Type II restriction modification system in Ureaplasma parvum OMC-P162 strain

Ureaplasma parvum serovar 3 strain, OMC-P162, was isolated from the human placenta of a preterm delivery at 26 weeks’ gestation. In this study, we sequenced the complete genome of OMC-P162 and compared it with other serovar 3 strains isolated from patients with different clinical conditions. Ten unique genes in OMC-P162, five of which encoded for hypothetical proteins, were identified. Of these, genes UPV_229 and UPV_230 formed an operon whose open reading frames were predicted to code for a DNA methyltransferase and a hypothetical protein, respectively. DNA modification analysis of the OMC-P162 genome identified N4-methylcytosine (m4C) and N6-methyladenine (m6A), but not 5-methylocytosine (m5C). UPV230 recombinant protein displayed endonuclease activity and recognized the CATG sequence, resulting in a blunt cut between A and T. This restriction enzyme activity was identical to that of the cultivated OMC-P162 strain, suggesting that this restriction enzyme was naturally expressed in OMC-P162. We designated this enzyme as UpaP162. Treatment of pT7Blue plasmid with recombinant protein UPV229 completely blocked UpaP162 restriction enzyme activity. These results suggest that the UPV_229 and UPV_230 genes act as a type II restriction-modification system in Ureaplasma OMC-P162.

Bacteria can incorporate extracellular DNA into their genome, which may be accompanied by a characterization or virulence change. U. parvum serovar 3 type strain ATCC700970 was isolated from a patient with nongonococcal urethritis [15], and the clinical strain SV3F4 of U. parvum serovar 3 was isolated from a Japanese patient with history of an infectious abortion [12]. In our previous study, the complete genome of the SV3F4 clinical strain was sequenced, and its genome was found to be the smallest among all 14 other human pathogenic serovars of Ureaplasma spp. [12]. This genome size variation might be caused by gene insertions and deletions (indels). A good understanding of the genetic variations among bacterial strains could contribute toward developing future strategies for the diagnosis and treatment of bacterial infections. However, this has not been investigated in Ureaplasma spp. In our present study, we determined the whole genome sequence of OMC-P162 strain and identified a novel type II restriction modification system in OMC-P162.

Bacterial strains and cultures
All of the U. parvum clinical strains used in this study and ATCC700970 were cultured at 37˚C in Ureaplasma medium [2], consisting of 2.4% Mycoplasma broth base, 2.5% yeast extract, 5% horse serum, 0.04% urea, 0.01% L-cysteine hydrochloride, 0.001% phenol red, and 1,000 U/ml penicillin. The Escherichia coli JM109 (Takara Bio, Shiga, Japan) and C43 (DE3) (Lucigen, WI, USA) strains were used for DNA manipulation and recombinant protein expression, respectively. These two strains were grown at 37˚C in Luria-Bertani (LB) medium. This study was conducted in accordance with the Declaration of Helsinki and was approved by the Osaka Medical Center for Maternal and Child Health (former name of the Osaka Women's and Children's Hospital) Ethics Committee. Written informed consent was obtained from individuals or legally acceptable representatives when the patients were infants.

