Differences in Resolution of mwr-Containing Plasmid Dimers Mediated by the Klebsiella pneumoniae and Escherichia coli XerC Recombinases: Pot
http://www.100md.com
《细菌学杂志》
Department of Biological Science, College of Natural Sciences and Mathematics, California State University Fullerton, Fullerton, California 92834-6850,Division of Molecular Genetics, Department of Biochemistry, University of Oxford, Oxford OX1 3QU, United Kingdom
ABSTRACT
Xer-mediated dimer resolution at the mwr site of the multiresistance plasmid pJHCMW1 is osmoregulated in Escherichia coli containing either the Escherichia coli Xer recombination machinery or Xer recombination elements from K. pneumoniae. In the presence of K. pneumoniae XerC (XerCKp), the efficiency of recombination is lower than that in the presence of the E. coli XerC (XerCEc) and the level of dimer resolution is insufficient to stabilize the plasmid, even at low osmolarity. This lower efficiency of recombination at mwr is observed in the presence of E. coli or K. pneumoniae XerD proteins. Mutagenesis experiments identified a region near the N terminus of XerCKp responsible for the lower level of recombination catalyzed by XerCKp at mwr. This region encompasses the second half of the predicted -helix B and the beginning of the predicted -helix C. The efficiencies of recombination at other sites such as dif or cer in the presence of XerCKp or XerCEc are comparable. Therefore, XerCKp is an active recombinase whose action is impaired on the mwr recombination site. This characteristic may result in restriction of the host range of plasmids carrying this site, a phenomenon that may have important implications in the dissemination of antibiotic resistance genes.
INTRODUCTION
A series of mechanisms ensure that plasmids are stably maintained within the host bacterial cells (53). One of these mechanisms consists of resolution of plasmid dimers, which are formed through recombination events in the cell. Dimerization of plasmids leads to multimer formation, a known cause of plasmid segregational instability by reduction of the number of molecules in the cell (25, 45, 46). Xer is a site-specific recombination system that catalyzes conversion of plasmid dimers into monomers, ensuring plasmid stability (47). This recombination mechanism is also responsible for resolution of chromosome dimers allowing proper chromosome segregation at cell division. It is also indirectly involved in virulence by mediating integration of phages harboring genes coding for virulence factors (28, 38, 52).
Xer recombination occurs through formation of a heterotetrameric complex that includes two related tyrosine family recombinases, XerC and XerD (11, 19). These proteins act at target sites found in plasmids and near the ter locus of the chromosome to resolve dimeric DNA replicons to monomers. Recombination proceeds by exchanging two pairs of strands with a Holliday junction being a reaction intermediate (7, 9-11, 15, 19, 20, 38, 55). In addition to XerC and XerD, other factors are required for Xer recombination as a consequence of the adaptation to the different biological functions. Chromosome dimer resolution at the dif site, which consists of two 11-bp binding sites for the recombinases XerC and XerD and a 6-bp central region, requires the protein FtsK (5-7, 13, 29). Recombination at plasmid sites such as psi (pSC101), cer (ColE1), or mwr (pJHCMW1) requires a core site related to dif plus an adjacent DNA stretch of about 180 bp known as accessory sequences (20, 46, 49). The proteins PepA and ArgR (in the case of cer and mwr) or PepA and ArcA (in the case of psi) bind the accessory sequences and induce the formation of a synaptic complex required for recombination and for ensuring that the reaction is exclusively intramolecular (1, 12, 17, 18, 27, 35). Xer recombination at plasmid sites starts with the exchange of one pair of strands catalyzed by XerC, resulting in formation of a Holliday junction. While XerD mediates the exchange of the second pair of strands when the recombination target site is psi, the Holliday junction intermediates at cer or mwr are resolved by Xer-independent cellular processes (3, 4, 18, 30, 49). Regardless of the path followed to resolve the Holliday junction, both XerC and XerD are essential for full dimer resolution. The C-terminal regions of these proteins must interact with their cognate binding regions within the partner recombinase to coordinate the catalysis process (24). XerC must interact with XerD to adopt the active conformation necessary to catalyze the exchange of the first pair of strands, which results in formation of the Holliday junction (4, 21, 24). The C-terminal domains of XerC and XerD also play other roles: they provide sequence-specific DNA recognition to the outer portion of the core recombinase-binding sites, include the catalytic amino acid residues, and contribute to cooperative DNA binding (23, 24, 40, 44). The N-terminal domains contact the inner five nucleotides of the binding site plus one or two nucleotides of the central region and participate in interactions between the monomers (14, 22, 23).
The plasmid pJHCMW1, isolated from a clinical Klebsiella pneumoniae strain, includes the Xer recombination site mwr, which has some unique characteristics (34, 49). This plasmid also harbors the transposon Tn1331, which specifies resistance to several aminoglycosides and -lactams (37, 48, 50, 51, 54). Resolution of dimers harboring mwr is inefficient when the Escherichia coli host cells are cultured in L broth (34). The low levels of mwr-mediated resolution observed proved insufficient to stabilize the plasmid (49). However, the levels of resolution are substantially increased in cells cultured in low-osmolarity broth (34). Resolution experiments using dimers harboring a hybrid site, including the accessory sequences of mwr and the core recombination site of cer or vice versa, demonstrated that the core recombination mwr site is responsible for the difference in recombination efficiency at low or high osmolarity (34). Furthermore, mutagenesis experiments showed that the mwr central region of the recombination core site plays an important role in osmoregulation of site-specific recombination (34). In this paper, we show that the resolution of dimers containing mwr in the presence of K. pneumoniae proteins is dependent on the osmolarity of the culture medium. However, the levels of plasmid dimer resolution at mwr in the presence of the XerC K. pneumoniae protein (XerCKp) are lower than those observed when the XerC E. coli protein (XerCEc) is present. We have identified amino acid residues responsible for this lower activity when mwr is the substrate. We discuss the possible implications of the presence of a site with the characteristics of mwr in dissemination of antibiotic resistance genes.
MATERIALS AND METHODS
E. coli strains and plasmids. The E. coli strains and plasmids used in this study are described in Table 1. The E. coli DS941 strain possesses all four Xer recombination genes intact, and it is designated "wild type" throughout the paper. A subscript indicates if the gene or protein is from K. pneumoniae or E. coli. DNA fragments containing the argRKp, pepAKp, xerCKp, xerCEc, xerDKp, and xerDEc genes were generated by DNA PCR amplification using as templates genomic DNA from K. pneumoniae MGH 7857 or recombinant clones harboring the E. coli genes from the laboratory collection. The DNA fragments were inserted into the appropriate plasmid vectors to generate the different recombinant plasmids (Table 1). Primers to amplify the K. pneumoniae genes were designed using the genome nucleotide sequence available at http://genomeold.wustl.edu/projects/bacterial/kpneumoniae/. These sequence data were produced by the Genome Sequencing Center at Washington University in St. Louis and can be obtained from ftp://genome.wustl.edu/pub/segmgr/bacterial/klebsiella/B_KPN. Plasmid pArgRKp was generated by removing the argREc gene from pCS349 and replacing it with argRKp. Plasmids pPepAKp and pCSXerCKp were generated by replacing the pepAEc with pepAKp and xerCKp, respectively, in pCS119. The inserts of all recombinant plasmids were sequenced to ensure accuracy.
Bacterial growth media and general DNA procedures. Bacteria were grown in Lennox L broth (1% tryptone, 0.5% yeast extract, 0.5% NaCl), and 2% agar was added in the case of solid medium. This is called "high-osmolarity" medium throughout the text. "Low-osmolarity" medium contained the same concentrations of tryptone and yeast extract with no NaCl added. Transformations were carried out as described by Cohen et al. (16). Restriction endonuclease and ligase treatments were carried out as recommended by the suppliers. Plasmid DNA preparations and DNA gel extractions were performed with the QIAspin miniprep kit and QIAquick gel extraction kit, respectively (QIAGEN). DNA fragments for cloning were generated by PCR using the QIAGEN Taq master mix. Protein secondary structure predictions were carried out using the PredictProtein software (36) (http://cubic.bioc.columbia.edu/predictprotein/). Amino acid sequence alignments were generated by subjecting the E. coli amino acid sequences to BLAST search (2) at the Genome Sequencing Center, Washington University, School of Medicine website (http://www.genome.wustl.edu/projects/bacterial/kpneumoniae/index.php). Nucleotide sequencing was performed at the DNA Sequencing Facility, Department of Biochemistry, University of Oxford. Site-directed mutagenesis was carried out with the Quikchange site-directed mutagenesis kit (Stratagene) as recommended by the supplier.
In vivo resolution assay. In vivo resolution assays were carried out as described by Pham et al. (34). Plasmid dimers were prepared as described before (34); plasmid DNA extracted from E. coli JC8679 harboring the plasmid was subjected to electrophoresis in a 0.7% agarose gel. DNA of the correct size to be a plasmid dimer was purified from the agarose gels and introduced by transformation into the XerC-deficient E. coli strain DS981. In this strain, dimers are not resolved by Xer recombination. Purified plasmid dimers were introduced by transformation into the indicated strains, and plasmid was purified after overnight growth at 37°C in the appropriate media containing 100 μg/ml ampicillin (AMP). The efficiency of Xer-mediated dimer resolution was analyzed by agarose gel electrophoresis.
