Structural basis for transport and inhibition of the human glucose-6-phosphate transporter G6PT

Xia, Zhanyi; Wang, Yaqi; Wu, Di; Chi, Cheng; Li, Chen; Chen, Ligong; Jiang, Daohua

doi:10.1038/s41467-025-64464-1

Download PDF

Article
Open access
Published: 24 October 2025

Structural basis for transport and inhibition of the human glucose-6-phosphate transporter G6PT

Zhanyi Xia^1,2^na1,
Yaqi Wang³^na1,
Di Wu ORCID: orcid.org/0000-0003-2271-4963^1,2^na1,
Cheng Chi⁴^na1,
Chen Li³,
Ligong Chen ORCID: orcid.org/0000-0002-7893-7173³ &
…
Daohua Jiang ORCID: orcid.org/0000-0002-4397-5807^1,2

Nature Communications volume 16, Article number: 9420 (2025) Cite this article

9782 Accesses
2 Citations
4 Altmetric
Metrics details

Subjects

Abstract

The human glucose-6-phosphate transporter (G6PT) moves glucose-6-phosphate (G6P) into the lumen of endoplasmic reticulum, playing a vital role in glucose homeostasis. Dysregulation of G6PT causes glycogen storage disease 1b. Despite its functional importance, the structure, G6P recognition, and inhibition mechanism of G6PT remain unclear. Here, we report the cryo-EM structures of human G6PT in apo, G6P-bound, and the specific inhibitor chlorogenic acid (CHA)-bound forms, elucidating the structural basis for G6PT transport and inhibition. The G6P pocket comprises subsite A for phosphate and subsite B for glucose. The CHA occupies the G6P site and locks G6PT in a partly-occluded state. Functional assays demonstrate that G6PT activity is enhanced by co-expression of glucose-6-phosphatase (G6PC), but G6PT does not form a complex with G6PC. Together, this study provides a solid foundation for understanding the structure‒function relationships and pathology of G6PT and sheds light on the future development of potential therapeutics targeting G6PT.

Introduction

Glucose metabolism provides energy for almost all domains of cellular activity¹. In humans, glucose is first converted into glucose-6-phosphate (G6P), which functions as a key hub for downstream carbohydrate metabolic pathways, including glycogen synthesis, the hexosamine pathway, the pentose phosphate pathway, glycolysis, and glucose production². The glucose-6-phosphate transporter (G6PT, encoded by SLC37A4) moves cytoplasmic G6P into the endoplasmic reticulum (ER) lumen, where G6P breaks down into glucose and phosphate by the glucose-6-phosphatase (G6Pase), the final step of both gluconeogenesis and glycogenolysis^3,4,5,6. G6PT is ubiquitously expressed in tissues; for example, it couples with G6Pase-α (also known as G6PC1) in the liver, kidney, and intestine to maintain blood glucose homeostasis⁷, and it also couples with G6Pase-β (also known as G6PC2) in neutrophils to maintain energy homeostasis^8,9,10. G6PT deficiency is associated with glycogen storage disease type 1b (GSD-1b)^11,12, an autosomal recessive disorder with a relatively low prevalence rate of approximately 1 in 500,000 individuals. However, GSD-1b is often accompanied by symptoms of hypoglycemia, excessive glycogen accumulation in the liver and kidney, growth retardation, nephromegaly, hepatomegaly, hyperlipidemia, lactic acidemia, and hyperuricemia^3,13, which severely impair health and quality of life. In addition, a striking difference between GSD-1b and other GSDs is immune deficiency characterized by neutropenia and myeloid dysfunction⁸. To date, 132 pathogenic variants of G6PT have been characterized, all of which are associated with GSD-1b¹⁴. Furthermore, studies suggest that G6P inhibition is a potential strategy for treating type 2 diabetes¹⁵ and antagonizing tumor metastasis^16,17. Chlorogenic acid (CHA), the natural ester of caffeic acid and (−) quinic acid synthesized by plants, is a specific and strong inhibitor of G6PT and has been clinically investigated for the treatment of glioblastoma and lung cancer^18,19,20,21. These findings highlight the physiological and pharmacological importance of G6PT.

G6PT belongs to the solute carrier 37 (SLC37) family, which consists of 4 isoforms, SLC37A1-4 (also called sugar‒phosphate exchangers SPX1-4)^7,22,23. Members of the SLC37A family are cognate to bacterial organophosphate/phosphate exchangers, such as the E. coli hexose-6-phosphate transporter (UhpT) and glycerol-3-phosphate (G3P) transporter (GlpT)^22,23,24. G6PT shares a low amino acid sequence identity (less than 25%) with SLC37A1-3²³. SLC37A1, SLC37A2, and G6PT are G6P transporters, whereas SLC37A3 cannot transport G6P, and its function remains unknown²⁵. In contrast to the functional coupling of G6PT and G6Pase, SLC37A1 and SLC37A2 are independent of G6Pase and insensitive to CHA²⁵. G6PT is localized in the ER membrane with a “KKXX” ER-retention signal motif at the C-termini, the mutation of which leads to mislocalization of G6PT and a congenital disorder of glycosylation^26,27. Protease digestion assays revealed that G6PT contains an even number of transmembrane helices with both the N- and C-termini facing the cytoplasm²⁸. Glycosylation scanning assays suggested that G6PT consists of 10 transmembrane helices (TMs)^28,29; however, G6PT was predicted to have 12 TMs by homology modeling on the basis of the crystal structure of GlpT³⁰. G6PT is highly specific for G6P and can also transport 1,5-anhydroglucitol-6-phosphate (1,5AG6P), a close analog of G6P³¹. G6PT has been proposed as a Pi-related antiporter because the inclusion of 50 mM Pi enhances the G6P-Pi and Pi-Pi antiporter activities in proteoliposomes³². However, the concentration of cytosolic Pi normally does not exceed 5 mM, and preloading native liver microsomal vesicles with or without 5 mM Pi did not affect the G6P uptake rate, suggesting that G6PT may operate as a uniporter³³. Consequently, the structure and G6P transport mechanism of G6PT have remained unclear.

In this work, we report the cryo-electron microscopy (cryo-EM) structures of human G6PT in apo, substrate G6P-bound, and inhibitor CHA-bound forms. Combining with in vitro biochemical and transport assays, we elucidate the detailed recognition sites for G6P and CHA and the mechanism of action of G6PT at the atomic level, offering crucial insights into G6PT function. Moreover, we analyze the effects of molecular disruption caused by known GSD-1b-related mutations on the basis of the G6PT structure, providing important insights into G6PT pathology. These findings could advance the future development of potential therapeutics targeting G6PT.

Results

Functional characterization of G6PT

We first assayed the G6P uptake activity of human wild-type (WT) G6PT overexpressed in HEK293F cells, which exhibited weak activity that was slightly greater than that of cells expressing GFP (Fig. 1a). Previous studies revealed that co-expressing G6PT with G6PC substantially increased G6P transport activity¹¹. We therefore coexpressed WT G6PT with human WT G6PC, which indeed yielded an approximately 5-fold increase in G6P uptake compared with that in cells expressing GFP or co-expressing GFP and G6PC (Fig. 1a). Additionally, G6P uptake can be completely abolished by the G6PT-specific inhibitor chlorogenic acid (CHA), confirming the G6P uptake activity of G6PT (Fig. 1a).

**Fig. 1: Functional characterization and structure of human G6PT.**

We next investigated how G6PC enhances G6PT activity. Our confocal imaging results revealed that both G6PC and G6PT are localized in ER membranes when coexpressed in HEK293F cells (Supplementary Fig. 1a). However, G6PC can barely be copurified with G6PT in the coexpressed cells (Supplementary Fig. 1b), indicating a weak interaction between G6PT and G6PC. This finding is in line with a previous study showing no Förster resonance energy transfer (FRET) signal between the two proteins³⁴. Thus, we speculate that G6PT does not form a complex with G6PC.

Structure determination of hG6PT

G6PT is a relatively small membrane protein (~46 kDa) that lacks a discernible extramembrane domain, making cryo-EM single-particle analysis difficult. To overcome this obstacle, we fused a short α-helical ALFA-tag (SRLEEELRRRLTEP)³⁵ to G6PT before Y6, which was designated G6PT-ALFA (Supplementary Fig. 1c). Functional analysis revealed that G6PT-ALFA exhibited similar G6P uptake activity to that of WT G6PT (Fig. 1a). To facilitate structural study, we purified G6PC1-ALFA in the presence of the ALFA-specific nanobody (NbALFA), which was eluted as a stable complex from size-exclusion chromatography (SEC) (Fig. 1b). The NbALFA works as an apparent fiducial marker, enabling accurate particle alignment (Fig. 1c and Supplementary Fig. 2b). After extensive cryo-EM data processing, the final density map of G6PT (G6PT-apo) was determined at a moderate resolution of 3.6 Å (Supplementary Table 1), partly because of preferred particle orientation, as reflected by a small cFAR score of 0.31 (Supplementary Fig. 2f). The density map allows the construction of a reliable model of G6PT comprising Y6 to E425 with the assistance of its Alphafold2³⁶ model (Fig. 1c, d and Supplementary Fig. 2c). Notably, we observed strong nonproteinaceous densities located in the cleft between the two domains, presumably belonging to the bound lipid or detergent molecules (Supplementary Fig. 3a, b).

