Structural Prediction and Functional Mechanistic Characterization of Farnesyl Diphosphate Synthase from Mycobacterium tuberculosis

Abstract

Tuberculosis, triggered by the bacterium Mycobacterium tuberculosis (Mtb), continues to pose a major global health challenge, further intensified by the rise in drug resistance. The enzyme (2E,6E)-farnesyl diphosphate synthase (FPPS) in Mtb catalyzes the synthesis of farnesyl diphosphate (FPP), which is crucial for processes such as arabinogalactan biosynthesis in the cell wall. However, the functional characterization of Mtb FPPS is not well understood. Here, a hydrodynamic radius of 5.4 ± 0.3 nm was obtained for monomeric FPPS. The molecular docking relied on a structure predicted using AlphaFold2, and site-specific mutagenesis was subsequently executed. Mutants K201A or K266A showed significantly reduced activity, while D112A, D116A, or R121A mutations completely abolished activity. Deletion of specific segments (Δ58–74, Δ117–126, or Δ249–268) also resulted in complete loss of activity. This work lays a crucial foundation for advancing innovative tuberculosis therapies that function via an alternative mechanistic pathway.

1. Introduction

Mycobacterium tuberculosis (Mtb), the causative agent of tuberculosis (TB), still constitutes a critical international health danger that carries weighty public safety concerns. Despite decades of intensive chemotherapy, approximately one-quarter of the world’s population is still infected with Mtb (World Health Organization, https://www.who.int/news-room/fact-sheets/detail/tuberculosis). TB is further complicated by comorbidities such as HIV and SARS-CoV-2 infections. The growing threat is intensified by drug-resistant Mtb strains that can withstand critical agents, among them isoniazid, pyrazinamide, and rifampicin. These resistant strains significantly hinder effective treatment, contributing to high mortality rates. , These challenges urgently demand a shift in drug discovery priorities toward uncovering fresh therapeutic targets and designing innovative drug candidates.

The type II terpenoid biosynthetic pathway, found in bacteria and some lower eukaryotes, is a critical point of vulnerability in Mtb physiology. , This pathway is essential for the production of isoprenoid precursors that serve as molecular scaffolds for a variety of vital cellular components. A pivotal enzyme in this pathway is (2E,6E)-farnesyl diphosphate synthase (FPPS), which drives the sequential head-to-tail condensation of dimethylallyl diphosphate with two isopentenyl diphosphate units to yield the 15-carbon isoprenoid farnesyl diphosphate (FPP). , FPPS contains highly conserved aspartate-rich motifs (DDXXD) that bind divalent magnesium ions (Mg²⁺), which are essential for its enzymatic activity (Figure ).

1.

Sequence alignment of the farnesyl diphosphate synthase family. The ClustalW default color scheme is applied, in which conserved residues are highlighted with greater color intensity than nonconserved ones. The following reference proteins are included in the alignment: P9WKH1, Mycobacterium tuberculosis (strain ATCC 25618/H37Rv); Q7XYS9, Artemisia spiciformis (Spiked big sagebrush; Artemisia tridentata spiciformis); Q7XYT0, Artemisia spiciformis (Spiked big sagebrush; Artemisia tridentata spiciformis); A0A6B9UF20, Lavandula x intermedia (Lavandin; Lavandula angustifolia x Lavandula latifolia); Q9NH03, Dictyostelium discoideum (Social amoeba); Q54XP1, Dictyostelium discoideum (Social amoeba). The Asp-Asp-X-X-Asp (DDXXD) motif, highlighted in a green frame, is critical for catalytic activity due to its role in binding magnesium ions (Mg²⁺).

FPPS is crucial for Mtb pathogenesis and persistence. ,, First, FPP serves as the precursor for the synthesis of decaprenylphosphoryl-β-d-ribose, which is directly involved in the production of cell wall arabinan and the elongation of meromycolate chains, ultimately leading to the formation of mycolic acids. These long-chain α-alkyl, β-hydroxy fatty acids are fundamental building blocks of the mycobacterial cell envelope, specifically comprising the inner leaflet of the asymmetrical outer membrane. The mycolic acid layer is a major factor in the impermeability of Mtb, providing resistance to hydrophobic antibiotics and enabling survival within hostile macrophage environments. Second, FPP is a vital precursor for menaquinone synthesis, which functions as lipophilic electron carriers in the mycobacterial respiratory chain, as well as for other polyisoprenoids involved in protein glycosylation and tRNA modification. Thus, FPPS represents a promising drug target. However, the functional mechanism of Mtb FPPS is still poorly understood.

