To avoid the influence of age and season (Ma et al., 2007), all female bats in this study were randomly collected from July to August 2020 in the Asian particolored bats habitat under the highway bridge of Acheng District, Harbin City, Heilongjiang Province, China (45°55'N, 126°94'E). Asian particolored bats mainly feed on insects and inhabit the roofs or eaves of bridges or old buildings. During reproduction, males and females inhabit separate locations, which provides ideal conditions for sampling. Bats were captured with a mist net as they returned to their habitat after hunting. We gently caught the bats from the net by hand with sterile gloves. We released the bats from the net as soon as possible and none were injured. We collected lactating (n=10) and non-lactating (n=14) Asian particolored bats in two periods. The determination of reproductive status was based on external morphological characteristics: the nipples of female bats become enlarged during lactation and the hairs around the nipples of female bats disappear because of suckling by their infants (Haarsma, 2008; Racey, 2009). Each captured individual was put into sterilized kraft bags to collect their excrement. After defecation, fecal samples were immediately put into an empty cryopreservation tube with sterile forceps and placed into dry ice for preservation until they were transported to the laboratory, where they were frozen at −80°C until DNA extraction. On completion of the sampling, the bats were immediately released back to their habitat.

FastDNA™ Spin Kit (Santa Ana, California, United States) was used to extract total insect and bacterial genomic DNA samples following the manufacturer’s instructions, and the extracted DNA was tested. The quantity of extracted DNA was measured using a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific, Waltham, MA, United States) and the quality was assessed using agarose gel electrophoresis.

Forward primer ZBJ-ArtF1cF and reverse primer ZBJ-ArtR2cR were used for Cytochrome oxidase fragments of PCR amplification, and forward primers 338F and reverse 806R were used for PCR amplification of the V3–V4 region of bacterial 16S rRNA gene. The PCR reaction system and conditions are described in detail in the Supplementary Material. PCR amplicons were purified with Agencourt AMPure Beads (Beckman Coulter, Indianapolis, IN, United States) and quantified using the PicoGreen dsDNA Assay Kit (Invitrogen, Carlsbad, CA, United States). After the individual quantification step, amplicons were pooled in equal amounts and pair-end 2×300-bp sequencing was performed using the Illlumina MiSeq platform with MiSeq Reagent Kit v3.

Raw fastq files were quality-filtered by Trimmomatic (Bolger et al., 2014) and merged by FLASH (Magoč and Salzberg, 2011). After performing quality control to obtain the optimized sequences, the operational taxonomic units (OTUs) were clustered using Usearch (Edgar, 2010) with singletons removed and chimera filtering. All optimized sequences were mapped to representative sequences, and sequences with >97% similarity with the representative sequence were selected to generate an OTU table. All analyses of the sequencing filtering and normalization were performed in Usearch. Sequencing effort coverage was visually assessed using alpha diversity rarefaction curves of number of OTUs.

Taxonomic identification of insect species was made by comparing a representative sequence from each OTU to reference sequences in the Barcode of Life Database (BOLD)¹ and the Genbank database.² By doing so, the insects in the bats’ diet could be identified to the species level. Sequence data analyses were mainly performed using QIIME v1.8.0 (Caporaso et al., 2010) and R packages (v3.5.1). OTU-level ranked abundance curves were generated to compare the richness and evenness of OTUs among samples. The Shannon diversity index and Chao1 richness index were calculated using the OTU table in QIIME to determine the diet alpha diversity for the groups of bats. We then used the Wilcoxon rank-sum test to calculate significant differences. Beta diversity was analyzed to investigate variation in the dietary composition of the different group samples using unweighted UniFrac distance metrics. Differences in group dispersion among bat groups were assessed using ADONIS, using R package, vegan (Jari Oksanen et al., 2019). Beta diversity was then visualized with principal coordinate analysis (PCoA; Ramette, 2007). The species composition and relative abundance of bat diets in each group were calculated at order taxonomic levels, and the composition of dominant species in different groups was visualized by community Pie plots. The relative abundance of insect species at order, family, and genus consumed by lactating and non-lactating bats were analyzed using Wilcoxon rank-sum test in R package, stats (R Core Team, 2019).

