Figure 1.
Schematic illustration of secretin and PilP domain structures.
A) Secretin domain organisation: a type IV pilus-dependent secretin (PilQ; top) is shown compared to type II and type III secretion system secretins (middle and bottom). The proposed domain names are based on secondary structure predictions and homology sequence alignments. B1 and B2 represent structural regions predicted to be rich in β-structure; N0, N1, N2 and N3 represent the consecutively arranged α/β domains described previously for the GspD [20] and EscC [21] secretins. The C-terminal domain is responsible for secretin oligomerization and is embedded in the outer membrane. The ss shaded area is the signal peptide sequence. The constructs discussed in this paper are indicated above, including residue numbering based on the full length protein. B) PilP domain organisation: the lipid moiety is covalently linked to a conserved cysteine at the N-terminus, and is separated from the globular C-terminal domain by a region which is predicted is to be in an unstructured state.
Figure 2.
Structure of the N. meningitidis PilQ B2 domain.
A) Two views of a ribbon plot (above) and structural ensemble (below) of the B2 domain (B2PilQ224–329). The ribbon plot and fold topology diagram (right), generated using Topdraw [62], are shown with a monochromatic gradient from the N- to C-terminus. B) Surface and ribbon plots of the β-domain, generated using CHIMERA [69], showing sequence conservation determined using CONSURF [35]. High sequence conservation is shown in purple, medium in white and low in light blue.
Figure 3.
Structure of the N. meningitidis PilQ N0/N1 domains.
A) Structure of the N0 domain (NmPilQ343–442), ribbon plot (above and structural ensemble (below); for clarity flexible residues 431–442 are omitted. The dashed circle indicates the proposed linker region. B) A composite model for the N0/N1 double domain structure (N0N1PilQ343–545), is shown with a indigo-green gradient from N- to C-terminus. The model is based on the NMR-derived structure for the first domain, and a CS-ROSETTA [65] model for the second domain and linker, with the relative orientation of the domains selected from optimal fit to the cryoelectron microscopy density map. The topology of the N0/N1 domains is outlined below, generated using TOPDRAW [62].
Table 1.
NMR structure calculation statistics.
Figure 4.
Structural model for the PilP C-domain bound to the PilQ N0 domain.
A) Peak attenuation mapped on to the PilP C-domain (PDB accession 2IVW) following titration with N0N1PilQ343–545. Ratios of PilP:PilQ were colored as follows: 1∶0.1, red; 1∶0.5, orange; 1∶0.8, yellow; 1∶1, pale-yellow. Left, ribbon plot with β1 stand marked; right, surface plot. B) Peak attenuation mapped on to N0PilQ343–442. Ratios of PilQ:PilP were colored as follows: 1∶0.1, dark-blue; 1∶0.2, blue; 1∶0.3, cyan; 1∶0.5, pale-blue. C) Model of the PilP77–164: N0PilQ343–442 complex generated from CNS1.2 [56], with PilP77–164 in gold and N0PilQ343–442 in blue. Flexible residues at the N- and C-termini have been removed for clarity. D) Surface plot of the N0PilQ343–442 domain generated using CONSURF [35] and CHIMERA [69], with the same color scheme as used in Figure 2B. A ribbon plot of the PilP C-domain structure [29] is shown in green. The same sequence set was used for CONSURF as employed in Figure 2 B).
Figure 5.
Determination of PilQ oligomer structure by single particle averaging: Raw data image and class averages.
A. Frozen-hydrated specimen of PilQ complexes resuspended in a dodecyl maltoside-containing buffer at 1 mg/ml. The particles in the 1300 nm diameter hole in the carbon support film were selected interactively for image processing and single particle alignment, and then a preliminary model was generated by selecting projections with bilateral symmetry as well as 12-fold rotational symmetry. This model was then used as an alignment reference to which all particles were aligned, after which a refined model was generated. After several rounds of iterative refinement, no further improvement in the model was detected (as judged by Fourier Shell Correlation between the Nth and (N-1)th iteration). B. Comparison between different orientations of the final model (even numbered projections) and averages of all the particles best corresponding to those projections (subsequent odd numbered projection averages). Deviations between the odd and subsequent even numbered projections reflect the errors in the processing procedure, especially with respect to classification and rotational and translational alignment, and are a useful separate measure of the resolution and reliability of the structural data.
Figure 6.
Cryoelectron microscopy structure of the PilQ dodecamer and comparison with Salmonella type III secretion system needle complex.
A) Left panel: surface-contoured map; scale bar = 100 Å; right panel, as left, but with the front half of the volume removed to reveal major domain boundaries. B) Superposition of the PilQ density map (dark gray) onto Vibrio cholerae GspD (light gray; EMD-1763). The periplasmic gate structure is outlined in blue. C) Superposition of the PilQ density map (dark gray) onto Salmonella T3SS needle complex map (light gray; EMD-1875). The approximate locations of the outer membrane (OM) and inner membrane (IM) are shown and the density attributed to the InvG secretin is highlighted in blue.
Figure 7.
Model for the PilP-PilQ assembly.
A) Docking of B2 domain (left panel), N0/N1 domain (middle panel) and both (right panel) into the PilQ cryoelectron density map, contoured at 2.9σ. Some parts of the density map and oligomers have been removed for clarity. Colors are as used in Figure 2A (B2 domain) and Figure 3B (N0/N1 domains). B) Reconstruction of PilQ N0/N1/B2 domain structures (colored as in A) with PilP C-domain (orange) bound. Left panel: side view with 6 oligomers; right panel: top view with 12 oligomers. C) Detail of two oligomers on opposing sides of the PilQ chamber. The scale bar is 60 Å and corresponds to the approximate dimensions of an assembled type IV pilus fiber.