ABSTRACT Opportunistic feeding and multiple other environment factors can modulate the gut microbiome, and bias conclusions, when wild animals are used for studying the influence of phylogeny and diet on their gut microbiomes. Here, we controlled for these other confounding factors in our investigation of the magnitude of the effect of diet on the gut microbiome assemblies of nonpasserine birds. We collected fecal samples, at one point in time, from 35 species of birds in a single zoo as well as 6 species of domestic poultry from farms in Guangzhou city to minimize the influences from interfering factors. Specifically, we describe 16S rRNA amplicon data from 129 fecal samples obtained from 41 species of birds, with additional shotgun metagenomic sequencing data generated from 16 of these individuals. Our data show that diets containing native starch increase the abundance of Lactobacillus in the gut microbiome, while those containing plant-derived fiber mainly enrich the level of Clostridium. Greater numbers of Fusobacteria and Proteobacteria are detected in carnivorous birds, while in birds fed a commercial corn-soybean basal diet, a stronger inner-connected microbial community containing Clostridia and Bacteroidia was enriched. Furthermore, the metagenome functions of the microbes (such as lipid metabolism and amino acid synthesis) were adapted to the different food types to achieve a beneficial state for the host. In conclusion, the covariation of diet and gut microbiome identified in our study demonstrates a modulation of the gut microbiome by dietary diversity and helps us better understand how birds live based on diet-microbiome-host interactions. IMPORTANCE Our study identified food source, rather than host phylogeny, as the main factor modulating the gut microbiome diversity of nonpasserine birds, after minimizing the effects of other complex interfering factors such as weather, season, and geography. Adaptive evolution of microbes to food types formed a dietary-microbiome-host interaction reciprocal state. The covariation of diet and gut microbiome, including the response of microbiota assembly to diet in structure and function, is important for health and nutrition in animals. Our findings help resolve the major modulators of gut microbiome diversity in nonpasserine birds, which had not previously been well studied. The diet-microbe interactions and cooccurrence patterns identified in our study may be of special interest for future health assessment and conservation in birds. INTRODUCTION Digestive tracts of animals contain microbial communities that are composed of different bacterial groups with various abundances and functional characteristics ([49]1). In addition to digestion and energy acquisition, many recent studies have shown that the animal gut microbiome also has important functions in host immune responses, detoxification, and behavior ([50]2, [51]3). The adaptive capacity and health status of an animal are not solely due to the host genome but also depend upon the vast genetic repertoire of its microbiome ([52]4). Birds exhibit the most diverse range of ecological functions among vertebrates ([53]5) and represent a highly evolved lineage that provides processes that are essential for ecological communities and agricultural ecosystems ([54]6, [55]7). Coevolution between host and microbial lineages played key roles in the adaptation of mammals to their diverse lifestyles, but this subject has been far less studied in birds ([56]8[57]–[58]11). Birds represent multiple different feeding groups including folivorous, nectar feeding, opportunistic, strictly carrion feeding, and others ([59]12). Exploitation of a new dietary niche is a powerful driver for changes in gut microorganisms and their coevolution with their animal host ([60]13). Many studies have shown the important effects of diet on birds, especially in Passeriformes, which represent more than half of all known species of birds ([61]14[62]–[63]17). To date, however, few studies have characterized the gut microbiomes of nonpasserine birds and their associations with their highly diverse dietary habits. The dynamic gut microecological system that formed in the adaption of birds to their environment is influenced by many factors, such as sex, reproductive status, age, geography, environment, human activity, and social