Insights into the distribution and abundance of the ubiquitous candidatus Saccharibacteria phylum following tag pyrosequencing

Abstract

The phylum candidatus Saccharibacteria formerly known as Candidate Division TM7 is a highly ubiquitous phylum with 16S rRNA gene sequences reported in soils, sediments, wastewater and animals, as well as a host of clinical environments. Here, the application of two taxon-specific primers on environmental and human-associated samples using bar-coded tag pyrosequencing revealed two new clades for this phylum to exist and we propose that the division consists of 2 monophyletic and 2 polyphyletic clades. Investigation into TM7 ecology revealed that a high proportion (58%) of phylotypes were sample specific, few were widely distributed and of those most widely distributed all belonged to subdivision 3. Additionally, 50% of the most relatively abundant phylotypes observed were also subdivision 3 members. Community analysis showed that despite the presence of a high proportion of unique phylotypes, specific groups of samples still harbor similar TM7 communities with samples clustering together. The lack of relatively abundant phylotypes from subdivisions 1, 2 and 4 and the presence of very few cosmopolitan members' highlights not only the site specific nature of this phylum but provides insight into why the majority of studies into TM7 have been biased towards subdivision 3.

Figure 1. Evolutionary dendogram of the Candidatus Saccharibacteria phylum.

Phylogeny constructed with 72 new sequences (OTUs defined at 0.08 dissimilarity), as well as 57 sequences from GreenGenes and the HMP representing previously defined class level clades 1, 3, MJK10, SC3 and the outgroup OP11. The reference sequences from GreenGenes were prefixed with the subdivision affiliation assigned by the GreenGenes taxonomy; Subdivision 1 (dark green), Subdivision 3 (light green), Class MJK10 (yellow), SC3 (dark blue) or OP11 (pink). The branches of the trees are colored according to the 4 proposed clades; 1 (dark green), 2 (blue), 3 (light green), 4 (red). Bootstrap confidence was calculated from 1000 replicates and displayed by circles at each node. The color of the circles, ranging from white to black corresponds to the bootstrap confidence values (between 50 and 100%).

Figure 2

Bar charts comparing the relative abundance of TM7 phylotypes per individual sample from the 4 proposed clades (A) OTUs recovered per sample. (B) Total reads obtained per sample from each subdivision. Environmental samples contained TM7 sequences from all 4 proposed clades with soil harboring the greatest number of phylotypes per sample analysed. OTUs from clades 2 and 4 were distributed widely, yet no reliable 16S rRNA gene sequences were present in the databases for these phylotypes. In most cases, a greater distribution and abundance of subdivision 3 OTUs was found in environmental samples.

Figure 3. Rarefaction curves for species-level TM7 OTUs (0.02 similarity) recovered using new taxon-specific primers and 454 bar coded tag pyrosequencing.

Each group of samples was sequenced to various depths with human oral samples sequenced to asymptote. The higher species richness of environmental samples, particularly soil and seal faecal samples is highlighted with the Australian soil requiring much more sequencing then the other environments.

Figure 4. nMDS plots showing the community similarity between sample groups.

Despite the presence of very few abundant phylotypes in more than one sample, TM7 communities from various hosts or environments were more closely related to each other than to the other environments analysed with clustering observed. At a level of 20% similarity most groups did not completely cluster, highlighting the variable nature of this enigmatic phylum.

Figure 5. Heatmap visualisation of the most dominant TM7 OTUs identified along with their proposed clade affiliation (A) the most highly abundant phylotypes and (B) the most widely distributed phylotypes observed.

Each gradient key displays the highest number of total reads obtained for each group of phylotypes with many OTUs present in relatively high abundance for one sample only (red/brown). The most widely distributed phylotypes all clustered with subdivision 3 sequences, all in relatively low abundance (<0.02% total reads) and were found in soils, sponge, seal gut and human oral samples. The most abundant phylotypes, representing 2–9% of total reads were present within clades 1, 2 or 3. Those affiliated with clade 1 were detected in Antarctic and Australian soils, with OUT 336 also found in seal gut and sponge samples. Clade 2 OTUs were primarily affiliated with skin or oral samples and soils. Clade 3 OTUs were affiliated with a range of samples, with OTU's 1 and 18 almost exclusively detected in oral human samples, OTU 309 was found in soils and seal gut samples, while OTU 343 was only detected in seal gut samples. For details on each sample and each OTUs closest match in the NCBI database see Tables S1 and 3.