Activity of XerCKp on dif in E. coli and coculture assay. To determine whether XerCKp together with XerDEc can mediate recombination at the E. coli dif locus, the recombinant plasmid pXerCKp was introduced by transformation into E. coli BS52. The transformant strain was cultured overnight, and the percentage of cells that lost the ability to resist kanamycin (KAN), an indication of recombination at dif, was determined. Coculturing assays were used to compare the fitness of E. coli cells harboring XerCKp or XerCEc. A 1:1 mixture of the strains to be tested, E. coli DS981(pXerCKp) or E. coli DS981(pXerCEc), and E. coli DS941 were grown in serial cultures, and their relative frequencies were determined every 20 generations (13, 33). A coculturing assay using E. coli DS981(pACYC184) and E. coli DS941 was carried out as a control.
RESULTS
Dimer resolution in the presence of K. pneumoniae proteins. The K. pneumoniae versions of all four genes involved in Xer-mediated dimer resolution, argRKp, pepAKp, xerDKp, and xerCKp, were cloned, and the recombinant clones were transferred to the corresponding E. coli mutants to assess recombination at mwr at low and high osmolarity. Figure 1a shows resolution of dimers of the plasmid pES, which consists of pUC18 with a DNA fragment including mwr (34), by E. coli DS956 (argR deficient) and E. coli DS9028 (xerD deficient) harboring the recombinant plasmid pArgRKp (coding for ArgRKp) or pXerDKp (coding for XerDKp), respectively. The levels of dimer resolution in the presence of either of these proteins were identical to those obtained with the E. coli proteins (compare with resolution in E. coli DS941 in Fig. 1a); resolution was less efficient in cells growing in high-osmolarity medium. Control resolution assays were carried out using E. coli DS9028 and E. coli DS956 as hosts (Fig. 1a). As expected, there was no resolution of dimers of pES in E. coli mutants lacking any of the four Xer resolution genes. The amino acid sequences of the E. coli and K. pneumoniae proteins are very similar. The ArgR amino acid sequences share 94% identity and 98% similarity, and the XerD amino acid sequences share 90% identity and 95% similarity (Fig. 1b and c).
Analysis of resolution of dimers at mwr in the presence of PepAKp or PepAEc was performed by transformation of the pepA-deficient strain E. coli DS957 with either pPepAKp or pCS119 followed by transformation with pES dimers. Figure 2a shows that both PepA versions have the same effect on resolution of pES dimers; nearly all dimer molecules were resolved at low or high osmolarity. We have shown before that the levels of PepA expressed from these clones are higher than normal, which increases the efficiency of resolution (34). Comparison between the amino acid sequences of PepAKp and PepAEc showed 93% identity (471/503) and 96% similarity (484/503) (Fig. 2b). As expected, no resolution was observed in the pepA-deficient E. coli DS957 strain.
Transformation of E. coli DS981 (xerC deficient) with pXerCKp resulted in a strain where pES dimers were resolved at lower levels as compared to resolution in the wild-type E. coli DS941 or in E. coli DS981 harboring pXerCEc, the plasmid coding for XerCEc (Fig. 3a). However, inspection of Fig. 3a also shows that while XerCKp mediates resolution of mwr-containing plasmid dimers with less efficiency than XerCEc, recombination is still dependent on the osmolarity of the growth medium. No resolution was observed when pES dimers were introduced in the plasmidless E. coli DS981 or when this strain harbored the plasmid vector pACYC184 (Fig. 3a). A comparison of the amino acid sequence of both XerC proteins showed that they share 87% identity (263/300) and 92% similarity (278/300) (Fig. 3b). The C-terminal portions of the proteins show high identity, while there is some divergence at the N termini within the region encompassing the -helices A, B, and C, as defined by secondary structure prediction.
These results indicate that all four K. pneumoniae proteins behave like those from E. coli with respect to the ability of mwr to undergo Xer site-specific recombination in an osmolarity-dependent manner. However, XerCKp mediates lower levels of dimer resolution compared to XerCEc.
XerCKp-mediated resolution of mwr-containing dimers in the presence of XerDKp or XerDEc. XerC must be activated by interaction with XerD to be able to mediate the exchange of the first pair of strands (24). To determine whether XerCkp's reduced ability to resolve pES dimers is due to improper interaction with XerDEc, we compared the levels of resolution of pES dimers by XerCEc or XerCKp in the presence of XerDEc or XerDKp. Dimers of pES were introduced in the xerC xerD-deficient E. coli DS9040 strain already transformed with pCSXerCKp and pXerDKp. Figure 3a shows that the relative amounts of dimers and monomers in E. coli DS981(pXerCKp) and E. coli DS9040(pCSXerCKp, pXerDKp) are nearly identical, indicating that regardless of the source of XerD, the levels of resolution promoted by XerCKp were lower than those mediated by XerCEc. On the other hand, XerCKp-mediated resolution of dimers at cer was efficient in the presence of either XerDEc or XerDKp (Fig. 4; see next section).
These results suggest that improper interaction with XerDEc is not the cause of the poor recombination activity exhibited by XerCKp when mwr is the recombination target. In addition, on the basis of our previous studies (34), we conclude that the levels of recombination at mwr mediated by XerCKp at low or high osmolarity are not sufficient to promote stable maintenance of the plasmid.
XerCKp activity on cer and dif. We determined if Xer recombination mediated by XerCKp using other target sites also occurred at levels detectably lower than those mediated by XerCEc. Xer recombination at the ColE1 cer site was tested by transformation of E. coli DS981(pXerCKp) or E. coli DS9040(pCSXerCKp, pXerDKp) with dimers of pKS492, a recombinant clone including cer. Figure 4 shows a comparison of resolution of dimers of the recombinant clones pES (mwr) and pKS492 (cer). While resolution of pES dimers was inefficient, no pKS492 dimers were detected, indicating that XerCKp mediates recombination at high levels when cer was the target site.
Xer recombination at dif is important to resolve chromosome dimers before cytokinesis. To determine whether XerCKp also mediates poor levels of recombination when dif is the target, we transformed E. coli BS52 with pXerCKp. This xerC-deficient E. coli strain carries a KAN resistance gene between two directly repeated dif copies in its chromosome. Therefore, the fraction of cells resistant to KAN gives an indication of the level of site-specific recombination at dif. All colonies tested had lost the KAN resistance gene after overnight culture of E. coli BS52(pXerCKp), suggesting that XerCKp is fully active when dif is the recombination target.
To compare the XerCKp and XerCEc efficiencies to resolve chromosome dimers, we carried out coculturing assays, a technique used for comparing the fitness of strains. It has been shown previously that efficient chromosome dimer resolution by Xer recombination results in a growth advantage with respect to strains impaired in that function, as detected in coculturing assays (13, 33). Table 2 shows coculturing of E. coli DS981(pXerCKp) or E. coli DS981(pXerCEc) with the wild-type E. coli DS941. The growth competition with E. coli DS941 showed no differences between the E. coli DS981 strains harboring either XerCKp or XerCEc. Conversely, coculturing E. coli DS981(pACYC184) and E. coli DS941 showed that the E. coli strain harboring the cloning vector pACYC184 was rapidly outcompeted (Table 2).
These results indicate that lower XerCKp activity compared to that of XerCEc was detectable only when mwr was the target for recombination. A comparison between the two XerC proteins acting on cer or dif showed no difference in levels of activity. Therefore, we conclude that XerCKp must be an efficient enzyme for resolving chromosome dimers in its natural host and for supporting stable maintenance of plasmids with a suitable target sequence such as ColE1.
Identification of the XerCKp region responsible for lower recombination activity when mwr is the target site. Most differences in the XerCKp and XerCEc amino acid sequences occur within the N-terminal region (Fig. 3b). To determine whether amino acids within the region encompassing the second half of the predicted -helix B and the beginning of the predicted -helix C are responsible for the differences observed between XerCKp and XerCEc, we generated mutant derivatives substituting for groups of amino acids in XerCKp with those present at the equivalent position in XerCEc. The three substitution mutants, ALAD40-43NFAS, DEA43-45SEN, and AAQ55-57VTM, were tested in pES dimer resolution assays. Figure 5a shows that both ALAD40-43NFAS and DEA43-45SEN were able to mediate complete dimer resolution, while the level of resolution in the presence of derivative AAQ55-57VTM was similar to that observed with the wild-type XerCKp. To identify the amino acid(s) responsible for the variation in levels of activity, we generated single-amino-acid substitutions within the stretch encompassed by residues 40 to 45. Only one of these substitutions, L41F, was able to mediate complete resolution (Fig. 5a). Two of the mutants, A40N and D43S, did not show significant difference from the wild-type XerCKp, and one of them, A45N, showed a substandard activity (Fig. 5a). These results indicate that this region plays an important role in the levels of XerC activity when mwr is the recombination target. Although the single-amino-acid substitution L41F could elevate the activity of the enzyme, the same change could be obtained when both D43 and Q45 were substituted for by those amino acids present in XerCEc. If L41 were the sole amino acid responsible for the difference in phenotypes, one would expect a reduced activity when the reverse mutation occurs in XerCEc. Figure 5b shows that this is not the case and the XerCEc substitution F39L still shows full or nearly full activity. A certain degree of modification of levels of dimer resolution at mwr mediated by the mutant proteins could also be due to changes in expression levels or stability of the derivatives.
