Downregulation of Endo-Beta-N-Acetylglucosaminidase in Caenorhabditis elegans Improves Stress Adaptivity Open Access

Protein N-glycosylation is a dynamic posttranslational modification process found in eukaryotic cells. It is involved in proteins’ folding, trafficking and digestion, cell surface signaling, microbial-host interactions, and immune responses [1, 2]. Its alterations may reflect physiological and/or pathological conditions and may serve as potential therapeutic targets [3].

ENGASE is an endo-deglycosylase for N-glycans on glycoproteins and free oligosaccharides. It hydrolyzes a β-1,4-glycosidic bond between two N-acetylglucosamines (GlcNAc) in the core pentasaccharide [4, 5]. In addition to its hydrolytic activity, a reverse synthesis activity has also been reported and deployed in engineering for glycoproteins with uniformed N-glycan [6, 7].

Previous studies suggested that Engase was a potential therapeutic target for Ngly1-deficient diseases, which often result in embryonic lethality, which can be rescued by Engase knockout [3, 8]. In shrimp, the cumulative mortality and viral copy number were significantly decreased when PmEngase (Penaeus monodon Engase) was silenced [9]. Filamentous fungus Trichoderma atroviride TaEngase18B (Trichoderma atroviride Engase B) knockout strain grew faster and was more resistant to abiotic stresses on cell wall, such as treatment with 1 mol/L NaCl or 0.025% SDS [10]. Accumulated evidences suggest that Engase inhibition plays a protective role under stress conditions.

Caenorhabditis elegans is an ideal multicellular eukaryotic animal model for life cycle and single cell analysis. Its genetic and developmental background is well characterized. Molecular analysis at single cell level has also been well established. In addition, it shares many basic biological processes with human beings, such as aging, digestion, and stress adaptations [11]. CeEngase (C. elegans eng-1) was discovered and characterized based on its sequence homology to Endo-M from Mucor hiemalis [12]. Following this study, CeEngase had been cloned, prepared, and subjected to in vitro assays for its enzymatic activity, substrate specificity, and optimal reaction conditions. A CeEngase deletion mutant was constructed by using conventional mutagenesis. In addition, free oligosaccharides and some basic biological indicators were examined under normal conditions [13, 14].

In this study, a frameshift mutant in exon 2 of CeEngase was constructed and subjected to biological and biochemical analysis. Survival rates of the mutant under stress conditions of the mutant were measured and compared to wild type. Total proteins were collected and subjected to a glycomic assay, and the results of the glycomic assay were analyzed and presented.

All C. elegans strains were cultured on nematode growth media (NGM) plates with OP50 standard food [15]. All assays were performed at 21°C unless otherwise stated. Worms were used in this study at day 1 adult stage unless specified. Detailed strain information for strains was listed in online supplementary Table 1 (for all online suppl. material, see https://doi.org/10.1159/000546244).

CRISPR-Cas9-guided genome editing was performed as previously described [16]. sgRNA for CeEngase gene (eng-1) knockout (5′-CAC​TCG​AAG​AGC​TCT​GGA​GT-3′) was designed by using benchling. The sgRNA was inserted into pDD162 vector [16] for the plasmid (pLXR1). Successful knockouts were screened by PCR and Sanger sequencing.

Genomic DNA was sequenced (primer: 5′-GCG​GGA​TCT​GGA​AAT​TGT​GCT​ATA​T-3′) for knockout. A second around single/sequencing was performed for diploid knockout. Sequencing confirmed line was outcrossed 4 times with wild-type (N2) worms, and the confirmed knockout after outcrossing was named as LXR1. Detailed information for plasmids and primers was listed in online supplementary Table 2.

Using genomic DNA as template, a fragment containing a promotor (2,000 bp) and a gene sequence (1,957 bp) for CeEngase was amplified and inserted into pSM vector [17]. After confirming the construction by sequencing, GFP was inserted into upstream of CeEngase stop codon. In worms with the plasmid (pLXR2), the fluorescent signals generated by CeEngase/GFP (PCeEngase::CeEngase::GFP) may represent the expression pattern of CeENGASE in worms. In addition, the expression of the CeEngase/GFP fusion protein may also affect the distribution of CeENGASE. Site-directed mutagenesis was conducted to generate a 4-base deletion and then detected on PCeEngase::CeEngase::GFP (pLXR3). Detailed information for plasmid and primers was listed in online supplementary Table 2.

