Gut Bacterial Communities of Lymantria xylina and Their Associations with Host Development and Diet

The gut microbiota of insects has a wide range of effects on host nutrition, physiology, and behavior. The structure of gut microbiota may also be shaped by their environment, causing them to adjust to their hosts; thus, the objective of this study was to examine variations in the morphological traits and gut microbiota of Lymantria xylina in response to natural and artificial diets using high-throughput sequencing. Regarding morphology, the head widths for larvae fed on a sterilized artificial diet were smaller than for larvae fed on a non-sterilized host-plant diet in the early instars. The gut microbiota diversity of L. xylina fed on different diets varied significantly, but did not change during different development periods. This seemed to indicate that vertical inheritance occurred in L. xylina mutualistic symbionts. Acinetobacter and Enterococcus were dominant in/on eggs. In the first instar larvae, Acinetobacter accounted for 33.52% of the sterilized artificial diet treatment, while Enterococcus (67.88%) was the predominant bacteria for the non-sterilized host-plant diet treatment. Gut microbe structures were adapted to both diets through vertical inheritance and self-regulation. This study clarified the impacts of microbial symbiosis on L. xylina and might provide new possibilities for improving the control of these bacteria.


Introduction
The symbiotic association between bacteria and insects is a well-known and universal phenomenon, and is important in the biological processes of host insects. In this symbiotic relationship, the gut microbiota is essential for maintaining insect health [1,2]. The intestinal bacterial community usually provides metabolic benefits to the host through the production of digestive enzymes and vitamins, thereby improving nutrient absorption [3]. In stinkbugs of the family Plataspidae, gut bacteria were vertically transmitted as a 'symbiont capsule'. When the 'symbiont capsule' was removed, the host insect showed high mortality rates [4]. The intestinal symbiotic bacteria of the oriental fruit fly Bactrocera dorsalis (Hendel) adjust the ecology and physiology of the host insect and even help host insects improve their resistance to insecticides [5].
In recent microbial studies, diet and environment were considered to be the main factors in the formation of gut microbiota in insects [6]. For example, xylophagous insects had more complex gut microbiota communities, while insects that fed on sap, such as aphids and psyllids, had the least complex gut microbiota communities [7]. The gut The potential effect of gut microbiota on L. xylina development was studied using aseptic feeding experiments. Collected eggs were assigned to feed in two diet-groups: a non-sterilized C. equisetifolia branch (NSC) and a sterilized artificial diet mainly composed of a C. equisetifolia branch (SAC). For the sterilized artificial feeding experiment (feeding on SAC), eggs were disinfected by an incubation in 75% ethanol for 3 min, then washed with sterile water 2-3 times to remove surface microbiota. To simulate natural feeding (feeding on NSC), other eggs were unsterilized. The eggs from the different treatment groups were transferred to high-temperature sterilized (121 • C, 20 min) closed transparent plastic (Polycarbonate) boxes (the upper mouth of the plastic boxes measured 17.2 × 11.3 cm, while the lower mouth measured 12.2 × 8.5 × 8.1 cm), and were stored in an incubator (MIR-154-PC Incubator, Panasonic, Kadoma, Japan) at 26 • C and 80% relative humidity with a photoperiod of 12:12 (light/dark) to break the egg diapause. After the hatching of first instar larvae, they were grouped into 5 boxes/diet groups, each containing 50 larvae from one of the two treatment groups. These were maintained in different incubators under the same conditions with different diets. In the NSC treatment, the first instar larvae were transferred to a breathable transparent plastic box, and the NSC diet was changed daily. In the SAC treatment, the first instar larvae were transferred to a new high-temperature sterilized transparent closed plastic box and fed on SAC. In order to maintain ventilation and sterility, air exchange and feed changeover were carried out every day in the sterile environment of the AlphaClean1300 system (AlphaClean1300, HealForce, Shanghai, China). All groups were fed until pupation. The artificial SAC diet was introduced in a sterile environment after high-temperature sterilization (121 • C, 20 min) of the food, while the diet formula