Estimation of a generalized wastewater treatment plant microbial community
In order to perform a deep taxonomic survey of microbial communities associated to wastewater treatment, we initially surveyed the EBI MGnify database11, collecting abundance profiles obtained by 16S amplicon based surveys of wastewater treatment communities.
We were able to roughly identify 1465 prokaryotic genera in 3433 samples from 49 studies (see Supplementary Data S1.1), with members of the archaea kingdom in about 22.5% of samples. When we restricted the analysis on sludges the number of studies was reduced to 33 with 1363 samples, however we identified 1379 genera and the number of samples showing archaea was about 40% (see supplementary data S1.2). Such observation underlined the relevance of archaea in wastewater environments. Interestingly, in the wastewater biome we found 128 samples (about 3.7%) from 24 studies (about 50%) showing evidence of anammox species, a percent that grew to about 5% in sludge samples from 10 studies (30%). This result manifests the need of properly taking into account anammox communities when estimating microbial abundance profiles in such environments.
Evaluation of existing primers
We then sought to verify whether existing primer pairs with established high performances and good coverage over the widest range of microbial species were able to appropriately cover wastewater associated communities, especially for the anammox components, using the most updated 16S RDP collection.
All Takahashi et al.10 and Albertsen et al.11 primers pairs were tested in silico using RDP ProbeMatch against updated 16S rRNA sequences from all genera available in the current RDP database. As shown in Fig. 1, we found that all Albertsen et al. primer pairs targeting the V1-V3 and V3-V4 and V4 only 16S region showed good performances for bacteria, but had relatively poor performances for archaeal species, that we have shown above to be relevant for wastewater associated communities12. On the contrary, Takahashi Pro pair (Pro341F and Pro805R) effectively showed high coverage for both bacteria and archaea, despite a surprisingly low performance for microbes highly relevant for the denitrification cycle, namely anammox bacteria especially of the Brocadiaceae family, Candidatus brocadia genus. Accordingly, our further efforts were focused on improving the Takahashi et al. primer pair.
Comparison of the overall theoretical performance in coverage (percent of members of the given rank mapped) of the different primer pairs used in this study.
Predicted improvement of coverage on RDP database
When specifically matched against Brocadiales sequences, we found the possibility of improving the coverage of the Takahashi PRO primer pair by introducing a purine degeneration in the forward primer Pro341F, so that most member of our community of interest was matched. To design this, we extracted from the RDP global dataset all high quality (> 1200 bp) 16S classified as Brocadiaceae at the family rank. On this dataset we simulated amplicons formation with RDP probeMatch, systematically imposing degenerations that could accommodate members of this family in the most complete as well as parsimonious way. We ended up with a modified primer, Pro341FB, that was paired with the original reverse primer Pro805R and tested in silico using a mismatch 0 approach and considering the taxonomic coverage as a selection metric. As shown in Fig. 1, the primer pair Pro341FB + Pro805R (TAKB_v3v4) proved a very modest 0.007% coverage increase for archaea with respect to primer Pro341F + Pro805R (TAK_v3v4), while we found a noticeable 1% coverage increase for bacteria. Primer Pro341FB was theoretically able to amplify a total of 59% of the approximately 3.2 million sequences present in the bacteria data bank. In particular, primer Pro341FB was found to target phyla that were completely ignored by the primer Pro341F. As shown in Fig. 2, phyla which received an increase in coverage of more than 25% were found to be Chlamydiae (41%), Lentisphaerae (76%), Omnitrophica (63%), Parcubacteria (44%), candidate division WPS-1 (46%) and, importantly for this study, Planctomycetes (46%). Descending the taxonomic tree from phylum Planctomycetes to genera involved in anaerobic ammonium oxidation (anammox) we systematically observed an increase in coverage (class Planctomycetia 45%, order Candidatus Brocadiales 28%, family Candidatus Brocadiaceae 28%, genus Candidatus Brocadia 75%). As shown in Fig. 3, all anammox bacteria (genera Candidatus Brocadia, Candidatus Kuenenia, Candidatus Anammoxoglobus, Candidatus Jettenia and Candidatus Scalindua), that were almost neglected by the original Pro341F primer (red bars, secondary y-axis), resulted, as expected, markedly more covered when the Pro341FB primer was used. Major numerical details on the results of this comparison are available in supplementary materials (Supplementary data S2).
Improvement of taxonomic coverage by the newly optimized primer Pro341FB. The coverage percent value refers to the proportion between the total RDP database sequences annotated with the specific taxonomic rank and those that proved to generate an amplicon using the currently optimized Pro341FB primer and the original Pro341F (white bars), paired with the common reverse primer Pro805R. Only taxonomies with a difference in coverage higher than 25% are shown. The suffixes P, C, O, F and G refers to the ranks phylum, class, order, family and genus, respectively. Black columns mark taxonomic ranks associated with anammox bacteria.
Comparison of the theoretical coverage performance between the new Pro341FB (black, left axis) and the original Pro341F (red, right axis). The Brocadiaceae family consists of 5 genera, 4 of which are represented in the figure. A further genus named Candidatus jettenia is not present since no high-quality sequence (i.e. > 1200 bp) was present in RDP database.
Testing primer variations by NGS on selected communities
In order to verify the increase in performances for anammox communities by our modified forward primer Pro341FB, we collected the microbial community samples from 5 different origins, namely activated sludges from a domestic WWTP plant (SCS), activated sludges from a tannery WWTP (CDS), aerobic granular sludge (AGS), and partial nitrification anammox granular sludge (PNA) from pilot scale reactors fed with domestic wastewater. For the two former plants, samples from their anaerobic digestion reactors were also collected (SCD and CDD, respectively). The samples were collected from bioreactors operated in widely different conditions (suspended vs biofilm and aerobic/anoxic vs anaerobic) and fed with various substrates, in order to allow the validation of the protocol in most of the selective conditions typical for microbial communities in wastewater treatment. The total DNA of all communities was extracted and amplicons were generated using the primer pairs Pro341F + Pro805R or Pro341FB + Pro805R. As shown in Fig. 4, NGS revealed that the percentage of identified phyla was almost the same in all samples but in the PNA, where anammox communities were largely underestimated by Pro341F with respect to Pro341FB. As a confirmation it has been recently reported that anammox species largely dominate the granule population1, underlining the underestimation by the original Pro341F primer.
NGS verification of the improvement of coverage percent for members of the Brocadiaceae family coverage by the optimized Pro341FB primer. The tested samples (CDS, CDD, SCS, SCD, AGS, PNA, see text for a description) were amplified with Pro341F (suffix 1) or Pro341FB (suffix 2). Samples marked with the suffix 2 are systematically higher in Brocadia associated ranks.
Source: Ecology - nature.com