What Shapes The Gut Microbiome of Lizards From Different Habitats?


 Host-gut microbiota interactions are complex and can have a profound impact on the ecology and evolution of both counterparts. Several host traits such as taxonomy, diet and social behavior, and external factors such as prey availability and local environment are known to influence the composition and diversity of the gut microbiota. In this study, we investigated the influence of taxonomy, sex, host size, locality/habitat on gut microbiota diversity in five lizard species from two different sites in Portugal. We also analyzed the potential levels of microbial transmission between species that live in sympatry and syntopy. We studied Podarcis bocagei and Podarcis lusitanicus from northern Portugal (Moledo); and two invasive species, Podarcis siculus and Teira dugesii, and the native Podarcis virescens from Lisbon. We used a metabarcoding approach to characterize the bacterial communities from the cloaca of lizards, sequencing the V4 region of the 16S rRNA. Habitat/locality was found to be the main driver of the differences in composition and structure of gut bacterial communities of the studied lizards, with host effects more evident at finer taxonomic scales. Additionally, lizards from urbanized environments had higher microbiome diversity than lizards from rural areas. We detected a significant positive correlation between size and gut bacterial alpha-diversity in the invasive species P. siculus, which could be due to higher exploratory behaviours. Moreover, estimates of bacterial transmission indicate that P. siculus may have acquired a high proportion of local microbiota. These findings indicate that a diverse array of host and environmental factors can influence lizards gut microbiota.


Introduction
A myriad of microorganisms can be found living in the gastrointestinal tract of all animals. These microorganisms have a signi cant impact on host biology and can in uence a variety of processes that affect host tness [1,2]. While certain variations in the composition of gut microbial communities can cause disease [3,4], gut microbiome may also increase resistance to pathogens, besides being important for xenobiotics metabolism, nutrient uptake and energy acquisition [e.g. 5,6]. Moreover, gut microbiota may also contribute towards host adaptation to environmental changes by enabling a response to new challenges, such as exploitation of novel food sources [7,8]. Ultimately, gut microbiome can have major impacts on host development, behaviour and tness, with cascading effects to the dynamics of ecosystems [2]. In turn, it can also be modulated by several host traits, such as host taxonomy, sex and size, as well as the external environmental, such as habitat or prey availability [e.g. 9, 10). In addition, social interactions between hosts can also in uence the gut microbiome in many animal species, although these mechanisms remain less studied [see review by 11].
Gut microbiome dynamics has been studied in many mammals [e.g., 2], birds [e.g., 12], shes [e.g., 13] and amphibians [e.g., 14]. Comparatively, fewer studies have been performed in reptiles, and only a handful of these addressed lizards. Nevertheless, host taxonomy and ecology were seen to be important drivers of gut microbiota diversity in reptiles. For example, feeding habits in uence the gut microbiota of the Chinese crocodile lizard, Shinisaurus crocodilurus Ahl 1930, with potential effects on host health due to the in uence of diet on the abundances of pathogenic or opportunistic gut bacteria [15]. Diet and habitat of the Australian water dragon, Intellagama lesueurii (Gray, 1831), also have an effect on its gut microbiome, with lizards living in urban areas presenting higher bacterial diversity than populations living in natural habitats [16]. Moreover, host taxonomy and habitat also in uence the gut microbiota of venomous snakes [17].
Here, we analyzed and compared the diversity and composition of gut bacterial communities of ve phylogenetic related lacertid species captured in Portugal: Podarcis siculus (Ra nesque-Schmaltz, 1810), Podarcis virescens Geniez, Sá-Sousa, Guillaume, Cluchier and Crochet, 2014, Podarcis bocagei (Lopez-Seoane, 1885), Podarcis lusitanicus Geniez, Sá-Sousa, Guillaume, Cluchier and Crochet, 2014 and Teira dugesi (Milne-Edwards, 1829). Podarcis species are considered as model organisms to study ecotoxicology, immune/histochemical reactions, among other processes [e.g. 18,19]; however, microbiome studies are still largely lacking, with a single study on the Balearic Podarcis lilfordi (Günther, 1874) showing that islet and time since islet separation from mainland are signi cant factors contributing to gut microbiome structure [20]. Our main objective was to determine whether locality (which also corresponded to different habitats) and host factors such as species, size and sex modulate the gut bacterial diversity of these ve lizards. To achieve this, we used non-invasive sampling (cloacal swabs) to obtain a proxy for gut bacterial communities which were characterized by sequencing the V4 region of the 16S rRNA gene.