The results shown in this section suggest that amino acids within the region encompassing the second half of the predicted -helix B and the beginning of the predicted -helix C are responsible for the lower XerCKp activity in site-specific recombination at mwr.
DISCUSSION
The multiresistance plasmid pJHCMW1, originally isolated from a clinical K. pneumoniae strain, includes the Xer recombination site mwr (37, 49, 54). Our previous work showed an inverse relationship between osmolarity of the growth medium and efficiency of resolution by Xer recombination at mwr in E. coli strains (34). Although we still do not understand how osmotically regulated recombination at mwr may help the stability of pJHCMW1, one can speculate that the mwr locus is not critical for stability under certain environments in which resolution occurs via the resolvase at res, but it is essential in others where the osmolarity is lower.
The XerDKp, ArgRKp, and PepAKp proteins behaved identically to those from E. coli with respect to regulation by the osmolarity of the milieu and levels of dimer resolution at mwr. The lower XerCKp activity on mwr could be due to an intrinsic property of this enzyme, such as improper binding or improper formation of the synaptic complex, or to improper interaction with XerDEc resulting in poor XerCKp activation. Our results suggest that improper interaction of XerCKP with the heterologous XerDEc is not the main cause of lower recombination efficiency at mwr. The levels of activity of XerCKp were identical in the presence of XerDEc or XerDKp. These results are in agreement with the fact that major XerC-XerD interactions occur at the C termini, and, as shown in Fig. 1b and 3b (amino acids in gray boxes), the XerC and XerD proteins from both bacteria are identical at these regions. In vitro recombination experiments in the presence of purified recombinases and accessory proteins will be necessary to confirm that there are not differences in activation of XerCKp by XerDKp or XerDEc.
Our results indicate that the lower efficiency of recombination promoted by XerCKp as compared to that of XerCEc is specifically detected when mwr is the target site. It is known that the XerC N-terminal domain contacts the inner nucleotides of the binding site, which are important for recombination activity (8, 22, 23). This domain of the recombinases, although dispensable for catalytic activity, plays a role in controlling strand exchange by the recombinases (22, 26). Inspection of the nucleotide sequences of all three core recombination sites, mwr, cer, and dif, shows a higher variability at the inner portion of the XerC binding site (Fig. 6). Also, inspection of the amino acid sequence of the N terminus of XerCKp shows the highest divergence compared to that of XerCEc, and the mutagenesis experiments described in this work indicate that the region encompassed between amino acids 41 and 45 is responsible for promoting lower levels of recombination at mwr. We speculate that, even under the best conditions, i.e., low osmolarity of the growth medium, XerCKp does not bind optimally to the inner nucleotides of the mwr XerC binding site. This could be due to the specific conformation of the synaptic complex determined by the mwr accessory sequences and core recombination site. The presence of an imperfect ARG box in the mwr accessory sequences (34) may lead to a suboptimal synaptic complex to which XerCKp cannot bind properly. Mutagenesis of the mwr core recombination site as well as in vitro recombination experiments will confirm if this is the case.
We have shown previously that levels of dimer resolution comparable to those observed in the presence of XerCKp in this work are not sufficient to confer plasmid stability (49). Furthermore, our previous assays indicated that pJHCMW1 stability by multimer resolution occurs through site-specific recombination mediated by the Tn1331 transposon resolvase at the res site when the host cells are cultured in L broth (49). These findings may have interesting implications with respect to the host range of plasmids and dissemination of antibiotic resistance genes. Possessing a target for dimer resolution that is not well recognized as a substrate by a given host's Xer recombination system would make an otherwise stable plasmid unstable, thereby narrowing its host range. On the other hand, since the other essential plasmid inheritance functions may operate properly in the host, acquisition of a dimer resolution system, such as insertion of a replicative transposon like Tn1331, would allow the plasmid to be stably maintained. Inclusion of the transposon, which usually carries antibiotic resistance genes, within the plasmid would result in widening of the plasmid's host range and consequently expanding the number of genera or species harboring the resistance genes. In the case of pJHCMW1, an original plasmid unstable in K. pneumoniae could have acquired Tn1331, a process that resulted in an expansion of the host range to include this pathogen. Interestingly, a Salmonella enterica plasmid, pFPBT1, has been recently shown to include a replication region nearly identical to pJHCMW1, a site highly related to mwr, and a replicative transposon (31). It will be of great interest to determine the efficiency of the S. enterica Xer recombination proteins to resolve dimers including the pFPBT1 Xer recombination site. To our best knowledge, there have not been reports of plasmids consisting of the pJHCMW1 replicon without a replicative transposon. Another interesting possibility is that a plasmid's Xer resolution site could undergo detrimental mutations after acquisition of a replicative transposon. In these cases, the loss of an efficient dimer resolution site would result in a dependency on the transposon for plasmid viability. Loss of the transposon, and concomitantly of resistance genes, by deletion, recombination, or other mechanisms would lead to loss of the plasmid.
ACKNOWLEDGMENTS
This work was funded by Public Health grants AI47115-02 (M.E.T., National Institute of Allergy and Infectious Diseases, National Institutes of Health), LA Basin Minority Health and Health Disparities International Research Training Program (MHIRT) 9T37MD001368-08 (National Center on Minority Health and Health Disparities), and a grant from Wellcome Trust (D.J.S.). D.B. and J.N. were supported by grants R25 GM56820-03 (MSD), MIRT 5T37TW000048-07, and MHIRT 9T37MD001368-08 from the National Institutes of Health. J.R. was supported by grant MIRT 5T37 TW000048-07 from the National Institutes of Health.
We thank Migena Bregu and Christophe Possoz for useful comments and suggestions and Rachel Baker for expert technical assistance. We are also indebted to The Genome Sequencing Center at Washington University for making the K. pneumoniae genome sequence available prior to publication.
REFERENCES
Alen, C., D. J. Sherratt, and S. D. Colloms. 1997. Direct interaction of aminopeptidase A with recombination site DNA in Xer site-specific recombination. EMBO J. 16:5188-5197.
Altschul, S. F., T. L. Madden, A. A. Schaffer, J. Zhang, Z. Zhang, W. Miller, and D. J. Lipman. 1997. Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res. 25:3389-3402.
Arciszewska, L., and D. Sherratt. 1995. Xer site-specific recombination in vitro. EMBO J. 14:2112-2120.
Arciszewska, L. K., R. A. Baker, B. Hallet, and D. J. Sherratt. 2000. Coordinated control of XerC and XerD catalytic activities during Holliday junction resolution. J. Mol. Biol. 299:391-403.
Aussel, L., F. Barre, M. Aroyo, A. Stasiak, A. Z. Stasiak, and D. Sherratt. 2002. FtsK is a DNA motor protein that activates chromosome dimer resolution by switching the catalytic state of the XerC and XerD recombinases. Cell 108:195-205.
Barre, F. X., M. Aroyo, S. Colloms, A. Helfrich, F. Cornet, and D. J. Sherratt. 2000. FtsK functions in the processing of a Holliday junction intermediate during bacterial chromosome segregation. Genes Dev. 14:2976-2988.
Barre, F.-X., and D. J. Sherratt. 2002. Xer site-specific recombination: promoting chromosome segregation, p. 149-161. In N. Craig, R. Craigie, M. Gellert, and A. M. Lambowitz (ed.), Mobile DNA II, vol. 1. ASM Press, Washington, D.C.
Blake, J. A., N. Ganguly, and D. J. Sherratt. 1997. DNA sequence of recombinase-binding sites can determine Xer site-specific recombination outcome. Mol. Microbiol. 23:387-398.
Blakely, G., S. Colloms, G. May, M. Burke, and D. J. Sherratt. 1991. Escherichia coli XerC recombinase is required for chromosomal segregation at cell division. New Biol. 3:789-798.
Blakely, G., A. Davidson, and D. J. Sherratt. 2000. Sequential strand exchange by XerC and XerD during site-specific recombination at dif. J. Biol. Chem. 275:9930-9936.
Blakely, G., G. May, R. McCulloch, L. Arciszewska, M. Burke, S. Lovett, and D. J. Sherratt. 1993. Two related recombinases are required for site specific recombination at dif and cer in E. coli K12. Cell 75:351-361.
Bregu, M., D. J. Sherratt, and S. D. Colloms. 2002. Accessory factors determine the order of strand exchange in Xer recombination at psi. EMBO J. 21:3888-3897.
Capiaux, H., C. Lesterlin, K. Perals, J. Louarn, and F. Cornet. 2002. A dual role for the FtsK protein in Escherichia coli chromosome segregation. EMBO Rep. 3:532-536.
Chen, Y., U. Narendra, L. E. Iype, M. M. Cox, and P. A. Rice. 2000. Crystal structure of a Flp recombinase-Holliday junction complex: assembly of an active oligomer by helix swapping. Mol. Cell 6:885-897.
