Introduction

Toxin–antitoxin (TA) systems are widely spread in chromosomes and plasmids of bacteria and archaea (Goeders and Van Melderen 2014; Schuster and Bertram 2013; Yamaguchi et al. 2011). So far, TA systems have been categorized into six sub-classes (I–VI) based on the mechanisms of action and chemical nature of the antitoxin (Page and Peti 2016; Unterholzner et al. 2013). A typical TA system is composed of a toxin and antitoxin gene. Toxin gene always encodes a stable protein in all of the characterized bacterial TA systems, while the antitoxin gene is either protein (type II, IV, V and VI) or small non-coding RNAs (type I and III) (Schuster and Bertram 2013; Yamaguchi et al. 2011).

TA systems were firstly discovered on low copy plasmids as plasmid maintenance modules that were able to kill the daughter cells who failed to inherit the plasmids (Gerdes et al. 1986; Winther and Gerdes 2011). To date, some researchers believed that TA systems contribute to the adaptation of hosts under stress environments and the formation of persisted cells and biofilms (Wang and Wood 2011). In addition, TA systems also involve in stabilization of genomic islands (Bustamante et al. 2014) and selfish alleles (Ramisetty and Santhosh 2017), regulation of gene expression (Landini et al. 2014; Wang and Wood 2011) and programmed cell death (Hazan and Engelberg-Kulka 2004).

Moreover, there is mounting evidence showed that some TA elements can provide phage resistance, like Hok/Sok (Pecota and Wood 1996), MazEF (Hazan and Engelberg-Kulka 2004), RnlAB and LsoAB (Otsuka and Yonesaki 2012). The abortive infection (Abi) systems, for example, ToxIN of Pectobacterium atrosepticum and AbiE of Lactococcus lactis or Streptococcus agalactiae, have been shown to exert their function via toxin–antitoxin mechanism (Dy et al. 2014; Fineran et al. 2009).

AbiE of L. lactis consists of an antitoxin AbiEi and a toxin AbiEii, which are characterized by two conserved domains COG5340 and DUF1814, respectively. Such kind of two-component systems with COG5340-DUF1814 domains are found in 430 sequenced genomes of most major lineages of archaea and bacteria (Makarova et al. 2013). Homology of the DUF1814 domain also has been discovered in AbiG families (Makarova et al. 2013; O’Connor et al. 1999). It has been confirmed that GTP can specifically binds to the DUF1814 protein and appear to function as the toxin element, whereas the COG5340 protein that contains a predicted HTH domain is an antitoxin (Dy et al. 2014). AbiE system had been proved mostly located in many mobile genetic elements (MGEs) and plays a role in MGEs maintenance (Dy et al. 2014). In addition, AbiE of L. lactis provides resistance against the 936-type phage family by preventing DNA packaging (Tangney and Fitzgerald 2002).

AbiE system functions via a non-interacting (type IV) bacteriostatic TA mechanism. It has been assigned as an important representative in type IV TA system. The other members of the type IV TA system also have been identified, including YeeU–CbtA (Masuda et al. 2012a; Tan et al. 2011) and CptB–CptA (YafY/SdhE-YafX) (Masuda et al. 2012b) systems from Escherichia coli. Unlike AbiE, these groups of type IV TA system share the same feature to control polymerization of cytoskeletal proteins MreB and FtsZ.

TA systems have been extensively studied in bacteria, for example, at least 33 TA systems have been identified in E. coli (Yamaguchi et al. 2011). However, most of the studies on the TA system of archaea are based on bioinformatics analysis to identify mainly type II TA systems. RelBE and VapBC have been predicted as type II TA systems located in the mobile elements of Thermococcales and Methanococcales (Krupovic et al. 2013). Additionally, at least 26 TA gene loci (VapBC family) have been annotated in the genome of hyperthermophilic crenarchaeon Sulfolobus solfataricus (Cooper et al. 2009; Tachdjian and Kelly 2006). Global transcriptional analysis of S. solfataricus revealed that expression of several vapBC genes were triggered by the thermal shift from 80 to 90 °C, suggesting a role of TA system played in heat-shock response.

