Impact of Francisella tularensis pilin homologs on pilus formation and virulence

Abstract

Francisella tularensis is a facultative intracellular bacterium and the causative agent of tularemia. Virulence factors for this bacterium, particularly those that facilitate host cell interaction, remain largely uncharacterized. However, genes homologous to those involved in type IV pilus structure and assembly, including six genes encoding putative major pilin subunit proteins, are present in the genome of the highly virulent Schu S4 strain. To analyze the roles of three putative pilin genes in pili structure and function we constructed individual pilE4, pilE5, and pilE6 deletion mutants in both the F. tularensis tularensis strain Schu S4 and the Live Vaccine Strain (LVS), an attenuated derivative strain of F. tularensis holarctica. Transmission electron microscopy (TEM) of Schu S4 and LVS wild-type and deletion strains confirmed that pilE4 was essential for the expression of type IV pilus-like fibers by both subspecies. By the same method, pilE5 and pilE6 were dispensable for pilus production. In vitro adherence assays with J774A.1 cells revealed that LVS pilE4, pilE5, and pilE6 deletion mutants displayed increased attachment compared to wild-type LVS. However, in the Schu S4 background, similar deletion mutants displayed adherence levels similar to wild-type. In vivo, LVS pilE5 and pilE6 deletion mutants were significantly attenuated compared to wild-type LVS by intradermal and subcutaneous murine infection, while no Schu S4 deletion mutant was significantly attenuated compared to wild-type Schu S4. While pilE4 was essential for fiber expression on both Schu S4 and LVS, neither its protein product nor the assembled fibers contributed significantly to virulence in mice. Absent a role in pilus formation, we speculate PilE5 and PilE6 are pseudopilin homologs that comprise, or are associated with, a novel type II-related secretion system in Schu S4 and LVS.

Fig. 1

PilE5 and PilE6 antisera are specific for their respective proteins. The six Schu S4 putative pilin genes (in addition to LVS pilE4) were cloned into expression vector pGST-Parallel1 and transformed into E. coli DH5α to express recombinant pilins bearing N-terminal GST tags. Whole E. coli lysates were transferred onto PVDF following 15% SDS-PAGE. Western blotting with anti-GST antibody (A) revealed that all recombinant pilins were expressed by E. coli. Anti-PilE5 antisera (B) and anti-PilE6 antisera (C) revealed that each antibody was specific for recognition of the respective pilin protein against which it was raised. The protein products expressed in E. coli migrated close to the predicted sizes of their respective GST fusions, which were: PilE1 39.5 kD, PilE2 41.4 kD, PilE3 47.7 kD, PilE4 47.9 kD, PilE4LVS 59.7 kD, PilE5 40.7 kD, PilE6 46.9 kD.

Fig. 2

Schu S4 and LVS nonpolar deletion mutants do not express their respective deleted proteins. Whole cell lysates representing approximately 3.5×107 (LVS) or 7×107 (Schu S4) bacteria were transferred onto PVDF following 15% SDS-PAGE. Membranes were incubated with anti-PilE5 (A) or anti-PilE6 (B) antisera to confirm protein expression in each wild-type strain, as well as their absence or reconstituted presence in deletion and complementation strains, respectively. Anti-GroEL antiserum (C) was used to approximate equal loading of samples. Deletion strains are designated “Δ”, while trans-complemented strains are designated “c”.

Fig. 3

PilE4 is essential for pilus-like fiber expression on Schu S4. Negative stain TEM demonstrated that wild-type Schu S4, Schu S4ΔPilE5, and Schu S4ΔPilE6 exhibited similar levels of piliation, while Schu S4ΔPilE4 demonstrated a complete lack of pili. The absence of pili in Schu S4ΔPilE4, coupled to their reconstitution in the trans-complemented Schu S4ΔPilE4c, suggested that PilE4 was either the major pilin subunit or an accessory protein necessary for the expression of type IV pili in F. tularensis. Images shown here at 6000x magnification are representative of a minimum of 20 fields of observation.

Fig. 4

PilE4 is essential for pilus-like fiber expression on LVS. Negative stain TEM demonstrated that wild-type LVS, LVSΔPilE5, and LVSΔPilE6 exhibited similar levels of piliation, while LVSΔPilE4 and the pilT::HimarFT mutant demonstrated a complete lack of pili. The reconstitution of pilus expression upon PilE4 trans-complementation (LVSΔPilE4c) suggested that, as in Schu S4, PilE4 was either the major pilin subunit or a necessary accessory protein for type IV pilus assembly in LVS. Images shown here at 6000x magnification are representative of a minimum of 20 fields of observation.

