The htrA Gene Plays an Important role During Yersinia Pseudotuberculosis Growth at High Temperature

htrA is a gene coding for the stress inducible HtrA protein, identied as a temperature stress response protein in several Gram positive and Gram negative bacteria. Growth rates at several temperatures (30ºC, 37ºC and 42ºC) were compared for Yersinia pseudotuberculosis YPIII wild strain and the isogenic mutant 1YPIII (htrA::Km), which was obtained by insertion of a kanamycin resistance cassette into the htrA gene. Y. pseudotuberculosis 1YPIII growth rates did not differ from the Y. pseudotuberculosis wild strain growth rates when cultivated at 30°C, which is consistent with a non-essential role for the HtrA protein at this temperature. However, 1YPIII mutant strain growth rate decreased by 18.73% at 37°C, and by 60.14% at 42°C, as compared to the Y. pseudotuberculosis YPIII wild strain growth rate. HtrA complementation in the strain 1YPIII/pAHTRA46 suppressed the differences in growth rates. Immunoblot analysis conrmed the absence of the HtrA protein in the 1YPIII mutant strain at any of the growth temperatures under analysis. In silico predictions were obtained for the three-dimensional structure of amino acid sequence belonging to HtrA from Y. pseudotuberculosis YPIII, Yersinia pestis CO92, using the protein data bank structure 1KY9:B from Escherichia coli, as template. The model's quality was found to be acceptable. Southern blot analysis shows a single htrA gene signal. These data indicate that the unique htrA gene in Y. pseudotuberculosis YPIII is required for the adaptive response of this species to high temperatures and although it is not a pathogenicity factor, it can be targeted by antibiotics.


Introduction
HtrA (high temperature requirement A, also called DegP) is a periplasmic heat-shock protein, with ATPindependent serine protease (Resto et al. 2000); (Phillips and Roop, 2001) and chaperone activities. The chaperone function is dominant at low temperatures (Spiess et al. 1999), whereas the protease activity becomes considerable between 32 °C and 42 °C (Skorko-Glonek et al. 1995); (Spiess et al. 1999). This protein can discriminate damaged proteins to be degraded, from partially unfolded proteins amenable to successful refolding and is a key factor in protein quality control in Gram negative bacteria (Pallen and Wren, 1997); (Spiess et al. 1999); (Young and Kim, 2005). Several proteins of the HtrA/DegP family have been identi ed in Yersinia enterocolitica (Li et al. 1996), Haemophilus in uenza (Craig et al. 2002), Bartonella henselae (Resto et al. 2000), Legionella pneumophila (Pedersen et al. 2001), Brucella abortus (Phillips and Roop, 2001), Salmonella typhimurium (Bäumler et al. 1994) and Helicobacter pylori The monomer HtrA/DegP from E. coli (PDBid 1KY9:B, the S210A mutant; Uniprot P0C0V0, DEGP_ECOLI, full CDS and used for the numbering given here) presents a signal peptide lost at secretion (amino acids 1 to 20) and three functional regions: i) a "Do" protease domain (residues 1-259), ii) core domain containing the catalytic triad of amino acids H105, D135 S210 required for enzymatic function (Clausen et al. 2002), in the carboxy terminus, the Arg-Gly-Asp (RGD) motif, including a conserved sequence signature GNSGG (residues 208-212) as catalytic domain, and iii) two PDZ domains at the carboxy terminal region. PDZ1 comprises residues 260-358 and PDZ2 resides at residues 359-448. These last PDZ domains are involved in protein-protein interactions and appear to determine target recognition (Table 1). In Y. enterocolítica, the gsrA gene (global stress requirement A) has been identi ed and sequenced. When the amino acid sequence of GsrA was compared to the HtrA of Escherichia coli and S. typhimurium, the signatures of the three characteristic HrtA domains were found. GsrA presents a potential signal peptide (residues 1 to 27), a catalytic domain with the sequence Gly-Asn-Ser-Gly-Gly (residues 239-243) ( Table 1) and near the carboxy terminus, the RGD motif (Yamamoto et al. 1996).
The mutant gsrA − of Y. enterocolítica O:8 showed increased sensitivity to oxidative stress, temperature stress and had a reduced virulence in murine yersiniosis bioassays (Li et al. 1996).
A Yersinia pestis htrA orthologous was cloned and sequenced its amino acid sequence showed high identity to Y. enterocolítica GsrA (95%), E. coli HtrA (86%) and S. typhimurium HtrA (87%). Y. pestis htrA − mutant, in contrast to the wild strain, failed to grow at 39 °C, but showed only a small increase in sensitivity to oxidative stress, and its virulence was only partially attenuated in the animal model. In addition, the protein expression pattern of the mutant differed from that of the wild strain, when grown at approach is used to gain insight on the possible role of the htrA orthologous gene product, and to explore the behaviour of an isogenic mutant of Y. pseudotuberculosis YP III strain.

