Adaptation Strategies to High Hydrostatic Pressures in Pseudothermotoga species Revealed by Transcriptional Analyses

Pseudothermotoga elfii strain DSM9442 and P. elfii subsp. lettingae strain DSM14385 are hyperthermophilic bacteria. P. elfii DSM9442 is a piezophile and was isolated from a depth of over 1600 m in an oil-producing well in Africa. P. elfii subsp. lettingae is piezotolerant and was isolated from a thermophilic bioreactor fed with methanol as the sole carbon and energy source. In this study, we analyzed both strains at the genomic and transcriptomic levels, paying particular attention to changes in response to pressure increases. Transcriptomic analyses revealed common traits of adaptation to increasing hydrostatic pressure in both strains, namely, variations in transport membrane or carbohydrate metabolism, as well as species-specific adaptations such as variations in amino acid metabolism and transport for the deep P. elfii DSM9442 strain. Notably, this work highlights the central role played by the amino acid aspartate as a key intermediate of the pressure adaptation mechanisms in the deep strain P. elfii DSM9442. Our comparative genomic and transcriptomic analysis revealed a gene cluster involved in lipid metabolism that is specific to the deep strain and that was differentially expressed at high hydrostatic pressures and might, thus, be a good candidate for a piezophilic gene marker in Pseudothermotogales.


Introduction
Several studies have highlighted the fact that hydrostatic pressure affects the activity of numerous key processes and cellular components of deep-living microorganisms [1][2][3]. Depending on the genera investigated, piezophilic adaptation requires modifications of the genome, fine tuning of the gene expression, modifications of the structure of key proteins, or stress-like physiological cell responses [4]. An important finding is that osmolytetype co-solutes, which are termed "piezolytes", accumulate in the cells and counteract the pressure effects [5]. Increasing hydrostatic pressure has been observed to lead to intracellular accumulations of β-hydroxybutyrate, ectoine, and mannosyl-glycerate in Photobacterium profundum, Alcanivorax borkumensis, and the hyperthermophilic piezophilic archaeon Thermococcus barophilus, respectively [6][7][8]. Glutamate has also been suggested to be a piezolyte in the meso-and piezophilic Desulfovibrio hydrothermalis and D. piezophilus species [9,10]. Recently, it has been suggested that glutamate optimizes the enzymatic activities under high hydrostatic pressure [11]. Furthermore, it has been shown that several microorganisms adapt to high pressure by changing their energy metabolism; for instance, in the anaerobic sulfate-reducing bacterium D. piezophilus, transcriptomic and biochemical analyses have shown that the metabolite cycling H 2 /Formate is required for energy-conservation at high hydrostatic pressure [10,12]. Transcriptome analyses of the piezo-hyperthermophile Thermococcus piezophilus and Pyrococcus yayanosii highlighted recommended by Tettelin and colleagues [27]. The Roary pipeline generated a core gene alignment and gene presence/absence matrix that were used by Roary scripts to retrieve strain-specific genes (not belonging to any orthologous cluster) of T. elfii DSM9442 and T. elfii subsp. lettingae DSM14385. Each strain-specific protein was carefully checked for functional annotation by using the Conserved Domains Database at NCBI (ftp://ftp.ncbi. nlm.nih.gov/pub/mmdb/cdd (accessed on 6 January 2020)), which includes the COG (Clusters of Orthologous Groups of proteins) functional database (rpsblast of the predicted proteins against COG profiles included in the COG annotation selected for E-value below 10 −5 ). The collinearity of genomes was evaluated via MAUVE tool alignment [28].

Growth under High Hydrostatic Pressure
P. elfii DSM9442 and P. elfii subsp. lettingae (DSM14385) were grown anaerobically at 65 • C in Hungate tubes (18 mL) or rubber-stoppered Schott bottles (100 mL) fully filled with the medium previously described [22], which were inoculated at a 1:10 ratio. Tubes or bottles were inserted into 500 mL high-pressure bottles (HPBs) (HPB-500; Top Industrie, Vaux-le-Pénil, France), and the hydrostatic pressure was controlled by using a piloted pressure generator (PMHP 600-600; Top Industrie, Vaux-le-Pénil, France) connected to the HPBs [29]. The hydrostatic pressure applied to the HPBs was transmitted to the culture via the septum of the Hungate tubes or via the Schott bottle screw cap (Duran group, Germany), which was modified to include a rubber septum (Fisher Scientific, France). HPBs were incubated at 65 • C. Two consecutive subcultures were grown from frozen stocks in fully filled 18 mL Hungate tubes for 60 h at 65 • C at hydrostatic pressures of 0.1, 20, 30, and 40 MPa for P. elfii DSM9442 and of 0.1 and 20 MPa for P. elfii subsp. lettingae. Then, each culture was used to inoculate 100 mL of fresh medium in Schott bottles for growth under the same temperature and hydrostatic pressures. Cells were harvested in the mid-exponential growth phase and used for total RNA preparation. Five replicates of each culture were performed.

