1. Introduction
Diet is known to affect the physiology and morphology of many taxa [
1]), including the composition of the gut microbiota [
2,
3,
4]. We use the term ‘microbiota’ to refer to the taxonomic diversity of Bacteria and Archaea assessed using marker genes, rather than ‘microbiome’, which refers to both taxonomic and functional diversity of the complete community [
5]. The microbial taxa hosted in the gut change during the development and aging of animals [
6], including humans [
7,
8,
9,
10]. The microbiota also changes over longer time scales in relation to diet [
2]. A correlation between host phylogeny and enteric microbiome composition and abundance [
11,
12] has been observed, suggesting coevolution of host and microbiome at deeper time scales. In addition to dietary and genetic variation, many other factors, such as geographical and physicochemical characteristics of the environment, are known to impact the gut microbiota (e.g., [
13,
14,
15]). Consequently, how fast gut microbiota can change and how strong the impact of diet is versus other environmental features in driving variation in gut microbiota remains relatively poorly understood.
To address how diet impacts the enteric microbiota, we investigated a unique system featuring rapid changes in diet at the population level. Forty-nine years ago, Nevo and colleagues (1972) designed a study to analyze the competition between
Podarcis siculus and
Podarcis melisellensis on islands [
16]. They introduced ten
P. siculus from the island of Pod Kopište to the island of Pod Mrčaru, and ten
P. melisellensis from Pod Mrčaru to Pod Kopište. During a follow-up study, it was noted that
P. melisellensis has disappeared from Pod Mrčaru, while the descendants of the ten
P. siculus introduced on this island have thrived [
17]. Moreover, the Pod Mrčaru lizards are currently largely omnivorous (consuming up to 80% plant matter in the diet in summer) and have undergone significant morphological changes in the hindgut, including the development of caecal valves [
17]. These changes are typically associated with the consumption of plant material in lizards (see [
18] for an overview). There are also subtle changes in digestive biochemistry between the lizards on the two islands, and these changes are largely isolated to the hindgut [
19], which houses the enteric microbiota engaged in the digestive process [
6,
20]. Given the known impact of diet on the gut microbiota in vertebrates and the fact that vertebrates are unable to endogenously digest plant fiber (e.g., cellulose; [
1]), we predicted there would be differences in the gut microbiota among the
P. siculus lizards from the two islands.
Previous studies on the highly specialized herbivorous marine and land iguanas of the Galápagos Islands showed that these animals had different microbiota from mammalian herbivores [
21,
22]. Interestingly, both species showed the conservation of microbial genes important to the breakdown process of plant material [
21,
22]. However, how general these changes are and how fast changes in the microbiota evolve remains unclear (e.g., the Galápagos iguanas diverged 4.5 million years ago; [
23]. A study experimentally manipulating the diet of an omnivorous species of
Liolaemus lizards showed that an experimental increase in the amount of plant matter in the diet over 40 days significantly impacted the gut microbiota, suggesting that some changes may happen rapidly [
20]. Kohl et al. (2016) further observed that lizards fed a plant-only diet had a higher gut microbial diversity than lizards fed a mixed diet [
20]. However, other studies have suggested that lizard gut microbiomes are at least partly derived from the local environment [
6,
24,
25]. For example, recent studies have demonstrated that altitude, geography, and insularity impacted the gut microbial diversity in lizards [
25,
26,
27]. Overall, these studies suggest that in lizards, diet is not the only driver of variation in the gut microbiota, but also that the local environment may play a crucial role in establishing its composition.
To explore the impact of the relatively rapid changes in diet as well as the influence of the local environment (population of origin) and sampling year (2014 vs. 2016) in natural populations of lizards, we used 16S rRNA gene sequencing to compare the gut microbial communities of 40
P. siculus sampled from the islands of Pod Kopište and Pod Mrčaru, and an insectivorous population from the mainland (Split). Thus, we compared recently derived populations from the islands and contrasted them with a more distantly related population of
P. siculus. We also compared the three populations across different years (2014 and 2016) to examine whether any differences among the populations were stable over time. We hypothesized that the lizards from Pod Mrčaru would have more microbial taxa associated with the digestion and metabolism of plant material [
6,
21] than the insectivorous populations, and that the island lizards would be different from the lizards on the mainland [
25]. The overall aim of the study was to better understand the possible drivers of gut microbial diversity in lizards with different diets (omnivorous vs. insectivorous) and from different localities (insular vs. mainland).