DNA extraction, plasmid construction and PCR mutagenesis
Ureaplasma genomic DNA was prepared by Genomic DNA Buffer Set (Qiagen, Hilden, Germany) and isolated by QIAGEN Genomic-tip 20/G (Qiagen, Hilden, Germany), which was then amplified by polymerase chain reaction (PCR) using Ex Taq (Takara Bio, Shiga, Japan) for plasmid construction. To prepare the PCR product for the pUPV-230 plasmid construction, the primer pair UP162-F15 (5ʹ-ctagctagcatgatggacaattatgttag-3ʹ) and UP162-R15 (5ʹccgctcgagatggcgtccattataaatatc-3ʹ) was used. The PCR protocol consisted of 2 min at 94˚C; 30 cycles of 15 s at 94˚C, 15 s at 45˚C, and 1 min at 72˚C; followed by 1 min at 72˚C.
To prepare a PCR product with a mutated TGA codon (Trp in Mycoplasma) to TGG (Trp) for pUPV-229 plasmid construction, overlap extension PCR was performed over two rounds of PCR. We designed three sets of primers. The outermost primers were methyl-F1 (5ʹ-gggcatatgaaaatatcgtacaataa-3ʹ) and methyl-R1 (5ʹ-ccgctcgagacataattgtccatcatttt-3ʹ), and the two sets of inner primers were methyl-F2 (5ʹ-atataatgagtgggttaaatc-3ʹ) and methyl-R2 (5ʹ-gatttaacccactcattatat-3ʹ), and methyl-F3 (5ʹ-ctgtagggtgggcattagaa-3ʹ) and methyl-R3 (5ʹ-ttctaatgcccaccctacag-3ʹ). Both sets of inner primers were sense and antisense oligonucleotides where the TGA codon was changed to TGG. During the first round of PCR, the products were amplified using the methyl-F1 and methyl-R2, methyl-F2 and methyl-R3, and methyl-F3 and methyl-R1 primers. The PCR protocol consisted of 2 min at 94˚C; 30 cycles of 15 s at 94˚C, 15 s at 45˚C, and 1 min at 72˚C; followed by 1 min at 72˚C. These three PCR products were purified and then mixed. The mixture was used as template in the second round of PCR using the outermost primers methy-F1 and methy-R1. The PCR protocol consisted of 2 min at 94˚C; 30 cycles of 15 s at 94˚C, 15 s at 50˚C, and 1 min at 72˚C; followed by 1 min at 72˚C. Both pUPV-229 and pUPV-230 plasmids were generated by using the NdeI and XhoI sites in the E. coli expression plasmid pET28a(+) (Novagen, MA, USA). To prepare substrates for UPV229 and UPV230, PCR products with a mutated CATG were constructed by three sets of specifically designed primers. The outermost primers were pT-F1 and pT-R3, and the two sets of inner primers were pT-F3 and pT-R7, and pT-F7 and pT-R2. Both sets of inner primers were sense and antisense oligonucleotides where the CATG motif had been mutated. During the first round of PCR, the products were amplified using pT-F1 and pT-R7, pT-F3 and pT-R2 and pT-F2 and pT-R3. PCR was performed as described above and generated product 8. pT-F1 and pT-R1 were used to amplify product 1, pT-F2 and pT-R2 for product 2, pT-F3 and pT-R3 for product 3, pT-F4 and pT-R4 for product 4, pT-F5 and pT-R5 for product 5, pT-F6 and pT-R6 for product 6, and pT-F4 and pT-R5 for product 7. S1 Table lists the sequences of the primers used.

Protein analysis
The proteins were analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and western blotting. Following the 10% SDS-PAGE, the samples were stained with Coomassie brilliant blue. For western blotting, the proteins in the SDS-PAGE gel were transferred onto a nylon sheet (GE Healthcare, Freiburg, Germany) and tagged with anti-His antibody (Cell Signaling Technology, MA, USA). Following treatment with horseradish peroxidase-conjugated secondary antibody (Rockland Immunochemicals, PA, USA), the tagged proteins were visualized via chemiluminescence (LAS-4000; Fujifilm, Tokyo, Japan). The function of proteins was predicted by InterPro analysis (http://www.ebi.ac.uk/interpro/index. html).

