Deep sequencing of a DMSP-degrading gene (dmdA) in a coastal bacterioplankton community

Sampling

Ten primer pairs targeting environmental clades of the dimethylsulfoniopropionate (DMSP) demethylase protein, DmdA, were designed using an iterative bioinformatic approach that took advantage of >1700 dmdA sequences captured in marine metagenomic datasets. Using the bioinformatically-optimized primer pairs, dmdA genes were amplified from free-living coastal bacterioplankton samples and sequenced using 454 technology. An average of 6,400 amplicons per primer pair represented almost 800 clusters of environmental dmdA sequences, with clusters defined conservatively at >90% nucleotide sequence identity (~95% protein identity). Systematic comparisons of primer performance showed that degenerate and inosine-based primers did not perform better than specific primer sets in retrieving dmdA diversity, and sometimes captured a lower diversity of sequences from the same DNA sample. The specific primer sets were used to compare dmdA diversity in free-living versus particle-associated bacteria in southeastern U.S. coastal waters. Hundreds of different dmdA clusters were found in both size fractions, with Roseobacter-like and SAR11-like sequences dominating both. The free-living fraction had a higher diversity of dmdA clusters overall, though clusters retrieved by a given primer set were largely shared (52-88%) across the two size fractions, and most sequences were affiliated with these shared clusters (~90%). Despite evidence from 16S rRNA-based taxonomic surveys that free-living bacterioplankton are a considerably less diverse subset of particle-associated bacteria at this site, this seems not to be the case for a widespread bacterial gene mediating sulfur transformations.

Method steps

1. Pipeline used: https://www.ebi.ac.uk/metagenomics/pipelines/4.1