Journal of Protein Chemistry, Vol. 18, No. 8, 1999
Molecular Dynamics Analysis of the Structures of ras-Guanine Nucleotide Exchange Protein (SOS) Bound to Wild-Type and Oncogenic ras-p2l. Identification of Effector Domains of SOS James M. Chen,1 Fred K. Friedman,2 Mark J. Hyde,3 Regina Monaco,4 and Matthew R. Pincus4,5,6
Received November 4, 1999
The X-ray crystal structure of the ras oncogene-encoded p21 protein bound to SOS, the guanine nucleotide exchange-promoting protein, has been determined. We have undertaken to determine if there are differences between the three-dimensional structures of SOS bound to normal and oncogenic (Val 12-p21) proteins. Using molecular dynamics, we have computed the average structures for both complexes and superimposed them. We find four domains of SOS that differ markedly in structure: 631-641, 676-691, 718-729, and 994-1004. Peptides corresponding to these sequences have been synthesized and found to be powerful modulators of oncogenic p21 in cells as described in an accompanying paper. We find that the SOS segment from 809-815 makes contacts with multiple domains of ras-p21 and can facilitate correlated conformational changes in these domains. KEY WORDS: ras-SOS; molecular dynamics; average structure; effector domains; jun kinase.
The single amino acid substitution induces regional changes in the three-dimensional structure of oncogenic forms of p21 compared with that for the activated normal protein (Monaco et al., 1995a, b). Peptides from a number of these regions, such as 35-47, 96-110, and 115-126, inhibit oncogenic, but not activated normal p21, indicating that each protein induces mitogenic signaling using different signal transduction pathways (Amar et al., 1997; Pincus et al., 2000; Adler et al., 1995, 1996; Chung et al., 1992, 1997).
1. INTRODUCTION
Oncogenic ras-p21 protein7 induces the malignant transformation of mammalian cells in culture, induces Xenopus laevis oocyte maturation (Birchmeier et al., 1985), and has been implicated in causing a high proportion of human malignancies (Barbacid, 1987; Pincus et al., 1996). It is the counterpart of the normal ras-protooncogene product and differs from it by containing a single amino acid substitution at a critical position in the sequence, such as at Gly 12. Computational Chemistry Division, Wyeth-Ayerst Corporation, Pearl River, New York 10965. 2 Laboratory of Molecular Carcinogenesis, National Cancer Institute, Bethesda, Maryland 20892. 3 Advanced Computing Services, DeWitt, New York 13214. 4 Department of Pathology and Laboratory Medicine, VA Medical Center, Brooklyn, New York 11209. 5 Department of Pathology, SUNY Health Science Center, Brooklyn, New York 11203.
To whom correspondence should be addressed at VA Medical Center, 800 Poly Place, Brooklyn, New York 11209; E-mail:
[email protected]. 7 Abbreviations: SOS, Son-of-sevenless guanine nucleotide exchange protein; ras-p21, Harvey ras-oncogene-encoded p21 protein; normal p21, wild-type p21; oncogenic p21, G12V-substituted Harvey-ras-p21 protein; RBD, ras-binding domain of the raf-p74 protein; PKC, protein kinase C; JN, jun kinase; rms, root mean square.
1
6
867 0277-8033/99/100-0867$16.00/0 C 1999 Plenum Publishing Corporation
Chen et al.
