Analysis of composition and structure of coastal to mesopelagic bacterioplankton communities in the northern gulf of Mexico

Abstract

16S rRNA gene amplicons were pyrosequenced to assess bacterioplankton community composition, diversity, and phylogenetic community structure for 17 stations in the northern Gulf of Mexico (nGoM) sampled in March 2010. Statistical analyses showed that samples from depths ≤100 m differed distinctly from deeper samples. SAR 11 α-Proteobacteria and Bacteroidetes dominated communities at depths ≤100 m, which were characterized by high α-Proteobacteria/γ-Proteobacteria ratios (α/γ > 1.7). Thaumarchaeota, Firmicutes, and δ-Proteobacteria were relatively abundant in deeper waters, and α/γ ratios were low (<1). Canonical correlation analysis indicated that δ- and γ-Proteobacteria, Thaumarchaeota, and Firmicutes correlated positively with depth; α-Proteobacteria and Bacteroidetes correlated positively with temperature and dissolved oxygen; Actinobacteria, β-Proteobacteria, and Verrucomicrobia correlated positively with a measure of suspended particles. Diversity indices did not vary with depth or other factors, which indicated that richness and evenness elements of bacterioplankton communities might develop independently of nGoM physical-chemical variables. Phylogenetic community structure as measured by the net relatedness (NRI) and nearest taxon (NTI) indices also did not vary with depth. NRI values indicated that most of the communities were comprised of OTUs more distantly related to each other in whole community comparisons than expected by chance. NTI values derived from phylogenetic distances of the closest neighbor for each OTU in a given community indicated that OTUs tended to occur in clusters to a greater extent than expected by chance. This indicates that "habitat filtering" might play an important role in nGoM bacterioplankton species assembly, and that such filtering occurs throughout the water column.

Keywords: bacterioplankton; diversity; northern Gulf of Mexico; phylogenetic community structure; thaumarchaeota.

Figure 1

(A) Relative abundance of phyla and classes for bacterioplankton samples obtained from depths ≤100 m as determined from PANGEA analysis, excluding chloroplast, and cyanobacterial sequences; minor phyla and classes represented by sequences accounting for ≤0.1% of the total for each sample are not shown. (B) As for (A), but data are for samples from >100 m. Station positions are mapped in Figure A1 in Appendix.

Figure 2

Variation with depth in the ratios of α-proteobacteria/γ-Proteobacteria sequences identified by PANGEA for each of 44 samples. The dashed line indicates a depth of 100 m. Shaded area indicates values <1.

Figure 3

(A) Depth profiles of the relative abundances of Bacteroidetes (closed symbols) and δ-Proteobacteria (open symbols). (B) Depth profiles of the relative abundances of Bacilli (closed symbols) and Clostridia (open symbols). (C) Depth profiles of the relative abundances of Thaumarchaeota. All identifications based on PANGEA analysis. The dashed line indicates a depth of 100 m.

Figure 4

(A) Depth profiles of the most abundant individual α-proteobacteria (Candidatus Pelagibacter ubique and a Rhodobacteraceae) and Bacteroidetes (Yeosuana sp.) OTUs identified by PANGEA. (B) Depth profiles of the most abundant individual γ-Proteobacteria (Alteromonas and Pseudoalteromonas) and Thaumarchaeota (Nitrosopumilis sp.) OTUs identified by PANGEA. The dashed line indicates a depth of 100 m.

Figure 5

(A) Results from a principal components analysis of unweighted UniFrac distances determined using OTUs identified by Mothur analysis (distance = 0.03) for each bacterioplankton sample. Open symbols represent samples from depths <100 m; closed symbols represent samples from depths ≥100 m. (B) As for (A), but using weighted UniFrac distances.

Figure 6

Results from a canonical correlation analysis using relative abundances of phyla and classes (e.g., as in Figure 1) as determined by PANGEA, and salinity, pH, depth, dissolved oxygen, fluorescence (a measure of chlorophyll concentration), and beam attenuation (a measure of particle content) for each sample.

Figure 7

(A) Shannon indices determined using Mothur analysis of OTUs (distance = 0.03) in normalized samples (equal read numbers); open circles represent theoretical minimum estimates, closed symbols represent observed values, and open diamonds represent theoretical maximum values (see text). (B) Depth profiles of NRI (closed symbols) and NTI (open symbols); values within the shaded region are not statistically significant.

Figure A1

Station locations for samples used in this study.

Figure A2

Thaumarchaeota sequences as a percentage of total sequences plotted versus total Archaea sequence percentages. Linear regression trend line indicates that Thaumarchaeota account for about 76% of all Archaea sequences (y = −1.14 + 0.755×, r² = 0.946) regardless of sample site or depth.

Figure A3

Depth distribution of Nitrosopumilales sequences as a percentage of Thaumarchaeota sequences. Dashed line indicates 100 m depth.

Figure A4

Principal components analysis of phylum and class compositions for all stations and depths; percentages were analyzed after an arcsine transformation. Open symbols represent samples from depths <100 m; closed symbols represent samples from depths ≥100 m.

Figure A5

Depth distributions of observed OTU richness (S_obs, d = 0.03) and the Chao1 index derived from Mothur.