Invasive freshwater snails form novel microbial relationships

Abstract Resident microbes (microbiota) can shape host organismal function and adaptation in the face of environmental change. Invasion of new habitats exposes hosts to novel selection pressures, but little is known about the impact on microbiota and the host–microbiome relationship (e.g., how rapidly new microbial associations are formed, whether microbes influence invasion success). We used high‐throughput 16S rRNA sequencing of New Zealand (native) and European (invasive) populations of the freshwater snail Potamopyrgus antipodarum and found that while invaders do carry over some core microbial taxa from New Zealand, their microbial community is largely distinct. This finding highlights that invasions can result in the formation of novel host–microbiome relationships. We further show that the native microbiome is composed of fewer core microbes than the microbiome of invasive snails, suggesting that the microbiota is streamlined to a narrower set of core members. Furthermore, native snails exhibit relatively low alpha diversity but high inter‐individual variation, whereas invasive snails have higher alpha diversity but are relatively similar to each other. Together, our findings demonstrate that microbiota comparisons across native and invasive populations can reveal the impact of a long coevolutionary history and specialization of microbes in the native host range, as well as new associations occurring after invasion. We lay essential groundwork for understanding how microbial relationships affect invasion success and how microbes may be utilized in the control of invasive hosts.

determined by host traits such as genotype (Bowen et al., 2017;Parker et al., 2017), sex (Cong et al., 2016), infection of the host by parasites or pathogens (Jani & Briggs, 2014), environmental conditions (Aires et al., 2016;Ramirez et al., 2019;Wong & Rawls, 2012) and genotype by environment interactions (Macke et al., 2017). Therefore, both host organisms and the microbes on which they depend for vital processes can be impacted by environmental changes, with potential consequences for host adaptation (Corbin et al., 2017).
Host-associated microbes are expected to facilitate rapid adaptation, but the extent to which they may help invaders adapt to new habitats is relatively understudied (except see: (Aires et al., 2016;Coats & Rumpho, 2014;Kelly et al., 2014;Zilber-Rosenberg & Rosenberg, 2008)). When species invade novel habitats, they are exposed to new environmental conditions and selection pressures. Because host and microbe fitness can be mutually dependent, invasion is predicted to influence both the host and its microbiome (Aires et al., 2016). The fitness effects of microbes on their host may affect invasion success and the potential for subsequent adaptation to a new habitat (Coats & Rumpho, 2014). The performance of introduced species could also be altered by the availability of microbes outside their native range if there are core microbes the host needs to thrive. If host-associated microbes are not all carried over from the native environment, are lost upon invasion, and/or unavailable in the new environment, the microbiome diversity of invasive lineages could be lower than their native counterparts. The invader's microbiome might thus only represent a subset of microbial taxa residing in the native host lineages (Minard et al., 2015;Zepeda-Paulo et al., 2018). This pattern is directly analogous to the loss of genetic diversity that often characterizes invasive lineages (Minard et al., 2015;Städler et al., 2005). Whether such subsampling occurs at the microbiome level remains an open question with implications for host fitness in the new environment.