Pyrococcus yayanosii CH1 is a strictly piezophilic hyperthermophilic archaeon isolated from sample of 4100 m depth hydrothermal vent in the mid-Atlantic Ridge (Birrien et al. 2011; Jun et al. 2011; Zeng et al. 2009). The strain CH1 can grow within a temperature range of 80–108 °C and a pressure range of 20–120 MPa, with optima at 98 °C and 52 MPa, respectively (Zeng et al. 2009). In our previous study, a genomic island PYG1 of P. yayanosii was identified. Genetic analyses suggest that PYG1 could affect the host adaptation to the high temperature and HHP environments (Li et al. 2016b). Two genes in PYG1 encompass a pair of prokaryotic orthologous group families COG5340-DUF1814 (PYCH_15320-PYCH_15330), which showed high similarity to the corresponding region of a predicted genomic island in T. barophilus MP (Li et al. 2016b), formed a putative toxin–antitoxin system (PygAT) with AbiE conserved domains.

In this study, we showed that PygAT is a functional gene operon of P. yayanosii CH1 that could response to environment stress such as HHP. Either disruption of antitoxin gene pygA in P. yayanosii or overexpression of toxin gene pygT in E. coli could inhibit the growth of host.

Materials and methods

Strains, plasmids and culture conditions

Strains and plasmids used in this study are listed in Table 1. Bacterial growth was measured in a UV–VIS spectrophotometer at 600 nm (OD600). P. yayanosii strains were cultivated in 100 ml serum bottles under anaerobic conditions at 95 °C and 0.1 MPa in 30 ml of TRM (Li et al. 2016b). After transformation, the strains were selected on TRM supplemented with 10 μM simvastatin (Sigma). Gelrite (1.5% w/v) was added to solidify the medium. The growth of archaea was monitored by cell counting using a Thomas chamber and light microscopy at a magnification of 40× (Li et al. 2016b; Zeng et al. 2009). Experiments were repeated in three biological replicates.

Table 1 Strains and plasmids used in the present study
Full size table

Bioinformatics analysis

Annotations of the gene function were confirmed by BLASTP analysis against NCBI database (https://www.ncbi.nlm.nih.gov/). Alignment of the amino acid sequences were performed using ClustalX2 and visualized by ESPript 3.0 (http://espript.ibcp.fr/ESPript/ESPript/). The second structure of proteins was analyzed using the web-based tool SWISS-MODEL (https://swissmodel.expasy.org/ interactive). The phylogenetic analysis was performed using the maximum likelihood phylogenetic tree algorithm with a software package for constructing evolutionary trees (MEGA, version 6.0).

Genetic manipulation and real-time quantitative PCR

Pyrococcus yayanosii A2 was transformed according to Li et al. (2016b). P. yayanosii A2 was cultivated in 50 ml TRM medium at 95 °C, 0.1 MPa for 12 h, and cells in the late exponential growth phase were harvested and subsequently resuspended in 200 µl of cold transformation buffer (80 mM CaCl2). The suspended cells were incubated on ice for 0.5 h under anaerobic conditions, and subsequently 3 μg homologous recombinant suicide plasmids were added to the suspension and incubated on ice for 1 h. After heat shock at 95 °C for 45 s, the suspension was incubated on ice for 10 min. The treated suspension was transferred to 8 ml of TRM medium (without simvastatin) and cultured for two generations. The culture was spread onto TRM solid medium supplemented with 10 μM simvastatin and further incubated for 3–5 days at 95 °C. Transformants were verified by PCR amplification. Subsequently, the native transformant was counter selected by 5-FOA. The resulting mutant was confirmed by PCR amplification.

The genomic DNA from P. yayanosii was extracted as previously reported (Li et al. 2015). The total RNA of P. yayanosii was extracted by Trizol. RNA purification was performed with DNase I (Thermo). The cDNA was prepared using a cDNA synthesis kit (Thermo). Plasmids were extracted from E. coli DH5α using the plasmid extraction kit (Omega). The DNA purification was conducted using a DNA gel extraction kit or the Cycle-Pure Kit (Omega). The restriction endonucleases and T4 ligase were purchased from Takara or NEB. DNA sequencing and oligonucleotide synthesis were performed at Sangon (Sangon Biotech). Real-time quantitative PCR was carried out with an ABI 7500 system. 16S rRNA gene was used as an internal standard. The 2−∆∆CT method was used to analyze the relative gene expression.