was the modified feed A diet described by Shen and colleagues: sawdust from C. equisetifolia branches (147 g), sorbic acid (2 g), vitamin mix (vitamins A, E, B1, B2, B3, B5, B6, B9, B12, C, H, chloride, inositol) (10 g), P-nitro benzoic acid (1 g), ferric citrate ( 0.0428 g), agar (15 g), and distilled water (1 L) [10]. Molting was taken as the age classification standard. Pictures were taken of the first molting larvae by fluorescence stereomicroscope (M205FA, Leica, Wetzlar, Germany). The head widths and body lengths of L. xylina larvae were measured by 'Measurement Tool' (Leica Application Suite Version 4.8.0). A minimum of 10 larvae per sample were measured for biological repetitions.
2.3. DNA Extraction, Bacterial 16s rRNA Gene Amplification, and High-Throughput Sequencing The insect bodies were disinfected with 70% ethanol for 3 min, and then washed twice in asepsis. In addition, the method proposed by Hu et al. was used to dissect and collect intestinal bacteria [15], whereby 10 µL 10 mM sterile phosphate-buffered saline (138 mM NaCl and 2.7 mM KCl, pH 7.4) was placed in a sterile culture dish. The posterior end of the larvae was cut off using alcohol-sterilized surgical scissors. The gut was pulled out under a stereomicroscope with sterile insect pins, transferred into a 1.5 mL centrifuge tube, and stored at −80 • C (MDF-DU502VXL refrigerator, Panasonic, JPN) for subsequent DNA extraction. All procedures were performed in a sterile environment. Non-sterilized eggs, disinfected eggs, and guts of larvae (seven instars), pupae (5th day), and adults were transferred into individual tubes for DNA extraction separately. There were 3 samples in each developmental stage, and a total of 57 samples were obtained (one gut = one sample). The bacterial DNA of all samples was extracted by the CTAB method. The DNA extraction of pupal gut bacteria fed on SAC failed. Purified bacterial DNA was diluted to 1 ng/µL with ddH 2 O. We then used the diluted genomic DNA as a template, and the V4 regions of the bacteria 16S ribosomal RNA genes were amplified by PCR using the following primers with barcodes: 515F 5 -GTGCCAGCMGCCGCGGTAA-3 and 806R 5 -GGACTACHVGGGTWTCTAAT-3. PCR reactions were carried out in 30 µL reaction solution with 15 µL of Phusion High-Fidelity PCR Master Mix (New England Biolabs, Beijing, China), 0.2 µM of forward and reverse primers, and about 10 ng template DNA. Thermal cycling consisted of initial denaturation at 98 • C for 1 min, followed by 30 cycles of denaturation at 98 • C for 10 s, annealing at 50 • C for 30 s, elongation at 72 • C for 30 s, and finally at 72 • C for 5 min. PCR products were purified, and the quality was assessed using a Qubit 2.0 Fluorometer. The amplificates were sequenced on an Ion S5TM XL (Thermofisher, Shanghai, China) platform was used. Samples were examined for contamination with a blank PCR control (sterile water) before sequencing. The library was constructed and sequenced by Novogene Technology (Fujian, China).

Division of Larval Instars
Regression analysis was performed on the indices measured on each instar larva to verify the rationality of instar division. The Brooks and Crosby indices were calculated according to Dyar's law and Crosby's growth law [16]. The Brooks index was calculated using the following formula: where n represents the larvae's instar; X n and X n−1 represent the means of the body length and head capsule width of the n th instar and n−1 th instar larvae, respectively. The Crosby index was calculated using the following formula: where n represents the larvae's instar; b n and b n−1 represent the Brooks index values of the n th and n−1 th instar larvae, respectively [17]. When the coefficient of variation is <20% or the Crosby index is <10%, the instar division is reliable.

Bioinformatic Analysis
To obtain raw reads, low-quality fragments, the barcode, and primer sequences were cut and filtered using Cutadapt (V1.9.1) [18]. The raw reads sequences were compared with the species annotation database to detect and remove the chimera sequence in order to obtain clean reads [19,20]. All clean reads of all samples were clustered using Uparse [21]. Sequences were clustered into operational taxonomic units (OTUs) with 97% similarity using Mothur. Annotation of the sequences was performed in the SSUrRNA database (threshold of 0.8-1) to obtain taxonomic information [22]. OTU abundance information was normalized using a standard sequence number corresponding to the sample with the least sequences. All Illumina reads were deposited in the NCBI Short Read Archive database (accession number: PRJNA660917).

