (TM1) arginine (R88A in VGLUT2) eliminates flux triggered at low pH by all three VGLUT isoforms25 and greatly impairs synaptic glutamate release34. Consistent with an important role for charge, we now find that conservative mutation R88K only modestly impairs glutamate efflux and fails to confer recognition of aspartate (Extended Data Fig. 2a,b). In TM7, R322A eliminated transport, whereas R322K was not expressed and transport activity could therefore not be determined (Fig. 1g and Extended Data Fig. 2).

Substrate release

To activate postsynaptic receptors, SVs accumulate high millimolar concentrations of neurotransmitters. This requires discharge into the SV lumen that already contains high concentrations of transmitters as the SV fills. To determine the structural changes associated with glutamate discharge by the VGLUTs, we compared two VGLUT2 structures: R184Q/E191Q with the substrate analog ACPD, and WT without substrate (Fig. 2). To simplify the comparison, we determined the WT structure under the same low Cl− condition to 3.3 Å resolution (Table 1 and Extended Data Fig. 6). Both structures open to the lumen and thus differ primarily in the presence or absence of substrate. In the ACPD structure, the lumenal half of the C-domain rotates slightly over ACPD (Fig. 2a,b). Substrate-binding residues from the N-domain do not change between the two structures, whereas TM7 of the C-domain exhibits the largest conformational change. L330, Y327 and F326, bulky residues on the lumenal side of TM7, rotate significantly to enclose ACPD in the central binding site. In addition, S323, R322 and N319 on the cytoplasmic side of TM7 translate toward the N-domain (Fig. 2b).

Fig. 2: Rearrangement of the substrate recognition site associated with release.

a, Structure of R184Q/E191Q VGLUT2 with ACPD bound adopts a similar but more closed lumen-open conformation than WT, with the lumenal section of TM7 exhibiting the largest change. N/C, N-domain/C-domain. b, Bulky sidechains on TM7 (Y327, F326 and L330) have the major role in closing the substrate binding site to the lumen. c, Magnified view of the substrate-binding site in R184Q/E191Q VGLUT2 with ACPD bound. d, Substrate binding site of the apo WT VGLUT2 structure. e, Starting point for additional molecular dynamics simulations of WT VGLUT2 with L-Glu in a lower-scoring docked pose. f, By the end of equilibration, L-Glu fails to engage R322, which moves to interact with S423 and E396, similar to the R184Q/E191Q VGLUT2 structure in c. Interactions are indicated with a yellow dashed line, and the distances are indicated (in Å).

Full size image

Differences between the two structures in the interaction network around the essential R322 indicate a mechanism that accompanies substrate release into the vesicle lumen (Fig. 2c,d). In the ACPD-bound structure, both Y223 and R322 coordinate the substrate carboxyl, whereas the η-NH2 of R322 forms a hydrogen bond with S423 (Fig. 2c). In the apo structure, by contrast, S423 is sequestered by E396, releasing R322 to form a hydrogen bond with Y223 (Fig. 2d). In this configuration, the substrate 1′-carboxyl group loses two of three critical interactions with the transporter, reducing the enthalpic component.

To test this mechanism of substrate release, we carried out simulations starting from two additional docked poses (Supplementary Videos 1–4). In the first of these poses, L-Glu starts engaged by both R88 and R322 and remains stably bound for the full 1 µs of simulation (Extended Data Fig. 5d). In the second docked pose, L-Glu engages only R88, and in the course of equilibration, R322 moves away to interact with S423 and E396, similar to the apo structure (Fig. 2e,f). Glutamate then exits from the lumenal surface of the transporter at 200 ns. To assess the reproducibility of this observation, we performed an independent simulation from the same starting pose and observed lumenal exit of the substrate again, at 700 ns. Therefore, the transporter is open wide enough to allow exit, suggesting that engagement by both arginines is critical for stable binding of glutamate.

An asymmetric transport pathway

In the absence of neural activity, SVs exhibit very low rates of glutamate leakage even with dissipation of ΔΨ, the driving force for glutamate uptake18,19. The structures of VGLUT2 reveal an asymmetry in the pathways leading to the central substrate binding site that may reduce leakage. In the lumen-open conformation, a series of acidic residues (D97, E122, E338, E339 and E344) line the mouth of the lumenal cavity, disfavoring an anion in the vestibule, and the path to the central cavity is much more hydrophobic (Fig. 3a,b,c). Taking advantage of the cytoplasm-open structure of DgoT14, we generated models of the cytoplasm-open conformation for both R184Q/E191Q and WT VGLUT2. In both models, the N-domain and C-domain of VGLUT2 adopt structures very similar to their respective domains in DgoT, with conserved residues at identical positions (Extended Data Fig. 7a,b). However, the surface electrostatics of VGLUT2 show strikingly different paths to the central cavity between cytoplasm-open and lumen-open conformations. In the cytoplasm-open conformation, a continuous positively charged path (Fig. 3d) electrostatically attracts entry of anionic glutamate from the cytoplasm. In general, residues lining both cavities show low sequence conservation across the SLC17 family; however, the vesicular nucleotide transporter, also driven by ΔΨ, has three acidic residues on the lumenal surface of TM7, similar to E338 and E339 in VGLUT2 and the homologous residues in VGLUT1 and VGLUT3 (Fig. 3a and Extended Data Fig. 1). By contrast, the H+-coupled symporter DgoT that drives substrate into the cytoplasm has a more symmetric binding cavity, with positive charge extending from the central recognition site to both sides of the membrane (Extended Data Fig. 7c–e). The negative charge and hydrophobicity of the VGLUTs impose a barrier to glutamate rebinding from the lumen, suggesting a mechanism for the low rates of leakage. This barrier may also impede glutamate release into the vesicle lumen, requiring other mechanisms for release, such as the change in conformation of R322 described above. To test these possibilities, we neutralized the negative charge on the five lumenally oriented acidic residues. The D97A/E122A/E338A/E339A/E344A (5M) mutant exhibits no transport activity; it also shows reduced expression but a fraction still localizes to the plasma membrane, and the Y327F mutant with similar expression confers glutamate efflux (Fig. 1g and Extended Data Fig. 2c). The negative charge of the five acidic residues thus appears to be important for flux.

Fig. 3: Asymmetrical paths to the central cavity facilitate directional transport driven by ΔΨ.

a, Structure of the binding cavity R184Q/E191Q VGLUT2 colored with sequence conservation as in Extended Data Fig. 1 (orange, most conserved; no color, least conserved). b, Electrostatic surface of the lumen-open R184Q/E191Q structure without ACPD. c, Electrostatic surface of the apo WT structure in lumen-open conformation. d, Electrostatic surface of a model for WT in the cytoplasm-open conformation. Locations of the substrate binding site in b–d are outlined by a dashed green circle.

Full size image
Palmitoylation regulates VGLUT trafficking

The high-quality map of R184Q/E191Q (Extended Data Fig. 4b) identifies a lipid density at conserved C62 of the cytoplasmic N terminus (Fig. 4a,b). A palmitate molecule is well-defined in the density, and another weaker density is also observed at adjacent residue C64 (Fig. 4b). Palmitoylation is a lipid modification that is observed on integral and peripheral membrane proteins that can regulate their association with membranes or membrane microdomains35. Palmitoylated proteins include the presynaptic cysteine string protein, t-SNAREs syntaxin, SNAP25 and v-SNARE synaptobrevin, but the role of palmitoylation in the exocytic and endocytic recycling of SV membrane proteins remains poorly understood36.

Fig. 4: VGLUT2 is palmitoylated at C60, C62 and C64.

a, Alignment of SLC17 protein cytosolic N-termini proximal to TM1, which contains the palmitoylated cysteine residues (red boxes). C62 and C64 are conserved in all three mammalian VGLUTs, whereas C60 is conserved in VGLUT2 and VGLUT3 but replaced by a valine in VGLUT1. Other SLC17 family members have a conserved cysteine closer to TM1 (green). Sequence alignment is colored to indicate conservation, as in Extended Data Fig. 1 (red, most conserved; no color, least conserved). b, Structure of R184Q/E191Q VGLUT2 with the C62 and C64 palmitoylate densities. The N-domain is blue and C-domain is pink. Two lipid densities (white) are resolved on C62 and C64. A palmitoyl group (C16) fits well into the density at C62. c, Primary hippocampal neurons expressing lumenally pHluorin-tagged WT or 3CS (C60S/C62S/C64S) VGLUT2 were stimulated at 10 Hz for 30 s, and the fluorescence was normalized to total pHluorin revealed by alkalinization in 50 mM NH4Cl (n = 13–14 coverslips). d,e, Hippocampal neurons expressing N-terminally cytoplasmic GFP-tagged WT or 3CS VGLUT2 were stimulated at 5 Hz and 25 Hz, and the fluorescence response over boutons (d) or axon shafts (e) was normalized to the baseline before stimulation. Analysis shows the time constants of reclustering over boutons (d) and the extent of recovery in axon shafts (e) after stimulation. n = 15 coverslips at 5 Hz and n = 16 coverslips at 25 Hz from three independent experiments (d); n = 20 coverslips from three independent experiments (e). The data indicate mean ± s.e.m. for individual coverslips containing 50–80 boutons each using a two-way ANOVA (d) with Tukey’s multiple comparisons test or unpaired two-tailed t-test (e).

