The response of Sphingopyxis granuli strain TFA to the hostile anoxic condition

Sphingomonads comprises a group of interesting aerobic bacteria because of their ubiquity and metabolic capability of degrading many recalcitrant contaminants. The tetralin-degrader Sphingopyxis granuli strain TFA has been recently reported as able to anaerobically grow using nitrate as the alternative electron acceptor and so far is the only bacterium with this ability within the sphingomonads group. To understand how strain TFA thrives under anoxic conditions, a differential transcriptomic analysis while growing under aerobic or anoxic conditions was performed. This analysis has been validated and complemented with transcription kinetics of representative genes of different functional categories. Results show an extensive change of the expression pattern of this strain in the different conditions. Consistently, the most induced operon in anoxia codes for proteases, presumably required for extensive changes in the protein profile. Besides genes that respond to lack of oxygen in other bacteria, there are a number of genes that respond to stress or to damage of macromolecules, including genes of the SOS DNA-damage response, which suggest that anoxic conditions represent a hostile environment for this bacterium. Interestingly, growth under anoxic conditions also resulted in repression of all flagellar and type IV pilin genes, which suggested that this strain shaves its appendages off while growing in anaerobiosis.


Sample preparation for high-throughput RNA sequencing:
For all growth conditions, TFA was previously grown aerobically at 30°C in mineral medium containing 40 mM BHB as the only carbon and energy source to the exponential phase (optical density at 600 nm, 0.8). Then, cells were diluted into a fresh medium to an initial optical density of about 0.1. All cultures were grown at 30°C until they reached a final optical density of 0.7, when samples for RNA extraction were taken. Aerobic conditions are represented as "Aerobiosis" in Supplementary Table S1. In the case of the anaerobic cultures ("Anaerobiosis" in Supplementary Table S1), the appropriate aliquots were transferred into standing stoppered bottles filled to the top and with sodium nitrate 20 mM as a final electron acceptor 1 . Anaerobic growth of TFA was routinely followed together with nitrate consumption and nitrite accumulation 1,2 . For the cultures in the presence of DETA-NO ("DETA-NO" in Supplementary Table S1), 4 mM of the NO releaser was added to the aerobic cultures and incubated just for one additional hour at 30 ºC to prevent differences in cell density with respect the standard aerobic culture due to NO toxicity. For tetralin-grown cells ("Tetralin" in Supplementary Table S1), tetralin was supplied as a sole carbon and energy source in the gas phase as previously described 3 .
The concentration of antibiotic used for TFA was streptomycin 50 μg/ml Total RNA extraction was carried out as previously described 4 from exponentially-growing TFA cells in each of the different growth conditions. DNase I treatment was performed with a DNA-free kit (Ambion). The samples were purified using RNAeasy columns (Quiagen) and RNA quality was confirmed by non-denaturing agarose gel electrophoresis.
The absence of contaminating DNA was then confirmed by PCR amplification. Equal amounts of RNA obtained from three independent biological replicates of each growth condition were mixed and used in the gene expression analysis. RNA samples were sent to the ASCIDEA Headquarters, Barcelona (Spain) (http://www.ascidea.com/lifesciencesservices.html) for library preparation and RNA sequencing. The cDNA libraries were sequenced using an Illumina HiSeq2000 machine.

Bioinformatic analysis
The bioinformatic analyses were carried out by ASCIDEA. Briefly, quality of the reads obtained by HiSeq2000 sequencing was checked with FastQC software (http:/www.bioinformatics.bbsrc.ac.uk/projects/fastqc/). Preprocessing of the reads was performed with fastx-toolkit (http://hannonlab.cshl.edu/fastx_toolkit/index.html) and aScidea specific perl scripts property of aScidea (http://www.ascidea.com) in order to filter regions of low quality. Adaptors were trimmed and low quality bases at the ends of sequences and reads with undetermined bases or with 80% of their bases with less than 20% quality score. Raw reads that passed the quality filter threshold were mapped using Bowtie2 2.0.6 5 . The reference genome used and genomic annotations were obtained from NCBI public databases (https://www.ncbi.nlm.nih.gov/nuccore/CP012199.1) 1 . The inner distance between mate pairs used was 50 bp and the rest of the parameters were used with the default values. Gene level counts were calculated and FPKM normalized using Cufflinks 2.0.2 software 6 . Differential transcript expression was then computed using Cuffdiff. Main statistical analyses were performed using the free statistical language R and the libraries developed for data analysis by the Bioconductor Project (www.bioconductor.org) 7 . FPKM values were used to normalise and quantify the gene expression level. Genes differentially expressed more than 3-fold between the different conditions studied were selected for further analysis. These genes were clustered according to the Cluster of Orthologous Groups of proteins (COGs) broad classification 8 .  Figure S1. Anaerobic incubation with nitrate of WT strain and ΔnarG mutant MPO253.

Supplementary
WT strain is represented by black symbols and ΔnarG mutant by white symbols. OD 600 (squares) was measured for both strains over time and nitrate (circles) and nitrite (triangles) concentrations were measured for the WT strain. The time in which the samples for the RNA-seq were taken is indicated by the arrow. Graphic represents the mean ± SD of 3-4 biological replicates.