Plant secondary metabolites (PSMs) are chemical compounds that play an important role in defense against herbivores, with likely effects on herbivore physiology and behavior (Freeland and Janzen, 1974; Hughes, 1988). Conversely, herbivores have evolved several strategies in response to PSMs, such as gut microbial detoxification (Jones and Megarrity, 1986; Dearing et al., 2005; Sundset et al., 2010; Johnson et al., 2018). Gastrointestinal microbes play a significant role in host metabolism and are essential modulators of body homeostasis and health (De Filippo et al., 2010; Sommer and Bäckhed, 2016). Gut microbial metabolites, such as short-chain fatty acids (SCFAs), have been shown to mediate the effects of gut microbiota on their hosts (Koh et al., 2016). Therefore, changes induced in the gut microbiota in response to PSM ingestion would have profound effects on herbivores, such as nutrient digestion and absorption and materials metabolism, not exclusively due to gut detoxification. However, the response of herbivore gut microbiota to PSMs has not been thoroughly studied, and the role of the gut microbiota in the interaction between plants and herbivore is thus not yet fully understood.

The PSM 6-methoxybenzoxazolinone (6-MBOA) is mainly produced during the early growth stages of plants belonging to the family Gramineae (Argandoña et al., 1981; Niemeyer, 1988; Acharya et al., 2021). Although 6-MBOA is mainly known to stimulate the reproduction of certain animals (Dai et al., 2016) such as rodents (Berger et al., 1981; Alibhai, 1986; Anderson et al., 1988; Nelson and Shiber, 1990; Martin et al., 2008; Dai et al., 2016) and rabbits (Rodríguez-De Lara et al., 2007), it has diverse effects and acts as a defense compound against insect feeding and digestion (Klun and Brindley, 1966; Campos et al., 1988; Houseman et al., 1992; Dowd and Vega, 1996; Maag et al., 2014), in addition to restraining the growth of microbes such as Escherichia coli, Proteus rulgaris, Cephalosporium gramineum, Fusarium oxysporum, Coprinus comatus, Rhizoctonia solani, and Pythium species (Wang et al., 2001; Wang and Ng, 2002; Martyniuk et al., 2006; Acharya et al., 2021). This metabolite can also alter root-associated microbiota in maize (Hu et al., 2018). Additionally, 6-MBOA can influence animal and human metabolism such as the ameliorating glucose tolerance in diabetic rats (Hameed et al., 2019) and obesity in humans (Fomsgaard et al., 2011). However, to date, little has been known about the effect of 6-MBOA on the gut microbiota in mammalian herbivores and its potential effects on the physiology of mammalian herbivores mediated by the gut microbiota.

Brandt’s vole (Lasiopodomys brandtii) is a small and seasonally reproductive mammal widely distributed in the grasslands of Inner Mongolia, China (Xie et al., 1994; Shi et al., 1999; Wan et al., 2002a). This species feeds on a wide variety of herbal plants, favoring Leymus chinensis, Medicago varia, Stipa krylovii, and Saussurea runcinata (Wang et al., 1992; Li et al., 2019). It has been reported that gut microbiota composition in Brandt’s vole can be affected by multi-factors, such as temperature, housing density, stress, diet, age, and gender (Bo et al., 2019; Li et al., 2020; Liu J. et al., 2020, 2021; Xu and Zhang, 2021). Previously, we detected 6-MBOA in the seedlings of L. chinensis, which is dominant in the grasslands of Inner Mongolia (Wang et al., 1992; Li et al., 2019), with the highest concentration of 6-MBOA exceeding the 100 mg/kg during seedling germination (Dai et al., 2014). Therefore, Brandt’s vole represents an ideal model for studying the effect of 6-MBOA on the structure and function of the gut microbiota in mammalian herbivores.

In this study, we investigated the effects of different doses of 6-MBOA on the body weight growth, cecal SCFAs, alpha and beta diversities, and functions of the cecal microbiota of adult male Brandt’s voles. These results provide new insights into the effect of 6-MBOA on the physiology of mammalian herbivores and the mechanisms of their responses to 6-MBOA. We illustrate the potential role of 6-MBOA in plant-mammalian herbivore interactions and contribute to the theory of coevolution between plants and animals.