Gut Microbiota Composition of L. xylina on Different Diets
From the sequence results, we analyzed the complexity of the species diversity for each sample through 4 indices, namely the Chao1, Shannon, Simpson, and ACE indices. All of the indices used for our samples were calculated with QIIME (Version 1.7.0). The Mothur method and the SSUrRNA database [23] of SILVA132 (http://www.arb-silva.de/ (accessed on 25 March 2019)) [24] were used to perform species annotation analysis (with threshold values of 0.8~1) to obtain taxonomic information and classification level counts in order to determine the community composition of each sample. This was conducted to show the diversity of the gut microbiota at the phylum and genus levels in L. xylina fed on different diets, using GraphPad Prism software (Version 6.01). This study focused on the most richly represented taxa, whereby the top 10 phyla and genera were selected for bacterial community analysis based on their relative abundance. Sequences from other than the first 10 phyla or genera were referred to as 'others'. We also compared beta diversity values to evaluate differences between samples in terms of species-diversity complexity using the weighted Unifrac distance metric and QIIME software (Version 1.7.0). We used LEfSe analysis to analyze the differences in species abundance data between groups using the rank sum test method, implemented a dimensionality reduction through LDA to evaluate the impacts of different species, and finally drew a histogram of the distribution of LDA values and evolutionary clades of different species. Principal coordinate analysis (PCoA) was performed to obtain principal coordinates from complex, multidimensional data.

Functional Annotation
We extracted the KEGG database prokaryotic whole genome 16S rRNA gene sequence and then applied the BLASTN algorithm to compare it to the SILVA SSU Ref NR database (BLAST bitscore > 1500) to establish the correlation matrix. The KEGG database prokaryotic annotation process was performed using UProC and PAUDA. The genomic function information was calibrated to the SILVA database to allow the SILVA database function annotation. Sequencing samples were clustered using the SILVA database sequence as the reference sequence, and then functional annotation information was obtained. According to the annotation results, we selected the grouping at the highest abundance of each annotation level to generate a functional relative abundance column stacking chart, so as to visually view the relatively high abundance of each sample at different annotation levels of functions and their proportions.

Statistical Analyses
The Shapiro-Wilk test was used to evaluate the normality of the data. The Kruskal-Wallis test was used for analysis of potentially significant differences for non-normality data, while the Student's t-test was used for normality data. Multi-sample comparisons were performed in accordance with the normal distribution through one-way analysis of variance (ANOVA). These statistical tests were analyzed using SPSS (Version 21.0.0).

Division of Larval Instars
A total of 214 L. xylina larvae were measured, in which process the head capsule widths and body length values were recorded. According to the variation index, the coefficient of variation was <20% for the head capsule width, but was >20% for the body length in the fifth and seventh instar (Tables 1 and A1); therefore, head capsule width was more reliable as an indicator of the larval instar of L. xylina than body length. Except for the fifth instar, the body lengths were significantly different between larvae fed on NSC and and those fed on SAC (p < 0.05). The head capsule widths did not differ significantly (p > 0.05) between the NSC and SAC treatments for the first and second instar larvae, while from the third to fifth instars, the head capsule widths of the L. xylina larvae fed on NSC were significantly higher than the larvae fed on SAC (p < 0.05). There were no significant differences in the width of the head capsules of the sixth instar L. xylina larvae between NSC and SAC diets (t-test, p = 0.605) ( Table 1). Note: A1-A7: First to seventh instar larvae of L. xylina under the SAC treatment; N1-N7: first to seventh instar larvae of L. xylina under the NSC treatment; p value: the t-test was used to compare the differences in larvae fed on two diets at the same instar; *: p < 0.05; **: p < 0.01; ***: p < 0.001; data were the means ± standard error.

