Activity of Vsr endonucleases encoded by Neisseria gonorrhoeae FA1090 is influenced by MutL and MutS proteins

Background The functioning of DNA repair systems is based on correct interactions between proteins involved in DNA repair. Very Short Patch (VSP) repair is a DNA repair system that corrects mismatches resulting from the deamination of 5-methylcytosine. The key enzyme in the VSP system is Vsr endonuclease, which can cleave mismatched DNA independently of accessory proteins. Until now, in vivo activity has only been shown for V.EcoKDcm - the only Vsr endonuclease in Escherichia coli. Additionally, the VSP system of E. coli is the only one for which interactions between proteins of the system have been demonstrated. Neisseria gonorrhoeae FA1090 is the first bacterium that we previously demonstrated to encode two active in vitro Vsr endonucleases: V.NgoAXIII and V.NgoAXIV. Results We elucidate the mutator phenotype of N. gonorrhoeae mutants with disrupted genes encoding V.NgoAXIII or V.NgoAXIV endonuclease. Furthermore, we investigate the interactions between gonococcal Vsr endonucleases and MutL and MutS proteins. The Vsr endonucleases physically interact with gonococcal MutL protein but not with MutS protein. In the presence of the MutL protein, the efficiency of DNA cleavage by both V.NgoAXIII and V.NgoAXIV endonucleases increases, resulting in a decrease in the amount of Vsr enzyme required to complete digestion of mismatched DNA. Both Vsr endonucleases are also stimulated in vitro by the MutL protein of E. coli. In turn, the gonococcal MutS protein hinders DNA cleavage by the Vsr endonucleases. However, this effect is overridden in the presence of MutL, and furthermore, the simultaneous presence of MutL and MutS causes an increase in the efficiency of DNA cleavage by the Vsr endonucleases compared to the reaction catalyzed by V.NgoAXIII or V.NgoAXIV alone. Conclusions For the first time, interactions between proteins of the DNA repair system encoded by N. gonorrhoeae that are responsible for the correction of mismatches resulting from the 5-methylcytosine deamination were identified. The increase in activity of Vsr endonucleases in the presence of MutL protein could allow for reduced synthesis of the Vsr endonucleases in cells, and the susceptibility of gonococcal Vsr endonucleases on MutL protein of E. coli implies a universal mechanism of Vsr stimulation by MutL protein. Electronic supplementary material The online version of this article (10.1186/s12866-018-1243-3) contains supplementary material, which is available to authorized users.


Background
Many bacteria use 5-methylcytosine (m5C), which is produced by DNA C 5 -methyltransferases (C5MTases), to distinguish between self and non-self DNA and to regulate gene expression [1]. However, being less stable than unmethylated cytosine, m5C can undergo spontaneous deamination to thymine, which cannot be removed by general repair mechanisms because this thymine is not recognized as erroneous in DNA. As a result, T:G mismatches arise and, if not corrected, lead to C→T transitions [2]. In Escherichia coli K-12 possessing one C5MTase, the mutagenic effects of spontaneous deamination of m5C are offset by the Very Short Patch (VSP) repair system. The key enzyme in the E. coli VSP system is a single Vsr endonuclease (V.EcoKDcm), which cleaves the mismatched DNA on the 5′-side of the mispaired thymine to allow its specific removal. The in vitro digestion of mismatched DNA by V.EcoKDcm endonuclease is affected by MutL and MutS proteins, which also take part in the post-replicative DNA mismatch repair path [3][4][5]. Furthermore, in E. coli K-12, mutations in mutL and mutS genes limit in vivo VSP repair; however, they do not completely impair it [5,6]. In vivo, DNA ligase and DNA polymerase I are also involved in the E. coli VSP system [7,8].
Apart from E. coli K-12 [9,10], the in vitro endonucleolytic activity of Vsr towards T:G mismatch has been shown for enzymes of Bacillus stearothermophilus H3 [11], Neisseria gonorrhoeae FA1090 [12] and Neisseria meningitidis [13]. Nonetheless, the E. coli K-12 Vsr endonuclease remains the only one for which the in vivo activity has been demonstrated and for which the physical and functional interactions with accessory proteins have been experimentally demonstrated [3,8,14].
N. gonorrhoeae FA1090 encodes five active C5MTases [15] that catalyze the creation of m5C. This bacterium is also the first and, thus far, the only microorganism shown to possess two in vitro active, phylogenetically-distant Vsr endonucleases, V.NgoAXIII and V.NgoAXIV. In contrast to the monospecific Vsr endonuclease of E. coli K-12, both gonococcal Vsr are multispecific. The former enzyme recognizes T:G mismatch in all nucleotide contexts, while the latter recognizes mismatches only in specific ones [12]. In addition to two Vsr endonucleases, N. gonorrhoeae FA1090 possesses one mutL and one mutS gene encoding proteins engaged in DNA repair [16]. However, it should be noted that unlike to E. coli K-12, our research model N. gonorrhoeae strain FA1090 does not encode MutH protein and Dam methyltransferase [17]. Moreover, many neisserial DNA repair pathways diverge from those of E. coli [18,19].
The aim of this work was to determine the in vivo activity of the gonococcal Vsr endonucleases and to examine the presence and potential effects of interactions between the gonococcal Vsr enzymes and accessory proteins -MutL and MutS. We show that inactivation of vsr genes results in a mutator phenotype and that the activities of both types of gonococcal Vsr endonucleases are affected by both MutL and MutS proteins.