Worms with GFP expression were anesthetized with 50 mm muscimol on 3% agarose pads on slides. Images were captured by Andor Dragonfly Spinning Disc Confocal Microscope with 10× objective lens and 488-nm laser. Image processing was performed with Imaris 10.0/Image J software.

Lifespan assays were performed following a standard protocol [18]. Briefly, synchronized larval stage 4 (L4) worms were placed on fresh NGM plates and incubated. Survived worms were counted daily. Worms that did not respond to external mechanical stimuli were counted as dead. Bagged, internal hatching, or missing worms were counted as censored. Worms were transferred to new NGM plates every other day. Sixty worms were checked for each test. Three tests were performed for this lifespan assay. Kaplan-Meier survival analysis with a log-rank test was used to analyze the lifespan of the worms.

The assay was conducted to check the impact of a specific treatment on worm’s mobility by measuring the head-swing ability in liquid. The reported protocol was followed [19]. Briefly, synchronized worms were placed on fresh NGM plates. Two hundred microliters of M9 solution was dripped into the plates to help worms swing freely. Times for head swings in 10 s were counted. Twenty worms were checked for each test. The tests were repeated 3 times.

Osmotic stress resistance assays were performed by following a published protocol with slight modifications [20]. Synchronized L4 stage worms were placed on fresh NGM plates for 24 h. Day 1 adults were transferred from standard 50 mm NaCl NGM plates to high salt NGM plates (≥400 mm NaCl) and incubated at 21°C for 24 h. Worms that did not respond to external mechanical stimuli were counted as nonsurvival and sensitive to osmotic stress. Sixty worms were checked for each test, and the test was repeated three times.

Heat stress assays were performed following the instruction that was slightly modified [21]. Synchronized L4 stage worms placed on fresh NGM plates were incubated at 21°C for 24 h. Day 1 adults were exposed to high-temperature stress at 35°C for 3 h or 7 h, respectively. Those worms that did not respond to external mechanical stimuli after the treatment were counted as sensitive. The percentage of survival worms was evaluated. Sixty worms were checked for each test, and the test was repeated.

RNAi feeding bacteria and plates were prepared following the previous study [22]. Eggs were collected from 20 synchronized day 1 adults for 2 h and fed them with RNAi bacteria. Day 1 adults were incubated at 35°C, 11 h for the stress assay. Survival rates were measured.

The structures of CeENGASE (Q8TA65) and hENGASE (Q8NFI3) were predicted by AlphaFold 3, respectively [23]. The alignment of the 2 was obtained with open ChimeraX access. The display option in ChimeraX was selected to mark critical residues.

CeEngase (eng-1) gene was cloned into pET28a vector (GenScript). The constructed pET28a-eng-1 plasmid was transformed into Escherichia coli BL21 (DE3) (Tiangen). Then single colony was cultured in Luria-Bertani medium, and protein expression was induced by IPTG (1 mm) at 21°C for 16 h. Protein purification was carried out using the traditional Ni Separose™ 6 Fast Flow (GE Healthcare) elution method [24]. The proteins on the column were eluted with a buffer containing 200–300 mm imidazole. The eluted proteins were collected, concentrated by 30 kDa ultrafiltration tube, re-eluted with PBS, and stored at 4°C for assays.

hEngase gene was cloned into pET28a vector (GenScript). Protein expression purification, mutant construction, and enzyme activity against glycoproteins of hENGASE were prepared following the previous study [24]. In brief, a totally 20 μL reaction system consisted of 10 μg RNaseB and 20 μg hENGASE was prepared. Then, the mixture was incubated at 37°C for 24 h. After that, 5 μL loading buffer was added to the reaction system, and the reaction was denatured for 10 min at 100°C. The treated proteins were conducted SDS-PAGE and Caucasian blue stain based on the shift of RNaseB to determine hENGASE enzymatic activity.

This assay was performed by following a published protocol with slight modifications [24]. In brief, a 200-μL reaction system was incubated at 21°C for 24 h, which contained 50 mm sodium acetate reaction buffer (pH 5.5), 100 μg RNaseB (Sigma), 100 μg CeENGASE, and 50 μM Rabeprazole sodium (Topscience) [25]. The reaction was terminated at 100°C for 10 min. After a spin (10,000 g 10 min), the free glycans in supernatant generated in the reaction were collected, purified by 3 kDa ultrafiltration, processed to remove salt by Supelclean™ EVNI-Carb™ SPE Tube (Sigma-Aldrich), condensed by lyophilization, dissolved in ddH₂O, and put forward to MALDI-TOF MS analysis. The internal reference N2H3F1X1 (GlcNAc2Hexose3Fucose1Xylose1) prepared from HRP was used as a semiquantitative control for mass spectrometry detection. The intensities of the peaks corresponding to the products and the internal control were collected. The ratio of the products/the internal control was calculated. The ratios detected with and without Rabeprazole sodium were compared to determine the impact of the assay on CeENGASE.