structure ([64]9, [65]18[66]–[67]23). The dominant drivers of gut microbiome diversity in birds appear to be host evolutionary history and diet ([68]10). Diet drives not only taxonomic diversity but also the functional content of the gut microbiome in mammals ([69]24, [70]25). While there have been many analyses of microbial taxa based on amplicon rRNA sequencing, fewer studies have focused on functional gut metagenome profiling in the dietary diversity of birds ([71]26). Moreover, it is difficult to address questions on the identity of host-specific microbes, as the microbial species were often collected in different habitat niches and environmental conditions (e.g., weather and season) that have roles in influencing the composition of the gut microbiota ([72]27). Therefore, to better reveal the complex relationship between diet and the gut microbiome in birds, it is necessary to minimize the influence of any interfering factors. In this study, fecal material was collected from birds housed at Guangzhou Zoo representing 35 species with various dietary habits at one point in time to minimize the influence of external factors such as geography, weather, and season. In addition, we also compared the gut microbiota from 6 species of domestic poultry that were fed different types of food to identify differences in their gut microbiota that can occur in a species as a response to different food types. To understand the dietary and phylogenetic effects on the taxonomic composition and metabolic function of gut microbiota in birds, a systematic analysis combining 16S rRNA amplification and metagenomic sequencing was used. RESULTS Gut microbial diversity of nonpasserine birds. To assess microbial diversity, we sequenced the V3-V4 regions of the 16S rRNA gene in 129 fecal samples from 41 species (classified in the orders Gruiformes, Psittaciformes, Anseriformes, Accipitriformes, Galliformes, Pelecaniformes, Ciconiiformes, Bucerotiformes, Struthioniformes, Casuariiformes, Columbiformes, and Charadriiformes) ([73]Fig. 1A and see [74]Table S1 in the supplemental material). In total, 128 samples passed our quality control process and 2,391 operational taxonomic units (OTUs) were identified. We first assessed the impact of general feeding habits on gut microbiome diversity, noting, however, that the food types of omnivores vary widely ([75]Fig. 1B and [76]Fig. S1). When we grouped according to seven classes of food types (fruits, corn-soy, grains, foliage, flesh, fish, and omnivore), we found that 135 OTUs were shared by all groups. The food type with the highest number of unique OTUs was the fruit food group (294 OTUs), followed by the omnivore group (228 OTUs) ([77]Fig. 1C). In contrast, the fewest unique OTUs were detected in the grain (1 OTU), foliage (13 OTUs), and corn-soy (34 OTUs) food groups. Moreover, most OTUs (90%) were detected only in fewer than 20% of samples ([78]Fig. S2). FIG 1. [79]FIG 1 [80]Open in a new tab Diet type influences microbial diversity. (A) Phylogenetic tree of the birds used in this study. (B) Flower plot shows shared and unique OTUs between the 6 feeding habit groups. (C) Flower plot shows shared and unique OTUs between 7 dietary type groups. (D) Differences in microbial diversity (Chao1, PD whole tree, and Shannon index) at the OTU level between 7 dietary type groups shown as box plots (t test). (E) PCoA plot based on the Bray-Curtis dissimilarities at the OTU level. Differences observed between the groups are based on the PERMANOVA test. Results show that dietary type is a predictor of microbial variance (r^2 = 0.21304, P = 0.0001). Each color corresponds to a dietary type. Ellipses are at the 70% confidence level. FIG S1 Feeding habits influence microbial diversity. (A) Venn diagram of the OTUs shared between the omnivorous, herbivorous, and carnivorous groups. (B) Differences in microbial diversity (Chao1, PD whole tree, and Shannon index) at the OTU level between the 6 feeding habit groups are shown as box plots. (C) PCoA plot based on the Bray-Curtis dissimilarities at the OTU level. Download [81]FIG S1, TIF file, 1.0 MB^(1MB, tif) . Copyright © 2021 Xiao et al. This content is distributed under the terms of the [82]Creative Commons Attribution 4.0 International license. FIG S2 OTUs (A) and microbiota genera (B) are sparsely distributed in the data set. The prevalence of OTU