Results
Inactivation of ngoAXIIIV, ngoAXIVV, mutL and mutS genes results in a mutator phenotype of N. gonorrhoeae The in vitro endonucleolytic activity of V.NgoAXIII and V.NgoAXIV protein towards DNA carrying a T:G mismatch [12] has been previously demonstrated. To study the in vivo activity of gonococcal Vsr endonucleases, gonococcal mutants were constructed by the gene replacement method [20]. To achieve this purpose, appropriate plasmids (Additional file 1: Table S2) were created and introduced into gonococcal cells to yield N. gonorrhoeae ngoAXIIIV::km, N. gonorrhoeae ngoAXIVV::km, N. gonorrhoeae mutL::km and N. gonorrhoeae mutS::km and complemented mutants.
Next, the frequency of spontaneous mutations in the obtained gonococcal mutant strains was assessed by screening for rifampicin [21][22][23] or nalidixic acid [16] resistance and compared with that of the parental wild-type strain FA1090. The strains lacking the individual Vsr endonucleases, MutL or MutS proteins showed elevated levels of spontaneous mutations upon rifampicin selection: the mutation frequency for the ngoAXIIIV::km strain was increased 2.97-fold, and those of the ngoAXIVV::km, mutL::km, mutS::km strains by 5.8-fold, 6.1-fold, and 5.7-fold, respectively, compared to the mutation frequency of the wild-type strain FA1090 (Fig. 1). The increase in mutation frequency was also observed upon nalidixic acid selection, by 3.5-fold, 4.65-fold, 5.2-fold and 6.7-fold, respectively, for gonococcal mutants with a disrupted ngoAXIIIV, ngoAXIVV, mutL or mutS gene (Additional file 1: Figure S1). Following mutant complementation by in cis insertion of intact copies of the respective gene in the mutant strain chromosome, the frequency of spontaneous mutations, both upon rifampicin and nalidixic acid selection, returned to the wild-type level ( For ngoAXIIIV::km and ngoAXIVV::km mutant strains, sequencing of the central conserved region of the rpoB gene of gonococcal rifampicin-resistant colonies confirmed the appearance of the C→T transitions. Additionally, the transversions G→T, C→A and A→T, but no deletion and insertion mutations, were identified (Additional file 1: Table S4) . We also noted that inactivation of the mutL gene in N. gonorrhoeae ngoAXIIIV::km or ngoAXIVV::km mutants increased the frequency of spontaneous mutations, respectively, 14-fold and 10-fold compared to the wild-type strain FA1090. In contrast, the frequency of spontaneous mutations in the mutants with disrupted mutS and ngoAXIIIV or mutS and ngoAXIVV genes was not significantly different compared to mutants with a single gene disruption (Fig. 1).
These observations confirm the involvement of gonococcal Vsr endonucleases in DNA repair in vivo and also may suggest an association between MutL and gonococcal Vsr endonucleases.

Gonococcal Vsr endonucleases physically interact with MutL but not with the MutS protein
The increase in mutation frequency in gonococcal mutants with disrupted ngoAXIIIV and mutL or ngoAXIVV and mutL genes led us to examine whether the MutL protein influences Vsr endonuclease activity via direct proteinprotein interactions.
First, the physical interactions between the V.NgoAXIII and V.NgoAXIV proteins and MutL and MutS proteins were studied using a bacterial two-hybrid system with lacZ as a reporter gene. The interactions between the investigated proteins were determined based on the value of the Miller Units of the β-galactosidase activity ratio calculated for each pair of proteins. If the two proteins interacted with each other, the ratio was less than 0.5 (Fig. 2).
The ratio obtained for the E. coli R721 strain harboring plasmids with the genes ngoAXIIIV and mutL Ngo or ngoAXIVV and mutL Ngo to that of the plasmid-less one was 0.18 (±0.019) and 0.23 (±0.02), respectively. This finding indicated a physical interaction between the V.NgoAXIII or V.NgoAXIV endonuclease and the gonococcal MutL protein (MutL Ngo ) (Fig. 2). The MutL Ngo protein also interacted with MutS protein, as demonstrated by the ratio of Miller Units for E. coli R721 carrying plasmids with the mutS and mutL Ngo genes to that of the plasmid-less strain of 0.39 (±0.01) (Fig. 2). In turn, the ratio of Miller Units for E. coli R721 harboring plasmids with the genes mutS and ngoAXIIIV or mutS and ngoAXIVV to that of the plasmid-less strain was 0.8 (±0.07) and 0.9 (±0.1), respectively (Fig. 2). This finding suggested that the MutS protein interacted neither with V.NgoAXIII nor V.NgoAXIV endonuclease. In the negative control, the ratio of Miller Units for E. coli R721 harboring a single plasmid to that of the plasmid-less strain was~1.0 (±0.046), whereas in the positive control it was~0.2 (±0.001).
The in vitro interactions between purified Vsr endonucleases and MutL Ngo protein, Vsr endonucleases and MutS protein and MutL Ngo and MutS proteins were studied by protein affinity chromatography. As is shown in Fig. 3, MutL Ngo protein was visible only in the elution but not in flow-through or wash fractions, after using resin with V.NgoAXIII or V.NgoAXIV endonucleases as ligand. This result indicated that the MutL Ngo protein was retained on the Vsr affinity column. Additionally, MutS protein was visible only in the elution but not in flow-through and wash fractions after using resin with MutL Ngo as ligand, supporting an interaction between these proteins.
In turn, MutS protein was in the flow-through and wash fractions when resin with V.NgoAXIII or V.NgoAXIV protein was used as ligand (Fig. 3), which indicated that MutS was not retained by endonucleases. In summary, these results support the conclusion that physical interactions occur between MutL Ngo protein and V.NgoAXIII, MutL Ngo and V.NgoAXIV and between MutL Ngo and MutS protein, but not between MutS protein and Vsr endonucleases.