868 Activation of normal and some oncogenic forms of p21 requires guanine nucleotide exchange in which GDP bound to ras-p2l is replaced by GTP. This exchange process is catalyzed by a ras-specific nucleotide exchange protein, identified in Drosophila as the sonof-sevenless (SOS) protein, the human form of which contains 1044 amino acids (Chardin et al., 1993). Binding of SOS to p21 therefore promotes dissociation of GDP; GTP replaces GDP because its intracellular concentration is approximately tenfold higher than that for GDP. SOS itself is activated by the adapter grb-2 protein that concurrently binds to a tyrosine kinase domain of a transmembrane growth factor receptor linking it to ras-p21 (Chardin et al., 1993). The X-ray crystal structure of wild-type p21 bound to residues 568-1044 of SOS has recently been determined (Boriak-Sjodin et al., 1998). SOS is largely an A-helical protein with an amino-terminal structural domain consisting of six A-helices 1-6 and a rasbinding domain consisting of residues 752-1044 containing helices A-K. Several A-helical segments of SOS have been found to interact with p21, most notably the H-helix (residues 925-946). SOS binds simultaneously to a number of different domains of ras-p21, including residues 25-40, 55-75, 95-105, and 118-123. Residues 25-40, the switch 1 region, include a domain that has been identified as being the effector domain that directly interacts with GTPase activating protein (GAP) (Scheffzek et al., 1997), the raf-p74 protein (Nassar et al., 1995), and the enzyme phosphatidylinositol-3-hydroxy kinase (PI3K) (reviewed in Pincus et al., 2000). Residues 55-75 constitute the switch 2 domain that is mobile and not well-defined in crystal structures of ras-p2l, free or bound to the RBD of raf or to GAP. Binding of these residues to SOS results in a welldefined structure for this segment. Residues 95-105 coincide with the peptide domain 96-110, which we have previously identified as being involved in the interaction of oncogenic p21 with jun kinase (JNK) and jun proteins (Adler et al., 1995, 1996; Amar et al., 1997). Residues 118-123 are involved in the binding of p21 to the guanine ring of GDP/GTP (Barbacid, 1987) and constitute part of the 115-126 domain, also implicated in the binding of oncogenic p21 to JNK (Adler et al., 1995, 1996). Binding of SOS to wild-type ras-p21 results in large changes in the conformations and dispositions of both switch 1 and 2 domains compared with those of the nucleotide-bound protein. For example, Phe 28 moves 9.6 A in the switch 1 region. This results in disruption of interactions of the residues of these regions with the
phosphate moieties of GDP/GTP and a magnesium ion chelated by p21 residues in these domains (BoriakSjodin et al., 1998). Overall, favorable interactions of p21 with SOS replace favorable interactions of p21 with GTP/GDP (Boriak-Sjodin et al., 1998). As we have performed previously for the binding of p21 to the ras-binding domain of raf (Chen et al., 1996, 1997; Chung et al., 1997), in this paper we explore the conformational effects of the binding of wildtype and oncogenic (Val 12-containing) ras-p21 on SOS, using molecular dynamics. Regional differences in the SOS structures of these complexes identify these domains as being potentially important in oncogenic ras-p21-induced mitogenic signaling. We identify four regions of SOS whose conformations differ markedly between the two complexes. In a companion paper, we present the results of studies in which these peptides were microinjected into oocytes induced to mature by oncogenic ras-p2l or by insulin-activated normal p21.
2. METHODS 2.1. Starting Structures for Wild-Type and Val 12-p21 Bound to SOS The atomic coordinates for the X-ray crystal structure of normal (wild-type) p21 bound to SOS were obtained from the Brookhaven National Laboratories Data Bank (Boriak-Sjodin et al., 1998; Abola et al., 1987). The structure contains residues 1-166 (of 189 amino acid residues) of p21 and residues 568-1044 that constitute the ras-binding domain of SOS. Segments of this truncated SOS protein were not well defined in the X-ray crystal structure at 2.8-A resolution and have been omitted from the coordinate list. The following SOS fragments were used in the computations described below: Arg 568-Glu 590; Gly 597-Pro 653; Arg 676-Lys 741; and Ile 752-Asn 1044.
2.2. Energy Minimization of the X-Ray Structure The X-ray structure was then subjected to energy minimization using both ECEPP (Nemethy et al., 1983) and AMBER (Weiner et al, 1986) programs. In these computations, p21 and SOS were allowed to undergo all internal degrees of freedom (dihedral angles), and p21 was also allowed to undergo the six external (three translational and three rotational) degrees of freedom. In the first set of iterations, the amino- and carboxyl-terminal residues for each SOS
Effector Domains of SOS Bound to ras-p2l fragment were constrained to remain as closely as possible to their starting positions. In a following set of iterations, this constraint was removed. The resulting energy-minimized structure was closely superimposable on that for the starting X-ray crystal structure.
