Site-specific and illegitimate recombination in the oriV1 region of the F factor. DNA sequences involved in recombination and resolution

Mobilization of the nonconjugative virulence plasmid from hypervirulent Klebsiella pneumoniae

Background

Klebsiella pneumoniae, as a global priority pathogen, is well known for its capability of acquiring mobile genetic elements that carry resistance and/or virulence genes. Its virulence plasmid, previously deemed nonconjugative and restricted within hypervirulent K. pneumoniae (hvKP), has disseminated into classic K. pneumoniae (cKP), particularly carbapenem-resistant K. pneumoniae (CRKP), which poses alarming challenges to public health. However, the mechanism underlying its transfer from hvKP to CRKP is unclear.

Methods

A total of 28 sequence type (ST) 11 bloodstream infection-causing CRKP strains were collected from Ruijin Hospital in Shanghai, China, and used as recipients in conjugation assays. Transconjugants obtained from conjugation assays were confirmed by XbaI and S1 nuclease pulsed-field gel electrophoresis, PCR detection and/or whole-genome sequencing. The plasmid stability of the transconjugants was evaluated by serial culture. Genetically modified strains and constructed mimic virulence plasmids were employed to investigate the mechanisms underlying mobilization. The level of extracellular polysaccharides was measured by mucoviscosity assays and uronic acid quantification. An in silico analysis of 2608 plasmids derived from 814 completely sequenced K. pneumoniae strains available in GenBank was performed to investigate the distribution of putative helper plasmids and mobilizable virulence plasmids.

Results

A nonconjugative virulence plasmid was mobilized by the conjugative plasmid belonging to incompatibility group F (IncF) from the hvKP strain into ST11 CRKP strains under low extracellular polysaccharide-producing conditions or by employing intermediate E. coli strains. The virulence plasmid was mobilized via four modes: transfer alone, cotransfer with the conjugative IncF plasmid, hybrid plasmid formation due to two rounds of single-strand exchanges at specific 28-bp fusion sites or homologous recombination. According to the in silico analysis, 31.8% (242) of the putative helper plasmids and 98.8% (84/85) of the virulence plasmids carry the 28-bp fusion site. All virulence plasmids carry the origin of the transfer site.

Conclusions

The nonconjugative virulence plasmid in ST11 CRKP strains is putatively mobilized from hvKP or E. coli intermediates with the help of conjugative IncF plasmids. Our findings emphasize the importance of raising public awareness of the rapid dissemination of virulence plasmids and the consistent emergence of hypervirulent carbapenem-resistant K. pneumoniae (hv-CRKP) strains.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13073-021-00936-5.

Identification of oriT and a recombination hot spot in the IncA/C plasmid backbone

Dissemination of multiresistance has been accelerating among pathogenic bacteria in recent decades. The broad host-range conjugative plasmids of the IncA/C family are effective vehicles of resistance determinants in Gram-negative bacteria. Although more than 150 family members have been sequenced to date, their conjugation system and other functions encoded by the conserved plasmid backbone have been poorly characterized. The key cis-acting locus, the origin of transfer (oriT), has not yet been unambiguously identified. We present evidence that IncA/C plasmids have a single oriT locus immediately upstream of the mobI gene encoding an indispensable transfer factor. The fully active oriT spans ca. 150-bp AT-rich region overlapping the promoters of mobI and contains multiple inverted and direct repeats. Within this region, the core domain of oriT with reduced but detectable transfer activity was confined to a 70-bp segment containing two inverted repeats and one copy of a 14-bp direct repeat. In addition to oriT, a second locus consisting of a 14-bp imperfect inverted repeat was also identified, which mimicked the function of oriT but which was found to be a recombination site. Recombination between two identical copies of these sites is RecA-independent, requires a plasmid-encoded recombinase and resembles the functioning of dimer-resolution systems.