The activity of both gonococcal Vsr endonucleases is stimulated by the gonococcal MutL protein
Previously, we have demonstrated the endonucleolytic activity of the gonococcal Vsr endonucleases in the absence of any accessory proteins. However, to complete digestion of the substrate DNA, a 10-fold excess of enzyme over DNA is required [12]. The occurrence of interactions Fig. 1 Spontaneous mutation frequency in N. gonorrhoeae mutants with disrupted ngoAXIIIV, ngoAXIVV, mutL Ngo or mutS genes. A sample of liquid culture (0.1 ml, 10 8 cells) of each strain was plated on GC agar supplemented with rifampicin or without antibiotics. After incubation at 37°C in 5% CO 2 for 48 h, the colony numbers were counted and frequency of spontaneous mutations causing rifampicin resistance determined. Asterisks indicate statistically significant differences compared to wild-type strain, which were calculated using the two-tailed heteroscedastic Student's t-test (p value < 0.05) between gonococcal Vsr endonucleases and the MutL Ngo protein prompted us to examine whether physical interactions influence the in vitro activity of V.NgoAXIII or V.NgoAXIV endonucleases. For this purpose, the first order rate constant (k st ) of the reactions catalyzed by the Vsr endonucleases in the presence of MutL Ngo protein and in the absence of MutS was estimated using a two-fold excess Vsr endonuclease over substrate DNA. Since it is assumed that Vsr enzymes act as monomers [24] and bacterial MutL proteins are functional homodimers [25,26], DNA cleavage by Vsr endonucleases in the presence of MutL Ngo was studied using a molar ratio purified proteins of 1:2 (Vsr:MutL Ngo ). Additionally, as ATP binding is required to instigate conformational changes in the MutL domain responsible for the regulation of interactions of MutL with other components in the cellular DNA repair machinery [25,26], each of the above reactions was carried out in two variants: in the presence of 1 mM ATP or without supplementation with ATP. The obtained data were compared to k st of the reaction catalyzed by Vsr endonucleases without MutL and ATP.
The V.NgoAXIII endonuclease alone most efficiently cleaves M.NgoA302P-sub substrate (GTTCGGT/ACCGGC) [12]. The efficiency of cleavage of this substrate by V.NgoAXIII in the presence of MutL Ngo and 1 mM ATP increased by 138.7% (k st value 0.0967) in comparison to the control reaction carried out by the endonuclease alone (k st = 0.0405) (Fig. 4). Since V.NgoAXIII is multispecific, the influence of the MutL Ngo protein on the DNA cleavage efficiency of all 11 substrates (Additional file 1: Table S3) recognized by Vsr was tested. As presented in Fig. 4 and Additional file 1: Table S5A, a similar effect as for M.NgoA302P-sub substrate of the MutL Ngo protein on DNA digestion by V.NgoAXIII was observed for all other tested DNAs. The efficiency of digestion of these DNAs by V.NgoAXIII increased by 62.6-107.1% (k st = 0.0298-0.0762) compared to the control (k st = 0.0160-0.0373).
In the absence of ATP, the level of stimulation of the V.NgoAXIII activity by MutL Ngo was reduced. Nonetheless, for all substrates, k st for the variants with MutL Ngo still remained higher compared with the reaction with endonuclease alone (Fig. 4, Additional file 1: Table S5A).
An analogous set of experiments was performed for the V.NgoAXIV endonuclease, and the obtained results were similar to those for V.NgoAXIII. For all nine substrates recognized by V.NgoAXIV, the highest Vsr cleavage rates were observed in the reactions in the presence of the MutL Ngo protein and 1 mM ATP. In such an experimental variant, cleavage efficiency for different substrates increased by 73.5-110% (k st = 0.0313-0.0924) compared to the control Fig. 2 The interactions between Vsr endonucleases, MutL and MutS proteins of N. gonorrhoeae FA1090. The study was carried out using the bacterial two-hybrid system with lacZ as a reporter gene as described in the Methods section. The β-galactosidase activity was estimated and presented as Miller Units, and the ratio between E. coli cells harboring plasmids with genes encoding proteins under investigation to plasmid-less E. coli was calculated. Black bars indicate the presence of interactions between the investigated proteins. The results are the mean values obtained from at least six independent repeats. Asterisks indicate statistically significant differences, calculated using the two-tailed heteroscedastic Student's t-test (p value < 0.05)  Table S5B). Without ATP, k st values ranged between 0.0282 and 0.0812, depending on the substrate DNA, but they were still higher than those obtained for the reaction with the endonuclease alone k st = 0.0165-0.0440 (Fig. 4, Additional file 1: Table S5B). Furthermore, as presented in Fig. 5, the increase in efficiency of DNA cleavage by V.NgoAXIII or V.NgoAXIV in the presence of the MutL Ngo protein indicated that the two-fold excess of Vsr endonuclease over substrate DNA was sufficient to complete DNA digestion instead of 10-fold in the control reaction (without accessory MutL Ngo protein). Such results were observed for all tested substrates for both V.NgoAXIII and V.NgoAXIV endonucleases ( Fig. 5 and Additional file 1: Figure S2).
The efficiency of the control reactions carried out by the gonococcal Vsr endonucleases in the presence of 1 mM ATP or BSA was the same as in the reaction with Vsr endonucleases alone. No DNA cleavage products were observed when the MutL Ngo alone was incubated with different substrate DNAs.
To conclude, under in vitro conditions, the gonococcal MutL protein significantly increased the efficiency of the reaction catalyzed by the gonococcal Vsr endonucleases, and the presence of ATP enhanced this effect.
The activity of gonococcal Vsr endonucleases is also stimulated by MutL protein of E. coli K-12 The high conservation of the N-terminal domain among proteins of the MutL superfamily and the 62% identity between N-terminal domains of MutL proteins from N. gonorrhoeae and E. coli led us to examine whether the in vitro activity of the gonococcal Vsr endonucleases could also be affected by the MutL protein of E. coli K-12 (MutL E.coli ). For this purpose, the efficiency and rate of reactions carried out by V.NgoAXIII or V.NgoAXIV in the presence of purified MutL E.coli at a molar ratio Vsr:MutL E.  Table S6). In the control reaction with MutL E.coli protein only (without Vsr endonucleases), no DNA cleavage products were observed.
In conclusion, the activity of the gonococcal V.NgoAXIII and V.NgoAXIV endonucleases may also be stimulated by MutL protein of E. coli K-12.