Culturing experiments under HHP

The procedure of HHP culturing experiments is consistent with that described previously (Li et al. 2016b). All manipulations before the pressurized culturing experiments were performed anaerobically inside an anaerobic glove box (Coy Lab). Cultivation of P. yayanosii was performed using a custom-built high-pressure/high-temperature incubation system. A 10-ml plastic syringe was used as the container of liquid medium. After inoculation, the needle head of the syringe was sealed tightly with a butyl rubber stopper. The syringe was then placed inside a titanium chamber that was pressurized to the appropriate hydrostatic pressure and maintained at high temperature.

Toxicity and antitoxicity assays

The putative toxin gene pygT and pygA–pygT gene pairs were amplified using primers listed in Supplementary Table 1. The products were cloned into pET28a and the resulting recombinants were transformed into E. coli BL21. E. coli were cultured in 20 ml LB with 50 µg/ml kanamycin at 37 °C. When the OD600 reached to about 0.8, 50 μl bacteria solution of different dilutions were inoculated on the LBA plates which contain or not contain inducer IPTG (final concentration 0.5 mM), respectively. These LBA plates were incubated overnight at 37 °C. Moreover, IPTG was also added to the remaining culture medium to induce toxin expression. 100 μl bacteria solution were diluted with phosphate-saline buffer (PBS) and spread onto the LBA plates to determine the number of colony-forming units (CFU/ml) every 2 h from OD600 reached to about 0.4.

Plasmid stability assays

Recombinant plasmids pUC19/pygAT and pLMO12103/pygAT contain the PygAT system were constructed. An E. coli plasmid pUC19 and P. yayanosii–E. coli shuttle plasmids pLMO12102 and pLMO12103 were control groups, respectively. These plasmids were transformed into E. coli DH5a and cultured in LB medium without any antibiotics at 37 °C and 200 rpm for 24 h. The above strains were serially sub-cultured for 5 days, and each day of bacterial solution was diluted and spread on antibiotics-free plate. About 100 colonies were randomly selected to replica on plates containing antibiotics to determine the plasmids maintenance.

Protein expression and purification

Escherichia coli BL21 (DE3) strain harboring pET28a/pygA plasmid was cultivated in 1 l of Luria–Bertani (LB) medium supplemented with 50 µg/ml kanamycin at 37 °C for 4 h. When the OD600 reached about 1.0, expression of the protein His6-PygA was induced with 0.5 mM isopropyl-b-d-thiogalacto-pyranoside (IPTG) and the culture was cultivated overnight at 16 °C, 200 rpm. Cells were harvested by centrifugation at 8000×g for 10 min at 4 °C. The cell pellet was resuspended in 25 ml lysis buffer (20 mM Tris–HCl, 300 mM NaCl, 10% glycerol, pH 8.0) and sonicated on ice for 30 min. Then, cells were centrifuged by centrifugation at 10,000×g for 30 min at 4 °C. Collect the supernatant and place it in a column containing 1 ml of Ni Sepharose High Performance resin (GE Healthcare). Unbound proteins were removed through washing three times with 10 ml of the wash buffer (20 mM Tris–HCl, 300 mM NaCl, 10% glycerol, 20 mM imidazole, pH 8.0). Protein His6-PygA was eluted with elution buffer (20 mM Tris–HCl, 300 mM NaCl, 10% glycerol, 300 mM imidazole, pH 8.0). Proteins were dialysed overnight with 20 mM Tris–HCl, 300 mM NaCl, 50% glycerol. The purified protein was stored at − 20 °C and its concentration was determined by the Bradford method (Sangon Biotech). 12% SDS-PAGE was used to confirm the purity of protein.