The assay was performed following the protocol with minor modifications [26, 27]. Synchronized L4 stage worms were treated with 1 mm Rabeprazole sodium [27] and solvent control (DMSO) in OP50 food, respectively. After one and half day’s treatment, plates containing these adults were incubated at 35°C, 7 h for stress assay. Survival rates were measured.

Around 5,000 day 1 adult worms on NGM plates were collected, suspended in M9 buffer, and washed 3 times with the same buffer. The precipitated worms in M9 were collected, treated in RIPA lysis buffer (Beyotime) plus PMSF (100 mm, 1:1,000) (Beyotime) on ice for 3 × 5 min with vortex in interval. The sample was further processed by ultrasonication (10 × 5 s) and a final ultra-high centrifugation (13,000 g, 10 min). Supernatant was collected following reductive alkylation and acetone precipitation [28].

The precipitates were solubilized in 25 mm ammonium bicarbonate solution, then trypsinized at an enzyme/protein ratio of 1/50, and digested at 37°C overnight. The digested peptides were first purified by C18 desalting column (Waters) and then processed for N-glycopeptide enrichment using the ZIC-HILIC Glycopeptide Enrichment Kit (Merck). The enriched N-glycopeptides were lyophilizated, then solubilized in 0.1% formic acid solution, and analyzed by Thermo Fisher Easy-nano LC II system. Mass spectrometry raw data were analyzed by proteome discoverer software (Thermo Fisher), pGlycoNovo, and pGlyco3.0 software [29, 30]. Quantitative analysis was performed by peak finding, extraction, and intensity quantitation functions in pGlycoQuant [31]. Wild-type (N2) and LXR1 worms were both enrolled in this assay, and three technical replicates were performed for each set of samples.

All assays were statistically analyzed by GraphPad Prism 9.5.0. Data comparison between the two groups was analyzed by an unpaired two-tailed Student’s t test. Values were expressed as mean ± standard error of the mean.

Briefly, pLXR1 construct for knockout was injected in gonadal cell together with Pmyo-3::mCherry and Pmyo-2::mCherry. The genomic DNA of offsprings was extracted and put into PCR reactions covering the first three exons of CeEngase. The PCR products were sequenced for candidate mutants. After 4 rounds of screening, a knockout mutation with a 4-base deletion inside the second exon of CeEngase was selected and named as LXR1 for further studies (Fig. 1). A fast PCR reaction that was positive only for wild-type genome and negative for LXR1 was designed for the quality control of LXR1 (online suppl. Fig. 1A). The impacts of the frameshift mutation in LXR1 were further confirmed at the mRNA level by a transcriptome sequencing analysis (online suppl. Fig. 1B) and at the protein level by an in vivo PCeEngase::CeEngase::GFP assay (Fig. 1b).

To explore the impacts of CeEngase on C. elegans, the lifespan and motricity were observed. Comparing to the wild type, the lifespan of LXR1 was improved, with a median extending 2 more days than the wild type (p < 0.05) (Fig. 2a). In addition, the body movement of worms was measured as an indicator for healthy living. Day 5 worms were selected due to its peak point for movement study determined by previous studies [32, 33]. Head swings of worms in M9 solution were counted. A higher frequency was observed in LXR1 (p < 0.01) at adult day 5 (Fig. 2b). Together, these results show that CeEngase knockout has a positive effect on worms.

To address the effect of CeEngase knockout in response to stress, LXR1 and wild type were treated to high osmotic and temperature conditions, respectively. First, we tried high osmotic conditions by incubating day 1 adults on NGM plates containing high salt (≥400 mm NaCl). After incubation at 21°C for 24 h, the survival rates were calculated. The results showed that under osmotic stress, the mean survival rate of LXR1 was 30.64% higher than the wild type (p < 0.05) (Fig. 2c). Next, we investigated whether CeEngase knockout enhanced worms’ resistance to heat. Day 1 adults were incubated at 35°C for 7 h. The results exhibited that the mean survival rate of LXR1 treated under 35°C for 7 h was 26.20% more than wild type (p < 0.01) (Fig. 2d). These experiments suggested that knockout of CeEngase increases adaptivity of worms to stress.

