Preferential protection of Borrelia burgdorferi sensu stricto by a Salp15 homologue in Ixodes ricinus saliva

Abstract

Background: Ixodes ticks are the main vectors for Borrelia burgdorferi sensu lato. In the United States, B. burgdorferi is the sole causative agent of Lyme borreliosis and is transmitted by Ixodes scapularis. In Europe, 3 Borrelia species-B. burgdorferi, B. garinii, and B. afzelii-are prevalent, which are transmitted by Ixodes ricinus. The I. scapularis salivary protein Salp15 has been shown to bind to B. burgdorferi outer surface protein (Osp) C, protecting the spirochete from antibody-mediated killing.

Methods and results: We recently identified a Salp15 homologue in I. ricinus, Salp15 Iric-1. Here, we have demonstrated, by solid-phase overlays, enzyme-linked immunosorbent assay, and surface plasmon resonance, that Salp15 Iric-1 binds to B. burgdorferi OspC. Importantly, this binding protected the spirochete from antibody-mediated killing in vitro and in vivo; immune mice rechallenged with B. burgdorferi preincubated with Salp15 Iric-1 displayed significantly higher Borrelia numbers and more severe carditis, compared with control mice. Furthermore, Salp15 Iric-1 was capable of binding to OspC from B. garinii and B. afzelii, but these Borrelia species were not protected from antibody-mediated killing.

Conclusions: Salp15 Iric-1 interacts with all European Borrelia species but differentially protects B. burgdorferi from antibody-mediated killing, putatively giving this Borrelia species a survival advantage in nature.

Figure 1

Borrelia antigens recognized by rabbit antisera.

Figure 2

Characterization of (recombinant) Salp15 Iric-1. A, Quantitative reverse-transcription polymerase chain reaction (PCR) performed on RNA obtained from salivary glands (SG) from semiengorged adult female Ixodes ricinus ticks with primers specific for Salp15 Iric-1, -2, and -3. The no. of Salp15 Iric-1, -2, and -3 copies were determined from the same cDNA sample in a single PCR run. Bars represent means ± SEs for 3 independent experiments. The no. of copies is normalized to the amount of input RNA. B, Salp15 Iric-1 cloned into the pMT/BiP/V5–His C expression vector and stably transfected into S2 Drosophila cells. Recombinant protein with a C-terminal His tag and V5 epitope was purified using nickel-charged columns. Recombinant Ixodes scapularis Salp15 was purified similarly to Salp15 Iric-1. Recombinant proteins were separated by SDS-PAGE, and the gel was subjected to Coomassie brilliant blue staining; lane 1, molecular weight (MW) marker; lane 2, I. scapularis Salp15; lane 3, Salp15 Iric-1. C, Investigation of the difference in apparent molecular weight. We determined whether the 2 proteins were differentially modified at the posttranslational level by assessing potential N-glycosylation sites using the Web-based software available at http://www.cbs.dtu.dk/services/NetNGlyc. The N-glycosylation potential threshold was set at 0.5. D, I. scapularis Salp15 and Salp15 Iric-1 deglycosylated by N-glycosidase F (New England Biolabs) in accordance with the manufacturer’s instructions under denaturing conditions, subjected to SDS-PAGE, and blotted onto an Immobilon-P membrane. Proteins were visualized with anti-V5– horseradish peroxidase antibody; lanes 1 and 2, I. scapularis Salp15; lanes 3 and 4, Salp15 Iric-1. Arrows indicate that samples were treated with N-glycosidase F.

