Investigation of the Genes Involved in Antigenic Switching at the vlsE Locus in Borrelia burgdorferi: An Essential Role for the RuvAB Branch Migrase

Most of the pathogens adopt the strategy of antigenic variation to evade immune responses by host. These variations occur through targeted genome rearrangements. However, the exact molecular details of the recombination processes that generate diversity in the antigen-expressing genes are not fully known.
Borrelia burgdorferi is a spirochete that causes lyme borreliosis. This disease progresses through three stages, early, disseminated and persistent. The persistent infection requires continual segmental gene conversion at the vlsE locus. This locus encodes a 35KDa membrane lipoprotein and is present on the linear plasmid Ip-28-1. When this locus is deleted or when the plasmid is not present, a productive murine infection occurs, but the organisms are cleared between 8-21days post infection. Adjacent to vlsE (also referred to as vls1), is present a contiguous upstream array of 15 silent cassettes separated from each other by 17 bp direct repeats. It has been suggested that during murine infection, information is transferred uni-directionally from the silent cassettes into the expression site to generate diversity at six regions (VR-1 to VR-6) with in the central region of the vlsE gene. VR-1 to VR-6 regions are believed to be prominently displayed antigenic areas. The antigenic diversity is generated through segmental gene conversion so that information from several silent cassettes can be transferred into the single vlsE locus to generate a mosaic gene. And thus, unique vlsE proteins can be generated. It has been shown that all silent cassettes are used as sequence donors in the gene conversion events at vlsE. The mammalian signal that triggers recombinational switching is unknown. The protein machinery that promotes recombinational switching at vlsE remains unknown. This fascinating work is from the lab of Dr. George Chaconas, Department of Microbiology and infectious disease, university of Calgary, Canada.
In this paper, the authors generated 17 mutants carrying disruptions in known DNA recombination, repair and replication genes to identify proteins involved in recombinational switching at vlsE.
The methodology used and major results obtained are as follows:

 Construction of DNA repair and replication gene disruptions in B. burgdorferi- 21 different DNA replication, repair and recombination genes were disrupted to investigate their role in vlsE recombination. For this, knockout plasmids carrying a one kb gentamicine cassette that replaced the central portion of the target gene, were constructed and used to transform the infectious B. burgdorferi B31 clone 5A4. Following transformation, allelic exchange results in successful disruption. However, recombination can also lead to integrative recombination leading to the formation of merodiploids. Thus, construct verification was done. For this, each gene disruption was subjected to four PCR assays. Firstly, all the samples were subjected to PCR for the presence of gentamicin cassette. Secondly, samples were subjected to PCR using knock out primers to confirm the absence of deleted gene. Thirdly, all the samples were subjected to PCR using target gene primers to confirm the size. Fourthly, combination of target gene primers and primers internal to gentamicin cassette were used to amplify the boundaries. Of the 21 DNA replication, repair and recombination gene knockout attempted, 17 were successful.

 Effect of gene disruption on mouse (C3h/HeN) infection- For each gene disruption, two clones were chosen, which contained the full plasmid complement required for infectivity. These were used to infect C3H/HeN immunocompetent mouse. Cultures were grown from blood samples at day 7 of infectivity. At days 14 and 21, ear biopsies were used for monitoring infection and switching at vlsE. A portion of the vlsE expression site containing the variable regions was amplified from organisms obtained from ear biopsies at day 21. The incorporation of new restriction sites from the cassettes into the variable region of vlsE provides a clear indication of the switching process. Thus, these were further subjected to RFLP analysis to study switching. The results obtained can be used to divide the 17 mutations in two groups. In the first group, >75% of blood cultures were positive at day 7 and same was observed at day 21. All these mutants also displayed switching. In the second group, the mutant strains displayed <75% positive cultures from ear biopsies at 21 days post-infection. In case of sbcD, sbcC and BBG32, switching was observed by RFLP. For the remainder of mutant strains in this group, the infection was allowed till day 35. At that time, mice were sacrificed and the spirochetes were cultivated from heart, bladder, joint and ear. The RFLP experiments demonstrated that switching could be observed in all the strains except those carrying ruvA or ruvB mutations. These genes encode for the two subunits of a Holliday junction branch migrase.