Herein, we measured the hydrodynamic radius of monomeric FPPS at 5.4 ± 0.3 nm, followed by molecular docking via an AlphaFold2 model and site-specific mutagenesis. The K201A or K266A mutants showed significantly reduced activity, while the D112A, D116A, or R121A mutations completely abolished activity. Deletion of some segments (Δ58–74, Δ117–126, or Δ249–268) also resulted in complete loss of activity. These findings enhance our mechanistic understanding of FPPS and provide a promising foundation for developing new therapeutic strategies to combat drug-resistant tuberculosis.

2. Results

2.1. Characterization of FPPS Using Dynamic Light Scattering

A hydrodynamic radius of 5.4 ± 0.3 nm was obtained for FPPS via dynamic light scattering (DLS) after centrifugation (Figure ). This indicated that the protein FPPS exists as a monomer.

2.

Dynamic light scattering (DLS) spectrum of the FPPS protein. The hydrodynamic radius of FPPS was measured to be 5.4 ± 0.3 nm, which is consistent with a monomeric state.

2.2. Predicting and Assessing the Quality of the Protein FPPS Model

The protein FPPS structural prediction was carried out with AlphaFold2 (Figure a). This approach diverges from classical homology modeling by applying deep learning algorithms, resulting in improved prediction accuracy and robustness. The quality of the predicted FPPS model was assessed using Ramachandran plot analysis (Figure b; Table ), which evaluates the backbone dihedral angles to ensure that they fall within sterically favorable regions, confirming proper folding. The analysis indicated that 91.3% of residues occupied the most favored regions, 8.1% fell within additional allowed regions, 0.6% were in generously allowed regions, and none were located in disallowed regions (Figure b; Table ). The structure also contained 2 terminal residues, 29 glycines, and 19 prolines (Figure b; Table ), further attesting to the high reliability of the predicted model. A structural analysis using the ProSA-Web server showed a Z-score of −10.5 (Figure c), indicating that the model falls within the expected range and has a stable conformation, confirming its quality.

3.

Modeling and evaluation of FPPS structure. (a) Ribbon diagram of the predicted FPPS structure from two orientations, depicting alpha-helices in cyan, beta-sheets in red, and loops in magenta. The model was predicted using AlphaFold2. (b) Ramachandran plot evaluating model conformation, with core allowed regions indicated in red and progressively less favorable regions in lighter shades. (c) ProSA analysis of model quality, yielding a Z-score of −10.5.

1. Ramachandran Plot Validation of the FPPS Model via PDBsum.

Residues	Residues in most favored regions		Residues in additional allowed regions		Residues in generously allowed regions		Residues in disallowed regions
Structural models	Number of residues	% of residues	Number of residues	% of residues	Number of residues	% of residues	Number of residues	% of residues
AlphaFold2	282	91.3	25	8.1	2	0.6	0	0

^aModels of good quality typically exhibit >90% of residues in the most favored regions.

^bExcluding Gly and Pro, there are 2 terminal residues: glycine residues, shown as triangles, total 29; proline residues number 19.

Overall, of the total residues, 91.3% were found to occupy the most favored region (exceeding the 90% threshold), showing an excellent model quality. Additionally, the ProSA Z-score of −10.5 (below −10.0) further confirms an excellent model quality. Therefore, the predicted model is suitable for further study.

2.3. FPPS-GPP Binding Sites

A high-confidence FPPS-GPP complex, indicated by a binding energy of −5.87 kcal/mol, was obtained through molecular docking with AutoDock, using the reliable FPPS model predicted by AlphaFold2 (Figure a). The cavity within the predicted structure, which carried a partial positive charge, appeared to be well-suited to accommodate GPP (Figure a). The catalytic activity of FPPS was suggested to depend on key conserved residues lining the binding pocket, specifically D112, D116, R121, K201, and K266 (Figure b).

4.