Invasion may affect the level of microbial community diversity in invasive hosts relative to hosts in the native range. Lower alpha diversity in invasive populations could result from the comparatively short time period for coevolution between hosts and newly acquired microbes (Minard et al., 2015;Zepeda-Paulo et al., 2018). If suitable beneficial microbes are not now available, the effects of colonization on microbial community composition could be harmful for the host. Alternatively, diversity may be higher within invasive hosts as they form new relationships with microbes (Himler et al., 2011) or if there is decreased immune-mediated control of novel microbes (Foster et al., 2017). By this logic, we might therefore expect native host individuals to have a less rich microbiota because established and stable microbial communities, in which novel competitive interactions between microbes have long since played out, are likely to be of lower diversity (Ghoul & Mitri, 2016). This prediction might be particularly likely to hold whether the host has evolved to cultivate beneficial associations (Foster et al., 2017), addressing whether these hypotheses require characterization of how invasion impacts microbial diversity and composition within and among populations.
Molluscs are one of the most species-rich phyla (Smith et al., 2011).
They represent an ecologically and economically important class of invaders (Alonso & Castro-Díez, 2008;Clusa et al., 2017); however, little is known about the relationships between molluscs and their resident microbes with the exception of a handful of recent studies (King et al., 2012;Neu et al., 2019). Here, we leveraged a powerful model system for host-parasite coevolution, the New Zealand freshwater snail Potamopyrgus antipodarum, to characterize the impact of invasion on host microbiota. There is substantial across-lake population variation in microbiome composition among native P. antipodarum (Takacs-Vesbach et al., 2016). Potamopyrgus antipodarum in native populations also feature wide individual-level variation in reproductive mode (sexual versus asexual) and ploidy level (sexual snails are diploid while asexual snails are polyploid (triploid or tetraploid) (Lively, 1987;Neiman et al., 2011)). They have also been demonstrated to be coevolving with sterilizing trematode parasites (Dybdahl & Lively, 1996;King et al., 2009). By contrast, invasive lineages of P. antipodarum in Europe and North America are polyploid asexuals with reduced genetic variation compared with the native range (Städler et al., 2005;Verhaegen et al., 2018) and have escaped from parasite infection (Alonso & Castro-Díez, 2012;Żbikowski & Żbikowska, 2009).
We conducted 16S rRNA sequencing of field-collected P. antipodarum across New Zealand and Europe to assess changes in microbiota diversity and composition across and within native and invasive host populations. We hypothesized several nonmutually exclusive possible outcomes: (a) microbial community composition (alpha diversity) of invasive P. antipodarum would contain a subset of the taxa observed in native snails (analogous to patterns often described for genetic variation postinvasion). Lower alpha diversity could consequently rise because invasive snails have less time to form new microbial relationships or some native microbial taxa are not now available to acquire; (b) microbe taxa would be at similar relative abundance in invasive populations if snails actively promote taxa in their native environment and these are available to acquire in the invasive range; and (c) invaders also could have fewer conserved microbes than native counterparts due to relatively recent contact in the new environment and less time for specialization. We also address associations between the fundamental and often-variable organismal traits of sex, reproductive mode and infection status on microbiota within native host populations. More broadly, this study provides a starting point to evaluate whether and how microbes could be utilized to understand the biology of invasive host species and potentially how to control them.