Materials And Methods
Bacterial strain and plasmids The bacterial strains and plasmids used in this work are listed in Table 2. The mutation was corroborated by three Southern blot hibridization, as described by Sambrook (Sambrook and Russell, 2001), using as probes the kanamycin cassette, the htrA gene from pAHTRA46 plasmid and pDS132 plasmid. The DNA was digested with EcoRI for the rst and third assay (data not shown), and on the second assay with AvaI and ClaI.
Characterization of the mutant Y. pseudotuberculosis 1YPIII Using genomic DNA of Y. pseudotuberculosis and speci c primers (Gene ID: 6090393 NCBI data bank) for the Yersinia genus. 16S gene primers are: Forward primer: U16S38D: (5'-GCG GCA GCG GGA AGT AGT TTA-3') and Reverse primer: D16S47R: (5'-GAT TAA GCG TAT TAA ACT CAA CCC C-3'). and also having canonical geometrical and chemical features. According to this evaluation, the model for the wild type HrtA protein was rated as a good prediction, likely to represent correctly the threedimensional structure of the real protein in most of the model, whit the exception of the rst 40-45 amino acids at the N-terminus. As expected, the model reconstructed the essential amino acids at the serin protease active site and provided a plausible prediction for the active geometry of these residues.
The model for the HtrA::Km chimera was made using the well know I-TASSER server (Zhang, 2008), and the resulting prediction was also evaluated with the tools provided by SWISS MODEL workspace (Arnold et al. 2006). In contrast to the previous models the I-TASSER predictions for the HtrA::Km chimera where all dramatically different and the Rd.HMM failed to match the target sequence, therefore all predictions were rated as incorrect.

Results
Analysis of the sequence of 2315 bp.
The complete sequence of htrA gene was obtained, thoroughly analysed and not showed any mutation (Fig. 1) and aligned to E. coli HtrA (1KY9:B) and Y. pestis (GI:115349010) with CLUSTALW program, revealing 99% sequence identity ( Fig. 2A). The amino acid sequence of HtrA from Y. pseudotuberculosis was determined by in silico translation of the gene sequence and analysed with the Smart program (http://smart.embl.de/) to identify those sections corresponding to the putative characteristic domains of the HtrA family proteases, i.e.: i) signal peptide; ii) catalytic domain; and iii) two PDZ domains (Fig.2B). In silico models of HtrA from Y. pseudotuberculosis YPIII and Y. pestis CO92 were created using Swiss-Pdb-Viewer program taking as template 1KY9:B and threaded onto the crystallographic structure of HtrA from E. coli. As expected the threading revealed many coincidences between the different structures of HtrA protein in E. coli, Y. pestis and Y. pseudotuberculosis, but a signi cant difference in the loop formed by R-44 to G-79 in E. coli and the loop formed by P-73 to G-108 in Y. pestis and Y. pseudotuberculosis. This last loop corresponds to the LA loop (containing Q-linker) in the proteolytic domain (Fig. 2C) The insertion of kanamycin cassette into the htrA gene was described in methods, and generated the pAHTK99 plasmid. This plasmid containing the mutated htrA CDS (htrA::Km) was digested with Bam HI (4.2 and 3.5 kb fragments), Hind III (7.7 kb fragment) and Eco RI (3.9, 2.5 and 1.2 kb fragments) to check its restriction enzyme pattern (Fig.3A), sequenced, analysed and showed the inserted sequence (Fig.3B).
The mutant Y. pseudotuberculosis 1YPIII was complemented by transformation with the original pAHTRA46 plasmid containing the wild-type copy of the htrA gene. These transformed colonies were recovered and tested for the presence of the pAHTRA46 plasmid. The resulting con rmed transformants were designated Y. pseudotuberculosis 1YPIII/pAHTRA46. Characterization of the Y. pseudotuberculosis 1YPIII htrAmutant.
In order to prove that the clone of Y. pseudotuberculosis 1YPIII did not produce the native HrtA protein, a Western blot analysis was performed using total protein extracts of YPIII wild type and 1YPIII mutant strains using polyclonal antibodies anti HtrA of Staphylococcus aureus. In the total proteins extracts from cultures at 30°C, 37°C and 42°C no signal was observed in the protein extracts form the clone Y. pseudotuberculosis 1YPIII (Fig. 4A). This was surprising, because no signal was observed at all, neither at the chimeric HtrA::Km expected size, nor not at the wild-type HrtA protein expected size. In contrast, a band with the size expected for HtrA wild type protein was observed for the wild-type strain Y. pseudotuberculosis YPIII.
Y.pseudotuberculosis 1YPIII (htrA -) mutant strain was grown at different temperatures (30ºC, 37ºC y 42ºC). The mutant growth rates were comparable to those of the wild type strain at 30ºC, but were reduced at 37ºC and severely affected at 42 ºC (Fig. 4B). The data suggest an important role for HtrA protein in the growth of the bacterium at 37 and 42ºC. To con rm the participation of HtrA in the growth at 37 and 42ºC, the growth rates of the complemented strain 1YPIII/pHTRA46 were determined. The complemented mutant showed wild-type like growth rates at 37 and 42°C (Fig. 4B). In agreement, the protein band corresponding to the HtrA wild type protein size was detected in the crude extracts from the complemented mutant to a similar level observed for the wild type strain (data not shown).
Southern blot type hybridizations, using a 2.1 kb fragment as a probe, which contains the htrA gen of Y. pseudotuberculosis, only one signal was obtained (Fig. 5), also was ampli ed the htrA gene and htrA::Km in the YPIII, 1YPIII strains, shown 2.3 kb and 3.5 kb fragments respectively, another ampli cation using primers 16S speci c for Yersinia U16S38D and D16S47R, shown 412 bp fragment (Fig. 6).

