Uncovering the microbiome of invasive sympatric European brown hares and European rabbits in Australia

Background European brown hares (Lepus europaeus) and European rabbits (Oryctolagus cuniculus) are invasive pest species in Australia, with rabbits having a substantially larger environmental impact than hares. As their spatial distribution in Australia partially overlaps, we conducted a comparative microbiome study to determine how the composition of gastrointestinal microbiota varies between these species, since this may indicate species differences in diet, physiology, and other internal and external factors. Methods We analysed the faecal microbiome of nine wild hares and twelve wild rabbits from a sympatric periurban reserve in Canberra, Australia, using a 16S rRNA amplicon-based sequencing approach. Additionally, we compared the concordance between results from Illumina and Nanopore sequencing platforms. Results We identified significantly more variation in faecal microbiome composition between individual rabbits compared to hares, despite both species occupying a similar habitat. The faecal microbiome in both species was dominated by the phyla Firmicutes and Bacteroidetes, typical of many vertebrates. Many phyla, including Actinobacteria, Proteobacteria and Patescibacteria, were shared between rabbits and hares. In contrast, bacteria from phylum Verrucomicrobia were present only in rabbits, while phyla Lentisphaerae and Synergistetes were represented only in hares. We did not identify phylum Spirochaetes in Australian hares; this phylum was previously shown to be present at high relative abundance in European hare faecal samples. These differences in the composition of faecal microbiota may be indicative of less discriminate foraging behaviour in rabbits, which in turn may enable them to adapt quicker to new environments, and may reflect the severe environmental impacts that this species has in Australia.


4 7
Cycling conditions were: 98 °C for 30 sec, followed by 28 cycles of 98 °C for 10 sec, 55 °C confirmed by agarose gel electrophoresis before being purified with AMPure XP beads 1 5 0 (Beckman Coulter), validated using the Tapestation D1000 high sensitivity assay (Agilent 1 5 1 Technologies), and quantified using the Qubit High Sensitivity dsDNA assay kit (Thermo Nanopore Technologies, Oxford, UK) as per the manufacturer's protocol, except 500 ng of 1 5 5 input DNA per sample was used for end preparation (44). For the addition of barcodes, 80 ng 1 5 6 of end-prepped DNA was used as input, and barcoded samples were pooled in equimolar 1 5 7 concentrations to obtain at least 400 ng of pooled DNA. We used a final library amount of  Nanopore raw reads in fast5 format were demultiplexed using deepbinner (45).

6 2
Basecalling, adapter trimming, and conversion into fastq format was performed in Guppy 1 6 3 2.3.7 (Oxford Nanopore Technologies). BLASTn was used to locally align 16S sequences 1 6 4 with a quality score higher than seven against the SILVA_132 reference database as and imported into QIIME2 for subsequent analyses. Nanopore data was rarefied at a We assessed whether the bacterial diversity of hares and rabbits was statistically 1 7 3 different by using permutation-based statistical testing (PERMANOVA) via the QIIME 1 7 4 diversity beta-group-significance pipeline within QIIME2 for both Illumina and Nanopore Kruskal Wallis test within QIIME2 (46). We also evaluated statistical differences between 1 7 7 hare and rabbit faecal samples for each observed bacterial phyla using a combination of 1 7 8 multiple tests. We performed analysis of variance (ANOVA) to estimate the variance of both populations and Student's t-test to identify statistical differences between the means of both  Interrogating sympatric hare and rabbit faecal microbiota using 16S rRNA sequencing 1 8 5 We performed Illumina short-read sequencing of the 16S rRNA V3-V4 region and 14,559,822 reads greater than Q30 (average 501,908 reads per sample, excluding ROC and 1 9 0 NTC). Reads were converted to ASVs, which were assigned using the SILVA_132 reference 1 9 1 database via BLAST+ in QIIME2. This produced 6,662 unique features at a frequency 1 9 2 greater than two. Long-read sequencing of 21 samples (no ROCs or NTC due to lack of 1 9 3 amplification) using the Nanopore MinION platform produced an aggregate of 6,544,770 1 9 4 reads (average of 386,924 reads per sample) above a quality threshold of seven across two 1 9 5 MinION runs. After aligning the reads against the SILVA_132 reference database using 1 9 6 BLAST, 47,100 unique features with a frequency greater two were obtained. Wild rabbits have greater faecal microbial diversity compared to sympatric hares 1 9 8 We conducted alpha and beta diversity analyses to assess the species richness and Wallis test) and beta diversity (p = 0.01 as measured by Bray-Curtis dissimilarity matrix and 2 0 2 PERMANOVA) than hare samples. We observed considerable variance in faecal microbial had similar bacterial diversity with relatively low variance between individuals. We observed 2 0 5 these effects in both the Illumina and Nanopore 16S sequencing data sets (Fig. 1). We did not 2 0 6 observe clear correlation between bacterial diversity and sex, reproductive status, or season of 2 0 7 sample collection, although our statistical power was low due to the relatively small sample 2 0 8 size.

0 9
Taxonomic classification of both the Illumina and Nanopore sequencing data sets 2 1 0 identified Firmicutes and, to a lesser extent Bacteroidetes, to be the two dominant phyla in 2 1 1 both hare and rabbit faecal samples (Fig. 2). On average, Firmicutes and Bacteroidetes 2 1 2 together comprised more than 85% of the faecal microbiome in hares and rabbits and across Bacteroidetes compared to rabbit faecal samples (p = 0.015 as measured by Student's t-test).