4. Discussion
We found support for each of our hypotheses, as specific microbial taxa in the lizards’ guts varied by diet, others with location, and some with time (
Figure 2). Overall, we did not see large-scale shifts in the entire microbiota (
Supplementary Table S2), as one may expect given that we investigated population-level differences (
Figure S4). The morphology of the Pod Mrčaru
P. siculus lizards changed rapidly (~30 years) in line with a dietary shift towards the consumption of more plant material (higher bite force, larger body size, evolution of caecal valves, longer guts; [
17]). Moreover, the Pod Mrčaru
P. siculus digest plant material more efficiently (by about 10%) than do the Pod Kopište lizards [
19]. Consistent with most of the digestive physiological differences among the Pod Mrčaru and Pod Kopište
P. siculus being isolated to the hindgut [
19], we observed differences in the hindgut microbiota among these lizard populations, and diet was a significant indicator of microbiota diversity (
Table 2). The microbial diversity of omnivorous Pod Mrčaru lizards was higher than the microbial diversity of insectivorous lizards (
Supplementary Figure S3), and the Archaean,
Methanobrevibacter, was significantly more abundant in omnivorous individuals, with diet being a major factor in their abundance (
Table 1). Although one species of
Methanobrevibacter (
M. smithii) is commonly found in the human gut microbiome [
59], this genus is also associated with shifts in microbiome function. For instance,
Methanobrevibacter is methanogenic and leads to more polysaccharide degradation by bacterial species, greater levels of microbial fermentation, and is often associated with obesity and type two diabetes in rodent models [
59,
60]. Perhaps this taxon is aiding the Pod Mrčaru lizards in acquiring sufficient energy from their plant-rich diet. One caveat here is that, based on the
Methanobrevibacter abundance, one would predict greater levels of microbial fermentation in the hindguts of the Pod Mrčaru lizards than in the insectivorous population from Pod Kopište, but in fact the opposite was found [
19]. Thus,
Methanobrevibacter may play other roles in the digestive process than just the production of short chain fatty acids in
P. siculus [
4]. More Pod Mrčaru
P. siculus should be screened for short chain fatty acid concentrations in their hindguts to determine whether fermentation matters for these lizards or not.
Bacteria in the family Peptostreptococcaceae were more abundant in the Pod Mrčaru omnivorous lizards (
Table 2;
Figure 3). Although the detailed function of these taxa is unknown, they have been associated with lower protein diets in porcine models [
61], and protein digestion more broadly [
62], suggesting that they could aid in amino acid scavenging. A previous study has shown that trypsin activity is higher in the hindguts of the Pod Mrčaru lizards, suggesting that protein scavenging is something that occurs in these animals [
19]. Interestingly, the genus
Helicobacter was abundant in the Pod Mrčaru lizards (>2.5% relative abundance), whereas this genus composed less than 0.2% of the reads in the insectivorous populations (
Figure 3;
Supplementary Table S2). Several
Helicobacter taxa are known from intestinal environments, including other
Podarcis species [
25] and other lizards [
6,
24], and they may be able to perform many biochemical functions [
63]. What these bacteria can do in these lizards requires further study, however. Other bacteria observed in the omnivorous lizards belong to the genus
Rickettsiella (
Figure 3) and were variable in abundance among years (
Table 1;
Figure 2).
Rickettsiella are intracellular parasites in a range of host organisms, including insects [
64] and mollusks [
65]. It remains, however, unknown what the role of this genus could be in
P. siculus. It is peculiar that it is only prominent in the omnivorous population, but also varied with time (
Figure 2). Finally, some of the microbial abundances in the Pod Mrčaru lizards (enriched in
Methanobrevibacter, depleted in
Akkermansia) are consistent with dysbiosis (type two diabetes, obesity) in mammalian models including humans [
66]. An altered metabolism, including insulin resistance and dyslipidemia, may be advantageous in nutrient-limited environments as seen in cavefishes [
67]. Perhaps something similar is happening the Pod Mrčaru lizards, and this could be the focus of future microbiome-host studies in these lizards.
In the island lizards more generally, but in the Pod Kopište lizards more specifically, a member of the genus
Bacillus appears to be abundant (
Table 1;
Figure 2 and
Figure 3). With over 266 named species, and many biochemical functions known, species in the genus
Bacillus can perform multiple functions ranging from enzyme secretion to short chain fatty acid (SCFA) synthesis [
68]. The island lizards generally have higher b-glucosidase activities in their guts than mainland lizards [
19], and perhaps the
Bacillus may be the source [
68].
Citrobacter, which are a member of the Enterobacteriacea, were more abundant in the insectivorous lizards (
Figure 3,
Supplementary Table S2). This organism is known to inhabit intestinal environments, and indeed, a wide range of habitats [
69], but their role in the gut also remains to be determined.