Next-generation sequencing, genome assembly, annotation, DNA modification analysis, and genome comparison
Genomic DNA from OMC-P162 was extracted and purified using a Qiagen genomic DNA kit (Qiagen, Hilden, Germany) according to the manufacturer's instructions. The insertion library of U. parvum was prepared and sequenced using a PacBio RS II single-molecule real-time sequencing (SMRT) system (Pacific Biosciences, CA, USA). We obtained 85,709 filtered reads from one SMRT Cell, with an average read length of 10,316 base pairs (bp). These were assembled into one contig of 732,031 bp using the Hierarchical Genome Assembly Process (HGAP) 2.0 (Pacific Biosciences) [19]. This contig was trimmed and circularized into a final complete genome of U. parvum OMC-P162 strain. A diagram of the overall genomic structure was generated using DNAPlotter 1.10 from Artemis 15.0 (Wellcome Sanger Institute, https://www. sanger.ac.uk/science/tools/artemis) [20]. The predicted CDSs, 6 rRNA genes, and 30 tRNA genes were annotated by a BLAST-based search (https://blast.ncbi.nlm.nih.gov/Blast.cgi) against the GenBank nucleotide and non-redundant databases. DNA modification analysis was carried out using RS Modification Detection 1 in SMRT Analysis v2.3.0 software and visualized with SMRT View software (PacBio,CA, USA; https://www.pacb.com/products-and-services/analytical-software/smrt-analysis/). Interpulse durations (IPDs) were measured as described previously [21]. The genome comparison was performed using wgVISTA (http:// genome.lbl.gov/cgi-bin/WGVistaInput). The prediction of promoters was searched on the website of BDGP (http://www.fruitfly.org/seq_tools/promoter.html). The genome sequence for the U. parvum strain OMC-P162 was deposited in the GenBank database (accession no. AP018561).

UpaP162 expression and purification
To purify the UpaP162 enzyme from OMC-P162, cells from a 1,000 ml culture were centrifuged at 26,000 × g for 60 min at 4˚C and gently suspended in 10 ml of ice-cold phosphatebuffered saline (PBS) consisting of 75 mM Na 3 PO 4 (pH 7.3) and 68 mM NaCl. The cells were recovered by centrifugation at 26,000 × g for 60 min at 4˚C, suspended in 1 ml PBS containing 1% Triton X-100 and 0.1 mM phenylmethylsulfonyl fluoride (PMSF) (Sigma-Aldrich, Steinheim, Germany), and incubated for 60 min on ice. The cell lysate was centrifuged at 26,000 × g for 60 min at 4˚C. Then, the supernatant was diluted 10-fold with buffer A (50 mM Tris-HCl, pH 7.5) and loaded on a Macro-Prep High S Support column (Bio-Rad, CA, USA) equilibrated with 1 ml of buffer A, washed with 2 ml of buffer A, and eluted with 1 ml of buffer A at various NaCl concentrations. The restriction enzyme-rich fraction was diluted with buffer A to a final concentration of 100 mM NaCl, and 1 ml was loaded on an equilibrated Heparin Sepharose 6 Fast Flow column (GE Healthcare, Uppsala, Sweden) filled with buffer B (10 mM Tris-HCl [pH 7.5], 0.1 mM EDTA, 10% glycerol, 1 mM dithiothreitol [DTT], and 100 mM NaCl). The column was washed with 2 ml of buffer B, and the fraction was eluted with 1 ml of buffer B at various NaCl concentrations. The restriction enzyme-rich fraction was recovered and dialyzed in buffer B and stored at −80˚C.
To induce the expression of the UPV230 protein, the E. coli strain C43 (DE3) harboring pUPV-230 was grown in 1,000 ml of culture medium at 37˚C until the optical density at 580 nm (OD 580 ) reached 0.5. Then, the expression was induced with 0.4 mM of isopropyl-β-D-1-thiogalactopyranoside (IPTG) and allowed to proceed at 25˚C overnight. The cells were pelleted by centrifugation at 6,000 × g for 30 min at 4˚C and then resuspended in 50 ml of icecold buffer C (50 mM Tris-HCl [pH 7.5] and 300 mM NaCl) containing 1 mM PMSF, 10 mM imidazole, and 10% glycerol. This was followed by 10 cycles of sonication on ice (one cycle consisted of 60 s of sonication and 60 s of resting) with a sonicator UD-200 (TOMY seiko, Tokyo, Japan) equipped with a microprobe at an output of 6. The suspension was centrifuged at 34,000 × g for 60 min at 4˚C, and the supernatant was passed through a 0.22 μm syringe filter. The filtered supernatant was loaded in a 6 ml equilibrated column of cOmplete His-Tag Purification Resin (Roche, Mannheim, Germany) with buffer C. The column was washed with 60 ml of buffer C containing 10 mM imidazole. Then, the protein was eluted with 6 ml of buffer C containing 500 mM imidazole. The eluted proteins were diluted with buffer A to a final concentration of 100 mM NaCl and loaded on 3 ml equilibrated Macro-Prep High S Support column by buffer D (50 mM Tris-HCl [pH 7.5] and 100 mM NaCl). The column was washed with 6 ml of buffer D and eluted with 3 ml of buffer D containing various concentrations of NaCl. The restriction enzyme-rich fraction was further purified using a sepharoseheparin-filled column as described above.
The purification of the UPV229 recombinant protein was performed as described for the purification of the UPV230 recombinant protein with some modifications. Briefly, the E. coli C43 (DE3) strain harboring the pUPV-229 plasmid was grown in 400 ml of LB medium at 37˚C until the OD 580 reached 0.5. The expression was induced by IPTG, and the cells were recovered and disrupted. Then, the supernatant was collected and filtered. The UPV229 protein was purified using a cOmplete His-Tag Purification Resin-filled column followed by a sepharose-heparin-filled column.