Concerted action of plasmid maintenance functions: partition complexes create a requirement for dimer resolution

Partition of prokaryotic DNA requires formation of specific protein-centromere complexes, but an excess of the protein can disrupt segregation. The mechanisms underlying this destabilization are unknown. We have found that destabilization by the F plasmid partition protein, SopB, of plasmids carrying the F centromere, sopC, results from the capacity of the SopB-sopC partition complex to stimulate plasmid multimerization. Mutant SopBs unable to destabilize failed to increase multimerization. Stability of wild-type mini-F, whose ResD/rfsF site-specific recombination system enables it to resolve multimers to monomers, was barely affected by excess SopB. Destabilization of plasmids lacking the rfsF site was suppressed by recF, recO and recR, but not by recB, mutant alleles, indicating that multimerization is initiated from single-strand gaps. SopB did not alter the amounts or distribution of replication intermediates, implying that SopB-DNA complexes do not create single-strand gaps by blocking replication forks. Rather, the results are consistent with SopB-DNA complexes channelling gapped molecules into the RecFOR recombination pathway. We suggest that extended SopB-DNA complexes increase the likelihood of recombination between sibling plasmids by keeping them in close contact prior to SopA-mediated segregation. These results cast plasmid site-specific resolution in a new role - compensation for untoward consequences of partition complex formation.

Purification and DNA binding of the D protein, a putative resolvase of the F-factor of Escherichia coli

The D protein encoded by plasmid mini-F promotes resolution of plasmid cointegrates or dimers of the F-factor or mini-F. In addition, two rfsF sequences are essential for this site-specific, recA-independent recombination event. The D gene was cloned into an expression vector and the gene product was overproduced in Escherichia coli and purified to homogeneity. The sequence of the N-terminus of the D protein was determined, thus permitting identification of the correct translational start codon in the nucleotide sequence that results in a 29.6 kDa protein. The binding site for the purified D protein is located within the mini-F NcoI-HpaI DNA fragment (192 bp). Binding seems to be affected by DNA methylation, since the protein did not bind to DNA isolated from a dam mutant of E. coli. The binding site, which is a region of approximately 28 bp and is located 160 bp downstream of the rfsF site, was identified by DNase I footprinting using fluorescence labelled DNA.

The replication origin of the Methylomonas clara plasmid pBE-2

The Methylomonas clara narrow host range plasmids pBE-2 and pBE-3 belong to the class of plasmids encoding a trans acting replication initiation factor. Characteristically for such plasmids, the sequence of the origin of pBE-2 and pBE-3 contains a number of large direct repeats (8 and a half iterons of 19 bp), which by analogy are putative binding sites of the trans acting replication factor. Several additional features typical for the majority of E. coli plasmids were found in the M. clara origin: These include sequences homologous to the E. coli DnaA-box, sequences resembling E. coli IHF binding-sites, an AT-rich region with short repeats (similar to those repeats of E. coli origins responsible for an initial DNA duplex opening), and an AT-rich bent DNA region containing inverted repeats which have homology to small repeated sequences found in several plasmid origins. In addition, in the M. clara plasmid origin, large potential hairpin structures are present and the sequence of one of these participates in site specific recombination.

Identification of a multimer resolution system involved in stabilization of the Salmonella dublin virulence plasmid pSDL2