or genera across all samples was calculated (found in at least one sample). Download [83]FIG S2, TIF file, 0.9 MB^ (945.9KB, tif) . Copyright © 2021 Xiao et al. This content is distributed under the terms of the [84]Creative Commons Attribution 4.0 International license. TABLE S1 Metadata for each sample used in this study. Download [85]Table S1, DOCX file, 0.02 MB^ (26.2KB, docx) . Copyright © 2021 Xiao et al. This content is distributed under the terms of the [86]Creative Commons Attribution 4.0 International license. Next, we used Chao1, phylogenetic diversity (PD whole tree index), and Shannon index to illustrate bacterial richness and diversity within the communities based on OTUs level. The α-diversity index of the grain food group was the lowest, and significantly lower than the fruit, corn-soy, flesh, and omnivore groups with the Chao1, PD, or Shannon index (t test, P < 0.05). No significant differences were observed among the other groups ([87]Fig. 1D and [88]Table S2). A principal-coordinate analysis (PCoA) based on the Bray-Curtis dissimilarity values was used to assess the differences in bacterial community structure between the samples. The results of this analysis revealed a significant clustering of gut microbiota by diet groups (permutational multivariate analysis of variances [PERMANOVA], P < 0.001), with food types separating the microbial communities along the first principal coordinate (PC1, 13.95% of variance) ([89]Fig. 1E). TABLE S2 Comparison of the α-diversity (Chao1 index, PD whole tree index, and Shannon index) between the pairwise groups. Download [90]Table S2, DOCX file, 0.02 MB^ (22.8KB, docx) . Copyright © 2021 Xiao et al. This content is distributed under the terms of the [91]Creative Commons Attribution 4.0 International license. Based on the Mantel test, both host diet (r = 0.1623, P value = 0.0001) and phylogeny (r = 0.09667, P value = 0.0001) affect the gut microbiota composition of birds, but the influence of diet seems to be greater. To further assess the effects of host phylogeny and diet on the gut microbiome, we then compared the host phylogenetic tree with a Bray-Curtis dissimilarity-based UPGMA (unweighted pair group method with arithmetic mean) tree of the gut microbiota, with the results showing that the gut microbiomes of the different bird species mainly clustered based on food type ([92]Fig. S3). Moreover, gut microbiota from domestic poultry species fed with different food types (e.g., Gallus gallus, Meleagris gallopavo, Anas platyrhynchos, and Cairina moschata) was diverse, and also mainly clustered according to their food type ([93]Fig. S3). MaAsLin2 analysis (adjusted P value < 0.05) further revealed that diet plays a major role and accounts for 89.1% of the microbiota features ([94]Fig. S4). FIG S3 Effects of host phylogeny on gut microbiome diversity. The host phylogenetic tree is shown on the left. UPGMA tree clustering based on Bray-Curtis dissimilarities of gut microbiome is on the right. Sample and host are connected by the straight lines. Download [95]FIG S3, JPG file, 1.9 MB^ (1.9MB, jpg) . Copyright © 2021 Xiao et al. This content is distributed under the terms of the [96]Creative Commons Attribution 4.0 International license. FIG S4 Effects of diet and phylogeny on the gut microbiome based on the MaAsLin2 analysis. (A) Pie chart of the significantly explained gut microbiome by diet and phylogeny. (B) Box plots show the BH-adjusted P values. Download [97]FIG S4, TIF file, 0.4 MB^ (432.1KB, tif) . Copyright © 2021 Xiao et al. This content is distributed under the terms of the [98]Creative Commons Attribution 4.0 International license. Dietary diversity affects predominant bacteria. OTUs and genera were sparsely distributed in all samples ([99]Fig. S2). However, the predominant bacterial phyla present in the feces of all birds were Firmicutes (mean abundance ranged from 9.19% to 76.49%), Proteobacteria (11.14% to 51.48%), Actinobacteria (1.53% to 20.22%), and Bacteroidetes (0.02% to 13.78%) ([100]Fig. 2A). Notably, the most abundant phylum in the flesh-eating birds was Proteobacteria, while Firmicutes was the most abundant in the others. At the genus level, the five most abundant genera were Lactobacillus, Clostridium, Enterococcus, Escherichia, and Turicibacter ([101]Fig. 2B). Consistent with the α-diversity characteristic, the number of