Determination of UPV230 restriction enzyme activity and its inhibitory activity of UPV229
To study the restriction enzyme activity in Ureaplasma cells, cultured OMC-P162 and C43 (DE3) strains in the mid-log growth phase were centrifuged at 17,800 × g for 20 min at 4˚C, gently washed once with ice-cold PBS, suspended in PBS containing 1% Triton X-100, and incubated for 10 min on ice. Then, the restriction enzyme activity was assessed by incubating 1 μl of cell lysate and 1 μg of closed circular pT7Blue plasmid (Novagen, MA, USA) in a total volume of 20 μl containing buffer L (100 mM Tris-HCl [pH 7.5], 100 mM MgCl 2 , and 10 mM DTT; Takara Bio Inc, Shiga, Japan) for 1 h at 37˚C. The reaction products were visualized by 1% agarose gel electrophoresis and ethidium bromide staining.
The endonuclease activity of the purified protein from the OMC-P162 and UPV_230 expressed in C43 (DE3) strains was assessed by incubating the purified protein and 1 μg of closed circular pT7Blue plasmid in a total volume of 20 μl containing L buffer at 37˚C for the various times indicated. To analyze the endonuclease activity of UPV230 on PCR products and the Ureaplasma genome, the reaction pH was adjusted to 9.2. NlaIII was purchased from New England Biolabs (Tokyo, Japan), and SphI was purchased from Takara Bio (Otsu, Shiga, Japan). The reaction products were visualized by 1% or 3% agarose gel electrophoresis and ethidium bromide staining.
The inhibitory activity of UPV229 against the endonuclease activity of UPV230 was assessed by incubating purified UPV229 recombinant protein and 1 μg of closed circular pT7Blue plasmid or 250 ng of PCR products in a total volume of 20 μl containing methylation buffer (20 mM Tris-HCl [pH 7.5], 0.2 mM DTT, 0.5 mM EDTA, 5% glycerol, and 4 μM S-adenosylmethionine) at 37˚C for 1 h. The plasmid or PCR products was recovered by phenol extraction and ethanol precipitation and digested with UPV230 protein. The DNA nucleotide sequence of the restriction site was determined using the BigDye Terminator v3.1 kit on an ABI 3130 Genetic Analyzer (Thermo Fisher Scientific, MA, USA).