Structural model of the FPPS-GPP (geranyl diphosphate) binary complex. The substrate, geranyl diphosphate (GPP), is rendered in stick form, and the magnesium ion (Mg²⁺) is shown as a yellow sphere. (a) Surface representation of the binary model, colored by electrostatic potential: red indicates negative, blue positive, and white neutral regions. (b) Ribbon diagram of the binary complex. The right panel provides a detailed view of the binding sites, with dashed lines denoting specific charged interactions.

2.4. Characterization of the FPPS Enzyme Activity

Following the molecular docking results, we performed site-specific mutagenesis. The K201A or K266A mutants showed significantly reduced activity (Figure ), while the D112A, D116A, and R121A mutations completely abolished activity (Figure ). The catalytic importance of the negatively charged residues D112 and D116 is highlighted, especially given their proximity to the GPP. D112 and D116 are ideally situated to electrostatically coordinate with Mg²⁺, thereby contributing to the stabilization and charge neutralization of the GPP phosphate group. The GPP substrate likely interacts with the FPPS binding pocket through electrostatic interactions involving Mg²⁺ and the side chains of the key residues. Consistent with previous studies, − these further confirm that negatively charged residues are conserved throughout the farnesyl diphosphate synthase family.

5.

Enzymatic activity assay of FPPS and its mutants. The K201A or K266A mutant exhibited dramatically reduced activity, whereas the D112A, D116A, or R121A mutation abolished activity entirely. Activity values are expressed relative to the wild-type (WT) protein, which was set to 100%.

2.5. The Effect of Triton X-100, EDTA, and Metal Ions on the Activity of Enzymes for FPPS

The influence of metal ions on the FPPS activity was evaluated, revealing near-complete inhibition of the enzyme (Figure ). Furthermore, pretreatment with EDTA followed by dialysis led to total activity loss, underscoring the indispensable role of Mg²⁺ in FPPS function. Complete inactivation was also observed upon Triton X-100 treatment, confirming that FPPS is not a membrane-associated protein (Figure ).

6.

Effects of Triton X-100, EDTA, and metal ions on FPPS activity. The activity was significantly reduced or even completely abolished. The activity of the WT (wild-type) FPPS with Mg²⁺ added was set to 100%.

2.6. Regions 58–74, 117–126, and 249–268 Are Essential for FPPS Activity

The regions (aa 58–74, 117–126, and 249–268) are located around the periphery of the catalytic cavity in FPPS, as indicated by molecular docking results (Figures and a). Deletion of these regions (Δ58–74, Δ117–126, Δ249–268) completely abolished activity (Figure b), suggesting their critical role in FPPS catalysis. This loss of activity is likely due to key amino acids within these regions that are involved in the enzyme’s catalytic function.

7.

Regulation of FPPS activity via peripheral catalytic pocket regions. (a) The segments comprising residues 58–74, 117–126, and 249–268 are shown in green, cyan, and magenta, respectively. (b) Deletion of any of these segments (Δ58–74, Δ117–126, or Δ249–268) abolished activity entirely. The activity of the wild-type (WT) FPPS was set to 100%.

2.7. Challenges in Optimizing FPPS Crystals

To probe the structure–function relationship of FPPS, we set out to determine its three-dimensional structure. Preliminary crystals were successfully obtained and subsequently optimized in 12% (w/v) PEG 6000 with 0.1 M citric acid (pH 4.5), leading to visibly enhanced crystals (Figure a). Despite these improvements, extensive efforts failed to overcome their poor diffraction, preventing further refinement (Figure b). Efforts are now underway to generate diffraction-quality crystals.

8.

Crystallization challenges and limited diffraction quality. (a) The FPPS crystals were of poor quality, resulting in (b) a diffraction pattern of low quality with few well-resolved spots.

3. Discussion

In this study, a hydrodynamic radius of 5.4 ± 0.3 nm was measured for the FPPS monomer. Molecular docking, based on an AlphaFold2-predicted structural model, was conducted, and site-specific mutations were introduced. The K201A or K266A mutants showed significantly reduced activity, while the D112A, D116A, or R121A mutations completely abolished activity. Deletion of any of the segments (Δ58–74, Δ117–126, or Δ249–268) resulted in complete loss of activity. These findings enhance our mechanistic understanding of FPPS and provide a valuable chemical foundation for developing new therapeutic strategies to combat drug-resistant tuberculosis.

