Community transcriptomic assembly reveals microbes that contribute to deep-sea carbon and nitrogen cycling

Abstract

The deep ocean is an important component of global biogeochemical cycles because it contains one of the largest pools of reactive carbon and nitrogen on earth. However, the microbial communities that drive deep-sea geochemistry are vastly unexplored. Metatranscriptomics offers new windows into these communities, but it has been hampered by reliance on genome databases for interpretation. We reconstructed the transcriptomes of microbial populations from Guaymas Basin, in the deep Gulf of California, through shotgun sequencing and de novo assembly of total community RNA. Many of the resulting messenger RNA (mRNA) contiguous sequences contain multiple genes, reflecting co-transcription of operons, including those from dominant members. Also prevalent were transcripts with only limited representation (2.8 times coverage) in a corresponding metagenome, including a considerable portion (1.2 Mb total assembled mRNA sequence) with similarity (96%) to a marine heterotroph, Alteromonas macleodii. This Alteromonas and euryarchaeal marine group II populations displayed abundant transcripts from amino-acid transporters, suggesting recycling of organic carbon and nitrogen from amino acids. Also among the most abundant mRNAs were catalytic subunits of the nitrite oxidoreductase complex and electron transfer components involved in nitrite oxidation. These and other novel genes are related to novel Nitrospirae and have limited representation in accompanying metagenomic data. High throughput sequencing of 16S ribosomal RNA (rRNA) genes and rRNA read counts confirmed that Nitrospirae are minor yet widespread members of deep-sea communities. These results implicate a novel bacterial group in deep-sea nitrite oxidation, the second step of nitrification. This study highlights metatranscriptomic assembly as a valuable approach to study microbial communities.

Figure 1

Abundance of major phyla based on classification of rRNA transcript reads. All 16S rRNA reads (total of 14 571 562) were mapped (>75% over half the read length cutoff) to a 16S rRNA gene database (SILVA).

Figure 2

Abundance of gene transcripts in plume and background based on mapping transcripts to the plume de novo metatranscriptomic assembly. Red filled circles are mRNAs that have high similarity to Altermonoas sp., yellow are those related to Nitrospirae and green are MGII. Gray filled circles are highly transcribed ammonium transporters, most of these belong to AOA, consistent with previous findings (Baker et al., 2012). The dotted line indicates equal representation of transcripts in plume and background.

Figure 3

Plot of gene transcript abundance vs coverage in the metagenomic assembly from Lesniewski et al. (2012). Abundance is the number of cDNA reads mapped to the transcript, normalized to the length of the gene. Top matches in the genomic DNA library assembly are greater than e-value of 1E⁻¹⁰.

Figure 4

Transcripts not present in accompanying metagenomic data but with similarity to sequences in public databases. Each circle represents an assembled mRNA contig. Plotted is percent similarity to NCBI sequence vs the number of plume cDNA reads recruited. Coloring is consistent with Figures 2 and 3; red are Alteromonas, yellow are Nitrospirae and green are MGII.

Figure 5

Schematic model and abundance of transcripts in the plume for proteins involved in nitrite oxidation and associated electron transfer. Colored proteins were detected in the plume cDNA libraries. Complexes in gray were not identified but are included in the model of electron transport for reference. Arrows show movement of electrons and protons. For transcript abundance, multiple circles for each gene represent multiple closely related gene sequence variants. Normalization is calculated as the number of cDNA reads mapped divided by lengths of the genes and multiplied by 1000.

Figure 6

Phylogeny of Nitrospira-like 16S rRNA genes from assembled transcripts. Trees were generated using the maximum likelihood method and Planctomycetes brasiliensis as the outgroup.

Figure 7

Abundance of assembled transcripts most closely related to Nitrospirae from the plume transcript assembly. Each circle represents a distinct gene sequence, with assigned functions listed on the left. Thus, multiple data points for each gene represent sequence variants present in the community. Abundance is based on the number of reads that mapped to the assembled transcript. Normalization is calculated as the number of cDNA reads mapped divided by lengths of the genes and multiplied by 1000.