3 3
The NTC microbiome comprised the phyla Proteobacteria and Bacteroidetes at 2 3 4 relative frequencies of 99% and 1%, respectively (Fig. S2). When analysing the ROCs, there between faecal samples and ROCs (Fig. S2). Additional bacterial phyla were also detected 2 3 7 only in ROCs, likely comprising the "reagent microbiome", including Plantomycetes,  We then compared the faecal microbiome of Australian wild hares to that of wild hares and rabbits (Fig. 4). Furthermore, phylum Patescibacteria was below our limit of both European and Australian hare samples at similar relative abundances (Fig. 4). We observed similar faecal bacterial diversity at the phylum, family, and genus levels 2 5 1 in both the Illumina and Nanopore datasets, although the relative abundances varied slightly 2 5 2 (Fig. 2, Fig. 3, Fig. S1). Most notably, phyla Actinobacteria and Synergistetes were present at and to a lesser extent Cyanobacteria, were present at higher abundance in the Nanopore Nanopore data (Fig. 3, Fig. S1). Even at the genus level there was generally good agreement 2 5 9 between datasets (Fig. 3). However, genera Butyricimonas, Mailhella, Ruminococcus 2, 2 6 0 Erysipelatoclostridium, and Sutterella were only detected in Illumina data and Parasutterella, 2 6 1 Schwartzia, an unassigned Victivallaceae, and the Ruminococcaceae NK4A214 group were 2 6 2 only identified in the Nanopore data. To date, studies investigating the gastrointestinal microbial diversity of lagomorphs 2 6 6 have been limited to domestic production rabbits in either Europe or China (14-17, 19, 23).

6 7
Currently only one study has focussed on wild lagomorphs, namely European brown hares in 2 6 8 their native European home-range (24). We were interested in understanding why rabbits, an 2 6 9 invasive pest species in Australia, were able to rapidly colonise over two thirds of the European brown hares, also a non-native species, remained relatively stable, despite both 2 7 2 species occupying a similar ecological niche. Differences in colonising potential of these two 2 7 3 species are likely multifactorial, for example, involving differences in behaviour, physiology, species is one method that can be used to investigate host-specific dietary preferences, which 2 7 6 may in turn reveal clues to the differences in environmental impacts that each species has. A 2 7 7 secondary aim was to use this study as an opportunity to compare short-read (Illumina) and 2 7 8 long-read (Nanopore) sequencing technologies for microbial diversity investigations.

7 9
We estimated the gastrointestinal microbiome from faecal samples collected from diversity in humans has been linked to a range of pathologies (47). Hares are known to be 2 8 7 more selective foragers than rabbits, which are considered more generalist (7). The observed indeed permit rabbits to consume feeds that hares cannot digest. It is interesting to speculate 2 9 1 that the high faecal diversity observed in rabbits may allow them to rapidly adapt to new 2 9 2 environments, contributing to this species being such a successful invader. The faecal microbiomes of both host species were dominated by the phyla Firmicutes 2 9 4 and Bacteroidetes, as is typical for other vertebrate species (48). However, the ratio of 2 9 5 Firmicutes to Bacteroidetes varied, being markedly higher in rabbits compared to hares. In humans, an increased Firmicutes to Bacteroidetes ratio has notoriously been associated with 2 9 7 age, diet, and obesity (49). A higher abundance of Bacteroidetes (genus Bacteroides) has 2 9 8 been linked to consumption of diets rich in protein and fat in humans (50), and indeed, hares are known to prefer diets rich in crude fat and protein (51). Despite having a higher daily 3 0 0 digestible nitrogen intake, hares tend to have less efficient protein digestion compared to 3 0 1 rabbits, potentially due to the absence of key microbes in their gastrointestinal tract (52).

0 2
Another apparent difference in the faecal microbiomes of these species was the presence of Akkermansia, which are noted to be positively influenced by dietary polyphenols (53, 54). of SCFAs than hares, with a higher ratio of butyrate to propionate (55-57). Again, it is  with "Western diets" rich in carbohydrates (59). The observed differences in faecal 3 1 8 microbiota could also be related to other known differences in digestive physiology between 3 1 9 rabbits and hares. For example, hares have a higher gastrointestinal passage rate compared to 3 2 0 rabbits, while rabbits retain digesta longer in order to maximise the efficiency of nutrient 3 2 1 extraction (52). Rabbits also have a greater ability to digest hemicelluloses and have a higher 3 2 2 rate of methanogenesis compared to hares (55-57, 60).

2 3
We also analysed raw data previously sequenced from faecal pellets of wild hares in with a non-pathogenic genus. The absence of this phylum of bacteria in Australian hares may 3 3 0 reflect geographical differences in diet between the populations studied, loss of this bacterial 3 3 1 species after introduction into Australia, or absence in the original founding hare population.

2
Despite these spirochaetes likely being non-pathogenic, it is worth nothing that the absence of were collected from healthy hares and rabbits, no candidate pathogens were investigated here. An additional aim of this study was to compare the bacterial diversity between two 3 3 7 emerging sequencing platforms. The Illumina platform is a popular approach for 16S rRNA sequencing, particularly because of its lower error rate (61). However, the short-read length 3 3 9 can make species level identification very challenging, especially between closely related 3 4 0 species (61). In contrast, the Oxford Nanopore platform has the ability to sequence very long 3 4 1 reads, however, it is prone to a relatively high error rate, again making accurate taxonomic 3 4 2 assignment challenging (61). In this study, we observed very similar bacterial diversity at the most likely due to PCR-based errors or bias, since different primer sets were used for each 3 4 7 sequencing platform, or bias during sequencing (61).

4 8
In conclusion, we observed notable differences in the microbiome of hares and rabbits Spirochaetes in Australian compared to European wild hares demonstrates considerable 3 5 7 geographical differences between populations, although whether these spirochetes are bacterial species in lagomorph microbiomes with particular plant species may provide further 3 6 0 insights into the impacts of wild rabbits and hares in Australia.