| Study System
Native New Zealand P. antipodarum are primarily found in freshwater lakes and streams (Winterbourn, 1968). These snails have gained prominence as a model for studying the evolutionary maintenance of sexual reproduction (Lively, 1987;Neiman et al., 2011), as well as host-parasite coevolution (King et al., 2009;Lively, 1987) with a sterilizing trematode parasite, Atriophallophorus winterbourni (formerly Microphallus livelyi) (Blasco-Costa et al., 2020;King et al., 2009;Lively, 1987). Multiple triploid and tetraploid asexual lineages have been separately derived from diploid sexual conspecifics (Lively, 1987;Neiman et al., 2011). A previous study suggested that the microbiome composition of these snails varies substantially among native populations and between sexual and asexual forms (Takacs-Vesbach et al., 2016). However, these snails were either laboratory-cultured or housed in a laboratory for several months before harvest (Maurine Neiman, personal communication), and whether the results in (Takacs-Vesbach et al., 2016) hold in snails sampled directly from the field remains unclear.

| Sample collection, processing and sequencing
We collected adult Potamopyrgus antipodarum during warm seasons from shallow (lake depth <1 m) rocks and vegetation from three New Zealand collection sites (early January 2015) and ten European collection sites (late May 2016) (Table 1; Figure S1). Snails were maintained in identical 15-L tanks separated by source population in a 16°C room with a 16:8 hr light:dark cycle for less than one month before dissection. Snails were fed dried Spirulina cyanobacteria ad libitum (Zachar & Neiman, 2013).
All snails were sexed and shells were removed prior to DNA extraction and assessed for A. winterbourni infection based on the presence of metacercariae cysts via dissection   Note: The coordinates and location names of sample sites for both native and invasive snails are list below. Ploidy is a proxy for reproductive mode; diploids are sexual, polyploids are asexual. We also include the sex, infection status and number of individuals sequenced from each condition and location and condition.