Aspartate Measurements
Intracellular aspartate quantitation was performed at 0.1 and 30 MPa on cells in mid-exponential growth cultured in 100 mL bottles at the indicated hydrostatic pressures (by using the same protocol as above for the RNA preparation). Triplicate cultures were performed. Cells were pelleted via centrifugation at 9000× g for 15 min at 4 • C and then washed once with 1 mL of 200 mM NaCl. After centrifugation, the pellet was suspended with 0.5 mL hot distilled water and incubated for 15 min at 100 • C, followed by incubation on ice for 10 min. Then, the extract was centrifuged at 15,000× g for 5 min at 4 • C to eliminate cell debris, and the supernatant was aliquoted, snap-frozen in liquid nitrogen, and stored at −80 • C. Aspartate was measured by using the fluorometric Aspartate assay Kit (Abcam, Cambridge, UK). Samples were plated in 96-well black microplates (Costar TM 96-well Assay Plates, Black Polystyrene) for fluorescence readings at Ex/Em 535 nm/ 587 nm on a TECAN-Spark Infinite M200 plate reader. In the meantime, the number of cells on each sample was counted as previously described [22] in order to relate the aspartate concentration to the cells amount. Experiments were performed in triplicate.

RNA-Seq Experiments
For RNA preparation, cells were grown in 120 mL flasks at the desired hydrostatic pressure until reaching mid-exponential growth. Then, cells were harvested via centrifugation at 6800× g for 15 min at 4 • C; the pellet was resuspended in 400 µL of 10 mM Tris-HCl (pH 8.0) and spun for 2 min at 12,000× g. The supernatant was removed, and the pellet was frozen in liquid nitrogen before being placed in −80 • C storage until use. RNA samples were prepared by using the Maxwell 16 LEV samples kit (Promega), and the quality was checked by using the Experion RNA analysis kit (Bio-Rad).
RNA library preparations and sequencing reactions were conducted at GENEWIZ, LLC. (South Plainfield, NJ, USA). Ribosomal RNA samples were depleted by using the Illumina Ribo-zero Plus rRNA depletion kit (Illumina, San Diego, CA, USA), and the sequencing libraries were prepared by using the NEBNext Ultra RNA Library Prep Kit for Illumina by following the manufacturer's recommendations (NEB, Ipswich, MA, USA). Sequencing libraries were validated by using a DNA Chip on the Agilent 2100 Bioanalyzer (Agilent Technologies, Palo Alto, CA, USA) and were quantified by using a Qubit 2.0 Fluorometer (Invitrogen, Carlsbad, CA, USA), as well as via quantitative PCR (Applied Biosystems, Carlsbad, CA, USA). The sequencing libraries were multiplexed and clustered on six lanes of a flowcell. After clustering, the flowcell was loaded on the Illumina HiSeq 2500 instrument according to the manufacturer's instructions. The multiplexed libraries were sequenced by using paired-end 150 cycle chemistry for the HiSeq 2500 (Illumina). Image analysis and base calling were conducted via the HiSeq Control Software (HCS) v2.2.68 on the HiSeq 2500 instrument. Raw sequence data generated from Illumina HiSeq 2500 was converted into fastq files and de-multiplexed by using Illumina CASAVA 1.8.2 software. One mismatch was allowed for index sequence identification. RNA-seq data were submitted to the NCBI Gene Expression Omnibus (GEO, https://www.ncbi.nlm.nih.gov/geo/ (accessed on 25 February 2023)) under the accession number GSE226101.
During the analysis, the selected replicates were submitted to a custom R script for differential expression analysis by using the "edgeR" package [38]. Two sets of samples, including all RNA-seq replicates for P. elfii DSM9442 and P. elfii subsp. lettingae were processed, revealing a very high overall quality of samples and an extensive coverage of reference genomes. Two samples from each set showed deviations from their corresponding replicates and were discarded from further analyses. All conditions retained at least 3 replicates for downstream analyses. Genes were considered significantly differentially expressed if their log 2 -fold change was above 1.3 and the Benjamini-Hochberg corrected p-value was below 0.01.
To provide functional information about differentially expressed genes (DEG), the full KEGG (Kyoto Encyclopedia of Genes and Genomes) database was downloaded [39][40][41]. Sequences for all annotations in reference genomes were aligned to KEGG Orthology (KO) annotation sequences by using "blastx" (E-value < 1 × 10 −10 , BitScore > 50) [42]. Resulting KO molecular function annotations were added to the original annotation files and reported as part of the results of differential expression analyses.

Comparative Genomics of the Deep and Surface P. elfii Strains
We conducted a comparative genomics study between the piezophilic strain P. elfii DSM9442 (accession number: AP014507) (optimal growth at 20 MPa) and the piezotolerant Microorganisms 2023, 11, 773 5 of 17 strain P. elfii subsp. lettingae DSM14385 (accession number: CP000812) (optimal growth at 0.1 MPa). The genetic closeness of the two genomes in combination with different sensitivities of the strains to high hydrostatic pressure provided a valuable framework to identify potential genetic determinants associated with the piezophilic lifestyle in the genus Pseudothermoga.