Source data

Full size image

To confirm the identity of lipid modification and the sites modified, we analyzed the purified VGLUT2 used for cryogenic electron microscopy (cryo-EM) by mass spectrometry. We found modification of three N-terminal cysteines (C60, C62 and C64) by palmitic acid (Extended Data Fig. 8), consistent with putative sites identified from mouse brain37. To determine the role of palmitoylation, we monitored exocytosis and endocytosis of VGLUT2 in cultured neurons by live cell imaging using a pH-sensitive form of GFP (pHluorin) fused to the lumenal surface of the transporter. Fluorescence of the pHluorin is quenched by the low pH inside SVs and increases with exposure to the higher external pH during exocytosis38,39. Mutation of all three cysteines to serine (3CS) does not impair the initial exocytic response of VGLUT2 to stimulation (Fig. 4c). However, the mutations result in a greater fluorescence decline, below baseline, after stimulation. This decline reflects both the re-acidification that accompanies endocytosis of SV proteins and their dispersion into the axonal plasma membrane. In the case of the 3CS mutant, the subsequent return back up to baseline takes longer than for WT, suggesting increased axonal dispersion that slows endocytic recycling, reclustering and the regeneration of SVs. To directly assess the effects of palmitoylation on the dispersion of VGLUT2, we expressed a fusion of GFP to the cytoplasmic N terminus of VGLUT2 in hippocampal neurons and used stimulation to trigger its dispersion into the axon, followed by reclustering. Monitoring fluorescence over the boutons, stimulation at 5 Hz and 25 Hz both result in the dispersion of VGLUT2, and the 3CS mutant does not appear to affect this result. However, the 3CS mutant delays reclustering of the dispersed VGLUT2–GFP with stimulation at higher frequency (Fig. 4d,e). Imaging over the adjacent axon supports the delay in presynaptic reclustering of GFP–VGLUT2. Thus, palmitoylation serves to anchor VGLUT2 near the site of exocytosis and promote the efficient retrieval required for high rates of sustained glutamate release.

The cytoplasmic gate and regulation by lumenal pH

The alternating access mechanism of transport generally involves the sequential opening and closure of gates on each side of the membrane. The high-quality density map for R184Q/E191Q VGLUT2 now allows us to unambiguously resolve solvent-exposed flexible residues, revealing an extensive ionic interaction network that appears to act as the cytoplasmic gate. This network comprises alternating positively charged (arginine) and negatively charged (glutamate or aspartate) residues that interact with each other and span both the N-domain and C-domain, as well as intracellular helix 1 (ICH1), located between TM6 and TM7 (Fig. 5a,c). E278 in ICH1 binds R70 in TM1, which also interacts with E210 at the beginning of TM5. All three residues are highly conserved through the entire SLC17 family (Fig. 5b). Consistent with their importance, mutation of the residue equivalent to VGLUT2 R70 in sialin (R39C) causes Salla disease by impairing lysosomal efflux of sialic acid12,13. In VGLUT2, the network continues with interaction of E210 and R213 at the beginning of TM5 and of R213 with D371 at the bottom of TM8, an interaction that bridges the N-domain and C-domain (Fig. 5c). D371 is in turn surrounded by three arginines, R374 and R385 from the C-domain, as well as R213. Similarly, R374 and R385 from the C-domain and R211 from the N-domain surround D436 from the C-domain. These extensive ionic interactions stabilize the closed cytoplasmic gate in this lumen-open conformation. D371 is part of MFS cytosolic motif A [Gxxx(D/E)(R/K)xGx(R/K)(R/K)], which has been shown to stabilize the outward-open conformation and promote transport activity40,41.

Fig. 5: Cytoplasmic ionic interaction network.

a, Overall structure of R184Q/E191Q VGLUT2 with the cytoplasmic interaction network (box) magnified in c. The N-domain is blue and C-domain is pink. b, Sequence alignment of the cytoplasmic interaction network in VGLUT homologs and other SLC17 family members. Color indicates conservation (red, most conserved; no color, least conserved). R213 and D371 (cyan triangles) interact across the N-domain and C-domain. c, Ionic interactions between the N-domain and C-domain, as well as the ICH1 helix on the cytoplasmic face of R184Q/E191Q VGLUT2 in the lumen-open conformation. Distances between interacting atoms are labeled in red and between potential interactions in black. d, In the cytoplasm-open conformation of VGLUT2 (modeled on DgoT), the interaction between D371 and R213 is broken. e, VGLUT2 activity measured as 3H-glutamate efflux from HEK293T cells expressing WT, R213S and D371N pmVGLUT2. R213S transport activity resembles WT, but D371N shows reduced activity at pH 5.5 and increased activity at pH 7.4. f, pH response curve for glutamate efflux from cells expressing WT, D371N pmVGLUT2 and untransfected controls. WT shows potentiation by protons (EC50, pH = 6.0 ± 0.025), whereas D371N shows inhibition by protons (IC50, pH = 6.2 ± 0.27). n = 3 biological replicates for e and f. g, Maximum inward currents (at −120 mV) from oocytes expressing WT, R213S, D371N pmVGLUT2-HA and uninjected controls (n = 6 oocytes per condition). h, Individual I–V curves for each of the conditions in g. The data indicate mean ± s.e.m.; n.s., P > 0.05, ∗P < 0.05; ∗∗P < 0.01, two-way ANOVA with correction for multiple comparisons.

Source data

Full size image

To determine the functional significance of the cytoplasmic interaction network, we neutralized by mutation the two interacting residues (R213 and D371) that bridge the N-domain and the C-domain. The backbone amine group of R213 interacts with the sidechain carboxyl of D371, and the sidechain of R213 may form a weaker interaction with the backbone carbonyl of D371 (Fig. 5c). We assessed the effect on glutamate transport using the HEK cell efflux assay. The R213S mutation does not affect glutamate transport, with activity similar to WT (Fig. 5e), consistent with an interaction that involves the backbone of R213. However, D371N both reduces glutamate efflux and changes its pH dependence. In contrast to WT, D371N VGLUT2 shows more activity at high rather than low external pH (Fig. 5e). Therefore, D371N reverses the pH dependence of the VGLUTs. WT VGLUT2 shows activation by protons with a sigmoidal dose response (EC50, pH = 6.0 ± 0.025), whereas D371N shows inhibition by protons but with a similar pH sensitivity (IC50, pH 6.2 ± 0.27) (Fig. 5f). The differential effect of R213S and D371N is consistent with a critical role for the interaction between the backbone of R213 and the sidechain of D371. Given that the VGLUTs exhibit an associated Cl− conductance that is not coupled to glutamate flux, we also tested the effect of the mutations on this property using oocytes expressing pmVGLUT2 (ref. 20). The R213S mutation has no effect on these currents but D371N eliminates them (Fig. 5g,h and Extended Data Fig. 9) even though this mutation does not eliminate glutamate transport (Fig. 5e,f). Therefore, the integrity of the cytoplasmic gate both influences the allosteric regulation by lumenal H+ and enables the Cl− conductance associated with VGLUTs.

Discussion

The structures reported here indicate how the VGLUTs recognize glutamate with the low apparent affinity but high specificity required for the rapid refilling of SVs. One arginine and two tyrosine residues contact each of the carboxyl groups, and the distance between these two sites dictates the length requirement for substrate recognition that excludes a role for aspartate as neurotransmitter29. The arginine in TM1 is conserved through many members of the SLC17 family, consistent with a role in anion transport (Extended Data Fig. 1). However, the arginine in TM7 occurs only in the VGLUTs and not in any other members of SLC17, which transport substrates such as urate, p-aminohippurate and sialic acid that carry a single negative charge. The amine in glutamate forms either a cation–π interaction with F326 or a hydrogen bond with N450, and flipping between these two orientations observed by molecular dynamics simulation is consistent with the low apparent affinity of the VGLUTs.

Comparison of the two VGLUT2 structures with and without substrate suggests a mechanism that accompanies substrate discharge into the SV lumen. In the apo structure, two of the three residues recognizing the 1′ carboxyl of ACPD interact with each other rather than with the substrate. This change may result from the loss of substrate, or it may promote substrate discharge, a possibility supported by simulation. A mechanism to prevent glutamate rebinding from the lumen may become increasingly important as the SV fills with glutamate. In addition, after glutamate discharge into the lumen, Y223 may serve to neutralize R322 to reduce electrostatic repulsion, thus promoting reorientation of the unloaded carrier to the cytoplasm. Similarly, substrate glutamate serves to neutralize the positive charges on R88 and R322, enabling translocation of the loaded carrier from the cytoplasm back to the lumen and completing a productive transport cycle. Neutralization of R88 and R322 may therefore have an important role in the reorientation of both loaded and unloaded VGLUT.

An asymmetry in the main cavity of VGLUT2 suggests a mechanism to regulate the rates of substrate rebinding in the lumen. As shown in the model based on DgoT, the main cavity of cytoplasmically oriented VGLUT2 contains a positively charged path to the substrate recognition site, which should promote anion entry. In the lumenal orientation, however, the main cavity is lined with hydrophobic residues and the lumenal face is lined with acidic residues. The negative charge and hydrophobicity will slow rebinding from the lumen and, along with the requirement for allosteric activation by H+, may contribute to the low rates of leakage observed after inhibition of the v-type H+-ATPase18,19. The negative charge may also slow substrate discharge into the lumen, but this could be overcome by other mechanisms such as the rearrangement of substrate-recognizing residues described above.

The structure also reveals palmitoylation of the VGLUT2 N terminus. Many SV proteins undergo palmitoylation, but the significance has remained unclear36. We now find that this modification serves to restrict the movement of VGLUT2 in the plasma membrane after exocytosis, thus promoting its rapid retrieval by endocytosis and the regeneration of SVs capable of glutamate storage. Presumably, palmitoylation restricts VGLUT2 to a membrane microdomain35 surrounding the release site, predicting a similar role in the endocytic recapture of other SV proteins.

Based on the structures, we propose the following mechanism for ΔΨ-driven transport by the VGLUTs (Fig. 6). In the lumen or outward-open conformation, the VGLUTs are stabilized by extensive ionic interactions across the N-domain and C-domain at the cytoplasmic face of the protein. Neutralization of negative charges on the lumenal side of the VGLUT by H+ binding enables closure of a putative lumenal gate, rupturing the cytoplasmic network and allowing reorientation toward the cytoplasm. Glutamate then enters the substrate binding site through the positively charged, cytoplasmically oriented main cavity. Transition through an occluded state that involves recognition of each glutamate carboxyl by both the N-domain and C-domain and then reorients the transporter to the SV lumen, re-establishing the cytoplasmic network. In this conformation, a change in the interaction with R322 disrupts glutamate recognition, allowing substrate to move down its electrochemical gradient into the SV lumen. The negative charge and hydrophobicity of the lumenal vestibule then slow glutamate rebinding from loaded SVs, even if ΔΨ decreases. The empty transporter then returns to the cytoplasm-open conformation with R322 bound to Y223.

Fig. 6: Proposed transport cycle for VGLUT2.