Clerget, M. 1991. Site-specific recombination promoted by a short DNA segment of plasmid R1 and by a homologous segment in the terminus region of the Escherichia coli chromosome. New Biol. 3:780-788.
Cohen, S. N., A. C. Chang, and L. Hsu. 1972. Nonchromosomal antibiotic resistance in bacteria: genetic transformation of Escherichia coli by R-factor DNA. Proc. Natl. Acad. Sci. USA 69:2110-2114.
Colloms, S., C. Alen, and D. J. Sherratt. 1998. The ArcA/ArcB two-component regulatory system of Escherichia coli is essential for Xer site-specific recombination at psi. Mol. Microbiol. 28:521-530.
Colloms, S., R. McCulloch, K. Grant, L. Neilson, and D. J. Sherratt. 1996. Xer-mediated site-specific recombination in vitro. EMBO J. 15:1172-1181.
Colloms, S. D., P. Sykora, G. Szatmari, and D. J. Sherratt. 1990. Recombination at ColE1 cer requires the Escherichia coli xerC gene product, a member of the lambda integrase family of site-specific recombinases. J. Bacteriol. 172:6973-6980.
Cornet, F., I. Mortier, J. Patte, and J.-M. Louarn. 1994. Plasmid pSC101 harbors a recombination site, psi, which is able to resolve plasmid multimers and to substitute for the analogous chromosomal Escherichia coli site dif. J. Bacteriol. 176:3188-3195.
Ferreira, H., B. Butler-Cole, A. Burgin, R. Baker, D. J. Sherratt, and L. K. Arciszewska. 2003. Functional analysis of the C-terminal domains of the site-specific recombinases XerC and XerD. J. Mol. Biol. 330:15-27.
Ferreira, H., D. J. Sherratt, and L. Arciszewska. 2001. Switching catalytic activity in the XerCD site-specific recombination machine. J. Mol. Biol. 312:45-57.
Guo, F., D. N. Gopaul, and G. D. van Duyne. 1997. Structure of Cre recombinase complexed with DNA in a site-specific recombination synapse. Nature 389:40-46.
Hallet, B., L. K. Arciszewska, and D. J. Sherratt. 1999. Reciprocal control of catalysis by the tyrosine recombinases XerC and XerD: an enzymatic switch in site-specific recombination. Mol. Cell 4:949-959.
Hallet, B., V. Vanhooff, and F. Cornet. 2004. DNA site-specific resolution systems, p. 145-180. In B. Funnell and G. Phillips (ed.), Plasmid biology. ASM Press, Washington, D.C.
Hayes, F., and D. J. Sherratt. 1997. Recombinase binding specificity at the chromosome dimer resolution site dif of Escherichia coli. J. Mol. Biol. 266:525-537.
Hodgman, T. C., H. Griffiths, and D. K. Summers. 1998. Nucleoprotein architecture and ColE1 dimer resolution: a hypothesis. Mol. Microbiol. 29:545-558.
Huber, K., and M. Waldor. 2002. Filamentous phage integration requires the host recombinases XerC and XerD. Nature 417:656-659.
Massey, T. H., L. Aussel, F. X. Barre, and D. J. Sherratt. 2004. Asymmetric activation of Xer site-specific recombination by FtsK. EMBO Rep. 5:399-404.
McCulloch, R., L. Coggins, S. Colloms, and D. J. Sherratt. 1994. Xer-mediated site-specific recombination at cer generates Holliday junctions in vivo. EMBO J. 13:1844-1855.
Pasquali, F., C. Kehrenberg, G. Manfreda, and S. Schwarz. 2005. Physical linkage of Tn3 and part of Tn1721 in a tetracycline and ampicillin resistance plasmid from Salmonella Typhimurium. J. Antimicrob. Chemother. 55:562-565.
Perals, K., H. Capiaux, J. B. Vincourt, J. M. Louarn, D. J. Sherratt, and F. Cornet. 2001. Interplay between recombination, cell division and chromosome structure during chromosome dimer resolution in Escherichia coli. Mol. Microbiol. 39:904-913.
Perals, K., F. Cornet, Y. Merlet, I. Delon, and J. M. Louarn. 2000. Functional polarization of the Escherichia coli chromosome terminus: the dif site acts in chromosome dimer resolution only when located between long stretches of opposite polarity. Mol. Microbiol. 36:33-43.
Pham, H., K. J. Dery, D. J. Sherratt, and M. E. Tolmasky. 2002. Osmoregulation of dimer resolution at the plasmid pJHCMW1 mwr locus by Escherichia coli XerCD recombination. J. Bacteriol. 184:1607-1616.
Rijns, M., Y. Lu, S. Leach, and S. Colloms. 2005. Mutagenesis of PepA suggests a new model for the Xer/cer synaptic complex. Mol. Microbiol. 57:927-941.
Rost, B., and J. Liu. 2003. The PredictProtein Server. Nucleic Acids Res. 31:3300-3304.
Sarno, R., G. McGillivary, D. J. Sherratt, L. A. Actis, and M. E. Tolmasky. 2002. Complete nucleotide sequence of Klebsiella pneumoniae multiresistance plasmid pJHCMW1. Antimicrob. Agents Chemother. 46:3422-3427.
Sherratt, D. J., L. Arciszewska, G. Blakely, S. Colloms, K. Grant, N. Leslie, and R. McCulloch. 1995. Site-specific recombination and circular chromosome segregation. Phil. Trans. R. Soc. London B 347:37-42.
Sherratt, D. J., B. Soballe, F. X. Barre, S. Filipe, I. Lau, T. Massey, and J. Yates. 2004. Recombination and chromosome segregation. Phil. Trans. R. Soc. Lond. B 359:61-69.
Spiers, A. J., and D. J. Sherratt. 1999. C-terminal interactions between the XerC and XerD site-specific recombinases. Mol. Microbiol. 32:1031-1042.
Spiers, A. J., and D. J. Sherratt. 1997. Relating primary structure to function in the Escherichia coli XerD site-specific recombinase. Mol. Microbiol. 24:1071-1082.
Stirling, C. J., S. D. Colloms, J. F. Collins, G. Szatmari, and D. J. Sherratt. 1989. xerB, an Escherichia coli gene required for plasmid ColE1 site-specific recombination, is identical to pepA, encoding aminopeptidase A, a protein with substantial similarity to bovine lens leucine aminopeptidase. EMBO J. 8:1623-1627.
Stirling, C. J., G. Szatmari, G. Stewart, M. C. Smith, and D. J. Sherratt. 1988. The arginine repressor is essential for plasmid-stabilizing site-specific recombination at the ColE1 cer locus. EMBO J. 7:4389-4395.
Subramanya, H. S., L. K. Arciszewska, R. A. Baker, L. E. Bird, D. J. Sherratt, and D. B. Wigley. 1997. Crystal structure of the site-specific recombinase, XerD. EMBO J. 16:5178-5187.
Summers, D., C. Beton, and H. Withers. 1993. Multicopy plasmid instability: the dimer catastrophe hypothesis. Mol. Microbiol. 8:1031-1038.
Summers, D., and D. J. Sherratt. 1984. Multimerization of high copy number plasmids causes instability: CoIE1 encodes a determinant essential for plasmid monomerization and stability. Cell 36:1097-1103.
Summers, D. K., and D. J. Sherratt. 1988. Resolution of ColE1 dimers requires a DNA sequence implicated in the three-dimensional organization of the cer site. EMBO J. 7:851-858.
Tolmasky, M. E. 1990. Sequencing and expression of aadA, bla, and tnpR from the multiresistance transposon Tn1331. Plasmid 24:218-226.
Tolmasky, M. E., S. Colloms, G. Blakely, and D. J. Sherratt. 2000. Stability by multimer resolution of pJHCMW1 is due to the Tn1331 resolvase and not to the Escherichia coli Xer system. Microbiology 146:581-589.
Tolmasky, M. E., and J. H. Crosa. 1993. Genetic organization of antibiotic resistance genes (aac(6')-Ib, aadA, and oxa9) in the multiresistance transposon Tn1331. Plasmid 29:31-40.
Tolmasky, M. E., and J. H. Crosa. 1987. Tn1331, a novel multiresistance transposon encoding resistance to amikacin and ampicillin in Klebsiella pneumoniae. Antimicrob. Agents Chemother. 31:1955-1960.
Val, M., M. Bouvier, J. Campos, D. J. Sherratt, F. Cornet, D. Mazel, and F. Barre. 2005. The single stranded genome of phage CTX is the form used for integration into the genome of Vibrio cholerae. Mol. Cell 19:559-566.
Wegrzyn, G. 2005. What does "plasmid biology" currently mean Summary of the Plasmid Biology 2004 Meeting. Plasmid 53:14-22.
Woloj, M., M. E. Tolmasky, M. C. Roberts, and J. H. Crosa. 1986. Plasmid-encoded amikacin resistance in multiresistant strains of Klebsiella pneumoniae isolated from neonates with meningitis. Antimicrob. Agents Chemother. 29:315-319.