As previously reported, P. elfii and P. elfii subsp. lettingae have intra-species relationships with an average nucleotide identity score of 99.12% [21]. We found a strong genome collinearity (results not shown), indicating a high degree of gene order conservation. The difference in genome size between P. elfii DSM9442 and P. elfii subsp. lettingae DSM14385 is almost negligible, with 2,169,860 and 2,135,342 base pairs, respectively, while the guanine-cytosine (GC) content is the same (39%). Consequently, the number of coding sequences (CDS) is very similar between the two strains, with 2063 and 2047 CDS for P. elfii and P. elfii subsp. lettingae, respectively.
A comparison of all CDS (i.e., the predicted proteome) between both species resulted in the identification of orthologous proteins shared by both strains (core-genome) and of those unique to each strain (strain-specific). Our main focus was the identification of the strainspecific genes/proteins in the piezophilic strain P. elfii DSM9442 and, therefore, lost (those originating from the ancestral genome) in the surface strain P. elfii subsp. lettingae. These potential piezophilic markers will be the particular focus of the transcriptomic analysis (see below).
We identified a total of 77 genes that were specific to P. elfii DSM9442 (absent in P. elfii subsp. lettingae). They were mostly gathered in clusters and operons that were related mainly to carbohydrate transport and metabolism, cell envelope biogenesis, and amino acid transport and metabolism (Table S1). In total, 16 strain-specific genes were annotated as hypothetical proteins. Several clusters contained genes encoding ABC transporters, which are known to be involved in nutrient uptake but could also be linked to environmental stress responses [43]. Clusters 1 to 4 provided different ABC transporters for the transport and metabolism of sugar (Xylulose, L-Arabinose, or D-ribose due to strong operon homology) and phosphate (under phosphate starvation via the uptake of Glycerol-2P and Glycerol 3-P), for dipeptide transport (Cluster 3) and the transport of branched-chain amino acids [44] including L-Leucine, L-Phenylalanine, L-Isoleucine, and L-Valine (Cluster 4, LivKHMGF operon). Interestingly, the LivKHMGF operon was present in multiple copies in both strains, although Cluster 4, which was present solely in the deep strain P. elfii DSM9442, bore little similarity with the other Liv operons (under the homology cutoff). This Liv operon could be a paralog (Liv duplicate) under a neo-functionalization or pseudogenization process, which would explain the sequence divergence with respect to the other Liv copies. We identified another transport system in Cluster 7 that exhibited a TRAP system permease (DctPQM). Previous studies have shown that this system has a high affinity for importing C4-dicarboxylates, such as malate succinate and fumarate [45].
Interestingly, the genes involved mainly in fatty acid biosynthesis (AcP, FaD, and FabG; Cluster 6) were found to be specific to the deep strain, P. elfii, with the closest related proteins (by using a similarity-based search) belonging to Firmicutes representatives and to the anaerobic thermophilic bacterium Caldithrix abyssi, which was isolated from a depth of 3000 m [46].
Defense mechanisms were represented in Cluster 8 by the CRISPR-associated RAMP cmr system, which protects bacteria against foreign RNA from plasmids or viruses. It has also been shown to possess the ability to cleave endogenous RNA of Pyrococcus furiosus in vivo [47]. We also found a specific McrA endonuclease acting against foreign DNA. In the identified clusters, we found several glycosyl transferases for the initiation and elongation of glycan chains, flippase for the translocation of lipid-linked oligosaccharides [48], and a transcriptional regulator mainly for drug resistance.
We identified a total of 43 genes that were specific to P. elfii subsp. lettingae, 19 of which were found in 4 clusters that are related mainly to cell envelope biogenesis, amino acid transport (Cluster 3), and carbohydrate metabolism (Table S2). Cluster 4 is composed of genes involved in poly-γ-glutamate (PGA) synthesis. While PGA can bind to peptidoglycan to facilitate nutrient uptake, it can also be released into the environment to sequester toxic metal ions, decrease salt concentration, and protect against adverse conditions [49].

3.2.
Overview of the Modifications Induced by Hydrostatic Pressure at the Transcriptome Level of P. elfii DSM9442 and P. elfii subsp. lettingae Adaptation to high hydrostatic pressure should involve transcriptional regulation processes. In order to elucidate the hydrostatic pressure adaptation strategies of P. elfii DSM9442 and P. elfii subsp. lettingae, the transcriptome of cells in mid-exponential growth was analyzed via RNA-seq. The DEG values of P. elfii DSM9442 (grown at 0.1, 20, 30, and 40 MPa) and those of P. elfii subsp. lettingae (grown at 0.1 and 20 MPa) are listed in Tables S3 and S4, respectively. By using 0.1 MPa as the reference for the deep strain P. elfii DSM9442, a total of 58 genes was differentially expressed at 20 MPa (optimal hydrostatic pressure for growth), 211 genes were differentially expressed at 30 MPa, and 124 genes were differentially expressed at 40 MPa (maximum hydrostatic pressure for growth) ( Figure 1). transcriptional regulator mainly for drug resistance.