 Effect of B. burgdorferi gene disruptions on SCID C3H/HeN mouse infections: The presence of an effective immune response might exert a selective pressure on antigenic variation, thus to remove this selective pressure the mutant strains of the second group were used to infect SCID mice. This removes the pressure as well as it allows the organisms to persist in the host for longer duration. In a wild type mouse, the strains defective in switching would be cleared by day 21. However, in SCID mice, this will not happen and thus analysis of switching can be done beyond 21 days. It was observed that all the mutant strains tested displayed wild type levels of infectivity and persistence throughout 35 days of infection. This showed that mutant strains that displayed reduced infectivity at 21 days were fully competent for the infection process in mice lacking an acquired immune response.


 Analysis of switching at the vlsE by DNA sequencing of mutant strains recovered from SCID mouse infection- Since RFLP analysis is not quantitative and because of the previous observations that switching is apparently less frequent in SCID mice, the investigators analyzed switching in a limited set of mutants by sequencing. They chose the two mutants (ruvA and ruvB) that were defective in switching by RFLP analysis and two mutants (recJ and mutL) that were shown to switch by RFLP at 35 days post infection, but displayed no organisms in 21 days. In these experiments, the PCR product used for RFLP and obtained from different tissue (heart, joints, bladder, and ear) was used for cloning and sequencing. Sequencing of 10 clones from each tissue type culture was performed for each mutation. Thus, total 40 clones were sequenced for each mutation. It was observed that for wild type clones, 10 out of the 10 clones showed switching and contained sequences similar to those in silent cassettes in the heart and bladder tissue culture. While 5/10 and 8/10 showed switching in the joint and ear tissue culture, respectively. These results are similar to a previously reported data. In case of ruvA mutants, it was observed that only one of the 40 clones differed from the wild type vlsE (that is one showed switching). In case of ruvB, all 40 clones were similar to the wild type vlsE sequences (that is no switching). These results further confirm the negative switching observed in the ruvA and ruvB strains by RFLP. In case of mutL mutants, 27.5% of clones showed nucleotide changes and thus switching, while in case of recJ mutants 57.5% of mutants displayed switching. The major conclusions of the study are as follows:

a. The investigators successfully disrupted 17 of the 21 genes involved in DNA replication, repair and recombination. It remains to be elucidated whether the reaming four genes are essential for B. burgdorferi or whether the gene disruption mechanisms used in this study are ideal for these loci.
b. By creating these mutants, the authors aimed to study the possible role of genes in question on recombinational switching at the vlsE locus. However, the effect of these 17 mutations on generalized recombination and DNA repair is not known and authors are studying it.
c. It was observed by RFLP that ruvA/ruvB disruption leads to no switching. The results were further confirmed by sequencing.
d. Several of the mutants exhibited an altered infectivity phenotype in C3H/HeN mice that was not attributable to the defect in vlsE switching. When these mutants were inoculated into SCID C3H/HeN mice, their infectivity was restored and was comparable to wild type strains. The exact mechanism for such changes is not known.
e. recA is not required for recombinational switching.
f. A role of ruvAB encoded branch migarse in recombinational switching at vlsE
g. A possible role for mutL in recombinational switching at vlsE- No clear-cut involvement of mutL in recombination switching of vlsE was observed and this needs to be studied further. It may be possible that mutL is actually required for switching at vlsE but some other B. burgdorferi protein can substitute for mutL because of functional redundancy.
h. Role of recJ in switching?- Similarly, the possible role of recJ in switching is not clear and needs to be studied further.
i. Comparison with other antigenic variation systems: Information about proteins involved in gene conversion events is available only from the studies on N. gonorrhoeae and Trypanosoma brucei. However, the process of antigenic variation by recombinational switching in B. burgdorferi at the vlsE locus differs dramatically from that of N. gonorrhoeae and further studies are required.