To reveal the possible mechanism, a glycomics analysis was conducted to identify the targets of CeENGASE in worms. Proteins from wild type and LXR1 were prepared, treated with trypsin, and subjected to LC-MS/MS. The data were analyzed by pGlyco3.0, pGlycoNovo, and pGlycoQuant for glycopeptides. A total of 672 and 664 glycopeptides were detected in wild type and LXR1, respectively. Among them, 9 was found only in wild type and 1 was found only of N-glycopeptides in LXR1 (Table 1). For the N-glycopeptides found in both groups, the ratio of LXR1/WT was calculated. Those significantly high (≥7) and low (≤0.05) were listed (Table 1). We noticed that in Table 1, N-glycopeptides generated from glycoproteins associated with basement membrane of extracellular matrix (ECM) were significantly enriched (22.1% up and 21.90% down), among which LAM-1, LAM-2, EPI-1, PAT-2, PAT-3, and MIG-6 were included (Fig. 3). This result indicated that ECM proteins were sensitive to the downregulation of CeENGASE.

The N-glycopeptides detected from wild-type and LXR1 worms in the glycomics assay were subgrouped by N2HX (GlcNAc2HexoseX) and compared. In the detected N-glycopeptides, 274 had the forms of N2HX. These N2HX were further divided into 10 subgroups. In each subgroup, the percentile of significantly up- and downregulation by CeEngase is calculated and listed in Table 2. The average percentile of up- and downregulation for all subgroups was 22% and 19%, respectively. Compared to the average, N2H1 was significantly higher (40.00%) in the up group and N2H7 was significantly lower (9.09%) in the down group. The amplitude of variation in each subgroup was defined by % of ups divided by % of downs. These results implied that N2H7 was more sensitive to CeENGASE regulation, owing to its highest ratio at 3.00. This prediction was consistent to early reports [12, 13].

To further validate whether the downregulation of CeEngase had similar heat stress tolerance in phenotype, CeEngase RNAi was performed. First, the knockdown efficiency of CeEngase RNAi was confirmed (Fig. 4a). Then, the stress assay was conducted at 35°C for 11 h. We found that the survival rate of CeEngase RNAi worms was significantly increased (25.26%, p < 0.01) (Fig. 4b). This result proved that reducing CeEngase expression could improve the resistance of C. elegans. The result could also be concluded that mechanism of heat tolerance in previous knockout worms was mainly benefited from the loss of CeEngase but not the other genes background.

Due to the structural flexibility of its substrate and other reasons, the cryoelectron microscopy structure of hENGASE remains elusive. The structures of CeENGASE and hENGASE were predicted by AlphaFold 3 (Fig. 5a, b) and aligned by ChimeraX. A common domain for substrate binding and enzymatic hydrolysis is presented in Fig. 5c. Seven conserved sites between hENGASE and CeENGASE were identified (His181, Gly184, Asn235, Glu237, Tyr271, Trp282, and Trp306 in hENGASE). Alanine scanning was conducted for these sites. The mutated proteins were expressed, purified, and subjected to an in vitro assay. The result demonstrated that His181 and Asn235 were critical for hENGASE activity (Fig. 5d). The result of His181 was reported for first time, and the observation regarding Asn235 was consistent with an earlier report [34]. His181 and Asn235 (purple), two conserved residues in the catalytic pocket of hENGASE, were identified by mutation analysis (Fig. 5c). The matched residues in CeENGASE were His99 and Asn152, respectively (beige) (Fig. 5c).

Early studies reported that hENGASE was inhibited by Rabeprazole sodium in vitro [27]. Given the structural and functional similarity between hENGASE and CeENGASE, we wished to know if Rabeprazole sodium can inhibit CeENGASE as well. We first examined the role of Rabeprazole sodium in inhibiting CeENGASE in vitro. CeENGASE expression in the prokaryotic system was prepared and subjected to enzyme activity assay. RNaseB was used as the substrate, and N1H7 (GlcNAc1Hexose7), the major product generated by CeENGASE, was detected. This result was in line with the in vivo finding generated in this study (Table 2). Also, the ratio of the main product (N1H7,1378) to internal reference (GlcNAc2Hexose3Fucose1Xylose1, N2H3F1X1, 1211) was decreased around 50 times by Rabeprazole sodium (50 μM) (Fig. 6a). The result demonstrated that Rabeprazole sodium had a significant inhibitory impact on CeENGASE.