genera with mean relative abundance of >1% in the omnivore food group was higher than in the other groups. FIG 2. [102]FIG 2 [103]Open in a new tab Microbial composition in the different dietary groups. (A and B) The bar plot shows taxa with average relative abundances higher than 1% at the phylum level (A) and the genus level (B). Remaining species are classified as other. (C) LEfSe analysis. Cladogram showing the differences in relative abundance of taxa at five levels between the 7 dietary groups. Plot showing the taxonomic levels represented by rings with phyla in the outermost ring and genera in the innermost ring. Circles with nonyellow color indicate that there is a significant difference in the relative abundance at the different taxon levels (Wilcoxon rank sum test, P < 0.01; LDA score > 4), and yellow circles indicate nonsignificant differences. Differences in the abundance of the bacterial taxa were determined through a linear discriminant analysis effect size (LEfSe) analysis. Differentially abundant taxa were considered when there was significant variation between any two groups. A total of 35 taxa, at different classification levels, were found to have significant differences (Wilcoxon rank sum test, P < 0.01) ([104]Fig. 2C). At the genus level, Lactobacillus, Clostridium, Oceanisphaera, and Cetobacterium were the dominant genera and were significantly more abundant in the grain, foliage, flesh, and fish food groups, respectively (Wilcoxon rank sum test, P < 0.01). At the order level, a significant enrichment of Bacteroidales was detected in the corn-soy food group. No significantly enriched taxa were observed in the omnivore food group. Microbial cooccurrence association patterns are influenced by diet. We next examined how bacterial species cooccur among the birds in our study, which might be due to dietary differences or microbe-microbe interactions. A network containing 285 nodes and 2,689 edges was constructed ([105]Fig. 3A and [106]Table S3). Based on the layout structure, this integrated network could be divided into 6 subnetworks, each with differing taxonomic compositions at the class level ([107]Fig. 3A and [108]Table S4). FIG 3. [109]FIG 3 [110]Open in a new tab Microbial community linkages and species coexistence in the gut microbiome of birds. (A) Colored cooccurrence network of microbial taxa. Each node represents one OTU, and each edge represents a strong (|ρ| > 0.6) and significant correlation (FDR P < 0.01) between the two nodes. The size of each node is proportional to the degree of the OTUs; the thickness of edges is proportional to the value of the Spearman correlation coefficient. Gray edge, positive correlation (ρ > 0.6); red edge, negative correlation (ρ < −0.6). (B) Box plot of the completeness and richness of each submicrobial group (SC) in the different dietary groups. Cooccurrence percentage represents the completeness, and total abundance represents the richness. TABLE S3 List of a set of topological metrics of the cooccurrence network. Download [111]Table S3, DOCX file, 0.01 MB^ (14.7KB, docx) . Copyright © 2021 Xiao et al. This content is distributed under the terms of the [112]Creative Commons Attribution 4.0 International license. TABLE S4 Taxonomic annotation of the OTUs in each subnetwork. Download [113]Table S4, DOCX file, 0.1 MB^ (54.3KB, docx) . Copyright © 2021 Xiao et al. This content is distributed under the terms of the [114]Creative Commons Attribution 4.0 International license. At the class level, the cooccurrence network was mainly composed of interactions of Clostridia, Bacteroidia, Gammaproteobacteria, and Bacilli. Subcommunity a (SC-a), subcommunity b (SC-b), and subcommunity c (SC-c) were dominated by Clostridia and Bacteroidia; subcommunity d (SC-d) and subcommunity e (SC-e) possessed more members of Actinobacteria, Gammaproteobacteria, and Alphaproteobacteria; SC-f was mostly composed of taxa from Bacilli. In addition, most of the interactions were positive, while negative interactions appeared only between Pseudomonas (classified in Gammaproteobacteria) in SC-d with Negativibacillus, Ruminococcaceae UCG-014, Intestinimonas, and Christensenellaceae R-7 group (classified in Clostridia) in SC-a ([115]Fig. 3A), suggesting that competitive inhibition occurs between these communities. We then calculated the cooccurrence percentage and