Metagenomic analyses of marine new production under elevated CO2 conditions

Abstract

A mesocosm experiment was carried out in a Norwegian fjord near Bergen in May 2006, with the main objective being the study of the effects of increasing concentrations of atmospheric CO2 (and associated effects such as increased acidification) on blooms of natural marine coastal plankton. Three mesocosms were bubbled with CO2(g) to achieve a high (~700ppm) CO2 concentration (pH ~7.8) to simulate predicted future conditions as a result of rising atmospheric CO2 concentrations. Another three mesocosms were treated as controls and bubbled with ambient air to represent a near pre-industrial scenario (atmospheric CO2 concentration ~300ppm, surface seawater pH ~8.15). Blooms in the mesocosms were stimulated by the addition of nutrients at a near-Redfield ratio ([N:P] ≈ [16:1]), and scientific measurements and analyses were carried out over the course of the blooms for approximately one month. Of particular interest in this study were the autotrophic plankton. The diversity and activities of these microorganisms under the two treatments was therefore investigated. By designing and using new degenerate primers specifically targeting ‘Green-type’ (Form IA and IB), ‘Red-type’ (Form IC and ID) and Form II RuBisCO, analysis of primary producers was carried out using PCR and either gDNA or cDNA (mRNA) templates from key time points spanning the complete duration of the blooms throughout the mesocosm experiment. Over 1250 novel RuBisCO large subunit sequences have been fully annotated and deposited in the NCBI GenBank® database. These sequences revealed distinct changes in the diversity of primary producers both over the courses of the blooms and between treatments. Particularly striking was the effect of acidification on the community structure of the eukaryotic picoplankton, Prasinophytes. A clade of prasinophytes closely related to Micromonas pusilla showed a distinct preference for the high CO2 conditions; a laboratory-based experiment confirmed the high tolerance of Micromonas pusilla to lower pH. Conversely, a clade related to Bathycoccus prasinos was almost entirely excluded from the high CO2 treatments. Clades of form II RuBisCO-containing dinoflagellates were also abundant throughout the experiment in both treatments. The high similarity of some of these clades to the toxin-producing species Heterocapsa triquetra and Gonyaulax polyedra, and apparent high tolerance of some clades to high CO2 conditions, is perhaps cause for concern in a high CO2 world and demands further research. In parallel with the RubisCO work, new primers were designed that target the gene encoding the Fe protein of nitrogenase (NifH). 82 Bergen genomic nifH sequences have been annotated and submitted to GenBank®. These sequences include those from organisms related to Alpha, Beta, and Gammaproteobacteria, and Cluster II and Cluster III sequences that align most closely with anaerobic Bacteria, Gram positive, and/or sulphur-reducing Bacteria. The biggest surprise, however, was the apparent abundance and significance of a Rhodobacter sphaeroides-like microorganism throughout the duration of the experiment in both treatments. Whilst this clade was unsurprisingly absent in the RuBisCO cDNA libraries, all but two of 128 nifH cDNA clones analysed were identical to the gene from Rhodobacter sphaeroides. This shows that this clade was potentially fixing N2 throughout the entire experiment, even in the presence of combined N added to both sets of mesocosms at the start of the experiment. A group of Rhodobacter sphaeroides-like microorganisms present at Bergen may therefore have been an unexpected source of new N during the experiment and contributed to the maintenance of the mesocosm communities as nutrients became depleted. One organism dominated the autotrophic communities after the blooms in both treatments. Synechococcus spp. Form IA rbcL clones most closely related to the coastal strain Synechococcus sp. strain CC9902 were recovered throughout the experiment but were particularly numerous toward the end of the experiment and dominated the “Green-type” libraries at this time. Initially, rbcL clones from these cyanobacteria were mostly derived from the ambient CO2 mesocosms but were equally distributed between treatments by the end of the experiment. This suggests that cyanobacteria related to strain CC9902 may be less tolerant of elevated CO2 (which was greatest at the beginning rather than the end of the experiment). However, despite the mesocosms being Pi-limited at the end of the experiment, several Synechococcus species (including those related to strain CC9902 and another coastal strain, CC9311) thrived. Following on from this observation, Pi uptake and assimilation mechanisms in a Synechococcus species were investigated in the laboratory. This led to the sequencing and characterisation of a pstS gene from the marine cyanobacterium Synechococcus sp. WH 8103. Unlike conventional pstS, it was discovered that the pstS II gene in this organism is constitutively expressed and unresponsive to or only weakly regulated by Pi supply. The use of PstS/pstS as a marker for P-limitation in natural samples, therefore, should be interpreted with caution.