In this study the gut microbiota composition in two species of Australian freshwater chelonians, the eastern longneck turtle and the Macquarie River turtle, are described. We found that in both species the predominant phylum was Proteobacteria, followed by Actinobacteriota, Deinococcota, Bacteroidota, and Firmicutes in the eastern longneck, and Deinococcota, Bacteroidota, Actinobacteriota, and Cyanobacteria in Macquarie River turtles. There is considerable variability of gut microbiota composition in chelonians, with Proteobacteria dominating in other species such as map turtles (Graptemys pseudogeographica) [9, 11], wild Beal’s eyed turtles (Sacalia bealei) [8], and also in all species of female sea turtles that were sampled during nesting [42]. However, Firmicutes has been reported as the most prevalent phylum in other chelonians including Seychelles giant tortoises (Aldabrachelys gigantea) [43], southern river terrapins (Batagur affinis) [7], captive Beal’s eyed turtles [8], painted turtles (Chrysems picta) [10], one population of flatback turtles (Natator depressus) [29], and a range of sea turtles under different age, sex and health classes [44]. Interestingly, a predominance of Firmicutes was also seen in Macquarie River turtles captured from undisclosed waterways in Queensland, Australia [6, 31]. Making meaningful comparisons between studies to explain observed variance is difficult as microbiota investigations are complicated by differences in methodology [45], phylogenetic history of species sampled [42], physiological state of hosts [46], and local environmental factors [47]. Therefore, until protocols to standardise field and laboratory practices have been agreed upon, caution should be taken when inferring which factors have physiological effects on microbiota based on differences between studies.
This investigation into the gut microbiota in freshwater chelonians has shown that bacterial communities differ significantly among turtles of the same species but originating from geographically distinct populations. These differences were apparent in both composition and diversity and highlight the importance of interpreting microbiota data in the context of locality. Geography has been shown to play a key role in driving gut microbial ecology in vertebrate hosts and is likely to be related to deterministic processes such as food availability, inter/intraspecific competition, host dispersal and local environmental factors [48–51]. However, these findings are not consistent across all vertebrate classes with location not an important factor in determining microbiota composition in eastern garter snakes (Thamnophis sirtalis sirtalis) and northern watersnakes (Neridia sipedon sipedon) [52]. Correctly identifying the processes responsible for observed disparities between populations of wild animals is difficult because how microorganisms colonise the gut of vertebrates is multifaceted and dependent on a suite of host, microbe, and spatial ecology features.
Differences in alpha diversity metrics existed for all populations of eastern longneck turtles. A consistent finding was that animals from urban ponds in the middle of Melbourne (DP and IW) scored lowest for all alpha diversity measures in comparison to the other study locations (Fig. 3A). These observations were particularly apparent in animals captured from DP, in which Observed ASVs, Chao1 and Shannon diversity were all much lower than the other localities. Both DP and IW are in areas of high public visitation and there may be significant public interaction with the animals at these sites. Interestingly, animals captured from the wastewater treatment plant at BD had higher alpha diversity metrics than both DP and IW. The ponds at BD are all manmade and constructed of concrete, there is no vegetation surrounding or within the ponds, and given that it is a facility used for processing human waste it would be easy to assume that this was the most disturbed habitat that we captured turtles. Human interaction with turtles is lower at BD than at DP and IW and only occurs when maintenance of the treatment ponds is necessary. Turtles from BW had the highest alpha diversity scores, and this was the least disturbed habitat with almost no human visitation. For longneck turtles it appears that presence of humans may be a more significant factor in reducing microbiota richness and diversity than habitat quality. Anthropogenic pressures have been shown to considerably alter gut microbiota of a range of wildlife including Tome’s spiny rats (Proechimys semispinosus) [53], swan geese (Anser cygnoides) [54], Kuhl’s pipistrelle bats (Pipistrellus kuhlii) [20], and olive baboons (Papio anubis) [55], and there are concerns that these disturbances in microbial ecology may have negative implications for vertebrate conservation [22, 56]. In contrast to eastern longneck turtles, no significant differences in alpha diversity due to location were found for Macquarie River turtles and it is possible that specific physiological traits, such as omnivory, in this species ensures a more stable environment for gut microbiotas.
The role that diet may be playing in driving differences in bacterial composition in our samples is unknown. Eastern longneck turtles are obligate carnivores that appear to have limited dietary preferences and will feed on almost any vertebrate or invertebrate that they can apprehend and subdue [57, 58]. In contrast, the Macquarie River turtle is omnivorous with much of their diet consisting of filamentous algae supplemented with carrion when available [34, 59]. Given the relative dietary plasticity seen in both of these species, geographic variation in food consumption likely results from local differences in prey abundance and this may be a significant factor in shaping gut microbiota composition. Other influences on dietary intake in Australian freshwater turtles are competition with fish [57], and inter-species antagonism with other chelonians [35]. In this investigation we made no attempt to assess food availability or heterospecific behaviour and therefore their roles in resource partitioning and as a determinant of microbial community composition remains unknown. One further potentially significant confounder on gut microbiotas is “Good Samaritan” feeding of turtles in public places. While this practice was not directly observed during trapping events, feeding of wildlife is common in Melbourne [60], and access to anthropogenic food sources has been shown to negatively impact the gut microbiotas of multiple vertebrate species [55, 61, 62].
An interesting finding in this investigation was the significant differences in both alpha and beta diversity between longneck turtles originating from ponds in the Darebin Parklands. These ponds are separated by approximately 200m in a straight line which should not be a significant obstacle for this species to overcome. Eastern longneck turtles are capable of long terrestrial migrations to escape unfavourable conditions [63, 64], made possible by their resistance to desiccation [65]. The results of this study may indicate that there is little seasonal fluctuation in resources available to turtles within these ponds negating the need for individuals to emigrate between waterbodies. In the event that turtles do translocate to new ponds, if and how long gut microbiotas take to assimilate to their new environments remains unknown.
For both species sampled, the bacterial family Comamonadaceae was routinely identified in almost all samples. These bacteria form a major group of the Beta-Proteobacteria and are characterised as being poly(3-hydroxybutyrate-co-3-hydroxyvalerate) (PHBV)-degrading denitrifying bacteria frequently isolated from sewerage sludge [66]. This result is not surprising given that many of our turtles originated from either active, or decommissioned sewerage and wastewater treatment plants. What is unclear is if these bacteria form a normal component of the gut microflora in these turtles, or if they are simply environmental species that have been inadvertently sequenced. However, we did identify this family in animals originating from MW, which is a natural waterbody with no history of use as a wastewater treatment facility, and they have also been identified as normal flora in a range of vertebrates including fish, birds and humans [67–69]. Further exploration of turtles from more pristine waterways is warranted to determine the prevalence of this bacterial family in chelonians, and its significance to turtle physiology.