QIIME software, Chao1 and Shannon indices were used to evaluate the alpha diversity of each sample. We then used the Wilcoxon rank-sum test to calculate significant differences between lactating and non-lactating bats. Beta diversity analysis was performed to investigate structural variation in microbial communities of the different group samples on the basis of unweighted UniFrac and weighted UniFrac distance metrics. Differences in group dispersion among bat groups were assessed using ADONIS. Beta diversity was then visualized with PCoA. Taxonomies were grouped at the phylum, family, and genus levels. Taxa abundances at the phylum, family, and genus levels were statistically compared among groups with using Wilcoxon rank-sum test, in R package, stats. To further identify the highly-dimensional gut microbes and characterize the differences between two or more biological conditions, we combined the linear discriminant analysis effect size (LEfSe) analysis.

We performed canonical correspondence analysis (CCA) to test the relative abundance of different insect species in the diet on microbial community at different taxonomic levels in lactating and non-lactating bats, applying forward selection and the Monte Carlo permutation test with 999 random permutations. In addition, we used Procrustes analysis to assess congruence between different insect species relative abundance dietary composition in different groups on microbial community structure and the significance of the Procrustes statistic was tested by 999 permutations with “protest” function.

After quality processing, a total of 293,784 effective sequences were obtained, with an average effective sequence of 12,241 per sample. We identified 276 OTUs from reads, 99 of which were only found in the diet of lactating bats and 138 of which were only found in the diet of non-lactating bats (Supplementary Table 1).

Dietary diversity differed significantly between the two reproductive periods. Compared with non-lactating bats, the alpha diversity of the diet based on the Shannon index was higher in lactating bats (Wilcoxon rank-sum test, p=0.004, Q=0.009; Figure 1A; Supplementary Figure 1A). However, the alpha diversity based on the Chao1 index was no different for lactating and non-lactating bats (Wilcoxon rank-sum test, p=0.23, Q=0.23; Figure 1B). The unweighted UniFrac metric data showed distinct clustering by lactating bat and non-lactating bat values, such that individuals from the two groups clustered separately. The results of ADONIS showed significant differences between the two groups (ADONIS: R²=0.14, p=0.001; Figure 2A).

Ten orders were found in the diet of lactating bats, with Diptera accounting for 67.64% and Lepidoptera for 23.74% (Figure 3A). Thirteen orders were found in the diet of non-lactating bats, with Diptera accounting for 40.24% and Lepidoptera for 26.94% (Figure 3B). We compared the relative read abundance of more than 1% of insect species and found that compared with non-lactating bats, lactating bats consumed more Diptera. The consumption of Coleoptera was significantly reduced (p=0.03), and the lactating bats even stopped consuming the larger Orthoptera (p=0.001; Figure 3C; Supplementary Table 3). At the family level, the proportion of Limoniidae (p=0.0002), Lasiocampidae (p=0.00002), and Limacodidae (p=0.04) were significantly increased and Carabidae (p=0.009) was significantly reduced in the diet of lactating bat diets (Figure 3D; Supplementary Table 3). At the genus level, the proportion of Rhipidia (p=0.000009), Symplecta (p=0.0002), Dicranomyia (p=0.02), and Dendrolimus (p=0.00003) were significantly increased and unclassified_f__Tipulidae (p=0.012), Amara (p=0.12), Hylaea_f__Geometridae (p=0.043; Figure 3E; Supplementary Table 3).

After quality inspection, 730,728 effective sequences were obtained (average of 30,447 reads/sample). We identified 995 OTUs from reads, 276 of which were only found in the gut microbiota of lactating bats and 380 of which were only found in the gut microbiota of non-lactating bats (Supplementary Table 1).

There was no significant difference in the alpha diversity of gut microbiotas between lactating bats and non-lactating bats (Wilcoxon rank-sum test, Chao1 index, p=0.62, Q=0.98, Shannon index, p=0.98, Q=0.98; Figures 1C,D; Supplementary Figure 1B). For the comparison of individual bat microbiotas, the PCoA results based on unweighted UniFrac metric data showed that individuals from the lactating and non-lactating groups clustered separately. The results of ADONIS showed significant differences between the two groups (ADONIS: R²=0.1387, p=0.001; Figure 2B). But weighted UniFrac measurements showed no difference between the lactating and non-lactating groups (ADONIS: R²=0.0755, p=0.137; Supplementary Figure 2).