Yates, J., M. Aroyo, D. J. Sherratt, and F. X. Barre. 2003. Species specificity in the activation of Xer recombination at dif by FtsK. Mol. Microbiol. 49:241-249.(Duyen Bui, Judianne Ramis)
ABSTRACT
Xer-mediated dimer resolution at the mwr site of the multiresistance plasmid pJHCMW1 is osmoregulated in Escherichia coli containing either the Escherichia coli Xer recombination machinery or Xer recombination elements from K. pneumoniae. In the presence of K. pneumoniae XerC (XerCKp), the efficiency of recombination is lower than that in the presence of the E. coli XerC (XerCEc) and the level of dimer resolution is insufficient to stabilize the plasmid, even at low osmolarity. This lower efficiency of recombination at mwr is observed in the presence of E. coli or K. pneumoniae XerD proteins. Mutagenesis experiments identified a region near the N terminus of XerCKp responsible for the lower level of recombination catalyzed by XerCKp at mwr. This region encompasses the second half of the predicted -helix B and the beginning of the predicted -helix C. The efficiencies of recombination at other sites such as dif or cer in the presence of XerCKp or XerCEc are comparable. Therefore, XerCKp is an active recombinase whose action is impaired on the mwr recombination site. This characteristic may result in restriction of the host range of plasmids carrying this site, a phenomenon that may have important implications in the dissemination of antibiotic resistance genes.
INTRODUCTION
A series of mechanisms ensure that plasmids are stably maintained within the host bacterial cells (53). One of these mechanisms consists of resolution of plasmid dimers, which are formed through recombination events in the cell. Dimerization of plasmids leads to multimer formation, a known cause of plasmid segregational instability by reduction of the number of molecules in the cell (25, 45, 46). Xer is a site-specific recombination system that catalyzes conversion of plasmid dimers into monomers, ensuring plasmid stability (47). This recombination mechanism is also responsible for resolution of chromosome dimers allowing proper chromosome segregation at cell division. It is also indirectly involved in virulence by mediating integration of phages harboring genes coding for virulence factors (28, 38, 52).
Xer recombination occurs through formation of a heterotetrameric complex that includes two related tyrosine family recombinases, XerC and XerD (11, 19). These proteins act at target sites found in plasmids and near the ter locus of the chromosome to resolve dimeric DNA replicons to monomers. Recombination proceeds by exchanging two pairs of strands with a Holliday junction being a reaction intermediate (7, 9-11, 15, 19, 20, 38, 55). In addition to XerC and XerD, other factors are required for Xer recombination as a consequence of the adaptation to the different biological functions. Chromosome dimer resolution at the dif site, which consists of two 11-bp binding sites for the recombinases XerC and XerD and a 6-bp central region, requires the protein FtsK (5-7, 13, 29). Recombination at plasmid sites such as psi (pSC101), cer (ColE1), or mwr (pJHCMW1) requires a core site related to dif plus an adjacent DNA stretch of about 180 bp known as accessory sequences (20, 46, 49). The proteins PepA and ArgR (in the case of cer and mwr) or PepA and ArcA (in the case of psi) bind the accessory sequences and induce the formation of a synaptic complex required for recombination and for ensuring that the reaction is exclusively intramolecular (1, 12, 17, 18, 27, 35). Xer recombination at plasmid sites starts with the exchange of one pair of strands catalyzed by XerC, resulting in formation of a Holliday junction. While XerD mediates the exchange of the second pair of strands when the recombination target site is psi, the Holliday junction intermediates at cer or mwr are resolved by Xer-independent cellular processes (3, 4, 18, 30, 49). Regardless of the path followed to resolve the Holliday junction, both XerC and XerD are essential for full dimer resolution. The C-terminal regions of these proteins must interact with their cognate binding regions within the partner recombinase to coordinate the catalysis process (24). XerC must interact with XerD to adopt the active conformation necessary to catalyze the exchange of the first pair of strands, which results in formation of the Holliday junction (4, 21, 24). The C-terminal domains of XerC and XerD also play other roles: they provide sequence-specific DNA recognition to the outer portion of the core recombinase-binding sites, include the catalytic amino acid residues, and contribute to cooperative DNA binding (23, 24, 40, 44). The N-terminal domains contact the inner five nucleotides of the binding site plus one or two nucleotides of the central region and participate in interactions between the monomers (14, 22, 23).
The plasmid pJHCMW1, isolated from a clinical Klebsiella pneumoniae strain, includes the Xer recombination site mwr, which has some unique characteristics (34, 49). This plasmid also harbors the transposon Tn1331, which specifies resistance to several aminoglycosides and -lactams (37, 48, 50, 51, 54). Resolution of dimers harboring mwr is inefficient when the Escherichia coli host cells are cultured in L broth (34). The low levels of mwr-mediated resolution observed proved insufficient to stabilize the plasmid (49). However, the levels of resolution are substantially increased in cells cultured in low-osmolarity broth (34). Resolution experiments using dimers harboring a hybrid site, including the accessory sequences of mwr and the core recombination site of cer or vice versa, demonstrated that the core recombination mwr site is responsible for the difference in recombination efficiency at low or high osmolarity (34). Furthermore, mutagenesis experiments showed that the mwr central region of the recombination core site plays an important role in osmoregulation of site-specific recombination (34). In this paper, we show that the resolution of dimers containing mwr in the presence of K. pneumoniae proteins is dependent on the osmolarity of the culture medium. However, the levels of plasmid dimer resolution at mwr in the presence of the XerC K. pneumoniae protein (XerCKp) are lower than those observed when the XerC E. coli protein (XerCEc) is present. We have identified amino acid residues responsible for this lower activity when mwr is the substrate. We discuss the possible implications of the presence of a site with the characteristics of mwr in dissemination of antibiotic resistance genes.
MATERIALS AND METHODS
E. coli strains and plasmids. The E. coli strains and plasmids used in this study are described in Table 1. The E. coli DS941 strain possesses all four Xer recombination genes intact, and it is designated "wild type" throughout the paper. A subscript indicates if the gene or protein is from K. pneumoniae or E. coli. DNA fragments containing the argRKp, pepAKp, xerCKp, xerCEc, xerDKp, and xerDEc genes were generated by DNA PCR amplification using as templates genomic DNA from K. pneumoniae MGH 7857 or recombinant clones harboring the E. coli genes from the laboratory collection. The DNA fragments were inserted into the appropriate plasmid vectors to generate the different recombinant plasmids (Table 1). Primers to amplify the K. pneumoniae genes were designed using the genome nucleotide sequence available at http://genomeold.wustl.edu/projects/bacterial/kpneumoniae/. These sequence data were produced by the Genome Sequencing Center at Washington University in St. Louis and can be obtained from ftp://genome.wustl.edu/pub/segmgr/bacterial/klebsiella/B_KPN. Plasmid pArgRKp was generated by removing the argREc gene from pCS349 and replacing it with argRKp. Plasmids pPepAKp and pCSXerCKp were generated by replacing the pepAEc with pepAKp and xerCKp, respectively, in pCS119. The inserts of all recombinant plasmids were sequenced to ensure accuracy.
Bacterial growth media and general DNA procedures. Bacteria were grown in Lennox L broth (1% tryptone, 0.5% yeast extract, 0.5% NaCl), and 2% agar was added in the case of solid medium. This is called "high-osmolarity" medium throughout the text. "Low-osmolarity" medium contained the same concentrations of tryptone and yeast extract with no NaCl added. Transformations were carried out as described by Cohen et al. (16). Restriction endonuclease and ligase treatments were carried out as recommended by the suppliers. Plasmid DNA preparations and DNA gel extractions were performed with the QIAspin miniprep kit and QIAquick gel extraction kit, respectively (QIAGEN). DNA fragments for cloning were generated by PCR using the QIAGEN Taq master mix. Protein secondary structure predictions were carried out using the PredictProtein software (36) (http://cubic.bioc.columbia.edu/predictprotein/). Amino acid sequence alignments were generated by subjecting the E. coli amino acid sequences to BLAST search (2) at the Genome Sequencing Center, Washington University, School of Medicine website (http://www.genome.wustl.edu/projects/bacterial/kpneumoniae/index.php). Nucleotide sequencing was performed at the DNA Sequencing Facility, Department of Biochemistry, University of Oxford. Site-directed mutagenesis was carried out with the Quikchange site-directed mutagenesis kit (Stratagene) as recommended by the supplier.
In vivo resolution assay. In vivo resolution assays were carried out as described by Pham et al. (34). Plasmid dimers were prepared as described before (34); plasmid DNA extracted from E. coli JC8679 harboring the plasmid was subjected to electrophoresis in a 0.7% agarose gel. DNA of the correct size to be a plasmid dimer was purified from the agarose gels and introduced by transformation into the XerC-deficient E. coli strain DS981. In this strain, dimers are not resolved by Xer recombination. Purified plasmid dimers were introduced by transformation into the indicated strains, and plasmid was purified after overnight growth at 37°C in the appropriate media containing 100 μg/ml ampicillin (AMP). The efficiency of Xer-mediated dimer resolution was analyzed by agarose gel electrophoresis.