The Salmonella dublin virulence plasmid pSDL2 is a low-copy-number plasmid that is highly conserved in its host. Deletion of the 8-kb EcoRI C fragment downstream of the virulence region leads to plasmid instability and formation of multimers. We identified a multimer resolution system in the EcoRI C fragment composed of a trans-acting resolvase gene and a cis-acting resolution site. The resolvase gene, rsd, maps within a 2-kb EcoRV fragment and appears to be part of a multicistronic unit together with at least two other genes of unknown function. The derived protein, 28.7-kDa in size, is almost identical to the D protein of miniF. The C-terminal region was shown to have substantial similarity to the conserved C-terminal domains of the site-specific recombinases of the integrase family. The cis-acting resolution site, crs, is located upstream of rsd within a 628-bp SmaI-HpaI fragment. It contains eight direct incomplete 17-bp repeats followed by a segment rich in indirect repeats, the latter being homologous to the oriV1 sequence of miniF. crs contains the crossover site for specific recombination and mediates bidirectional promoter activity. A replicative function in analogy to that of oriV1 of F could not be demonstrated. The multimer resolution system was shown to stabilize pACYC184 and is dependent on the recA-mediated formation of multimeric plasmids. Screening different Salmonella serovars with a pSDL2-specific recombination assay revealed that only strains harboring a virulence plasmid encode for resolvase activity. Our results suggest that site-specific recombination contributes to the stable inheritance of pSDL2 and other Salmonella virulence plasmids.

The 41 carboxy-terminal residues of the miniF plasmid CcdA protein are sufficient to antagonize the killer activity of the CcdB protein

The ccd operon of plasmid F encodes two genes, ccdA and ccdB, which contribute to the high stability of the plasmid by post-segregational killing of plasmid-free bacteria. The CcdB protein is lethal to bacteria and the CcdA protein is an antagonist of this lethal action. A 520 bp fragment containing the terminal part of the ccdA gene and the entire ccdB gene of plasmid F was cloned downstream of the tac promoter. Although the CcdB protein was expressed from this fragment, no killing of host bacteria was observed. We found that the absence of killing was due to the presence of a small polypeptide, CcdA41, composed of the 41 C-terminal residues of the CcdA protein. This polypeptide has retained the ability to regulate negatively the lethal activity of the CcdB protein.

Linear multimer formation of plasmid DNA in Escherichia coli hopE (recD) mutants

The hopE mutants of Escherichia coli, which cannot stably maintain a mini-F plasmid during cell division, have mutations in the recD gene coding for subunit D of the RecBCD enzyme (exonuclease V). A large amount of linear multimer DNA of mini-F and pBR322 plasmid accumulates in these hopE mutants. The linear multimers of plasmid DNA in the hopE (recD) mutants accumulate in sbc+ genetic backgrounds and this depends on the recA+ gene function. Linear plasmid multimers also accumulated in a recBC xthA triple mutant, but not an isogenic xth A mutant or an isogenic recBC mutant. The recBC xth A mutant is defective in the conjugative type of recombination. Linear plasmid multimers were not detected in the recBC strain. We propose models to account for linear multimer formation of plasmids in various mutants.

Involvement of DnaK protein in mini-F plasmid replication: temperature-sensitive seg mutations are located in the dnaK gene

The seg mutants (seg-1 and seg-2) of Escherichia coli cannot support the replication of the F factor and mini-F plasmids at 42 degrees C. We cloned the wild-type E. coli chromosomal DNA fragment complementing the seg-1 and seg-2 mutations and found that both mutations were complemented by the wild-type dnaK gene coding for a heat shock protein. Transduction with phage P1 indicated that the seg-2 mutation is located at about 0.3 min in the region containing the dnaK gene in the order trpR--thrA--seg-2--leuB, consistent with the locus of the dnaK gene. Cloning and sequencing of the dnaK gene of the seg mutants showed that there was one base substitution within the dnaK gene in each mutant causing an amino acid substitution. These results indicate that the seg gene in which the seg-1 and seg-2 mutations occurred is identical to the dnaK gene. The mini-F plasmid pXX325 did not transform a dnaK null mutant to ampicillin resistance at 30 degrees C in contrast to plasmids pBR322, pACYC184 and pSC101, which did. The active dnaK (seg) gene product is therefore essential for replication of the mini-F plasmid at both 30 degrees and 42 degrees C.