Fig. 5

LVS pilE4, pilE5, pilE6, and pilT mutants demonstrate increased adherence to J774A.1 cells. LVSΔPilE4, LVSΔPilE5, LVSΔPilE6, and the pilT::HimarFT mutant demonstrated increased adherence compared to wild-type LVS in adherence assays with cytochalasin D-treated J774A.1 cells. Adherent bacteria were enumerated via recovered CFU counts, and the ratio of adherence capacity for each mutant strain compared to wild-type was calculated using the General Linear Model. A value of “1” represents wild-type LVS adherence to cells (approximately 0.07% of starting inoculum) (A). CFU enumeration results were confirmed qualitatively by confocal microscopy (B), in which bacteria adherent to cells were stained yellow, and quantitatively in (C). For these measurements, the number of bacteria associated with 50 mammalian cells was counted and then averaged for each strain. Notably, both nonpiliated (LVSΔPilE4, pilT::HimarFT) and piliated (LVSΔPilE5, LVSΔPilE6) mutant strains demonstrated increased adherence compared to wild-type LVS, indicating adherence was independent of piliation. The data shown represent the cumulative results of four independent experiments in which MOIs were approximately 100:1.

Fig. 6

Adherence of Schu S4 pilE4, pilE5, and pilE6 mutants to J774A.1 cells. Schu S4 pilE4, pilE5, and pilE6 mutants demonstrated adherence similar to wild-type Schu S4 in adherence assays with cytochalasin D-treated J774A.1 cells. Adherent bacteria were enumerated via recovered CFU counts, and the ratio of adherence capacity for each mutant strain compared to wild-type was calculated using the General Linear Model. A value of “1” represents wild-type Schu S4 adherence to cells (approximately 0.60% of starting inoculum) (A). CFU enumeration results were qualitatively confirmed by confocal microscopy (B) in which bacteria adherent to J774A.1 cells were stained yellow. Notably, both nonpiliated (Schu S4ΔPilE4) and piliated (Schu S4ΔPilE5) mutant strains demonstrated statistically similar levels of adherence compared to wild-type Schu S4 (p>0.06 and p>0.08, respectively), indicating adherence was independent of piliation. The data shown represent the cumulative results of seven independent experiments in which MOIs were approximately 100:1.

Fig. 7

PilE5, PilE6, and PilT contribute to LVS virulence in vivo. Balb/C mice were inoculated subcutaneously (1×10⁸ CFU) or intradermally (5×10⁸ CFU) with wild-type LVS, LVSΔPilE4, LVSΔPilE5, LVSΔPilE6, or LVS pilT::HimarFT bacteria. Mice infected subcutaneously with LVSΔPilE4 did not survive bacterial challenge, though mice infected intradermally with LVSΔPilE4 demonstrated a slight increase (36%) in survival compared to wild-type LVS. Mice infected with LVSΔPilE5, LVSΔPilE6, or pilT::HimarFT via either route demonstrated significantly attenuated disease progression (at least 70% survival) compared to mice infected with wild-type LVS. There were no deaths after day 9. Surviving mice were followed for more than 21 days.

Fig. 8

Dissemination to and replication of LVS pilE4, pilE5 and pilE6 mutants in the liver. Mice were intradermally infected with wild-type LVS (3.9 ×10⁶ CFUs), LVSΔPilE4 (2.2 ×10⁶ CFUs), LVSΔPilE5 (7.1 ×10⁶ CFUs), or LVSΔPilE6 (3.0 ×10⁷ CFUs). Livers were harvested on days 3, 6, and 9 post-infection, and bacterial loads determined by serial dilution and plate count. Means statistically different from that of LVS are indicated by “*” (p<0.05) and “**” (p < 0.02).

Fig. 9

PilE4, PilE5, and PilE6 do not significantly contribute to Schu S4 virulence in vivo. C57BL/6 mice were inoculated subcutaneously or intradermally with approximately 20 CFU of wild-type Schu S4, Schu S4ΔPilE4, Schu S4ΔPilE5, or Schu S4ΔPilE6 bacteria. Mice receiving Schu S4ΔPilE5 and Schu S4ΔPilE6 demonstrated no significant difference in time to death from those receiving the parent Schu S4 strain via either route of infection. Four mice receiving Schu S4ΔPilE4 survived infection challenge; these mice were challenged with only half the infectious dose (9 CFU) of those mice that had died or were euthanized (20 CFU) indicating PilE4, in addition to PilE5 and PilE6, is not significantly important to Schu S4 virulence in vivo. There were no deaths after day 9. Surviving mice were followed for more than 21 days.