Ras: Do We Know Its Structure Well Enough?

May 27, 2015, by Carla Mattos

The first crystal structures of Ras were solved over 25 years ago^1,2. Its architecture was noted and the first structures of oncogenic mutants soon followed in which local changes in the active site were observed with no global perturbations in structure^3,4. As a community we felt empowered by these structures and efforts to translate this knowledge into structure-based drug discovery programs soon followed at major pharmaceutical companies. The fact that these efforts failed was discouraging and approaches other than directly targeting Ras quickly came into play. We knew the structure of Ras, but did we know it well enough?

Ras has a catalytic domain tethered to the membrane through a posttranslationally modified C-terminus. We knew that the catalytic domain is composed of a 6-stranded β-sheet flanked by 5 α-helices with 10 connecting loops and that the nucleotide-binding pocket has a few sequence motifs that are highly conserved across the GTPase family, critical for function. We also knew that GAPs are major players in increasing the slow Ras hydrolysis rates to biologically relevant levels and by the end of the 1990s we knew the structure of the transition state of this reaction⁵. I have the sense that by then the Ras research community had adopted the structural view of Ras that is mainstream today. Because those concepts have not facilitated drug discovery, Ras structure fell somewhat out of fashion, with only few labs continuing to study its features.

In 1999, when I began what became a career focused on the structure of Ras, I faced, and continued to face over the years, the question of why I chose to work on a protein whose structure has been around for so long and for which we already know so much. Why not work on a sexier, more cutting edge system? The truth is that at the time I was a protein structure aficionado unfamiliar with the details of Ras biology and viewed it as an excellent small intracellular protein for application of the Multiple Solvent Crystal Structures method that I had been involved with developing⁶. For me, Ras has turned out to be an exciting and challenging research problem, with every new structure paving the way for new questions.

As it turned out, we did know quite a bit about the structure of Ras, but perhaps became satisfied too soon. In those early days we knew little about possible involvement of dynamics and subtle conformational states of Ras-GTP with Ras function, and were focused on larger conformational changes that occur between Ras-GTP and Ras-GDP. In terms of the hydrolysis reaction those days were colored by the Ras-GTP structures of the time, with particular conformations of the switch regions stabilized by crystal contacts shaping our interpretation of the role of essential catalytic residues. The fact that Ras-GTP has a disordered active site in solution^7,8 makes it difficult to determine the role that catalytic residue Q61 plays in intrinsic hydrolysis. Its role is better understood in the GAP-catalyzed reaction due to the available structure of the transition state, with the side chain of Q61 interacting simultaneously with the nucleophilic water molecule and with the backbone carbonyl of the arginine finger⁵. Residue Q61 is unlikely to be critical in positioning the nucleophilic water molecule for intrinsic hydrolysis as this water is fully present in the ground state in a structure with a disordered switch II and no electron density for Q61 (ref. 9). Given the view that intrinsic hydrolysis is too slow to be biologically relevant, the focus was on the GAP-catalyzed reaction and although understanding the intrinsic hydrolysis mechanism in detail might have been an interesting academic pursuit, it also, like structure, fell out of fashion.  In spite of diminished focus in this area, research has continued to make significant contributions to our understanding of Ras catalysis^10-13.

New studies on the activation of intrinsic hydrolysis through allosteric modulation of the active site has revealed that in addition to known catalytic residues, switch II, helix 3, loop 7 and helix 4 may modulate hydrolysis, particularly in the complex with Raf¹⁴. These structural elements are affected by Ca2+ binding at an allosteric site remote from the nucleotide-binding pocket and thought to interact with the membrane¹⁵. We have recently shown that the Q61L mutant affects not only distal portions of Ras, but also portions of Raf-RBD in the complex¹⁶. Remarkably, the binding of Raf-RBD to wild type HRas promotes a decrease in flexibility of residues that coordinate Ca2+ at the remote allosteric site, adding to our proposed mechanism that interaction with the membrane primes the site for calcium binding to enhance intrinsic hydrolysis in the complex with Raf. Thus, our current hypothesis is that Raf-RBD works synergistically with the membrane to render the allosteric site receptive to Ca2+ binding, triggering GAP-independent hydrolysis to turn off signaling through Ras/Raf. A requirement for involvement of both the membrane and of Raf would explain the fact that NMR studies of Ca2+ and Mg2+ binding to the allosteric site of Ras in solution showed a relatively flexible allosteric site with no specificity for Ca2+ over Mg2+ and low affinity for both¹⁷.