At neutral pH in the plasma membrane, VGLUT is stabilized in the lumen-open conformation with the cytoplasmic gate closed (upper right panel). Protonation of a lumenal site (indicated here with a minus inside a circle) at the low pH of SVs enables closure of the lumenal gate, disrupting the interaction between R213 and D371, rupturing the cytoplasmic gate and driving reorientation to the cytoplasm (lower middle panel). In this conformation, the anionic substrate glutamate enters the binding site through the positively charged cavity (lower panels) and binds to R88 and R322, driving reorientation back to the lumen-open conformation (upper left panel). Rearrangement of R322 to interact with Y223 accompanies glutamate release into the SV lumen. The balance between lumenal and cytoplasmic gates controls conformational dynamics and hence transport activity. In WT VGLUT, the cytoplasmic gate stabilizes the lumen-open conformation. In this case, lumenal protonation lowers the energy barrier for closure of the lumenal gate, enabling reorientation to the cytoplasm and alternating access. However, weakening of the cytoplasmic gate (for example, by the D371N mutation) shifts the equilibrium to favor the cytoplasm-open conformation, and reorientation to the lumen becomes rate-limiting. In this case, lumenal deprotonation favors opening of the lumenal gate, allowing completion of the transport cycle. The cytoplasmic gate thus tunes allosteric regulation from the lumen, enabling pH to coordinate glutamate transport with the SV cycle. In addition, weakening of the gate by D371N eliminates the VGLUT-associated Cl− conductance, suggesting that the mutation destabilizes a transient intermediate required for the conductance.

Full size image

Since their identification, the VGLUTs have been recognized to undergo potent allosteric regulation2,22,26, but the mechanisms have remained unclear. Lumenal protons activate both glutamate transport and the associated Cl− conductance20,32, restricting the activity of the VGLUTs to SVs and preventing nonquantal release. Our data suggest a mechanism for activation by lumenal pH that depends on tuning of the strength of lumenal and cytoplasmic gates. The D371N mutation weakens the cytoplasmic gate between the N-domain and the C-domain and reverses the lumenal pH dependence of glutamate transport by VGLUT2. The midpoint of pH activation does not differ from WT, suggesting involvement of the same titratable residue. Lumenal H+ binding inhibits transport by the D371N mutant but activates WT VGLUT2. This supports an allosteric role for H+ rather than the coupling of glutamate flux to H+ that has recently been proposed28,31.

We questioned how a cytoplasmic mutation can affect the role of lumenal protonation. At lumenal pH ~7, cytoplasmic gate closure stabilizes the outward-open orientation of WT VGLUTs, and we reason that protonation promotes closure of a lumenal gate required for reorientation to the cytoplasm. D371N weakens the cytoplasmic gate, thereby promoting closure of the lumenal gate and eliminating the requirement for activation by lumenal H+. Indeed, activation of D371N at higher pH suggests that deprotonation weakens the lumenal gate, thereby restoring the balance of strength between the two gates required for alternating access. Therefore, D371N and WT VGLUT2 both undergo protonation in the lumen, but the strength of the cytoplasmic gate determines the overall effect on transport. In the bacterial permease LacY, the mutation of periplasmic residues rescues cytoplasmic motif A mutations40, supporting a role for the balance between opposing gates, although not in the context of regulation. The allosteric regulation of VGLUTs by Cl− may involve a related mechanism. Chloride binds to a similar cytoplasmic interaction network in the cytoplasm-open conformation of two other MFS transporters (the glucose transporter GLUT1 and the plant sugar transporter STP10) but not in their outward-open conformation with cytoplasmic gates closed42,43. In the VGLUTs, cytoplasmic Cl− may thus be required to weaken the cytoplasmic gate and hence promote the cytoplasm-open orientation required for substrate binding. The associated Cl− conductance was abolished by the D371N mutation, showing that it requires a conformation disfavored by weakening of the cytosolic gate. The two functions of the VGLUTs are thus separable: the alternating access required for glutamate transport depends on the appropriate balance between cytoplasmic and lumenal gates, whereas the Cl− conductance involves a specific intermediate, indicating the potential for differential regulation of these two activities in the course of SV filling. The insight into mechanism provided by structure will enable future investigation into the role of these properties in neurotransmission.

Methods
Molecular biology

The R184Q mutation was introduced using Q5 Polymerase (New England Biolabs), sense primer 5′-GACTGGCTGCTAACCGGGTCTTTGGGGC-3′ and antisense primer 5′-GCCCCAAAGACCCGGTTAGCAGCCAGTC-3′. The E191Q mutation was introduced using sense primer 5′-GCAAGGACTTGTGCAGGGCGTCACCTACCC-3′ and antisense primer 5′-GGGTAGGTGACGCCCTGCACAAGTCCTTGC-3′. All fragments were inserted into the PpuMI and XhoI sites of pFastBac Twin-Strep-Tag VGLUT2 and pIRES2-EGFP pmVGLUT2. Sequencing confirmed the mutations and no other changes.

The D371N mutation was introduced using sense primer 5′-GGGCAAATTGCAAACTTTCTAAGGAGCAAG-3′ and antisense primer 5′-GCTCCTTAGAAAGTTTGCAATTTGCCCCCC-3′. The R213S mutation was introduced using sense primer 5′-TTGGAGAGGAGTAGTTTGGCTACCACC-3′ and antisense primer 5′-TGGTAGCCAAACTACTCCTCTCCAAAG-3′.

Other mutations in VGLUT2 were introduced by PCR amplification (using PfuUltraII Fusion High Fidelity Polymerase; Agilent) of two fragments surrounding the site of mutation and subcloned into rat VGLUT2-IRES-GFP using NheI and XhoI and ligation-independent cloning. The outside primers were VG2 FWD: TGAACCGTCAGATCCGCTAGCCGCCA and VG2 REV: AATTCGAAGCTTGAGCTCGAGTTATGA. Inside primers used to produce the mutations were:

R322A FWD: GCCAGTTGGACTTTTTATTTACTGCTC

R322A REV: ATAAAAAGTCCAACTGGCGCAGAAGT

R322K FWD: AAGAGTTGGACTTTTTATTTACTGCTC

R322K REV: ATAAAAAGTCCAACTCTTGCAGAAGTT

S323T FWD: ACCTGGACTTTTTATTTACTGCTCATC

S323T REV: TAAATAAAAAGTCCAGGTCCTGCAGA

F326A FWD: GCTTATTTACTGCTCATCAGTCAACCA

F326A REV: GATGAGCAGTAAATAAGCAGTCCAAC

Y327F FWD: TTCTTACTGCTCATCAGTCAACCAGCT

Y327F REV: ACTGATGAGCAGTAAGAAAAAAGTCC

Y327S FWD: AGGAGTTGGACTTTTTCCTTACTGCTCATC AGTCAACCAGCTTATTTC

Y327S REV: AAAAGTCCAACTCCTGCAGAAGTT

Y327T FWD: AGGAGTTGGACTTTTACCTTACTGCTCATCAGTCAACCAGCTTATTTC

Y327T REV: AAAAGTCCAACTCCTGCAGAAGTT

The GFP–VGLUT2 construct was produced by PCR amplification with the following primers to introduce a HindIII at the 5′ end and EcoRI at the 3′, with a GGS linker between GFP and the N terminus of VGLUT2.

GFP-HindIII-F: CTTTTTTGTTAGACGAAGCTTGCCACCATGGTGAGCAAGGGCGAGGAGCTGTTC where the italicized nucleotides indicate the HindIII site and the underlined the Kozak sequence

GFP-GGS linker-R1: GTTTTACCGACTCCATTGAGCCTCCCTTGTACAGCTCGTCCAT where the underlined sequence encodes GGS

VGLUT2-F1: ATGGAGTCGGTAAAACAAAGGATTTTGGC

VGLUT2-EcoRI-R: CGATAAGCTTGATATCGAATTCTTATGAATAATCATCTCGGTCCTTATAG

The PCR products were subcloned into pFUGW (Addgene, 14883) at HindIII and EcoRI sites by ligation-independent cloning.

The palmitoylation mutations (that is, 3CS) were introduced into pFUGW-VGLUT2-pH by PCR mutagenesis using forward primer 5′-AGAAGAAGGCTCCGCTATGCGACTGTACGTGCTTCGGCCTGCCGCGCCGC-3′ and reverse primer 5′-GCGGCGCGGCAGGCCGAAGCTCGTACTGTCGCTTAGCGGAGCCTTCTTCT-3′, followed by insertion at the EcoRI and BamHI sites of pFUGW-VGLUT2-pH.

The 3CS mutations were introduced into GFP–VGLUT2 by PCR amplification using the GFP-HindIII-F and VGLUT2-EcoRI-R primers above and

3CS-F: TCCGACTCTACGTCCTTCGGCCTGCCGCGCCGCTACATCATAGC

3CS-R1: GACGTAGAGTCGGATAGCGGAGCCTTCTTCTCAGGCAC

The products were subcloned into the HindIII and EcoRI sites of FUGW as above.

Antibodies

A monoclonal antibody to VGLUT2 was obtained from Synaptic Systems (135411) and used at a dilution of 1:2,000 for western blot analysis. The secondary antibody was a goat anti-mouse antibody (IRDye800) obtained from Rockland Immunochemicals and used at 1:10,000. Western blotting was performed by electrophoretic separation through 4–20% precast SDS-acrylamide gels (Bio-Rad), electrotransfer to nitrocellulose and blocked with 10% nonfat dry milk before incubation with primary and secondary antibodies in Tris-buffered saline containing 0.1% Tween-20 and fluorescent detection (Odyssey, v.2.0.3; LI-COR Biosciences).