869 ergies of the last 50 structures on the trajectory for each complex were the same, the atomic coordinates of each of the 50 structures were numerically averaged to obtain the average structure. 2.6. Superposition of Average Structures
2.3. Energy-Minimized Val 12-p21 Structure A Val residue was generated in place of the Gly residue in the X-ray structure, using the same set of backbone dihedral angles as for Gly as described previously (Monaco et al., I995a, b). The energy of this starting structure was then initially minimized such that only the dihedral angles of Val 12 and its four nearest neighbor residues on both amino- and carboxylterminal ends were allowed to vary. Following this step, the energy-minimization procedures were identical to those described in the preceding paragraph.
The purpose of the present study is to determine the regions of SOS that differ in structure when it is complexed to wild-type or to oncogenic (Val 12-) p21 protein to identify possible effector domains uniquely affected by the binding of oncogenic p21. Therefore, the average structures computed for the two complexes were superimposed on one another such that the rms deviations of the backbone atoms were a minimum. Regions that were found to differ between the two structures were considered as candidates as possible effector domains. 3. RESULTS AND DISCUSSION
2.4. Molecular Dynamics Calculations For each of the two complexes MD calculations were performed using the program AMBER (Weiner et al., 1986) with a distance-dependent dielectric constant. Each structure was heated from OK to 300K in increments of 5K. At each temperature, molecular dynamics trajectories were computed for 50 fsec in which velocities were assigned randomly to each atom from a Maxwellian distribution of molecular velocities characteristic of the particular temperature. After heating to 300K, the system was allowed to equilibrate for an additional 50 psec so that the computed average temperature for the complex was 300K. Each of the atoms of the two temperatureequilibrated complexes was assigned random atomic velocities from the Maxwell distribution of velocities for 300K. The dynamics trajectory for the system was then followed for 2 nsec. In all of the computations, integrations for the trajectories were performed over 1-fsec intervals using the Verlet integration algorithm. In the dynamics trajectory, it was found that the energy converged after about 50 psec. Structures were saved after each 1-psec interval, and average molecular properties (see below) were computed from the structures with converged energies, i.e., the last 50 structures generated in the dynamics procedure. 2.5. Average Structures and Residue Fluctuations These were computed for the two complexes were computed as described previously (Liwo et al., 1994; Monaco et al., 1995a, b). Since the conformational en-
3.1. Residue Fluctuations Average normalized residue coordinate fluctuations are plotted in Fig. 1, where it can be seen that there are significant differences between oncogenic and normal p21-SOS protein complexes. The maximum fluctuation for the wild-type protein bound to SOS was 2.4 A at Thr 124 of the p21 protein (1.7 A for the corresponding Val 12-p21 protein complex), while the maximum fluctuation for the Val 12-p21-SOS complex was 3.2 A at Ile 598 of SOS (0.8 A for the corresponding Gly 12-p21-SOS complex). Fluctuations for more than 90% of the residues were less than 1 A for the wildtype p21-SOS complex, while over half the residues for the oncogenic p21-SOS complex were less than 1 A, and more than 90% of the residues had fluctuations less than 1.5 A, as seen in Fig. 1, indicating that both structures converged to well-defined average structures.
3.2. Oncogenic p21-SOS Complex Is More Flexible than the Wild-Type p21-SOS Complex Overall the fluctuations for the wild-type protein complex are lower than for the oncogenic protein complex, suggesting that the oncogenic protein may be more flexible in that it can adopt alterative lowenergy conformations around the average structure that differ from one another regionally. This conclusion is in agreement with the results of prior computations on the low-energy structures for two mutant
Chen et al.