As a member of the prenyltransferase family, FPPSs (farnesyl diphosphate synthases) are characterized as short-chain isoprenoid synthases. Their enzymatic activity depends on two conserved aspartate-rich motifs (Asp-Asp-X-X-Asp) that chelate divalent Mg²⁺. − The low sequence identity in the FPPS family suggests that FPPSs may not share a universal catalytic mechanism.

Dynamic light scattering (DLS) analysis gave a hydrodynamic radius of approximately 5.4 nm, consistent with the theoretical value for the monomeric FPPS. Notably, DLS reports the size distribution of scattering particles and does not directly resolve oligomeric states. The observed hydrodynamic radius therefore supports a monomeric species but does not rigorously exclude small populations of dimers or transient higher-order oligomers. The monodisperse size distribution and low polydispersity index nonetheless indicate that the predominant species in solution behaves as a monomer under the experimental conditions. Resolving the oligomeric state unambiguously would be strengthened by the use of supplementary methods, for instance, size-exclusion chromatography paired with multiangle light scattering or analytical ultracentrifugation.

The key findings of this study arise from the systematic mutagenesis of residues within the predicted C-terminal domains. The complete loss of activity upon mutation of D112, D116, or R121 confirms this region as a critical component of the catalytic machinery. D112 and D116 are likely part of or structurally support the second aspartate-rich motif (DDXXD), which binds the isopentenyl diphosphate substrate and Mg²⁺ in class I isoprenyl synthases. Their absolute requirement further validates our structural model. The role of R121 is particularly intriguing as it is not part of the canonical motifs and may be essential for substrate orientation, transition-state stabilization, or maintaining the structural integrity of the active site pocket.

The complete loss of activity in the Δ58–74, Δ117–126, and Δ249–268 variants does not unequivocally demonstrate that these regions participate directly in catalysis. Because each deletion removes 10–17 residues, it may compromise protein folding, conformational stability, solubility, or overall structural integrity. Consequently, the observed inactivation could equally result from indirect structural perturbations that prevent the variants from adopting a native-like functional conformation, rather than from the absence of catalytically essential residues. The biochemical data presented here cannot distinguish between a direct catalytic role and an indirect structural contribution for these segments. We therefore refrain from assigning specific mechanistic functions and emphasize that further biophysical and structural characterizationsuch as circular dichroism spectroscopy, thermal stability assays, limited proteolysis, or high-resolution structure determination of the variantsis required to clarify their precise roles. Nevertheless, the complete inactivation caused by these deletions unambiguously identifies residues 58–74, 117–126, and 249–268 as indispensable for enzymatic activity, whether their role is direct or indirect.

Additionally, the significant reduction in activity observed for the K201A and K266A mutants underscores the functional importance of the positively charged residues distant from the active site. These residues may facilitate long-range electrostatic interactions that stabilize the protein tertiary structure, assist in positioning flexible loops during catalysis, or participate in binding the diphosphate groups of the substrates/products. Their partial loss-of-function phenotype suggests that they modulate rather than strictly dictate the chemical step.

Deletion analysis provides compelling evidence for the essential roles of three specific structural segments (Δ58–74, Δ117–126, and Δ249–268). The segment 117–126, which contains the critical residues D112, D116, and R121, is confirmed as a key catalytic element. The indispensable nature of segments 58–74 and 249–268 suggests a more complex structural mechanism. Their deletion likely leads to local unfolding or global destabilization, preventing the formation of a functional active site. This implies that FPPS catalytic competence depends not only on the precise geometry of active site residues but also on the structural integrity of peripheral catalytic pocket regions (residues 58–74, 117–126, and 249–268).

While aspartate-rich motifs are conserved, Mtb FPPS possesses a unique extended loop between helices D and E, absent in human and E. coli enzymes. , The predicted model suggests this loop protrudes near the active site, consistent with our kinetic data showing distinct bisphosphonate sensitivity. The enzyme also exhibits pronounced substrate inhibition at elevated IPP concentrations, a feature less prominent in eukaryotic FPPS. These structural and functional distinctions, coupled with its essentiality for mycobacterial cell wall synthesis, highlight the enzyme’s divergence and validate it as a selective antitubercular drug target.

A limitation of the current study is the reliance on an AlphaFold2-predicted model, although its accuracy is strongly supported by the mutagenesis data. Our future work should focus on determining the high-resolution crystal structures of wild-type and key mutant FPPS proteins. This would enable the direct visualization of the C-terminal domain architecture and clarify the specific roles of R121, K201, and K266. Additionally, biochemical assays, such as thermal shift analysis, could be used to differentiate between mutations that directly affect catalytic activity and those that primarily destabilize protein folding.