Akkermansia is a common intestinal denizen that is more abundant in Pod Kopište than in Pod Mrčaru lizards (
Figure 3).
Akkermansia live in the mucosal layer of mammalian intestines, digest mucins, and play roles in immune function, as well as mucus and peptide secretion [
70,
71]. A member of this genus was abundant in all lizards, but more so in the insectivorous populations (
Figure 3,
Supplementary Table S2). Interestingly,
Akkermansia can become more abundant during starvation in many host taxa [
62,
72] since they can digest mucus [
70], and there is a negative relationship between
Akkermansia abundance and obesity in rodent models: fewer
Akkermansia equates with the obese phenotype [
71]. With
Methanobrevibacter abundance positively associated with obesity in mammalian models [
59,
60], and
Akkermansia negatively so [
71], the Pod Mrčaru lizards display both patterns, again, suggesting that they rely on some level on microbial help to gain sufficient energy from their plant-rich diet.
Desulfovibrio, sulfur reducing bacteria [
73], were also more abundant in the island lizards than in the mainland ones (
Table 1,
Figure 2). Members of this genus have been shown to be more active when exposed to chitin breakdown products [
73] and were found in greater abundance in omnivorous lizards fed more plant material in the laboratory [
20]. Although all of the lizard populations ingest chitin (part of the exoskeleton of their insect prey), the Pod Kopište and mainland lizards presumably consume more chitin than the Pod Mrčaru lizards. Nevertheless, a previous study showed that only the Pod Kopište lizards displayed elevated N-acetyl-β-D-glucosaminidase (NAG) activity in their guts, indicating potential elevated breakdown of chitin [
19].
Previous studies on highly specialized herbivorous lizards like Galápagos marine and land iguanas found differences in the fecal microbiota of these two lizards with a greater diversity of OTUs in the land iguana [
21]. Land iguanas would consume more fibrous material (e.g., cellulose) than the marine iguana, consuming marine algae [
74]. A subsequent metagenomic study found that the fecal microbiomes between these two lizards were more similar to each other than either was to mammalian herbivores, and that genes for enzymes that could aid in fiber digestion (e.g., cellulases) were abundant in both species, suggesting a functional role of the microbiome in digestion. Moreover, two omnivorous liolaemid lizard species had microbiomes that were more similar to each other than to a distantly related herbivorous lizard [
6]. This example, and the iguana studies [
21,
22], agree with the tenets of Phylosymbiosis—i.e., that there is a correlation between host identity and microbial community [
12]. In this study, we found similar to Baldo et al. (2018) for
P. lilfordi that the collection locality matters for
P. siculus (
Figure 4) [
25]. Location was a significant predictor for
Methanobrevibacter,
Bacillus, and
Delsulfovibrio (
Table 1;
Figure 2). The microbiotas of the Pod Mrčaru lizards consistently grouped separately from the other populations, and those of the Pod Kopište lizards were separate from those of the mainland lizards from Split (
Figure 4). Indeed, the two insectivorous populations showed the most heterogeneity amongst them, depending on the analyses used (
Figure 4). The gut microbial diversity of insectivorous lizards overall was more heterogeneous than the gut microbial diversity of omnivorous lizards (
Figure 4) because insectivorous lizards living on the continent had different gut microbial communities than those living on Pod Kopište. Likewise,
Bacillus and
Desulfovibrio were more abundant in the gut microbiotas in insular lizards (
Figure 2). The local environment is likely very important in structuring the microbiota in lizards, as a strong similarity between gut microbiota and those observed on plants eaten by lizards was observed in a previous study [
20]. The enteric microbial communities of
P. lilfordi in Menorca were structured by the age of isolation of the islet and the local environment [
25] in accordance with our results. Other potential drivers of differences in gut microbial diversity such as sex or year of sampling were largely non-significant as observed in other studies of lizard gut microbiota [
11,
75].
The vast majority of the bacteria observed in previous microbiome studies in lizards were also observed in
P. siculus ([
6,
20,
22,
24,
25,
27,
75];
Supplementary Table S2). Our results showed moderate but significant differences in the microbiota between omnivorous and insectivorous lizards, and the omnivorous lizards can digest plant material moderately (~10%), but significantly better than lizards from the insectivorous population [
19]. Given that other studies have demonstrated changes in microbiota when feeding lizards different diets in captivity [
20], this likely does not reflect a time constraint but rather the availability of different microbiota in the immediate environment of the lizards [
6], combined with flexibility in gut structure and function [
20].