Activity of XerCKp on dif in E. coli and coculture assay. To determine whether XerCKp together with XerDEc can mediate recombination at the E. coli dif locus, the recombinant plasmid pXerCKp was introduced by transformation into E. coli BS52. The transformant strain was cultured overnight, and the percentage of cells that lost the ability to resist kanamycin (KAN), an indication of recombination at dif, was determined. Coculturing assays were used to compare the fitness of E. coli cells harboring XerCKp or XerCEc. A 1:1 mixture of the strains to be tested, E. coli DS981(pXerCKp) or E. coli DS981(pXerCEc), and E. coli DS941 were grown in serial cultures, and their relative frequencies were determined every 20 generations (13, 33). A coculturing assay using E. coli DS981(pACYC184) and E. coli DS941 was carried out as a control.
RESULTS
Dimer resolution in the presence of K. pneumoniae proteins. The K. pneumoniae versions of all four genes involved in Xer-mediated dimer resolution, argRKp, pepAKp, xerDKp, and xerCKp, were cloned, and the recombinant clones were transferred to the corresponding E. coli mutants to assess recombination at mwr at low and high osmolarity. Figure 1a shows resolution of dimers of the plasmid pES, which consists of pUC18 with a DNA fragment including mwr (34), by E. coli DS956 (argR deficient) and E. coli DS9028 (xerD deficient) harboring the recombinant plasmid pArgRKp (coding for ArgRKp) or pXerDKp (coding for XerDKp), respectively. The levels of dimer resolution in the presence of either of these proteins were identical to those obtained with the E. coli proteins (compare with resolution in E. coli DS941 in Fig. 1a); resolution was less efficient in cells growing in high-osmolarity medium. Control resolution assays were carried out using E. coli DS9028 and E. coli DS956 as hosts (Fig. 1a). As expected, there was no resolution of dimers of pES in E. coli mutants lacking any of the four Xer resolution genes. The amino acid sequences of the E. coli and K. pneumoniae proteins are very similar. The ArgR amino acid sequences share 94% identity and 98% similarity, and the XerD amino acid sequences share 90% identity and 95% similarity (Fig. 1b and c).
Analysis of resolution of dimers at mwr in the presence of PepAKp or PepAEc was performed by transformation of the pepA-deficient strain E. coli DS957 with either pPepAKp or pCS119 followed by transformation with pES dimers. Figure 2a shows that both PepA versions have the same effect on resolution of pES dimers; nearly all dimer molecules were resolved at low or high osmolarity. We have shown before that the levels of PepA expressed from these clones are higher than normal, which increases the efficiency of resolution (34). Comparison between the amino acid sequences of PepAKp and PepAEc showed 93% identity (471/503) and 96% similarity (484/503) (Fig. 2b). As expected, no resolution was observed in the pepA-deficient E. coli DS957 strain.
Transformation of E. coli DS981 (xerC deficient) with pXerCKp resulted in a strain where pES dimers were resolved at lower levels as compared to resolution in the wild-type E. coli DS941 or in E. coli DS981 harboring pXerCEc, the plasmid coding for XerCEc (Fig. 3a). However, inspection of Fig. 3a also shows that while XerCKp mediates resolution of mwr-containing plasmid dimers with less efficiency than XerCEc, recombination is still dependent on the osmolarity of the growth medium. No resolution was observed when pES dimers were introduced in the plasmidless E. coli DS981 or when this strain harbored the plasmid vector pACYC184 (Fig. 3a). A comparison of the amino acid sequence of both XerC proteins showed that they share 87% identity (263/300) and 92% similarity (278/300) (Fig. 3b). The C-terminal portions of the proteins show high identity, while there is some divergence at the N termini within the region encompassing the -helices A, B, and C, as defined by secondary structure prediction.
These results indicate that all four K. pneumoniae proteins behave like those from E. coli with respect to the ability of mwr to undergo Xer site-specific recombination in an osmolarity-dependent manner. However, XerCKp mediates lower levels of dimer resolution compared to XerCEc.
XerCKp-mediated resolution of mwr-containing dimers in the presence of XerDKp or XerDEc. XerC must be activated by interaction with XerD to be able to mediate the exchange of the first pair of strands (24). To determine whether XerCkp's reduced ability to resolve pES dimers is due to improper interaction with XerDEc, we compared the levels of resolution of pES dimers by XerCEc or XerCKp in the presence of XerDEc or XerDKp. Dimers of pES were introduced in the xerC xerD-deficient E. coli DS9040 strain already transformed with pCSXerCKp and pXerDKp. Figure 3a shows that the relative amounts of dimers and monomers in E. coli DS981(pXerCKp) and E. coli DS9040(pCSXerCKp, pXerDKp) are nearly identical, indicating that regardless of the source of XerD, the levels of resolution promoted by XerCKp were lower than those mediated by XerCEc. On the other hand, XerCKp-mediated resolution of dimers at cer was efficient in the presence of either XerDEc or XerDKp (Fig. 4; see next section).
These results suggest that improper interaction with XerDEc is not the cause of the poor recombination activity exhibited by XerCKp when mwr is the recombination target. In addition, on the basis of our previous studies (34), we conclude that the levels of recombination at mwr mediated by XerCKp at low or high osmolarity are not sufficient to promote stable maintenance of the plasmid.
XerCKp activity on cer and dif. We determined if Xer recombination mediated by XerCKp using other target sites also occurred at levels detectably lower than those mediated by XerCEc. Xer recombination at the ColE1 cer site was tested by transformation of E. coli DS981(pXerCKp) or E. coli DS9040(pCSXerCKp, pXerDKp) with dimers of pKS492, a recombinant clone including cer. Figure 4 shows a comparison of resolution of dimers of the recombinant clones pES (mwr) and pKS492 (cer). While resolution of pES dimers was inefficient, no pKS492 dimers were detected, indicating that XerCKp mediates recombination at high levels when cer was the target site.
Xer recombination at dif is important to resolve chromosome dimers before cytokinesis. To determine whether XerCKp also mediates poor levels of recombination when dif is the target, we transformed E. coli BS52 with pXerCKp. This xerC-deficient E. coli strain carries a KAN resistance gene between two directly repeated dif copies in its chromosome. Therefore, the fraction of cells resistant to KAN gives an indication of the level of site-specific recombination at dif. All colonies tested had lost the KAN resistance gene after overnight culture of E. coli BS52(pXerCKp), suggesting that XerCKp is fully active when dif is the recombination target.
To compare the XerCKp and XerCEc efficiencies to resolve chromosome dimers, we carried out coculturing assays, a technique used for comparing the fitness of strains. It has been shown previously that efficient chromosome dimer resolution by Xer recombination results in a growth advantage with respect to strains impaired in that function, as detected in coculturing assays (13, 33). Table 2 shows coculturing of E. coli DS981(pXerCKp) or E. coli DS981(pXerCEc) with the wild-type E. coli DS941. The growth competition with E. coli DS941 showed no differences between the E. coli DS981 strains harboring either XerCKp or XerCEc. Conversely, coculturing E. coli DS981(pACYC184) and E. coli DS941 showed that the E. coli strain harboring the cloning vector pACYC184 was rapidly outcompeted (Table 2).
These results indicate that lower XerCKp activity compared to that of XerCEc was detectable only when mwr was the target for recombination. A comparison between the two XerC proteins acting on cer or dif showed no difference in levels of activity. Therefore, we conclude that XerCKp must be an efficient enzyme for resolving chromosome dimers in its natural host and for supporting stable maintenance of plasmids with a suitable target sequence such as ColE1.
Identification of the XerCKp region responsible for lower recombination activity when mwr is the target site. Most differences in the XerCKp and XerCEc amino acid sequences occur within the N-terminal region (Fig. 3b). To determine whether amino acids within the region encompassing the second half of the predicted -helix B and the beginning of the predicted -helix C are responsible for the differences observed between XerCKp and XerCEc, we generated mutant derivatives substituting for groups of amino acids in XerCKp with those present at the equivalent position in XerCEc. The three substitution mutants, ALAD40-43NFAS, DEA43-45SEN, and AAQ55-57VTM, were tested in pES dimer resolution assays. Figure 5a shows that both ALAD40-43NFAS and DEA43-45SEN were able to mediate complete dimer resolution, while the level of resolution in the presence of derivative AAQ55-57VTM was similar to that observed with the wild-type XerCKp. To identify the amino acid(s) responsible for the variation in levels of activity, we generated single-amino-acid substitutions within the stretch encompassed by residues 40 to 45. Only one of these substitutions, L41F, was able to mediate complete resolution (Fig. 5a). Two of the mutants, A40N and D43S, did not show significant difference from the wild-type XerCKp, and one of them, A45N, showed a substandard activity (Fig. 5a). These results indicate that this region plays an important role in the levels of XerC activity when mwr is the recombination target. Although the single-amino-acid substitution L41F could elevate the activity of the enzyme, the same change could be obtained when both D43 and Q45 were substituted for by those amino acids present in XerCEc. If L41 were the sole amino acid responsible for the difference in phenotypes, one would expect a reduced activity when the reverse mutation occurs in XerCEc. Figure 5b shows that this is not the case and the XerCEc substitution F39L still shows full or nearly full activity. A certain degree of modification of levels of dimer resolution at mwr mediated by the mutant proteins could also be due to changes in expression levels or stability of the derivatives.
