PMC

Methods and results for FRET measurement between Arp2/3 subunits tagged with GFP and TR-labelled CA peptide. (A–C) FCCS with 488 nm excitation for eGFP and 561 nm excitation for TR was used to measure the percentage of Arp2/3 complex bound with peptide at a range of peptide concentrations. As an example, representative curves are shown for ARPC3–GFP (A) titrated with TR-CA at the concentrations indicated in (B). The concentration of ARPC3–GFP was held constant (B). FCS curves of TR-CA are shown at the concentrations labelled. In FCS, the amplitude of the autocorrelation curve is inversely proportional to concentration. At each concentration of TR-CA, the cross-correlation was calculated between ARPC3–GFP and TR-CA. An example cross-correlation curve is shown in blue (C) for the concentrations of TR-CA and ARPC3–GFP. (D) The relative amplitude of the cross-correlation compared with the autocorrelation curves (example in C) is used to calculate the percentage of ARP2/3 bound to peptide at each peptide concentration. Representative data are shown for ARPC3–GFP titrated with TR-CA. (E) At each concentration of the TR-labelled peptide, after the FCCS measurement, the 561-nm laser was turned off and eGFP data were acquired and fit with moment-based brightness analysis. The curve of average brightness of each GFP-labelled Arp2/3 complex as a function of fraction bound to labelled CA was fit to extract bound (quenched) and unbound (unquenched) brightness of GFP, which was converted to FRET efficiency. Representative data are shown for Arp3–GFP titrated with TR-CA, which are the pair with the highest FRET. (F) Tables showing FRET efficiency (top) and corresponding distances (bottom) between the C-terminal GFP of indicated subunits and TR at C- or N-terminus of the CA peptide.

Schematic snapshots of early key stages of the pathway of Arp2/3 complex mediated branch formation. (1) VCA binds to inactive Arp2/3 complex, triggering reorientation of Arp2 (red arrow) to form a short-pitch actin-filament-like arrangement with Arp3 as well as partial cleft closure in Arp3 (light grey arrow) and complete cleft closure in Arp2. (2) VCA is bound at the pointed end of the Arp2/Arp3 heterodimer (black arrowhead). This conformation does not support growth of a new filament from the Arp2/Arp3 dimer. In addition, the bound VCA partially occludes the Arp2/3 complex binding site for actin filaments. (3) Because the binding site is not fully accessible, VCA-bound Arp2/3 complex can bind actin filaments only weakly, if at all, possibly through electrostatic interactions. (4) Several changes need to take place to allow binding of Arp2/3 complex to the side of an actin filament. These changes should include closing of the Arp3 nucleotide-binding cleft (light grey arrow) and rearrangement of two mother-filament subunits as observed in the fully assembled branch junction (). In addition, the affinity of the binding site for actin-bound VCA at the barbed end of Arp3 needs to increase to allow binding and delivery of the first daughter-filament subunit (). Finally, dissociation of the first VCA that is bound to Arp2/3 complex at the filament-binding site (red arrow) is necessary to allow tight binding of Arp3/3 complex to the actin filament.

Molecular model of the NPF-bound conformation of Arp2/3 complex. (A, B) Four orthogonal views of the molecular model of the NPF-bound conformation of Arp2/3 complex in cartoon representation and a transparent surface representation of the class II reconstruction from which the NPF density was removed are shown (see Materials and methods model building section for details). The colour scheme of the Arp2/3 subunits follows that in . The excellent visual fit indicates that the conformational changes induced by NPF binding are well accounted for our modelling at the resolution of the reconstruction. (C) Fits of Arp3/ARPC3 (orange, magenta, respectively) models with various states of the Arp3 nucleotide-binding cleft into the transparent surface representation of the density module for these subunits. The view corresponds to the pointed end view in . The models were aligned to subdomain 1 of Arp3 and the contour level of the density presentation was chosen to be slightly lower than that displayed in (A, B) for illustration purposes. Filament and crystal (1K8K) conformations denote closed and open nucleotide-binding cleft, respectively. The fits indicate that Arp3 adopts an intermediate cleft conformation. (D) Fit of two Arp2 models into the Arp2 density module. Subdomains 1 and 2 of Arp2 crystal model were completed using the structure of an actin monomer (1J6Z) overlaid with subdomains 3 and 4 of Arp2. The fits indicate that Arp2 adopts a conformation similar to a subunit in an actin filament.

