Page 11 - Read Online

Mueller et al. Microbiome Res Rep 2024;3:33 https://dx.doi.org/10.20517/mrr.2024.09 Page 5 of 18

RESULTS
Genomic analysis of multiple Akkermansia isolates supports the establishment of new species and
A. muciniphila subspecies
To perform our analysis, we retrieved 221 publicly available genome sequences originally identified as
T
A. muciniphila, including Muc , from NCBI. In addition to these, we added 13 new Akkermansia strains
that we isolated from human and animal fecal samples. The final collection included 207 genomes from
human Akkermansia isolates, 87 of which had been previously phylotyped as AmIa, AmIb, AmII, AmIII,
AmIV, and AmV [19-22,60] , and 26 genomes derived from non-human sources.

The average nucleotide identity (ANI) between two pairs of genomes has become the gold standard for
identifying the genetic boundary of species from genome sequencing data [61,62] . Our pairwise analysis ANI
between Akkermansia genomes reinforced the previously established phylogroup structure [19,20,22] and added
a new phylogroup [Figure 1A and Supplementary Table 2]. This phylogroup, which we termed AmVI, is
represented by Akkermansia isolated from a patient undergoing cancer immunotherapy and a patient with
amyotrophic lateral sclerosis. Current standards use 95% ANI as a threshold to define new species [61,62] , and a
98% ANI threshold has been used to define sub-species . We calculated the average ANI between each pair
[63]
of phylogroups and determined that the previously proposed phylogroups of A. muciniphila should be
considered as separate species based on the 95% threshold, with Akkermansia AmI being considered
A. muciniphila sensu stricto. At a 98% threshold, A. muciniphila splits into two sub-species: AmIa and AmIb
[Figure 1B]. Two genomes that belonged to the AmIII group, BAA-2869 (isolated from squirrel) and
CSUN-56 (isolated from human), each share an average ANI of less than 90% with all other phylogroups,
indicating that these represent additional Akkermansia species [Supplementary Figure 1]. We have chosen
to exclude the BAA-2869 and CSUN-56 genomes from further analyses, as they are likely single
representatives of new species.

We next refined the evolutionary relatedness among Akkermansia genomes by identifying a total of 137
single-copy, core gene clusters among Akkermansia groups. We included A. glycaniphila to anchor into a
deeper branch of the Akkermansia evolutionary tree. The aligned amino acid sequences predicted to be
encoded by these gene clusters were concatenated to generate a phylogenetic tree. After rooting the tree to
A. glycaniphila, we found a similar clustering of Akkermansia core proteins into the same phylogroups as
predicted by ANI [Figure 1C].

As further evidence of the distinctiveness of Akkermansia species, an ANOVA test indicated that there are
statistically significant differences in their average genome sizes [F(6,225) = 92.77, P < 0.0001].
A. muciniphila genomes were the smallest, with AmIa and AmIb displaying averages of 2.743 Mbp and
2.816 Mbp, respectively. Similarly, AmV had an average genome size of 2.809 Mbp. A. massiliensis, AmIII,
and AmVI were larger with average genome sizes of 3.112, 2.956, and 3.161 Mbp, respectively [Figure 1D].
A. biwaensis had the largest genomes with an average size of 3.213 Mbp.

The divergence of 16S rRNA sequence identity is commonly used to distinguish between species [64,65] .We
successfully extracted full-length representative 16S rRNA gene sequences for 217 genomes and generated a
phylogenetic tree rooted on A. glycaniphila. Genomes for which our analysis did not yield full-length 16S
rRNA gene sequences were of contig or scaffold assembly levels, suggesting incomplete assembly or low
sequencing quality for these genomes. As with genome level ANI and core protein divergence, Akkermansia
16S rRNA sequences clustered by phylogroup, except for AmIa and AmIb sub-phylogroups, which could
not be resolved. While these observations further support the phylogroup-level separation by ANI and core
phylogeny, the average 16S rRNA gene sequence identity between any pair of phylogroups does not meet