We identified a total of 43 genes that were specific to P. elfii subsp. lettingae, 19 of which were found in 4 clusters that are related mainly to cell envelope biogenesis, amino acid transport (Cluster 3), and carbohydrate metabolism (Table S2). Cluster 4 is composed of genes involved in poly-γ-glutamate (PGA) synthesis. While PGA can bind to peptidoglycan to facilitate nutrient uptake, it can also be released into the environment to sequester toxic metal ions, decrease salt concentration, and protect against adverse conditions [49].

Overview of the Modifications Induced by Hydrostatic Pressure at the Transcriptome Level of P. elfii DSM9442 and P. elfii subsp. lettingae
Adaptation to high hydrostatic pressure should involve transcriptional regulation processes. In order to elucidate the hydrostatic pressure adaptation strategies of P. elfii DSM9442 and P. elfii subsp. lettingae, the transcriptome of cells in mid-exponential growth was analyzed via RNA-seq. The DEG values of P. elfii DSM9442 (grown at 0.1, 20, 30, and 40 MPa) and those of P. elfii subsp. lettingae (grown at 0.1 and 20 MPa) are listed in Tables S3 and S4, respectively. By using 0.1 MPa as the reference for the deep strain P. elfii DSM9442, a total of 58 genes was differentially expressed at 20 MPa (optimal hydrostatic pressure for growth), 211 genes were differentially expressed at 30 MPa, and 124 genes were differentially expressed at 40 MPa (maximum hydrostatic pressure for growth) (Figure 1).   Figure 1). The number of DEG was higher at 30 and 40 MPa compared to 20 MPa, which shows that the response of the cells to hydrostatic pressures exceeding the optimal pressure for growth requires a larger variation in the transcriptome. It can be explained by the phenotypic response previously described, showing that at 30 and 40 MPa, the final biomass and growth rates were lower than at 20 MPa and 0.1 MPa [22].
In the surface strain, P. elfii subsp. lettingae, 31 genes were differentially expressed between 0.1 MPa and 20 MPa, with 21 being over-and 10 being under-expressed at 20 MPa (Table S4). This lower number of DEG (compared to P. elfii DSM9442) in response to pressure could be due to the more similar growth characteristics of P. elfii subsp. lettingae at 0.1 MPa and 20 MPa [22] 3.3. Differentially Expressed Genes in P. elfii DSM9442 and P. elfii subsp. lettingae in Response to Hydrostatic Pressure We conducted a systematic comparison between detected transcripts of P. elfii DSM9442 that were cultured at 0.1 MPa (atmospheric pressure) versus detected transcripts when the strain was grown at 20 MPa (optimal pressure for growth), 30 MPa, and 40 MPa (maximum pressure for growth). Overall, the expression patterns for P. elfii DSM9442 showed few overlaps for the different pressures ( Figure 1). Only 13 genes were found to be similarly differentially expressed for all 3 pressures (Table S3, Figure 1), with 6 overand 7 under-expressed. This included two genes involved in amino acid metabolism, namely, TEL01S_RS09135, which encodes a protein belonging to the alanine-glyoxylate aminotransferase family, and TEL01S_RS06965, which encodes an aspartate kinase. None of the orthologs of these genes was found to be differentially expressed in the surface strain P. elfii subsp. lettingae, suggesting that this response is specific to the deep strain.
When the transcriptomes of cells grown at 0.1 MPa and 20 MPa are compared, 21 genes were specifically over-and 16 were under-expressed at 20 MPa (Figure 1). Among the overexpressed genes, 14 (locus tags TEL01S_RS00165 to TEL01S_RS00175, TEL01S_RS00525 to TEL01S_RS00535, and TEL01S_RS08670 to TEL01S_RS08705) encode proteins of sugar transport and metabolism, while two genes (locus tags TEL01S_RS06345 and TEL01S_RS06350) encode an ABC-2 type transporter system. Furthermore, five of the under-expressed genes encode proteins linked to sugar transport and metabolism, and a cluster of five genes (locus tags TEL01S_RS05715 to TEL01S_RS05735) encode proteins involved in a branched-chain amino acid transport system (Table S3).
At even higher pressures of 30 and 40 MPa, we found a total of 67 genes similarly differentially expressed that were not differentially expressed at 20 MPa (Figure 1), 29 of which were over-expressed and involved mainly in the transport system and amino-acid metabolism. They include genes with locus tags TEL01S_RS00940 to TEL01S_RS00950, which encode a peptide/nickel transport system; TEL01S_RS01770 to TEL01S_RS01785, which are genes linked to the arginine biosynthesis pathway; TEL01S_RS03985 to TEL01S_ RS03995, which encode a branched-chain amino acid transport system; TEL01S_RS03635, which encodes an aspartate kinase LysC; and TEL01S_RS06945, which encodes the aspartatesemialdehyde dehydrogenase Asd linked to the glycine, serine, and threonine metabolism. Genes involved in amino acid metabolism and transport systems were also identified among the 38 genes under-expressed at 30 and 40 MPa (Table S3). These results support our hypothesis that the transport systems and amino acid metabolism play a pivotal role in the adaptation mechanisms of P. elfii DSM9442 to high hydrostatic pressures.