870
Table I. RMS Deviations for Superposition of Structures for SOS-p21 Complexes
Fig. 1. Average residue fluctuations in A plotted against residue number for wild-type (dotted lines) and oncogenic (solid line) rasp21-SOS complexes. The fluctuations for the 166 residues of p21 are plotted in the lower left part of the figure. The abscissa is then broken at the junction of the last p21 residue and the first crystallographically defined residue of SOS (Arg 568). The abscissa is then broken between consecutive crystallographically defined SOS segments, Arg 568-Glu 590, Gly 597-Pro 653, Arg 676-Lys 741, and He 752-Asn 1044.
non-nucleotide-binding p21 proteins, one of which transforms cells and the other which does not induce cell transformation (Liwo et al., 1994). 3.3. Superposition of Structures Superposition of the energy-minimized starting structures for SOS bound to Val 12-p21 on that for SOS bound to wild-type p21 resulted in an overall backbone rms deviation of 0.3 A, i.e., the two structures were identical. Molecular dynamics of each of these structures, however, resulted in average structures that differed regionally from their respective starting (energy-minimized) structures, as summarized in Table I. In this table, superposition of the average oncogenic (Val 12) p21-SOS complex structure on its starting structure resulted in an rms deviation of 8 A, while the rms deviation of the average structure for the wild-type protein on that for its starting structure was 5 A (entries 1 and 2 in Table I, respectively). The average structure of the p21 protein component in the wild-type p21-SOS complex superimposes on that for the p21 component of the starting structure with an average rms deviation of 2.1 A (entry 3 in Table I). In contrast, superposition of the
Structure
RMS deviation (A)
1. Average structure of SOS-Val 12-p21 on energy-minimized complex 2. Average structure of SOS-Gly 12-p21 on energy-minimized complex 3. Average Gly 12-p21 from complex on energy-minimized Gly 12-p21 in complex with SOS 4. Average Val 12-p21 from complex on energy-minimized Val 12-p21 in complex with SOS 5. Average structure of Val 12-p21 alone on that for Gly 12-p21 alone from SOS complexes 6. Average structure of SOS-Val p21 on average SOS-Gly 12-p21 for whole complex
8.0 5.0 2.1
4.0
3.7
6.4
average structure of the Val 12-p21 component on that for the Val 12-p21 component in the starting energy-minimized complex results in an rms deviation of 4 A, entry 4 of Table I, substantially higher than for the wild-type protein. A similar value for the rms deviation was found when the average structure for the Val 12-p21 protein was superimposed on that for the wild-type protein (entry 4 in Table I). Therefore, much of the overall deviations for the oncogenic p21-SOS complex from its energy-minimized starting structure arises from differing conformations of the p21 protein component. In contrast, for wild-type p21-SOS, the major deviations in structure between the average and starting structures occur in the SOS component. Superposition of the average structure for the whole Val 12-p21-SOS on the corresponding structure for the wild-type (Gly 12-containing) p21-SOS complex results in an rms deviation of 6.4 A (entry 6 in Table I). These deviations reflect differences in structure between normal and oncogenic p21 and significant differences in the regional structures of SOS in each complex. That there are significant regional changes in the structures of oncogenic and normal p21 complexed with SOS can be seen in Fig. 2, in which the computed average structures for the two complexes are superimposed on one another. As can be seen in this figure, while the overall chain fold for p21 and the SOS proteins is similar, there are major differences in the orientation of discrete segments. Quantitative comparison of the two superimposed structures as aver-
Effector Domains of SOS Bound to ras-p21
Fig. 2. Stereoview of the superposition of the average structure for Val 12-p21 (red) (oncogenic) bound to SOS (blue) on that for wildtype (Gly 12-) p21 (green) bound to SOS (purple). The amino- and carboxyl-termini of ras-p21 structures are labeled, as are each of the crystallographically defined SOS segments defined in the legend to Fig. 1.