OMC-P162 genome, its modification, and genome comparison
The length of the OMC-P162 genome was 732,031 bp with a G+C content of 25.47%. A total of 582 CDSs, 6 rRNA genes, and 30 tRNA genes were predicted (Fig 1A). The total number of post filtered subreads for DNA modification analysis was 920,879,592 bp. The estimated depth of coverage was 1,258, sufficient for m4C, m6A and m5C detection by IPD ratio. The recommended depths of coverage were 25, 25 and 250, respectively (https://www.pacb.com/wpcontent/uploads/2015/09/WP_Detecting_DNA_Base_Modifications_Using_SMRT_ Sequencing.pdf). The modified numbers of m6A nucleotides were 6,824, of m4C were 41,396 and of m5C were 0, while the numbers of unclassified modified nucleotides were estimated to be 225,981 (Fig 1B and S2, S3, S4 and S5 Tables). The comparison of the genomes of OMC-P162, ATCC700970, and SV3F4 identified four DNA fragments containing 14 gene insertions and seven accompanying gene deletions. Additionally, four DNA fragments containing 32 genes were deleted from the genome without gene insertion in the OMC-P162 genome. Three DNA fragments containing 19 genes were inserted into the genome and were accompanied by 11 gene deletions. Furthermore, five DNA fragments containing 36 genes were deleted from the genome without gene insertion in the SV3F4 genome (Fig 2 and S6  Table). Based on these data, we determined that the indel genes in the genome were clustered, and that 10 genes were unique to the OMC-P162 strain and were not present in ATCC700970 and SV3F4 strains (Fig 3A). Among these genes, UPV_229 and UPV_230 formed an operon structure. UPV_229 was predicted to code for a DNA (cytosine-5)-methyltransferase (Fig 3B) while UPV_230 might code for a restriction enzyme.

Expression of restriction enzymes in U. parvum serovar 3
To evaluate the restriction enzyme activity of the six clinical strains isolated in Japan and ATCC700970 strain, cells were recovered from cultures at the exponential phase, washed in Type II restriction modification system in U. parvum PBS, lysed with 1% Triton X-100, mixed with a pT7Blue plasmid and L buffer (Takara Bio, Shiga, Japan), and incubated at 37˚C for 1 h. The products were separated using 1% agarose gel electrophoresis (Fig 4). The result indicated that pT7Blue plasmid was cleaved by five of the seven strain lysates, suggesting that restriction enzymes were expressed in the U. parvum serovar 3 strains. Of these five strains, SN5 was isolated from an endotracheal aspirate of a preterm infant, F26 was isolated from a vaginal swab after an infectious abortion, OMC-P162 was isolated from the human placenta of a preterm delivery at 26 weeks' gestation, S104 was isolated from a vaginal swab at 8 weeks' gestation, and S29 was isolated from amniotic fluid at 32 weeks' gestation. The cleavage patterns from the cell lysates derived from F26, OMC-P162, S104, and S29 were consistent suggesting that these four strains might express the same restriction enzyme.

Identification of the restriction enzyme protein
To confirm that UPV_230 coded for a restriction enzyme, we expressed the UPV_230 fulllength sequence (S1 Fig) in E. coli as a His-tagged fusion protein. First, UPV230 protein was expressed in E. coli cells. SDS-PAGE of the cell lysates revealed that UPV230 His-tagged Type II restriction modification system in U. parvum protein was expressed in IPTG-induced cells but not in cells without IPTG induction (S2 Fig). Next, we purified this protein using a Ni-NTA-sepharose-filled columns and heparin-sepharose-filled columns as described in the Materials and Methods. The protein purification was confirmed by the anti-His antibody (Fig 5A). A final amount of 66 μg protein was obtained from 1,000 ml of culture.
Next, the endonuclease activity of the purified protein was examined. The pT7Blue plasmid was treated with the purified protein at a final concentration of 22.4-89.6 μg/ml at 37˚C for 1 h and separated by electrophoresis on 1% agarose. The DNA cleavage was observed, and the cleavage patterns were dependent on the protein concentration (Fig 5B). To identify the cleavage site, the individual cleaved bands were cut out and purified from the agarose gel and sequenced. Five sequences whose information can be used to identify the cleavage site were obtained (Fig 5B). Based on the pT7Blue plasmid sequence, we found that the ends of the fragments of UPV230 recombinant protein-digested sequences from 5ʹ to 3ʹwere CA (Fig 5C), suggesting that the enzyme cleavage site could be CATG, with a blunt cut between A and T. To confirm this prediction, pT7Blue plasmid treated by the enzyme was recovered, ligated to the EcoRV site of pT7Blue plasmid, and transformed into competent E. coli JM109 strain. Sixteen randomly selected positive colonies and plasmids were isolated, and the insertion DNA fragments were sequenced. The corresponding positions of these sequences on the pT7Blue plasmid were identified (Fig 5D). Four sequences corresponded to the plasmid position 726-1,119, three sequences to 1,155-1,233, one sequence to 1,233-1,243, four sequences to 1,243-1,734, one sequence to 1,734-2,454, and three sequences to 2,454-2,800. Based on the plasmid sequence information, we determined that recombinant UPV230 protein cleaved the DNA at the same site (CATG), suggesting that this enzyme is a type II restriction enzyme.

