Experimental Materials and Sampling
Seawater samples were collected near an oil drilling platform located at 38°26’31.7” N, 118°24’45.2” W. Samples were divided into 13 equal portions, each of which comprised 50 L. For 12 of them, different culture conditions comprising high- and low-nutrient conditions, different aeration conditions, and conditions with or without oil were used to evaluate differences in gene transcription among treatments; the rest one was set as control (named as ORI group; the condition was kept as same as of the sampled seawater around oil drilling platform). Different ventilation conditions: Every day in the morning and evening, the 50 L seawater of barrel of each different aeration group was ventilated for half an hour with a steel cylinder filled with CO2, O2, or air. All flow rates are controlled at 5 mg L− 1. Detailed cultivation parameters are described in Table 1.
On the 1st, 5th, 15th, and 20th day of such culture, the bacteria in the seawater were filtered with a glass sand filter device to a 0.45µm and then a 0.22µm filter membrane, respectively. Each filter membrane was enriched with bacteria from 1000ml of seawater. The filter membrane, after bacterial enrichment, was divided into different 50ml centrifuge tubes according to different samples, marked, and stored in a refrigerator at -80℃. Gene transcription was evaluated by extracting RNA samples of treatment cultures on the 20th day of cultivation.
Metatranscriptomic Analyses
RNA libraries were subjected to fragment screening prior to establishing the final RNA library. A quality-controlled library was then sequenced on the Illumina Hiseq high-throughput sequencing platform, and the raw data obtained from sequencing was used for subsequent transcriptomic analysis. Raw data was quality filtered to obtain clean data for evaluating taxonomic and functional differences between samples. The assembled sequencing data is available in the IMG/JGI (Integrated Microbial Genomes & Microbiomes, DOE’s Joint Genome Institute, USA) database (GOLD Analysis Project IDs are, ORI: Ga0439915, OOL: Ga0439914, OOH: Ga0439913, ONL: Ga0439912, ONH: Ga0439911, COL: Ga0439910, COH: Ga0439909, CNL: Ga0439908, CNH: Ga0439504, AOL: Ga0439503, AOH: Ga0439502, ANL: Ga0439501, ANH: Ga0439470). The abbreviations mean specific transcriptomes from certain conditions: ORI, marine bacteria, directly from seawater, origin; OOL, marine bacteria, cultured with oxygen, oil(petroleum), and low nutrient content; OOH, oxygen, oil(petroleum), and high nutrient content; ONL, oxygen, no oil(petroleum), and low nutrient content; ONH, oxygen, no oil(petroleum), and high nutrient content; COL, carbon dioxide, oil(petroleum), and low nutrient content; COH, carbon dioxide, oil(petroleum), and high nutrient content; CNL, carbon dioxide, no oil(petroleum), and low nutrient content; CNH, carbon dioxide, no oil(petroleum), and high nutrient content; AOL, air, oil(petroleum), and low nutrient content; AOH, air, oil(petroleum), and high nutrient content; ANL, air, no oil(petroleum), and low nutrient content; ANH, air, no oil(petroleum), and high nutrient content.
BLAST (The Basic Local Alignment Search Tool) alignment (e value threshold of ≤ 1e-5 for positive hits) of the quality filtered data against the NCBI non-redundant (NR) protein database was conducted. The lowest common ancestor (LCA) classification algorithm implemented in MEGAN (Metagenome Analyser) [11] was used to assign taxonomic classifications based on alignments, since each sequence might have multiple alignment results. Specifically, the taxonomic classification based on the first branch served as the species annotation for the sequence in question. Relative abundances of the ten most abundant classes are shown in the results, with the largest relative abundances in each sample shown and the remaining taxa being grouped as “Others”.
Functional annotation of the quality filtered data was conducted using BLAST alignment of protein sequences against the assembled transcriptomic and protein annotation databases. Since multiple results could arise from each sequence alignment, the alignment results of each sequence were screened by calculating the alignment coverage ratio of each gene based on the reference and query lengths (i.e., the BLAST coverage ratio; BCR) to ensure that the BCR of the reference and query in each alignment exceeded 40%. The corresponding functional annotation information was then compiled from the databases.