In total, 131 genes are specifically, differentially expressed between 0.1 and 30 MPa. Among those what were over-expressed, 9 encode proteins of the sugar transport systems; 14 encode other ABC-type transport systems, with those with locus tags TEL01S_RS02380 to TEL01S_RS02390 encoding a putative polar amino acid transport system; and 6 genes are involved in amino acid metabolism (Table S3). Interestingly, genes TEL01S_RS06225 to TEL01S_RS06235-which encode proteins AcpP, FadD, and FabG, which are involved in fatty acid (FA) biosynthesis and are absent in P. elfii subsp. lettingae genome-and TEL01S_RS09220-which encodes the UDP-N-acetylmuramoyl-tripeptide-D-alanyl-D-alanine ligase MurF, which is involved in peptidoglycan biosynthesis-were under-expressed.
A total of 36 genes was specifically differentially expressed between 0.1 and 40 MPa. A fabG homolog (TEL01S_RS00260) was over-expressed, while the cluster TEL01S_RS08165 to TEL01S_RS008175, which is absent in the P. elfii subsp. lettingae genome and which encodes an ABC-type peptide transport system, and genes TEL01S_RS09090-TEL01S_RS09095, which encode a 2-oxoacid:ferredoxin oxidoreductase linked to the TCA cycle and to gluconeogenesis, were under-expressed. Hence, those genes involved in FA and peptidoglycan synthesis, in the transport of peptide, and in amino acid metabolism tend to respond to the highest pressures.
In P. elfii subsp. lettingae, a total of 31 genes was found to be differentially expressed at 20 MPa (maximum pressure for growth), 21 of which were over-expressed, while 10 were under-expressed (Table S4). Of the former, nine genes of the cluster TEL01S_RS00925-TEL01S_ RS00980 encode proteins involved in ABC membrane transport and the metabolism of sugars (pentose-phosphate pathway), gene TEL01S_RS01250 encodes a protein linked to peptidoglycan biosynthesis, and gene TEL01S_RS08015 encodes a short-chain FA trans-porter. Of the 10 under-expressed genes, TEL01S_RS01020-TEL01S_RS01025 encode a peptide/nickel transport system.
Overall, these results show that genes that are involved in membrane transport play pivotal roles in the response of both P. elfii strains to high hydrostatic pressures. Notably, these gene expression patterns highlight the reshuffle of the branched-chain amino acid transport systems in the deep strain P. elfii DSM9442 in response to pressure. Moreover, the response in this strain also involved numerous genes linked to amino acid metabolism. In contrast, few of the differentially expressed genes in the surface strain P. elfii subsp. lettingae were linked to amino acid metabolism. Of the 31 genes differentially expressed in P. elfii subsp. lettingae, 11 orthologs were found to be differentially expressed in P. elfii DSM94442 as well (Table S5), although 9 of them were found to be expressed in opposing directions, which suggests a different pressure response altogether in these two strains.
3.4. DEG-Related KEGG Pathway Distribution in P. elfii DSM9442 and P. elfii subsp. lettingae To obtain a more integrated view of the transcriptomic variations in response to hydrostatic pressure, we analyzed the distribution of KEGG pathways that are related to the list of differentially expressed genes. A KEGG pathway was considered significantly enriched if its percentage was higher among the DEG-related vs. genome-related KEGGpathways (one-tail z-test, p-value < 0.05). By examining the patterns of DEG-related KEGG-pathway enrichments for P. elfii DSM9442 as a function of hydrostatic pressure, we found that the pathways involved in amino acid metabolism, the biosynthesis of other secondary metabolites, the metabolism of terpenoid and polyketides, and the metabolism of cofactors and vitamins were among the most enriched pathways, irrespective of the hydrostatic pressure (Figure 2). At 20 MPa, DEG-related KEGG pathways involved in membrane transport and amino acid metabolism were the most enriched (Figure 2A). At 30 and 40 MPa, the total number of enriched DEG-related KEGG-pathways was more than twice the number at 20 MPa ( Figure 2B,C vs. Figure 2A). A similar KEGG-pathway enrichment was highlighted between 30 and 40 MPa. In addition, the KEGG pathways involved in lipid metabolism were enriched only for hydrostatic pressures ≥30 MPa.
In P. elfii subsp. Lettingae, the enrichment pattern of DEG-related KEGG pathways revealed the membrane transport pathway as the most enriched pathway as for for P. elfii DSM9442 at 20 MPa (Figure 3).
The pathway involved in folding, sorting, and degradation was among the most enriched in P. elfii subsp. lettingae but not in P. elfii DSM9442. The latter could be linked to the piezotolerant capability of P. elfii subsp. lettingae [22], which allows bacteria to fold and/or degrade macromolecules at high hydrostatic pressures. Interestingly, DEGrelated KEGG pathways involved in amino acid metabolism were among the most enriched pathways in P. elfii DSM9442 but not in P. elfii subsp. lettingae.