Purification of the UpaP162 restriction enzyme from the OMC-P162 strain
To establish whether the restriction enzyme activity of the UPV230 recombinant protein was identical to that of the cultivated OMC-P162, we purified the restriction enzyme from the OMC-P162 strain. The cells were recovered from a culture at the exponential growth phase, washed with PBS, and lysed with 1% Triton X-100. The supernatant was recovered, diluted 10-fold with PBS, and subjected to ion-exchange chromatography using Macro-Prep High S Support. The restriction enzyme was eluted using gradient concentrations of NaCl and the restriction enzyme activity of each fraction was analyzed (Fig 6A). Next, the restriction enzyme-rich fraction was subjected to affinity chromatography in a heparin-sepharose-filled column, and the restriction enzyme-rich fraction was obtained as described above (Fig 6B). Next, we examined the restriction enzyme activity of the cultivated OMC-P162 strain. The isolated restriction enzyme was mixed with 1 μg pT7Blue plasmid containing L buffer, incubated at 37˚C for various times, and the product was separated by 1% agarose gel electrophoresis. The results indicated that the pT7Blue plasmid was digested within the incubation time, and the cleavage pattern was similar between 2 and 3 h of incubation ( Fig 6C). Next, we identified the cleavage site of this enzyme. The cleavage pattern of this purified restriction enzyme from OMC-P162 strain was the same as that of the UPV230 recombinant protein. Briefly, enzymetreated pT7Blue plasmid was recovered, ligated to the EcoRV site of pT7Blue plasmid, and the insertion DNA fragments were sequenced. Finally, the cleavage site was identified as CATG, suggesting that the restriction enzyme activity of the UPV230 recombinant protein was identical to that of the cultivated OMC-P162 strain. We designated UPV_230 gene encoded Type II restriction modification system in U. parvum restriction enzyme as UpaP162. Next, we examined the digestion pattern of the pT7Blue plasmid after treatment with UpaP162 or NlaIII (a neoschizomer of UpaP162) at 37˚C for 1 h. The cleavage patterns by UpaP162 were similar to those of NlaIII; however, the pT7Blue plasmid was only partially digested by UpaP162 (Fig 7).  Table), and the bands used to identify the cleavage site were marked by arrows. (C) Sequence at the end of fragment of Enzyme-digested pT7Blue plasmid. The ends of sequenced DNA fragments in C were shown in red and their adjacent sequences in pT7Blue plasmid were shown in black, and primers used for sequence were indicated at left. Recognition site was underlined. (D) Enzyme-digested pT7Blue plasmid cloned into the EcoRV site of pT7Blue plasmid. pT7Blue plasmid was shown as linear solid line. Nine sites of CATG were marked by solid arrows and their A nucleotide positions in pT7Blue plasmid were shown above the arrows. Insertion DNA fragments were shown as dotted lines and obtained sequences numbers were shown. https://doi.org/10.1371/journal.pone.0205328.g005