The overall structure of G6PT adopts the canonical MFS fold³⁷. The 12 TM helices of G6PT are organized into an N-domain (TM1-6) and a C-domain (TM7-12) with a central pseudo twofold symmetry axis perpendicular to the ER membrane plane (Fig. 1d, e). The N- and C-domains are connected by a long amphipathic loop, namely, L6-7 (37 amino acids long). Part of L6-7 (P203 to K212) is invisible to cryo-EM, presumably due to its conformational flexibility (Fig. 1d, e). The N- and C-domains engage in extensive contacts on the cytosolic side and separate from each other at a long distance of ~31 Å on the ER luminal side, indicating that G6PT-apo is captured in an outward-open conformation (Fig. 1d). TMs 1, 4, 7, and 10 line the central cavity, which is supported by TMs 2, 5, 8, and 11 (Fig. 1e). The banana-like TM5 and TM11 couple with TM8 and TM2, forming an open cavity toward the ER lumen as large as 32° (Fig. 1d, e). The C-terminal loop (C-loop) is positioned in a groove of the C-domain (Fig. 1e). The interactions between the C-loop and the C-domain may also play a role in regulating G6PT function. The last four residues KKAE at the C-terminus, which act as a conserved ER retrieval signal motif for the localization of membrane proteins in the ER membrane²⁶, are invisible in our G6PT structure. Previous studies have revealed that C-loop deletion at R415 greatly reduces the G6P transport activity of G6PT³⁸, whereas the deletion of the C-loop at R423 leads to mislocalization of G6PT to the Golgi and causes a congenital disorder of glycosylation²⁷.

G6P recognition in hG6PT

To understand how hG6PT recognizes G6P, we purified G6PT-ALFA in the presence of NbALFA and G6P and solved the cryo-EM structure of G6PT in complex with G6P (G6PT-G6P) at a moderate resolution of 3.6 Å (Supplementary Figs. 1d, e, 4). The final reconstruction map of G6PT-G6P has a cFAR score of 0.52, suggesting a slight issue of preferred particle orientation (Supplementary Fig. 4g). Similar to the G6PT-apo structure, the G6PT-G6P structure adopts an outward-open conformation with its central cavity facing the ER lumen, representing a state of G6P release (Fig. 2a–c). Structural superposition of G6PT-G6P and G6PT-apo reveals that the two structures are essentially identical, with a small root mean square deviation (r.m.s.d.) value of 0.248 Å over 359 aligned Cα atoms (Supplementary Fig. 5a–c). In contrast to the empty cavity of G6PT-apo, strong electron microscopy density was observed in the central pocket of G6PT-G6P, which fits a G6P molecule well (Fig. 2d and Supplementary Fig. 4e).

**Fig. 2: Structural basis of G6P recognition by G6PT.**

In the outward G6PT-G6P, G6P is positioned at the bottom of the highly positive cavity nearly parallel to the membrane plane, which is approximately 14 Å from the cytosolic layer and 28 Å from the luminal layer (Fig. 2c, e). A closer look reveals that G6P is recognized by a specific binding site, which we termed subsite A for phosphate recognition and subsite B for glucose recognition (Fig. 2f, g). In subsite A, the phosphate group of G6P engages in potential polar interactions with Y60 and K64 on TM2 and H366 on TM11 (Fig. 2f, g). In subsite B, the glucose moiety participates in potential polar interactions with S142 and W138 on TM5, Y233 on TM7, the backbone carbonyl of I346 on TM10, and H366 on TM11, as well as van der Waals interactions with W118 on TM4 (Fig. 2f, g). To validate the role of these interacting residues in G6P recognition, we generated the mutants Y60F, K64A, W118A, W118F, W138A, S142A, Y233F, and H366A and assayed their biochemical profiles and G6P transport activity. These mutants presented similar expression levels, fluorescence-detection SEC (fSEC) profiles, and ER membrane localization to those of WT hG6PT (Supplementary Figs. 6a–g, 7a, b, 8). However, the mutations Y60F and K64A in subsite A greatly reduced G6P uptake activity by approximately 75% and 60%, respectively (Fig. 2h). On the other hand, the W118A, W118F, W138A, S142A, Y233F, and H366A mutations in subsite B also greatly decreased G6P transport activity by at least 70% compared with that of WT G6PT (Fig. 2h). Moreover, the W118R and W118C mutations, which are associated with GSD-Ib¹⁴, almost completely abolished G6P uptake (Fig. 2h). Sequence alignment revealed that Y60 and W118 are conserved in SLC37A1-3, whereas K64, W138, S142, Y233 and H366 are not conserved (Supplementary Fig. 9). These structural and functional findings indicate that both the glucose and the phosphate are critical for G6P recognition and specificity in G6PT.

Although the outward G6PT-G6P structure cannot be well superimposed with the E. coli glycerol-3-phophate transporter GlpT (PDB code: 1PW4)³⁹ (Supplementary Fig. 5d), the central cavity of GlpT displays a positive potential surface, which is similar to that of G6PT (Supplementary Fig. 5e). Structural superposition of the substrate binding sites revealed that Y60, K64 and W118 of subsite A of G6PT are replaced by equivalent residues of Y76, K80 and W138 in GlpT, whereas residues in subsite B are not conserved (Supplementary Fig. 5f), in good accordance with the same phosphate moieties of glycerol-3-phosphate and G6P as well as the different structures of glycerol and glucose. These observations suggest an evolutionary relationship between bacterial GlpT and human G6PT.

G6PT inhibition by CHA

Chlorogenic acid (CHA), an ester formed from cinnamic acid and quinic acid, is a reversible and competitive inhibitor of G6PT¹⁹ (Fig. 3a). The G6P transport activity of rat liver microsomes and proteoliposomes containing recombinant G6PT proteins was effectively inhibited by CHA^11,32. Additionally, neither SLC37A1 nor SLC37A2 showed sensitivity to CHA²⁵, suggesting high specificity of CHA for G6PT. We assayed the concentration-dependent inhibition of G6P uptake by CHA in G6PT- and G6PC1-coexpressing HEK293F cells. The resulting half-maximal inhibitory concentration (IC₅₀) for CHA is 45 ± 10 μM (Fig. 3b), which is ~5-fold more potent than the competitive inhibition of G6P transport by CHA via liver microsomes from 20-h starved male Wistar rats (IC₅₀ value of ~226 μM)⁴⁰.

**Fig. 3: Structural basis of G6PT inhibition by chlorogenic acid (CHA).**

To dissect how CHA inhibits G6PT, we purified G6PT-ALFA in the presence of NbALFA and CHA and determined the cryo-EM structure of G6PT in complex with CHA (G6PT-CHA) at 3.4 Å resolution (Fig. 3c, d, and Supplementary Figs. 1f, g, 10). The cFAR score of G6PT-CHA map is 0.53, suggesting a slight issue of preferred particle orientation (Supplementary Fig. 10g). The density map permits the unambiguous assignment of protein side chains to the placement of CHA (Supplementary Fig. 10c, e). In sharp contrast to the outward-open G6PT-apo and G6PT-G6P structures, the luminal gate of G6PT-CHA is completely closed, whereas the cytoplasmic gate is partly open, indicating an inward-occluded state (Fig. 3c, d). Strong electron microscopy density was observed in the partially opened central cavity of G6PT-CHA, which accommodates a CHA molecule unequivocally (Fig. 3e and Supplementary Fig. 10e).

CHA comprises a caffeic acid moiety and a quinic acid moiety (Fig. 3a). In the G6PT-CHA structure, CHA is vertically positioned in the pocket, with the quinic acid moiety buried in the cavity and the caffeic acid group facing the cytosol, which is very close to the cytoplasmic layer (Fig. 3d). More precisely, the catechol ring of CHA is located in the space between TM5 and TM10, forming edge-to-face π‒π stacking interactions with F134 on TM5 and van der Waals interactions with V351 on TM10, as well as a hydrogen bond interaction with N354 from TM10 (Fig. 3f, g). The middle linker of CHA is sandwiched by W138 on TM5 and H366 on TM11, which participate in hydrogen bond interactions and van der Waals interactions. The quinic acid moiety forms multiple polar interactions with K64, W118, and H366 (Fig. 3f, g). Sequence alignment revealed that most of the residues participating in CHA-binding are not conserved in SLC37A1-3, which explains the specificity of CHA for G6PT (Supplementary Fig. 9).