Protein expression and purification

WT and R184Q/E191Q VGLUT2 were expressed in Spodoptera frugiperda (Sf9) cells (Expression System, 94-0015). High-titer recombinant baculovirus (>109 viral particles per ml) was obtained using the Bac-to-Bac Baculovirus Expression System (Thermo Fisher Scientific) as previously described34. Sf9 cells at a cell density of (2–3) × 106 cells per ml were infected with P1 virus at a multiplicity of infection of three. Cells were collected by centrifugation 48 h after infection and stored at −80 °C until use. The frozen cell pellets were gently sonicated in lysis buffer containing 50 mM Tris (pH 7.4), 400 mM NaCl, 5 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, DNase and EDTA-free protease inhibitor cocktail tablets (Sigma-Aldrich). Then, 5 M NaCl stock solution was added to the broken cells to a final concentration of 1 M, and the cells were washed by repeated homogenizing with a Dounce homogenizer. Membranes were then sedimented at 200,000g and resuspended in buffer A (50 mM Tris, 400 mM NaCl, 10% glycerol). Subsequently, 1% (w/v) n-dodecyl-β-d-maltopyranoside (DDM; Anatrace) and 0.2% (w/v) cholesteryl hemisuccinate (CHS; Sigma-Aldrich) were added to solubilize the membranes at a detergent-to-protein ratio of 0.22 (w/w) by stirring at 4 °C for 4 h. WT protein with a 10× His tag at the C terminus was purified as previously described34. R184Q/E191Q VGLUT2 with a Twin-Strep-Tag at the N terminus was purified as previously described for reconstitution20, except that DDM/CHS was used for solubilization and purification at the same concentrations as for WT VGLUT2. Purified proteins were concentrated with a 100 kDa Amicon concentrator (Millipore) to 1–2 mg ml−1 and frozen at −80 °C in aliquots of ~25–50 μl until used to form a complex with the Fab. VGLUT2 proteins were incubated with the ligand in buffer for 30 min before adding the Fab (8E11 or 9C6) to avoid interference of the Fab with conformational change. Purified Fab was then added to the mixture and incubated on ice for another 30 min before purifying the complex on a Superdex 200 Increase column (GE Healthcare) in size exclusion chromatography (SEC) buffer (20 mM HEPES, 150 mM NaCl, pH 7.4, adjusted with NaOH and 0.01% glycol-diosgenin) with additional ligand. For the R184Q/E191Q-ACPD complex, 30 mM ACPD in SEC buffer was used for the incubation and 1 mM ACPD was added to the SEC running buffer. For WT VGLUT2, 2.77 mM clodronate was used during incubation, and 263 µM clodronate was added to the SEC running buffer, given that clodronate has been suggested to compete with the cytosolic Cl− that activates the vesicular nucleotide transporter44. However, no additional density for clodronate was observed in the structure. We therefore denote the structure as WT apo.

Cryo-EM sample preparation and data acquisition

Freshly purified VGLUT2–Fab complex (3 μl) was concentrated to 0.5–1 mg ml−1 and applied to Quantifoil holey carbon grids, glow-discharged with an EMS 700 glow discharge device (at the University of California San Francisco (UCSF) EM Core) using mixed air at 10–20 mA for 30 s. Copper 1.2/1.3–400 mesh grids were blotted with grade 595 standard Vitrobot filter paper for 4–7 s at 10 °C and 100% humidity using a Vitrobot Mark IV, followed by rapid plunging into liquid ethane cooled by liquid nitrogen. The dataset for R184Q/E191Q-ACPD was collected at UCSF, and the dataset for WT VGLUT2 was collected through S2C2 at SLAC with the parameters shown in Table 1.

Image processing

Micrographs were corrected for beam-induced drift using MotionCor2 (ref. 45). The contrast transfer function parameters for each micrograph were determined using CTFFIND4 (ref. 46), and micrographs of poor quality (contrast transfer function lower than 6 Å) were discarded. All subsequent processing was carried out in RELION 3.0 (ref. 47) and cryoSPARC48. Similar strategies were used for WT and R184Q/E191Q VGLUT2 data as illustrated in Extended Data Figs. 4 and 6. In brief, particles were picked from good micrographs using templates previously generated. Two to three rounds of 2D classification with 4× binned particles were used to remove junk particles. The particle stack was then re-extracted as 2× binned particles and aligned to a single class in RELION. Six parallel 3D classifications were then initialized from parameters at iterations 20–25 from the previous 3D classification. For the R184Q/E191Q dataset, good, unique particles from each run were selected, combined and subjected to another round of multi-reference 3D classification. For the WT dataset, two classes that resulted in the highest resolution reconstructions were selected and imported into cryoSPARC. Each class was further classified into two classes using the ab initio reconstruction scheme. Particles classes with high-resolution features were then selected and refined in RELION using gold-standard 3D auto-refine and Bayesian polishing strategy49. Polished particles were then auto-refined in RELION with a soft mask applied to the protein region, including VGLUT2 and the Fab. The same polished particle stack was also imported into CryoSPARC and refined in parallel using the non-uniform refinement strategy. Final maps obtained in both RELION and CryoSPARC achieved similar resolution and quality with minor variations in the flexible loop 1 region. Maps generated in RELION generally resolved more density in this region, as it is included in the manually generated mask. Statistics of the final maps generated were calculated using the Fourier shell correlation criterion and a threshold of 0.143 using Phenix_comprehensive_validation tool50 and are reported in Table 1.

Model building and structure refinement

Final maps generated by RELION were sharpened with Phenix Auto-sharpening before being used for model building and refinement in Phenix. Sharpened maps generated in cryoSPARC were used directly. The structure of WT VGLUT2 in high Cl− (150 mM) (PDB 6V4D) was used as the initial model for building in Coot51. The local areas of the built models were first refined within Coot using real-space refinement with Ramachandran and geometry constraints. The full models of WT and R184Q/E191Q VGLUT2 were then refined using Phenix.real_space_refine without imposed constraints51. The density corresponding to loop one between TM1 and TM2 is of much lower quality. As a result, only the main chain was built in this area. Figures were prepared using Pymol and Chimera52.

Docking

The Small Molecule Drug Discovery Suite (release 2017-4; Schrödinger) was used for all docking calculations. For docking to the WT protein, the previously published high Cl− structure (PDB 6V4D) was used alongside the R184Q/E191Q structure reported here. Both proteins were prepared for docking, with protonation states assigned at pH 7.4 ± 0.1 and missing sidechains and loops modeled using the Protein Preparation Wizard workflow. ACPD coordinates were extracted from the R184Q/E191Q map and saved as a separate PDB, then protonation state and tautomers were assigned at pH 7.4 ± 0.1 using Epik in Maestro LigPrep with the OPLS3 forcefield. Glutamate was drawn with Maestro (v.11.4), followed by identical processing with Maestro LigPrep. For docking ACPD into R184Q/E191Q with a flexible mode, a 20 × 20 × 20 Å3 docking grid was centered around the putative binding site. The flexible mode allowed bond rotation and ring conformation sampling while penalizing non-planar amide groups. To dock glutamate into WT VGLUT2, a docking grid was centered around the same putative binding site. In all instances, the protein was held rigid while the Glide XP scoring function was used, and a maximum of ten optimal poses were returned. Representative poses were picked based on a combination of visual inspection, chemical intuition and score.

Molecular dynamics simulations

To compare the stability of each ligand in the binding pocket, three simulation systems were built: R184Q/E191Q VGLUT2 and ACPD, WT VGLUT2 and glutamate, and WT VGLUT2 with aspartate. Initial atomic coordinates for R184Q/E191Q VGLUT2 and ACPD were taken directly from the cryo-EM structure (PDB 6V4D), while the glutamate and WT protein coordinates were taken from the Glide docking outputs. The aspartate coordinates were generated using Amber18 tleap as described below (‘Alchemical free energy calculations’) and placed inside the same WT protein structure from Glide. Additionally, a fourth simulation was started with no ligand to probe the effect of D371N on chloride; the updated WT structure (PDB 8SBE) was used as a starting point, and the D371N mutation was done using the CHARMM-GUI bilayer builder module. The steps described below apply to all systems built with CHARMM-GUI. The bilayer builder module was used to place each protein in a 1:1 DOPC membrane patch with 30 Å of water above and below and 0.05 mM KCl in solution. The final systems had ~185 DOPC lipids, ~20,000 water molecules and initial dimensions of 90 × 90 × 120 Å3. Simulations were run using the ff14sb, Lipid17, GAFF and TIP3P force fields in the Amber18 engine. Equilibration was done by lowering the restraint stepwise from 5.0 to 0.05 kcal mol−1 Å−2 over 3 ns for the protein backbone, 1 ns for sidechains and 100 ps for lipid headgroups and heavy water atoms. Next, a voltage of −80 mV was applied to the system using eight 10 mV steps of 5 ns each. Finally, unrestrained production simulations were run for 1,048 ns (WT and Glu), 848 ns (R184Q/E191Q and ACPD) and 819 ns (WT and aspartate). All equilibration simulations used the Berendsen barostat with semi-isotropic pressure coupling and a target pressure of 1 atm in the NPT ensemble, whereas voltage ramping and subsequent production simulations were run in the NVT ensemble. All simulations used a hard cutoff of 10 Å for Lennard–Jones forces and a 2 fs timestep; long-range electrostatic interactions were treated using the Particle Mesh Ewald method.

Alchemical free energy calculations

Relative binding free energies were calculated using the Thermodynamic Integration feature from Amber18.17. Two systems were set up to convert from ACPD into glutamate and from glutamate into aspartate. The R184Q/E191Q protein, bilayer, water and ions generated by CHARMM-GUI were used as the starting point for all Thermodynamic Integration simulations. To generate the starting glutamate and aspartate coordinates, the ‘common atoms’ from ACPD were copied and given an updated residue name (GLU or ASP), then missing heavy atoms were generated using Amber18 tleap. Equilibration was carried out after 10,000 steps of minimization by lowering the restraint stepwise over 300 ps from 5.0 to 2.5 kcal mol−1 Å−2 for the protein backbone, 5.0 to 0.0 kcal mol−1 Å−2 for sidechains and heavy water atoms and held at 5.0 kcal mol−1 Å−2 for lipid headgroups and ligand. Single-step transformations were done using pmemd.cuda.MPI, with 11 evenly spaced lambda windows of 15 ns each. Default values of alpha (0.5) and beta (12A2) were used.