total abundance of each submicrobial community to estimate the microbial coexistence in the different diet groups. The presence and abundance of OTUs from each subnetwork differed substantially among the groups ([116]Fig. 3B). SC-d was generally the most prevalent in all birds, while wide differences in the prevalence and abundance of SC-a and SC-f occurred among the groups, suggesting certain host dietary specificity of these microbial consortia. The abundance of SC-a in the corn-soy group was significantly higher than in the other groups, while the abundance of SC-d in the corn-soy group was the lowest among all groups (Wilcoxon rank sum test, P < 0.05) ([117]Table S5). This phenomenon is consistent with antagonism between SC-a and SC-d as described above. We also compared the gut microbiome of some birds in our study with those previously reported from wild birds (Cairina moschata and Dromaius novaehollandiae) ([118]8, [119]10). A greater abundance of Proteobacteria was observed in the wild birds (data not shown). Compared with birds fed commercial feed, there were more Gram-negative bacteria in the gut microbiome of zoo birds and wild birds, which may be related to their complex food types and environment. TABLE S5 Comparison of the cooccurrence percent and total abundance of the microbiota communities between the pairwise groups. BH-adjusted P values of <0.05 are colored (red, percent; green, abundance). Download [120]Table S5, DOCX file, 0.03 MB^ (27.7KB, docx) . Copyright © 2021 Xiao et al. This content is distributed under the terms of the [121]Creative Commons Attribution 4.0 International license. Adaptive evolution of microbial functions to fit food types. To further investigate the functional capacities of the gut microbial communities in birds, a metagenomic analysis was conducted. The 16 samples used for the metagenomic analysis are distributed across the 6 food type groups (corn-soy, flesh, foliage, fruit, grain, and “omni”). In total, 2,420,003 assembled genes (92.3%, 2,621,823) were identified from the prokaryotic microbes and fungi by searches against the NCBI NR database. Of this total, 1,729,192 (65.95%) and 52,051 (1.99%) were annotated in the KEGG and CAZy (carbohydrate-active enzyme) databases, respectively. Detailed KEGG orthology (KO) annotation information is listed in [122]Fig. S5A. Notably, 2,508 (28.47%) of the KOs were annotated in the global and overview metabolism pathways and 741 KOs were annotated in carbohydrate metabolism. The number of KOs with an average relative abundance higher than 0.01% in the different groups ranged from 1,329 to 2,779, and the total abundance of those KOs in each group was higher than 85% ([123]Fig. S5B). This indicates that the high-abundance KOs cover most of the microbial functions. FIG S5 Functional annotation of gut microbiome of nonpasserine birds. (A) Pathway classification of all annotated KOs in the metagenomic sequencing. (B) Count and total abundance of KOs with relative abundance higher than 0.01% in each group. Download [124]FIG S5, TIF file, 1.9 MB^ (2MB, tif) . Copyright © 2021 Xiao et al. This content is distributed under the terms of the [125]Creative Commons Attribution 4.0 International license. To compare the microbial functions between each group, we first tested the KEGG pathway enrichment analysis based on the top-abundance KOs (mean relative abundance > 0.01%) in each group. The top 20 significantly enriched KEGG metabolism pathways in each group are shown in [126]Fig. 4A. Only 6 pathways were shared among the 6 groups. Due to differences in host diet, the metabolic pathway enrichment for each group was different. For example, lipid metabolism, including glycerolipid metabolism and fatty acid biosynthesis, was enriched, while some amino acid biosynthesis functions were not, in the corn-soy group. In addition, starch and sucrose metabolisms were enriched in all groups except the flesh group, a group of birds that do not eat plant-derived polysaccharides. FIG 4. [127]FIG 4 [128]Open in a new tab Microbial functional differences between groups. (A) KEGG pathway enrichment of the high-abundance KOs (average relative abundance > 0.01%) in each group. The top 20 most enriched pathways in each group are shown. The count equals the number of KOs in this pathway. (B) Heatmap shows the top 20 highest-abundance CAZy families