A total of 1,211,102 effective sequences were used in abundance and diversity analyses, as well as KEGG pathway comparisons. A total of 1,994 OTUs were identified, with 776-1,236 OTUs in each sample. The numbers of total tags, taxon tags, and OTUs were not significantly different among the three groups (Table 1). The average Goods coverage was as high as 99.63% among the three groups, showing that the majority of bacteria present in the samples were identified in our study (Table 1). The dominant phyla identified were Firmicutes (68.76%), followed by Bacteroidetes (22.69%), Proteobacteria (2.76%), and Spirochaetes (2.15%) (Figure 1).

Statistical analysis using the Kruskal-Wallis test showed that after 15 days of treatment, 6-MBOA had significant effects on Chao1 (P = 0.007), ACE indices (P = 0.016), and observed species (P = 0.035), with each parameter being higher in the low 6-MBOA dose group than in the control group (P = 0.002, 0.004, and 0.011, respectively) (Figures 2A–C). The 6-MBOA had no significant effects on the Shannon (P = 0.113) (Figure 2D) or Simpson indices (P = 0.412) (Figure 2E).

The PERMANOVA results showed that 6-MBOA administration for 15 days significantly altered the beta diversity of the cecal microbial community (F = 1.3778, P = 0.035), with the 6-MBOA dose group being significantly different from the control group (P = 0.027). Additionally, the profile of dissimilarities in the structure of the cecal microbiome after 6-MBOA administration for 15 days was depicted using PCoA of Bray–Curtis dissimilarity (Figure 3A). Principal coordinates 1 and 2 (PCo1 and PCo2) explained 16.97 and 11.98% of the variation, respectively. Although the clustering of the control, low 6-MBOA dose, and high 6-MBOA dose groups was not completely separated in the ordination space, ANOSIM revealed that the divergence in the structure of cecal microbiome between the three groups was greater than the divergence within the low 6-MBOA dose group (R = 0.184, P = 0.038) (Figure 3B).

To identify the differential abundance and categories of the intestinal flora and to identify significantly different biomarkers among the three groups—referring to statistical tests and biological relevance—we used LEfSe (LDA Effect Size) (Score > 2). The enriched biomarkers in the low 6-MBOA dose group (compared to the control group) included the phylum Tenericutes, classes Mollicutes and Negativicutes, order Selenomonadales, families Ruminococcaceae, and Veillonellaceae, genera Quinella, Caproiciproducens, Anaerofilum, Harryflintia, and unidentified Spirochaetaceae. All of these belong to the phylum Firmicutes, except the Mollicutes class (phylum Tenericutes), and the unidentified genus Spirochaetaceae (phylum Spirochaetes) (Figures 4A,B). The enrichment of unidentified order Melainabacteria, families Melainabacteria and Prevotellaceae, and genus Melainabacteria was reduced in the low 6-MBOA dose group (compared to the control group).

Fifteen days of 6-MBOA treatment resulted in a significant change in the relative abundance of the phylum Tenericutes (P = 0.032), with an increase in the low 6-MBOA dose group over the control and high 6-MBOA dose groups (P = 0.020 and 0.030, respectively) (Figure 5A). 6-MBOA also significantly changed the relative abundance of the Mollicutes and Negativicutes classes (P = 0.032 and 0.043, respectively), with increases in the low 6-MBOA dose group over both the control and high 6-MBOA dose groups for the Mollicutes class (P = 0.020 and 0.027, respectively) (Figure 5B and Supplementary Figure 1), and just the control group for the Negativicutes class (P = 0.016) (Figure 5B and Supplementary Figure 1). 6-MBOA significantly altered the relative abundance of the Selenomonadales and unidentified Melainabacteria orders (P = 0.043 and 0.036, respectively), with an increase in the low 6-MBOA dose group over the control (P = 0.016) for Selenomonadales (Figure 5C and Supplementary Figure 2), and a decrease in the low 6-MBOA dose group over the control (P = 0.019) for unidentified Melainabacteria (Figure 5C). 6-MBOA significantly altered the relative abundance of the Ruminococcaceae (P = 0.014) and Veillonellaceae (P = 0.043) families, with an increase in the low 6-MBOA dose group over the control group (P = 0.008 and 0.016, respectively) (Figure 5D and Supplementary Figure 3). The relative abundance of the unidentified Melainabacteria family significantly altered (P = 0.036), with an increase in the low 6-MBOA dose group (compared to the control and high 6-MBOA dose groups) (P = 0.019 and 0.036, respectively) (Figure 5D). The relative abundance of the Prevotellaceae family was also significantly altered (P = 0.022) with a decrease in the low 6-MBOA dose group (compared to the control group) (P = 0.007) (Figure 5D and Supplementary Figure 3). Fifteen-day treatment with 6-MBOA significantly altered the relative abundance of the genera Caproiciproducens, Quinella, unidentified Spirochaetaceae, Anaerofilum and Harryflintia (P = 0.045, 0.040, 0.037, 0.044 and 0.038, respectively), with an increase in the low 6-MBOA dose group (compared to the control group) (P = 0.013, 0.014, 0.014, 0.013, and 0.016, respectively) (Figure 5E and Supplementary Figure 4). The relative abundance of the unidentified genus Melainabacteria was also significantly altered (P = 0.036), with a decrease in both the low and high 6-MBOA dose groups (compared to the control group) (P = 0.019 and 0.036, respectively) (Figure 5E).