The results shown in this section suggest that amino acids within the region encompassing the second half of the predicted -helix B and the beginning of the predicted -helix C are responsible for the lower XerCKp activity in site-specific recombination at mwr.
DISCUSSION
The multiresistance plasmid pJHCMW1, originally isolated from a clinical K. pneumoniae strain, includes the Xer recombination site mwr (37, 49, 54). Our previous work showed an inverse relationship between osmolarity of the growth medium and efficiency of resolution by Xer recombination at mwr in E. coli strains (34). Although we still do not understand how osmotically regulated recombination at mwr may help the stability of pJHCMW1, one can speculate that the mwr locus is not critical for stability under certain environments in which resolution occurs via the resolvase at res, but it is essential in others where the osmolarity is lower.
The XerDKp, ArgRKp, and PepAKp proteins behaved identically to those from E. coli with respect to regulation by the osmolarity of the milieu and levels of dimer resolution at mwr. The lower XerCKp activity on mwr could be due to an intrinsic property of this enzyme, such as improper binding or improper formation of the synaptic complex, or to improper interaction with XerDEc resulting in poor XerCKp activation. Our results suggest that improper interaction of XerCKP with the heterologous XerDEc is not the main cause of lower recombination efficiency at mwr. The levels of activity of XerCKp were identical in the presence of XerDEc or XerDKp. These results are in agreement with the fact that major XerC-XerD interactions occur at the C termini, and, as shown in Fig. 1b and 3b (amino acids in gray boxes), the XerC and XerD proteins from both bacteria are identical at these regions. In vitro recombination experiments in the presence of purified recombinases and accessory proteins will be necessary to confirm that there are not differences in activation of XerCKp by XerDKp or XerDEc.
Our results indicate that the lower efficiency of recombination promoted by XerCKp as compared to that of XerCEc is specifically detected when mwr is the target site. It is known that the XerC N-terminal domain contacts the inner nucleotides of the binding site, which are important for recombination activity (8, 22, 23). This domain of the recombinases, although dispensable for catalytic activity, plays a role in controlling strand exchange by the recombinases (22, 26). Inspection of the nucleotide sequences of all three core recombination sites, mwr, cer, and dif, shows a higher variability at the inner portion of the XerC binding site (Fig. 6). Also, inspection of the amino acid sequence of the N terminus of XerCKp shows the highest divergence compared to that of XerCEc, and the mutagenesis experiments described in this work indicate that the region encompassed between amino acids 41 and 45 is responsible for promoting lower levels of recombination at mwr. We speculate that, even under the best conditions, i.e., low osmolarity of the growth medium, XerCKp does not bind optimally to the inner nucleotides of the mwr XerC binding site. This could be due to the specific conformation of the synaptic complex determined by the mwr accessory sequences and core recombination site. The presence of an imperfect ARG box in the mwr accessory sequences (34) may lead to a suboptimal synaptic complex to which XerCKp cannot bind properly. Mutagenesis of the mwr core recombination site as well as in vitro recombination experiments will confirm if this is the case.
We have shown previously that levels of dimer resolution comparable to those observed in the presence of XerCKp in this work are not sufficient to confer plasmid stability (49). Furthermore, our previous assays indicated that pJHCMW1 stability by multimer resolution occurs through site-specific recombination mediated by the Tn1331 transposon resolvase at the res site when the host cells are cultured in L broth (49). These findings may have interesting implications with respect to the host range of plasmids and dissemination of antibiotic resistance genes. Possessing a target for dimer resolution that is not well recognized as a substrate by a given host's Xer recombination system would make an otherwise stable plasmid unstable, thereby narrowing its host range. On the other hand, since the other essential plasmid inheritance functions may operate properly in the host, acquisition of a dimer resolution system, such as insertion of a replicative transposon like Tn1331, would allow the plasmid to be stably maintained. Inclusion of the transposon, which usually carries antibiotic resistance genes, within the plasmid would result in widening of the plasmid's host range and consequently expanding the number of genera or species harboring the resistance genes. In the case of pJHCMW1, an original plasmid unstable in K. pneumoniae could have acquired Tn1331, a process that resulted in an expansion of the host range to include this pathogen. Interestingly, a Salmonella enterica plasmid, pFPBT1, has been recently shown to include a replication region nearly identical to pJHCMW1, a site highly related to mwr, and a replicative transposon (31). It will be of great interest to determine the efficiency of the S. enterica Xer recombination proteins to resolve dimers including the pFPBT1 Xer recombination site. To our best knowledge, there have not been reports of plasmids consisting of the pJHCMW1 replicon without a replicative transposon. Another interesting possibility is that a plasmid's Xer resolution site could undergo detrimental mutations after acquisition of a replicative transposon. In these cases, the loss of an efficient dimer resolution site would result in a dependency on the transposon for plasmid viability. Loss of the transposon, and concomitantly of resistance genes, by deletion, recombination, or other mechanisms would lead to loss of the plasmid.
ACKNOWLEDGMENTS
This work was funded by Public Health grants AI47115-02 (M.E.T., National Institute of Allergy and Infectious Diseases, National Institutes of Health), LA Basin Minority Health and Health Disparities International Research Training Program (MHIRT) 9T37MD001368-08 (National Center on Minority Health and Health Disparities), and a grant from Wellcome Trust (D.J.S.). D.B. and J.N. were supported by grants R25 GM56820-03 (MSD), MIRT 5T37TW000048-07, and MHIRT 9T37MD001368-08 from the National Institutes of Health. J.R. was supported by grant MIRT 5T37 TW000048-07 from the National Institutes of Health.
We thank Migena Bregu and Christophe Possoz for useful comments and suggestions and Rachel Baker for expert technical assistance. We are also indebted to The Genome Sequencing Center at Washington University for making the K. pneumoniae genome sequence available prior to publication.
REFERENCES
Alen, C., D. J. Sherratt, and S. D. Colloms. 1997. Direct interaction of aminopeptidase A with recombination site DNA in Xer site-specific recombination. EMBO J. 16:5188-5197.
Altschul, S. F., T. L. Madden, A. A. Schaffer, J. Zhang, Z. Zhang, W. Miller, and D. J. Lipman. 1997. Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res. 25:3389-3402.
Arciszewska, L., and D. Sherratt. 1995. Xer site-specific recombination in vitro. EMBO J. 14:2112-2120.
Arciszewska, L. K., R. A. Baker, B. Hallet, and D. J. Sherratt. 2000. Coordinated control of XerC and XerD catalytic activities during Holliday junction resolution. J. Mol. Biol. 299:391-403.
Aussel, L., F. Barre, M. Aroyo, A. Stasiak, A. Z. Stasiak, and D. Sherratt. 2002. FtsK is a DNA motor protein that activates chromosome dimer resolution by switching the catalytic state of the XerC and XerD recombinases. Cell 108:195-205.
Barre, F. X., M. Aroyo, S. Colloms, A. Helfrich, F. Cornet, and D. J. Sherratt. 2000. FtsK functions in the processing of a Holliday junction intermediate during bacterial chromosome segregation. Genes Dev. 14:2976-2988.
Barre, F.-X., and D. J. Sherratt. 2002. Xer site-specific recombination: promoting chromosome segregation, p. 149-161. In N. Craig, R. Craigie, M. Gellert, and A. M. Lambowitz (ed.), Mobile DNA II, vol. 1. ASM Press, Washington, D.C.
Blake, J. A., N. Ganguly, and D. J. Sherratt. 1997. DNA sequence of recombinase-binding sites can determine Xer site-specific recombination outcome. Mol. Microbiol. 23:387-398.
Blakely, G., S. Colloms, G. May, M. Burke, and D. J. Sherratt. 1991. Escherichia coli XerC recombinase is required for chromosomal segregation at cell division. New Biol. 3:789-798.
Blakely, G., A. Davidson, and D. J. Sherratt. 2000. Sequential strand exchange by XerC and XerD during site-specific recombination at dif. J. Biol. Chem. 275:9930-9936.
Blakely, G., G. May, R. McCulloch, L. Arciszewska, M. Burke, S. Lovett, and D. J. Sherratt. 1993. Two related recombinases are required for site specific recombination at dif and cer in E. coli K12. Cell 75:351-361.
Bregu, M., D. J. Sherratt, and S. D. Colloms. 2002. Accessory factors determine the order of strand exchange in Xer recombination at psi. EMBO J. 21:3888-3897.
Capiaux, H., C. Lesterlin, K. Perals, J. Louarn, and F. Cornet. 2002. A dual role for the FtsK protein in Escherichia coli chromosome segregation. EMBO Rep. 3:532-536.
Chen, Y., U. Narendra, L. E. Iype, M. M. Cox, and P. A. Rice. 2000. Crystal structure of a Flp recombinase-Holliday junction complex: assembly of an active oligomer by helix swapping. Mol. Cell 6:885-897.
Clerget, M. 1991. Site-specific recombination promoted by a short DNA segment of plasmid R1 and by a homologous segment in the terminus region of the Escherichia coli chromosome. New Biol. 3:780-788.
