To our knowledge, this is the first study to characterize the gut microbiome, or any microbiome, of a member of the land snail family Oreohelicidae. In these results, we will present the overall taxonomic composition of O. strigosa gut microbiomes and members of the core microbiome. Then, we present differences between the three treatment groups of adult, fetal, and starved snails both taxonomically, and in terms of diversity metrics.
Overall Species Microbiome:
Sequencing Statistics:
Next-generation sequencing of 16S rRNA sequences was employed to assess the gut microbiomes of 50 non-starved adult snails, 6 starved adult snails, and 12 fetal snails. We obtained 2,714,330 reads in total from the 68 samples. Of those reads, 7,056 unique OTUs were identified. The number of OTUs per sample ranged from 23 to 1,559. The number of reads per OTU ranged from 0 to 39,012 reads (this was an OTU from genus Mycoplasma), and the average number of reads per OTU was 384.68. The alpha- diversity indices indicate a high diversity of the O. strigosa bacterial community, as compared with other snail species. For example, there were 7,065 total OTUs in O. strigosa, versus 1,196 OTUs in freshwater snail R. Auricularia [53]. The invasive land snail Achatina fulica only had 228 OTUs [8].
These discrepancies may be driven by the snails’ differential dietary and habitat needs. O. strigosa primarily feeds on decaying lignocellulosic matter, rather than fresh vegetation. Lignocellulosic matter may be more difficult to digest, thus calling for a more diverse set of commensal bacteria to assist digestion. R. auricularia is a freshwater species that likely has different dietary needs than a terrestrial snail, consuming more aquatic plants and algae than woody matter and dry leaf litter. Dietary diversity may also be directly correlated with habitat diversity and may in turn impact bacterial diversity in guts. The land snail Achatina fulica does have a wide-ranging diet, as it is an invasive species that adapts to a broader range of environments. In Cardoso et al. [8], A. fulica was collected from one locality in its invasive range. This narrow sampling range might be driving the lower number of OTUs recovered. Further broader sampling may reveal different microbiome diversity patterns.
Taxonomic Composition of Gut Bacteria:
The taxonomic composition of Oreohelix strigosa gut microbiomes proved to be highly diverse (Fig. 2A). A total of five phyla accounted for 96.48% of the total sequences across all samples. Proteobacteria (61.29%) and Bacteroidetes (25.44%) were the most dominant bacterial phyla followed by Verrucomicrobia (5.36%), Firmicutes (3.22%), and Actinobacteria (1.17%). A total of 31 other classified phyla each comprised less than 1% of the relative abundance; and one phylum was not classified. Proteobacteria contained the largest number of OTUs (2,835) which belonged to the following classes: alpha-, gamma-, beta-, delta-, epsilon-, and zeta- proteobacteria, with alphaproteobacterial contributing the majority of OTUs (1,613), followed by Bacteroidetes with 1,146 OTUs and Actinobacteria with 956 OTUs.
Our results showed that within the phylum Proteobacteria, gammaproteobacteria constitutes the largest proportion, and within gammaproteobacteria, the order Enterobacteriales is the most abundant. Members of Enterobacteriales are widely common gut bacteria, having been reported across animal phyla including vertebrates and invertebrates [54].
There were 20 identifiable bacterial families with >1% abundance across all samples, which accounted for 82.03% of the total sequences. Among them, Enterobacteriaceae (41.95% of sequences), Sphingobacteriaceae (15.84%), Flavobacteriaceae (4.75%), Pseudomonadaceae (4.48%), and Verrucomicrobiaceae (4.43%) were the most common families.
Over half (53.15%) of genera were not classified. There were 10 genera with >1% abundance across all of the samples; these included Sphingobacterium, Pseudomonas, Flavobacterium, Serratia, Pedobacter, Acinetobacter, Sphingomonas, Yersinia, Enterococcus, andLuteolibacter, with abundances ranging from 1.14% to 13.06%.