To substantiate CHA-binding, we mutated the key interacting residues and assessed their effects on CHA potency. The mutations K64A, W118A/F, F134A, and H366A greatly impair G6P uptake (Figs. 2h, 3h) and thus are not feasible for this evaluation. In contrast, W138F and N354A did not affect G6P transport activity (Fig. 3h), both of which also presented similar expression levels, fSEC profiles and ER membrane localization to those of WT G6PT (Supplementary Figs. 6g, 7b, 8). We further investigated the concentration-dependent inhibition of G6P transport by CHA in the W138F and N354A mutants. The IC₅₀ value for W138F is 523 ± 210 μM, which is approximately 11.5-fold greater than that of CHA in WT G6PT; for N354A, the highest tested CHA concentration of 20 mM only reduced G6P uptake by ~20% (Fig. 3i), strongly indicating that both residues play crucial roles in CHA inhibition.

Owing to the binding of CHA close to the cytosolic layer of the ER membrane, an empty cavity is present in the vicinity of the quinic acid moiety (Fig. 3j), providing ample space for accommodating optimized inhibitors. Indeed, synthetic derivatives of CHA, such as S3483 and S4048 (Fig. 3k), are 2–3 orders of magnitude more potent than CHA in inhibiting G6PT^41,42. Most likely, the 2-(4-chloro-phenyl)-cyclopropylmethoxy moieties of S3483 and S4048 can be placed in the deep cavity above the CHA, and the benzimidazole moiety of S4048 can fit into the empty cavity next to the CHA in the N-domain of G6PT (Fig. 3j, k). These moieties of CHA derivatives could establish additional interactions with G6PT, thereby improving their affinity for G6PT.

The inward-occluded G6PT-CHA cannot be well superimposed with the outward-open G6PT-G6P because of their different conformations (Supplementary Fig. 11a). Interestingly, the quinic acid moiety of CHA is placed in a similar position to that of G6P and displays overlapping interactions with K64, W118, W138, and H366 (Supplementary Fig. 11b, c), indicating that the quinic acid moiety may act as a G6P surrogate, which is in line with the competitive property of CHA¹⁹. Compared with the inward-open E. coli GlpT (PDB code: 1PW4)³⁹, CHA-binding induces an approximately 20° rotation of the N-domain relative to the C-domain (Supplementary Fig. 11d); however, the catechol ring of CHA is stuck between TM5 and TM10, preventing closure of the cytosolic gate. Therefore, we propose that CHA inhibits G6PT not only by partly occupying the G6P-binding site but also by locking it in the inward-occluded state.

State transitions of G6PT

To gain insight into the domain movements of G6PT in state transitions, we compared the structure of the inward-occluded G6PT-CHA with that of the outward-open G6PT-G6P (Fig. 4a, b). Structural alignments of the two N-domains or the two C-domains from the two distinct states reveal only minor local conformational shifts; however, when the C-domains are superimposed, the N-domain of G6PT-G6P swings approximately 50° and expands over 20 Å on the luminal side relative to that of G6PT-CHA (Fig. 4b), suggesting rigid body-like motion during G6P transport. These observations are consistent with the shared rocker-switch-like alternating-access mechanism of MFS transporters^43,44.

**Fig. 4: Structural basis of human G6PT state transition.**

In the inward-occluded G6PT-CHA structure, the closed luminal gate is stabilized by multiple hydrogen-bond and salt-bridge interactions between the N- and C-domains. For example, K46 and D47 from TM2 establish a hydrogen bond network with S385, T386, A388, and K389 on TM11 and W393 on TM12; K29 and S32 on TM1 participate in electrostatic interactions with D245 and Q248 on TM7 (Fig. 4c–e). Notably, the pathogenic mutation S385R causes GSD-Ib¹⁴, suggesting that disruption of the hydrogen-bond networks may affect conformational equilibrium during state transitions, thereby impairing G6PT activity. In the outward-open G6PT-apo structure, the closed cytosolic gate is stabilized by three pairs of charge‒helix dipole interactions: D285 of TM8 and the N-termini of TM5, R126 of TM4 and the C-termini of TM10, and D72 of TM1 and the N-termini of TM11 (Fig. 4f, g). In addition, R300 on TM9 forms electrostatic interactions with the backbone carbonyl of P131 on TM5 and E355 on TM10, contributing to stabilizing the closed cytosolic gate. The R300C/H mutations are related to GSD-Ib¹⁴^, and a previous study reported that R300C/H retained less than 10% G6P transport activity compared with WT G6PT⁴⁵.

To examine these structural observations, transport assays revealed that the D72A and D285A mutations greatly reduced G6P transport by more than ~90% and that R126D only reduced G6PT activity by ~20% in comparison with WT G6PT, whereas the triple-site mutation D72A/R126D/D285A completely abrogated G6P transport (Fig. 4h). The expression level and ER membrane localization of these mutants are similar to those of WT G6PT; however, their stability is impaired, as reflected by their broad fSEC profiles and the results of the GFP fluorescence-based thermal shift stability assay (Supplementary Figs. 7e, 12a).

Transport mechanism of G6PT

G6PT possesses an unexpectedly large and positively charged central cavity in both the inward-occluded and outward-open states (Fig. 5a, b). R28 on TM1, K64 on TM2, and K240 on TM7 define the electropositive potential surface of the central cavity and form a charge relay system, which is responsible for attracting, binding and translocating negatively charged G6P. These membrane-embedded positive residues are stabilized by nearby polar residues and experience local conformational shifts in different states (Fig. 5c, d). The R28A, R28H, K64A, K240A and K240C mutants presented similar expression, fSEC profiles, and cellular localization to those of WT G6PT (Supplementary Figs. 6d–h, 7a–d, 8). However, R28A completely abrogated G6P uptake, and the GSD1b-causing mutation R28H greatly reduced G6P uptake by approximately 75% (Fig. 5e), in agreement with the findings of a previous study⁴⁶; the K240A mutation decreased G6P activity by 40%, whereas the pathogenic mutation K240C almost fully abolished the transport activity (Fig. 5e), which is also in line with the findings of a previous study²⁹. These findings highlight the functional importance of the three positive residues in the cavity.

**Fig. 5: Transport and inhibition model of human G6PT.**

On the basis of the existing knowledge and the findings of this study, we propose a transport model for G6PT (Fig. 5f). When G6PT assumes a cytoplasm-opening (inward) conformation, cytosolic G6P is attracted by the positively changed cavity and binds to the G6P recognition site, whereas glucose and other metabolic derivatives cannot form proper interactions and are thus excluded (Fig. 5f, left). The binding of G6P induces local conformational rearrangements, resulting in closure of the cytosolic gate and transmission to an occluded state (Fig. 5f, upper-middle). Because the occluded state of MFS transporters is generally metastable⁴⁷, the occluded G6PT with G6P bound opens the luminal gate and switches to a lumen-opening (outward) state. Moreover, the upper positive R28 and K240 values provide electrostatic strength that may accelerate the movement of G6P from its binding site, resulting in the release of G6P to the ER lumen (Fig. 5f, right). Notably, the K64-R28-K240 relay of G6PT is conserved in the E. coli glycerol-3-phosphate transporter GlpT (K80-R45-R269), suggesting a conserved role of this relay in transporting phosphate-conjugated substrates.

Previous studies¹¹ and our results revealed that the coexpression of G6PC1 and G6PT substantially increases G6P uptake by G6PT (Fig. 1a). However, the coexpression of G6PC and G6PT resulted in no FRET signal³⁴ and no copurification (Supplementary Fig. 1b), suggesting functional coupling between them. Proteoliposome transport assays demonstrated that 50 mM internal Pi greatly enhanced G6PT activity³². Therefore, we speculate that after transiting into the lumen-facing conformation, luminal Pi may access the central cavity and interact with the charge relay to completely facilitate G6P release. Notably, pretreatment of native liver microsomal vesicles with or without 5 mM Pi did not affect the G6P uptake rate³³, suggesting that whether Pi acts as a counterion transported by G6PT remains unclear (Fig. 5f). Considering that G6PT is strongly colocalized with G6PC in the ER membrane, we hypothesize that the G6PC-stimulated transport activity of G6PT arises from the fact that the hydrolysis of G6P by G6PC not only increases the local Pi concentration but also decreases the G6P concentration.

Pathogenic mutations in G6PT

To date, more than one hundred pathogenic variants of G6PT have been associated with GSD-1b¹⁴. To understand the structural pathology of G6PT, we mapped 55 missense pathogenic mutations onto our G6PT structure (Fig. 6a), which demonstrated that these mutations are broadly distributed in both the N-domain (20 mutations) and the C-domain (35 mutations). Some mutations, such as W118R/C, directly alter the G6P recognition site and dramatically reduce G6P uptake (Fig. 2h). The R28H/C mutation disrupted the charge relay and abolished G6P transport activity (Fig. 5e)⁴⁶. Some mutations take place at the N- and C-domain interfaces or the helix‒helix packing interface, which may affect state transitions or protein stability (Supplementary Fig. 12a). For example, the Y24H and R166L mutations eliminate the hydrogen bond interactions between these residues and their respective nearby interacting residues (Fig. 6b–d). Our transport assay demonstrated that Y24H and R166L reduced G6P uptake by approximately 60% and 70%, respectively (Fig. 6e). The protein stability of Y24H and R166L is decreased compared with that of WT G6PT (Supplementary Fig. 12a). In addition, mutations, including G19R, G20D, G50R/E, G68R, G83E, G115R, G122E, G135D, G150R, G273S, G335E, G339D/C and G376S, replace glycine residues in TMs with larger side chain residues, which may lead to the loss of helical dynamics and potential disruption of helical packing (Fig. 6c). The mutations G339D and G339C not only greatly impaired G6PT activity but also decreased protein stability (Fig. 6e and Supplementary Fig. 12a). Although the G335E mutant has similar G6P uptake activity to WT G6PT, it presents reduced protein expression and stability (Fig. 6e and Supplementary Figs. 6h, 12a).