Because our molecular docking analysis used a computationally predicted structure rather than an experimentally determined one, the accuracy of the proposed binding poses is inherently limited. Docking scores alone do not validate any specific binding mode. Consequently, the docking results presented here are strictly hypothesis-generating and served only to guide the selection of residues for site-directed mutagenesis. The conclusions regarding the catalytic mechanism and the roles of individual residues rest entirely on convergent evidence from mutagenesis and kinetic experiments independent of the docking scores.

The structural interpretations in this study rely heavily on an AlphaFold2-derived model. Although this model provides a valuable framework for hypothesis generation, it remains a computational prediction that has not been experimentally validated at high resolution. The putative roles assigned to particular residues in substrate recognition, transition-state stabilization, or catalysis should consequently be treated with caution until substantiated by direct structural evidence from methods including X-ray crystallography and/or cryoelectron microscopy. Our mutagenesis and biochemical data strongly support the functional importance of these residues, but the precise molecular mechanisms inferred from the model are hypotheses, rather than established facts. Driven by these unresolved issues, we set out to investigate how Mtb FPPS operates mechanistically.

In conclusion, our findings suggest that, rather than focusing solely on the active site, targeting the stability and architecture of the entire catalytic domain presents a promising strategy for developing a new generation of highly specific antitubercular agents.

4. Materials and Methods

4.1. FPPS Expression and Purification

FPPS constructs, expression, and purification were performed based on previously reported methods with some modifications. Primers for recombinant FPPS (Mycobacterium tuberculosis, UniProt ID P9WKH1; Figures S1 and S2) were designed for PCR and synthesized by Shanghai Sangon Biotechnology (China). The recombinant FPPS gene was inserted into the pET28a vector, and PCR amplification was carried out using Q5 polymerase. DNA sequencing by Shanghai Sangon Biotechnology (Shanghai, China) confirmed the correct cloning sites. To purify FPPS proteins, cell pellets were resuspended in 45 mL of buffer and homogenized at 650 MPa for 3 min. After centrifugal removal of cell debris, the clarified extract was passed over the Ni-NTA resin. The resin was washed, and FPPS was eluted by using 300 mM imidazole. The eluted protein was concentrated and subsequently fractionated by SEC (size exclusion chromatography) using a Superdex 75 10/300 GL column.

4.2. Dynamic Light Scattering Measurements

The oligomeric state of FPPS was evaluated using DLS (dynamic light scattering) via a Dynapro DLS instrument (Malvern Zetasizer, Malvern, UK). After being concentrated to approximately 2.3 mg/mL, the FPPS protein was centrifuged at 18,000 rpm for 5 min at 4 °C. The prepared sample was then placed in a 1 cm cuvette and measured over 30 acquisitions, with a 120 s equilibration interval before each run. Data analysis employed Zetasizer software (Version 6.20), producing regularized histograms that enabled diameter tracking and determination across the measurement period.

4.3. Prediction and Assessment of the Structural Model of FPPS

The structural model of FPPS was predicted using AlphaFold2 with the protein sequence (UniProt code P9WKH1) as input. Molecular graphics were visualized with PyMOL v2.3.4 (https://www.pymol.org/2/). Geometric analysis was employed to validate the structure. Backbone dihedral angles (φ/ψ) and overall stereochemical quality were evaluated through a Ramachandran plot generated by PDBsum, which distinguished energetically favorable conformations from unfavorable residue orientations, thereby offering quantitative support for the model’s reliability. The structure was further assessed using ProSA (Protein Structure Analysis), a well-established tool for protein model validation. ProSA specifically analyzes X-ray and NMR spectroscopy data.

ClustalW with default settings was employed to carry out multiple sequence alignments, and the results were visualized with ESPript 3.0.

4.4. Molecular Docking

Geranyl diphosphate (GPP) was docked to the FPPS binding pocket using AutoDock 4.2.6, , with genetic algorithm settings of 1000 runs, a population size of 200, 30,000 generations, and 3,000,000 evaluations. The resulting FPPS-GPP interactions were visualized in PyMOL v2.3.4, and the calculated binding energy of −5.87 kcal/mol confirmed the docking reliability. As a measure of FPPS-GPP affinity, more negative binding energies indicate tighter binding and a more stable complex. This metric is crucial for assessing the functional modulation potential and guiding the identification of bioactive compounds for further investigation.