To address the effect of Rabeprazole sodium on in vitro study, wild-type worms were subjected to 35°C stress with the drug dissolved in DMSO mixed in the food, while worms treated without the drug were used as the control (DMSO). The result showed that the survival rate of the drug group was significantly improved with 36.60% compared to 10.12% in the control group (p < 0.05) (Fig. 6b). This result indicated that Rabeprazole sodium might inhibit CeENGASE function and further supported the hypothesis that downregulation CeENGASE might serve as a new target for stress adaptation.

A CeEngase knockout with a 4-base frameshift deletion in its second exon was generated and confirmed at DNA, RNA, and protein levels (Fig. 1). In addition, the deletion was independently confirmed by the data generated from a transcriptomic analysis (online suppl. Fig. 1). The knockout was subjected to a panel of functional analysis, including longevity, head swings, heat stress, and high osmotic stress assays. The results demonstrated that worms’ adaptivity to environment was significantly improved by CeEngase knockout (Fig. 2).

Stress adaptivity is a key to survival and evolution. Several stress-regulating pathways, such as insulin/insulin-like factor-1, c-Jun N-terminal kinase, and mitogen-activated protein kinase, have been characterized in C. elegans. The molecular details of these pathways associated with oxidative stress, heat stress, infection, and longevity have been reported [35, 36]. DAF-16 in the insulin/insulin-like factor-1 pathway could help to extend lifespan and regulate resistance to heat and oxidative stress by interacting with several transcription factors, such as HSF-1 and SKN-1 [37‒39]. In addition, HSP-12.6, HSP-16.2, SOD-3, and CTL-1 have been implicated in heat and oxidative adaptation [40]. More than 300 osmotic sensitive proteins have been discovered, including collagen protein DPY-7 and hypodermal secretory protein OSM-8. Deletion of these two genes could activate GPDH-1 and glycerol accumulation, thereby promoting resistance to hypertonic environments [41‒43]. The data collected in this paper suggested a novel stress adaptation mechanism via CeENGASE.

Early studies have indicated that ECM may be involved in stress regulation. In Drosophila, ECM protein Perlecan, Nidogen, Collagen IV, and Laminin were implicated in osmotic regulation [44]. In mice, it was found that pain sensitivity, anxiety, and depression were induced when laminin β1 (LAMB1) in anterior cingulate cortex was changed [45]. In C. elegans, heat stress was regulated via HSF-1 and calcium-binding protein PAT-10 through cytoskeleton and ECM [46, 47]. In addition, oxidative stress has been associated with sulfated chondroitin in ECM [48, 49]. Although early studies conducted in different models suggested that glycoproteins in ECM may represent a general pathway for stress adaptation [50‒52], however, the molecular mechanism for this regulation remains unclear. Our analysis illustrated that several glycoproteins in the basement membrane of ECM (LAM-1, LAM-2, and EPI-1) were relatively more sensitive to CeENGASE regulation, indicating a potential role of CeENGASE for survival and stress adaptation by regulating the structure and function of ECM (Table 1; Fig. 3).

The impact of cytosolic enzymes on secreted ECM glycoproteins has been reported previously [53‒55]. It was found that ECM glycoproteins such as LAMA4 were altered in fibroblasts cells established from NGLY1-CDDG patients [53]. The glycosylation profile of AGRIN, LAMB2, and LAMA5 in ECM was altered in STT3A/3B knockout cells [54], and the potential mechanism behind this observation was investigated [55]. In addition, ENGASE products were purified from extracellular fluids [56, 57]. Finally, it has been reported that knocking out of gh18-10 (Neurospora crassa Engase) has a significant impact on secreted proteins, likely including the ones in ECM [58]. Our observation was consistent to all these previous reports. The Engase knockout impact on glycosylation profile of ECM glycoproteins could be mediated indirectly via its role in glycosylation system of the cells. More studies will be conducted to find out the mechanism of Engase knockout on ECM.

To further validate the impact of CeENGASE downregulation as a potential antistress strategy, RNAi on CeEngase was performed (Fig. 4a). The result showed that knockdown CeEngase could improve heat tolerance of worms (Fig. 4b) and was consistent with the result of knockout.

