What they found

Crystal structures of WH2 domains from WASP, WAVE2, and WIP in complex with G-actin (ternary complex with DNase I as crystallization cofactor). Parallel ITC biochemistry comparing WH2 vs thymosin-β (Tβ) binding to actin. First structural evidence that short WH2 domains (~17 aa) can coexist with intersubunit contacts in F-actin; the canonical G-actin binding site (barbed-end groove between subdomains 1 and 3) is partly accessible on the filament surface.

Numbers that matter

WH2 G-actin binding affinities — ITC, 25°C, G-buffer (2 mM Tris pH 7.5, 0.2 mM CaCl₂, 0.2 mM ATP) — VALUES FROM FIGURE 1C (2026-04-25 update):

Figure 1C bar graph values read directly from main-text figure (MinerU image parse, 2026-04-25). SI Table 2 still inaccessible (PMC proof-of-work challenge blocks curl/fetch; PNAS 403); Fig. 1C displays the same ITC dataset graphically with error bars.

ConstructKd (µM)Kd (nM)Notes
Tβ4 (2-44)0.76 ± 0.12760 ± 120Full-length Tβ4, reference
Tβ4 (2-33)3.10 ± 0.943100 ± 940N-terminal half only
Tβ4 (18-44)no bindingC-terminal half only
Ciboulot (10-43)21.55 ± 3.8021,550Tβ-family, weak
Ciboulot (54-82)53.48 ± 9.1853,480Tβ-family, very weak
Ciboulot (87-119)6.25 ± 0.116,250Tβ-family
MIM (724-755)0.23 ± 0.04230 ± 40Long WH2, strong
WIP (29-60)0.16 ± 0.01160 ± 10Long WH2, strongest
WIP (29-46)2.15 ± 0.462,150N-terminal half
WIP (46-63)>100>100,000C-terminal half, negligible
WASP (430-458)0.25 ± 0.03250 ± 30Short WH2, weakest full-domain
WAVE2 (433-464)0.052 ± 0.00352 ± 3Short WH2, strongest (Arg-Arg)
WAVE2 (450-464)26.53 ± 13.4226,530C-terminal half, confirms N-helix drives binding

Bold = constructs whose crystal structures are in this paper; these are the therapeutically relevant WH2 comparators.

Key ratios (cross-check with text):

  • WAVE2/WASP = 0.25/0.052 = 4.8× (text: “~5-fold” — consistent)
  • Tβ4/WASP = 0.76/0.25 = 3.0× (text says “~10-fold higher” — text likely refers to WAVE2/Tβ4 = 14.6× or to average WH2 family vs Tβ4 average including ciboulot)

CRITICAL NOTE on model value: WH2_KD_GACTIN_M = 200 nM is between WIP (160 nM) and WASP (250 nM) — a conservative mid-range for de novo WH2 constructs of unknown sequence type. WAVE2 (52 nM) is the favorable outlier due to Arg-Arg in LKKT forming salt bridges with Asp-24/Asp-25 on actin.

SI Table 2 retrieval history:

  • Anna’s Archive (DOI, title+supplementary, MD5 searches): SI not separately indexed — 4 attempts, 0 hits
  • PMC (pnas_0507021102_2.pdf): proof-of-work challenge blocks non-browser access
  • PNAS publisher: 403 on all SI URL patterns
  • Resolution: Fig. 1C in main text is the same ITC dataset — values read directly from bar graph image. Status upgraded ⚠ → ✅.

WH2 × F-actin side-binding — NOT MEASURED in this paper:

  • The paper shows WH2 binding site (barbed-end groove, subdomains 1/3) is ALSO the site of longitudinal actin–actin contacts in the filament long-pitch helix
  • Short WH2 domains (~17 aa) can “coexist with intersubunit contacts in F-actin” structurally, but this means nucleation (WH2 templating monomers along a strand), not side-binding of preformed filaments
  • No co-sedimentation or F-actin binding Kd measured

Tβ4 × F-actin: weak cooperative binding, Kd = 5–10 mM (cited from Husson 2010) This is 3 orders of magnitude weaker than G-actin binding — the C-terminal α-helix in Tβ4 interferes with filament contacts, but the same region being absent in WH2 doesn’t mean WH2 side-binds better; it just means WH2 doesn’t lock monomer as completely.

Why WH2 × F-actin side-binding is structurally disfavored

The N-terminal amphipathic helix of WH2 binds in the cleft between actin subdomains 1 and 3. In the filament (Holmes model), this same cleft is occupied by longitudinal actin–actin contacts along the long-pitch helix — it is partially buried. Residual surface accessibility does not constitute a high-affinity independent binding site. No evidence that WH2 can bind to the side of a preformed filament with affinity < 1 mM. The model’s WH2_KD_FACTIN_M = 5 μM is optimistic; the real value is likely 1–100 mM or unmeasurable.

Connections