Podarcis bocagei and P. lusitanicus were collected from a semi-natural habitat in Moledo, northern Portugal ( Fig. 1d) (41°50'19.2"N 8°52'24.5"W), where they live in syntopy (i.e., occurrence of two species in the same habitat at the same time). This location has limited human disturbance and has lots of vegetation with natural and arti cial shelters (e.g., walls of agricultural properties) that can be used by lizards. Ecological adaptation is considered a major factor favoring the isolation between these two species; P. lusitanicus lives more on rocks, while P. bocagei is ground-dwelling [21]. The diet of these two species is mainly composed by prey belonging to Hemiptera, Coleoptera, Diptera, Hymenoptera and Araneae, with minimal differences between species or sexes [22]. Podarcis siculus and P. virescens were collected in Lisbon, at Parque das Nações (Fig. 1a, b) (38°76'22.4"N, 9°09'44.3 W), where both live in sympatry (sharing habitat type). This is a highly urbanized area near the Tejo river, characterized by large residential and commercial areas, with considerable daily human disturbance. While P. virescens is native to this location, P. siculus is an invasive species introduced about two decades ago [23]. Its plasticity in spatial use of habitat, morphology, behaviour, and versatile diet explains its successful colonization of multiple locations outside its native range [24][25][26][27]. This invasive species can present a more versatile diet, as it can also consume fruits and nectar [28] and have a more herbivorous diet [e.g. 24], while P. virescens is known to be insectivorous and to feed mainly on individuals of the class Arachnida and the orders Hymenoptera, Hemiptera, Coleoptera and Diptera [29]. Finally, we collected Teira dugesii in a nearby area in Lisbon, in the Alcantara docks, close to the city port area (38°70'33.8"N, 9°16'54.1"W). Similar to the other Podarcis spp. captured in Lisbon, T. dugesii occupies an anthropogenic area, although less busy, close to railway tracks with limited vegetation cover (Fig. 1c). This species is thought to have been accidentally introduced via transport ships from Madeira Island three decades ago, in 1992 [30]. Teira dugesii feeds preferentially on insects but also on small fruits [31].
All individuals were captured using nooses. Lizards were carefully immobilized, avoiding any human contact with the cloaca. We quickly inserted a sterile cotton swab into the entrance of the cloaca to obtain individual microbial samples. The tips of the swabs were cut into individual tubes and stored in ice boxes in the eld, and then stored at -80°C upon arrival in the laboratory. After the microbial sampling, each lizard was sexed, and the snout-vent length was measured (SVL; from head to cloaca) using a digital caliper (± 0.01mm error).
In the laboratory, we extracted the DNA from the swabs using the DNeasy® PowerSoil® Kit ( All analysis were performed using the R Software v.4.1.1 [33]. Raw FASTQ les were denoised using the DADA2 pipeline [34] in R with the parameters for ltering and trimming: trimLeft = 20, truncLen = c(220,200), maxN = 0, maxEE = c(2,2), truncQ = 2; and the SILVA 138 database [35,36] was chosen for taxonomic assignment. After quality control and taxonomic assignment, sequences identi ed as Archaea, Eukaryota, Mitochondria, Chloroplast, as well as sequences unassigned to bacteria were removed from the dataset. An Amplicon Sequence Variant (ASV) frequency table was constructed using the R package phyloseq [37] and normalized read counts were obtained using the negative binomial distribution implemented in DESeq2 [38]. ASVs with a count of less than 0.001% of the total number of reads (3586752 [total number of reads]x 0.001% = 36) and that were present in a single sample were also removed.
Bacterial taxonomic alpha-diversity (intra-sample) and beta-diversity (inter-sample) were estimated using the phyloseq package. Alpha-diversity was estimated using the number of observed ASVs, and the Shannon, Faith's Phylogenetic Diversity (PD). Beta-diversity was measured using the Bray-Curtis index and the Unifrac phylogenetic weighted and unweighted distances. Principal Coordinate Analysis (PCoA) were used to visually assess dissimilarity among groups.
Statistical differences in alpha-diversity between locality/habitat, species and sex were analyzed using a linear model (lm(alpha-diversity ~ locality + species + sex)). Given the signi cant effect of locality (which also corresponded to semi-natural and urbanized habitats) on alpha-diversity (see results section), differences in the proportions of the most abundant taxa at the phyla and genera level (represented by ≥ 1% on average of all sequences) were assessed between species and sex for each locality/habitat separately using a linear model (lm (bacterial phyla/genus ~ species * sex)). The effects of locality, species and sex on microbial beta-diversity were assessed using permutational analysis of variance (PERMANOVA) with 10000 permutations, implemented using the adonis function of the R vegan package [39] (adonis(beta-diversity ~ local + species + sex)). Correlations between individual size and bacterial alpha-diversity were also tested using the Pearson correlation test for each species, using the ggpubr package [40].
Bacterial transmission between each pair of species from sympatric populations living in Moledo and Parque das Nações was estimated using the FEAST software [41], by testing the contribution of each species (source) to the microbial diversity to its sympatric congener (sink). To this end, the nonnormalized ASV frequency table was used, and due to differences in the number of samples between P. bocagei and P. lusitanicus, only a fraction of the individuals of P. bocagei was included (with the most similar sex and SVL ratios to the P. lusitanicus samples as possible), following the FEAST developers' recommendations to avoid overestimation of transmission.
Pearson correlation test only showed signi cantly positive correlations between SVL and bacterial alphadiversity (for Shannon indice) for males of the in the invasive species P. siculus (online resource 4).