Proteoliposome uptake assay

VGLUT reconstitution and glutamate uptake assay were performed as previously described20. In brief, purified WT and R184Q/E191Q rVGLUT2 were reconstituted into proteoliposomes together with a His-tagged ATP synthase holoenzyme (TF0F1) from the thermophilic Bacillus sp. PS3 with the following protein to lipid ratios (mol/mol): >1:40,000 for TF0F1 and 1:2,000 for VGLUT2. After dialysis overnight against 150 mM KCl, 5 mM MgCl2 and 2 mM HEPES (pH 7.3), the buffer was replaced with the same fresh buffer, and dialysis continued for an additional 2 h. To remove residual detergent, the sample was incubated with 100 mg ml−1 Bio-Beads SM-2 resin (Bio-Rad) for 2 h and used immediately for the uptake assay. To measure glutamate uptake driven by a proton electrochemical gradient, 25 µl proteoliposomes containing the TF0F1 ATP synthase and 2–3.5 µg VGLUT2 were diluted into 775 µl uptake buffer (150 mM K gluconate, 5 mM HEPES-KOH pH 7.3, 3 mM Mg gluconate, 0.2 mM glutamate, 10 µCi ml−1 L-[3,4-3H]-glutamic acid (PerkinElmer) and 2 mM Mg-ATP) for a final Cl− concentration of 5 mM. The reactions were incubated at 30 °C for 15 min and stopped by rapid filtration through 0.45 µm HAWP discs (Millipore). The filters were washed three times with ice-cold buffer (150 mM K gluconate, 5 mM HEPES-KOH pH 7.3 and 3 mM Mg gluconate), dried, solubilized in CytoScint (MP Biomedicals) and the bound radioactivity measured using an LS 6000SC scintillation counter (Beckman).

In-solution digestion and mass spectrometry analysis

Aliquots (5 µl) containing 8.4 µg purified R184Q/E191Q VGLUT2 in 150 mM NaCl, 20 mM Tris pH 7.4 and 0.003% lauryl maltose neopentyl glycol buffer were mixed with 15 µl 0.2% Rapigest (Waters) in 100 mM triethylammonium bicarbonate buffer at pH 8. Then, 1 µl 100 mM tris-(2-carboxyethyl) phosphine was added to a final concentration of 4.76 mM and incubated at 56 °C for 30 min. This was followed by the addition of 0.9 µl 500 mM iodoacetamide and a 30-min incubation at room temperature (22–25 oC) in the dark. For digestion, 1 µg of the proteolytic enzymes, either trypsin (Pierce, MS-grade, Thermo Scientific), chymotrypsin (Sequencing grade, Roche) or GluC (Sequencing grade, Promega), was then then added to the samples, resulting in a 25 µl final digestion volume, and the samples were incubated for 12 h at 37 °C. After that, another 1 µg aliquot of enzyme was added, and the digestion continued for an additional 4 h. After digestion, the samples were acidified with formic acid to a final concentration of 10% and incubated at 37 °C for 30 min. The digests were then desalted using C18 ZipTip tips (Millipore) following the manufacturer’s instructions and eluted in two consecutive 7 µl aliquots of 50% acetonitrile/0.1% formic, followed by a third elution in 7 µl 0.1% formic in acetonitrile. Combined eluates were dry-evaporated and resuspended in 0.1% formic acid at 0.4 µg µl−1.

For mass spectrometry analysis, samples (1 µg digest) were run in a 3 µm, 75 µm i.d. × 15 cm PepMap RSLC C18 EasySpray column (Thermo Scientific). Acetonitrile gradients were used to separate peptides at a flow rate of 300 nl min−1 as follows: 2–5% B for 3 min, 5–30% B for 62 min, 30–50% B for 5 min and 50–95% B for 5 min, where A is 0.1% formic acid in water and B is 0.1% formic acid in acetonitrile. Mass spectrometry analysis was performed using an Orbitrap Lumos Fusion (Thermo Scientific) in positive ion mode. MS spectra were acquired between 375 and 1,500 m/z with a resolution of 120,000. For each MS spectrum, multiply charged ions over the selected threshold (2 × 104) were selected for tandem mass spectrometry in cycles of 3 s, with an isolation window of 1.6 Th. Precursor ions were fragmented by HCD using a normalized collision energy of 30. Tandem mass spectrometry spectra were acquired in centroid mode with a resolution of 30,000 from m/z = 110. A dynamic exclusion window was applied, which prevented the same m/z from selection for 30 s after its acquisition. Each sample was reanalyzed in a second liquid chromatography–mass spectrometry injection using EThcD as the fragmentation method. In these runs, we used the same parameters as described above, except that the nine most intense multiply charged ions over the selected threshold were selected for HCD (normalized collision energy as before) and then EThcD (supplemental collision energy 25).

Peak lists were generated using PAVA in-house software53. All peak lists were searched against the rat subset of the SwissProt database (SwissProt.2019.07.31), using Protein Prospector54 with the following parameters: enzyme specificity was set as trypsin, chymotrypsin or GluC, as appropriate for the particular sample, and up to two missed cleavages per peptide were allowed. Carbamidomethylation of cysteine residues, N-acetylation of the N terminus of the protein, loss of protein N-terminal methionine, pyroglutamate formation from peptide N-terminal glutamines, oxidation of methionine and palmitoylation of cysteine residues were allowed as variable modifications, and three variable modifications were allowed per peptide. Mass tolerance was 10 ppm in mass spectrometry and 30 ppm in tandem mass spectrometry. The false positive rate was estimated by searching the data using a concatenated database that contains the original SwissProt database, as well as a version of each original entry in which the sequence has been randomized. A 1% false discovery rate was permitted at the protein and peptide level. Spectra for peptides containing palmitoylated cysteine residues were manually inspected.

The mass spectrometry data have been deposited in the ProteomeXchange Consortium through the PRIDE partner repository, with dataset identifier PXD061514.

Live cell imaging and analysis

All animal experiments were approved by the Institutional Animal Care and Use Committee at UCSF and performed as required by the National Institutes of Health (NIH) Guide for Care and Use of Laboratory Animals. C57Bl/6 mice were maintained on a 12 h day, 12 h night cycle at 25 °C and 40–60% humidity. Dissociated hippocampal neurons from postnatal day 0–1 mice were prepared and grown as previously described55. On day-in-vitro six, the cells were transduced with a lentivirus encoding WT or C60S/C62S/C64S (3CS) VGLUT2-pHluorin. At day-in-vitro 13–15, the cells were imaged using an inverted Nikon TE300 fluorescence microscope in a laminar flow perfusion and stimulation chamber. Images were collected by epifluorescence microscopy using the following bandpass filters: 470/40 nm excitation and 525/50 nm emission. Images were acquired at 1 Hz with 300 ms exposure time. At frame 15, action potentials (10 Hz for 30 s) were evoked by passing 1 ms bipolar current pulses through platinum–iridium electrodes, to yield fields of 5–10 V cm−1. Cells were continuously perfused with Tyrode’s buffer (119 mM NaCl, 25 mM HEPES, 2 mM CaCl2, 2 mM MgCl2, 2.5 mM KCl and 30 mM glucose at pH 7.4) containing 10 μM 6-cyano-7-nitroquinoxaline-2,3-dione and 10 μM 3-(2-carboxypiperazin-4-yl)propyl-1-phosphonic acid at 35 °C. To image the pHluorin response to stimulation, cells at frame 15 were stimulated at 10 Hz for 30 s. After 140 s, the total unquenched pHluorin signal was revealed by applying Tyrode’s buffer with 50 mM NH4Cl substituted for 69 mM NaCl. To image GFP dispersion and reclustering, cells were stimulated at 5 Hz and 25 Hz for 20 s.

The images were analyzed using ImageJ56. For each coverslip imaged, 35–80 synapses were manually selected using the fluorescence images obtained after NH4Cl application with a round region of interest and a diameter of seven pixels for pHluorin and three pixels for GFP. For each time point, the region of interest mean fluorescence intensity was recorded and background subtracted for pHluorin and GFP. pHluorin data were then represented as ΔF/ΔFNH4Cl, where ΔF is the mean fluorescence intensity for a given time point minus the mean fluorescence before stimulation, and ΔFNH4Cl is the peak fluorescence after NH4Cl addition, also subtracting the mean fluorescence intensity before stimulation. The normalized data were visualized using GraphPad Prism.

Glutamate efflux

Glutamate efflux experiments were performed as previously described32 with a few modifications. HEK293T cells were plated into 24-well plates precoated with poly-L-lysine (0.1 mg ml−1), and cells were transfected the following day with 395 ng WT and mutant pmVGLUT2 DNA in pIRES2-EGFP together with 1.19 µl FUGENE-HD per well. The media was replaced 1 day later with fresh growth media. Glutamate flux was measured 2 days after transfection. First, the cells were washed twice in Ringer’s solution, pH 7.4 (145 mM NaCl, 10 mM glucose, 10 mM HEPES (pH 7.4), 4 mM KCl, 2 mM CaCl2, 1 mM MgCl2) before loading with 50 μM 3H-glutamate (2 µCi) in 200 µl Ringer’s solution for 10 min at 34 °C. The cells were then washed three times in ice-cold Ringer’s solution with 0.5 mM aspartate. Glutamate efflux was monitored in 200 µl warm Ringer’s solution or Ringer’s solution at pH 4.5–8.0 (145 mM NaCl, 10 mM glucose, 4 mM KCl, 2 mM CaCl2, 1 mM MgCl2 with 10 mM HEPES for pH 7.0–8.0 or 10 mM MES for pH 4.5–6.5) with 0.5 mM aspartate for 5 min at 34 °C. A total of 150 µl of the medium was collected, and the radioactivity was measured by scintillation counting. Residual Ringer’s solution was aspirated, the cells lysed by the addition of 400 µl 1% SDS, and the radioactivity measured by scintillation counting. The percent glutamate efflux was calculated as follows:

$$\% \,{\rm{Glutamate}}\, {\rm{efflux}}=\frac{{\rm{CPM}}\left({\rm{efflux}}\, {\rm{buffer}}\right)\times 1.33}{{\rm{CPM}}\left({\rm{efflux}}\, {\rm{buffer}}\right)\times 1.33+{\rm{CPM}}\,({\rm{cell}}\, {\rm{lysate}})}\times 100,$$

where CPM is the 3H counts per minute for the indicated samples. The data were plotted and analyzed in GraphPad Prism. Each measurement was made in duplicate, and experiments were performed three times independently.