TA B L E 1 Sample collection sites
( Table 1). Snails containing non-A. winterbourni infections were excluded from the study. All metacercariae were removed from infected snails using a micropipette. After dissection, we used flow cytometry (Krist et al., 2017;Neiman et al., 2011)  Methods in Appendix S1 for more details about sample collection and processing.
The 16S rRNA V4 region was amplified from the P. antipodarum microbiome gDNA using the 515F Golay-barcoded primers and 806R primers (Apprill et al., 2015;Caporaso et al., 2011). Samples were prepared in accordance with the standard Earth Microbiome Project 16S rRNA protocol (Caporaso et al., 2012). Please see our Electronic Supplementary Methods in Appendix S1 for PCR mixtures and thermocycler conditions. gDNA was quantified using the

| Computational and statistical analyses
We removed PhiX sequences from sequencing libraries using We employed the Analysis of Composition of Microbes (ANCOM) algorithm to conduct a differential count analysis (Mandal et al., 2015). These analyses were limited to the Lake Alexandrina

| RE SULTS
Variation in P. antipodarum microbiota was in large part driven by whether a snail was collected from Europe or New Zealand (Figure 1; ANOSIM; R = .97; p < .01; Permutations: 999; ADONIS; R 2 = .33, p < .01; weighted UniFrac dissimilarity distances are presented in Figure S2). To validate this analysis, we performed a Mantel test on coordinate and microbiota distance matrices and found geographic distance was significantly associated with microbiome distance (Mantel test; Observation = 0.13; p < .01). Within the two European and New Zealand regions, we did not find significant differences in microbial variation among sample sites. We thus attribute the geographic distance-to-microbiome distance correlation to major differences between microbiomes from the most geographically distant regions of Europe and New Zealand, rather than differences among sites within the two larger regions. We also observed that the average distance from beta diversity centroids was higher in New Zealand  Figure S3). A bar chart of microbial taxa at the family level for each sample site is presented ( Figure S4).
To identify taxa that significantly predicted ecosystem cluster- European or New Zealand ecosystems with 99.6% accuracy. We corroborated these results using a reference frames approach (Morton et al., 2019) and identified taxa associated with snails from the two regions. Eight phyla were exclusively associated with European snails and none with New Zealand snails (Table 2; Figure 1). However, there were more Firmicutes taxa associated with snails from New Zealand (9 taxa) than Europe (6 taxa) ( Table 2; Table S1).
To test the degree to which microbiota exhibit host population-specific distributions, we calculated the core microbiome of snails from Europe and New Zealand, defined at the tip agglomerated level. European snails had a core microbiome (found in >=90% of samples) of 30 taxa while snails from New Zealand had one core taxon (Figure 2a), Arenimonas, also in the European snail sore microbiome. Our results thus departed from the prediction of higher core microbiome specificity for in the New Zealand region. We found that the core microbiome of European snails constituted, on average, 13.8% of their microbiome. By contrast, the core microbiome of New Zealand snails constituted 70.1% of their microbiome, a marked and statistically significant difference between native and invasive snails (Figure 2b; t test; p < .01). Proteobacteria were the main constituent of the core microbiota in both snail populations, making up 100% (1/1) of the New Zealand snail core and 68.9% of the taxa in the European snail core (20/29).
We next evaluated whether there were differences in species richness between European and New Zealand snail microbiota. We found that European snail microbiota (mean: 116 ASVs, se: 2.51) had on average 3.52x more ASVs than those in New Zealand (mean: 32.9 ASVs, se: 2.56) (Figure 3; t test; adj-p < .01; t = −23.1). There was significantly higher Chao1 and PD Whole Tree diversity among European snail microbiota, suggesting higher richness of rare species at relatively low abundance and higher phylogenetic diversity ( Figure 3; t test; adj-p-values < 0.01; Chao1 t = −25.2; PD t = −24.8).
Shannon diversity was also significantly higher in European snail microbiota (Figure 3; t test; adj-p < .01; t = −4.52). Our prediction that New Zealand snail microbiota may have higher Pielou's evenness than the European snails was upheld (Figure 3; t test; adj-p < .01; t = 22.4). To confirm that alpha diversity differences between snail microbiota in Europe and New Zealand were not due to uneven sampling, we permuted beta diversity tests on all combinations of three, randomly selected Europe sites versus the three New Zealand sites ( Figure S3). Among the 600 combination tests, 98.3% remained significant (Table S2; t test; adj-p < .05). The nonsignificant findings were for Shannon diversity comparisons between European sites including West Holme and Melle compared with New Zealand sites.
There were thus more species, more rare species and higher phylogenetic richness in European snail microbiota, but higher species evenness in New Zealand snail microbiota.
Our comparisons of snails from Lake Alexandrina show that sex (R 2 = .147; adj-p = .001), reproductive mode (R 2 = .0467; adjp = .001) and infection status (R 2 = .0330; adj-p = .005) all are associated with microbiota composition (Table 3; ADONIS; Permutations: 999). We controlled for these covariate effects and highlight taxa that significantly differed in abundance by sex and reproductive mode ( Figure S5; ANCOM; adj-p < .05). No taxa significantly differed in abundance based on infection status. We identified ten ASVs (all Xanthomonadaecae) that were significantly higher in abundance in asexual versus sexual snails ( Figure S5). We observed 50 ASVs (Niabella (n = 16), Bacillus (n = 15) and OM60 (n = 19)) that were significantly more abundant in male versus female snails. There was significantly higher ASV richness in uninfected compared with infected snails (t test; adj-p < .01; Figure S6). There were no significant differences in observed ASV richness when comparing reproductive mode or sex ( Figure S6).