Rabeprazole is a well-known proton pump inhibitor used for stomach disease [59, 60]. It has been reported to have an inhibitory effect on human ENGASE in vitro [27]. We explored whether Rabeprazole sodium work through CeENGASE as well. Recombinant CeENGASE expressed in E. coli was prepared for an in vitro assay. The CeENGASE activity was significantly inhibited by Rabeprazole sodium (Fig. 6a). This result was well aligned with the previous report [27]. This in vitro result was further supported by an in vivo assay (Fig. 6b). Based on these results, we concluded that Rabeprazole sodium could improve stress resistance through CeENGASE, and CeENGASE could serve as a novel antistress target. A placebo effect of DMSO on the stress response was observed in the in vivo test (Fig. 2d; Fig. 6b). This observation was consistent with previous studies [61, 62]. In addition, using wild-type worms expressing PCeEngase::CeEngase::GFP, no significant decrease in GFP fluorescence was observed in the group treated with Rabeprazole sodium compared to the control group (online suppl. Fig. 2). Therefore, we expected that Rabeprazole sodium might act by inhibiting the enzymatic activity of CeENGASE. We are establishing an enzymatic activity assay for intracellular CeENGASE to help us determine the enzymatic activity of CeENGASE more accurately.

In this study, a preliminary analysis of N-glycopeptides was conducted for impacts of knockout on glycoproteins. Out of our expectation, the product of CeENGASE (N1H0 glycopeptide) was not detected in wild type. This could be an indication that the sensitivity of the assay needs to be improved, and/or the product was engaged in a dynamic process in vivo. To figure out the impact of CeEngase knockout on its substrates cell, the detected glycopeptides were subgrouped via structure of glycans, such as N2H0, N2H1, N2H2…till N2H9. The percentile of each subgroup in total was calculated (Table 2). Compared to the wild type, the N2H7 glycopeptides were relatively accumulated in LXR1 worms, suggesting that glycoproteins with N2H7 glycan were a selective substrate for CeENGASE in vivo (Table 2). This result was consistent with an observation in the RNaseB assay (Fig. 6a). Although N2H5 is the dominant glycan on RNaseB detected by PNGase F [63, 64], the major product generated by CeENGASE is N1H7 (Fig. 6a).

Protein N-glycosylation is a new set of biological codes that follow the central dogma and are critical for regulating healthy life [65]. Protein N-glycosylation and deglycosylation are a comprehensive posttranslational process accomplished by a series of sequential reactions, including the one catalyzed by ENGASE. N-glycans on glycoproteins can act as a biological recognition signals for glycoproteins folding, trafficking, and localization. Removing glycan signals on glycoproteins by ENGASE may induce systemic negative effects on cells and microenvironment. Early studies indicated that knockout Engase function was a way to rescue Ngly1 defect in mice [3, 8, 66] and knockout Engase function could neutralize the damage caused by a viral infection [9]. In this study, we demonstrated that in addition to biological stresses, CeEngase may also be involved in adaptation to environmental stresses, suggesting that CeEngase downregulation may represent a general stress management strategy for survival.

We honestly thank Haojie Lu and Liming Wei (Institutes of Biomedical Sciences, Fudan University) for suggestions.

An ethics statement was not required for this study type since no human or animal subjects or materials were used.

The authors have no conflicts of interest to declare.

This work was supported by National Key Research and Development Program of China (2021YFC 2700800), the Ministry of Science and Technology, People’s Republic of China (2021YFA 0909300), and National Natural Science Foundation of China (32170828, 32370877).

Conceptualization: Li Chen, Guiqin Sun, Zhiyong Shao, and Xinrong Lu; data acquisition and analysis: Xinrong Lu, Yongliang Tong, Li Chen, Guiqin Sun, Mengting Wu, Shaoxian Lyu, Jiale Fan, Junyu Zheng, Lin Zou, Danfeng Shen, and Lin Rao; writing – original draft: Xinrong Lu and Li Chen; writing – review and editing: Xinrong Lu, Linlin Hou, Cuiying Chen, Xunjia Cheng, Guiqin Sun, Zhiyong Shao, and Li Chen. All authors have read and agreed to the published version of the manuscript.

Xinrong Lu and Yongliang Tong contributed equally to this work.

The authors confirm that the data supporting the findings of this study are openly available within the article and its supplementary materials. Further inquiries can be directed to the corresponding author.