The digestion efficiency of UpaP162 depends on the sequences around CATG motifs
UpaP162 partially digested the CATG site within the pT7Blue plasmid (Fig 7). To investigate the mechanism of this incomplete digestion, the bases flanking the CATG motifs in the pT7Blue plasmid were compared using various PCR products. Among the PCR products, column, washed by buffer and eluted by gradient concentrations of NaCl. Lanes contained pT7Blue plasmid as follows: lane1, pT7Blue plasmid (negative control). Lanes 2-4, Digestion products of pT7Blue plasmid by using fraction of enzyme rich fraction from (A), passed through and washed fraction. Lanes 5-8 Digestion products of pT7Blue plasmid by using elution factions containing 200 mM, 300 mM, 400 mM and 500 mM of NaCl. M, marker. (C) pT7Blue plasmid was digested with purified enzyme at 37˚C for indicated time, and products were separated by 1% agarose gel electrophoresis.
https://doi.org/10.1371/journal.pone.0205328.g006 Type II restriction modification system in U. parvum products 1, 4 and 5 were designed to contain one CATG motif, and 2, 3 and 6 were designed to contain two (Fig 8A). PCR products 1 to 6 were digested with UpaP162, and products 1 and 4, which contain one CATG motif, were only partially digested (Fig 8B). Both of these CATG motifs were flanked by A or T nucleotides. Next, we investigated whether dual CATG motifs in a PCR product influence UpaP162 restriction enzyme activity. PCR product 7 was amplified from positions 1,251 to 2,779 of pT7Blue, which includes two CATG motifs localized between 1,734 and 2,454. PCR product 8 was amplified from between 169 and 1,734 of pT7Blue, which includes two CATG motifs localized between 1,233 and 1,243. The CATG sites at 1,155, 1,233 and 1,243 were mutated by overlap PCR (Fig 8C). Neither of these dual CATG motifs enhanced UpaP162 activity, suggesting that UPaP162 is not a type II E restriction enzyme [22]. TCATGA and ACATGT motifs were not preferentially digested by UpaP162 activity.

Expression of UPV229 and its inhibitory activity against UpaP162 restriction enzyme
UPV_229 was predicted to code for a DNA (cytosine-5)-methyltransferase by gene annotation. To confirm the function of UPV229 protein, we expressed the UPV_229 full-length sequence (S1 Fig) in E. coli as a His-tagged fusion protein. The UPV229 recombinant protein was purified by a Ni-NTA-sepharose-filled columns and a heparin-sepharose-filled column (Fig 9A). Next, we examined the methyltransferase activity of this protein. The pT7Blue plasmid was treated with the recombinant protein UPV229 at a final concentration of 3×10 −6 to 3×10 −1 mg/ml, recovered, and digested by UpaP162 (Fig 9B). The pT7Blue plasmid that was treated Type II restriction modification system in U. parvum with 0.3 mg/ml of recombinant UPV229 protein completely blocked the UpaP162 restriction enzyme activity, suggesting that the UPV229 protein might methylated the DNA. Taken together, the UPV_230 and UPV_229 played the role of a type II restriction-modification system in the U. parvum OMC-P162 strain.

Sensitivity of Ureaplasma genomes to digestion by UpaP162
Given that the UPV229 gene is unique to the OMC-P162 strain and not present in the ATCC7 00970 and SV3F4 strains (Fig 3A), we isolated the genomes of OMC-P162, ATCC700970 and SV3F4, respectively, and each was digested by UpaP162 or NlaIII. The OMC-P162 genome was resistant to the both of UpaP162 and NlaIII, while the genomes of both ATCC700970 and SV3F4 were susceptible to both of the enzymes (Fig 10A). Additionally, UPV229 was protective from NlaIII. Indeed, pT7Blue plasmid that had been treated by UPV229 was also completely protected from NlaIII activity (Fig 10B). NlaIII is known to be blocked by m6A methylation, suggesting that UPV229 may be an m6A motif-methylation enzyme. Next, we used the SphI restriction enzyme to determine whether m6A is methylated by UPV229. SphI recognizes GCATGC and is inhibited by the presence of m6A. An unmethylated PCR product containing one GCATGC site was digested by both UpaP162 and SphI, while prior UPV229 treatment blocked the SphI activity (S3A Fig), suggesting the UPV229 methylated the m6A of GCATGC. Next, we examined whether the OMC-P162 genome is also resistant to SphI by treating OMC-P162 genomic DNA with UpaP162, SphI or EcoRI, respectively. As expected, the OMC-P162 genome was resistant to both UpaP162 and SphI but susceptible to EcoRI (S3B Fig).