Three major bacterial phyla with relative abundance greater than 1% were identified in Asian particolored bat microbiotas, and most sequences were classified as Firmicutes (93.14%), followed by Proteobacteria (3.32%) and Actinobacteria (3.02%; Figure 4A).

In terms of microbial composition, the relative abundances of some dominant microbial groups at the level of phylum, family, and genus changed significantly. At the phylum level, compared with the gut microbiotas of non-lactating bats, Actinobacteria significantly increased (p=0.004) and Firmicutes decreased (p=0.01) in lactating bats (Figures 4A,D; Supplementary Figure 3; Supplementary Table 3). At the family level, Corynebacteriaceae (p=0.0001), Staphylococcaceae (p=0.0001), Mycoplasmataceae (p=0.001), and Rhizobiaceae (p=0.02) significantly increased and Lactobacillaceae (p=0.001) significantly decreased in lactating bats (Figures 4B,E; Supplementary Figure 3; Supplementary Table 3). At the genus level, Mycoplasma (p=0.001), Corynebacterium (p=0.05), Jeotgalicoccus (p=0.0004), unclassified_f__Corynebacteriaceae (p=0.0001) significantly increased and Weissella (p=0.03) and Lactobacillus (p=0.001) significantly decreased in lactating bats (Figures 4C,F; Supplementary Figure 3; Supplementary Table 3).

Because effects of dietary composition on gut microbiome community may vary with taxonomic level, we summarized gut microbiome community and dietary datasets at multiple taxonomic levels, and performed correlation score for each rank. CCA suggests that there was a significant correlation between diet and gut microbiota in lactating and non-lactating bats at OTU and family levels (p<0.05; Figure 5; Supplementary Table 4). At the genus level, diet was found to be only correlated with the OTU level of gut microbiome (p=0.004), but no significant correlations were observed between genus and family level of gut microbiome (p>0.05; Figure 5; Supplementary Table 4). The Procrustes analysis showed that there was a strong correlation between the diet and gut microbiome of lactating bats compared with non-lactating bats (p<0.05; Supplementary Table 5).

Lactation is the most energetically costly activity during reproduction in mammalian females (Vanslambrouck et al., 2021). Bats are the only truly flying mammals and during lactation, in addition to the high energy expenditure required to produce milk, females have to carry their young while flying. This has been shown to require staggering energy expenditure (Kunz et al., 1995; McLean and Speakman, 1999). Our study suggests that to meet the high energy requirements of lactation, female V. sinensis have a significantly higher diversity of insect species in their diet than non-lactating females, and the proportion of specific groups in their diet changed significantly. For example, lactating bats relied heavily on Diptera and decreased their consumption of Orthoptera and Coleoptera. The consumption of Lasiocampidae and Limacodidae in lactating bats was also significantly higher than that in non-lactating bats. There are several reasons for these results. First, Diptera have a higher protein content than Lepidoptera (Churchward-Venne et al., 2017). Lactating bats increased their dietary diversity and chose high-protein foods to ensure their nutritional needs were met and that lactation functioned properly to provide nutrient factors that are essential for infant growth and development (Lee et al., 2016; Lee and Kelleher, 2016; Marangoni et al., 2016). Thus, it is evident that high-quality food selection, rather than bulk feeding, is the most important component of nutrient acquisition when nursing (Vesterinen et al., 2016). Second, OFT suggests that an animal will attempt to gain the greatest energetic benefit for the lowest energetic cost while foraging to maximize fitness. Coleoptera have large exoskeleton that take time to handle, so bats tend to eat less of them during lactation (Dufour and Sauther, 2002). Although Orthoptera has a higher protein content, most species are active during the day. Few nocturnal species are active on the ground, which makes them more difficult for bats to hunt (Whitby et al., 2020). Finally, lactating females consumed significantly more species of certain insect families of the order Lepidoptera than did non-lactating females. We speculated that these particular species may have higher quality proteins that are more effective in supplementing the nutritional and energetic needs of lactating females, but evidence for this is lacking. On the basis of these findings, the OFT is supported, such that during lactation, Asian particolored bats support their energy demands by eating more nutritious food, which increases their nutritional intake.