**Fig. 6: Pathogenic mutations in human G6PT.**

Discussion

In this work, we investigated the structural mechanism of the transport and inhibition of human G6PT and the coupling mechanism of G6PT and G6PC. By means of the ALFA-tag protein fusion strategy, we determined the cryo-EM structures of human G6PT in complex with the substrate G6P and the inhibitor CHA in the outward-open and inward-occluded states, respectively. The G6P-bound structure defines a G6P recognition site that is closer to the cytosolic side. The G6P recognition site consists of subsite A for phosphate recognition and subsite B for glucose recognition, both of which are essential for G6P transport. Moreover, the G6PT structure and the G6P site are similar to those of the bacterial GlpT³⁹, suggesting a conserved functional and evolutionary relationship between G6PT and GlpT.

Previous studies revealed that G6PT is highly specific for G6P³² and can also transport 1,5-anhydroglucitol-6-phosphate (1,5AG6P), a close analog of G6P that lacks the C-1 hydroxyl group³¹. Disrupted rat liver microsomes catalyze the hydrolysis of G6P, 2-deoxy-D-glucose-6-phosphate, glucosamine-6-phosphate and mannose-6-phosphate (the C-2 epimer of G6P) at nearly identical rates; however, intact microsomes catalyze the hydrolysis of the latter three compounds at rates less than 10% of that of G6P^48,49, strongly indicating that compartmentation confers substrate specificity for G6PC and that G6P is a favorable substrate for G6PT. Since we have not resolved the G6P-bound G6PT structure in a cytoplasm-open state, the detailed molecular mechanism for the substrate specificity of G6PT deserves further investigation.

Our structural analysis revealed that G6PT adopts the canonical alternating-access mechanism³⁷: the N- and C-domains move like a rigid body around the G6P-binding site to open the cytosolic gate for G6P loading and subsequently open the luminal gate for G6P release (Fig. 5f). The specific and competitive inhibitor CHA¹⁹ binds to the cytoplasm-open G6PT, occupies the G6P site and induces conformational changes to lock it in the inward-occluded state, thereby preventing state transitions of G6PT (Fig. 5f, middle center). We further identified that there is extra space in the cavity of the CHA-bound structure, providing room for optimized CHA-based inhibitors or new classes of inhibitors. G6PT and G6PC are colocalized in the ER membrane, and their coexpression strongly enhances G6PT activity; however, previous FRET³⁴ and our pull-down assays indicated negligible binding between them. Therefore, we postulate that G6PT is functionally related to G6PC but does not form a complex with it. A previous study revealed that a high concentration of Pi enhances G6PT activity³². G6PC hydrolyzes G6P, resulting in an increase in the local Pi concentration and a decrease in G6P, both of which may contribute to stimulating G6PT activity.

In summary, this study provides important mechanistic insights into the substrate recognition, inhibition, and pathogenic mutation of human G6PT and the coupling mechanism of G6PT and G6PC, which should aid in the future development of potential therapeutics for treating G6PT-related diseases.

Methods

[¹³C]G6P transport assay of G6PT and its mutants

The full-length human wild-type G6PT gene (UniProt accession code: O43826) was subcloned and inserted into a pEG2 BacMam expression vector with a Twin-Strep tag, green fluorescent protein (GFP), followed by an HRV 3C protease site at the N-terminus (G6PT-GFP). The full-length human wild-type G6PC1 gene (UniProt accession code: P35575) was subcloned and inserted into a pEG2 BacMam expression vector with an HRV 3C protease site, green fluorescent protein (GFP), followed by an 8 × His tag at the C-terminus (G6PC1-GFP). With G6PT-GFP as a template, all the G6PT mutants were generated via a standard PCR-based strategy, and all the constructs were verified via DNA sequencing. All the primers are summarized in Supplementary Table 2. Bacmids were generated by transforming plasmids into DH10bac competent cells (Thermo Fisher). After that, the Bac-to-Bac baculovirus expression system (Invitrogen, USA) was used to produce baculoviruses in Spodoptera frugiperda (Sf9) insect cells (Invitrogen, USA). P2 viruses were collected and used for transfecting HEK293F cells (FreeStyle 293F, Gibco, USA). HEK293F cells were cultured at 37 °C with 5% CO₂ in OPM-293 CD05 medium (OPM Biosciences). When the cell density reached 2.5 × 10⁶ cells/mL, the cells were infected with P2 viruses, and the viability was greater than 95%. The cells were cultured in suspension medium supplemented with 1% (v/v) heat-inactivated fetal bovine serum (FBS, Gibco) in a 37 °C incubator with 5% CO₂ operating at 120 rpm. After 12 h, sodium butyrate (10 mM final) was added to the culture mixture to increase protein expression. The cells were subsequently incubated for 48 h at 37 °C. The cells were collected by centrifugation at 1000 rpm for 3 min and washed once with wash buffer A1 [1× Hank’s balanced salt solution (HBSS, Gibco), 2% FBS], followed by permeabilization for 10 min at 37 °C with permeabilizing buffer A2 (1 × HBSS, 10 µM digitonin).

The cells were subsequently washed with wash buffer A3 (1×HBSS) to remove the remaining permeabilization buffer. The uptake of G6P was initiated by immediately adding a final concentration of a 1 mM G6P (MedChemExpress) mixture ([¹³C]G6P:G6P at a molar ratio of 1:1). After incubation for 10 min at 37 °C, the cells were washed twice with 500 µL of ice-cold ending buffer A3, after which [¹³C]G6P and [¹³C]glucose were extracted via 500 µL of prechilled 80% (v/v) methanol and stored in a −80 °C freezer overnight. The extract was separated by centrifugation at 12,000 rpm for 20 min, and 400 µL of the supernatant was transferred to a new 1.5 mL EP tube and stored at −80 °C until quantitative mass spectrometry analysis was performed. The precipitate was lysed in 400 µL of 0.1 M KOH, and its protein concentration was determined with a BCA protein assay kit (Thermo Scientific).

To calibrate [¹³C]G6P uptake, the molar concentrations of [¹³C]G6P (A) and [¹³C]glucose (B) in the extract were determined via quantitative mass spectrometry analysis. The protein concentration (C) of the precipitate was determined with a BCA protein assay kit. The final statistics are calculated via the following formula:

$$Y=\frac{A+B}{C}$$

(1)

To calibrate the specific [¹³C]G6P uptake of G6PT variants, cells were infected with 2% (v/v) P2 viruses of wild-type G6PT or variants and 1% (v/v) P2 viruses of G6PC1. The detection of [¹³C]G6P uptake (Y1) of G6PT variants was performed as described above. In addition, 2% (v/v) P2 viruses expressing GFP and 1% (v/v) P2 viruses expressing G6PC1 were used as negative controls to calibrate the effect of endogenous G6PT on G6P uptake (Y0). The specific [¹³C]G6P uptake (%) of G6PT variants (YS) was calculated via the following formula:

$${YS}(\%)=\frac{Y1-Y0}{Y1({WT})-Y0}\times 100$$

(2)

For the chlorogenic acid (CHA) inhibition assay, the reaction buffer [1 mM G6P mixture ([¹³C]G6P:G6P 1:1)] that initiated the transport reaction was supplemented with different concentrations (20 nM to −20 mM) of chlorogenic acid. The cells were infected with 2% (v/v) P2 viruses of wild-type G6PT or its variants and 1% (v/v) P2 viruses of G6PC1 when the cell density reached 2.5 × 10⁶ cells/mL and the viability was above 95%. In addition, 2% (v/v) P2 viruses expressing GFP and 1% (v/v) P2 viruses expressing G6PC1 were used as negative controls to calibrate the effect of endogenous G6PT on G6P uptake. The uptake assay was performed as described above. The molar concentrations of [¹³C]G6P and [¹³C]glucose in the extract were normalized to the protein concentration determined via the BCA method. Each assay was carried out in three independent experiments. The statistics for G6P uptake by G6PT and its variants were analyzed, and the IC₅₀ values were calculated via nonlinear regression in Prism 8.0.1.

