Main

Small G proteins belonging to the Ras superfamily act as signalling molecules controlling a wide range of biological processes. Crucial to their signalling function is the conformational switching that takes place between their active, GTP-bound and inactive, GDP-bound forms1. They have a slow, intrinsic rate of GTP hydrolysis which is substantially accelerated by GAPs specific to each subfamily. Crystallographic coordinates for rhoGAP and rasGAP are now available2,3 and we have compared them using the program LSQMAN (ref. 4).

Superposition of the two GAPs shows that they share a core structure made up of seven α-helices (Fig. 1). The α-helices making up this core pack against each other in a related, but not identical, fashion. However, the automated alignment puts the catalytic arginine residues (R85 in rhoGAP and R789 in rasGAP) into approximately the same position and, by implication, locates the active sites of the bound G proteins.

Figure 1: The common core structures of the catalytic domains of p50rhoGAP (top) and p120rasGAP (bottom), as aligned by the computer program LSQMAN, in helical-tube representation.
figure 1

Helical segments forming the core structure (helices A-G and 1c to 7c in rhoGAP and rasGAP, respectively) are coloured from blue at their amino termini to red at their carboxy termini. Helical insertions that do not form part of the core motif are grey. The positions of the catalytic arginines R85(rhoGAP) and R789(rasGAP) are indicated by yellow circles.

Full size image

It is particularly interesting to note that in both structures the catalytic arginine residue is located on a surface loop and is preceded by two bulky, hydrophobic residues (isoleucine and phenylalanine at residues 83 and 84, respectively, in rhoGAP, and leucine and phenylalanine at positions 787 and 788, respectively, in rasGAP). These residues are held in a hydrophobic clamp made up of residues from the B(2c), E(5c) and F(6c) helices that act to anchor the catalytic arginine loop to the core of the domain.

The published structures of the transition-state-analogue complexes of RhoGDP.AlF4−/p50rhoGAP (ref. 5) and RasGDP.AlF3/p120rasGAP (ref. 6) show that the two GAP domains bind their respective G proteins in related but distinct ways. rhoGAP binds to Rho primarily through a surface made up by the helices B and F, whereas rasGAP binds to Ras through the antiparallel helices 6c and 7c.

These differences in G-protein-binding surfaces arise because of the different spatial relationship between helices A-B and E-F of rhoGAP and 1c-2c and 5c-6c of rasGAP. Relative to the rhoGAP helices E and F, the carboxy terminal (or top half) of 5c bends back into the plane of the page while the whole of 6c rotates further in a similar direction.

Nonetheless, both GAP domains present their catalytic arginine residues to the active sites of the G protein in very similar ways, which is a consequence of the equivalence in function of the arginine loop in these two systems. It will be intriguing to see whether other GAPs associated with the Ras superfamily contain the same core structure, obey the same topology, and make similar use of hydrophobic clamping to orient the catalytic arginine loop.