The functions of 17 KEGG pathways in the cecal bacterial communities changed following the administration of 6-MBOA for 15 days (Table 2 and Figure 6). Dioxin and xylene degradation (metabolism), and retinol metabolism (metabolism) were more dominant in the cecal bacterial community in the high 6-MBOA dose group than in the control and low 6-MBOA dose groups (P = 0.040, 0.008; P = 0.040, and 0.005; P = 0.045 and 0.004; respectively). Protein digestion and absorption (organismal systems) were more dominant in the control cecal bacterial community than in the low and high 6-MBOA dose groups (P = 0.017 and 0.045, respectively). Cellular antigens (environmental information processing) (P = 0.002), glycosaminoglycan degradation (metabolism) (P = 0.003), glycosphingolipid biosynthesis-ganglio series (metabolism) (P = 0.007), glycosphingolipid biosynthesis-globo series (metabolism) (P = 0.027), N-glycan biosynthesis (metabolism) (P = 0.005), and lipoic acid metabolism (metabolism) (P = 0.003) were more dominant in the control cecal bacterial community than in the low 6-MBOA dose group. Ascorbate and aldarate metabolism (metabolism) (P = 0.013), chloroalkane and chloroalkene degradation (metabolism) (P = 0.008), ethylbenzene degradation (metabolism) (P = 0.004), flavonoid biosynthesis (metabolism) (P = 0.005), inorganic ion transport and metabolism (unclassified) (P = 0.011), inositol phosphate metabolism (metabolism) (P = 0.009), N-glycan biosynthesis (metabolism) (P = 0.027), glycosphingolipid biosynthesis-globo series (metabolism) (P = 0.031), and naphthalene degradation (metabolism) (P = 0.002) were more dominant in the cecal bacterial community in the high 6-MBOA dose group than in the low 6-MBOA dose group.

Fifteen days of 6-MBOA treatment significantly increased the butyrate content (P = 0.008) in the low and high 6-MBOA dose groups (compared to control) (P = 0.016, 0.027, respectively) (Figure 7A), while the contents of other five SCFAs were not significantly affected by 6-MBOA (Figures 7B–F) (P > 0.05). ANOVA showed that 6-MBOA administration for 15 days did not significantly affect the relative food consumption [F_{(2, 21)} = 3.104, P = 0.066] (Figure 8A), while it significantly affected the growth rate in terms of body weight in male Brandt’s voles [F_{(2, 21)} = 5.838, P = 0.010] (Figure 8B). The growth rate in the control group was significantly higher than that in the low and high 6-MBOA dose groups (P = 0.029, P = 0.014) (Figure 8B).

In this study, the dominant phyla were Firmicutes and Bacteroidetes. Consistently, these were also the dominant phyla in the fresh feces of wild Brandt’s voles (Li et al., 2020; Xu and Zhang, 2021). However, the ratio of Firmicutes to Bacteroidetes decreased sharply from 9.7 in the previous study (Li et al., 2020) to approximately 3.0, in the present study. The abundance of Bacteroidetes increased in the guts of captive primates when the dietary fiber content decreased (Clayton et al., 2016). In the study by Li et al. (2020), wild voles feeding on fresh grassland plants yielded a plant fiber content in the diet of approximately 30%. In our study, as the voles were fed rodent chow, which only contained approximately 5% crude fiber, the fiber content in the diet was significantly less than that in the food of wild voles. Therefore, we infer that the large decrease in the fiber content in the diet contributed to the sharp increase in the abundance of Bacteroidetes, and thus the large decrease in the Firmicutes to Bacteroidetes ratio in the gut of voles in our study in comparison to that in the wild voles in the study by Li et al. (2020).