The proportional dominance of members of these taxonomic levels is consistent with other snail gut microbiome studies. The phyla Proteobacteria, Bacteroidetes, and Actinobacteria were also the dominant phyla in Planorbid snail intestines [55]. Proteobacteria (and within, specifically alpha- and gammaproteobacterial classes) similarly has been identified as the dominant bacterial phylum in the gut microbiome in other snails, including Biomphalaria pfeifferi, Bulinus africanus, Helisoma duryi [55], Achatina fulica [8, 9], Helix pomatia [11], and Radix auricularia [53]. However, our results show a striking higher relative abundance of Proteobacteria (61.29%) compared to that of other snails. For example, in the big-ear radix, Radix auricularia, Proteobacteria only accounts for 36.0% of sequences in juvenile snails and 31.6% in adults [53]. Other terrestrial snail guts also contain a majority of Enterobacteriaceae members, including the genera Butiauxella, Citrobacter, Enterobacter, and Kluyvera [56]. One species from genus Sphingobacterium, S. multivorum, has been isolated from the giant African land snail A. fulica [57, 23]. The higher proportion of Proteobacteria could be attributed to any number of life history traits, such as diet, trophic level, farmed versus wild species, method of reproduction, and infection with parasites.
Members of the gut microbiome are often expected to aid in the food ingestion, digestion and nutrient absorption of the host [9]. In snails, food is scraped by the radula and combined with salivary gland secretions after being ingested by the buccal mass and digested in the stomach [23]. In this process, gut bacteria play an important role in capturing energy from digested plant biomass. Snails may use their resident gut bacteria to degrade and ferment cellulose, hemicellulose, and lignin, all of which are common to many of their diets [23].
Overall, the known symbionts associated with the O. strigosa gut are expected to aid digestive functioning consistent with the inferred diet of this snail. As with some other land snails, O. strigosa feeds preferentially on decaying wood and leaf litter rather than fresh leafy greens [58]. As such, bacterial symbionts with functional ability to help the host digest complex molecules, like lignocellulosic matter, are consistent with the needs of this snail host. Cellulose degrading bacteria isolated from the gut of different snails are also found in O. strigosa, these include many members of Enterobacter, Bacillus, and members of genus Sphingobacterium, including S. multivorum. Lactic acid bacteria, responsible for fermentation in other snail guts, were similarly also found in our samples, including members of Enterobacter, Lactococcus, Butiauxella, and Enterococcus (like E. casseliflavus) [23]. Although we cannot test bacterial functions directly with our microbiome profiling data, it is highly likely that O. strigosa gut microbiome aids its digestive and fermentative processes.
Core Gut Microbiome:
Venn analyses found that 11.34% (800 OTUs) were common to all three treatment groups of the total 7,056 OTUs identified (Fig. 2B). The non-starved adult snails share 22.15% and 49.42% OTUs with the fetal snails and starved snails respectively.
To further determine the members of the core microbiome, which is the most stable part of the microbiome, we first identified the families and genera that were present in all treatment groups of individuals. Of the 800 common OTUs, the largest proportion of OTUs came from families Sphingomonadaceae, Chitinophagaceae, Flavobacteriaceae, and Sphingobaceteriaceae, and within those, genera Sphingomonas, Flavobacterium, and Pedobacter. The most common genus found across all groups was Sphingomonas (4.34% of the common OTUs).
We then determined which OTUs were present in 100% of samples. Only four OTUs were common to every sample’s gut microbiome: OTU_9 and OTU_10, unidentified members of family Enterobacteriaceae; OTU_1, Sphingobacterium faecium; and OTU_17, an unidentified member of genus Sphingomonas. Together, these four OTUs contributed to roughly 50% of total reads. 19 OTUs were found in at least 90% of all samples, which made up roughly 62% of all reads. The 43 OTUs found in at least 80% of all samples made up 68% of all reads.