Overall, this analysis reveals the common traits of adaptation to increasing hydrostatic pressure in both strains, namely, variations in transport membrane or carbohydrate metabolism. There were also several species-specific adaptations: variations in amino acid metabolism, lipid metabolism, or biosynthesis of secondary metabolites were mostly found in the deep strain P. elfii DSM9442, while changes to the folding, sorting, and degradation pathways were found in the surface strain P. elfii subsp. lettingae. This suggests that members of the genus Pseudothermotoga activate both similar and specific global functions in response to the hydrostatic pressure increases.

Variations in the Amino Acid Metabolism in P. elfii in Response to Pressure Changes
As mentioned above, the deep strain P. elfii DSM9442 responds to increases in hydrostatic pressure through changes in the metabolic pathways involved in amino acid metabolism. Twenty-eight genes encoding proteins that are involved in amino acid metabolism were detected as differentially expressed depending on the hydrostatic pressure (Tables 1 and S4), corresponding mostly to pathways linked to arginine biosynthesis, glycine serine threonine metabolism, cysteine methionine metabolism, and lysine biosynthesis.  In P. elfii subsp. Lettingae, the enrichment pattern of DEG-related KEGG pathways revealed the membrane transport pathway as the most enriched pathway as for for P. elfii DSM9442 at 20 MPa (Figure 3). The pathway involved in folding, sorting, and degradation was among the most enriched in P. elfii subsp. lettingae but not in P. elfii DSM9442. The latter could be linked to the piezotolerant capability of P. elfii subsp. lettingae [22], which allows bacteria to fold and/or degrade macromolecules at high hydrostatic pressures. Interestingly, DEG-related KEGG pathways involved in amino acid metabolism were among the most enriched pathways in P. elfii DSM9442 but not in P. elfii subsp. lettingae.
Overall, this analysis reveals the common traits of adaptation to increasing hydrostatic pressure in both strains, namely, variations in transport membrane or carbohydrate metabolism. There were also several species-specific adaptations: variations in amino acid metabolism, lipid metabolism, or biosynthesis of secondary metabolites were mostly found in the deep strain P. elfii DSM9442, while changes to the folding, sorting, and degradation pathways were found in the surface strain P. elfii subsp. lettingae. This suggests that members of the genus Pseudothermotoga activate both similar and specific global functions in response to the hydrostatic pressure increases.

Variations in the Amino Acid Metabolism in P. elfii in Response to Pressure Changes
As mentioned above, the deep strain P. elfii DSM9442 responds to increases in hydrostatic pressure through changes in the metabolic pathways involved in amino acid metabolism. Twenty-eight genes encoding proteins that are involved in amino acid metabolism were detected as differentially expressed depending on the hydrostatic pressure (Tables 1 and S4), corresponding mostly to pathways linked to arginine biosynthesis, glycine serine threonine metabolism, cysteine methionine metabolism, and lysine biosynthesis.   Notably, the lysC gene TEL01S_RS03635 coding for an aspartate kinase and the asd gene TEL01S_RS06945 encoding an aspartate-semialdehyde dehydrogenase, were overexpressed at both 30 and 40 MPa, while the thrA gene TEL01S_RS06965 encoding a bifunctional aspartokinase/homoserine dehydrogenase was over-expressed at 20, 30, and 40 MPa (Table 1). These gene products are involved in the synthesis of aspartate-4-semialdehyde from aspartate (Figure 4).
Aspartate-4-semialdehyde is a metabolite that is involved in the biosynthesis of methionine, threonine, and lysine. In addition, genes dapA, B, and D (TEL01S_RS00130, TEL01S_RS03640, and TEL01S_RS10080), which are involved in lysine biosynthesis, were over-expressed at 30 MPa ( Figure 4, Table 1), while genes metB, metH, mtnK, and mtnA (TEL01S_RS05650, TEL01S_RS06970, TEL01S_RS08685, and TEL01S_RS08690), which are involved in methionine metabolism, were over-expressed at 20 and 30 MPa. The gene cluster TEL01S_RS01770-TEL01S_RS01785, which encodes ArgGHCJ, was over-expressed at 30 and 40 MPa ( Table 1). The N-acetylglutamate synthase/glutamate N-acetyltransferase encoded by argJ and the N-acetyl-gamma-glutamyl-phosphate reductase encoded by argC are involved in the synthesis of ornithine, a component of the urea cycle from glutamate. The arginosuccinate synthase and the arginosuccinate lyase encoded, respectively, by argG and argH are enzymes of the urea cycle, allowing arginine synthesis from aspartate ( Figure 4). These variations in gene expression suggest that the arginine biosynthesis pathway from aspartate was boosted at high pressures. However, as the urea cycle produces fumarate, we could also hypothesize that feeding of the citrate cycle via fumarate is boosted when cells are grown at elevated hydrostatic pressure. This would not only provide energy to the