Determinants of site-specific recombination in the lambdoid coliphage HK022. An evolutionary change in specificity

The temperate bacteriophage HK022, like its relative lambda, inserts its chromosome into a specific site in the bacterial chromosome during lysogenization and excises it after induction. However, we find that the recombinational specificities of the two phages differ: they use different bacterial sites, and neither promotes efficient insertion or excision of the other phage chromosome. In order to determine the basis for this difference in specificity, we sequenced the HK022 elements that are involved in insertion and excision, and compared them to the corresponding lambda elements. The location, orientation, size and overall arrangement of the int and xis genes and the phage attachment sites are nearly identical in the two genomes, as is common for other functionally related elements in lambdoid phages. The Xis proteins of the two phages are functionally interchangeable, and their predicted amino acid sequences differ by but one residue. In contrast, the two Int proteins are not functionally interchangeable, and their sequences, although similar, differ at many positions. These sequence differences are not uniformly distributed: the amino-terminal 55 residues are completely conserved, but the remaining 302 show a pattern of differences interspersed with identities and conservative changes. These findings imply that the specificity difference between HK022 and lambda site-specific recombination is a consequence of the inability of the respective Int proteins to recognize pairs of heterologous attachment sites. The two phage attachment sites are remarkably similar, especially the two "arm" segments, which in lambda contain binding sites for Int, Xis and integration host factor. They are less similar in the segment between the two arms, which in lambda contains the points of recombinational strand exchange and a second class of binding site for Int protein (the "core-type" sites). The two bacterial attachment sites are quite different, although both have a short stretch of perfect homology with their respective phage partners at the points of strand exchange. We propose that the two Int proteins recognize similar or identical sites in the arms of their cognate attachment sites, and that differences in binding or action at the core-type sites is responsible for the divergent specificities. Genetic experiments and sequence comparisons suggest that both proteins recognize different but overlapping families of core-type sites, and that divergence in specificity has been achieved by an alternating succession of small, mutually compatible changes in protein and site.

Construction of large DNA segments in Escherichia coli

Recombinant DNA clones containing large pieces of DNA are useful in the study of large genetic units, but these are difficult to make in most bacterial cloning vectors. A strategy is described that uses general and site-specific recombination to construct large pieces of eukaryotic DNA from smaller cloned segments. The large clones are propagated on F factor-based plasmids in Escherichia coli. They can be easily modified to introduce mutations or rearrangements. These techniques were applied to the construction of large DNA segments from the bithorax complex of Drosophila.

Transposon Tn7. cis-Acting sequences in transposition and transposition immunity

We have identified and characterized the cis-acting sequences at the termini of the bacterial transposon Tn7 that are necessary for its transposition. Tn7 participates in two kinds of transposition event: high-frequency transposition to a specific target site (attTn7) and low-frequency transposition to apparently random target sites. Our analyses suggest that the same sequences at the Tn7 ends are required for both transposition events. These sequences differ in length and nucleotide structure: about 150 base-pairs at the left end (Tn7L) and about 70 base-pairs at the right end (Tn7R) are necessary for efficient transposition. We also show that the ends of Tn7 are functionally distinct: a miniTn7 element containing two Tn7R ends is active in transposition but an element containing two Tn7L ends is not. We also report that the presence of Tn7's cis-acting transposition sequences anywhere in a target replicon inhibits subsequent insertion of another copy of Tn7 into either an attTn7 target site or into random target sites. The inhibition to an attTn7 target site is most pronounced when the Tn7 ends are immediately adjacent to attTn7. We also show that the presence of Tn7R's cis-acting transposition sequences in a target replicon is necessary and sufficient to inhibit subsequent Tn7 insertion into the target replicon.

