An NMR Comparison of the Changes Produced by Different Guanosine 5′-Triphosphate Analogs in Wild-Type and Oncogenic Mutant p21^ras

We have used nuclear magnetic resonance spectroscopy to compare the conformational changes produced by replacement of bound GDP by the GTP analogs guanosine 5′-O-(3-thiotriphosphate) (GTP-S) and guanylyl (β,γ-imido)diphosphate (GMPPNP) in wild-type p21^ras as well as the oncogenic mutant (G12D)p21^ras. We have used isotope-edited nuclear magnetic resonance spectroscopy to observe the amide resonances of selectively [¹⁵N]glycine and [¹⁵N]isoleucine labeled p21^ras-nucleotide complexes. We find that eight of the nine resonances that respond strongly to GTPγS and GMPPNP binding are the same but that the nature of the effect appears different. With GTPγS, seven new resonances replace the eight resonances specifically associated with GDP-p21^ras, but in GMPPNP-p21^ras only two resonances replace the GDP-specific resonances that are lost. The resonance of Gly 60 is clearly shown to be responsive to replacement of GDP by GMPPNP, in addition to glycines 10, 12, 13, 15, and 75 and isoleucines 36, 21, and one other, that were found to respond to GTPγS by Miller et al. [Miller, A.-F., Papastavros, M. Z., & Redfield, A. G. (1992) Biochemistry 31, 10208–10216). The two GMPPNP-specific resonances observed appear in positions similar to GTPγS-specific resonances, and the GTPγS-specific resonances, although not lost altogether, are weaker than the GDP-specific resonances they replace. Thus, the two GTP analogs have similar effects on the spectrum of p21^ras, suggesting that the effects are due to features common to both analogs. We propose that active site resonance intensities are specifically attenuated when GTP analogs are bound because interactions with the γ-phosphate of GTP analogs couple the flexible loops 2 and 4 to the rigid loop 1 of the active site. The conformational heterogeneity and dynamics of loops 2 and 4 would be constrained by loop 1 but also transmitted to it. Coupled conformational exchange on a common intermediate time scale could explain the simultaneous loss of resonances from all three loops in the active site. In our comparison of wild-type and (Gl2D) GDP-p21^ras, we find that the resonance of He 36 is not visible in (G12D)p21^ras. In (G12D)p21^ras, replacement of GDP by GTPγS causes the resonances of glycines 10, 13, 15, 60, and 75 and isoleucine 21 and four others to shift from their GDP-specific positions. GTPγS-specific resonances are observed for all but two of these. The assigned responsive resonances all correspond to residues in the active site or connected to it. Largely the same resonances respond to GMPPNP binding, but only four corresponding resonances specific to GMPPNP-(G 12D)p21^ras are observed. Thus, replacement of glycine 12 by aspartate only slightly alters the responsiveness of the ground state of p21^ras to nucleotide replacement. Furthermore, the observed GTPγS-specific resonances of (G12D)p21^ras are close to GTPγS-specific resonances of wild-type p21^ras, even though several of the GDP-specific (G12D)p21^ras resonances differ significantly from those of wild-type p21^ras.