Cell surface expression of VGLUT2

HEK293T cells were plated into 24-well plates precoated with poly-L-lysine (1 mg ml−1), transfected the next day with 0.4 µg of pm WT, R322A, R322K, S323T, F326A, Y327F, Y327S, Y327T, D97A/E122A/E338A/E339A/E344A (5M) VGLUT2 or EV and 1.2 µl FUGENE-HD per well, and the media replaced 1 day later. Then, 2 days after transfection, the cells were washed twice with PBS, once with 150 mM NaCl, 1 mM MgCl2, 2 mM CaCl2 and then incubated with 150 mM NaCl, 1 mM MgCl2, 2 mM CaCl2, and 1 mg ml−1 NHS-LC-biotin for 60–90 min at 4 °C to label VGLUT2 expressed on the plasma membrane, the fraction of VGLUT2 active in the efflux assay. To stop the biotinylation, the cells were washed once with quenching buffer containing 100 mM glycine, 1 mM MgCl2, 0.1 mM CaCl2, and then incubated with the same buffer for 10 min at 4 °C. The cells were then washed twice with calcium-magnesium-free PBS, lysed with 150 mM NaCl, 50 mM Tris-HCl (pH 7.4), 5 mM EDTA, 1% Triton X-100, 0.5% deoxycholate, 0.1% SDS (RIPA buffer) containing ThermoScientific Pierce EDTA-Free protease inhibitors and the Roche PhosSTOP EasyPack phosphatase inhibitors, and maintained at 4 °C for 30 min. The cell lysates were then sedimented at 21,000g for 10 min at 4 °C, the supernatant collected and 20 µl mixed with 20 µl 2× Laemmli buffer and stored at −80 °C as an input control. Then, 20 µl streptavidin-magnetic beads washed and resuspended in RIPA buffer were added to the cell lysates, and the mixture was mixed by inversion for 2 h at 4 °C. The beads were magnetically separated, washed twice in RIPA buffer, resuspended in 30 µl 2× Laemmli buffer, sedimented at 1,000g for 15 s and the streptavidin–biotin complex denatured by heating the resuspended beads at 65 °C for 5 min. The input control and biotinylated fraction were then separated by electrophoresis through SDS-polyacrylamide and immunoblotted for VGLUT2 with fluorescence detection (Odyssey; LI-COR).

Oocyte recordings

Recordings were performed as previously described20. In brief, Xenopus laevis oocytes were obtained from Ecocyte, injected with 50 ng pmVGLUT2-HA, R213S pmVGLUT2-HA or D371N pmVGLUT2-HA cRNA and incubated in ND96 (96 mM NaCl, 2 mM KCl, 1.8 mM CaCl2, 1 mM MgCl2, 5 mM HEPES (pH 7.4)) with 50 μg ml−1 tetracycline and gentamicin at 16 °C for 5 days. The oocyte currents were recorded by standard two-electrode voltage clamp in Ca2+-free ND96 and ND96, pH 5.5 (96 mM NaCl, 2 mM KCl, 1 mM MgCl2, 5 mM MES (pH 5.5)). Steady-state current–voltage relations were obtained using a protocol of 300 ms voltage steps from −120 to +60 in 10 mV steps from a holding potential of −30 mV. Data were analyzed using Clampfit and GraphPad Prism.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Data availability

The atomic coordinates of rVGLUT2 have been deposited in the Protein Data Bank (PDB) under accession codes PDB 7T3N for the R184Q/E191Q structure and PDB 7T3O for the WT structure. The corresponding maps have been deposited in the Electron Microscopy Data Bank (EMDB) under accession codes EMD-25665 and EMD-25666, respectively. The mass spectrometry data, including search results and peptide identification, are available at the ProteomeXchange Consortium through the PRIDE partner repository, with dataset identifier PXD061514. All data supporting the findings of this study are available from the corresponding authors on reasonable request. Source data are provided with this paper.

References

Katz, B. Quantal mechanism of neural transmitter release. Science 173, 123–126 (1971).

Article
 
CAS
 
PubMed
 
Google Scholar

Naito, S. & Ueda, T. Characterization of glutamate uptake into synaptic vesicles. J. Neurochem. 44, 99–109 (1985).

Article
 
CAS
 
PubMed
 
Google Scholar

Maycox, P. R., Deckwerth, T., Hell, J. W. & Jahn, R. Glutamate uptake by brain synaptic vesicles. Energy dependence of transport and functional reconstitution in proteoliposomes. J. Biol. Chem. 263, 15423–15428 (1988).

Article
 
CAS
 
PubMed
 
Google Scholar

Bellocchio, E. E., Reimer, R. J., Fremeau, R. T. & Edwards, R. H. Uptake of glutamate into synaptic vesicles by an inorganic phosphate transporter. Science 289, 957–960 (2000).

Article
 
CAS
 
PubMed
 
Google Scholar

Takamori, S., Rhee, J. S., Rosenmund, C. & Jahn, R. Identification of a vesicular glutamate transporter that defines a glutamatergic phenotype in neurons. Nature 407, 189–194 (2000).

Article
 
CAS
 
PubMed
 
Google Scholar

Fremeau, R. T. et al. The expression of vesicular glutamate transporters defines two classes of excitatory synapse. Neuron 31, 247–260 (2001).

Article
 
CAS
 
PubMed
 
Google Scholar

Gras, C. et al. A third vesicular glutamate transporter expressed by cholinergic and serotoninergic neurons. J. Neurosci. 22, 5442–5451 (2002).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Fremeau, R. T. et al. Vesicular glutamate transporters 1 and 2 target to functionally distinct synaptic release sites. Science 304, 1815–1819 (2004).

Article
 
CAS
 
PubMed
 
Google Scholar

Wojcik, S. M. et al. An essential role for vesicular glutamate transporter 1 (VGLUT1) in postnatal development and control of quantal size. Proc. Natl Acad. Sci. USA 101, 7158–7163 (2004).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Li, F., Eriksen, J., Finer-Moore, J., Stroud, R. M. & Edwards, R. H. Diversity of function and mechanism in a family of organic anion transporters. Curr. Opin. Struct. Biol. 75, 102399 (2022).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Drew, D. & Boudker, O. Ion and lipid orchestration of secondary active transport. Nature 626, 963–974 (2024).

Article
 
CAS
 
PubMed
 
Google Scholar

Morin, P., Sagné, C. & Gasnier, B. Functional characterization of wild-type and mutant human sialin. EMBO J. 23, 4560–4570 (2004).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Wreden, C. C., Wlizla, M. & Reimer, R. J. Varied mechanisms underlie the free sialic acid storage disorders. J. Biol. Chem. 280, 1408–1416 (2005).

Article
 
CAS
 
PubMed
 
Google Scholar

Leano, J. B. et al. Structures suggest a mechanism for energy coupling by a family of organic anion transporters. PLoS Biol. 17, e3000260 (2019).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Shioi, J. & Ueda, T. Artificially imposed electrical potentials drive L-glutamate uptake into synaptic vesicles of bovine cerebral cortex. Biochem. J. 267, 63–68 (1990).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Ishikawa, T., Sahara, Y. & Takahashi, T. A single packet of transmitter does not saturate postsynaptic glutamate receptors. Neuron 34, 613–621 (2002).

Article
 
CAS
 
PubMed
 
Google Scholar

Omote, H., Miyaji, T., Juge, N. & Moriyama, Y. Vesicular neurotransmitter transporter: bioenergetics and regulation of glutamate transport. Biochemistry 50, 5558–5565 (2011).

Article
 
CAS
 
PubMed
 
Google Scholar

Cavelier, P. & Attwell, D. Neurotransmitter depletion by bafilomycin is promoted by vesicle turnover. Neurosci. Lett. 412, 95–100 (2007).

Article
 
CAS
 
PubMed
 
Google Scholar

Takami, C., Eguchi, K., Hori, T. & Takahashi, T. Impact of vesicular glutamate leakage on synaptic transmission at the calyx of Held. J. Physiol. 595, 1263–1271 (2017).

Article
 
CAS
 
PubMed
 
Google Scholar

Eriksen, J. et al. Protons regulate vesicular glutamate transporters through an allosteric mechanism. Neuron 90, 768–780 (2016).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Tabb, J. S., Kish, P. E., Van Dyke, R. & Ueda, T. Glutamate transport into synaptic vesicles. Roles of membrane potential, pH gradient, and intravesicular pH. J. Biol. Chem. 267, 15412–15418 (1992).

Article
 
CAS
 
PubMed
 
Google Scholar

Hartinger, J. & Jahn, R. An anion binding site that regulates the glutamate transporter of synaptic vesicles. J. Biol. Chem. 268, 23122–23127 (1993).

Article
 
CAS
 
PubMed
 
Google Scholar

Wolosker, H., de Souza, D. O. & de Meis, L. Regulation of glutamate transport into synaptic vesicles by chloride and proton gradient. J. Biol. Chem. 271, 11726–11731 (1996).

Article
 
CAS
 
PubMed
 
Google Scholar

Schenck, S., Wojcik, S. M., Brose, N. & Takamori, S. A chloride conductance in VGLUT1 underlies maximal glutamate loading into synaptic vesicles. Nat. Neurosci. 12, 156–162 (2009).

Article
 
CAS
 
PubMed
 
Google Scholar

Chang, R., Eriksen, J. & Edwards, R. H. The dual role of chloride in synaptic vesicle glutamate transport. Elife 7, e34896 (2018).

Article
 
PubMed
 
PubMed Central
 
Google Scholar

Hori, T. & Takahashi, T. Kinetics of synaptic vesicle refilling with neurotransmitter glutamate. Neuron 76, 511–517 (2012).

Article
 
CAS
 
PubMed
 
Google Scholar

Preobraschenski, J., Zander, J. F., Suzuki, T., Ahnert-Hilger, G. & Jahn, R. Vesicular glutamate transporters use flexible anion and cation binding sites for efficient accumulation of neurotransmitter. Neuron 84, 1287–1301 (2014).