Electrophoretic mobility shift assay (EMSA)

The putative promoter region of pygAT was amplified with specific primers (From upstream 226 bp of the initiation codon ATG to the downstream 22 bp of it.) and purified with Cycle Pure kit (Omega). The EMSA reaction system (20 μl) is as follows: 0.5 μl DNA was incubated with different amounts of protein His6-PygA in 4 μl EMSA buffer (100 mM HEPES, pH 7.5; 5 mM EDTA; 5 mM DTT and 1% w/v Tween 20,150 mM KCl), 1 μl Poly d[I-C] and 13.5 μl ddH2O at room temperature for 20 min. The mixtures were then loaded onto 6% native polyacrylamide gel electrophoresis at 100 V for 1 h at room temperature and then the gel was stained with GelRed.

DNase I footprinting assay

DNA probe which contains pygAT promoter region was amplified with FAM-labeled primers (He et al. 2015; Li et al. 2016a). The reaction system was the same as EMSA experiment. The mixture was treated with 0.02 units of DNase I (ThermoFisher) at room temperature for 10 min. The reaction was stopped by adding 0.25 μl 0.5 M EDTA and incubating in a water bath at 75 °C for 15 min. Then, the treated DNA was purified and mixed with GeneScan-LIZ500 size standard, assayed with Applied Biosystems 3730XL DNA analyzer (identified by Sangon Biotech) (He et al. 2015; Li et al. 2016a). Electropherograms were analyzed using the GeneMarker software. Protein binding sites were determined by analyzing the decrease of peaks after binding of different concentrations of proteins to DNA.

Results

A putative TA system was identified in the genomic island PYG1 of P. yayanosii

In the genomic island PYG1, a putative TA system was identified as a bicistronic operon based on the blast search of the conserved domains (COG5340-DUF1814) against NCBI database. PygA (PYCH_15320) possess a winged helix-turn-helix (wHTH) domain of AbiEi-related protein (Aravind et al. 2005; Dy et al. 2014), and it has a predicted secondary structure of H1-S1-H2-t-H3-S2-W1-S3-W2 (H = helix. S = strand, and W = Wing) that represent of typical antitoxin protein (Fig. 1a). Multiple sequence alignment of the putative toxin protein PygT (PYCH_15330) revealed that the enzyme shared four conserved motifs with other homologs that belongs to the AbiEii family from different sources (archaea and bacteria) (Fig. 1b). Either antitoxin protein PygA or toxin protein PygT exhibit low sequence identity (11.15 and 17.36%, respectively) to the antitoxin protein AbiEi (Accession no. WP_012477749.1) and the toxin protein AbiEii (Accession no. WP_012477748.1) of Lactococcus lactis (Fig. S1). The phylogenetic analysis of PygA and PygT shows that they are located in a different branch compared with their counterparts from bacteria (Fig. S2).

Fig. 1
figure 1

Schematic diagram of conserved domains COG5340-DUF1814 of PygAT a PYCH_15320 sharing a wHTH domain with AbiEi (Accession no. WP_012477749.1) of Lactococcus lactis and Rv2827c (Accession no. NC_000962.3) of Mycobacterium tuberculosis. b Multiple amino acid sequence alignment of PYCH_15330 with homologs showing four motifs (I–IV). WP_013468141.1: hypothetical protein from Thermococcus barophilus MP; WP_013479616.1: hypothetical protein from Asticcacaulis excentricus; WP_002992222.1: hypothetical protein from Myroides odoratus; WP_013617187.1: hypothetical protein from Bacteroides salanitronis; WP_012584350.1: hypothetical protein from Thauera sp. MZ1T; WP_049765176.1: hypothetical protein from Olsenella uli; WP_012477748.1: AbiEii from Lactococcus lactis. Black boxes represent motifs

Full size image

pygA and pygT were organized into an operon

In PYG1, the putative toxin gene pygT overlap with its upstream antitoxin gene pygA by 11 nucleotides, and were separated by a 13 bp region with PYCH_15340 (Fig. 2a). Amino acid sequence of PYCH_15340 shared 93% identity with the N-terminal end of a putative protein TERMP_01870 of Thermococcus barophilus MP (Accession no. WP_013468141.1). TERMP_01870 and PygT have 95% amino acid sequence identity.

