Picoplankton Bloom in Global South? A High Fraction of Aerobic Anoxygenic Phototrophic Bacteria in Metagenomes from a Coastal Bay (Arraial do Cabo--Brazil)

Abstract

Marine habitats harbor a great diversity of microorganism from the three domains of life, only a small fraction of which can be cultivated. Metagenomic approaches are increasingly popular for addressing microbial diversity without culture, serving as sensitive and relatively unbiased methods for identifying and cataloging the diversity of nucleic acid sequences derived from organisms in environmental samples. Aerobic anoxygenic phototrophic bacteria (AAP) play important roles in carbon and energy cycling in aquatic systems. In oceans, those bacteria are widely distributed; however, their abundance and importance are still poorly understood. The aim of this study was to estimate abundance and diversity of AAPs in metagenomes from an upwelling affected coastal bay in Arraial do Cabo, Brazil, using in silico screening for the anoxygenic photosynthesis core genes. Metagenomes from the Global Ocean Sample Expedition (GOS) were screened for comparative purposes. AAPs were highly abundant in the free-living bacterial fraction from Arraial do Cabo: 23.88% of total bacterial cells, compared with 15% in the GOS dataset. Of the ten most AAP abundant samples from GOS, eight were collected close to the Equator where solar irradiation is high year-round. We were able to assign most retrieved sequences to phylo-groups, with a particularly high abundance of Roseobacter in Arraial do Cabo samples. The high abundance of AAP in this tropical bay may be related to the upwelling phenomenon and subsequent picoplankton bloom. These results suggest a link between upwelling and light abundance and demonstrate AAP even in oligotrophic tropical and subtropical environments. Longitudinal studies in the Arraial do Cabo region are warranted to understand the dynamics of AAP at different locations and seasons, and the ecological role of these unique bacteria for biogeochemical and energy cycling in the ocean.

FIG. 1.

Percent fraction of AAPs in ten samples of the GOS dataset with the highest AAP frequencies and our two samples from Arraial do Cabo (unassembled samples). Sample P, Arraial do Cabo; GS033, Punta Cormorant, Floreana Island (Hypersaline Lagoon); GS108b, Coccos Keeling, Inside Lagoon (>0.8 μm fraction); GS003, Browns Bank, Gulf of Maine; Sample E, Arraial do Cabo (>0.8 μm fraction); GS112, Indian Ocean; GS111, Indian Ocean; GS108a,Coccos Keeling, Inside Lagoon; GS117a, St. Anne Island, Seychelles; GS034, North Seamore Island (Galapagos); GS035, Wolf Island (Galapagos); and GS008, Newport Harbor, RI.

FIG. 2.

Percent fraction of AAPs in ORFs (calculated from “reads equivalents” of each ORF) from the selected ten samples of the GOS dataset and our two samples from Arraial do Cabo. Sample P (free-living), Arraial do Cabo; GS033, Punta Cormorant, Floreana Island (Hypersaline Lagoon); GS108b, Coccos Keeling, Inside Lagoon (0.8 μm fraction); GS003, Browns Bank, Gulf of Maine; Sample E (particle-associated), Arraial do Cabo (0.8 μm fraction); GS008, Newport Harbor, RI; GS117a, St. Anne Island, Seychelles; GS035, Wolf Island (Galapagos); GS112, Indian Ocean; GS111, Indian Ocean; GS108a, Coccos Keeling, Inside Lagoon; and GS034, North Seamore Island (Galapagos).

FIG. 3.

Worldwide distribution of sample sites of the ten GOS datasets with the highest AAP % fraction and the samples from Arraial do Cabo.

FIG. 4.

Phylogenetic tree of pufM genes from all ten GOS samples, our two Arraial do Cabo samples, and all reference sequences retrieved from NCBI and KEEG (KO). Only sequences with more than 700 nucleotides were used. The tree was obtained by Bayesian analysis on Mr Bayes 3.2, using the GTR model and gamma distribution. Two executions were carried out with four parallel chains and 10 millions of executions. The highlighted clades refer to the different AAP phylo-groups defined by Yutin et al., (2007).

FIG. 5.

Relative abundance of each phylo-group retrieved from the different analyzed metagenomic samples. Number of read equivalents for each obtained ORF was counted and percentages were calculated by using the classification of the phylogenetic tree generated by ARB.