One of the most abundant and important organosulfur molecules in the marine environment is a molecule called dimethylsulfoniopropionate (otherwise known as DMSP). DMSP and its catabolite dimethyl sulfide (DMS) are key marine nutrients and signalling molecules. They also effect atmospheric chemistry, cloud formation and, potentially, climate. These molecules are produced in huge amounts by marine phytoplankton, seaweeds, corals, some plants and, as was recently described, bacteria. Until this discovery that bacteria can make DMSP, it was widely assumed that DMSP and DMS production was mostly limited to oxic and photic environments, specifically the surface oceans, as these are the domain of most marine eukaryotes that produce it. Given that heterotrophic bacteria are not beholden to light or oxygen, it seemed to us that the process of DMSP production might occur at significant levels in marine environments where these factors are limiting and bacteria are abundant, such as in marine sediment.
Happily, many such environments are situated just down the road from us here at the University of East Anglia, UK, in the form of coastal sediment from local estuaries and saltmarshes. For example, our main study site was on the North Norfolk coast, at Stiffkey saltmarsh (see figure). Such saltmarshes are known to be highly productive for DMSP and DMS, but publications to date mainly attribute this to the DMSP-producing cordgrass Spartina that often grows at these sites. Working on these coastal sediments we found that DMSP levels and DMSP and DMS synthesis rates were far higher in the surface sediment than in the overlying seawater. We also found that these high DMSP levels were not dependent on the presence of Spartina, with similar levels being detected at Yarmouth Estuary, a site lacking this plant. DMSP-producing diatoms were present in the sediment and we isolated a strain from the most abundant genus Asterionellopsis, finding it to produce DMSP, albeit at low levels. We therefore proposed that bacteria with the dsyB gene (a reporter for bacterial DMSP production) were key contributors to the high DMSP levels in the sediment. To identify some of these DMSP-producing bacteria we developed media conditions that enriched for DMSP production and isolated many diverse proteobacteria and actinobacteria as a result. Most DMSP-producing isolates contained dsyB, but several key strains lacked this reporter gene for DMSP production. We discovered that some of these bacteria, notably Novosphingobium, make DMSP through a completely distinct pathway to dsyB, one that in fact closely resembles the pathway used by DMSP-producing plants. These bacteria contain a gene we termed mmtN, which, like dsyB, encodes a sulfur methyltransfase enzyme that is essential for DMSP synthesis, and is a robust reporter of bacterial DMSP synthesis. By using these two genes as markers for DMSP synthesis in bacteria, we estimated that ~108 bacteria per gram sediment possess the genetic machinery to produce DMSP in Stiffkey surface mud.
During the project we developed gene probes to dysB and mmtN, which we used to confirm that the abundance of DMSP-producing bacteria (containing dsyB and mmtN), and the transcript levels of dsyB were far higher in surface sediment than in surface seawater. This supports our theory that bacteria play a more significant role in DMSP production in the sediment compared to the phytoplankton-dominated water column. This was also evident in our metagenomics analysis of Stiffkey sediment, which contained much higher incidences of both dsyB and mmtN compared to those found in published surface seawater metagenomes. Incidentally, we saw very low levels of the eukaryotic DSYB gene (a reporter for DMSP synthesis in phytoplankton) in our metagenomes, and although we don’t yet know how diatoms such as Asterionellopsis produce DMSP, it is a future question we are eager to answer.
A final comment on something that we find very exciting: DMSP concentrations and the abundance of DMSP-producing bacteria and their dsyB transcripts were far higher in 4 km deep surface sediment from the Mariana Trench (totally dark, cold and under high pressure) than in surface seawater from the same site, leading us to propose that marine sediment in general, and not just coastal sediment, is highly productive for DMSP and DMS. The caveat to this, of course, is that most marine sediment will have little direct influence on atmospheric DMS levels.
As with most studies, this work has posed many interesting questions that require future focus. For example:
- Anoxic sediments had far lower stocks of DMSP compared to surface samples, but these were still higher than those in the surface pond water, and we have little idea of the organisms that are producing DMSP in these environments.
- There are still many bacterial and diatom isolates which produce DMSP and lack known DMSP synthesis genes in their genomes and/or transcriptomes.
This work is the culmination of Beth Williams’s PhD at UEA and was certainly a team effort, with contributions from international scientists working in distinct disciplines including molecular microbiology, evolutionary ecology, protein biochemistry, bioinformatics and analytical science. We feel that together we have created a well-rounded piece of work that ties together genetics, protein and culture analysis with large scale metagenomics and rate experiments, giving us a glimpse into the role of bacteria in global DMSP production. We might have only scratched the surface in our study of bacterial DMSP synthesis, but we intend to keep on digging.
This paper is published in Nature Microbiology here:
The following authors of the paper contributed to writing this blog: Dr Beth Williams and Dr Jonathan Todd