Discussion
Y. pseudotuberculosis has been subclassi ed into 21 serotypes, which are considered as causative agents of several human and animal diseases (Skurnik, 1999). Y. pseudotuberculosis is the occasional etiologic agent of gastroenteritis, resulting in severe abdominal pain, fever and headache (Putzker et al. 2001).
Y. pseudotuberculosis HtrA protein shares the proteolytic and the two PDZ domains with its orthologues from E. coli (Wessler et al. 2017), Legionella fallonii, and Thermotoga maritime (Hansen and Hilgenfeld, 2013). A regulatory role for the PDZ domains has been demonstrated for HtrA and DegS (Krojer et al. 2010), but currently, there is no evidence for a similar regulation of the Y. pseudotuberculosis HtrA protein.
According to in silico modeling, the differences in the amino acid sequence between the Y. pseudotuberculosis HtrA protein and its orthologues from Y. pestis, Chlamydia trachomatis and E. coli (Gloeckl et al. 2012) cluster at the LA loop within the proteolytic domain. The alteration in the coding region introduced in the htrA::Km chimeric gene interrupts the gene at the beginning of the proteolytic domain and encodes an 836 amino acid long polypeptide which is coincident only in the rst 70 codons and has less than 25% similarity (at the amino acid level thereafter). Contrary to our expectations, the signal of this large chimeric HtrA::Km polypeptide, or any degradation fragment could not be detected with Abs anti HtrA of S. aureus. As expected, the band corresponding to the size of the HrtA wild-type protein was not present either. Either the chimeric transcript encoding HtrA::Km chimera is not transcribed, it is not translated, it is rapidly degraded at the mRNA or protein level, or all the antigenic determinants are lost due to the mutation. Instead, the HtrA protein is present in the Y. pseudotuberculosis wild type grown at 30, 37 and 42 °C, and under this conditions, it has been reported to have proteolytic activity (Lopes et al. 2009).
In contrast to the mutation by deletion of H. pylori In Salmonella enterica serovar Typhimurium, there is a protein homologue to E. coli HtrA called DegQ, in addition to classic S. enterica HtrA. The protease DegQ has identical activity to HtrA in vitro, but the null mutation of degQ gene does not attenuate the virulence of this strain, whereas the double null mutant in htrA and degQ present virulence attenuation (Farn and Roberts, 2004).
Two proteases of the type HtrA are found in S. aureus and their coding genes are called htrA 1 and htrA 2 .
In this case only HtrA 1 was found to have a function in the protection against thermal stress, but its protease activity is low or absent, and the role of HtrA 1 as a protease may be compensated by a different, yet unidenti ed protease (Rigoulay, 2005).
Mycobacterium leprae has the gene ML0176, which codes a predicted HtrA-like protease, conserved in other species of mycobacteria (Lopes et al. 2009) and in Mycobacterium tuberculosis. This gene, known as pepD, is directly regulated by the stress-responsive two-component signal transduction system MprAB and indirectly by the so called extracytoplasmic function (ECF) sigma factor, SigE (White et al. 2011).
In Campylobacter jejuni the HtrA has protease and chaperone activities and participates in the interaction between C. jejuni and mammalian host cells (Baek et al. 2011).
In Southern blot type hybridizations were performed, using a 2.1 kb fragment as a probe, which contains the htrA gen of Y. pseudotuberculosis, and only one signal was obtained. This last result suggests a unique gene encoding HtrA function in Y. pseudotuberculosis YPIII strain.