The fact that the four core gut microbes contribute to half of the relative abundance across all samples might reflect their ecological importance as beneficial symbionts to the snail host. It is worthy to note that these bacterial strains were found in the unborn fetal snails, suggesting that they may be passed down through vertical transmission (see discussion in fetal and adult comparison section) from parent to offspring. Therefore, these strains may confer some evolutionary advantage at birth for these snails. Members of Enterobacteriaceae and Sphingobacteriaceae are known cellulose-degrading bacteria in other snail species [23]. While less functional information is known regarding Sphingomonas species, members of the genus have been found in the freshwater snail Biomphalaria glabrata and are thought to play a role in immune functioning, specifically in parasite defense [59]. In the land snail Cornu aspersum, cellulose-degrading bacteria and lactic acid bacteria persisted in every life stage of the snail examined, which is consistent with the four core bacteria found here [60]. This demonstrates that O. strigosa likely maintains an obligate group of bacterial symbionts that are important to the snails’ survival.
Comparison Between Treatment Groups:
Taxonomic Composition
Fetal vs. Adults
Fetal gut microbiome samples consisted of fewer OTUs than their adult counterparts. In fetal samples, a total of three phyla accounted for 97.32% of the total sequences. Proteobacteria (86.92%) was the most dominant phylum followed by Bacteroidetes (8.03%) and Firmicutes (2.37%). The lack of Verrucomicrobia at abundances >1% in fetal samples, as opposed to the phylum’s larger presence in every other group, indicates taxa from this phylum might be horizontally acquired in adults. Some snails are known to eat soil to augment their gut microbiome [23]. Verrucomicrobia is generally abundant in soil and freshwater samples, and so could likely be found in the soil substrate that O. strigosa live and burrow in. Other phyla that were common in similar abundances across snail groups, like Proteobacteria, Bacteroidetes, and Firmicutes, may be vertically transmitted from parent to offspring. A small number of OTUs were present only in fetal samples and not the other treatment groups, this indicates that vertical transmission of highly transient bacteria may exist, explaining why these OTUs were not found in adults.
The clustering of gut bacteria by host life stage was highly significant in the PERMANOVA analysis. A non-metric multi-dimensional dimensional scaling (NMDS) plot showed that fetal samples formed a distinct cluster but could not be separated from adult snail samples (Fig. 2C, upper). The R2-value indicates that the variability attributed to life stage as a factor is about 9.5% (PERMANOVA: p-value < 0.001, R2-value = 0.095).
Fetal O. strigosa are an important group for studying how life stage influences gut microbiota because unlike many other land snails, these snails are ovoviviparous and give live birth. Rather than extracting DNA from an egg mass, we can dissect the fetal snails directly from the parent to reduce the impacts of any environmental contamination. In other animals, fetal microbiomes are similarly present, whether acquired directly from the mother’s oviduct or through environmental seepage, like through eggshell pores [17]. The majority of existing literature looking at microbiome changes over varying life stages use vertebrate species as models. Therefore, much of the following discussion compares our study to these vertebrates. For example, chickens show a similar pattern of higher relative abundance of Proteobacteria earlier in life. Members of Proteobacteria in chicks and in snails may be poor competitors that are unable to compete and persist in a mature gut microbiome [4]. As young chickens advance in age, even beginning as early as seven days old through 42 days, their microbiome becomes more diverse, showing increased relative abundances from more phyla [17]. This is because gut microbiome succession is dependent on nutrition and the establishment of new bacteria through exogenous food sources [4]. As a case in point, when rabbits are fed only milk in the first days of life, they possess no cellulolytic bacteria and are unable to digest plant matter [61]. Since some bacteria is found in unborn, fetal snails, they may have an advantage in already possessing some mutualistic gut bacteria. As snails age and ingest food, the gut microbiome likely becomes more and more diverse until it reaches the composition of the adult microbiome. Further studies of varying life stages of Oreohelix could provide more detailed insight into how changes in diet and age can affect the makeup of the gut microbiome.
Means of reproduction may also influence microbiome composition. Asexually reproducing New Zealand mud snails (Potamopyrgus antipodarum) are dominated by a strain from the genus Rhodobacter, while sexually reproducing individuals are dominated by a strain from the order Rickettsiales. This association suggests the snail’s reproductive strategy has certain levels of influence on the assembly of its microbiota [62]. Oreohelix strigosa can self-fertilize or mate, but seem to prefer mating as little research has investigated self-fertilization [63, personal observation]. Means of reproduction may be an important factor that alters the microbiome of fetal snails in Oreohelix, but this needs further investigation.
