The transcriptome response of the ruminal methanogen Methanobrevibacter ruminantium strain M1 to the inhibitor lauric acid

Objective Lauric acid (C12) is a medium-chain fatty acid that inhibits growth and production of the greenhouse gas methane by rumen methanogens such as Methanobrevibacter ruminantium. To understand the inhibitory mechanism of C12, a transcriptome analysis was performed in M. ruminantium strain M1 (DSM 1093) using RNA-Seq. Results Pure cell cultures in the exponential growth phase were treated with 0.4 mg/ml C12, dissolved in dimethyl sulfoxide (DMSO), for 1 h and transcriptomic changes were compared to DMSO-only treated cells (final DMSO concentration 0.2%). Exposure to C12 resulted in differential expression of 163 of the 2280 genes in the M1 genome (maximum log2-fold change 6.6). Remarkably, C12 hardly affected the expression of genes involved in methanogenesis. Instead, most affected genes encode cell-surface associated proteins (adhesion-like proteins, membrane-associated transporters and hydrogenases), and proteins involved in detoxification or DNA-repair processes. Enrichment analysis on the genes regulated in the C12-treated group showed a significant enrichment for categories ‘cell surface’ and ‘mobile elements’ (activated by C12), and for the categories ‘regulation’ and ‘protein fate’ (represssed). These results are useful to generate and test specific hypotheses on the mechanism how C12 affects rumen methanogens. Electronic supplementary material The online version of this article (10.1186/s13104-018-3242-8) contains supplementary material, which is available to authorized users.


Introduction
Ruminal methane-producing archaea acquire attention because ruminant livestock is estimated as the most important source of anthropogenic emission of the greenhouse gas methane [1]. Among the most-promising anti-methanogenic compounds are two medium chain fatty acids (MCFA), lauric acid (C 12 ) and myristic acid (C 14 ), which were shown to inhibit methanogenesis in vivo when supplemented to the diet of ruminants [2][3][4], in vitro in rumen fluid [5] and in methanogenic cultures [6]. MCFA cause leakage of K + ions and decrease survival of Methanobrevibacter ruminantium, a optical density at 600 nm and by methane (CH 4 ) formation after 24, 48, 60 and 61 h. The culture was inoculated with 5 ml of an exponentially growing pre-culture (OD 600 ~ 0.64) to 45 ml of medium. Cell survival was detected with the LIVE/DEAD BacLight Bacterial Viability Kit for microscopy and quantitative assays (Kit L7012; Invitrogen GmbH, Darmstadt, Germany) [6]. Lauric acid (≥ 97% purity) was obtained from Sigma-Aldrich (Buchs, Switzerland), and a stock solution with 200 mg/ml was prepared by dissolving the C 12 in sterile dimethyl sulfoxide (DMSO) (Sigma-Aldrich), a commonly used solvent for water-insoluble substances [8]. After 60 h of incubation, when cells reached the exponential phase, three bottles were supplemented with 0.1 ml of the C 12 stock solution to reach a final concentration of 0.4 mg C 12 / ml (treatment group), three bottles were supplemented with 0.1 ml of DMSO (final concentration: 0.2%) (control group), and three bottles received no supplement (blank group). The concentration of C 12 and the exposure time of 1 h chosen were in a range where most cells remained alive and where CH 4 formation was clearly but not completely inhibited. It was verified that, at 61 h of incubation, CH 4 formation rates and proportion of living cells did not differ between DMSO-exposed control cultures (measured: 0.71 ± 0.03 µmol/ml × h and 97 ± 0.3%, respectively) and untreated blank cultures (0.74 ± 0.04 µmol/ml × h and 99 ± 1.2%). At 61 h, i.e. after 1 h of exposure to C 12 , CH 4 formation rates in the hour after exposure were suppressed by 40 ± 6% compared to the control cultures (P < 0.05), and cell viability was reduced down to 71 ± 1.8% when compared to the control cultures (P < 0.05). At this time point, three samples per group (each 50 ml of culture) were anaerobically collected at 4 °C after centrifugation at 5000×g for 6 min. Cell pellets were immediately frozen in liquid nitrogen and stored at − 80 °C until RNA extraction.

RNA isolation
Total RNA was isolated from the frozen cell pellets by using TRIzol ® Reagent (ThermoFisher, Waltham, MS, USA), according to the manufacturer's protocol. In order to remove genomic DNA from total RNA samples, a DNA digestion was performed with the RNase-Free DNase Set (Qiagen, Hilden, Germany) following manufacturer's instructions. Quantity and quality of extracted RNA were determined by a Qubit ® 1.0 fluorometer with a Qubit RNA BR (Broad Range) assay kit (Invitrogen, Carlsbad, CA, USA) and by an Agilent 2200 TapeStation with the Agilent RNA ScreenTape assay (Agilent Technologies, Santa Clara, CA, USA), respectively. Nine purified total RNA samples with a yield of at least 5 µg and RNA integrity numbers (RIN) in a range of 5.6-7.6 were used for sequencing. These included three replicates per group: three DMSO-dissolved C 12 -treated samples (T1,  T2 and T3), three samples with DMSO supplementation  alone (control samples C1, C2, C3) and three samples  without supplement (blank samples B1, B2, B3).