Article
 
CAS
 
PubMed
 
Google Scholar

Martineau, M., Guzman, R. E., Fahlke, C. & Klingauf, J. VGLUT1 functions as a glutamate/proton exchanger with chloride channel activity in hippocampal glutamatergic synapses. Nat. Commun. 8, 2279 (2017).

Article
 
PubMed
 
PubMed Central
 
Google Scholar

Herring, B. E., Silm, K., Edwards, R. H. & Nicoll, R. A. Is aspartate an excitatory neurotransmitter? J. Neurosci. 35, 10168–10171 (2015).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Naito, S. & Ueda, T. Adenosine triphosphate-dependent uptake of glutamate into protein I-associated synaptic vesicles. J. Biol. Chem. 258, 696–699 (1983).

Article
 
CAS
 
PubMed
 
Google Scholar

Kolen, B. et al. Vesicular glutamate transporters are H+-anion exchangers that operate at variable stoichiometry. Nat. Commun. 14, 2723 (2023).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Eriksen, J., Li, F., Stroud, R. M. & Edwards, R. H. Allosteric inhibition of a vesicular glutamate transporter by an isoform-specific antibody. Biochemistry 60, 2463–2470 (2021).

Article
 
CAS
 
PubMed
 
Google Scholar

Winter, H. C. & Ueda, T. The glutamate uptake system in presynaptic vesicles: further characterization of structural requirements for inhibitors and substrates. Neurochem. Res. 33, 223–231 (2008).

Article
 
CAS
 
PubMed
 
Google Scholar

Li, F. et al. Ion transport and regulation in a synaptic vesicle glutamate transporter. Science 368, 893–897 (2020).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Fukata, Y., Murakami, T., Yokoi, N. & Fukata, M. Local palmitoylation cycles and specialized membrane domain organization. Curr. Top. Membr. 77, 97–141 (2016).

Article
 
CAS
 
PubMed
 
Google Scholar

Prescott, G. R., Gorleku, O. A., Greaves, J. & Chamberlain, L. H. Palmitoylation of the synaptic vesicle fusion machinery. J. Neurochem. 110, 1135–1149 (2009).

Article
 
CAS
 
PubMed
 
Google Scholar

Collins, M. O., Woodley, K. T. & Choudhary, J. S. Global, site-specific analysis of neuronal protein S-acylation. Sci. Rep. 7, 4683 (2017).

Article
 
PubMed
 
PubMed Central
 
Google Scholar

Miesenböck, G., De Angelis, D. A. & Rothman, J. E. Visualizing secretion and synaptic transmission with pH-sensitive green fluorescent proteins. Nature 394, 192–195 (1998).

Article
 
PubMed
 
Google Scholar

Voglmaier, S. M. et al. Distinct endocytic pathways control the rate and extent of synaptic vesicle protein recycling. Neuron 51, 71–84 (2006).

Article
 
CAS
 
PubMed
 
Google Scholar

Cain, S. M., Matzke, E. A. & Brooker, R. J. The conserved motif in hydrophilic loop 2/3 and loop 8/9 of the lactose permease of Escherichia coli. Analysis of suppressor mutations. J. Membr. Biol. 176, 159–168 (2000).

Article
 
CAS
 
PubMed
 
Google Scholar

Jiang, D. et al. Structure of the YajR transporter suggests a transport mechanism based on the conserved motif A. Proc. Natl Acad. Sci. USA 110, 14664–14669 (2013).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Bavnhoj, L., Paulsen, P. A., Flores-Canales, J. C., Schiott, B. & Pedersen, B. P. Molecular mechanism of sugar transport in plants unveiled by structures of glucose/H+ symporter STP10. Nat. Plants 7, 1409–1419 (2021).

Article
 
PubMed
 
Google Scholar

Custodio, T. F., Paulsen, P. A., Frain, K. M. & Pedersen, B. P. Structural comparison of GLUT1 to GLUT3 reveal transport regulation mechanism in sugar porter family. Life Sci. Alliance 4, e202000858 (2021).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Kato, Y. et al. Identification of a vesicular ATP release inhibitor for the treatment of neuropathic and inflammatory pain. Proc. Natl Acad. Sci. USA 114, E6297–E6305 (2017).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Zheng, S. Q. et al. MotionCor2: anisotropic correction of beam-induced motion for improved cryo-electron microscopy. Nat. Methods 14, 331–332 (2017).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Rohou, A. & Grigorieff, N. CTFFIND4: fast and accurate defocus estimation from electron micrographs. J. Struct. Biol. 192, 216–221 (2015).

Article
 
PubMed
 
PubMed Central
 
Google Scholar

Scheres, S. H. RELION: implementation of a Bayesian approach to cryo-EM structure determination. J. Struct. Biol. 180, 519–530 (2012).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Punjani, A., Rubinstein, J. L., Fleet, D. J. & Brubaker, M. A. cryoSPARC: algorithms for rapid unsupervised cryo-EM structure determination. Nat. Methods 14, 290–296 (2017).

Article
 
CAS
 
PubMed
 
Google Scholar

Zivanov, J., Nakane, T. & Scheres, S. H. W. A Bayesian approach to beam-induced motion correction in cryo-EM single-particle analysis. IUCrJ 6, 5–17 (2019).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Afonine, P. V. et al. Real-space refinement in PHENIX for cryo-EM and crystallography. Acta Crystallogr. D 74, 531–544 (2018).

Article
 
CAS
 
Google Scholar

Emsley, P., Lohkamp, B., Scott, W. G. & Cowtan, K. Features and development of Coot. Acta Crystallogr. D 66, 486–501 (2010).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Pettersen, E. F. et al. UCSF ChimeraX: structure visualization for researchers, educators, and developers. Protein Sci. 30, 70–82 (2021).

Article
 
CAS
 
PubMed
 
Google Scholar

Guan, S., Price, J. C., Prusiner, S. B., Ghaemmaghami, S. & Burlingame, A. L. A data processing pipeline for mammalian proteome dynamics studies using stable isotope metabolic labeling. Mol. Cell Proteomics 10, M111.010728 (2011).

Article
 
PubMed
 
PubMed Central
 
Google Scholar

Clauser, K. R., Baker, P. & Burlingame, A. L. Role of accurate mass measurement (±10 ppm) in protein identification strategies employing MS or MS/MS and database searching. Anal. Chem. 71, 2871–2882 (1999).

Article
 
CAS
 
PubMed
 
Google Scholar

Silm, K. et al. Synaptic vesicle recycling pathway determines neurotransmitter content and release properties. Neuron 102, 786–800.e5 (2019).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Schneider, C. A., Rasband, W. S. & Eliceiri, K. W. NIH Image to ImageJ: 25 years of image analysis. Nat. Methods 9, 671–675 (2012).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Morin, A. et al. Collaboration gets the most out of software. Elife 2, e01456 (2013).

Article
 
PubMed
 
PubMed Central
 
Google Scholar

Bond, C. S. & Schüttelkopf, A. W. ALINE: a WYSIWYG protein-sequence alignment editor for publication-quality alignments. Acta Crystallogr. D 65, 510–512 (2009).

Download references

Acknowledgements

We thank D. Bulkley, Z. Yu, G. Gilbert and M. Harrington at the UCSF cryo-EM facility and C. Hecksel, P. Mitchell, L.-M. Joubert and L. Dunn at Stanford-SLAC Cryo-EM Center (S2C2) for their support in data acquisition and computation. We thank D. Cawley at the Vaccine and Gene Therapy Institute of Oregon Health & Science University for generating the monoclonal antibody and for advice on working with the antibodies. This work was supported by R01NS089713 to R.M.S. and R.H.E., R37MH50712 to R.H.E. and R01GM089740 to M.G. F.L. was supported by postdoctoral fellowships from the American Heart Association (17POST33660928) and National Institute of Mental Health (K99MH119591) and Y.K.G. by the National Science Foundation Graduate Research Fellowship Program (2034836). The UCSF EM facility was supported by NIH grants S10OD020054 and S10OD021741. Some of this work was performed at the Stanford-SLAC Cryo-EM Center (S2C2), which is supported by the NIH Common Fund Transformative High-Resolution Cryo-Electron Microscopy program (U24 GM129541). The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH. Structural biology applications used in this project were compiled and configured by SBGrid57. Mass spectrometry was provided by the Mass Spectrometry Resource at UCSF (A.L. Burlingame, Director), supported by the Dr. Miriam and Sheldon G. Adelson Medical Research Foundation (AMRF) and the UCSF Program for Breakthrough Biomedical Research.

Author information
Author notes

Fei Li

Present address: Department of Structural Biology, Amgen Research, South San Francisco, CA, USA

Jacob Eriksen

Present address: Nine Square Therapeutics, South San Francisco, CA, USA

These authors contributed equally: Fei Li, Jacob Eriksen.

Authors and Affiliations

Department of Biochemistry and Biophysics, UCSF School of Medicine, San Francisco, CA, USA

Fei Li, Janet Finer-Moore, Phuong Nguyen, Alisa Bowen, Andrew Nelson & Robert M. Stroud

Departments of Neurology and Physiology, UCSF School of Medicine, San Francisco, CA, USA

Fei Li, Jacob Eriksen, Hongfei Xu, Surabhi Hareendranath, Poulomi Das & Robert H. Edwards

Department of Pharmaceutical Chemistry, UCSF School of Medicine, San Francisco, CA, USA

Juan A. Oses-Prieto, Yessica K. Gomez, Alma Burlingame, Michael Grabe & Robert M. Stroud

Contributions

F.L., J.E., R.M.S. and R.H.E. conceived the project. P.N., A.N., A. Bowen, J.E. and F.L. expressed the protein. F.L. purified the protein, prepared all samples for EM, acquired cryo-EM data at UCSF and the SLAC cryo-EM facilities and determined the structures. J.E., H.X., S.H., P.D. and F.L. performed the functional experiments. J.A.O.-P. and A. Burlingame performed the mass spectrometry experiments. Y.K.G. and M.G. performed the docking and molecular dynamics simulations. F.L., J.F.-M. and R.M.S. analyzed the structure. F.L., R.H.E. and R.M.S. wrote the paper with input from all authors.

Corresponding authors

Correspondence to Fei Li, Robert M. Stroud or Robert H. Edwards.