Gonococcal MutS protein decreases the efficiency of the hydrolytic reactions catalyzed by gonococcal Vsr endonucleases
The potential impact of the MutS protein on the efficiency of DNA cleavage catalyzed by the V.NgoAXIII and V.NgoAXIV endonucleases was estimated based on the k st values obtained for the reactions carried out in the presence of the MutS protein at a molar ratio of 1:2 and in the presence of 1 mM ATP. Similarly, as for the study concerning the influence of MutL proteins on Vsr endonuclease activity, the impact of MutS protein was also tested for all substrates recognized by V.NgoAXIII and V.NgoAXIV endonucleases.
The activity of the V.NgoAXIII endonuclease decreased in the presence of MutS and ATP for all tested substrates (Fig. 7, Additional file 1: Table S7A). This enzyme, in the presence of the MutS protein and 1 mM ATP, digested various substrate DNAs less efficiently, by 74.9-90.6% (k st = 0.00246-0.0091), compared to the cleavage reaction performed in the absence of MutS protein (k st = 0.0160-0.0440) (Fig. 7, Additional file 1: Table S7A).
In conclusion, the activity of both gonococcal Vsr endonucleases was significantly repressed by MutS protein towards substrates containing a T:G mismatch in all nucleotide sequence context.  In summary, the reaction rate of DNA cleavage by V.NgoAXIII and V.NgoAXIV was enhanced in the simultaneous presence of MutL, MutS and ATP compared to Vsr endonucleases alone. In this analysis, 0.3 μM of V.NgoAXIII bound 10% of the substrate DNA in the presence of ATP, and a further increase in the amount of endonucleases did not increase the amount of bound DNA (Fig. 9a). Additionally, 0.75 μM of V.NgoAXIV (the second gonococcal Vsr endonuclease) bound 15% of the substrate DNA in the presence of ATP (Fig. 9b). The same results were obtained for the reaction carried out in the absence of ATP (data not shown).
In turn, 0.75 μM of MutS protein bound 90% of the substrate DNA. However, the DNA-MutS protein complexes were formed only in the presence of ATP ( Fig. 9c  and d).
In all cases, we did not observe binding of the competitor DNA by gonococcal Vsr endonucleases or MutS protein, indicating the specificity of the DNA binding by the studied proteins.
In conclusion, both gonococcal Vsr endonucleases and MutS protein bound the same mismatched sequences; however, the efficiency of DNA binding by MutS protein was dramatically increased compared with Vsr endonucleases, but only in presence of ATP.

Discussion
DNA repair systems are crucial for all living organisms because they ensure genomic stability and integrity. The correct functioning of such systems relies on the interactions between proteins involved in DNA repair. In this work, for the first time, we demonstrated the mutator phenotype of gonococci with a disrupted ngoAXIIIV or ngoAXIVV gene, thereby indicating the in vivo engagement of V.NgoAXIII and V.NgoAXIV endonucleases in DNA repair. And by demonstration of the mutator phenotype of N. gonorrhoeae mutL::km and N. gonorrhoeae mutS::km mutants we confirmed the results of Criss et al. [16], who showed the involvement of MutL and MutS protein in DNA repair. However, for E. coli mutants with disrupted ecoKDcmV, mutS or mutL genes, strong mutator effects were noted [27,28], and we have observed a relatively modest increase in the frequency level of spontaneous mutation for ngoAXIIIV, ngoAXIVV, mutL and mutS mutants of N. gonorrhoeae. An analogous level of the frequency of spontaneous mutation has also been noted for mutL and mutS mutants of N. gonorrhoeae by Criss et al. [16] and for mutS and mutL mutants of N. meningitidis (Additional file 1: Figure S3) [18,29,30]. Differences in mutation frequency between N. gonorrhoeae and E. coli with disrupted vsr, mutL or mutS genes may result from differences between the DNA repair pathways of these microorganisms. We also cannot rule out the possibility that in gonococcus, a few DNA repair pathways are engaged in the prevention of spontaneous mutations.
Furthermore, we investigated the interactions between the Vsr endonucleases, MutL and MutS proteins of N. gonorrhoeae FA1090. Thus, N. gonorrhoeae FA1090 is the first β-proteobacterium and the second of all microorganisms for which the physical and functional interactions between a Vsr endonuclease and MutL and MutS proteins have been demonstrated. As mentioned above, until now, the influence of MutL and MutS proteins on Vsr endonuclease activity has only been investigated for E. coli proteins. However, the results for E. coli proteins indicate a discrepancy concerning the involvement of the MutS protein in in vitro DNA cleavage. According to Monastiriakos et al. [3], MutL alone stimulates Vsr endonuclease activity, but as is demonstrated in [4], the activity of V.EcoKDcm endonuclease can be stimulated by MutL protein only when both MutS and ATP are also present in the reaction. Here, we have shown that, although the gonococcal Vsr endonucleases digested DNA in the absence of any accessory proteins [12], the efficiency of in vitro DNA cleavage catalyzed by V.NgoAXIII and V.NgoAXIV was significantly increased in the presence of MutL Ngo protein. This stimulatory effect was demonstrated for all tested DNA substrates; thus, we can conclude that the MutL Ngo protein alone is sufficient to stimulate DNA cleavage by the gonococcal Vsr endonucleases, irrespective of the MutS protein and the nucleotide context of the T:G mismatch. As a consequence of the stimulation of V.NgoAXIII and V.NgoAXIV activity by MutL proteins, only a double molar excess of Vsr endonuclease over the substrate DNA was sufficient to completely digest the substrate DNA instead of the 10-fold excess required in the absence of MutL [12]. Therefore, MutL could allow for reduced expression and synthesis of Vsr endonucleases in cells. This finding is particularly important since N. gonorrhoeae FA1090 encodes two Vsr endonucleases and five active C5MTases that catalyze the creation of m5C, representing putative hot spots and a source of T:G mismatches. Lowering the amount of Vsr endonuclease required for complete digestion of the substrate DNA may allow efficient removal of T:G mismatches without the requirement for increased synthesis of Vsr. Although both gonococcal Vsr endonucleases are stimulated by MutL proteins, there is a difference in the level of stimulation, which could result from the phylogenic distance of the V.NgoAXIII and V.NgoAXIV endonucleases, which belong to different groups [12]. V.NgoAXIV endonuclease, which belongs to the same group as the V.EcoKDcm endonuclease, is less susceptible to stimulation by MutL proteins. In turn, V.NgoAXIII endonucleases belonging to the second group of Vsr enzymes are more sensitive to the presence of MutL proteins. This difference is especially important because V.NgoAXIII recognizes the T:G mismatch in all nucleotide contexts, thus providing more comprehensive protection from the effects of m5C deamination; however, its lack of specificity may result in the need for more stringent regulation of its activity.
For the first time, we also demonstrated that the endonucleolytic activity of the Vsr endonucleases is stimulated not only by the MutL protein of the same organism, but also by that encoded by a distantly-related bacterial species. The capability of E. coli MutL protein to stimulate gonococcal Vsr endonucleases may result from the high similarity of amino acid sequences between N. gonorrhoeae and E. coli MutL proteins and the high evolutionary conservation of MutL proteins. Such conservation of the N-terminal domain has been observed for MutL proteins from different bacterial species [31] and may indicate a universal mechanism of Vsr activity regulation by MutL protein that involves its N-terminal domain. Indeed, in the E. coli MutL protein, the N-terminal domain is responsible for physical interactions with V.EcoKDcm endonuclease [4].
MutS protein also influences the efficiency of the reaction catalyzed by V.NgoAXIII and V.NgoAXIV. As we demonstrated in vitro, the presence of MutS hindered or completely inhibited DNA cleavage by gonococcal Vsr endonucleases. Such an effect was not observed for E. coli Vsr activity when the purified MutS protein was added in vitro to V.EcoKDcm, even at five-fold excess of MutS to V.EcoKDcm [4]. As we demonstrated, the inhibitory effect of the gonococcal MutS on the efficiency of the reaction catalyzed by Vsr endonucleases did not result from physical interactions between the endonucleases and the MutS protein, as these proteins did not interact with each other. Furthermore, analysis of the structure and spatial architecture of the domains responsible for binding T:G mismatches in the MutS and Vsr proteins indicates that simultaneous binding of mismatches by these proteins is impossible [5]. However, it may result from competition between Vsr endonuclease and MutS protein and from the higher affinity of MutS protein to the mismatch site than the V.NgoAXIII and V.NgoAXIV endonucleases. Thus, MutS can physically block the site for Vsr binding. The observed inhibitory effect of MutS on DNA digestion by Vsr is offset in the presence of MutL and ATP. The different levels of Vsr endonuclease activity exerted by gonococcal MutL and MutS proteins observed in vitro should be further investigated in vivo in a physiological context.