Next generation sequencing
Enriched RNA samples were used to produce library constructs by following the Illumina TruSeq ® Stranded total RNA protocol (Part#15031048 Rev. C, September 2012) with the Illumina TruSeq Stranded total RNA Sample Preparation Kit. Libraries were quantified and quality checked using qPCR with Illumina adapter specific primers (Roche LightCycler ® system, Roche Diagnostics, Basel, Switzerland) and by the Agilent Technologies 2100 Bioanalyzer with DNA-specific chips, respectively. Diluted indexed libraries (10 nM) were pooled, used for cluster generation (Illumina TruSeq SR Cluster Kit v4-cBot-HS reagents) and further sequenced (Illumina TruSeq SBS Kit v4-HS reagents) on the Illumina HiSeq 2500 instrument in the high output mode according to the manufacturer's recommendations. Illumina single read approach (1 × 125 bp) was used to generate raw sequencing reads with a depth of approximately 20-30 million reads per sample.

RNA-sequencing data analysis
Data analyses were performed as described by Tanner et al. [9]. Shortly, reads (125 bp) were mapped against the genome of M. ruminantium M1 using the CLC Genomics Workbench 6.5.1 (CLC, Aarhus, Denmark). Statistical analysis was performed using Bioconductor EdgeR software package in R. A false discovery rate (FDR) value < 0.05 was used as cutoff for significance of differentially expressed genes and log 2 fold change > 1 and < −1 was used as cutoff for differential transcription of genes higher (positive log 2 -fold change values) or lower (negative log 2 -fold change values) expressed in cultures [10]. To test for significant enrichment in each category listed in Table 1, a two-tailed Fisher test was performed at http://www.langs rud.com/fishe r.htm.