The gut microbiota plays an important role in the nutrition and health of the host (Al-Asmakh et al., 2014; Gentile and Weir, 2018). Dietary components have the capability to modulate the composition of this biota (Wu et al., 2011; Barrett et al., 2018). Although, we found differences in dietary composition between lactating and non-lactating bats, we did not detect significant differences in gut microbial diversity. The results suggest that healthy mammals tend to stabilize the diversity of their gut microbes in adulthood (Albenberg and Wu, 2014). Our results are consistent with studies in a variety of primates that confirmed that the alpha diversity of the gut microbial community remained stable throughout lactation. Studies on the human gut microbiota have shown that markers of microbial stability, such as richness and diversity, are often used as indicators of gut health (Gentile and Weir, 2018). In a few cases, they confirmed that the diversity of the intestinal microbiota of mammalian females changed with reproductive status (Gaona et al., 2019; Sun et al., 2020). However, these studies did not emphasize the role of dietary changes during lactation in shaping the structural diversity of gut microbiota (Phillips et al., 2012). On the contrary, they suggested that the diversity of gut microbiota in lactating females may be more susceptible to other factors besides diet, such as hormones and metabolism (Liu et al., 2019; Coleman et al., 2021).

Although, the overall diversity of gut microbes was unrelated to diet, we found that changes in diet had influences on specific, related microbiota. In human studies, it has been confirmed that dietary fat consumption is positively correlated with abundant Actinobacteria and negatively correlated with abundant Firmicutes (Abecia et al., 2007; Wu et al., 2011; Jin et al., 2021). Compared with non-lactating bats, the lactating bats had a higher relative abundance of Actinobacteria and lower relative abundance of Firmicutes, which would be useful for consuming more fat and reducing fat storage, thus improving the survival rate of offspring (Tarnaud, 2006; Hamady et al., 2008). Lactobacillus and Weissella mainly degrade macromolecules with complex structures to promote gut absorption (Famularo et al., 2005; Safika et al., 2019). Because of the reduced consumption of insects with large bodies and hard shells during lactation, it is speculated that the significant decrease in the relative abundance of Lactobacillus and Weissella in the intestines of V. sinensis during lactation is related to the decrease in the proportion of Coleoptera in the diet. Lactobacillus and Weissella participates in the production of short-chain fatty acids and maintains the balance of the gut and maintaining normal host health (Markowiak-Kopec and Slizewska, 2020). The relative abundance of Staphylococcaceae increased in lactating bats, which increased microflora nutrient acquisition by improving the hydrolysis of indigestible proteins and polysaccharides in the diets, and improving the energy efficiency of females to breastfeed their infants (Collado et al., 2008; Hunt et al., 2012; Rowland et al., 2018). In conclusion, by analyzing the correlation between diet and the gut microbiota at different classification levels, we confirmed that diet does have a regulatory effect on gut microbiota of lactating bats. Although there is no correlation between insect species at the genus level and gut microbes, this may be related to some unclassified species, but diet is only one of the factors regulating gut microbiota during lactation.

Mammals have evolved many behavioral and physiological strategies to meet the energy and nutrient requirements of their altered reproductive state. These strategies include eating longer, eating faster, eating more nutritious food, and enhancing metabolic efficiency (Bell and Bauman, 1997), as well as using seasonal reproduction (Bronson, 2009). In this study, we found that during lactation, Asian particolored bats did eat more nutritious food by consuming a diet that yields higher levels of protein. In addition, the abundance of specific microbes in their gut changed to help them adapt to changes to the composition of their diet. Because of limitations of sample size and the variables we considered, we were not able to fully reveal other behavioral and physiological changes caused by changes in mammalian reproductive status. Future studies should incorporate more factors to reveal the characteristics of adaptive evolution in different reproductive states of mammals, especially the adaptive mechanism of lactation, which is the reproductive period requiring the highest energy investments.