Bacterial Community Structures and OTUs
To describe the intestinal bacterial community structure of L. xylina, 57 samples were sequenced. After quality control, an average of 75,329 valid data points were obtained for each sample. The quality control efficiency reached 94.49%. This sequence was aggregated into OTUs with 97% identity, and a total of 3073 OTUs were obtained. Pairs of samples between the different diets at each stage shared some common OTUs, with L. xylina larvae sharing the most OTUs with 1328 OTUs (Figure 1). There was no significant difference between L. xylina eggs and adults' OTU numbers on either diet (p > 0.05), while the OTUs of L. xylina larvae fed on NSC were lower than those that received the SAC treatment. The total number of OTUs for L. xylina fed on SAC (2563 OTUs) was significantly higher than on the NSC diet (2178 OTUs) (t-test, df = 52, p = 0.009), with 1668 OTUs in common.
The 3073 OTUs reported in this study were assigned to phylum 38, class 58, order 131, family 246, genus 619, and species 432. At the phylum level, Firmicutes was predominant, with a mean relative frequency of 66.4%, followed by Proteobacteria (25.9%) and Actinobacteria (20.6%). At the genus level, the Enterococcus genus was predominant, with a mean relative frequency of 55.0%, followed by Acinetobacter (11.0%) and Weissella (6.6%). The proportions of Enterococcus in the seven instars of larvae fed on NSC were 67.9%, 63.9%, 36.7%, 23.2%, 68.2%, 90.2%, and 61.6%, while for larvae fed on NSC, the values were 0.8%, 51.4%, 76.3%, 34.6%, 59.0%, 88.98%, and 89.90%. Both groups showed a trend of decreasing first and then increasing; however, for larvae fed on SAC, the most abundant gut bacteria of the first instar larvae were Acinetobacter (33.5%), followed by Sphingomonas (12.5%) and Acidovorax (12.1%) (Figure 2). between the different diets at each stage shared some common OTUs, with L. xylina larvae sharing the most OTUs with 1328 OTUs (Figure 1). There was no significant difference between L. xylina eggs and adults' OTU numbers on either diet (p > 0.05), while the OTUs of L. xylina larvae fed on NSC were lower than those that received the SAC treatment. The total number of OTUs for L. xylina fed on SAC (2563 OTUs) was significantly higher than on the NSC diet (2178 OTUs) (t-test, df = 52, p = 0.009), with 1668 OTUs in common. The 3073 OTUs reported in this study were assigned to phylum 38, class 58, order 131, family 246, genus 619, and species 432. At the phylum level, Firmicutes was predominant, with a mean relative frequency of 66.4%, followed by Proteobacteria (25.9%) and Actinobacteria (20.6%). At the genus level, the Enterococcus genus was predominant, with a mean relative frequency of 55.0%, followed by Acinetobacter (11.0%) and Weissella (6.6%). The proportions of Enterococcus in the seven instars of larvae fed on NSC were 67.9%, 63.9%, 36.7%, 23.2%, 68.2%, 90.2%, and 61.6%, while for larvae fed on NSC, the values were 0.8%, 51.4%, 76.3%, 34.6%, 59.0%, 88.98%, and 89.90%. Both groups showed a trend of decreasing first and then increasing; however, for larvae fed on SAC, the most abundant gut bacteria of the first instar larvae were Acinetobacter (33.5%), followed by Sphingomonas (12.5%) and Acidovorax (12.1%) ( Figure 2).    Table A2). The relative abundance levels of OTU members were different during the different L. xylina development stages on both SAC and NSC diets. The proportion of Proteobacteria (47.23%) was larger than Firmicutes (44.33%) for L. xylina eggs; however, during the development of the larvae into adults, the proportion of Firmicutes gradually increased to become the most dominant population and reached the highest value at the sixth instar (90.29%), which was significantly higher than in the adults and in larvae at other instar points (t-test, df = 52, p = 0.000) (Table A3). At the genus level, the dominant bacteria from eggs were Enterococcus (23.98%), followed by Acinetobacter (19.61%) and Weissella (17.48%), while the proportion of Enterococcus increased (except for in the fourth instar larvae) with the progression of instars for the L. xylina larvae. In the gut of L. xylina larvae, Enterococcus accounted for the highest proportion of microbiota (58.05%), followed by Acinetobacter (6.15%) and Weissella (5.42%). In the gut of L. xylina adults, Enterococcus was also the most dominant microbiota (79.91%), followed by Acinetobacter (2.46%) and Acidovorax (0.83%) (Table A4).