Conclusions
The gonococcal Vsr endonucleases are engaged in DNA repair in vivo, and the activity of gonococcal endonucleases is influenced by the accessory proteins MutL and MutS. Thus, these proteins may be involved in the correct functioning of the VSP system by impacting Vsr endonuclease activity. Moreover, the influence of MutL protein of E. coli on the activity of the gonococcal Vsr endonucleases implies a universal mechanism regulation of DNA cleavage by the Vsr endonucleases.
Prior to each experiment, the piliation and Opa phenotypes of the N. gonorrhoeae strains employed were determined by microscopic examination of colony morphology. Only piliated gonococci were utilized in experiments and all strains exhibited the same piliation state.

Construction of N. gonorrhoeae mutants and complemented strains
N. gonorrhoeae mutants and complemented strains were achieved by the gene replacement method [20]. To achieve this purpose, N. gonorrhoeae FA1090 cells were transformed with the plasmids described below.
All plasmids were prepared in E. coli ER1821 cells. First, the gonococcal ngoAXIVV (flanked by a sequence of~500 bp), ngoAXIIIV, mutL and mutS genes were amplified from the chromosomal DNA of N. gonorrhoeae FA1090. The amplified DNA fragments were digested with appropriate restriction endonucleases and cloned into vector DNAs cleaved with the same enzymes (Additional file 1: Table S2). The ngoAXIIIV gene was cloned into pBluescript KS II (+), the ngoAXIVV gene into pUC19 and the mutL and mutS genes into the vector pMPMA4Ω [34]. This procedure generated the following plasmids: pBluescript KS II (+)::ngoAXIIIV, pUC19::ngoAXIVV, pMPMA4Ω::mutL and pMPMA4Ω:: mutS (Additional file 1: Table S2).
Next, a kanamycin resistance (Km R ) or chloramphenicol (Cm R ) gene cassette was inserted into the cloned genes. The Km R cassette was excised from plasmid pDIY-km [35] using either BamHI or SmaI. The Km R cassette obtained by excision with BamHI was used for inactivation of the ngoAXIIIV, mutL, and mutS genes, while that with SmaI used to inactivate the ngoAXIVV gene. Thus, pBluescript KS II (+)::ngoAXIIIV + km, pUC19::ngoAXIVV + km, pMPMA4Ω::mutL + km and pMPMA4Ω::mutS + km were obtained. The Cm R cassette was excised from pKRP10 plasmid [36] by SmaI and used to generate the pMPMA4Ω:: mutL + cm and pMPMA4Ω::mutS + cm plasmid constructs (Additional file 1: Table S2). Where necessary, the cohesive termini of the DNA fragments used to construct the suicide vectors were modified by treatment with Klenow Fragment of DNA polymerase I or T4 DNA polymerase prior to ligation.
Transformants were selected by their resistance to kanamycin or chloramphenicol, and the mutants and occurrence of double cross-over homologous recombinants were verified by PCR and Southern blot analysis. The presence of only one PCR product that corresponded in size to the sum of the mutated gene (414,423,1977 and 2595 bp for ngoAXIIIV, ngoAXIVV, mutL and mutS gene, respectively) and the Km R cassette gene (950 bp) or Cm R (842 bp) was treated as proof of successful mutagenesis. The amplified fragments were also fully sequenced to further verify the desired insertions. In addition, Southern blotting was carried out using non-radioactively labeled antibiotic resistance gene cassette (Ab R ) probes (DIG-High Prime, Roche) to confirm the presence of single copy insertions.
For complementation of the N. gonorrhoeae mutants, wild-type copies of the appropriate genes were inserted into the chromosomes of the mutants in the intergenic region between the ngo0275 and ngo0274 genes [37]. This intergenic region plus~500 bp flanking sequences was amplified from the N. gonorrhoeae FA1090 chromosome using Phusion High-Fidelity DNA Polymerase (Thermo Scientific) with primers iga and trpB (Additional file 1: Table S1). The PCR amplicon was digested with SalI and EcoRI, cloned into vector pMPMA4Ω cleaved with the same enzymes. The resulting construct, pMPMA4Ω:: IgaTrpB, was then used as template in ExSite PCR with the primers exit1left and exit1right. Simultaneously, the Opa promoter was amplified from the chromosome of the N. gonorrhoeae FA1090 strain by PCR using Phusion High-Fidelity DNA Polymerase (Thermo Scientific) with primers opaleft and oparight (Additional file 1: Table S1). Both amplicons were digested with HindIII and Mph1103I and then ligated to generate the construct pMPMA4Ω::IgaTrpBOpa.
This plasmid was then digested with HindIII and ligated to a Cm R cassette with HindIII cohesive termini excised from pKRP10 [36] to generate plasmid pMPMA4Ω:: IgaTrpBOpaCm. This construct was used as template in ExSite PCR with primers smatrpB and Nheiga. The amplicon was then digested with SmaI and NheI and ligated to DNA fragments representing the ngoAXIVV, ngoAXIIIV, mutL or mutS genes amplified from the chromosome of N. gonorrhoeae FA1090 with the appropriate primers (Additional file 1: Table S1) and digested with the same restriction enzymes. These ligations produced constructs pMPMA4Ω::IgaTrpBOpaCmVNgoAXIIIV, pMPMA4Ω:: IgaTrpBOpaCmVNgoAXIV, pMPMA4Ω::IgaTrpBOpaCm MutL and pMPMA4Ω::IgaTrpBOpaCmMutS, in which the respective gonococcal genes were fused to the constitutive Opa promoter.
These plasmids were linearized by digestion with EcoRV (ngoAXIIIV, mutL or mutS gene constructs) or BamHI (ngoAXIVV gene construct) and used to transform piliated N. gonorrhoeae mutant strains according to the method reported by Dillard [20]. In each case, the plasmid used to transform the mutant strain carried the gene required to complement that particular mutation. Transformants isolated by selection with kanamycin and chloramphenicol were then examined by PCR. In cis complementation of the N. gonorrhoeae mutants was confirmed by the presence of two PCR products: one large, corresponding in size to the sum of the ngoAXIIIV, ngoAXIVV, mutL or mutS gene and the Ab R cassette, and the other small, corresponding in size to the wild-type gene alone. The same primers that were used for amplification of particular genes were applied (Additional file 1: Table S1). Insertion of a functional copy of the gene of interest into the gonococcal chromosome between the ngo0275 and ngo0274 genes was also verified by PCR using primers iga and trpB. A 1134-bp fragment was amplified from the untransformed N. gonorrhoeae mutant strains, but following complementation, the size of the amplicon was increased by the combined size of the Cm R cassette (842 bp), Opa promoter (243 bp) and the gene of interest, i.e., 414, 423, 1977 and 2595 bp for the ngoAXIIIV, ngoAXIVV, mutL and mutS gene, respectively.
Determination of the spontaneous mutation frequency and nature of the mutations N. gonorrhoeae wild-type, mutant and complemented mutant strains were cultivated on GC agar for 16 h. The cells were then harvested from the plates, suspended in GC broth (10 9 cells/ml of each strain) and 0.1 ml aliquots of each strain were plated on GC agar base supplemented with 0.12 μg/ml rifampicin or 1 μg/ml nalidixic acid. In parallel, the gonococcal cells were diluted 10 − 6 , and 0.1 ml aliquots were plated on GC agar base without antibiotics to determine the total viable cell number. All plates were incubated at 37°C in 5% CO 2 for 48 h, and then the colony numbers were counted. The proportion of rifampicin-resistant cells was determined for the wild-type strain and gonococcal mutants to assess whether disruption of the ngoAXIIIV or ngoAXIVV genes resulted in an increase in the spontaneous mutation frequency. The mean frequency of spontaneous mutations was determined from 6 independent experiments. The significance of any differences was calculated using the two-tailed heteroscedastic Student's t-test (p < 0.05).
To determine the nature of the mutations, gonococcal wild-type and mutants were cultivated for determination of the spontaneous mutation frequency and plated on GC agar supplemented with 0.12 μg/ml rifampicin. Next, the central conserved region of the rpoB gene [21][22][23] was PCR-amplified from rifampicin-resistant colonies using Phusion High-Fidelity DNA Polymerase. The amplified fragments were sequenced using primers rpoBNGF and rpoBNGR, and the nucleotide sequences were compared to the "wild-type" rpoB gene sequence. Thirty rifampicin-resistant colonies of each N. gonorrhoeae strain were analyzed in this manner.