in each group. The relative abundance of each CAZy family is colored according to the row z-score. Moreover, we explored the distribution of CAZymes in the different groups. The top 20 abundant CAZymes in each group are shown in a z-score normalized heatmap ([129]Fig. 4B). These data showed that most of the high-abundance CAZymes were detected in the corn-soy, flesh, and grain groups. Groups that have diets containing plant-derived fiber (fruit, “omni,” and foliage) had similar enzyme profiles and clustered together. DISCUSSION Host evolutionary history and diet are suggested to be the main factors modulating microbial community composition in the vertebrate gut ([130]10, [131]25). When wild animals are used to study the influence of diet and phylogeny in the composition of the gut microbiome, other factors, such as habitat, weather, and season, can bias the analysis and conclusions. In this study, we collected fecal samples from multiple species of nonpasserine birds at one time to minimize these confounding factors. In our study, we found that diet has a greater impact on the gut microbiome than host phylogeny based on the Mantel test and a MaAsLin2 analysis (see [132]Fig. S4 in the supplemental material). For example, different species of birds eating the same food composition tended to have similar gut microbiomes ([133]Fig. S3). Although not all samples clustered according to food characteristics, our results support the conclusion that food source is a major factor determining the differences in intestinal microbial composition ([134]25). The bioactivity and bioavailability of diet are two aspects driving the patterns of the nonpasserine gut microbiome assembly ([135]28). The predominant bacteria found in the intestinal tracts of birds fed different types of food vary greatly. The relative abundance of the Lactobacillus, classified in Bacilli, in the grain-fed group reached more than 60% of the total microbes and was significantly higher than any other group (Wilcoxon rank sum test, P < 0.05). Indeed, the ranking of groups, from high to low, in Lactobacillus abundance was grain, corn-soy, fruit, and foliage, followed by flesh and fish ([136]Fig. 2). Starch is the major storage polysaccharide in cereal grains, legumes, and many roots and tubers ([137]29), and the above result indicates that Lactobacillus was positively associated with the intake of starch-rich foods. It has been reported that starch is the only polysaccharide hydrolyzed by the extracellular enzymes (amylopullulanase) of Lactobacillus ([138]30). Interestingly, functional metagenome analysis showed that genes involved in starch and sucrose metabolism were more greatly enriched in the plant-derived polysaccharide intake groups (grain, corn-soy, fruit, foliage, and “omni” food groups) than in the carnivore groups (flesh and fish groups) ([139]Fig. 4A). Although birds can secrete pancreatic amylase, they have limited ability to digest native starch as it is highly organized ([140]31). Taken together, we conclude that a high abundance of Lactobacillus in the gut microbiota of birds is essential for the metabolism of native starch. Plant-derived fiber is the main energy source for folivores, but dietary fiber utilization by birds is inefficient and variable due to the absence of enzymes that can digest fiber ([141]32). We found that Clostridium was significantly enriched in the foliage food group (Wilcoxon rank sum test, P < 0.05), followed by the fruit and omnivore food groups ([142]Fig. 2C), which is consistent with the results observed in passerine birds that have plant-based diets ([143]33). A similar CAZyme profile appeared in these 3 groups (foliage, fruit, and omnivore), whose diets contain high levels of plant cellulose ([144]Fig. 4B). Enzymes in Clostridium digest fibers and produce various metabolites such as short-chain fatty acid (SCFA) that can be used by the host ([145]34, [146]35); therefore, the enriched level of Clostridium in the guts of birds that eat plant-derived fiber could make up for the lack of fiber digestive enzymes in the host. Foods can be regarded as a possible source of some microbes, such as the lactic acid bacteria found in the human gut microbiome ([147]36), but it is not yet known to what extent the microbes ingested by birds with different food preferences become members of their gut microbiome.