Sortase-mediated assembly and surface topology of adhesive pneumococcal pili

Abstract

The rlrA genetic islet encodes an extracellular pilus in the Gram-positive pathogen Streptococcus pneumoniae. Of the three genes for structural subunits, rrgB encodes the major pilin, while rrgA and rrgC encode ancillary pilin subunits decorating the pilus shaft and tip. Deletion of all three pilus-associated sortase genes, srtB, srtC and srtD, completely prevents pilus biogenesis. Expression of srtB alone is sufficient to covalently associate RrgB subunits to one another as well as linking the RrgA adhesin and the RrgC subunit into the polymer. The active-site cysteine residue of SrtB (Cys 177) is crucial for incorporating RrgC, even when the two other sortase genes are expressed. SrtC is redundant to SrtB in permitting RrgB polymerization, and in linking RrgA to the RrgB filament, but SrtC is insufficient to incorporate RrgC. In contrast, expression of srtD alone fails to mediate RrgB polymerization, and a srtD mutant assembles heterotrimeric pilus indistinguishable from wild type. Topological studies demonstrate that pilus antigens are localized to symmetric foci at the cell surface in the presence of all three sortases. This symmetric focal presentation is abrogated in the absence of either srtB or srtD, while deletion of srtC had no effect. In addition, strains expressing srtB alone or srtC alone also displayed disrupted antigen localization, despite polymerizing subunits. Our data suggest that both SrtB and SrtC act as pilus subunit polymerases, with SrtB processing all three pilus subunit proteins, while SrtC only RrgB and RrgA. In contrast, SrtD does not act as a pilus subunit polymerase, but instead is required for wild-type focal presentation of the pilus at the cell surface.

Fig. 1

Structural characterization of the pneumococcal pilus by AFM and EM. Pili on D39∇(rlrA-srtD) cells were visualized by AFM (A–G) or by transmission EM (H–L). A–E. Low-magnification (A, 910 nm scale bar shown) and high-magnification (B, 170 nm scale bar) AFM images of pili, with a matching topographic projection shown in (C) (with 160 nm scale bar). Insert shows a terminal pilus, at which two height measurements were taken over the length of lines shown, with height (z-axis) deflection plots in nm in (D) and (E). Note that the pilus shaft diameter is estimated to be approximately 2 nm, assuming radial symmetry. Also note that the terminal ‘tip’ is estimated to be approximately 4.5 nm, suggesting different or additional structures than observed in the pilus shaft. F. Thickness projection and magnification of a subset of (C), showing the thickness projection of the structure measured in (D) and (E). G. D39∇(rlrA-srtD)ΔrrgABC serves as a negative-control strain and does not generate pili (1.0 μm scale bar). H and I. Low-magnification (H, scale bar 500 nm) and high-magnification (I, scale bar 100 nm) images of pili by EM. J. A digital magnification of a subset of (I) showing a pilus fibre with a red line indicating a measurement site. K. Greyscale value (arbitrary units) is plotted against distance in nm over the line shown in (J), indicating that this pilus fibre is approximately 1.8 nm in diameter. L. D39, the parental strain that lacks the pilus islet transgene inserted into D39∇(rlrA-srtD), served as a negative control, and did not produce detectable extracellular fibres (scale bar 500 nm). M. We tested whether RrgC composed the ‘tip knobs’ by determining tip knobs per pilus fibre ratios in high-magnification AFM fields, comparing strains D39∇(rlrA-srtD) (‘D39∇’), D39∇(rlrA-srtD)ΔrrgA (‘ΔrrgA’) and D39∇(rlrA-srtD)ΔrrgC (‘ΔrrgC’). The rrgC mutant exhibited fewer ‘tip knobs’ per fibre, supporting a model whereby RrgC is the predominant species in the pilus tip complex. N and O. Double labelling for both antigens permitted identification of ‘patches’ including both RrgA and RrgC in D39∇ at both low- (N, 500 nm scale bar) and high-magnification (O, 200 nm scale bar). P = part of figure N is shown in higher magnification in O.

Fig. 2

Redundant role of individual sortases in pilus biogenesis. Production of polymeric high-molecular-weight cell wall-associated pili in isogenic T4 mutants was evaluated by immunoblotting for RrgA (A), RrgB (B) and RrgC (C). In all cases, cell wall proteins were separated by gradient SDS-PAGE, transferred to PVDF and probed. Approximate molecular weights in kDa are indicated on left, based on marker proteins. A. Immunoblotting for RrgA in wild-type T4 (‘WT’), T4ΔrlrA-srtD (‘T4-8’), the trans-complemented srtBCD triple-mutant strains T4ΔsrtBCD+lacE::srtD (‘ΔsrtBCD+srtD’), T4ΔsrtBCD+lacE::srtC (‘ΔsrtBCD+srtC’) and T4ΔsrtBCD+lacE::srtB (‘ΔsrtBCD+srtB’), the triple-mutant T4ΔsrtBCD (‘ΔsrtBCD’), the trans-complemented single sortase mutant strains T4ΔsrtD+lacE::srtD (‘ΔsrtD+srtD’), T4ΔsrtC+lacE::srtC (‘ΔsrtC+srtC’) and T4ΔsrtB+lacE::srtB (‘ΔsrtB+srtB’), T4ΔsrtD (‘ΔsrtD’), T4ΔsrtC (‘ΔsrtC’) and T4ΔsrtB (‘ΔsrtB’). The predicted 90 kDa RrgA monomer is indicated by asterisk 1. B and C. (B) Immunoblotting for RrgB (67 kDa RrgB monomer is indicated by asterisk 2), and immunoblotting for RrgC (predicted 38 kDa RrgC monomer is indicated by asterisk 3) (C), with samples loaded as in (A). Lack of polymer formation in T4ΔsrtBCD shows that at least one of the three pilus-associated sortases, SrtB, SrtC or SrtD, is required for pilus biogenesis. SrtB and SrtC are sufficient for pilin polymerization, as shown by T4ΔsrtBCD+lacE::srtB and T4ΔsrtBCD+lacE::srtC, while SrtD is not capable of pilin polymerization (T4ΔsrtBCD+lacE::srtD). Finally, these data demonstrate that SrtB is necessary to conjugate RrgC to a pilus polymer.