In contrast to Y. pseudotuberculosis YPIII wild type strain increase of the temperature strongly reduces the growth of the corresponding 1YPIII htrA − mutant. This result is in agreement with the data reported by Yamamoto (Yamamoto et al. 1996), for a mutant of the gsrA gene of Y. enterocolítica, and with the Williams' results (Williams et al. 2000) for Y. pestis htrA − mutant. In none of these last two reports did the mutant grow at temperatures higher than 39 °C, is agreement with the data reported by Zarzecka for H. pylori ).
In addition, HtrA from M. leprae displayed maximum proteolytic activity at temperatures above 40 °C (Lopes et al. 2009), HtrA from H. pylori showed temperature-dependent oligomer dissociation (Hoy et al. 2013) and DegP was important during high-temperature bacterial growth (Kim and Robert, 2012). In our case, hrtA − mutant showed virtually a total growth arrest after 4 h at 42ºC, and when compared to the wild strain at 12 h, a culture OD difference of 77.82% was observed. As already mentioned, in mutants of the htrA gene of bacteria such as Streptococcus pneumonie (Musa et al. 2004) and S. enterica (Lewis et al. 2009), the growth at 42 °C is attenuated, while the growth at lower temperatures was normal. Therefore, HrtA must be performing a role at 37ºC and 42ºC which is required for normal growth and that cannot be ful lled by any other chaperones and/or proteases in any of these bacterial species.
In a similar fashion, mutations in the htrA gene of S. pneumonie (Musa et al. 2004), Klebsiella pneumonie (Cortes et al. 2002) and Listeria monocytogenes 10403S (Wonderling et al. 2004) did have no effect in the growth rates at 37ºC, but did reduce the growth and/or survival rates at higher temperatures and in other bacteria as mentioned by Wessler (Wessler et al. 2017). In L. monocytogenes 10403S the sensitivity to high osmolarity was also increased in the mutant.
In the present work, there seem to be an increase in the expression of the HtrA band with the growth temperature in the Y. pseudotuberculosis YPIII wild type strain, (Fig. 3A). This result is in agreement with the increase in the rpoE gene with an increasing growth temperature. This gene encodes a transcriptional regulator which is related to the transcription of the htrA gene in E. coli and is likely to ful ll a similar role in Y. pseudotuberculosis. Yersinia pseudotuberculosis HtrA protein is important for cell growth at 37ºC, it is essential at 42ºC, and its molecular features suggest an unshared involvement in protein folding homeostasis at high temperatures.

Declarations
Ethics approval and consent to participate Not applicable.

Consent for publication
Not applicable.

Availability of data and materials
Please contact to the authors for all request.  amino acids long, with a predicted molecular mass of 49.8 kDa and a predicted signal peptide. The mature protein was predicted to be 47.1 kDa without signal peptide, and have a pI 8.76. C.-In silico models of HtrA from Y. pseudotuberculosis (brown) and 1KY9 S210A:B Escherichia coli (light blue). The models were prepared by threading using Swiss-Pdb-Viewer. Rd.HMM scores were 102.7, E-value 1.2x10-24, ratio score to sequence length 0.233, for Y. pestis prediction, and 110.6, with E-value 5.1 x 10-27 and score ratio to sequence length 0.251 for Y. pseudotuberculosis prediction. The HMMer alignment showed coincidence between the backbone and the threaded residues in all but a small 40 amino acid segment at the amino-terminus. The model is therefore imperfect in this side.