Starved vs. Non-Starved
Starved snails had ten phyla showing relative abundances greater than 1%, compared with five in the adult group and three in the fetal group. Starved snails showed increased relative abundances of the phyla Acidobacteria, Planctomycetes, Cyanobacteria, Spirochaetes, and Tenericutes, which are present in the other two treatment groups but at low abundances.
A non-metric multidimensional scaling (NMDS) plot showed that adult starved gut microbial compositions were not distinctively separate from adult non-starved samples. Visually, the starved samples clustered within the greater cluster of non-starved samples (Fig. 2C, lower). PERMANOVA analyses showed the effect of starvation, though significant, explained only 3.08% of the variance between samples (PERMANOVA: PERMANOVA: p-value < 0.002, R2-value = 0.038).
Starved sample microbiomes appear to be largely a subset of adult sample microbiomes. Only 119 OTUs were unique to starved samples, compared with over 2,000 in adults. Other studies have found decreased OTU diversity in starved organisms. The gut of tunicate Ciona intestinales shows a similar pattern of OTU presence in starved and non-starved individuals; a relatively small amount of OTUs was shared between these two groups [64]. C. intestinales also similarly showed a greater number of unique OTUs in only the non-starved group, compared with a smaller number (still greater than the number of OTUs found in both groups) found only in the starved group [64]. This trend is also shown in fresh-water crayfish, where starved samples showed genera (largely from genus Vibrio) that were present, but rarer in fed samples in high relative abundances [65]. Unlike our study, these genera came to make up the majority of the starved microbiome, rather than simply becoming more abundant [65]. This discrepancy may be due to the much longer starvation period in the crayfish study that may have allowed Vibrio species the time needed to replace core bacterial species (four weeks, compared with our one week period). Prolonged starvation periods may give the microbiome more time for rarer species to take over the community, giving way to a rapid shift in bacterial diversity and relative abundance of some species. However, it is likewise difficult to denote a universal response against starvation, as tolerance levels vary across different host species [65, 66]. Future studies should attempt to increase the length of the starvation period to see if similar trends occur in Oreohelix.
Lastly, we conducted the same PERMANOVA test using only samples found in one location (MRS), to validate the results without the impact of locality. In this test, the significant results of age and starvation hold up, with age explaining roughly 9.3% of the variation between samples (p-value < 0.001) and starvation explaining 5.5% (p-value < 0.05).
Diversity Metrics
We used the non-parametric Kruskal-Wallis test to test for significant differences among the means of the richness, evenness, and Shannon index of the three groups (non-starved, starved, and fetal). The tests returned significant p-values for each metric (Evenness: Kruskal-Wallis chi-squared = 10.464, df = 2, p-value = 0.005342; Richness: Kruskal-Wallis chi-squared = 13.233, df = 2, p-value = 0.001338; Shannon Index: Kruskal-Wallis chi-squared = 13.156, df = 2, p-value = 0.001391, Fig. 2D). The starved group show significant higher evenness compared to others, and the fetal group show significant lower mean richness and Shannon diversity. In the bootstrap analyses where we compared fetal and starved samples with random samples of 6 and 12 non-starved adults, the significantly higher microbiome evenness in starved snails was found in 48% of the bootstrapped replicates; the lower Shannon Index scores in fetal samples was true for 52% of the replicates; and the lower fetal microbiome richness was supported by 93% of the replicates.
Fetal vs. Adult
Microbial ecology theory predicts that microbial species passed vertically from parent to offspring are more likely to be beneficial to the host, while species gained horizontally from the environment are opportunistic and more variable [67]. This is consistent with our finding of a significant lower richness in fetal snails compared with adult snails.