Fig. 3

Topology of pilus antigen in S. pneumoniae. Immunofluorescence (IF) microscopy of piliated pneumococci from two clonal lineages reveals non-homogenous pilus antigen topology, consistent with ‘bands’ or ‘rings’. EM and AFM preparations were adapted for IF staining for anti-RrgB (A and B) and anti-RrgA (C) antibodies, shown in red in these images. Cells were also stained with antibodies against the polysaccharide capsule, in green, and nucleoids were stained with DAPI, in blue. Examples of RrgB topology in the T4 and serotype 19F strain BHN100 clonal lineages were visualized by confocal microscopy (A). T4ΔmgrA is a pilus-overexpressing strain described elsewhere (Hemsley et al., 2003). Multiple examples of RrgB (B) and RrgA (C) topology imaged by conventional microscopy are shown for T4. RrgB is found in discretely concentrated foci, often paired across the division plane, or in a ring encircling the cell. Similar examples of RrgB and RrgA topology in 19F BHN100 are shown in Fig. S5.

Fig. 4

Pilus topology is not determined by pilins. RrgB topology was studied in T4ΔrrgA (A) and T4ΔrrgC (B), and representative images show RrgB immunostaining in red, capsule immunostaining in green and the nucleoid stained with DAPI in blue. Note that in both mutants, RrgB is found in symmetric foci, similar to observations in wild-type T4 (Fig. 3). RrgA is not detectable in T4ΔrrgB (C, top right), making it impossible to determine if RrgA topology is determined by RrgB expression. We hypothesized that RrgA is obscured by the capsule, and therefore evaluated RrgA topology in an unencapsulated mutant of T4, T4R (C, bottom left), and an rrgB mutant of T4R, T4RΔrrgB (C, bottom right). RrgA is preserved as symmetric rings in T4R, although the topology is less discrete than observed in the presence of a capsule (C, top versus bottom left), possibly due to RrgA monomers in the wall. We therefore conclude that RrgA topology is not dependent on rrgB expression. Furthermore, this demonstrates that the pilus shaft functions to extend RrgA and RrgC monomers beyond the capsule. Therefore, pilin antigens detected by IF in an encapsulated organism are not monomers, and are instead higher-order multimeric species.

Fig. 5

Roles of pilus-associated sortases in pilus topology. RrgB topology was examined in isogenic sortase mutants of T4. RrgB immunostaining is shown in red, capsule immunostaining in green and the nucleoid stained with DAPI in blue. A. RrgB focus formation was observed in T4ΔsrtB, but the foci are not co-ordinated into pairs or rings. B. T4ΔsrtC exhibited RrgB topology indistinguishable from wild-type cells. C. T4ΔsrtD exhibited a large number of small, poorly organized RrgB foci diffusely distributed along cell chains. D. Complementation of srtB in trans, strain T4ΔsrtB+lacE::srtB (‘T4ΔsrtB+srtB’), restores discrete symmetrical RrgB distribution, proving the necessity of SrtB in determining pilus topology. E. The necessity of SrtD in determining pilus topology was proven by complementation of srtD in trans, strain T4ΔsrtD+lacE::srtD (‘T4ΔsrtD+srtD’). F. Inactivation of all three pilus-associated sortases in T4ΔsrtBCD results in failure to detect any RrgB on the surface of the pneumococci. G. Insertion of a second copy of srtB on a transgene in strain T4ΔsrtBCD+lacE::srtB (‘T4ΔsrtBCD+srtB’) restores detectable RrgB on the surface of T4. Moreover, this strain displays diffuse RrgB foci, like T4ΔsrtD, supporting a role for SrtD in organizing pilus antigen. H. Strain T4ΔsrtBCD+lacE::srtC (‘T4ΔsrtBCD+srtC’) expressing only srtC shows a similar phenotype like T4ΔsrtD with diffuse RrgB foci. I. No RrgB can be detected after expression of srtD in trans in a triple-sortase mutant (‘T4ΔsrtBCD+srtD’). The effect of srtB and srtD mutations on RrgA topology in T4 is similar, and shown in Fig. S5. Moreover, srtB and srtD mutations have similar effects on RrgB and RrgA topology in 19F BHN100 (Fig. S6).

Fig. 6

The active-site cysteine in SrtB is important for incorporation of RrgC into the pilus polymer in the presence of SrtC and SrtD. A. The role of the active-site cysteine of SrtB was investigated using a SrtB mutant where SrtB, with a substitution of the active-site cysteine for an alanine, was inserted in the lacE locus, in accordance with the other trans-complemented strains used in this study. B. The topology of RrgB was demonstrated to be affected by exchanging the active-site cysteine for an alanine whereby the regular pattern of foci found in the wild-type SrtB was exchanged to an irregular distribution.