cells but also contribute to gluconeogenesis via the production of oxaloacetate, as aspartate has been shown to be a glucogenic amino acid [50]. The gluconeogenesis pathway end-product, glucose-6-phosphate, is used to form six-carbon or five-carbon sugars needed for cell envelope synthesis. These include precursors of the peptidoglycan layer located in the periplasmic space, the lipopolysaccharides present in the outer membrane, and several sugars located in other cell compartments [51,52]. In a previous report, Roumagnac and collaborators [22] observed that under increasing hydrostatic pressure, P. elfii strains formed chains containing up to 15 cells per chain, possibly as a protective mechanism. An increased gluconeogenesis from aspartate at high pressures would, thus, allow cells to synthesize those molecules required for the cell envelope to facilitate chain formation. Notably, the lysC gene TEL01S_RS03635 coding for an aspartate kinase and the asd gene TEL01S_RS06945 encoding an aspartate-semialdehyde dehydrogenase, were overexpressed at both 30 and 40 MPa, while the thrA gene TEL01S_RS06965 encoding a bifunctional aspartokinase/homoserine dehydrogenase was over-expressed at 20, 30, and 40 MPa (Table 1). These gene products are involved in the synthesis of aspartate-4semialdehyde from aspartate ( Figure 4). Overall, the biosynthesis pathways of lysine, arginine, methionine, threonine, glycine, and serine from aspartate appeared to be boosted when cells were grown at high hydrostatic pressures, which is indicative of the pivotal role played by aspartate in the pressure response of the deep strain P. elfii DSM9442. In addition, the cluster TEL01S_ RS02380-TEL01S_RS02390, which encodes a polar amino acid transport system, was overexpressed at 30 MPa vs. 0.1 MPa, which may suggest that this transport system is activated at 30 MPa to optimize the import of aspartate into the cells. To determine whether the intracellular abundance of aspartate varied with hydrostatic pressure, we quantified the intracellular aspartate concentration in P. elfii DSM9442 at 0.1 and 30 MPa, i.e., the pressure conditions for which major differences in gene expression were found. At 30 MPa, intracellular aspartate in P. elfii DSM9442 was 4 times more abundant compared to 0.1 MPa (3.5 vs. 14 nmol/10 6 cells) ( Figure 5). The transcriptomic data provided no evidence for any increase in the aspartate biosynthesis pathway. Therefore, the higher aspartate abundance in cells grown at elevated hydrostatic pressure may be attributed to a more efficient aspartate import. These data support our hypothesis that aspartate is indeed a key metabolite for several pathways and adaptation to high pressures in P. elfii DSM94442. In the surface strain P. elfii subsp. lettingae, only one gene of the amin metabolism, which is involved in glycine biosynthesis (gene soxA, loc TEL01S_RS05600), was found to be over-expressed at 20 MPa vs. 0.1 MPa, wh orthologous gene in P. elfii DSM9442 (TEL01S_RS05590) was also over-expressed 30 and 40 MPa. Hence, in contrast to P. elfii DSM9442, amino acid metabolism in subsp. lettingae does not seem to play an important role in the response of the s high hydrostatic pressures.
Under these experimental conditions, peptides and/or amino acids conta biotrypcase are the only sources of carbon and energy available to the cel Fermentation of these compounds leads to the production of acetate and propio major products and, in lower quantities, formate, succinate, isobutyrate, and isov In P. elfii DSM9442, the production of these fermentation end-products is unaffe pressure, except for propionate, for which production increases slightly with inc hydrostatic pressure [22]. It is known that the catabolic pathways of some amin (valine, isoleucine, threonine, and methionine) can lead to the production of pro and ATP via propionyl-CoA [53]. As mentioned above, the threonine and met biosynthesis pathways from aspartate were boosted at high hydrostatic pressure that valine and isoleucine are branched-chain amino acids, it is worth pointing o the genes TEL01S_RS03990-TEL01S_RS03995, which code for a branched-chain acid transport system, were over-expressed at 30 MPa and 40 MPa vs. 0.1 MPa. Sim genes TEL01S_RS06345-TEL01S_RS06350, which encode another branched-chain acid transport system, were over-expressed at 20 MPa and 30 MPa. Over-expres these genes may lead to a higher rate of import of branched-chain amino acids i cells at high hydrostatic pressures, and through their catabolism, this could, i In the surface strain P. elfii subsp. lettingae, only one gene of the amino acid metabolism, which is involved in glycine biosynthesis (gene soxA, locus tag TEL01S_RS05600), was found to be over-expressed at 20 MPa vs. 0.1 MPa, while the orthologous gene in P. elfii DSM9442 (TEL01S_RS05590) was also over-expressed at both 30 and 40 MPa. Hence, in contrast to P. elfii DSM9442, amino acid metabolism in P. elfii subsp. lettingae does not seem to play an important role in the response of the strain to high hydrostatic pressures.