Bacterial two-hybrid system
The bacterial two-hybrid system is based on the repression of β-galactosidase activity, as developed and described by Di Lallo et al. [32]. In this system, a chimeric operator is recognized by a hybrid repressor formed by two chimeric monomers, one domain of which is composed of heterologous proteins. Only if these proteins efficiently dimerize in vivo is the functional repressor formed, binding the chimeric operator and terminating the synthesis of a downstream lacZ gene.
First, mutL Ngo , mutS, ngoAXIIIV or ngoAXIVV genes were amplified by PCR using the chromosomal DNA of N. gonorrhoeae FA1090 as template. Primer names and sequences are listed in Additional file 1: Table S1. These amplicons were cloned into the SalI-BamHI sites of vectors pC434 and pC22 [32] using a standard cloning procedure. Genes ngoAXIIIV and ngoAXIVV were cloned into pC22, mutL Ngo into pC434, and mutS into the pC434 and pC22 vectors.
Subsequently, E. coli R721 cells were transformed with pC434-and pC22-derived plasmids (Additional file 1: Table S2). In this strain, there was a chimeric operator upstream of the lacZ gene encoding β-galactosidase, formed by two hemi-sites of P22 and a 434 phage operator. This operator was recognized and bound only by a hybrid repressor, consisting of two chimeric monomers. One of them contained the N-terminal domain of the repressor of phage 434, and the other contained that of phage P22. These domains were fused with the sequence of the heterologous proteins, the interaction abilities of which were investigated. If the studied proteins physically interacted with one another, a functional chimeric repressor was formed, and the N-terminal part of the hybrid repressor bound to the sequence of the chimeric operator, leading to the repression of β-galactosidase synthesis. If the tested proteins did not interact with one another, a functional repressor was not formed and the β-galactosidase gene was expressed.
To examine the interactions between gonococcal Vsr endonucleases, MutL and MutS proteins, the following combinations of two plasmids (Additional file 1: Table S2) were used for co-transformation: (i) pAK24 and pAK26 to investigate the interactions between V.NgoAXIII and MutL Ngo , (ii) pAK25 and pAK26 to study the interactions between V.NgoAXIV and MutL Ngo , (iii) pAK24 and pAK28 to investigate the interactions between V.NgoAXIII and MutS, (iv) pAK25 and pAK28 to study the interactions between V.NgoAXIV and MutS, and (v) pAK26 and pAK27 to analyze the interactions between MutL Ngo and MutS proteins.
Next, six colonies of fresh transformants were used for inoculation of 10 ml LB supplemented with 0.1 mM IPTG and the appropriate antibiotics, and grown at 37°C until the optical density at 600 nm (OD 600 ) of the culture reached 0.3-0.4. Subsequently, a β-galactosidase assay was performed to analyze protein-protein interactions, and the ratio of β-galactosidase activity for the E. coli R721 strain with both plasmids carrying genes for the two investigated proteins to that of the plasmid-less R721 strain was estimated. All experiments were repeated at least six times, and their results were averaged. The significance of any differences from six independent experiments was calculated using the two-tailed heteroscedastic Student's t-test (p < 0.05).