The Diversity Differences for Bacterial Communities with Different Development Diets for L. xylina
There were no significant differences between the Chao1 values of microbiota from the disinfected L. xylina eggs and those of microbiota from non-sterilized eggs (t-test, df = 4, p = 0.301). The gut microbiota OTU numbers for L. xylina fed on SAC (ACE = 521.6) during the larval stages were higher than for those fed on NSC (ACE = 308.8), while the third instar larvae of L. xylina fed on SAC contained the largest number of OTUs (1108 OTUs, Chao1 = 718.128). In the adult stage, there were no significant differences between the OTU numbers of L. xylina bacteria on NSC and those on SAC diets (t-test, df = 40, p = 0.527) (Figure 3).
At the phylum level, on comparing the abundance of gut microbiota of L. xylina fed on different diets, the most dominant microbiota in/on non-sterilized eggs was Proteobacteria (75.57%), which was significantly more abundant than Firmicutes (11.94%). The abundance of Firmicutes (98.34%) increased significantly for first instar larvae fed on NSC (t-test, df = 4, p = 0.011). In contrast, the most abundant phylum in/on disinfected eggs of L. xylina was Firmicutes (76.7%). The relative abundance of Firmicutes decreased significantly (t-test, df = 4, p = 0.018) for the first instar larvae fed on SAC, and Proteobacteria became the most dominant bacteria (Figure 4a). At the genus level, Enterococcus was the most abundant in all samples, accounting for an average of 56.7% of each sample, followed by Acinetobacter (7.2%) and Weissella (6.2%). Enterococcus was also the most abundant bacteria in/on disinfected eggs (37.9%), followed by Weissella (35%) and Acinetobacter (13.7%). At the phylum level, with the increase in age, the diversity of the intestinal microbiota in the NSC and SAC treatments were the same at the genus level, showing a trend of decreasing first and then increasing. There was no significant difference in the gut microbiota of L. xylina adults on different diets (p > 0.05) (Figure 4b).
4, p = 0.301). The gut microbiota OTU numbers for L. xylina fed on SAC (ACE = 5 during the larval stages were higher than for those fed on NSC (ACE = 308.8), whil third instar larvae of L. xylina fed on SAC contained the largest number of OTUs ( OTUs, Chao1 = 718.128). In the adult stage, there were no significant differences betw the OTU numbers of L. xylina bacteria on NSC and those on SAC diets (t-test, df = 40 0.527) (Figure 3). At the phylum level, on comparing the abundance of gut microbiota of L. xylin on different diets, the most dominant microbiota in/on non-sterilized eggs was Proteo teria (75.57%), which was significantly more abundant than Firmicutes (11.94%). abundance of Firmicutes (98.34%) increased significantly for first instar larvae fed on (t-test, df = 4, p = 0.011). In contrast, the most abundant phylum in/on disinfected eg L. xylina was Firmicutes (76.7%). The relative abundance of Firmicutes decreased sig cantly (t-test, df = 4, p = 0.018) for the first instar larvae fed on SAC, and Proteobac became the most dominant bacteria (Figure 4a). At the genus level, Enterococcus wa most abundant in all samples, accounting for an average of 56.7% of each sample, follo by Acinetobacter (7.2%) and Weissella (6.2%). Enterococcus was also the most abundant teria in/on disinfected eggs (37.9%), followed by Weissella (35%) and Acinetobacter (13 At the phylum level, with the increase in age, the diversity of the intestinal microbio  According to the results of the linear discriminant analysis (LDA), the main differences caused by different diet treatments are shown in Figure 5. The generic biomarkers were Sphingomonas in L. xylina larvae on the SAC diet and Weissella and Providencia in L. xylina larvae on the NSC diet. According to the results of the linear discriminant analysis (LDA), the main differences caused by different diet treatments are shown in Figure 5. The generic biomarkers were Sphingomonas in L. xylina larvae on the SAC diet and Weissella and Providencia in L. xylina larvae on the NSC diet.
bacterial phyla in/on the L. xylina's eggs and of larvae in the first instar fed on different diets; (b) variations in bacterial communities in/on the L. xylina's eggs and across seven instars for larvae fed on different diets at the genus level. The bars marked with different letters for the same bacterial phylum are significantly different in their abundance based on statistical analysis (p < 0.05).
According to the results of the linear discriminant analysis (LDA), the main differences caused by different diet treatments are shown in Figure 5. The generic biomarkers were Sphingomonas in L. xylina larvae on the SAC diet and Weissella and Providencia in L. xylina larvae on the NSC diet.