Fig. 2
figure 2

Co-transcription analysis of pygAT operon in the P. yayanosii genome. a Schematic diagram of the bicistronic structure of pygAT. b Confirmation of the co-transcript of pygA and pygT using primers q1532F/q1533R and c primers q1533F/q1534R

Full size image

Using reverse transcripted total RNA of P. yayanosii as templates, cDNAs were amplified by specific primers to analyze whether the putative TA loci gene pairs were co-transcribed or not (Fig. 2a). A band with the expected size was obtained for pygA–pygT (Fig. 2b), but not for pygT-PYCH_15340 (Fig. 2c). Genomic DNA and RNA also were used to amplify as positive and negative controls, respectively. These results suggested that pygAT was organized into an operon.

pygA and pygT constitute a functional toxin–antitoxin module

Recombinant expression plasmids pET28a/pygT and pET28a/pygAT were transformed into E. coli BL21 (DE3) to verify that the PygAT is a functional TA system. The assumed toxin gene pygT was overexpressed and can inhibit cell growth after IPTG induction (Fig. 3a). Whereas, strains that contained a complete PygAT system or only contained empty vector pET28a did not inhibit cell growth (Fig. 3a). Meanwhile, the normal growth also was observed without IPTG induction. Then, the strains that transformed with pET28a and pET28a/pygT plasmids were cultured in liquid LB medium. We found that the growth of cells was inhibited after the toxin gene pygT was induced by IPTG at OD600 = 0.8 (Fig. 3b). In addition, when the OD600 reached to about 0.4, the bacteria was diluted and coated on the LBA plates containing ampicillin resistance, we can see that the number of bacterial colonies overexpressed the toxin gene pygT was lower than that of the control group (Fig. 3c). The results above suggested that PygAT is a functional toxin–antitoxin module.

Fig. 3
figure 3

Toxic effect of PygT on the growth of E. coli BL21 (DE3). a Overexpression of pygT inhibited cell growth. Plasmids pET28a and pET28a/pygAT were control groups in this experiment. b Strains which contain plasmids pET28a and pET28a/pygT were cultured in liquid LB medium and the toxin gene pygT was induced by IPTG at OD600 = 0.8. c Bacteria was diluted and coated on the LBA plates containing ampicillin resistance and then was counted on the flat plate. Data shown are mean ± SD of three biological replicates

Full size image

PygAT plays a role in maintaining P. yayanosii–E. coli shuttle plasmids in E. coli

The AbiE family TA systems usually present in mobile genetic elements (MGEs), such as plasmids or genomic islands, and almost all types of TA systems have the ability to stabilize the extrachromosomal genetic elements. To test whether the PygAT system can also provide stabilization for MGEs, we transformed several recombinant plasmids pLMO12103/pygAT and pUC19/pygAT (including PygAT) into E. coli (DH5a). An E. coli plasmid pUC19 and P. yayanosii–E. coli shuttle plasmids pLMO12102 and pLMO12103 were control groups, respectively. The transformants were cultured in liquid LB medium without any antibiotics. After 5 days, the survival rates of the two strains were calculated by colony counting method on the LBA plates with ampicillin resistance. We found that the numbers of the strain carrying pLMO12102 and pLMO12103 plasmids declined sharply after successive transfer culturing on the LBA plate with ampicillin resistance. The survival rate on the fifth day was only about 0.625–2.2%, while strains carrying pLMO12103/pygAT, pUC19/pygAT and pUC19 plasmids in LBA plate did not affect the survival ability (Fig. 4). Shuttle plasmid pLMO2102 was constructed on the basis of pUC19 by adding the P. abyssi plasmid pGT5 replication protein Rep75. These results indicated that PygAT system plays a crucial role in maintaining of P. yayanosii–E. coli shuttle plasmids containing archaeal plasmid replication protein in E. coli.