Results and discussion
The Ribo-Zero ™ rRNA Removal Kit can be used to efficiently remove the rRNA fraction from total RNA samples isolated from the archaeon M. ruminantium M1. The Epicentre probes (directed to bind rRNA from a broad spectrum of bacteria species) reduced the rRNAs in all samples tested, which resulted in 40-85% of non-rRNA sequencing reads in the samples (Fig. 1). More than 10 million mRNA sequencing reads per sample were mapped to the genome of M. ruminantium M1 (Fig. 1), which is a sufficient coverage for transcriptome analyses [11].
First, we compared the untreated cultures to the control cultures treated with DMSO. DMSO affected the expression of 97 out of 2280 genes in the M1 genome (Additional file 1). DMSO induced changes in gene expression of cell surface-related proteins, cell membrane-associated transporters and intracellular proteins; the latter maybe related to the observation that DMSO penetrates cell membranes [8]. DMSO-regulated genes included genes encoding proteins related to the cell envelope, mainly adhesion-like proteins (six genes; four Table 1 Number of genes significantly differential expressed within functional categories a Significant functional enrichment in a Fisher exact test (p < 0.05) b Significant functional underrepresentation in a Fisher exact test (p < 0.05) c Non-conserved hypothetical genes and RNAs are omitted in the classification [11]. Treatment: with DMSO-dissolved C 12 , control: with DMSO alone, blank: without C 12  . Note that the y-axis is non-linear down-regulated, two up-regulated). Others were classified as mobile genetic elements (38 genes including hypothetical genes; all down-regulated), and genes involved in energy metabolism, mainly hydrogen metabolism [nine genes, six up-regulated (frhA/B1/D/G, mtrA2, DsbD), three down-regulated (hypA/B, adh3)]. Genes involved in metabolism of vitamins and cofactors (six genes; four down-regulated, two up-regulated) as well as of amino acids (four genes, all down-regulated) were regulated. Moreover, cation transporters (five genes; four of five up-regulated), amino acid transporters (two genes; down-regulated), and other transporters (three genes, up-regulated) showed differential expression when untreated cultures were compared to DMSO-supplemented cultures. Overall, the set of genes regulated in the DMSO control group compared to the blank group was enriched for genes assigned to categories: 'Mobile elements' , 'Transporters' , and 'Regulation' , whereas genes assigned to 'protein synthesis' and genes of unknown function were significantly underrepresented ( Table 1). The comparison between the C 12 + DMSO-treated and the untreated cultures revealed 42 genes differentially regulated (Additional file 2), 26 of these also found in the DMSO-treated versus untreated comparison (Additional file 3).
Thereafter the transcriptome of the C 12 + DMSOtreated and DMSO-treated cultures were compared to identify the mechanisms how MCFA affect methanogenesis. A total of 147 genes, 6.4% of all 2280 genes, were differentially regulated ( Table 2).
The subcellular localization of the encoded protein could be identified for 75% of the regulated genes. Predominantly, genes associated with the cell envelope were affected, namely trans-membrane proteins or membraneassociated proteins. Enrichment analysis showed that, with C 12 exposure, mainly adhesion-like proteins (category 'cell surface') and phage-related proteins ('mobile elements') were significantly enriched in the regulated genes data set (Table 1). This supports earlier suggestions that MCFA primarily target the cell envelope and processes that occur at the cell membrane [12]. For example, upon exposure to C 12 in the present study, the mRNA abundance of 26 adhesion-like proteins (ALPs) (part of the cell envelope [13]), i.e. of 25% of all ALPs of M. ruminantium, and of two proteins involved in biosynthesis of teichoic acid and pseudomurein which are cell-wall related [14], were up-regulated compared to the DMSO control group (Table 2).
Two subunits of the membrane-bound energy-converting hydrogenase (Eha), which is involved in hydrogenotrophic methanogenesis [13,15], were down-regulated by log 2 1.6-and 1.7-fold in cultures exposed to C 12 , whereas two cytoplasmic hydrogenases (Frh, Mvh) were not. A gene encoding ferredoxin, a trans-membrane iron-sulfur protein involved in electron transfer from hydrogen, was up-regulated (log 2.6-fold upon C 12 exposure). Expression of 3 genes encoding trans-membrane 4Fe-4S binding domain-containing proteins was affected by C 12 exposure. Two subunits of the methyl-H4MPT:coenzyme M methyltransferase (Mtr), which is membrane-bound and plays a crucial role in the methanogenesis pathway [15,16], were down-regulated by log 2 2.1-and 1.2-fold upon C 12 exposure. In total 13 genes encoding mainly transporters of amino acids and cations displayed differences in transcript abundance after C 12 exposure ( Table 2). For example, several genes encoding subunits of cations transporters, like the nickel ABC transporter permease proteins or nickel ABC transporter ATP-binding proteins, NikA2, NikB1, NikB2, NikC2 and NikD2, were differentially regulated. These cation transporters belong to a large family of ABC transporters (peptide/nickel transporter family) in ABC-type nickel transporter system, which is composed of a periplasmic binding protein (NikA), two integral membrane proteins (NikB and NikC) and two ABC proteins (NikD and NikE) [17]. One P-type ATPase, which are membrane-bound efflux pumps involved in metal homeostasis of microorganisms [18], was down-regulated. In prokaryotes, ABC transporters and P-type ATPases have important functions in maintaining appropriate concentrations of transition metals such as Ni, Co, Fe, Cu, and Zn, which are essential components of many prokaryotic enzymes [18]. Two transmembrane cobalt transport proteins (mru_0540; mru_0539), and two membrane-associated proteins involved in molybdate transport (mru_0200, mru_0201) [19], were up-regulated.
In addition, genes encoding intracellular proteins were affected by C 12 exposure. These data support earlier observations that exposure to C 12 causes leakage of intracellular K + ions in M. ruminantium [6,7], thus damages the cell envelope. Amongst the regulated genes, mostly genes encoding proteins involved in DNA repair, and genes controlling transcription/translation and redox homeostasis were affected. For example, thioredoxins and rubrerythrins showed an altered expression; they are considered to form a system protecting Archaea against oxidative stress [20,21]. Thioredoxin-like proteins exhibit biochemical activities similar to thioredoxin and help methanogens maintain redox homeostasis [7]. Genes which were up-regulated by C 12 included genes encoding proteins that are involved in nucleic acid metabolism and repair and in translation include a helicase (mru_0981), an exonuclease (mru_2097, recJ1), an anaerobic ribonucleosid-triphosphate reductase nrdD (mru_0241), a nucleotidase (mru_2104; SurE1), and a RNA-metabolizing metallo-beta-lactamase (mru_1978). Several genes   involved in translation or post-translational modification were down-regulated, e.g. a staphylococcal nuclease domain-containing protein (mru_1383), a nascent polypeptide-associated complex protein (mru_1511), an RNA-binding protein (mru_0519) and a signal peptidase (mru_1581).

Conclusion
The transcriptional response of M. ruminantium to the fatty acid C 12 does not involve repression of specific pathway such as the methanogenesis pathway. Instead, it implies that C 12 provokes broad transcriptional changes, and targets primarily cell surface associated adhesion-like proteins, phage-related proteins, and transmembrane proteins. How this response affects methanogens remains unclear. Future studies may investigate how different dosages of and prolonged exposure to C 12 affect gene and protein expression and survival of M. ruminantium.

Limitations
One limitation of our study is the low number of replicates per group. In addition, only one dosage of C 12 was tested and samples for RNA sequencing were collected only at one time point; this precludes generalization to C 12 -treated cultures were compared to DSMO-exposed control cultures (significant change with log 2 fold changes < 1 and > 1 and a false discovery rate < 0.05). The list does not include the 71 regulated hypothetical proteins. The M. ruminantium (mru) open reading frame (ORF) codes are adopted from the Kyoto Encyclopedia of Genes and Genomes