Compared to the control group, the low 6-MBOA dose group showed an increase in observed species, Chao1, and ACE indices of cecal microbiota, as well as an enrichment in biomarker OTUs; furthermore, a separation of the cecal microbiota structure in the low 6-MBOA dose group from the control group were also observed. This suggests that 6-MBOA could enhance the abundance of microbial species in the cecum of Brandt’s vole, thus improving enrichment and altering the structure of the cecal microbiota. However, the differences in these parameters between the high 6-MBOA dose group and the control group were not significant, indicating that the effect of 6-MBOA on the cecal microbiota was dose-dependent.

In the present study, glycan biosynthesis and metabolism functions were concurrently inhibited in the low 6-MBOA dose group, including functions, such as glycosaminoglycan degradation, glycosphingolipid biosynthesis-ganglio series, glycosphingolipid biosynthesis-globo series, and N-glycan biosynthesis. Members of the Prevotellaceae family can activate intestinal gluconeogenesis to ameliorate glucose homeostasis (De Vadder et al., 2016). Veillonellaceae bacteria have been found to be associated with higher serum insulin concentrations (Cheng et al., 2018). In this study, the abundance of Prevotellaceae in the low 6-MBOA dose group was lower than that in the control group, whereas that of Veillonellaceae was higher in the low 6-MBOA dose group. 6-MBOA was reported to improve glucose tolerance in diabetic rats (Hameed et al., 2019). Thus, we infer that 6-MBOA administration can vary the enrichment of pathways pertaining to glucose metabolism by reducing the abundance of Prevotellaceae and increasing the abundance of Veillonellaceae to help regulate glucose homeostasis in voles. Enrichment of bacteria belonging to the phylum Tenericutes was found in healthy individuals (compared to that in patients with metabolic syndromes, such as those with obesity) (Lim et al., 2017). Lower abundance of Anaerofilum has been found in obese patients (Del Chierico et al., 2018; Koo et al., 2019). In this study, the abundance of Tenericutes and Anaerofilum in the low 6-MBOA dose group was higher than that in the control group. Thus, enhancement of Tenericutes and Anaerofilum abundance in the cecum induced by 6-MBOA may contribute to body mass control in voles.

Negativicutes bacteria can produce propionate (Hino and Kuroda, 1993). Selenomonadales bacteria belonging to the class Negativicutes have been reported to ferment carbohydrates into acetate (Vargas et al., 2017). Veillonellaceae bacteria belonging to the order Selenomonadales are also thought to be related to fatty acid metabolism (Zeng et al., 2019). Quinella bacteria belonging to the Veillonellaceae family can produce propionate (Kittelmann et al., 2014). Mollicutes bacteria belonging to the phylum Tenericutes also help produce acetate (Sapountzis et al., 2018). Propionate produced by microbial fermentation in the large intestine may contribute to human health maintenance (Reichardt et al., 2014). Acetate aids host substrate metabolism by enabling the secretion of gut hormones and is also involved in the regulation of glucose homeostasis and appetite (Hernández et al., 2019). In the present study, the abundance of classes Mollicutes and Negativicutes, order Selenomonadales, family Veillonellaceae, and genus Quinella was concurrently higher in the low-dose group than in the control group. Meanwhile, increased acetate and propionate content was detected in the low 6-MBOA dose group (compared to the control group), although the differences were not significant. Thus, we inferred that 6-MBOA can enhance the abundance of these bacteria in the cecum to help regulate the production of acetate and propionate and thus benefit overall health. As Caproiciproducens can produce butyrate (Bengelsdorf et al., 2019; Flaiz et al., 2020), its higher abundance may partially contribute to the higher butyrate content in the cecum of the low 6-MBOA dose group. Besides, butyrate-producing Butyrivibrio (Kim et al., 2020) and Acetatifactor (Pfeiffer et al., 2012) enriched in low and high 6-MBOA dose group, respectively (not significantly; Supplementary Figure 4) may also help enhance the product of butyrate in the cecum. Butyrate is beneficial to humans and animals because of its anti-inflammatory properties (Sossai, 2012). Therefore, we speculate that 6-MBOA improves the health of voles by enhancing the butyrate levels in the cecal microbiota. Caproiciproducens, Anaerofilum, and Harryflintia belong to the family Ruminococcaceae, and therefore the increased abundance of these three genera in the low 6-MBOA dose group contributed to the higher abundance of the Ruminococcaceae family in the low 6-MBOA dose group (compared to the control group).