Fig. 4
figure 4

PygAT improved P. yayanosii–E. coli shuttle plasmid stability. E. coli DH5α carrying plasmids with and without the PygAT operon (pLMO12103/pygAT, pUC19/pygAT and pLMO12102, pLMO12103 and pUC19, respectively) were used to verify the plasmid stability without antibiotic selection. Ratio of plasmids maintenance was calculated by counting the difference in colony numbers between replica plates containing or not containing antibiotics

Full size image

PygT affects HHP adaptation of P. yayanosii

We constructed several gene knockout mutant strains of P. yayanosii including △pygT, △pygA and △pygAT to evaluate the physiological function of PygAT system. These strains were cultured at 0.1, 52 MPa (the optimum pressure) and 80 MPa. The growth curves of strains showed that there were no significant differences between the mutant and the parent strains at 0.1 MPa (Fig. 5a). However, comparing 52 and 80 MPa, we found that the lag phase of mutant strain △pygA was significantly prolonged. The mutant strain △pygA entered the logarithmic growth phase after growing 16 h (52 MPa) or 24 h (80 MPa). The gene complementary strain C△pygA was constructed and the growth of the strain was recovered to wild type (Fig. 5b, c). We analyzed the transcription level of pygT gene in wild strain A2, mutant strain △pygA and complementary strain C△pygA under 0.1 and 52 MPa. The results showed that the transcription level of pygT was derepressed in △pygA when the antitoxin was missing, while the transcription level of pygT was repressed in C△pygA when the antitoxin was complemented (Fig. 5d). These data confirmed that the transcription of pygT could respond to HHP stress in P. yayanosii, and the antitoxin gene pygA can regulate the expression of pygT in vivo.

Fig. 5
figure 5

PygAT affected the growth of P. yayanosii under different hydrostatic press conditions. The growth curves of P. yayanosii strains (A2, △pygA, △pygT, △pygAT, C△pygA, C△pygT) under a 0.1 MPa, b 52 MPa and c 80 MPa. d Transcription levels of pygT gene in (A2, ∆pygA, C∆pygA) under 0.1 and 52 MPa

Full size image

PygA specifically binds to two regions in the promoter of the pygAT operon

EMSA was performed to assess the direct interaction of the His6-PygA protein (Fig. S3) with pygAT operon promoter region. In reaction system, the promoter DNA showed obvious blockage with the increase of His6-PygA protein concentration, while no blocking was observed in the control groups (Fig. 6a). Furthermore, DNase I footprinting technique was used to identify the exact banding sites of His6-PygA protein in promoter region. We found that the peaks of two regions decreased or disappeared after binding to DNA with different concentrations of protein. It suggested that these two regions in the pygAT operon promoter were apparently protected by His6-PygA protein. We also found a pair of palindromes (TTTGAATAACAACCA and TGGTTGTTATTCAAA) in these two regions (Fig. 6b). All these results suggested that the antitoxin gene pygA can regulate the expression of pygT through specifically banding to the promoter region of pygAT operon.

Fig. 6
figure 6

Antitoxin protein PygA specifically binds to the putative PygAT promoter region. a EMSA experiment was performed to verify the binding of PygA protein (from 0 to 68 pmol) with the putative promoter DNA. b DNase I footprinting assay on the binding site of PygA and promoter region. Black arrows represent a pair of palindrome sequences (TTTGAATAACAACCA and TGGTTGTTATTCAAA). Black box (TATT) was predicted as a putative TATA-box

Full size image

Discussion

The AbiE family was firstly recognized as a pair of genes with COG5340-DUF1814 conserved domains (Garvey et al. 1995) and has been assigned as type IV TA system in several bacteria, such as L. lactis and S. agalactiae (Dy et al. 2014). To date, the understanding of the bacterial AbiE TA system is still very limited and almost there is no information of archaeal AbiE TA system.

There are 11 potential TA loci in the genome of P. yayanosii CH1 as predicted by a toxin–antitoxin database TADB (Shao et al. 2011). The putative TA system PygAT we studied here was not included in TADB but was found by a Blast search. Although the sequence homolog between PygAT on the genomic island PYG1 of P. yayanosii to conserved domains of AbiE family proteins of L. lactis was relative low, it provides good opportunity to study the archaeal counterparts in a phylogenetic point of view of type IV TA systems.

Our experiments indicated that the putative PygAT system can work as a functional TA module in E. coli. Interestingly, the toxin showed slight inhibition effect when the strains were cultured in liquid medium. This phenomenon is similar to the results of previous study by Dy et al. (2014). Research on the AbiE system of L. lactis also has been verified that it is a reversible bacteriostatic TA system. Four hours after the AbiEii was induced, the cells could be restored to the initial cell numbers (Dy et al. 2014). These results indicate that the inhibition of the toxin on the growth of the bacteria could not be a lethal effect, but a stress response in the face of external stimulus.