Pull-down assay of G6PT and G6PC1

The full-length human wild-type G6PT gene (UniProt accession code: O43826) was subcloned and inserted into a pEG2 BacMam expression vector with a Twin-Strep tag, red fluorescent protein (mCherry), followed by an HRV 3C protease site at the N-terminus (G6PT-mCherry). Bacmids were generated by transforming plasmids into DH10bac competent cells (Thermo Fisher). After that, the Bac-to-Bac baculovirus expression system (Invitrogen, USA) was used to produce baculoviruses in Spodoptera frugiperda (Sf9) insect cells (Invitrogen, USA). P2 viruses of G6PT-mCherry and G6PC1-GFP were collected and used for transfecting HEK293F cells (Gibco, USA). HEK293F cells were cultured in OPM-293 medium (OPM, China) in a 37 °C incubator with 5% CO₂. P2 baculoviruses of G6PT-mCherry and G6PC1-GFP (0.5 mL) were added to 50 mL of 293F cells at a density of 2.5 × 10⁶ cells per mL, and then 10 mM sodium butyrate (Sigma, USA) was added to the culture mixture after 12 h of culture. After another 48 h, the cells were collected via centrifugation, and the cell pellets were frozen in liquid nitrogen and stored at −80 °C. The cell pellets were thawed on ice and resuspended in buffer A containing 20 mM HEPES (pH 7.5), 150 mM NaCl, 5 mM β-mercaptoethanol (β-ME), and a protease inhibitor cocktail including 1 mM phenylmethylsulfonic acid acyl fluoride (PMSF), 2 μM pepstatin, 4.2 μM leupeptin and 0.8 μM aprotinin. The cell membranes were collected by ultracentrifugation at 158,600×g for 30 min and solubilized in buffer A supplemented with 2 mM ATP, 5 mM MgCl₂, 1% (w/v) lauryl maltose neopentyl glycol (LMNG, Anatrace, USA), and 0.1% (w/v) cholesterol hemisuccinate (CHS, Anatrace, USA) for 1.5 h at 4 °C. After ultracentrifugation at 158,600×g for 20 min, the supernatant was collected and loaded onto Streptactin beads (Smart-Lifesciences, China) at 4 °C. The resin was washed with 5 column volumes of buffer B (buffer A supplemented with 2 mM ATP, 5 mM MgCl₂, 0.01% (w/v) LMNG and 0.001% (w/v) CHS) and 15 column volumes of buffer C (buffer A supplemented with 0.01% (w/v) LMNG and 0.001% (w/v) CHS). The protein was eluted with buffer D (buffer C supplemented with 5 mM desthiobiotin (Sigma, USA)), and the eluted samples were collected. The supernatant and eluted samples were then assayed by in-gel fluorescence imaging (GFP: Ex 488 nm, Em 512 nm; mCherry: Ex 587 nm, Em 610 nm).

Immunofluorescence assays

For immunostaining, HEK293F cells were cultured in confocal dishes precoated with PDL. To determine the localization of the G6PT variants within the cell, the cells were infected with 2% (v/v) P2 viruses of wild-type G6PT-GFP or its variants when the cell density reached 5 × 10³ cells per dish. In addition, 2% (v/v) P2 viruses expressing GFP were used as controls. After the cells were incubated for 48 h at 37 °C, the endoplasmic reticulum was stained with ER-Tracker Red (1:1000, C1041M, Beyotime), and the cell nuclei were stained with Hoechst 33342 (1:100, C1027, Beyotime) for 30 min at 37 °C with 5% CO₂. Images were acquired with a confocal microscope (Zeiss LSM980).

To detect the colocalization of G6PT with G6PC1 in cells, cells were infected with 2% (v/v) P2 viruses of wild-type G6PT-mCherry and GFPC1-GFP when the cell density reached 5 × 10³ cells per dish. After the cells were incubated for 48 h at 37 °C, the cell nuclei were stained with Hoechst 33342 (1:100, C1027, Beyotime) for 30 min at 37 °C with 5% CO₂. Images were acquired with a confocal microscope (Zeiss LSM980) (Presented in Supplementary Fig. 8).

Western blot analysis

To characterize the biochemical properties of the G6PT mutants, two milliliters of cultured HEK293F cells expressing GFP alone or the WT G6PT or G6PT mutants were collected via centrifugation. The cell pellets were solubilized in 250 μL of buffer containing 20 mM HEPES pH 7.5, 150 mM NaCl, 5 mM β-ME, and a protease inhibitor cocktail including 1 mM PMSF, 2 μM pepstatin, 4.2 μM leupeptin, 0.8 μM aprotinin, 1% (w/v) LMNG and 0.1% (w/v) CHS for 1.5 h at 4 °C. The insoluble material was removed via ultracentrifugation, and the supernatant containing the G6PT mutants was collected and then assayed via fluorescence-detection size-exclusion chromatography (fSEC) (Ex 488 nm, Em 512 nm).

In addition, the samples were separated via SDS‒PAGE. The gels were then subjected to in-gel fluorescence imaging and excited at 488 nm via a Bio-Rad ChemiDoc MP Imaging System. Following imaging, the separated proteins were transferred onto PVDF membranes (Millipore) via a standard electroblotting procedure. Prior to antibody incubation, the PVDF membranes were blocked with a solution of 7% nonfat milk in TBST to minimize nonspecific binding. The membranes were subsequently incubated overnight at 4 °C with primary antibodies, including mouse GFP tag Monoclonal antibody (catalog # 66002-1-Ig, Proteintech, diluted 1:20,000) and mouse Beta Actin Monoclonal antibody (catalog # 66009-1-Ig, Proteintech, diluted 1:2000). After overnight incubation, the membranes were washed thoroughly and then incubated with the appropriate secondary antibodies Anti-mouse IgG (catalog # 7076, Cell Signaling Technology, diluted 1:20,000) for 40 min at room temperature. Following three rigorous washes with TBST, immunoblotting was performed via Immobilon Western (Millipore), and the proteins were visualized with a Tanon 5200 Chemiluminescent Imaging System to detect the target proteins.

GFP fluorescence-based thermal shift stability assay

We assessed the protein stability of WT G6PT and the selected mutants via a GFP fluorescence-based thermal shift stability assay^50,51. Protein samples of WT G6PT and the selected mutants were kept at 4 °C, treated at 37 °C for 10 min, treated at 43 °C for 10 min, and treated at 49 °C for 10 min. After that, the fluorescence-detection size-exclusion chromatography (fSEC) profiles of these samples were collected. The fSEC peak height of WT G6PT or its mutants at 4 °C was defined as 100%, and all other data were normalized to that value.

Expression and purification of the ALFA nanobody

The nanobody of ALFA-tag (NbALFA) was expressed and purified similarly to previous study^6,52. The NbALFA sequence was subcloned and inserted into a pET-21a vector with a GST tag, followed by an HRV 3C protease site at the N-terminus. The plasmid was transformed into the E. coli BL21 (DE3) strain (Solarbio, China). The E. coli cells were then cultured in terrific broth media at 37 °C for 4 h. After that, 0.5 mM IPTG was added to the media, and the E. coli cells were cultured for another 17 h at 18 °C. The E. coli cells were harvested via centrifugation and were lysed in buffer containing 50 mM HEPES (pH 7.5), 150 mM NaCl, 2 mM β-ME, 1 mM PMSF and 10% glycerol and ultracentrifuged at 158,600×g for 50 min. The supernatant was loaded onto glutathione beads (Smart-Lifesciences, China) at 4 °C, washed with 10 column volumes of buffer containing 50 mM HEPES (pH 7.5), 500 mM NaCl, 2 mM β-ME and 1 mM PMSF, 10% glycerol, and eluted with buffer containing 50 mM HEPES (pH 7.5), 150 mM NaCl, 2 mM β-Me, 10% glycerol, and 20 mM reduced L-glutathione. The eluted fraction was incubated with PreScission protease (1:50 w/w ratio), and the GST tag was removed by incubating the samples with glutathione beads. The NbALFA was further purified by size-exclusion chromatography, using a Superose 6 increase 10/300 GL column (Cytiva) equilibrated in buffer containing 50 mM HEPES (pH 7.5), 150 mM NaCl and 2 mM β-ME. The peak fractions were collected, concentrated to 5 mg/mL and stored at −80 °C.

Expression and purification of human G6PT

For cryo-EM analysis, the first five residues of G6PT (Met1-Gly5) were replaced by an ALFA-tag (SRLLEEELRRRTEP), namely, G6PT-ALFA. Bacmids were generated by transforming plasmids into DH10bac competent cells (Thermo Fisher). After that, the Bac-to-Bac baculovirus expression system (Invitrogen, USA) was used to produce baculoviruses in Spodoptera frugiperda (Sf 9) insect cells (Invitrogen, USA). P2 viruses were collected and used for transfecting HEK293F cells (Gibco, USA). HEK293F cells were cultured in OPM-293 medium (OPM, China) in a 37 °C incubator with 5% CO₂. Two milliliters of P2 baculovirus was added to 200 mL of 293F cells at a density of 2.5 × 10⁶ cells per mL, and then 10 mM sodium butyrate (Sigma, USA) was added to the culture mixture after 12 h of culture. After another 48 h, the cells were collected via centrifugation, and the cell pellets were frozen in liquid nitrogen and stored at −80 °C.

