Borrelia burgdorferi RevA Significantly Affects Pathogenicity and Host Response in the Mouse Model of Lyme Disease

Abstract

The Lyme disease spirochete, Borrelia burgdorferi, expresses RevA and numerous outer surface lipoproteins during mammalian infection. As an adhesin that promotes bacterial interaction with fibronectin, RevA is poised to interact with the extracellular matrix of the host. To further define the role(s) of RevA during mammalian infection, we created a mutant that is unable to produce RevA. The mutant was still infectious to mice, although it was significantly less well able to infect cardiac tissues. Complementation of the mutant with a wild-type revA gene restored heart infectivity to wild-type levels. Additionally, revA mutants led to increased evidence of arthritis, with increased fibrotic collagen deposition in tibiotarsal joints. The mutants also induced increased levels of the chemokine CCL2, a monocyte chemoattractant, in serum, and this increase was abolished in the complemented strain. Therefore, while revA is not absolutely essential for infection, deletion of revA had distinct effects on dissemination, arthritis severity, and host response.

FIG 1

Construction of the revA-deficient mutant. (A) Schematic of mutant construction. See Materials and Methods for details of cloning. (B) PCR confirmation of revA deletions by use of flanking primers for the revA6 locus (lanes 1 to 4) and primers for aad (lanes 5 to 7). Lanes 1 and 5, B31-A3ΔrevA1ΔrevA6; lanes 2 and 6, the parental strain, B31-A3; lane 3, cloning plasmid control (with the whole flanking region plus revA and the kan cassette); lane 4, no template; lane 7, cloning plasmid control (with the whole flanking region plus revA and the aad cassette). L, ladder. (C) Schematic of construction of the revA complement vector. MCS, multiple cloning site; RBS, ribosome-binding site. (D) Immunoblotting for RevA protein in whole-cell lysates. Lane 1, B31-A3 (wild type [WT]); lane 2, overloaded B31-A3ΔrevA1ΔrevA6 (knockout [KO]); lane 3, complemented strain. OspC serves as a loading control.

FIG 2

revA-deficient mutants are impaired in colonization of the heart but not the joint. One month after infection, hearts (A) and joints (B) from mice infected with either B31-A3ΔrevA1ΔrevA6 (revA mut), the parental strain, B31-A3 (A3), or the complemented strain (revA comp) were collected, DNA was extracted, and levels of B. burgdorferi DNA were determined by qPCR. Data are expressed as copies of B. burgdorferi recA per femtogram of mouse nidogen DNA. Samples were analyzed in triplicate, and each data point represents an individual animal. Two-way analysis of variance was used to calculate statistical differences between the groups infected with the parental strain or the B31-A3ΔrevA1ΔrevA6 mutant. The difference in the ability to colonize heart tissues (A) was statistically significant (P = 0.003).

FIG 3

Effects of RevA deficiency on ankle thickness. Three- to 4-week-old female C3H/HeN mice were injected in the left footpad with 10³ (A) or 10⁴ (B) bacteria of strain B31-A3, the B31-A3ΔrevA1ΔrevA6 mutant, or the complemented strain. Ankle thickness was measured with calipers (triplicate measurements) at baseline and at 2 and 4 weeks postinfection. Two-way analysis of variance was used to calculate statistical differences between the groups injected with the parental strain or the B31-A3ΔrevA1ΔrevA6 mutant. For mice injected with 10³ bacteria (A), the difference was significant (P < 0.0001) at 4 weeks postinfection.

FIG 4

Proliferative response in periarticular connective tissue. Tibiotarsal sections from infected mice were stained for collagen. (A) Section from a B31-A3-infected mouse with patent synovial spaces and clear demarcations of bone and dense, regular connective tissue. (B) Section from a B31-A3ΔrevA1ΔrevA6 mutant-infected mouse with apparent degeneration of bone and infiltration of connective tissue into the joint space. (C) Section from a mouse infected with the complemented strain. Three individual sections from 3 infected animals were observed per infection strain; representative slides are shown. Arrows indicate collagen staining at the edge of the joint space.

FIG 5

The revA-deficient mutant induces higher systemic CCL-2 levels than the wild-type strain. Blood was collected at the time of sacrifice from mice infected with B. burgdorferi (B31-A3, the B31-A3ΔrevA1ΔrevA6 mutant, or the complemented strain; subcutaneous dose of 10³, 10⁴, or 10⁵ bacteria). Red blood cells were pelleted, and serum was frozen at −80°C. A commercial sandwich ELISA was used to assay CCL-2 levels according to the manufacturer's instructions (R&D Systems). Samples from 6 to 8 animals per inoculum were analyzed in triplicate. Two-way analysis of variance was used to calculate statistical differences between the groups infected with the parental strain or the B31-A3ΔrevA1ΔrevA6 mutant (P = 0.014).