Across both vertebrates and invertebrates, there are many examples of vertical transmission of microbiomes. Demonstrated examples include insects [68, 69], sponges [70], bivalves [71, 72, 73], and cephalopods [74, 75, 76, 77]. In sponges, members of up to ten bacterial phyla and one archael phylum have been vertically transmitted from adult sponge to other life stages, including oocytes of oviparous sponges and embryos of viviparous sponges [78]. In chemosynthetic vesicomyid clams, vertical transmission also appears to be the main mechanism for maintaining thioautotrophic bacterial symbionts [71, 78]. In many of these animals, the deposited microbiome is not as rich as the adult microbiome, and compositional changes happen throughout maturation likely due to further horizontol transmission. For example, amylolytic bacteria are vertically transmitted in the land snail Cornu aspersum, whereas transient proteolytic and cellulolytic bacteria are gained through environmental augmentation when the adult snail is active [23, 59]. Importantly, the fetal snails examined here were unborn and therefore completely unexposed to their surrounding environment, so examining more life stages of these snails, like juveniles, could shed more light on which bacteria are adopted from parents versus environments.
Starved vs. Non-Starved Snails
Starved samples showed significantly greater evenness than non-starved adults, but not always in the bootstrap analyses. Starved samples also showed nonsignificant differences to richness and Shannon Index scores compared with adults.
Limited studies have assessed the effect of starvation on the gut microbiome in animals. The gut microbiome in cod [79], seabass [80], and shrimp [81] indicate that they can rapidly adapt to starvation stress. In shrimp, diversity indices between starved and non-starved groups were similarly not significantly different from one another. This lack of overall difference in community structure is hypothesized to be attributed to a few microbial members being more sensitive to disturbance, while the greater community is resistant [82]. These few members could be microbial “gatekeepers,” which contribute disproportionately to the functioning of the gut microbiome and the overall health of the host. If these gatekeepers are lost due to starvation stress, profound shifts in function of the gut microbiome could occur [81]. So, while diversity measures like richness between starved and non-starved groups are non-significantly different in this study, there may be significant changes happening in the actual functioning of the gut microbiome. Future studies should utilize -omics methods (e.g. metagenomics) to investigate how functional roles of the microbiome change due to starvation stress.
In addition, the lack of significant responses of richness and diversity, compared with higher evenness may be due to the effects of dysbiosis in the snail. Particularly, this may be consistent with the “Anna-Karenina Principle” for bacterial microbiomes in response to stress [83]. Following this principle, a host may not be able to regulate its microbiome when experiencing stress, and thus its microbial community takes on a more randomly distributed community structure. Thus, host individuals will not show consistent microbial community shifts in response to disturbance, though an overall separation between stressed and non-stressed populations may be obvious [84]. In invertebrates, stressed coral populations have also been reported to show increased dispersion in their microbial communities [83]. Similarly, ocean acidification can increase microbiome variability in sea sponges [85]. The effects of dysbiosis in starved organisms are important to elucidate as a transformed gut microbiome can lead to altered, poorer immune functioning in hosts and increase risk of disease from pathogenic bacteria [65, 86]. Signs of dysbiosis in snails may be particularly useful for researchers of montane ecosystems because snails act as bioindicators – their highly sensitive disposition reflects the impact of climatic stress on their environment [87].
In addition, there does not seem to be consistent microbiome responses to starvation across different animal groups. For example, in contrast to our results, the evenness of starved abalone was lower than fed abalone [88]. Due to a wide variety of microbiome responses across animals, the effect of starvation may need to be examined on a species-by-species basis. As with other studies [89], it may be useful to conduct a starvation study across a diverse array of gastropods to ascertain any shared responses.
Another reason why a universal response to starvation cannot be detected is because many host species go through fasting periods in their natural life cycle. For example, crustaceans must fast while they molt [90]. The Chinese alligator [91] is a natural hibernator that can fast for several months due to lower temperatures and unavailability, as are many mammals including bears [92] and squirrels [93]. In all of these species, microbial diversity has been impacted due to seasonal fasting, though not necessarily in a negative way. Oreohelix strigosa typically enters a dormant state of aestivation during periods of hotness or dryness that would lead to desiccation [94]. Therefore, there are many times when feeding is not a daily occurrence. The unpredictability of O. strigosa’s lifestyle may enable more gut microbiome adaptability to periods of fasting, therefore not showing dramatic changes in diversity patterns.