For the purification of G6PT-ALFA in complex with NbALFA, the cell pellets were thawed on ice and resuspended in buffer A containing 50 mM NaH₂PO₄/Na₂HPO₄ (pH 7.5), 150 mM NaCl, 5 mM β-ME, and a protease inhibitor cocktail including 1 mM PMSF, 2 μM pepstatin, 4.2 μM leupeptin and 0.8 μM aprotinin. The cell membranes were collected by ultracentrifugation at 158,600×g for 40 min and solubilized in buffer A supplemented with 2 mM ATP, 5 mM MgCl₂, 1% (w/v) lauryl maltose neopentyl glycol (LMNG, Anatrace, USA), and 0.1% (w/v) cholesterol hemisuccinate (CHS, Anatrace, USA) for 1.5 h at 4 °C. After ultracentrifugation at 158,600×g for 30 min, the supernatant was loaded onto Streptactin beads (Smart-Lifesciences, China) at 4 °C. The resin was washed with 5 column volumes of buffer B (buffer A supplemented with 2 mM ATP, 5 mM MgCl₂, 0.01% (w/v) LMNG and 0.001% (w/v) CHS) and 15 column volumes of buffer C (buffer A supplemented with 0.01% (w/v) LMNG and 0.001% (w/v) CHS). The protein was eluted with buffer D (buffer C supplemented with 5 mM desthiobiotin (Sigma, USA)). The eluted samples were collected and incubated with PreScission protease (1:50 w/w ratio) and NbALFA (1:2 molar ratio) on ice for 12 h. The protein was concentrated and further purified by size-exclusion chromatography on a Superose 6 Increase 10/300 GL column (Cytiva) equilibrated in buffer E (50 mM NaH2PO4/Na2HPO4 pH 7.5, 150 mM NaCl, 5 mM β-ME, 0.005% (w/v) LMNG and 0.0005% (w/v) CHS). The peak fractions were concentrated to 10 mg/mL for cryo-EM grid preparation.

G6PT-ALFA in complex with NbALFA and G6P or CHA (MedChemExpress) was purified similarly to that described above. The 50 mM NaH₂PO₄/Na₂HPO₄ in buffers A, B, C, D, and E was replaced with 20 mM HEPES, pH 7.5. For G6PT-G6P, buffer A was supplemented with 1 mM G6P, and buffer E was supplemented with 0.25 mM G6P. Before preparing the cryo-EM grids, 25 mM G6P was added to the concentrated sample and incubated on ice for 1 h. For G6PT-CHA, buffer A was supplemented with 2 mM CHA, and buffer E was supplemented with 0.01 mM CHA. Before preparing the cryo-EM grids, 5 mM CHA was added to the concentrated sample, which was subsequently incubated on ice for 1 h.

Cryo-EM sample preparation and data collection

The purified samples were subsequently centrifuged at 15,600 × g for 10 min at 4 °C. After that, aliquots of 4.0 μL samples were added to glow-discharged grids (Quantifoil, Cu, R1.2/1.3, 300 mesh). The grids were blotted for 3.5–4.5 s at 4 °C with 100% humidity and plunge-frozen in liquid ethane cooled by liquid nitrogen with Vitrobot Mark IV (Thermo Fisher Scientific).

Cryo-EM data were collected on a 300 kV Titan Krios microscope (Thermo Fisher Scientific) equipped with a K3 direct electron detector and a GIF quantum energy filter (slit width of 10 eV). All movie stacks, each containing 40 frames, were automatically acquired via EPS at a nominal magnification of × 105,000 with a pixel size of 0.85 Å and defocus values ranging from −1.0 to −2.0 μm. The dose rate was adjusted to 15 e⁻/pixel/s. A total of 1556, 4216 and 2804 movie stacks were collected for G6PT-apo, G6PT-G6P and G6PT-CHA, respectively.

Cryo-EM data processing

The data-processing pipelines are shown in the Supplementary Figs. 2, 4, and 10. All the data were processed via RELION-3⁵³ and CryoSPARC⁵⁴.

For G6PT-apo, the 1556 movie stacks were motion corrected, binned 2-fold, and dose-weighted via MotionCorr2⁵⁵^, and the defocus values of each summed micrograph were estimated via Gctf⁵⁶. After that, micrographs were used for autopicking. A total of 1,771,198 particles were picked and extracted in bin2 and a box of 128 pixels. The particles were subsequently imported into CryoSPARC⁵⁴ and subjected to two rounds of 2D classification to remove junk particles, resulting in 488,473 and 291,829 good particles in the first and second rounds, respectively. The good particles from the second round of 2D classification underwent four rounds of hetero refinement. A total of 50,530 particles from the best class of hetero refinement were retracted in bin1 and a box of 256 pixels and then subjected to nonuniform refinement with C1 symmetry imposed, yielding a map at 4.5 Å resolution. Using this map as a model, good particles from the first round of 2D classification underwent 4 rounds of hetero refinement. A total of 55,039 particles from the best class of hetero refinement were then retracted in bin1 and a box of 320 pixels and then subjected to nonuniform refinement with C1 symmetry imposed, yielding a refined map at 3.9 Å resolution. Further local refinement with a mask covering only the transmembrane helices of G6PT yielded a refined map at 3.6 Å resolution.

For G6PT-G6P, the data were processed in total in CryoSPARC⁵⁴. The 2235, 1088, and 893 movie stacks were motion corrected and binned 2-fold via patch motion, and contrast transfer function (CTF) parameters were estimated via patch CTF estimation. After that, micrographs were used for blob particle picking. A total of 3,595,409, 991,106, and 890,566 particles were picked and extracted in bin 2 and a box of 128 pixels, respectively. After two rounds of 2D classification, 523,967, 101,252, and 226,738 particles were selected from the last round of 2D classification, respectively. The 523,967 particles were used ab initio, which then generated three classes of particles, each containing a 3D map. The best map containing clearly resolved transmembrane helical density was then used as a model for hetero refinement. The 523,967, 101,252, and 226,738 particles selected from the last round of 2D classification were subjected to three rounds of hetero refinement. A total of 151,770 particles from the best class of hetero refinement were retracted in bin 1 and a box of 320 pixels. Then, the 151,770 particles were subjected to nonuniform refinement with C1 symmetry imposed, yielding a refined map at 3.5 Å resolution. Further local refinement with a mask covering only the transmembrane helices of G6PT without the detergent micelle yielded a refined map at 3.6 Å resolution.

For G6PT-CHA, the 2804 movie stacks were motion corrected, binned 2-fold, and dose-weighted via MotionCorr2⁵⁵^, and the defocus values of each summed micrograph were estimated via Gctf⁵⁶. After that, micrographs were used for autopicking. A total of 3,141,154 particles were picked and extracted in bin2 and a box of 128 pixels. The particles were subsequently imported into CryoSPARC⁵⁴ and subjected to two rounds of 2D classification to remove junk particles. A total of 511,424 good particles were selected and then used for ab initio reconstruction. After two rounds of ab initio reconstruction, the best class has a map containing clearly resolved transmembrane helical density, which was then used as a model for hetero refinement. Furthermore, 738,270 good particles were selected from the 2D classification and underwent three rounds of hetero refinement via the model generated ab initio. A total of 101,915 particles from the best class of hetero refinement were then retracted in bin1 and a box of 320 pixels and then subjected to nonuniform refinement with C1 symmetry imposed, yielding a refined map at 3.8 Å resolution. Further local refinement with a mask covering only the transmembrane helices of G6PT without the detergent micelle yielded a refined map at 3.4 Å resolution.

Model building and refinement

The predicted AlphaFold2 model of G6PT was fitted into the cryo-EM density map of G6PT-apo, G6PT-G6P, and G6PT-CHA in Chimera⁵⁷ and was manually inspected and adjusted in Coot⁵⁸. For G6PT-G6P and G6PT-CHA, G6P and CHA were manually fitted into the extra densities within the maps in Coot⁵⁸. Refinement of the G6PT-apo, G6PT-G6P, and G6PT-CHA models against the cryo-EM map in real space was performed via the phenix.real_space_refine in PHENIX⁵⁹. The models were subsequently checked and corrected in Coot. The model vs. map FSC curve was calculated via Phenix.mtrage. The statistics of the cryo-EM data collection and model refinement of G6PT-apo, G6PT-G6P, and G6PT-CHA are summarized in Supplementary Table 1.

All figures were prepared with ChimeraX⁶⁰ and PyMol (Schrödinger, LLC).

Reporting summary

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

Data availability

The three-dimensional cryo-EM density maps of human G6PT-apo, G6PT-G6P, and G6PT-CHA have been deposited into the Electron Microscopy Data Bank (EMDB) under accession codes EMD-62584, EMD-62602, and EMD-62585, respectively. The atomic coordinates of human G6PT-apo, G6PT-G6P, and G6PT-CHA have been deposited into the Protein Data Bank (PDB) under accession codes 9KUY, 9KVV, and 9KV0, respectively. Previously published data for the cryo-EM structure of GlpT are available with PDB accession code 1PW4. The source data underlying Figs. 1a, b, 2h, 3b, h, I, 4h, 5e, 6e, and Supplementary Figs. 1b, 1d–g, 6, 7, and 12 are provided as a Source Data file. Source data are provided with this paper.