Microbiotas in the large intestine that ferment proteins are important for protein metabolism (Xie et al., 2020). In the present study, both protein digestion and absorption were down-regulated in low and high 6-MBOA dose groups, which corresponds to inhibitory effect of 6-MBOA on protein digestion in Ostrinia nubilalis (Houseman et al., 1992). The abundance of the phylum Tenericutes in the ileal of pigs decreased as the dietary protein content reduced (Qiu et al., 2018). The abundance of the family Prevotellaceae in the rumen of Bos indicus decreased with the protein supplement (Latham et al., 2018). Consistently, the abundance of Tenericutes increased while that of the family Prevotellaceae reduced, which hints that the content of undigested protein left in the cecum by the 6-MBOA administration was high. In a parallel study, we observed that 6-MBOA significantly inhibited the in vitro activity of the intestinal trypsin in male Brandt’s voles and increased the nitrogen content of male Brandt’s voles feces (unpublished). Altogether, 6-MBOA could inhibit the protein digestion in Brandt’s voles. Moreover, in the 6-MBOA groups, the growth rate was reduced, with an observable food consumption decrease. Acetate, butyrate and propionate are known to stimulate the release of peptide YY inhibiting appetite (Karaki et al., 2006; Charrier et al., 2013; Hernández et al., 2019). The concentration of serum ghrelin was also reduced by 6-MBOA administration in Brandt’s voles in the parallel study (unpublished). Thus, we infer that restricted protein digestion, reduced serum ghrelin, and the enhanced SCFAs in cecum by 6-MBOA may together contribute to the reduction in food intake. Altogether these results indicate that 6-MBOA administration may lead to a reduced growth rate. Altered abundance of cecum microbiota related to glucose homeostasis regulation and body mass in response to 6-MBOA might also contribute to the reduced growth rate. Consistently, 6-MBOA-containing cereal grain products were used for obesity treatment in human (Fomsgaard et al., 2011). The restricted the protein digestion and absorption of the cecal microbiota, the decreased trend of relative food consumption, and the reduced growth rate in Brandt’s vole suggest that 6-MBOA can potentially act as a digestion-inhibiting PSM in the interaction between mammalian herbivores and plants.

Gut microbes can help the human body transform industrial chemicals and pollutants, alter their toxicities and half-lives, and are linked to health benefits (Koppel et al., 2017). Nitroreductases in the cecal contents of Sprague–Dawley rats produced by cecum microbiota can transform the xenobiotic nitrazepam (Takeno and Sakai, 1991). Woodrats (Neotoma lepida) consuming diets containing creosote resin (a PSM with aromatic rings) harbored microbes with a higher abundance of genes associated with the metabolism of aromatic compounds (Kohl et al., 2014). Notably, 6-MBOA also has one aromatic ring. Xylene, dioxin, naphthalene, and ethylbenzene are aromatic xenobiotics for animals. Therefore, the upregulated degradation of xylene and dioxin in the high 6-MBOA dose group (compared to that in the control group), and the enhanced degradation of dioxin, xylene, naphthalene, and ethylbenzene in the high 6-MBOA dose group (compared to that in the low 6-MBOA dose group), indicate that 6-MBOA induced a change in the metabolism of aromatic xenobiotics in the cecal bacteria and that the intensity of the change was dose-dependent. It has been reported that 6-MBOA can be transformed by fungi such as Fusarium moniliiforme and Gaumannomyces graminis (Fomsgaard et al., 2004) to N-(2-hydroxy-4-methoxyphenyl) malonamic acid (HMPMA), and caterpillars (Spodoptera littoralis and S. frugiperda) to MBOA-N-Glc in the gut (Maag et al., 2014). Therefore, intense metabolism of aromatic xenobiotics by cecum microbiota in the high 6-MBOA dose group might result in more metabolized 6-MBOA. This effect could partially explain why a high dose of 6-MBOA did not significantly influence the cecal microbiota community and indicate that the cecum microbiota can metabolize 6-MBOA. This also implies that a high dose of 6-MBOA might be harmful to voles due to the reinforced protective mechanism of xenobiotics in the cecal microbiota. We infer that the abundance of bacteria enhanced by 6-MBOA in the present study should be 6-MBOA tolerant or involved in the 6-MBOA metabolism. However, the exact cecal bacteria and mechanism by which 6-MBOA is metabolized in Brandt’s voles need more in-depth research. Reduced abundances in unidentified Melainabacteria, family of the unidentified Melainabacteria and Prevotellaceae, and the unidentified genus Melainabacteria suggest that those bacteria are susceptible to the administration of 6-MBOA, which verifies the antimicrobial properties of 6-MBOA (Wang et al., 2001; Wang and Ng, 2002; Martyniuk et al., 2006). Chloroalkane and chloroalkene degradation, a pathway also involved in xenobiotic biodegradation and metabolism (Liu F. et al., 2021), was consistently observed to be dose-dependently improved by 6-MBOA administration in cecal microbiota of Brandt’s voles.