Given the global effects of the Q61L mutation, my group is focused on understanding communication between the active site and distant structural elements of Ras in both mutant and wild type proteins.  We have recently deciphered networks of communication between the active site of HRas in the first half of the catalytic domain and the membrane-interacting regions elucidated by Hancock, Gorfe and co-workers in the second half of the catalytic domain^{15, 18-20}. Intriguingly, several sites identified by us as hot spots of protein-ligand interactions²¹ coincide with sites that have the most differences between H, K and NRas²². Thus, while intramolecular networks of communication originating in switch I and switch II and leading to these sites²³ are likely common to the three Ras isoforms, they may be affected differently by ligand binding in each isoform. Furthermore, the intramolecular networks of communication are affected differently by specific oncogenic mutants and an understanding of these effects will help guide efforts in directly targeting specific Ras mutants against human cancers²⁴.

Now, with an improved understanding of Ras structure, we can place the isoform specific differences in sequence within a three-dimensional view associated with function in order to formulate hypotheses that can be tested in the more aggressively oncogenic isoforms. By deciphering the nuances of structure in sufficient detail we can hope to discover innovative and exciting venues for targeting Ras.

About the Author

Carla Mattos earned her Ph.D. in Chemistry from MIT with Greg Petsko and did postdoctoral work at Harvard (Martin Karplus) and Brandeis (Greg Petsko and Dagmar Ringe).  She was on the faculty at North Carolina State University before moving to Northeastern University in 2012.

References

1. Milburn MV, et al. Science 1990; 247: 939-945.
2. Pai EF, et al. EMBO J 1990; 9: 2351-2359.
3. Franken SF, et al. Biochemistry 1993; 32: 8411-8420.
4. Krengel U, et al. Cell 1990; 62: 539-548.
5. Scheffzek K.  Science 1997; 277: 333-338.
6. Mattos C, et al. J Mol Biol 2006; 357: 1471-1482.
7. Ito Y, et al. Biochemistry 1997; 36: 9109-9119.
8. O'Connor C, Kovrigin EL. Biochemistry 2008; 47: 10244-10246.
9. Buhrman G, Wink G, Mattos C. Structure 2007; 15:1618-1629.
10. Du X, et al. PNAS 2004; 101: 8858-8863.
11. Du X, Sprang SR. Biochemistry 2009; 48: 4538-4547.
12. Klahn M, Schlitter J, Gerwert K. Biophys J 2005; 88: 3829-3844.
13. Kotting C, et al. PNAS 2008; 105: 6260-6265.
14. Buhrman G, et al. PNAS 2010; 107:4931-4936.
15. Gorfe AA, et al. J Med Chem 2007; 50:674-684.
16. Fetics SK, et al. Structure 2015; 23: 505-516.
17. O'Connor C, Kovrigin EL. Biochemistry 2012; 51: 9638-9646.
18. Abankwa D, Gorfe AA, Hancock JF. Cell Cycle 2008; 7: 2667-2673.
19. Abankwa D, et al. PNAS 2010; 107: 1130-1135.
20. Abankwa D, et al. EMBO J 2008; 27: 727-735.
21. Buhrman G, et al. J Mol Biol 2011; 413: 773-789.
22. Parker JA, Mattos C. Mol Cancer Res 2015; 13: 595-603.
23. Kearney BM, et al. J Mol Biol 2014; 426: 611-29
24. Marcus K, Mattos C. Clin Cancer Res 2015; 21: 1810-1818.