Localization of the NPF on the Arp2/3 complex. (A) The pointed end view of the entire Arp2/3 complex, (B) the Arp3 view of the entire complex, (C) the side view of the Arp2/Arp3 heterodimer in the NPF-bound conformation and (D) the Arp3 subunit with Arp2 removed for clarity. The Arp subdomains are marked in the first row of (C, D). The first row shows a low-resolution representation of the NPF-bound Arp2/3 conformation (model of class II, ) in the colour scheme defined in . The row labelled cortactin shows the density attributable to the bound cortactin (red) mapped onto the low-resolution representation shown in the first row (white). The extra density at ARPC1 (in pink) can largely be attributed for by the ∼50 residues missing in the crystal structure from ARPC1 (∼30 residues) and the C-terminus of ARPC2 (∼20 residues). The row labelled ‘N-WASp’ shows the density attributable to the bound N-WASp/Nck (red). Most of the N-WASp/Nck is not visible in our maps and appears to be disordered as suggested by the large variance peak in this region (peak 3 in ). The row labelled ‘Scar’ shows the density attributable for by Scar-VCA (red) and the additional density present in the Scar-VCA construct with an N-terminal MBP tag (orange). This location marks the N-terminus of the V region (V). The density does not account for all of MBP, indicating that it is flexibly attached. The row labelled ‘FRET’ shows the location of the C-terminus of the A region (blue, marked A) and the N-terminus of the C region (cyan, marked C). The peaks are contoured to contain 0.5 of the entire probability density. The location determined by FRET is fully consistent with the NPF densities obtained by electron microscopy. The MBS row shows the mother-filament contacts determined previously by electron tomography of fully assembled Arp2/3-mediated branch junctions (). It is clear that the NPF locations determined here overlap with the mother-filament contacts at the MBS2-binding site (green) while the MBS1 and MBS3-binding sites (magenta) remain accessible.

3D reconstructions of Arp2/3 complexes bound to different NPFs. (A–C) Different views of the reconstructions. Views looking towards the pointed end (A), the barbed end (B) and the Arp2 side (C) of the complex are shown. Crystal structure column: the crystal structure column shows a low-resolution representation of the crystal structure of inactive bovine Arp2/3 complex (PDB code: 1K8K). Subdomains 1 and 2 of Arp2 were completed using the structure of an actin monomer (1ATN) overlaid with subdomains 3 and 4 of Arp2. All samples segregated into two classes. Class I column: The class I column shows a surface representation of the class common to all samples. The one shown was obtained from budding-yeast Arp2/3 complex in the presence of N-WASp/Nck. The differences between class I and the low-resolution density calculated from the completed crystal structure were not significant, suggesting that no NPFs are bound in that conformation. Class II columns: The class II columns show surface representations of the second class of the respective samples. In general, all reconstructions are significantly different from the crystal structure, and in some regions from each other. Arrows point out some differences, colour coded according to region. The grey arrow points at changes attributed to Arp2 repositioning. The reconstruction in the Scar-VCA column was obtained from Acanthamoeba Arp2/3 complex in the presence of Scar-VCA tagged at the N-terminus with MBP. (D) Colour mapping for the Arp2/3 subunits depicted in the crystal structure columns of (A–C). The same colour scheme applies to and . (E) Fourier shell correlation for Acanthamoeba Arp2/3 complex with cortactin (blue), Scar-VCA (cyan) and budding-yeast Arp2/3 complex with N-WASp/Nck (magenta). The 0.5 cutoff criterion for the Fourier shell correlation (dotted lines) indicates 1.9–2.3 nm resolutions for these reconstructions. (F) Representation of the variance analysis of class II projections. The variance (red) was mapped into 3D and overlaid with surface representations of the Arp2 view (C) of the respective reconstructions of class II. The location of peaks 2 and 3 overlaps for Scar-VCA and N-WASp/Nck while peak 1 is only seen in the reconstruction of Acanthamoeba Arp2/3 complex in the presence of cortactin.