High doses of 6-MBOA, which exerts negative effects on the growth, food consumption, and protein digestion and absorption of cecal microbiota, and stimulates the degradation of xenobiotics in cecal microbiota without improving enrichment or altering the structure, hint that daily ingestion of more than 100 μg of 6-MBOA may diminish its benefit and exert potential harmful effect on Brandt’s voles. Desulfovibrio is known as sulfate-reducing bacteria, which is responsible for infections and diarrheas in mammals (Velasco-Galilea et al., 2020). Helicobacter is associated with severe gastric disease (Ferreira et al., 2018). Decrease in Roseburia abundance may damage various metabolic pathways and incur irritable bowel syndrome (Tamanai-Shacoori et al., 2017). The abundance of Desulfovibrio and Helicobacter enhanced and Roseburia reduced in high 6-MBOA dose group (not significantly, Supplementary Figure 4) furthermore indicates that high dose of 6-MBOA may impair the health of Brandt’s voles. Consistently, further intraperitoneal injection of 6-MBOA conversely weakened its stimulating effect on the reproduction of male Brandt’s vole under short photoperiods (Dai et al., 2016). Thus, we propose that 6-MBOA could protect L. chinensis from Brandt’s voles if the daily 6-MBOA intake dramatically exceeds 100 μg. Further ingestion of 6-MBOA by L. chinensis may impair voles in grasslands. These may indicate that 6-MBOA could act as a defensive secondary metabolite against mammalian herbivores in L. chinensis seedlings (Dai et al., 2014). In contrast, our results demonstrate that cecal bacteria can assist Brandt’s voles in eliminating ingested xenobiotics, which confirms the detoxifying strategy of the gut microbiota during the adaptation of herbivores to PSMs in the coevolution between plants and herbivores (Dearing et al., 2005).

In summary, our study demonstrated that 6-MBOA altered the enrichment of the microbial community and the abundance of bacteria in the cecum of Brandt’s voles, thereby inducing changes in the community structure of the cecal microbiota, SCFA content, and pathway enrichment. 6-MBOA mainly affected functions related to glucose metabolism, protein digestion and absorption, and body weight control of cecal microbiota in male Brandt’s voles. Metabolism of xenobiotics dose-dependently stimulated by 6-MBOA hints that cecal microbiota can aid the elimination of ingested xenobiotics in male Brandt’s voles. The digestive functions of the cecal microbiota and body growth rate restriction due to 6-MBOA administration suggest that in addition to its role as a reproduction stimulator in mammalian herbivores, 6-MBOA can potentially act as a digestion-inhibiting (PSMs) in the interaction between mammalian herbivores and plants.

The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author/s.

The animal study was reviewed and approved by the Animal Care and Use Committee of the Faculty of Veterinary Medicine of Yangzhou University.

WW supplied the funding and designed the experiment. SY supplied the funding and revised the manuscript. XD performed the statistical analyses and wrote the manuscript. LC, ML, YL, SJ, and TX performed the experiment. AW provided technological support for this experiment. All authors contributed to the article and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.