References

Adeva-Andany, M. M., Perez-Felpete, N., Fernandez-Fernandez, C., Donapetry-Garcia, C. & Pazos-Garcia, C. Liver glucose metabolism in humans. Biosci. Rep. https://doi.org/10.1042/BSR20160385 (2016).
Rajas, F., Gautier-Stein, A. & Mithieux, G. Glucose-6 phosphate, a central hub for liver carbohydrate metabolism. Metabolites. https://doi.org/10.3390/metabo9120282 (2019).
Chou, J. Y., Matern, D., Mansfield, B. C. & Chen, Y. T. Type I glycogen storage diseases: disorders of the glucose-6-phosphatase complex. Curr. Mol. Med. 2, 121–143 (2002).
Article CAS PubMed Google Scholar
Ghosh, A., Shieh, J. J., Pan, C. J., Sun, M. S. & Chou, J. Y. The catalytic center of glucose-6-phosphatase : HIS¹⁷⁶ is the nucleophile forming the phosphohistidine-enzyme intermediate during catalysis. J. Biol. Chem. 277, 32837–32842 (2002).
Article CAS PubMed Google Scholar
Foster, J. D. & Nordlie, R. C. The biochemistry and molecular biology of the glucose-6-phosphatase system. Exp. Biol. Med. 227, 601–608 (2002).
Article CAS Google Scholar
Xia, Z. et al. Structural insights into glucose-6-phosphate recognition and hydrolysis by human G6PC1. Proc. Natl. Acad. Sci. USA 122, e2418316122 (2025).
Article ADS CAS PubMed PubMed Central Google Scholar
Chou, J. Y. & Mansfield, B. C. The SLC37 family of sugar-phosphate/phosphate exchangers. Curr. Top. Membr. 73, 357–382 (2014).
Article CAS PubMed PubMed Central Google Scholar
Chou, J. Y., Jun, H. S. & Mansfield, B. C. Neutropenia in type Ib glycogen storage disease. Curr. Opin. Hematol. 17, 36–42 (2010).
Article CAS PubMed PubMed Central Google Scholar
Chou, J. Y., Jun, H. S. & Mansfield, B. C. Glycogen storage disease type I and G6Pase-beta deficiency: etiology and therapy. Nat. Rev. Endocrinol. 6, 676–688 (2010).
Article CAS PubMed PubMed Central Google Scholar
Jun, H. S. et al. Lack of glucose recycling between endoplasmic reticulum and cytoplasm underlies cellular dysfunction in glucose-6-phosphatase-beta-deficient neutrophils in a congenital neutropenia syndrome. Blood 116, 2783–2792 (2010).
Article CAS PubMed PubMed Central Google Scholar
Hiraiwa, H., Pan, C. J., Lin, B., Moses, S. W. & Chou, J. Y. Inactivation of the glucose 6-phosphate transporter causes glycogen storage disease type 1b. J. Biol. Chem. 274, 5532–5536 (1999).
Article CAS PubMed Google Scholar
Gerin, I., Veiga-da-Cunha, M., Achouri, Y., Collet, J. F. & Van Schaftingen, E. Sequence of a putative glucose 6-phosphate translocase, mutated in glycogen storage disease type Ib. FEBS Lett. 419, 235–238 (1997).
Article CAS PubMed Google Scholar
Kishnani, P. S. et al. Diagnosis and management of glycogen storage disease type I: a practice guideline of the American College of Medical Genetics and Genomics. Genet. Med. 16, e1 (2014).
Article CAS PubMed Google Scholar
Wang, Z. et al. Three novel SLC37A4 variants in glycogen storage disease type 1b and a literature review. J. Int. Med. Res. 51, 3000605231216633 (2023).
Article ADS CAS PubMed Google Scholar
Madsen, P. & Westergaard, N. Glucose-6-phosphatase inhibitors for the treatment of type 2 diabetes. Expert Opin. Ther. Pat. 11, 1429–1441 (2001).
Article CAS Google Scholar
Belkaid, A., Currie, J. C., Desgagnes, J. & Annabi, B. The chemopreventive properties of chlorogenic acid reveal a potential new role for the microsomal glucose-6-phosphate translocase in brain tumor progression. Cancer Cell Int. 6, 7 (2006).
Article PubMed PubMed Central Google Scholar
Tahanian, E. et al. Inhibition of tubulogenesis and of carcinogen-mediated signaling in brain endothelial cells highlight the antiangiogenic properties of a mumbaistatin analog. Chem. Biol. Drug Des. 75, 481–488 (2010).
Article CAS PubMed Google Scholar
Wei, Q., Qian, Y., Yu, J. & Wong, C. C. Metabolic rewiring in the promotion of cancer metastasis: mechanisms and therapeutic implications. Oncogene 39, 6139–6156 (2020).
Article CAS PubMed PubMed Central Google Scholar
Hemmerle, H. et al. Chlorogenic acid and synthetic chlorogenic acid derivatives: novel inhibitors of hepatic glucose-6-phosphate translocase. J. Med. Chem. 40, 137–145 (1997).
Article CAS PubMed Google Scholar
Arion, W. J. et al. Chlorogenic acid and hydroxynitrobenzaldehyde: new inhibitors of hepatic glucose 6-phosphatase. Arch. Biochem. Biophys. 339, 315–322 (1997).
Article CAS PubMed Google Scholar
Wang, L. et al. The biological activity mechanism of chlorogenic acid and its applications in food industry: a review. Front. Nutr. 9, 943911 (2022).
Article PubMed PubMed Central Google Scholar
Bartoloni, L. & Antonarakis, S. E. The human sugar-phosphate/phosphate exchanger family SLC37. Pflug. Arch. 447, 780–783 (2004).
Article CAS Google Scholar
Chou, J. Y., Sik Jun, H. & Mansfield, B. C. The SLC37 family of phosphate-linked sugar phosphate antiporters. Mol. Asp. Med. 34, 601–611 (2013).
Article CAS Google Scholar
Maloney, P. C., Ambudkar, S. V., Anatharam, V., Sonna, L. A. & Varadhachary, A. Anion-exchange mechanisms in bacteria. Microbiol. Rev. 54, 1–17 (1990).
Article CAS PubMed PubMed Central Google Scholar
Pan, C. J. et al. SLC37A1 and SLC37A2 are phosphate-linked, glucose-6-phosphate antiporters. PLoS ONE 6, e23157 (2011).
Article ADS CAS PubMed PubMed Central Google Scholar
Gao, C. et al. Retention mechanisms for ER and Golgi membrane proteins. Trends Plant Sci. 19, 508–515 (2014).
Article CAS PubMed Google Scholar
Marquardt, T. et al. SLC37A4-CDG: mislocalization of the glucose-6-phosphate transporter to the Golgi causes a new congenital disorder of glycosylation. Mol. Genet. Metab. Rep. 25, 100636 (2020).
CAS PubMed PubMed Central Google Scholar
Pan, C. J., Lin, B. & Chou, J. Y. Transmembrane topology of human glucose 6-phosphate transporter. J. Biol. Chem. 274, 13865–13869 (1999).
Article CAS PubMed Google Scholar
Pan, C. J., Chen, S. Y., Lee, S. & Chou, J. Y. Structure-function study of the glucose-6-phosphate transporter, an eukaryotic antiporter deficient in glycogen storage disease type Ib. Mol. Genet. Metab. 96, 32–37 (2009).
Article CAS PubMed Google Scholar
Almqvist, J., Huang, Y., Hovmoller, S. & Wang, D. N. Homology modeling of the human microsomal glucose 6-phosphate transporter explains the mutations that cause the glycogen storage disease type Ib. Biochemistry 43, 9289–9297 (2004).
Article CAS PubMed Google Scholar
Veiga-da-Cunha, M. et al. Failure to eliminate a phosphorylated glucose analog leads to neutropenia in patients with G6PT and G6PC3 deficiency. Proc. Natl. Acad. Sci. USA 116, 1241–1250 (2019).
Article ADS CAS PubMed PubMed Central Google Scholar
Chen, S. Y. et al. The glucose-6-phosphate transporter is a phosphate-linked antiporter deficient in glycogen storage disease type Ib and Ic. FASEB J. 22, 2206–2213 (2008).
Article CAS PubMed Google Scholar
Marcolongo, P. et al. The glucose-6-phosphate transport is not mediated by a glucose-6-phosphate/phosphate exchange in liver microsomes. FEBS Lett. 586, 3354–3359 (2012).
Article CAS PubMed Google Scholar
Soty, M. et al. New insights into the organisation and intracellular localisation of the two subunits of glucose-6-phosphatase. Biochimie 94, 695–703 (2012).
Article CAS PubMed Google Scholar
Gotzke, H. et al. The ALFA-tag is a highly versatile tool for nanobody-based bioscience applications. Nat. Commun. 10, 4403 (2019).
Article ADS PubMed PubMed Central Google Scholar
Jumper, J. et al. Highly accurate protein structure prediction with AlphaFold. Nature 596, 583–589 (2021).
Article ADS CAS PubMed PubMed Central Google Scholar
Drew, D., North, R. A., Nagarathinam, K. & Tanabe, M. Structures and general transport mechanisms by the major facilitator superfamily (MFS). Chem. Rev. 121, 5289–5335 (2021).
Article CAS PubMed PubMed Central Google Scholar
Chen, L. Y., Lin, B., Pan, C. J., Hiraiwa, H. & Chou, J. Y. Structural requirements for the stability and microsomal transport activity of the human glucose 6-phosphate transporter. J. Biol. Chem. 275, 34280–34286 (2000).
Article CAS PubMed Google Scholar
Huang, Y., Lemieux, M. J., Song, J., Auer, M. & Wang, D. N. Structure and mechanism of the glycerol-3-phosphate transporter from Escherichia coli. Science 301, 616–620 (2003).
Article ADS CAS PubMed Google Scholar
Schindler, P. W. et al. Identification of two new inhibitors of the hepatic glucose-6-phosphatase system. Drug Dev. Res. 44, 34–40 (1998).
Article CAS Google Scholar
Arion, W. J. et al. Chlorogenic acid analogue S 3483: a potent competitive inhibitor of the hepatic and renal glucose-6-phosphatase systems. Arch. Biochem. Biophys. 351, 279–285 (1998).
Article CAS PubMed Google Scholar
Herling, A. W. et al. Alterations of carbohydrate and lipid intermediary metabolism during inhibition of glucose-6-phosphatase in rats. Eur. J. Pharm. 386, 75–82 (1999).
Article CAS Google Scholar
Drew, D. & Boudker, O. Shared molecular mechanisms of membrane transporters. Annu. Rev. Biochem. 85, 543–572 (2016).
Article CAS PubMed Google Scholar
Wu, D. et al. Transport and inhibition mechanisms of human VMAT2. Nature 626, 427–434 (2024).
Article ADS CAS PubMed Google Scholar
Chen, S. Y., Pan, C. J., Lee, S., Peng, W. & Chou, J. Y. Functional analysis of mutations in the glucose-6-phosphate transporter that cause glycogen storage disease type Ib. Mol. Genet. Metab. 95, 220–223 (2008).
Article CAS PubMed PubMed Central Google Scholar
Chen, L. Y., Pan, C. J., Shieh, J. J. & Chou, J. Y. Structure-function analysis of the glucose-6-phosphate transporter deficient in glycogen storage disease type Ib. Hum. Mol. Genet. 11, 3199–3207 (2002).
Article CAS PubMed Google Scholar
Yaffe, D., Forrest, L. R. & Schuldiner, S. The ins and outs of vesicular monoamine transporters. J. Gen. Physiol. 150, 671–682 (2018).
Article CAS PubMed PubMed Central Google Scholar
Arion, W. J., Ballas, L. M., Lange, A. J. & Wallin, B. K. Microsomal membrane permeability and the hepatic glucose-6-phosphatase system. Interactions of the system with D-mannose 6-phosphate and D-mannose. J. Biol. Chem. 251, 4891–4897 (1976).
Article CAS PubMed Google Scholar
Arion, W. J., Wallin, B. K., Lange, A. J. & Ballas, L. M. On the involvement of a glucose 6-phosphate transport system in the function of microsomal glucose 6-phosphatase. Mol. Cell Biochem. 6, 75–83 (1975).
Article CAS PubMed Google Scholar
Nji, E., Chatzikyriakidou, Y., Landreh, M. & Drew, D. An engineered thermal-shift screen reveals specific lipid preferences of eukaryotic and prokaryotic membrane proteins. Nat. Commun. 9, 4253 (2018).
Article ADS PubMed PubMed Central Google Scholar
Chatzikyriakidou, Y., Ahn, D. H., Nji, E. & Drew, D. The GFP thermal shift assay for screening ligand and lipid interactions to solute carrier transporters. Nat. Protoc. 16, 5357–5376 (2021).
Article CAS PubMed Google Scholar
Wang, Y. et al. Transport and inhibition mechanism for VMAT2-mediated synaptic vesicle loading of monoamines. Cell Res. 34, 47–57 (2024).
Article ADS CAS PubMed PubMed Central Google Scholar
Zivanov, J. et al. New tools for automated high-resolution cryo-EM structure determination in RELION-3. Elife. https://doi.org/10.7554/eLife.42166 (2018).
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
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
Zhang, K. Gctf: real-time CTF determination and correction. J. Struct. Biol. 193, 1–12 (2016).
Article ADS CAS PubMed PubMed Central Google Scholar
Pettersen, E. F. et al. UCSF Chimera—a visualization system for exploratory research and analysis. J. Comput. Chem. 25, 1605–1612 (2004).
Article ADS CAS PubMed Google Scholar
Emsley, P. & Cowtan, K. Coot: model-building tools for molecular graphics. Acta Crystallogr. D Biol. Crystallogr. 60, 2126–2132 (2004).
Article ADS PubMed Google Scholar
Afonine, P. V. et al. Real-space refinement in PHENIX for cryo-EM and crystallography. Acta Crystallogr. D Struct. Biol. 74, 531–544 (2018).
Article ADS 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