Under these experimental conditions, peptides and/or amino acids contained in biotrypcase are the only sources of carbon and energy available to the cells [22]. Fermentation of these compounds leads to the production of acetate and propionate as major products and, in lower quantities, formate, succinate, isobutyrate, and isovalerate. In P. elfii DSM9442, the production of these fermentation end-products is unaffected by pressure, except for propionate, for which production increases slightly with increasing hydrostatic pressure [22]. It is known that the catabolic pathways of some amino acids (valine, isoleucine, threonine, and methionine) can lead to the production of propionate and ATP via propionyl-CoA [53]. As mentioned above, the threonine and methionine biosynthesis pathways from aspartate were boosted at high hydrostatic pressure. Given that valine and isoleucine are branched-chain amino acids, it is worth pointing out that the genes TEL01S_RS03990-TEL01S_RS03995, which code for a branched-chain amino acid transport system, were over-expressed at 30 MPa and 40 MPa vs. 0.1 MPa. Similarly, genes TEL01S_RS06345-TEL01S_RS06350, which encode another branched-chain amino acid transport system, were over-expressed at 20 MPa and 30 MPa. Over-expression of these genes may lead to a higher rate of import of branched-chain amino acids into the cells at high hydrostatic pressures, and through their catabolism, this could, in turn, induce a higher production of propionate in the deep strain P. elfii DSM9442.

Adaptation Involving Modifications in Membrane Composition
A cluster of three genes (TEL01S_RS06225-TEL01S_RS06235) that are specific to the deep strain P. elfii DSM9442 and which encode proteins AcpP, FadD, and FabG was under-expressed at 30 MPa vs. 0.1 MPa. Similarly, the gene TEL01S_RS04130, which encodes a FabG protein, was under-expressed at 30 and 40 MPa, while the paralog of fabG (TEL01S_RS00260) was over-expressed at 40 MPa (Table S3). This fabG redundancy is likely due to gene duplication or addition events [54]. The multi-enzymes FabG, FabA, FabI, and FabB perform the elongation steps of FA synthesis [55]. FabG is a dehydrogenase that interacts with the FA chain that is bound to the acyl-carrier protein (Acp), which, in turn, binds to substrates of varying lengths. In addition, FabG often substitutes other dehydrogenases to produce secondary metabolites that may play a role in membrane permeability. It is also involved in polyhydroxyalkanoate (PHA) production, a form of carbon and energy storage under stress conditions [56]. In view of these putative functions, the differential expression of the fabG paralogs may explain the slight modification of the FA membrane composition in response to increasing hydrostatic pressure and/or the microscopy observations of the chain morphotype that was previously reported at the highest hydrostatic pressure [22]. It should be noted that no variation in the expression of these genes was detected in P. elfii subsp. lettingae, which is in agreement with the very low number of chains observed in the culture of this strain at high hydrostatic pressure [22]. The regulation of the fabG gene expression may govern modifications of the cell membrane and/or the toga, making cells more resistant to the highest hydrostatic pressures (>20 MPa). This leads us to hypothesize that each of the three FabG proteins may be structurally adapted to be active at a specific pressure (e.g., TEL01S_RS00260, which is over-represented at 40 MPa, could be adapted at high-pressure conditions). Such adaptation strategies have been previously reported, notably for the flagellar systems of the piezophilic Photobacterium profundum SS9 strain, in which one system was preferentially expressed at high pressure, while another was preferentially expressed at low (atmospheric) pressure [57].

Conclusions
In this paper, we reported for the first time the adaptation at the transcriptomic level to hydrostatic pressure increases of two strains from the same species, namely, the deep strain P. elfii DSM9442 and the surface strain P. elfii subsp. lettingae. The genetic closeness of the two strains provides a valuable framework to identify the genetic determinants that are associated with the piezophilic lifestyle in the genus Pseudothermoga. In that way, the gene cluster which is involved in lipid metabolism (TEL01S_RS06225-TEL01S_RS06235), which is specific to the deep strain, was differentially expressed at high hydrostatic pressures (30 and 40 MPa) and might be a good candidate for a piezophilic gene marker in the Pseudothermotogales. Our analysis revealed the common traits of adaptation to increasing hydrostatic pressure in both strains as well as several species-specific adaptations. Membrane transport appears to play an important role in the adaptation of both strains to 20 MPa. A key outcome of the transcriptomic analysis was that amino acid metabolism and transport appear to be crucial for the adaptation to high pressures in the deep strain P. elfii DSM9442. In addition, our analysis revealed that several metabolic pathways for amino acid biosynthesis that involve aspartate are boosted at high hydrostatic pressures, which suggests that a reshuffling of the amino acid metabolism may take place in response to pressure increases, which would represent an important adaptation mechanism in the deep strain P. elfii DSM9442. Given that the effects of high pressure on microorganisms inhabiting deep-oil reservoirs have rarely been studied, our work provides valuable new information and insights regarding the dynamics of oil reservoir microbiomes.

Supplementary Materials:
The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/microorganisms11030773/s1, Table S1: List of genes specific to the deep strain P. elfii DSM9442 compared to the surface strain P. elfii subsp. lettingae; Table S2: List of genes specific to the surface strain P. elfii subsp. lettingae compared to the deep strain P. elfii DSM9442; Table S3: List of differentially expressed genes in the deep strain P. elfii DSM9442 grown at 20 and 30 MPa vs. cells grown at 0.1 MPa; Table S4: List of differentially expressed genes in the surface strain P. elfii subsp. lettingae grown at 20 MPa vs. 0.1 MPa; Table S5: List of differentially expressed genes in the surface strain P. elfii subsp. lettingae that have differentially expressed orthologs in the deep strain P. elfii DSM9442.