UpaP162 is a novel restriction enzyme
This is the first report to analyze the restriction-modification (RM) system in Ureaplasma. In this study, UPV_229 and UPV_230 genes act as a type II restriction-modification system in U. parvum OMC-P162. UPV_230 codes for an endonuclease (UpaP162) while UPV_229 inhibits UPV230. Although a type II restriction enzyme in Ureaplasma has been previously reported, the pair genes that code for restriction and modification were not identified [23]. To date, over 3,000 restriction enzymes have been studied in detail [24], and UPV_230 did not exhibit any homology to those reported in our Blast search. Furthermore, there was no predicted endonuclease domain for UPV_230 using InterPro analysis, suggesting that UpaP162 might be a unique restriction enzyme. There are nine CATG sequences in the pT7Blue plasmid, the Type II restriction modification system in U. parvum longest distance between two CATG sequences being 767 bp. However, the longest fragment from the treated pT7Blue plasmid by UpaP162 was >1 kb (Fig 4C), suggesting that pT7Blue plasmid did not cut at all CATG sequences. In the present study, UpaP162 activity on CATG motifs was examined with PCR products, and we found that TCATGA and ACATGT sequences influence its activity. Characterization of Ureaplasma genome has shown that it contains high levels of A and T nucleotides (Fig 1). The GC content of OMC-P162 is only 25.47%. The Ureaplasma genome is thought to have reduced its genome size by so-called evolutionary regression, perhaps because an AT rich genome is more unstable than a GC rich structure because of the lesser number of the hydrogen bonds. To protect the AT rich Ureaplasma genome from invaders such as phages, it might be advantageous to be able to digest DNAs with contents that are higher in GC than AT. Although the pT7Blue plasmid could not be cut by the cell lysate derived from both SV3F4 and ATCC700970 strains (Fig 1), we cannot rule out the possibility that these two strains could express restriction enzyme activity depending on the buffer used.

Methylation motifs of UPV229
Homology searching predicted UPV229 to be an m5C methyltransferase; however, as Fig 1B shows, we could only find m4C and m6A methylation in the OMC-P162 genome. This could be explained by m5C methyltransferase being better documented than m4C and m6A. This observation, together with the restriction fragment length polymorphism analysis using NlaIII and SphI, suggests that the methyltransferase encoded by the UPV_229 gene is an m6A methylase. Sulfolobus acidocaldarius is an organism that does not have an m5C methylase but does contain an m6A [25]. The methylation status by UPV229 methylase could be unraveled at the molecular level by future mass-spectrometry or NMR analyses.

Pathogenicity of methylation protein
In this study, we identified different restriction modification activities among Ureaplasma spp. (Fig 4). It is known that the DNA methylation proteins play a role in the virulence of bacteria. In Streptococcus pneumoniae, hsds subunits of Spn556II type -I R-M system causes DNA inversion and excision, and generated variation of Hsd proteins. Hsd proteins alter methylation patterns and the resulting epigenetic switches dictate the regulation of capsule production and modify their virulence [26]. It was observed in a study involving variable surface proteins (Vsa) of Mycoplasma pulmonis, a rat pathogen that phylogenetically closely related to Ureaplasma spp., play a role in the virulence to escape from host immune system and survive in the host. The restriction and modification enzymes specifically expressed during animal infection are associated with changes in the production of Vsa proteins [27]. Although it is unclear if the deduced methylation protein coded for by UPV_229 influences virulence in Ureaplasma spp., there are 1,450 CATG motifs in the OMC-P162 genome and 33 of them are localized at the predicted promoter regions (S7 Table). This includes the promoter regions of 50S ribosomal proteins L21, L11, and L10, for 30S ribosomal protein S20 and for a potential virulence protein of type I restriction enzyme S protein. The virulence of UPV229 protein will be the subject of future study.