The MosAT system located in the Vibrio cholera SXT ICE can promote the genomic island maintenance, which has a conserved domain COG2253-COG5340 similar to AbiE (Wozniak and Waldor 2009). Although we have no direct evidence to prove the effect of PygAT on the stability of the genomic island, but we have confirmed that PygAT enhances the genetic stability of shuttle plasmid carrying archaeal plasmid replication protein Rep75 in E. coli. PygAT has a close homologue, i.e., TbgAT, which is also located in a genomic island TBG1 of T. barophilus MP (Li et al. 2016b). We assume that the presence of the either PygAT or TbgAT contributes to the stabilization of the genomic islands PYG1 and TBG1 in these two piezophilic archaeon, respectively.

To find whether there is a relationship between physiological function of PygAT and HHP stress in P. yayanosii, we constructed a series of gene knockout mutants, and measured their growth curves under different pressure conditions. We found that the lag phase of mutant strain △pygA was significantly prolonged with the pressure was improved to 52 and 80 MPa. The transcription level of toxin gene pygT was significantly up-regulated in △pygA strain. Subsequently, the EMSA and DNase I footprinting experiments confirmed that PygA can regulate the expression level of pygT in P. yayanosii by interacting with two DNA regions of in the promoter region of pygAT operon. These results indicated that regulation on PygAT is part of the adaptation to HHP stress response in P. yayanosii.

In bacteria, type II TA system MqsR/MqsA regulate the general stress response (GSR). The GSR allows cells to survive long periods of starvation and environmental stresses. And it is accompanied by a significantly reduced growth rate, and it appears that TA systems are the means by which growth is slowed. Specifically, during stress, Lon protease degrades MqsA, which leads to induction of the stress regulating factor RpoS in E. coli, which in turn increases the concentration of the second messenger 3,5-cyclic diguanylic acid (c-di-GMP), inhibits motility and increases cell adhesion and biofilm formation (Wang et al. 2011). Moreover, AbiEii (DUF1814) of L. lactis functions as a nucleoside transferase (NTase) belong to the DNA polyβ superfamily, which can specifically bind GTP and transfer nucleotide to an unknown target (Dy et al. 2014). In our study, we found that PygT can specifically bind GTP when purified PygT was incubated with 5′-FAM labeled GTP and ATP (Fig. S4). According to these, we suspect that the antitoxin is deleted when the cells in high pressure uninhibited toxin PygT transferred nucleotides as a signaling molecule to the target, and further affects the biological processes that are significantly affected by the effect of pressure.

A lot of processes like transporters, motility and respiratory chain components, directly related to the membrane composition are also affected by HHP in piezophiles (Simonato et al. 2006). In P. yayanosii CH1, the biological processes that are significantly affected by the effect of pressure may involve in translation, chemotaxis, energy metabolism (hydrogenases and formate metabolism) and CRISPR sequences associated with cellular apoptosis susceptibility proteins (Michoud and Jebbar 2016). Meanwhile, due to the potential substrate diversity of NTase, it is difficult to accurately predict the toxin protein substrate. Therefore, to explain the mechanism of the phenomenon of △pygA under high pressure, it is necessary to do further research.

Meanwhile, the very low biomass and metabolic activity during the lag phase of the bacterial or archaeal growth cycle also resulted in a very limited understanding of this stage (Rolfe et al. 2012). CbtA is a toxin gene of type IV TA system YeeU–CbtA which can inhibit the polymerization of bacterial cytoskeletal proteins (MreB and FtsZ) and influence the shape of cells (Masuda et al. 2012a; Tan et al. 2011). We found the cell morphology of the mutant strains (△pygA, △pygT and △pygAT) did not change significantly (Fig. S5).

In conclusion, we reported a functional AbiE family TA system PygAT that located on the genomic island PYG1 of P. yayanosii. This is the first archaeal example of a type IV TA system, in which both physiological functions of stabilizing plasmid and participating in HHP stress response were suggested.