Figure 3

Interaction of Salp15 Iric-1 with Borrelia burgdorferi strain N40. A, Solid-phase overlays with B. burgdorferi strain N40 and B. burgdorferi strain 297 deficient for ospC as well as Ixodes scapularis Salp15 and Salp15 Iric-1. The left panel shows Coomassie blue staining of lysates from B. burgdorferi strain N40 (7.2 μg) and strain 297 deficient for ospC (5.1 μg), separated by SDS-PAGE and blotted onto an Immobilon-P membrane. The molecular weight (MW) was determined using a molecular weight marker. Western blot analysis using a monoclonal antibody, L22 1F8, recognizing outer surface protein (Osp) C from B. burgdorferi sensu lato confirmed that the ~22-kDa band was indeed OspC. Binding of recombinant I. scapularis Salp15 or Salp15 Iric-1 to native OspC was assessed by a solid-phase overlay with 1 μg/mL I. scapularis Salp15, 1 μg/mL Salp15 Iric-1, or vehicle control as a negative control, and binding was visualized using mouse anti-V5 and rabbit anti–mouse IgG– horseradish peroxidase (HRP) antibodies. B, Purified Escherichia coli– expressed recombinant OspC from B. burgdorferi strain N40, separated by SDS-PAGE and visualized by Coomassie blue staining. Lane 1, Molecular weight marker; lane 2, N40-OspC. C, OspC-Salp15 ELISA. N40-OspC (400 ng/well) was coated onto ELISA plates and incubated with different concentrations of I. scapularis Salp15 (black line) and Salp15 Iric-1 (gray line). Binding was visualized by incubation with a secondary and tertiary antibody, anti-V5-HRP and anti–mouse IgG–HRP antibodies, respectively. After addition of the substrate, absorbance was read at 450 nm. As a control, bovine serum albumin (BSA) was coated, and background signals did not exceed the indicated dashed line. This graph is representative of 5 independent experiments. Error bars represent SEs of triplicates from a single experiment. D, Surface plasmon resonance N40-OspC and I. scapularis Salp15 and Salp15 Iric-1. N40-OspC was bound to a CM5 sensor chip and different molar concentrations (ranging from 0 to 4500 nmol/L, represented by individual lines, with the highest line representing 4500 and the lowest line representing 0 nmol/L) of I. scapularis Salp-15 or Salp15 Iric-1 were injected for 300 s. Resonance units (RUs) depicted are after subtraction of the RUs from the BSA reference channel. Binding constants were calculated by 1:1 Langmuir fitting and confirmed by steady-state affinity plotting (see table 2). In addition, E. coli– expressed tumor necrosis factor–α with a C-terminal His tag and vehicle control did not show any binding (data not shown). E, In vitro protection assay in which spirochetes were preincubated with I. scapularis Salp15, Salp15 Iric-1 (final concentration, 40 μg/mL), or BSA as a control and then subjected to serum (diluted 1:200) from an N40-immunized rabbit (α-N40) for 16 h. Viable spirochetes were visualized by dark-field microscopy. Bars represent means ± SEs for 3 independent experiments. NRS, normal rabbit serum. F, In vivo experiment showing N40 burden in mice immune to N40 that were reinfected with Barbour-Stoenner-Kelly medium (SHAM) (2 mice), N40 preincubated with BSA (N40 plus BSA) (8 mice), or Salp15 Iric-1 (N40 plus Iric-1) (7 mice) and killed 2 weeks after reinfection. N40 burden was determined by quantitative polymerase chain reaction and was corrected for the amount of β-actin. Bars represent means ± SEs. As a positive control, naive mice (n = 3) were infected with N40. The N40 burdens in different tissues in these mice were comparable to those in immune mice rechallenged with N40 preincubated with Salp15 Iric-1 (data not shown). G, Carditis severity scores in hematoxylin-eosin–stained sagittal sections of mouse hearts. In comparison, 3 naive mice infected with N40 and killed 2 weeks after infection had an average carditis score of 2.7 ± 0.5 (data not shown). A 2-sided Mann-Whitney U test was performed to determine statistical differences between the groups. *P < .05 and **P < .01. NS, not significant.

Figure 4

Binding of Salp15 Iric-1 to outer surface protein (Osp) C from Borrelia burgdorferi strain B31.

Figure 5

Protection of Borrelia strains by Salp15 proteins at lower concentrations, as shown by protection assays using the different Borrelia strains.

Figure 6

Interaction of Salp15 Iric-1 with Borrelia garinii strain PBi and Borrelia afzelii strain pKo. A, Solid-phase overlay with the different Borrelia species. Lane 0, molecular weight (MW) marker; lane 1, Borrelia burgdorferi strain N40; lane 2, B. garinii strain PBi; lane 3, B. afzelii strain pKo; lane 4, ospC-deficient strain 297. The amounts loaded on SDS-PAGE gel for B. burgdorferi strain N40, B. garinii strain PBi, B. afzelii strain pKo, and the outer surface protein (Osp) C mutant B. burgdorferi strain 297 were 7.2, 8.14, 8.62, and 5.1 μg, respectively. After transfer to an Immobiolon-P membrane, Coomassie blue staining, Western blot, and solid-phase overlay using 1 μg/mL Salp15 Iric-1 were performed. B, Coomassie blue staining of recombinant N40-OspC (lane 2), PBi-OspC (lane 3), and pKo-OspC (lane 4). Lane 1, molecular weight marker. Western blot analysis using monoclonal antibody L22 1F8, recognizing OspC from B. burgdorferi sensu lato, confirmed that the ~22-kDa recombinant proteins were indeed OspC (data not shown). The affinity of Salp15 Iric-1 for the different OspCs was assessed by OspC ELISA (C) and by surface plasmon resonance (D). E, In vitro protection assay in which N40, PBi, or pKo spirochetes were preincubated with Salp15 Iric-1 (final concentration, 40 μg/mL) or with bovine serum albumin (BSA) as a control and then subjected to serum from a rabbit immunized with the indicated Borrelia species for 16 h (anti-N40, diluted 1:200; anti-PBi and anti-pKo, diluted 1:300). Bars represent means ± SEs for 3 independent experiments. The average amount of viable spirochetes per dark-field microscope field in the well incubated with normal rabbit serum (NRS) was set at 100%. F, Borrelia burden determined by quantitative polymerase chain reaction in tissues from an in vivo experiment with B. afzelii strain pKo as described for N40 in the legend to figure 3. Eight to 9 mice per group were used. As a positive control, naive mice (n = 3) were infected with pKo. These mice had high spirochete numbers in skin (688 ± 491 spirochetes/1 × 10⁶ β-actin DNA copies), bladder (228 ± 100 spirochetes/1 × 10⁶ β-actin DNA copies), and ankle tissue (1210 ± 813 spirochetes/1 × 10⁶ β-actin DNA copies). These data are not depicted in the graphs, but the Y-axis of the graphs is based on these values. Both immune and naive mice were killed 2 weeks after (re)infection. G, Carditis severity scores in hematoxylin-eosin–stained sagittal sections of mouse hearts. In comparison, 3 naive mice infected with pKo and killed 2 weeks after infection had an average carditis score of 3 ± 0 (data not shown). A 2-sided Mann-Whitney U test was performed to determine statistical differences between the groups. *P < .05, **P < .01, and ***P < .001. NS, not significant.