Ethics declarations
Competing interests

The authors declare no competing interests.

Peer review
Peer review information

Nature Structural & Molecular Biology thanks Albert Guskov and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Primary Handling Editor: Katarzyna Ciazynska and Melina Casadio, in collaboration with the Nature Structural & Molecular Biology team.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Extended data
Extended Data Fig. 1 Sequence alignment of VGLUT homologs and human SLC17 family members.

Sequence alignment of selected VGLUT homologs (line 1-5), other human SLC17 family members (line 6-11) and DgoT from E. coli demonstrates conservation across evolution and the SLC17 family of organic anion transporters. The sequences are listed in order of their sequence identity to rat VGLUT2, as in Fig. 4a. Residues are colored based on their sequence conservation, with red the most conserved and no color the least. This figure was generated with ALINE58.

Extended Data Fig. 2 VGLUT2 does not transport aspartate.

Untransfected (UN) HEK293T cells and HEK293T cells transfected with internalization-defective wild-type (WT) VGLUT2, R88K VGLUT2 and empty vector (EV) were loaded with 3H-aspartate (a) or 3H-glutamate (b) and the efflux monitored at pH 7.4 and 5.5. Neither WT nor R88K VGLUT2 mediate the flux of aspartate. c, Effect of mutations on the surface expression of VGLUT2. HEK cells were transfected with WT, mutant VGLUT2 or empty vector (EV), biotinylated and the total protein (input) or the biotinylated fraction (isolated with streptavidin) immunoblotted for VGLUT2. The quantitation of fluorescent western analysis from 3 biological replicates is shown below as % WT. The data indicate mean ± SEM. **, p<0.01; ****, p<0.0001; ns, not significant by one-way ANOVA.

Source data

Extended Data Fig. 3 Biochemical characterization of VGLUT2-WT and R184Q/E191Q mutant.

a, b, Size exclusion chromatography (SEC) and SDS-PAGE of WT (a) and R184Q/E191Q VGLUT2 (b) in complex with an Fab show high purity and homogeneity. Green indicates Fab alone, red VGLUT2 with Fab. c, Glutamate uptake by purified VGLUT2 reconstituted into liposomes with TF0F1 ATP synthase. The R184Q/E191Q mutant shows significantly less glutamate uptake than WT. The activity was measured as ATP-driven 3H-glutamate uptake for 15 min at 29 °C. The ionophores nigericin and valinomycin (100 nM each) were used to dissipate ΔpH and ΔΨ, respectively. d, The assay for efflux from transfected HEK cells shows that the R184Q/E191Q double mutant confers robust glutamate flux. Error bars indicate SEM. n=3 biological replicates.

Source data

Extended Data Fig. 4 Structure determination of VGLUT2-R184Q/E191Q-Fab complex.

a, Details of the data processing strategy are described in the Methods section. Briefly, 3 rounds of 3D classifications were carried out in RELION to select high quality particles in the same conformation. The resulting particle stack was then refined and polished in RELION. To obtain the final map, polished particles were refined using RELION 3D auto-refine with a manually generated mask including only the VGLUT2 and Fab or using the nonuniformal refinement strategy in cryoSPARC with an automatically generated mask. Both maps resulted in similar resolution and quality. Final maps were then sharpened using Phenix Auto-sharpening and used for refinement and model building. Statistics of the final map were calculated using Phenix_comprehensive_validation tool51 and reported in Table 1. b, Sharpened coulombic density map of R184Q/E191Q was of sufficient quality to resolve the side chains of most residues. The coulombic potential density map is shown in gray mesh and model colored as in Fig. 2.

Extended Data Fig. 5 Docking and molecular dynamics (MD) simulation of ACPD, glutamate, and aspartate show distinct interaction patterns.

a and b, Docking of ACPD and glutamate. ACPD pose from the solved R184Q/E191Q-ACPD complex is shown in transparent green, interacting sidechains in blue (N domain) or pink (C domain) for R184Q/E191Q VGLUT2 (a), and green (N domain) or orange (C domain) for wild-type VGLUT2 (b). Best ACPD poses from flexible and rigid mode docking into R184Q/E191Q VGLUT2 are overlaid in opaque green and dark green, respectively. These poses are virtually identical, with all three functional groups in the same location as the solved pose (transparent). b, Best glutamate pose from flexible mode docking into WT VGLUT2, shown in opaque gray. The location within the binding pocket is similar to the solved pose of ACPD (transparent green); there is an average distance of 1.5 Å between the corresponding carboxyl groups of glutamate and ACPD near R88 and 2 Å between the corresponding carboxyl groups of glutamate and ACPD near R322. c, Position of aspartate at the beginning of an 800 ns simulation, determined by first placing functional groups in positions similar to the solved ACPD then choosing favorable rotamers. d-g, Tracking substrate recognition by MD simulation. d, Distance between the amine group of the ligand and the ring of F326 (dotted trace) or the sidechain oxygen atom of N450 (solid trace) is shown for the 800 ns R184Q/E191Q VGLUT2-ACPD simulation (green), the 1000 ns WT VGLUT2-glutamate simulation (gray), and the 800 ns WT VGLUT2-aspartate simulation (cyan). Gray bars indicate the time at which the snapshots on the left (e-g) were taken. ACPD adopts two stable poses during the simulation, with the amine pointed either towards the F326 ring (e, top) or away from it and towards N450, which also removed one interaction between the carboxyl and R88 (e, bottom). f, Glutamate adopts a single stable pose during the simulation. g, Aspartate is unable to interact with both arginines simultaneously and does not adopt a single stable pose. h, Root mean squared deviation (RMSD) of each ligand during its simulation. Numbers indicate Å.

Extended Data Fig. 6 Structure determination of VGLUT2-WT-Fab complex.

a, Details of the data processing strategy are described in the Methods section and similar to that used for VGLUT2-R184Q/E191Q (Extended Data Fig. 4a). b, Sharpened coulombic density map of WT VGLUT2 was of sufficient quality to resolve the side chains of most residues. The coulombic potential density map is shown in gray mesh and model colored as in Fig. 2.

Extended Data Fig. 7 Pathways leading to the central binding cavity in DgoT are more symmetrical than those in VGLUT2.

Models of cytoplasmic-open WT (a) and R184Q/E191Q (b) VGLUT2 were produced by overlapping the N and C domains separately with the crystal structure of DgoT in the cytoplasm-open conformation. The sidechains of R88 and R322 in VGLUT2 overlap well with the equivalent residues in DgoT. c, Structure of the binding cavity in DgoT is colored by sequence conservation (Extended Data Fig. 1), as in Fig. 3. d, Electrostatic surface of the inward/cytoplasm-open WT DgoT structure. e, Electrostatic surface of model of WT DgoT in outward-open conformation. The outward-open conformation of DgoT was generated by overlaying N and C domains separately with the structure of WT VGLUT2 in outward/lumen-open conformation.

Extended Data Fig. 8 Mass spectrometry identification of palmitoylated C60, C62, and C64.

MS/MS spectra of tryptic peptides spanning amino acids K56 to R69 (a, b), and A57 to R69 (c) of R184Q/E1891Q VGLUT2, obtained by HCD fragmentation of precursor ions 626.0019+3, 626.0023+3 and 874.4525+2, respectively. Excess mass of 238.229 relative to the unmodified sequence, corresponding to a palmitoyl group, was observed in the precursor and in sequence ions including particular cysteine residues. a, Masses of b5 [Roepstorff-Fohlmann-Biemann nomenclature] and higher and y10 and higher indicate palmitoylation at C60 in peptide K56 to R69. b, Masses of b7 and higher and y8 and higher pinpoint palmitoylation at C62 in peptide K56 to R69. c, masses observed for y6 and higher indicate palmitoylation of C64 in peptide K57 to R69. Experimental masses of the most representative sequence ion peaks are labelled in the spectra, further indicating the fragment type. The position of fragmentation events generating these ions are indicated in the sequences over the spectra. A table indicating the theoretical masses of sequence ions according to the proposed modified sequences is shown on the right, with red indicating the observed fragments.

Extended Data Fig. 9 Current traces from pmVGLUT2 WT and mutant recordings.

Representative current traces from oocytes injected with WT, R213S, D371N pmVGLUT2-HA cRNA and uninjected controls. Recordings were performed in steps of 10mV from −120 to 60 mV. Holding current was −30 mV.

Supplementary information
Reporting Summary
Supplementary Video 1

All-atom MD simulation of WT VGLUT2 with L-Glu starting in a top-scoring docked pose (pose 1).

Supplementary Video 2

All-atom MD simulation of WT VGLUT2 with L-Glu starting in the second-best docked pose (pose 2).

Supplementary Video 3

All-atom MD simulation of WT VGLUT2 with L-Glu starting in a lower-scoring docked pose (pose 3).

Supplementary Video 4

All-atom MD simulation of WT VGLUT2 with L-Glu starting in a lower-scoring docked pose (pose 3) (run #2).

Source data
Source Data Fig. 1

Transport assays.

Source Data Fig. 4

VGLUT2-pH (c) and GFP–VGLUT (d,e) imaging.

Source Data Fig. 5

Transport assays and I–V curves.

Source Data Extended Data Fig. 2

Transport assays.

Source Data Extended Data Fig. 2

Western blots.

Source Data Extended Data Fig./Table 3

Transport assays.

Rights and permissions

Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.

Reprints and permissions

About this article
Cite this article

Li, F., Eriksen, J., Oses-Prieto, J.A. et al. Substrate recognition and allosteric regulation of synaptic vesicle glutamate transporter VGLUT2. Nat Struct Mol Biol 32, 1479–1487 (2025). https://doi.org/10.1038/s41594-025-01568-8

Download citation

Received
26 July 2022

Accepted
17 April 2025

Published
03 June 2025

Version of record
03 June 2025

Issue date
August 2025

DOI
https://doi.org/10.1038/s41594-025-01568-8

Share this article

Anyone you share the following link with will be able to read this content:

Get shareable link

Provided by the Springer Nature SharedIt content-sharing initiative

Subjects
Biochemistry
Cryoelectron microscopy
Molecular biology
Molecular modelling

Cancel

Preview · live