| D ISCUSS I ON
As organisms invade new habitats, resident microbes can be lost (Minard et al., 2015;Zepeda-Paulo et al., 2018). Microbes might alternatively be gained if new associations are readily formed, possibly with less host control (Foster et al., 2017). We predicted that invasion by a host organism would be reflected in changes to the composition and/or diversity of its microbiota. We found that native New Zealand and invasive European snails had distinct microbiotas.
While invasive snails carried many core taxa from New Zealand, they also harboured more core members overall and higher phylogenetic Note: We used a multinomial regression to predict whether snails were from Europe or New Zealand samples sites.

TA B L E 2 Europe and New Zealand microbiome phyla predictors
Represented here is a summary of counts of phyla that were associated with snails from EU or NZ.
diversity. These results suggest that invaders can form many new microbial associations, which may play a role in the success of this worldwide invasive snail. The results also suggest that microbes inhabiting invasive snails may be opportunistic and less specialized.
This pattern may be a consequence of the comparatively short time for host-microbiome co-adaptation, relative to the potential for longer coevolutionary relationships between New Zealand snails and microbes in the native range. In new environments, host-control mechanisms may not have yet evolved to keep microbial diversity in check (Foster et al., 2017).
The single core microbial taxon from native New Zealand snails (Arenimonas) was found in European invasive snails. There could thus be selection to maintain certain symbiont associations in both ranges, which may be important to host biology. Because we were unable to obtain environmental samples due to insufficient amounts of DNA in water samples, and it is unclear whether some core microbes are inherited, we cannot say whether core microbes found in European and New Zealand snail populations were present in both environments, or carried over from the latter. Nevertheless, European samples had 30 times more core taxa overall, and their most predictive taxa were absent or relatively infrequent in native snails. In other systems, it has been shown that invaders have lower microbial diversity and/or a subset of microbes relative to that present in native populations (Minard et al., 2015;Zepeda-Paulo et al., 2018). Invasive P. antipodarum populations feature low genotypic diversity relative to the native range (Donne et al., 2020;Dybdahl & Drown, 2011;Verhaegen et al., 2018) and mitochondrial haplotypes of invaders represent genetic subsets of native mitochondrial haplotype diversity (Städler et al., 2005). However, this genetic subsampling in inva- Violin plot summarizing the proportion of amplicon reads mapped to core. Significance is denoted as *** = p < .001 (t test) microbiome down to specialized taxa (Foster et al., 2017). A similar phenomenon occurs in a number of insect systems, whereby the host harbours few core microbes that are key players in organismal function (Chandler et al., 2011;Raymann & Moran, 2018). Conversely, other insect taxa are host to relatively diverse core microbial populations (Tinker & Ottesen, 2016).

Reciprocal antagonistic selection between P. antipodarum and
A. winterbourni is expected to be strong because infected snails are completely sterilized (Winterbourn, 1973), and parasites die if they encounter resistant individuals (King et al., 2011). In other systems, the potential for beneficial microbes to be involved in host defence and even host-parasite coevolution has been demonstrated (Dillon et al., 2005;Ford & King, 2016;Vorburger & Perlman, 2018 Because asexual P. antipodarum are more successful invaders than sexuals, whether the microbiome could influence the success of asexual invaders is of interest. Here, we found that reproductive mode was associated with microbiome composition and ten ASVs that were significantly more abundant in asexuals than sexuals. Takacs-Vesbach et al. (2016) observed significant differences in microbiome composition between sexuals and asexuals and among lake populations of P. antipodarum. Because the mechanism(s) leading to transitions from diploid sexuality to polyploid asexuality in P. antipodarum are still unclear, future work could focus on the possible links between microbes and reproductive mode and/or polyploidization.
For example, Wolbachia-mediated transitions to asexuality are well characterized in arthropods (Boivin et al., 2014). In regard to potential effects of ploidy level, Takacs-Vesbach et al. (2016) did not observe significant differences in microbiome composition between asexual triploids and tetraploids, and we were unable to obtain sufficient sampling of tetraploids to perform ploidy level comparisons.
Furthermore, besides a handful of studies (Cavé-Radet et al., 2019; Forrester & Ashman, 2017), little is known about links between host ploidy level and microbiome composition, especially in animals.
While it is established that resident microbes have important effects on host fitness, the relationship between microbes and invaders has received little attention. Here, we find that the process of invasion is associated with some carryover of a core microbiota, but mostly involves the formation of numerous new microbial relationships, indicating that successful invaders, like P. antipodarum, might take advantage of available microbes. Despite some impact of biological trait variation on the microbiota in the native range, there is a consistent and strong pattern highlighting that long-term hostmicrobiota interactions, and possibly coevolution (or host evolution alone (Foster et al., 2017)), could streamline the microbiota. Future research will need to establish the fitness impacts of those new microbes on invaders, and whether the invasive snails have adapted to accommodate them. Notwithstanding this possibility, the potential for invasion to alter the host microbiome is clear. Endeavours to control invasive species could focus on these new associations being formed with microbes and investigate their contribution to invader success. The "make or break" of an invasion may hinge upon the ability of an introduced species to take advantage of the new microbes it encounters.

CO N FLI C T O F I NTE R E S T S
None declared.

DATA AVA I L A B I L I T Y S TAT E M E N T
Raw sequencing data are publicly available on the European Nucleotide Archive under project accession PRJEB40881, and sample accessions ERS5219283-ERS5219479 (SAMEA7461104-SAMEA7461300).