4.5. Enzymatic Activity Assay

A modified version of a previously established method ,, was employed to measure the catalytic activity of FPPS and its mutant variants (Table S1; Figures S1 and S2). The assay involved incubating 50 nM FPPS with 3 U/mL PPase in 20 mM Tris-HCl (pH 7.2), 2 mM MgCl₂ and 50 mM KCl at 25 °C for 20 min, during which PPase converted diphosphate to phosphate (Figure S3). BIOMOL Green reagent was then added and incubated for 30 min, after which the optical density at 620 nm was measured. A phosphate standard curve was used for quantification (Figure S4).

4.6. Assay of the Enzymatic Activity of FPPS Mutants

Site-directed mutagenesis of FPPS was accomplished using PCR primers designed in our lab and synthesized commercially (Shanghai Sangon Biotechnology, China) (Table S1; Figures S1 and S2). The pET28a-FPPS construct served as the amplification template with Q5 polymerase driving the mutagenesis reaction. Nucleotide sequencing (Shanghai Sangon Biotechnology) confirmed the mutations, after which the altered genes were ligated into pET28a. Both expression and purification of the mutant proteins adhered to the wild-type FPPS protocol, and all enzymatic assays were conducted under the matched conditions.

4.7. Crystal Growth and X-ray Data Collection

Preliminary crystal screening was carried out by vapor diffusion at 25 °C, employing six sparse-matrix screens (Index, JCSG, PEG-Rx, PEG-screen, PEG-Ion, and PEG-salt). Drops were set up by combining 0.5 μL of 5.5 mg/mL FPPS with 0.5 μL of crystallization buffer and equilibrating over 65 μL of reservoir solution. Hits were optimized, and crystals destined for data collection were cryoprotected with reservoir solution supplemented with 26% (v/v) glycerol prior to flash-cooling in liquid nitrogen. Data collection took place on beamline BL19U1 at the Shanghai Synchrotron Radiation Facility, Shanghai, China.

4.8. Statistical Analysis

Each experiment was repeated a minimum of three times, and all values are reported as mean ± SD. Statistical computations were carried out in Origin 8.5, Microsoft Excel 2013, and SPSS 19.0. One-way ANOVA (one-way analysis of variance) combined with the Bartlett test was employed to evaluate significance. Assumptions were checked beforehand: the Shapiro–Wilk test assessed residual normality, and Levene’s test examined variance homogeneity across groups. Throughout the analyses, p < 0.05 denoted statistical significance, and p < 0.01 indicated high statistical significance.

5. Conclusions

This study demonstrates that the complete loss of activity upon deletion of segments 58–74, 117–126, or 249–268 reveals that these regions form an essential architectural framework for catalysis. Mutagenesis identified D112, D116, and R121 as indispensable residues, while K201 and K266 played significant modulatory roles. These results offer a novel mechanism of farnesyl diphosphate synthase in Mtb, providing a crucial step toward developing next-generation antitubercular therapeutics.

All supporting data are provided in the article and Supporting Information.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsomega.6c04601.

• Figure S1: The amino acid sequence of FPPS (UniProt code P9WKH1); Figure S2: The optimized sequence of gene FPPS; Figure S3: Phosphate ion formation was examined via a two-step enzymatic sequence; Figure S4: The standard curve for measuring phosphoric acid concentrations; Table S1: Primers used for generating site-directed mutants of FPPS (PDF)

Conceptualization, D.L.; Data curation, D.L.; Formal analysis, D.L.; Investigation, D.L., N.L.; Methodology, D.L.; Resources, D.L.; Software, D.L.; Validation, D.L.; Writingoriginal draft, D.L.; Writingreview and editing, D.L. All authors have read and approved the final version of the manuscript.

Our research work is financially supported by grants from the third batch of the “Tianchi Talent” Young Doctoral Research Grant, Xinjiang Autonomous Region (2025QNBS001), and the Talent Team Support Initiative for Scientific Research and Innovation Platforms of the Xinjiang Talent Development Fund (XJRCXYCTD2025-2).

All authors have read and agreed to the published version of the manuscript.

The authors declare no competing financial interest.