Bacteria are a major biological force in the marine carbon cycle: approximately half of the ocean’s primary production enters the marine food web through heterotrophic bacteria. The speed and amount of microbial conversion of organic matter can thus have significant effects on the ocean ecosystem. We are currently studying the degradation of macroalgae by natural microbial communities addressing three principal questions: (i) For microbial communities, how does loss of diversity affect community production? (ii) Are there combinations of microbes that are especially productive at degrading algal material? (iii) What are the relevant mechanisms that shape these combinations?
We performed a microcosm experiment where marine microbial communities with different diversity and composition were generated from a natural seawater community using a removal-of-species-by-dilution method. All communities were supplemented with macroalgal (Fucus) extract as the major carbon source. Production was measured as the community CO2 produced and maximum cell density reached by each community in 160h. Community composition before and after the addition of the supplement carbon was determined via 16s rRNA amplicon sequencing with OTUs clustered at 99% sequence similarity.
For both measurements, loss of diversity from a natural seawater community did not have a significant effect on average production until there was a ~2/3 loss in initial species richness, after which average production decreased. However, loss of biodiversity caused the variance of community production to increase when the mean community production remained stable, and to decrease when mean production decreased. This indicates that communities of lower diversity would fluctuate more in production upon disturbances such as a seaweed bloom. Communities with the highest production were not those with the highest stationary phase diversity, but those dominated by 3-4 bacterial species. Bacteria species from the order Oceanospirillales, Rhodobacterales, or Alteromonadales were identified as important predictors of community production using a random forest model. Interestingly, some of the top predictors were almost always lower than 5% in abundance in all communities, suggesting that these bacteria may affect community production by interference. We are currently in the process of identifying the relative contributions of individual species, selection, resource partitioning, and interactions to the production of each community, and study how these contributions change with the diversity of the community.