Download references

Acknowledgements

We thank Dr. B. Xu at the Cryo-EM Center of the School of Advanced Agricultural Sciences of Peking University and D. Sun at the Cryo-EM Facility at the Institute of Physics, Chinese Academy of Science & Beijing Branch of Songshan Lake Materials Laboratory for support in cryo-EM sample preparation and data collection. We thank Wei Fan for her research assistance service. This work is funded by the National Natural Science Foundation of China (32271272 and T2221001 to D.J.), the Ministry of Science and Technology of China National Key R&D Programs (2022YFA0806503 and 2024YFA1306103 to L.C.), and the Taishan Scholars Program of Shandong Province (tsqnz20231243 to C.C.).

Author information

These authors contributed equally: Zhanyi Xia, Yaqi Wang, Di Wu, Cheng Chi.

Authors and Affiliations

Beijing National Laboratory for Condensed Matter Physics, Laboratory of Soft Matter Physics, Institute of Physics, Chinese Academy of Sciences, Beijing, China
Zhanyi Xia, Di Wu & Daohua Jiang
School of Physical Sciences, University of Chinese Academy of Sciences, Beijing, China
Zhanyi Xia, Di Wu & Daohua Jiang
State Key Laboratory of Membrane Biology, School of Pharmaceutical Sciences, Key Laboratory of Bioorganic Phosphorus Chemistry and Chemical Biology (Ministry of Education), Tsinghua University, Beijing, China
Yaqi Wang, Chen Li & Ligong Chen
Peking University Institute of Advanced Agricultural Sciences, Shandong Laboratory of Advanced Agricultural Sciences at Weifang, Weifang, China
Cheng Chi

Authors

Zhanyi Xia
View author publications
Search author on:PubMed Google Scholar
Yaqi Wang
View author publications
Search author on:PubMed Google Scholar
Di Wu
View author publications
Search author on:PubMed Google Scholar
Cheng Chi
View author publications
Search author on:PubMed Google Scholar
Chen Li
View author publications
Search author on:PubMed Google Scholar
Ligong Chen
View author publications
Search author on:PubMed Google Scholar
Daohua Jiang
View author publications
Search author on:PubMed Google Scholar

Contributions

D.J. and L.C. conceived and designed the experiments. Z.X. prepared samples for the cryo-EM study and made all the constructs. D.W., Z.X., and C.C. collected the cryo-EM data. D.W. and Z.X. processed the data and built and refined the models. Y.W., C.L., and Z.X. performed the G6P transport assays. Z.X. and Y.W. prepared the figures. Z.X., Y.W., D.W., C.C., L.C., and D.J. analyzed and interpreted the results. D.J. and Z.X. wrote the paper, and all the authors reviewed and revised the paper.

Corresponding authors

Correspondence to Ligong Chen or Daohua Jiang.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Nature Communications thanks Adrian Teo, who co-reviewed with Hwee Hui Lau, Ronen Marmorstein, and the other anonymous reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Additional information

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

Supplementary information

Supplementary Information (download PDF )

Reporting Summary (download PDF )

Transparent Peer Review file (download PDF )

Source data

Source Data (download ZIP )

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Cite this article

Xia, Z., Wang, Y., Wu, D. et al. Structural basis for transport and inhibition of the human glucose-6-phosphate transporter G6PT. Nat Commun 16, 9420 (2025). https://doi.org/10.1038/s41467-025-64464-1

Download citation

Received: 13 February 2025
Accepted: 15 September 2025
Published: 24 October 2025
Version of record: 24 October 2025
DOI: https://doi.org/10.1038/s41467-025-64464-1