Discussion
The symbiotic association between bacteria and insects has attracted extensive attention. In our study, the gut microbiota of L. xylina fed on different diets were significantly different, and the diversity of the gut microbiota of L. xylina fed on SAC (2563 OTUs) was more abundant than for L. xylina fed on NSC (2178 OTUs). The associated bacteria of L. xylina fed on SAC were extremely abundant; however, with the effects of the bacteria from the host plant and environment, the bacterial community became simpler; this might have been related to bacterial recruitment to balance the benefits and costs associated with environmental acquisition, which was associated with higher growth rates and higher metabolic costs. However, the bacteria numbers detected could be reduced because the result was the relative abundance of bacteria detected by high-throughput sequencing.
The symbiotic association was due to the influence of diet, but also may have been related to vertical transmission. As there is currently little evidence concerning the presence of insect-specific gut bacteria in/on eggs, it is difficult to determine whether bacteria in the gut is spread by vertical or horizontal transmission. The use of eggs as vectors for vertical transmission of Lepidoptera remains speculative [25]. Acinetobacter was the dominant microbiota for non-sterilized eggs. Acinetobacter was also the most dominant microbiota in/on Cnaphalocrocis medinalis eggs and first instar larvae [22]. However, Enterococcus was the dominant bacteria for disinfected eggs. Disinfecting the eggs with ethanol was not sufficient to kill all bacteria. The bacterial community structure was also affected by the varying susceptibilities of bacteria to ethanol. Enterococcus could live long-term on environmental surfaces, and were tolerant to heat and some alcohol preparations [26]. The relative abundance of Acinetobacter in the first instar larvae on the SAC diet was significantly higher than those on the NSC diet. The predominance of Acinetobacter may have been caused by maternal inheritance. Acinetobacter in the intestinal tract of fifth instar larvae of H. armigera shows strong esterase activity, which promotes the metabolism of the insecticide cypermethrin, thereby enhancing insect resistance [27]. Additionally, Acinetobacter, which is also isolated in the intestine of Plutella xylostella larvae, may inhibit Phytophthora capsici and improve the nitrogen-fixation capacity and phosphorus content [28]. The presence of diazotrophic bacteria in the gut tract of insects is helpful in promoting nitrogen absorption and enhancing immunity [29]. In the SAC treatment, larvae seemed to obtain gut microbes through vertical transmission, while in the NSC treatment, larvae seemed to obtain gut microbes through their diet and environment. The phenomenon of the vertical transmission of gut bacteria was also reported by Hosokawa and colleagues in their research on stinkbugs [4]. In Pediculus humanus, the intestinal commensal bacteria Candidatus Riesia pediculicola leave the host insect through a hole in the symbiotic spores, gather in the lateral oviducts, and enter the eggs when the eggs are discharged into the lateral oviducts [30].
In this study, in the early stages of development, the head widths of L. xylina larvae fed on SAC were narrower than those of larvae fed on NSC, which may have been because L. xylina larvae without bacteria from the host plant could not effectively obtain nutrients from the host plant. After the fifth instar, the head width growth rate for the larvae fed on the SAC diet was faster than for those fed on the NSC diet. Until the 7th instar, there were no significant differences between the respective head widths of L. xylina larvae fed on the two different diets. L. xylina had higher growth rates when fed on host plants than when fed on the artificial diet in the short term. This study suggested that L. xylina without bacteria derived from their diets needed more time to adjust their gut bacterial community structure and adapt to the environment [10]. The surfaces of newly deposited eggs were also a symbiotic bacteria resource for the first instar larva [31]. The destruction of the symbiotic bacteria by disinfecting the surfaces of newly deposited eggs caused the hosts to suffer from growth retardation, lower reproduction success rates, and higher mortality [32,33]. The change of the intestinal bacterial community structure from Diabrotica virgifera virgifera helped promote the rapid adaptation of insects in a managed ecosystem [34]. When Pieris canidia was treated with antibiotics for gut-associated bacteria, the host's weight was reduced after the treatment, and the effect increased