Nucleotide sequence and replication characteristics of RepFIB, a basic replicon of IncF plasmids

A second autonomous replicon of P307, RepFIB, has been isolated that has significant homology with other replicons in IncFI group plasmids. Eleven homologous repeats of 21 base pairs are present on the sequence and flank an open reading frame capable of coding for a protein of about Mr = 40,000. This protein was identified by maxicell analysis of cloned RepFIB. A series of deletion mutations of RepFIB were inserted into a DNA polymerase I-dependent vector and examined for their replication proficiency in a polA1 strain. These experiments defined a minimal replication region of 1.6 kilobases which includes the three repeats immediately upstream and downstream of the open reading frame. Deletion of a second set of repeats further downstream doubled the copy number of a chimeric plasmid replicating under RepFIB control. It was concluded that these repeats control the copy number of the replicon. Incompatibility tests showed that all three sets of repeats could express incompatibility with a resident RepFIB plasmid.

Identification of plasmid and Bacillus subtilis chromosomal recombination sites used for pE194 integration

The plasmid pE194 (3.7 kilobases) is capable of integrating into the genome of the bacterial host Bacillus subtilis in the absence of the major homology-dependent RecE recombination system. Multiple recombination sites have been identified on both the B. subtilis chromosome and pE194 (J. Hofemeister, M. Israeli-Reches, and D. Dubnau, Mol. Gen. Genet. 189:58-68, 1983). The B. subtilis chromosomal recombination sites were recovered by genetic cloning, and these sites were studied by nucleotide sequence analysis. Recombination had occurred between regions of short nucleotide homology (6 to 14 base pairs) as indicated by comparison of the plasmid and the host chromosome recombination sites with the crossover sites of the integration products. Recombination between the homologous sequences of the plasmid and the B. subtilis genome produced an integrated pE194 molecule which was bounded by direct repeats of the short homology. These results suggest a recombination model involving a conservative, reciprocal strand exchange between the two recombination sites. A preferred plasmid recombination site was found to occur within a 70-base-pair region which contains a GC-rich dyad symmetry element. Five of seven pE194-integrated strains analyzed had been produced by recombination at different locations within this 70-base-pair interval, located between positions 860 and 930 in pE194. On the basis of these data, mechanisms are discussed to explain the recombinational integration of pE194.

vir-induced recombination in Agrobacterium. Physical characterization of precise and imprecise T-circle formation

Induction of Ti plasmid virulence (vir) gene expression during the early stages of plant cell transformation by Agrobacterium tumefaciens initiates the generation of several T-DNA-associated molecular events: (1) site-specific nicks at T-DNA border sequences (border nicks); (2) free, unipolar, linear, single-stranded T-DNA copies (T-strands); and (3) double-stranded, circular T-DNA molecules (T-circles). The first two T-DNA products have been detected in A. tumefaciens, while T-circles have only been detected following Escherichia coli transformation or transduction. The relationship between the three events has not been evaluated since the genesis of T-circles in A. tumefaciens has not been clarified. Evidence is presented here that T-circles are not an artefact of E. coli transformation, but are present as free, double-stranded molecules in A. tumefaciens resulting from site-specific reciprocal recombination between the left and right 25-base-pair border sequences that flank the T-DNA. Furthermore, the frequency of T-circle formation correlates with the frequency of formation of its reciprocal product, the Ti plasmid deleted in the T-DNA region. Several types of recombinant T-DNA circles arise after activation of vir gene expression, a major class representing precise site-specific recombination between both T-DNA borders, and a minor class representing recombination events either utilizing only one T-DNA border sequence and other Ti plasmid sequences, or utilizing only Ti plasmid sequences (i.e. no T-DNA borders). Nucleotide sequence analyses show that when one (nicked) border recombines with other Ti plasmid sequences, a small stretch (16 to 17 base-pairs) of local homology suffices to allow crossing over.

Mapping and regulation of the pifC promoter of the F plasmid

The pif region of the F plasmid, which causes abortive infection of Escherichia coli by T7 bacteriophage, is autogenously controlled by the product of the pifC gene. Here we describe the identification of the pif operon promoter by S1-nuclease mapping, and show that it is autoregulated at the transcriptional level and that its activity is modulated by integration host factor.