Results from FEAST software indicate that the level of bacterial transmission between sympatric species in both populations (Parque das Nações and Moledo) was high. Nevertheless, while between the syntopic P. lusitanicus and P. bocagei bacterial transmission was balanced in both directions (P. bocagei ◊ P. lusitanicus ~ 71% on average, and P. lusitanicus ◊ P. bocagei ~ 69% on average), the other two sympatric species showed a more biased transmission, with P. virescens having a higher contribution towards P. siculus (P. virescens ◊ P. siculus of 72% on average, and P. siculus ◊ P. virescens of about 55% on average).

Discussion
In this study, we characterized the gut bacterial microbiota of ve lizard species from Portugal (the native Podarcis virescens, P. bocagei and P. lusitanicus, and the introduced P. siculus and Teira dugesii) using a metabarcoding approach. Our results showed that locality was the main predictor of microbial diversity, signi cantly in uencing microbiota composition and structure. Moreover, host sex and size also had an effect, albeit more discrete, in gut bacterial communities. All lizards shared the same most abundant bacteria phyla with results being in accordance with what has been found in other studies in lizards [e.g. 42,43].
The two habitats in which lizards were captured are very different, with lizards from Lisbon living in an urbanized and arti cial habitat, with greater environmental disturbance, compared to lizards from Moledo, which live in a semi-natural habitat. Plausibly, differences in habitat may lead to differences in the composition and diversity of the gut microbiome [44]. We detected higher microbiota diversity in the more urbanized environment which could be explained by the higher variety of diet items, which may also include human food waste. Additionally, environmental microbiota, which could be horizontally transferred to lizards, may also be more diverse in urban habitat than in semi-natural ones [16]. We hypothesize that habitat disturbance, the co-existence with humans and urban animals (such as cats, dogs and rats), may in uence dietary behaviour and contribute to a higher bacterial load in the environment, which can then be acquired by lizards. These results agree with those on the Australian water dragon, where the gut microbiome of lizards in urban areas was more diverse compared to those residing in semi-natural habitats [16].
The proportion of some of the most abundant bacterial phyla and genera found in our study differed between lizard species at each locality. Additionally, we found that the species and sex also had an effect in the abundance of some of the gut microbiota components in P. virescens and P. siculus. The in uence of host taxonomy in gut microbiota, which is a proxy not only for host genetics but also its general ecology, has been reported in many animals [45,46] including reptiles [17]. The in uence of sex in the abundance of major bacterial groups, has also been reported in striped Plateau lizards (Sceloporus virgatus Smith, 1938), with sex-speci c cloaca microbiomes being related to different hormone levels [47]. In our study, the in uence of sex was small when compared to the effect of habitat and host taxonomy. Nonetheless, the in uence it exerts could also be linked to differences in the ecology of the two sexes. Importantly, the difference in size between female and male lizards may lead to slight differences in feeding behavior, and consequently on the gut microbial diversity.
Interestingly, we found a positive correlation between lizard lenght and alpha bacterial diversity in males from P. siculus. This lizard is larger than the other studied species and is also a very successful invasive species [25,26]. Indeed, P. siculus can be more aggressive than native Podarcis species [48], and also more exploratory, bolder, and better at exploiting food resources when compared to the native P. virescens in our study location [26,27]. These behaviours can be associated with the displacement of P. virescens from gardens now inhabited by P. siculus [49] and can also be leading to a wider ecological and trophic niche, and consequentially to a higher microbiome diversity in P. siculus.
Finally, our analysis of potential bacterial transmission between the lizards living in sympatry, indicates a balanced transmission between species at Moledo and unbalanced transmission between species in Lisbon, with the invasive P. siculus estimated to receive a higher proportion of bacteria from the native P. virescens. These differences in estimates of bacterial transmission may re ect different non-exclusive aspects of the ecology of the hosts, such as the similarity in dietary niches and habitat occupancy (i.e., horizontal bacterial transmission from the environment). Differences in the microbiome of P. siculus and P. virescens could be related to an increased habitat occupancy and successful adaptation to the environment by the invasive species, which facilitated the acquisition of a higher quantity of local microbiota upon its arrival. These results could also be re ecting an increased ability to exploit a variety of food resources, or most likely a combination of both. The populations of P. siculus and P. virescens live in sympatry, occupying roughly the same area, but rarely in syntopy, although sightings of these two species within 50 m of each other have been recorded [49, pers. obs.]. On the other hand, both Podarcis species from Moledo are considered syntopic and may have greater overlap and similarity in their habitat occupancy. Moreover, is very likely they consume the same or very similar prey items [22], and also encounter each other more frequently. All these factors may explain the much more balanced transmission we found between these two species.
The present study contributes to the existing knowledge on the effects of environmental and host factors on the dynamics of the gut microbiome of lizards. Our results also set the stage for future research exploring the in uence of other factors on the microbiome, particularly diet, as well as the use of sympatric Podarcis lizards as models to test the effects of behaviour on lizard microbial composition.  Boxplots of the alpha-diversity indices (Faith's phylogenetic diversity, Shannon diversity and the number of observed ASVs) for the gut microbiome of the studied lizards.