Gene cloning and protein expression and purification
Cloning of the ngoAXIIIV and ngoAXIVV genes into pET28a(+) (Novagen) and pQE-30 (Qiagen) vectors, respectively, has been previously described in [12], along with the protocol for expression and purification of the V.NgoAXIII and V.NgoAXIV endonucleases by metal affinity chromatography.
The genes for MutL Ngo , MutL E.coli and MutS proteins were cloned in such a way that the resulting proteins contained a vector-encoded amino terminal His-Tag.

Protein affinity chromatography
Protein affinity chromatography was carried out analogously to [39][40][41] with slight modifications. To achieve this purpose, purified V.NgoAXIII (250 μg), V.NgoAXIV (250 μg) or MutL Ngo (250 μg) proteins were coupled to Affi-Gel ® 10 (ester activated agarose) according to the supplier (Bio-Rad) in 0.1 M NaHCO 3 , pH 8.9 (V.NgoAXIII or V.NgoAXIV) or 0.1 M MOPS pH 7.5, 50 mM CaCl 2 (MutL Ngo ). Then, residual active groups of the resin were blocked by incubating the gel with 1 M ethanolamine, pH 8.9. After transfer to a chromatography column, excess soluble protein was removed by washing the gel with buffer A (25 mM Tris pH 7.5, 1 M NaCl), and the column was then equilibrated with buffer B (25 mM Tris pH 7.5, 50 mM NaCl). Then: (i) MutL Ngo protein (250 μg in buffer B with 2 mM ATP and 5 mM MgCl 2 ) was applied to the V.NgoAXIII, V.NgoAXIV or control columns; (ii) MutS protein (250 μg in buffer B with 2 mM ATP and 5 mM MgCl 2 ) was applied to MutL Ngo , V.NgoAXIII, V.NgoAXIV or control columns, which were subsequently washed with buffer B. Proteins were eluted with buffer C (25 mM Tris pH 7.5, 250 mM NaCl). Proteins were separated by electrophoresis on SDS-polyacrylamide gels and blotted onto an Immun-Blot™ PVDF membrane (Bio-Rad). Since the proteins contained a His-Tag, Western blot analysis was carried out using a 6×-His Tag Monoclonal Antibody (ThermoFisher) and Goat Anti-Mouse IgG, Alkaline Phosphatase conjugated (ThermoFisher) antibodies according to the manufacturer's recommendations. As a control resin, BSA-coupled gel was used.

Preparation of substrate DNA for Vsr activity assessment
Preparation of substrate DNA for Vsr endonuclease activity and specificity study was performed as described and presented in [12,13]. Construction of DNA substrates containing two T:G mismatches and the principle of an assay to demonstrate the activity of Vsr endonuclease are presented on Additional file 1: Figure S5. Briefly, two oligonucleotides (Sigma-Aldrich), 5′ATATTCAAACTGG CGCCGAGCGTATGCCGCATGACCTTTCCCATCTTG GCTTCCTTGCTGGTCAGATTGGTCGTCTTATTACC ATTTCAACTACTNNNNNNGATATCNNNNNNCGAC TCC 3′ (120-mer) and 5′AGCAAGGCCACGACGCAAT GGAGAAAGACGGAGAGCGCCAACGGCGTCCATCT CGAAGGAGTCGNNNNNNGATATCNNNNNNAGTA GTTG 3′ (90-mer), were used to prepare the substrate DNA. The DNA sequence of these oligonucleotides was a derivative of the M13 bacteriophage DNA. Each oligonucleotide contained a 33 nt-long complementary region (underlined). NNNNNN indicates the sequence recognized by the studied Vsr endonucleases. Additionally, the sequence in which thymine is mismatched to guanine is indicated in italics. These oligonucleotides were mixed and hybridized at 95°C for 5 min in 1×SSC (15 mM sodium citrate pH 7.2, 150 mM NaCl) and then slowly cooled to room temperature. Subsequently, single-stranded ends were filled using Polymerase I Klenow Fragment (Thermo Scientific) in the presence of dNTPs, under the conditions recommended by the manufacturer. Each substrate DNA contained two copies of the same sequence carrying a T:G mismatch (Additional file 1: Table S3). Additionally, substrate DNA containing a T:T mismatch was created. The control substrate DNAs did not contain any mismatches, and the AGCGCC sequence was substituted for the region indicted by NNNNNN. All substrates used are listed in Additional file 1: Table S3.

Analysis of the activity of Vsr endonucleases in the presence of MutL and MutS proteins
The in vitro endonucleolytic activity of Vsr endonucleases (V.NgoAXIII, V.NgoAXIV) in the presence of MutL Ngo or MutL E.coli and MutS proteins was assayed by incubation of 0.15 μM substrate DNA with 0.3 μM of a Vsr endonuclease, in a final volume of 20 μl containing either 10 mM Tris-HCl (pH 7.5), 10 mM MgCl 2 and 0.1 mg/ml BSA (for V.NgoAXIII), or 66 mM Tris-acetate (pH 7.9 at 37°C), 20 mM magnesium acetate, 132 mM potassium acetate and 0.2 mg/ml BSA (for V.NgoAXIV), at 37°C for 60 min. The effect of MutL and MutS proteins on the activity of Vsr endonucleases was studied using 0.3-0.6 μM of the purified MutL and MutS proteins. Additionally, when needed, 1 mM ATP was added. Control reactions were carried out with (i) a Vsr endonuclease and BSA, with or without 1 mM ATP; (ii) BSA only, with or without 1 mM ATP; or (iii) a MutL Ngo protein with or without 1 mM ATP. The cleavage products were analyzed using 10% neutral polyacrylamide gels, and DNA was visualized by ethidium bromide staining. The appearance of two product bands (~110 and~80 bp) after treatment of the substrate DNA (~190 bp) with Vsr endonuclease supported its hydrolytic activity. The amount of DNA in substrate and product bands was determined using a Gel/ChemiDoc (Bio-Rad) analyzer, quantified with the accompanying software package (Quantity One) (Bio-Rad), and expressed as a fraction of the total intensity of DNA as visualized by ethidium bromide staining. Then, the electrophoresis profiles were used to quantify the efficiency of cleavage of the substrate DNA by Vsr endonucleases. The efficiency of the reaction was calculated based on analysis of the rate constants determined by fitting the data to a first-order rate equation: % product t = 100[1-exp(−k st )] (where t is the time and k st is the first-order rate constant) [3,4,12,42,43] using Origin 6.1 software (OriginLab, USA). The significance of any differences from five independent experiments was calculated using the two-tailed heteroscedastic Student's t-test (p < 0.05).
Additionally, to demonstrate the nicking activity and specificity of the gonococcal Vsr endonucleases, the reaction products were also separated in a denaturing gel as was used for the characterization of Vsr endonuclease encoded by E. coli [10,44,45] (Additional file 1: Figure S6).