with increased concentrations of the antibiotics [35]. The gut bacterial communities resulting from NSC and SAC treatments were assessed to determine the relationships between the gut microbiota, diet, and the selfregulation ability of L. xylina. Changes of the gut microbiota from Lepidoptera have been determined via different factors including diet, the environment, gut physiology, and insect development stages, which could work alone or in concert [36]. There were no significant differences in the gut microbiota of L. xylina after the second instar, whether fed on NSC or SAC. A previous study focusing on the herbivore Spodoptera littoralis showed highly conservative intestinal bacterial community structures and member compositions when fed on different diets at the same developmental stages and under uniform conditions [37]. The gut microbiota of L. xylina shared some common taxa and similar community structures between the different diets at the same instars; this phenomenon has also been found for other herbivores [38]. The main bacteria for L. xylina were Enterococcus, Acinetobacter, and Weissella at the genus level; these bacteria have also been found present in 70% of other Lepidoptera insects [25]. Enterococcus comprises the largest proportion of gut microbiota of Helicoverpa armigera (Hubner), whether in the laboratory or in the field [39]. Enterococcus have also been found in L. dispar fed on different diets [40]. The gut microbiota of L. xylina larvae in the current study were mainly Enterococcus. Enterococcus is also the most abundant bacteria associated with two other larva types from Lepidoptera (African cotton leaf worm and H. armigera) [41]. Enterococcus is commonly found in a wide range of insect gut communities, benefiting the health and growth of the insect via various functions, including B vitamin biosynthesis, pheromone production, and the degradation of hostplant compounds [42]. Gut-associated Enterococcus may reduce the pH value of the gut microenvironment, enabling L. dispar to maintain acid-base balance [43]. Enterococcus may also protect the gut against pathogenic toxins that are activated in alkaline conditions by adapting the pH, which could improve gut immunity [44,45]. As a lactic acid bacterium, Weissella is responsible for the fermentation of food in the intestines, and directly affects the production of organic acids, esters, and alcohols.
According to the functional analysis, the functions of the gut microbiota of L. xylina included metabolism, genetic information processing, and environmental information processing, while secondary functions were transmembrane transport, carbohydrate metabolism, replication, and repair. The most abundant gene was related to the membrane transport function in the first instar larvae fed on NSC. The number of genes related to carbohydrate metabolism in L. xylina samples fed on SAC was less than for samples fed on NSC. L. xylina samples fed on NSC may have had greater metabolic demands. Moreover, there were more genes related to the function of endocrine and metabolic diseases and the immune system in L. xylina samples fed on NSC than those fed on SAC. L. xylina may need to adapt to more hostile environments in the wild. The increase in microorganisms in insects may increase the breadth of the insect's feeding range [46]. Increasing the abundance of gut microbes in insects is critical, as gut commensal bacteria are key to the digestion of foods and may be particularly important for regulating the host's pathogenicity (both positively and negatively) [47]. Gut-associated bacteria play an important role in the development of L. xylina larvae by facilitating the degradation of plants and other organic matter consumed by L. xylina larvae.

Conclusions
This study revealed the gut bacterial community structures of L. xylina and the relationships between gut bacteria and host insect development. The effects of diet on symbiotic microbes were also clarified in this study. The gut microbiota of L. xylina used vertical transmission and self-regulation to better consume the host's nutrients.

Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Written informed consent has been obtained from the patient(s) to publish this paper.

Data Availability Statement:
The data presented in this study are openly available in Zenodo at https://doi.org/10.5281/zenodo.5235663. Note: A1-A7: first to seventh instars of larvae of L. xylina under the SAC treatment; N1-N7: first to seventh instars of larvae of L. xylina under the NSC treatment; p values: the t-test was used to compare the differences in larvae fed on two diets at the same instar; *: p < 0.05; **: p < 0.01; ***: p < 0.001; data are the means ± standard error.