Cohen, S. N., A. C. Chang, and L. Hsu. 1972. Nonchromosomal antibiotic resistance in bacteria: genetic transformation of Escherichia coli by R-factor DNA. Proc. Natl. Acad. Sci. USA 69:2110-2114.
Colloms, S., C. Alen, and D. J. Sherratt. 1998. The ArcA/ArcB two-component regulatory system of Escherichia coli is essential for Xer site-specific recombination at psi. Mol. Microbiol. 28:521-530.
Colloms, S., R. McCulloch, K. Grant, L. Neilson, and D. J. Sherratt. 1996. Xer-mediated site-specific recombination in vitro. EMBO J. 15:1172-1181.
Colloms, S. D., P. Sykora, G. Szatmari, and D. J. Sherratt. 1990. Recombination at ColE1 cer requires the Escherichia coli xerC gene product, a member of the lambda integrase family of site-specific recombinases. J. Bacteriol. 172:6973-6980.
Cornet, F., I. Mortier, J. Patte, and J.-M. Louarn. 1994. Plasmid pSC101 harbors a recombination site, psi, which is able to resolve plasmid multimers and to substitute for the analogous chromosomal Escherichia coli site dif. J. Bacteriol. 176:3188-3195.
Ferreira, H., B. Butler-Cole, A. Burgin, R. Baker, D. J. Sherratt, and L. K. Arciszewska. 2003. Functional analysis of the C-terminal domains of the site-specific recombinases XerC and XerD. J. Mol. Biol. 330:15-27.
Ferreira, H., D. J. Sherratt, and L. Arciszewska. 2001. Switching catalytic activity in the XerCD site-specific recombination machine. J. Mol. Biol. 312:45-57.
Guo, F., D. N. Gopaul, and G. D. van Duyne. 1997. Structure of Cre recombinase complexed with DNA in a site-specific recombination synapse. Nature 389:40-46.
Hallet, B., L. K. Arciszewska, and D. J. Sherratt. 1999. Reciprocal control of catalysis by the tyrosine recombinases XerC and XerD: an enzymatic switch in site-specific recombination. Mol. Cell 4:949-959.
Hallet, B., V. Vanhooff, and F. Cornet. 2004. DNA site-specific resolution systems, p. 145-180. In B. Funnell and G. Phillips (ed.), Plasmid biology. ASM Press, Washington, D.C.
Hayes, F., and D. J. Sherratt. 1997. Recombinase binding specificity at the chromosome dimer resolution site dif of Escherichia coli. J. Mol. Biol. 266:525-537.
Hodgman, T. C., H. Griffiths, and D. K. Summers. 1998. Nucleoprotein architecture and ColE1 dimer resolution: a hypothesis. Mol. Microbiol. 29:545-558.
Huber, K., and M. Waldor. 2002. Filamentous phage integration requires the host recombinases XerC and XerD. Nature 417:656-659.
Massey, T. H., L. Aussel, F. X. Barre, and D. J. Sherratt. 2004. Asymmetric activation of Xer site-specific recombination by FtsK. EMBO Rep. 5:399-404.
McCulloch, R., L. Coggins, S. Colloms, and D. J. Sherratt. 1994. Xer-mediated site-specific recombination at cer generates Holliday junctions in vivo. EMBO J. 13:1844-1855.
Pasquali, F., C. Kehrenberg, G. Manfreda, and S. Schwarz. 2005. Physical linkage of Tn3 and part of Tn1721 in a tetracycline and ampicillin resistance plasmid from Salmonella Typhimurium. J. Antimicrob. Chemother. 55:562-565.
Perals, K., H. Capiaux, J. B. Vincourt, J. M. Louarn, D. J. Sherratt, and F. Cornet. 2001. Interplay between recombination, cell division and chromosome structure during chromosome dimer resolution in Escherichia coli. Mol. Microbiol. 39:904-913.
Perals, K., F. Cornet, Y. Merlet, I. Delon, and J. M. Louarn. 2000. Functional polarization of the Escherichia coli chromosome terminus: the dif site acts in chromosome dimer resolution only when located between long stretches of opposite polarity. Mol. Microbiol. 36:33-43.
Pham, H., K. J. Dery, D. J. Sherratt, and M. E. Tolmasky. 2002. Osmoregulation of dimer resolution at the plasmid pJHCMW1 mwr locus by Escherichia coli XerCD recombination. J. Bacteriol. 184:1607-1616.
Rijns, M., Y. Lu, S. Leach, and S. Colloms. 2005. Mutagenesis of PepA suggests a new model for the Xer/cer synaptic complex. Mol. Microbiol. 57:927-941.
Rost, B., and J. Liu. 2003. The PredictProtein Server. Nucleic Acids Res. 31:3300-3304.
Sarno, R., G. McGillivary, D. J. Sherratt, L. A. Actis, and M. E. Tolmasky. 2002. Complete nucleotide sequence of Klebsiella pneumoniae multiresistance plasmid pJHCMW1. Antimicrob. Agents Chemother. 46:3422-3427.
Sherratt, D. J., L. Arciszewska, G. Blakely, S. Colloms, K. Grant, N. Leslie, and R. McCulloch. 1995. Site-specific recombination and circular chromosome segregation. Phil. Trans. R. Soc. London B 347:37-42.
Sherratt, D. J., B. Soballe, F. X. Barre, S. Filipe, I. Lau, T. Massey, and J. Yates. 2004. Recombination and chromosome segregation. Phil. Trans. R. Soc. Lond. B 359:61-69.
Spiers, A. J., and D. J. Sherratt. 1999. C-terminal interactions between the XerC and XerD site-specific recombinases. Mol. Microbiol. 32:1031-1042.
Spiers, A. J., and D. J. Sherratt. 1997. Relating primary structure to function in the Escherichia coli XerD site-specific recombinase. Mol. Microbiol. 24:1071-1082.
Stirling, C. J., S. D. Colloms, J. F. Collins, G. Szatmari, and D. J. Sherratt. 1989. xerB, an Escherichia coli gene required for plasmid ColE1 site-specific recombination, is identical to pepA, encoding aminopeptidase A, a protein with substantial similarity to bovine lens leucine aminopeptidase. EMBO J. 8:1623-1627.
Stirling, C. J., G. Szatmari, G. Stewart, M. C. Smith, and D. J. Sherratt. 1988. The arginine repressor is essential for plasmid-stabilizing site-specific recombination at the ColE1 cer locus. EMBO J. 7:4389-4395.
Subramanya, H. S., L. K. Arciszewska, R. A. Baker, L. E. Bird, D. J. Sherratt, and D. B. Wigley. 1997. Crystal structure of the site-specific recombinase, XerD. EMBO J. 16:5178-5187.
Summers, D., C. Beton, and H. Withers. 1993. Multicopy plasmid instability: the dimer catastrophe hypothesis. Mol. Microbiol. 8:1031-1038.
Summers, D., and D. J. Sherratt. 1984. Multimerization of high copy number plasmids causes instability: CoIE1 encodes a determinant essential for plasmid monomerization and stability. Cell 36:1097-1103.
Summers, D. K., and D. J. Sherratt. 1988. Resolution of ColE1 dimers requires a DNA sequence implicated in the three-dimensional organization of the cer site. EMBO J. 7:851-858.
Tolmasky, M. E. 1990. Sequencing and expression of aadA, bla, and tnpR from the multiresistance transposon Tn1331. Plasmid 24:218-226.
Tolmasky, M. E., S. Colloms, G. Blakely, and D. J. Sherratt. 2000. Stability by multimer resolution of pJHCMW1 is due to the Tn1331 resolvase and not to the Escherichia coli Xer system. Microbiology 146:581-589.
Tolmasky, M. E., and J. H. Crosa. 1993. Genetic organization of antibiotic resistance genes (aac(6')-Ib, aadA, and oxa9) in the multiresistance transposon Tn1331. Plasmid 29:31-40.
Tolmasky, M. E., and J. H. Crosa. 1987. Tn1331, a novel multiresistance transposon encoding resistance to amikacin and ampicillin in Klebsiella pneumoniae. Antimicrob. Agents Chemother. 31:1955-1960.
Val, M., M. Bouvier, J. Campos, D. J. Sherratt, F. Cornet, D. Mazel, and F. Barre. 2005. The single stranded genome of phage CTX is the form used for integration into the genome of Vibrio cholerae. Mol. Cell 19:559-566.
Wegrzyn, G. 2005. What does "plasmid biology" currently mean Summary of the Plasmid Biology 2004 Meeting. Plasmid 53:14-22.
Woloj, M., M. E. Tolmasky, M. C. Roberts, and J. H. Crosa. 1986. Plasmid-encoded amikacin resistance in multiresistant strains of Klebsiella pneumoniae isolated from neonates with meningitis. Antimicrob. Agents Chemother. 29:315-319.
Yates, J., M. Aroyo, D. J. Sherratt, and F. X. Barre. 2003. Species specificity in the activation of Xer recombination at dif by FtsK. Mol. Microbiol. 49:241-249.(Duyen Bui, Judianne Ramis)