DNA binding by gonococcal V.NgoAXIII and V.NgoAXIV endonucleases and MutS protein
Protein-DNA complexes were detected using a gel electrophoresis mobility shift assay (EMSA). Binding of DNA by the investigated proteins was studied by incubation of 0.15 μM substrate DNA with 0.15-0.75 μM protein in a final volume of 20 μl, which contained 10 mM Tris-HCl (pH 7.5), 10 mM CaCl 2 , 0.1 mg/ml BSA and a nonspecific competitor DNA (0.032 μM of DNA without any mismatches). The competitor DNA was obtained by PCR amplification using chromosomal DNA of N. gonorrhoeae FA1090 as template, primers 16S RT F and 16S RT R, and PfuUltra DNA Polymerase (Stratagene). Additionally, when needed, 1 mM ATP was added. Next, 4 μl of 5% sucrose was added to the reaction mixture before loading the DNA-protein complexes in a gel. The complexes were separated on neutral 5% (w/v) polyacrylamide (19:1 acrylamide:bisacrylamide) gels supplemented with 1 mM MgCl 2 , 0.5 mM DTT, cast in 22.5 mM Tris base, 22.5 mM boric acid, 0.5 mM EDTA (pH 8.0) and run in 22.5 mM Tris base, 22.5 mM boric acid, 0.5 mM EDTA (pH 8.0). DNA-protein complexes were visualized by ethidium bromide staining. All experiments were repeated at least three times, and the results were averaged.

Standard molecular biology procedures
The primers for DNA amplification were synthesized at the Institute of Biochemistry and Biophysics, Poland. The PCR reactions were carried out using Pfu DNA Polymerase (Thermo Scientific) or PfuUltra DNA Polymerase (Stratagene), according to the manufacturer's recommendations.
All standard methods were carried out in accordance with the protocols described in [33].
In silico analysis DNA and protein sequences were compared with GenBank and SWISS-PROT databases on the BLAST server hosted by the National Center for Biotechnology Information (www.ncbi.nlm.nih.gov/blast).

Enzymes and chemicals
Restriction enzymes, T4 DNA ligase, IPTG (isopropylβ-D-thiogalactopyranoside), as well as DNA and protein markers were purchased from Thermo Scientific and used under conditions recommended by the manufacturers. Kits for DNA clean-up and plasmid DNA preparation were purchased from A&A Biotechnology, Poland. Ni-NTA Agarose was purchased from Qiagen. All other chemicals were purchased from Sigma-Aldrich, unless otherwise noted.

Additional file
Additional file 1: Table S1. Primers used in the work. Table S2. Plasmids used in the work. Table S3. DNA substrates used in the work for the study the activity of gonococcal Vsr endonucleases. Table S4. The mutations occurred in the rpoB gene in the gonococcal mutants with disrupted vsr genes. Table S5. The activity of gonococcal Vsr endonucleases in the presence of the gonococcal MutL protein. The k st values for all DNA substrates recognized by Vsr endonucleases that were used for plotting the graphs in Fig. 4C. Table S6. The activity of gonococcal Vsr endonucleases in the presence of MutL protein of E. coli. The k st values for all DNA substrates recognized by Vsr endonucleases that were used to plot the graphs in Fig. 6. Table S7. The activity of gonococcal Vsr endonucleases in the presence of the gonococcal MutS protein. The k st values for all DNA substrates recognized by Vsr endonucleases that were used to plot the graphs in Fig. 7C. Figure S1. Spontaneous mutation frequency in N. gonorrhoeae mutants with disrupted ngoAXIIIV, ngoAXIVV, mutL Ngo or mutS genes assayed by nalidixic acid resistance. Figure S2. The presence of the MutL Ngo protein decreases amount of Vsr endonuclease required to complete DNA digestion. The results for DNA substrates that are not presented in Fig. 5. Figure S3. Comparison of the increase in mutation frequency in different bacterial species with disrupted mutL, mutS or vsr genes. Figure S4. Purification of the V.NgoAXIII endonuclease, V.NgoAXIV endonuclease, MutL Ngo and MutS protein of N. gonorrhoeae FA1090 and the MutL protein of E. coli. Figure S5. Construction of DNA substrates containing two T:G mismatches and principle of an assay to demonstrate the activity of Vsr endonuclease. Figure S6. Construction of DNA substrates containing one T:G mismatch and principle of an assay to demonstrate the activity of Vsr endonuclease. (PDF 2718 kb)

Abbreviations
C5MTase: C 5 -methyltransferasem5C5-methylcytosineMutL E.coli MutL protein of E. coli K-12MutL Ngo MutL protein of N. gonorrhoeae FA1090VSPVery Short Patch repair system Availability of data and materials The datasets generated and/or analysed during the current study are available from the corresponding author on reasonable request.
Authors' contributions MAP involved in revising the manuscript for important intellectual content, interpretation of data; KB acquisition of data; AK conception and design of the experiments, acquisition of data, analysis and interpretation of data, drafting the manuscript. All authors have read and approved the manuscript.
Ethics approval and consent to participate Not applicable.

Consent for publication
Not applicable.