Mechanism of Nwasp Inhibition by Wiskostatin

Since wiskostatin inhibits the activity of full-length N-WASP, but not that of the isolated VCA domain, this suggested that wiskostatin might perturb the conforma-tional changes in N-WASP that accompany activation. How it might do so was of particular interest because N-WASP is a scaffolding protein not previously known to bind a small molecule. Available three-dimensional structural data for N-WASP that could facilitate prediction of binding sites for wiskostatin are limited to crystal structures of the N-terminal EVH1 domain of N-WASP (Volkman et al. 2002; Peterson et al. 2007). This domain does not, however, appear to play a role in either Cdc42 binding or Arp2/3 complex activation (Prehoda et al. 2000; Rohatgi et al. 2000). N-WASP is highly conserved in sequence and regulation with its hematopo-etic-cell-specific paralogue, WASP, for which additional structures are available. For example, an NMR structure has been reported of a truncated WASP fragment comprising the GBD linked to a C-terminal element involved in Arp2/3 complex activation (Kim et al. 2000). This structure and accompanying biochemical experiments revealed the core interactions responsible for maintaining the autoinhibited conformation of WASP (and presumably N-WASP). Given the availability of chemical shift assignments for the autoinhibited core of WASP and the 68% sequence identity with the corresponding domains of N-WASP, we chose to test whether binding of wiskostatin to WASP could be detected by NMR. Michael Rosen's group (UT Southwestern) conducted NMR titration experiments with wiskostatin and the autoinhibited core domains of WASP. Titrations of this WASP construct with wiskostatin resulted in dose-dependent and saturable amide proton chemical shift changes for a limited number of residues in the GBD of WASP. These initial experiments demonstrated the existence of a discrete wiskostatin-binding site in WASP.

Rosen's group then determined the structure of wiskostatin bound to the autoinhibited core of WASP by NMR (Fig. 1b). Remarkably, the structure of this region of WASP was largely unchanged by wiskostatin binding. The average backbone root mean square deviation between the liganded and free forms of WASP was 1.51 Á within the structured domains. The compound occupied a shallow pocket within the hydrophobic core of the WASP GBD that was well defined by 30 intermolecular nuclear Overhauser effects between the wiskostatin carbazole ring and the GBD. Importantly, 14 of the 15 amino acid residues most proximal to wisko-statin in WASP are conserved in N-WASP, suggesting that these paralogues likely both interact with the compound similarly.

Previous studies have demonstrated that the GBD of WASP is conformationally plastic, alternately adopting one conformation in autoinhibited WASP and another, mutually exclusive conformation when bound to Cdc42 (Abdul-Manan et al. 1999; Kim et al. 2000). This conformational change underlies the transition between autoinhibited WASP and active WASP in which Cdc42 binding to the GBD releases the VCA element, allowing Arp2/3 complex activation (Fig. 1a). A recombinant polypeptide comprising the isolated GBD is unstructured in aqueous buffer, indicating that folding of this segment is driven by these intra- or intermolecular interactions, respectively (Kim et al. 2000). The lack of significant changes in the conformation of the GBD of autoinhibited WASP on wiskostatin binding suggested that, instead of perturbing the conformation of WASP, the compound might act by stabilizing this inherently inactive conformation.

To test whether wiskostatin could stabilize the autoinhibited conformation of the GBD, the otherwise unstructured, isolated GBD polypeptide was titrated with wiskostatin and conformational changes in the GBD were monitored by NMR. Wiskostatin, but not an inactive derivative, induced folding of the GBD as evidenced by greater chemical shift dispersion in 'H-15N- and 'H-13C HSQC experiments. At saturating doses of wiskostatin, the resulting spectrum evidenced many characteristics of the spectrum of GBD when present in the autoinhibited conformation, strongly suggesting that wiskostatin alone could drive folding of the GBD into the autoinhibited conformation. This observation suggests that wiskostatin inhibits activation of N-WASP by binding to its autoinhibited conformation and antagonizing the conformational change mediated by Cdc42 that is required for N-WASP activation of Arp2/3 complex.

Activation of N-WASP by Cdc42 can be seen as the modulation of a conforma-tional equilibrium in the GBD from a predominantly "closed," inactive form to an

Wiskostatin Gbd

Fig. 1 The small-molecule N-WASP inhibitor wiskostatin antagonizes activation of N-WASP by Cdc42. (a) Intramolecular interactions between the GTPase-binding domain (GBD) and the ver-prolin homologous, cofilin homologous, acidic amino acid sequences (VCA) domain maintains N-WASP in an autoinhibited conformation (left). Binding of activated Cdc42 causes a restructuring of the GBD that relieves these interactions, leading to N-WASP activation (right). The small-molecule inhibitor wiskostatin binds the GBD and pulls the conformational equilibrium of N-WASP toward the autoinhibited state. (b) Structure of wiskostatin bound to the autoinhibited conformation of WASP (left side of a). A representative conformer from the NMR structure of the relevant domains of WASP bound to wiskostatin. Wiskostatin and side chains exhibiting nuclear Overhauser effects to wiskostatin are shown in ball-and-stick representation. Corresponding sequence elements are colored as in (a). Panel (b) is reproduced from Peterson et al. 2004

Fig. 1 The small-molecule N-WASP inhibitor wiskostatin antagonizes activation of N-WASP by Cdc42. (a) Intramolecular interactions between the GTPase-binding domain (GBD) and the ver-prolin homologous, cofilin homologous, acidic amino acid sequences (VCA) domain maintains N-WASP in an autoinhibited conformation (left). Binding of activated Cdc42 causes a restructuring of the GBD that relieves these interactions, leading to N-WASP activation (right). The small-molecule inhibitor wiskostatin binds the GBD and pulls the conformational equilibrium of N-WASP toward the autoinhibited state. (b) Structure of wiskostatin bound to the autoinhibited conformation of WASP (left side of a). A representative conformer from the NMR structure of the relevant domains of WASP bound to wiskostatin. Wiskostatin and side chains exhibiting nuclear Overhauser effects to wiskostatin are shown in ball-and-stick representation. Corresponding sequence elements are colored as in (a). Panel (b) is reproduced from Peterson et al. 2004

active conformation. Our results suggest that wiskostatin and Cdc42 can influence this structural change by binding to the opposing GBD conformational states and titrating the equilibrium toward each state (Fig. 1). Consistent with such an underlying equilibrium, we found that inhibition by wiskostatin can be overcome by increasing the concentration of activators Cdc42 and PIP2 in in vitro pyrene-actin assays. Thus, wiskostatin exploits an inherent regulatory equilibrium in N-WASP by binding and stabilizing a native, autoinhibited conformation. Because the structure of the GBD when bound to VCA is incompatible with the GBD-Cdc42 binding (Kim et al. 2000), we infer that wiskostatin prevents the interaction of Cdc42 with N-WASP. Wiskostatin therefore mediates its inhibitory activity, in part, by preventing Cdc42-N-WASP binding.

0 0

Post a comment