age residue rms deviations (Liwo et al., 1994; Monaco et al., 1995a, b) is shown in Fig. 3. 3.4. Contact Regions in ras-p21 and SOS in the Two Complexes From Fig. 2, in both structures, as in the X-ray crystal structure, the p21 protein is bound to a dis-
871 crete surface of the SOS protein in the region involving SOS residues 752-1044. The interaction interfaces can be seen where the two ras-p21 structures (green trace, wild-type; red trace, oncogenic) contact SOS (purple trace, wild-type; blue trace, oncogenic). SOS interacts with p21 in a number of different domains, including the switch 1 and 2 regions (residues 25-50 and 55-75, respectively), residues 85-107, and 115-126 (guanine ring-binding region). Contact regions between the two proteins in each complex are summarized in Table II. For the contact residues of SOS, the specific A-helices in which these residues are found are also listed. Despite differences in the conformation of regions of p21 and SOS between oncogenic and wildtype protein complexes, discussed below, several important contacts between domains of p21 and SOS, observed in the X-ray structure, are preserved. Thus, for example, Leu 938 from the H A-helix of SOS lies near Ala 59, from the switch 2 region of both p21 proteins, resulting in a hydrophobic pocket, and Glu 942, also from the H A-helix of SOS, has polar interactions with Ser 17 from the p21 switch 2 domain. Both of these contacts are thought to disrupt the magnesiumbinding site, destabilizing the interaction of p21 with nucleotide (Boriak-Sjodin, et al., 1998). 3.5. Comparison of ras-p2l Switch 2 Contacts with SOS in the Two Complexes Examination of the contacts between both p21 proteins and SOS, as summarized in Table II, reveals that, in p21, the switch 2 region (residues 54-74) makes the largest number of contacts with SOS. The nature of Table II. Domains of ras-p21 that Contact Domains of SOS in the Wild-Type and Oncogenic Complexes
Fig. 3. Average rms deviations in A between the superimposed average structures for the oncogenic p21-SOS complex on that for wild-type p21-SOS complex plotted against residue number. The abscissa is broken between p21 and SOS and between discontinuous crystallographically defined SOS segments as explained in the legend to Fig. 1.
Domain of WT-p21
Domain of V12-p21
Domains of SOSa
12-18 21-41 54-74
12-18 21-41 54-74
85-107
85-107
NCb NCb
118-121 147-149
809-815, 929-947 904-913, 929-947, 963 809-815, 824-836, 876-885, 904-913, 929-947, 1001-1010 809-815, 876-885, 1001-1010 809-815 929-947
The SOS domains are from the following structural units in both complexes: 809-815 is between AA and AB; 929-947, AH; 904-913 spans AF-AG; 824-836, AB; 875-885, AD; 1001-1010, AK. * No contacts with SOS for the corresponding segment of p2 1 .
a
872 these contacts is shown in Figs. 4 and 5 for wild-type and Val 12-(oncogenic) p21 proteins, respectively. Both figures show that the conformation of residues 54-67 is similar and consists of a semiextended region from residues 54-59, followed by a turn involving residues 60-62, followed by an A-helix from residues 63-67, as we predicted using conformational analysis prior to the determination of the X-ray structure of ras-p2l (Pincus et al., 1987). For normal (wild-type) p21 protein this helix continues through residue 74, while for the Val 12 (oncogenic) p21 protein these residues are nonhelical (Fig. 5). Despite the similarity in conformation of the 54-67 segment of the switch 2 domain for the two complexes, comparison of the interactions made between the switch 2 domain in the wild-type and oncogenic p21 proteins with SOS reveals a number of different contacts. These differences in contacts are caused by differing dispositions of this segment in each protein. For example, Glu 63 in the oncogenic protein forms
Fig. 4. Stereoview of the contacts between the wild-type ras-p21 switch 2 region (residues 54-74) shown in green with those of SOS (purple) where the contacting residues of SOS are labeled in gold. Asp 54 and Thr 74 are labeled in green in p21. View is from the average structure of wild-type p21 bound to SOS.
Chen et al. an extensive number of hydrogen-bonded interactions with Trp 809, Lys 814, Arg 826, Thr 829, Asn 830, Ser 1000, and Glu 1002, while Glu 63 in the wild-type protein hydrogen-bonds with Arg 826 only. On the other hand, Met 67 in the wild-type protein forms hydrophobic contacts with Phe 929 and Ile 932 of SOS, also observed in the X-ray structure, which are not present in the oncogenic p21 protein-SOS complex. Other specific critical interactions between p21 and SOS in the switch 2 region observed in the X-ray crystal structure, such as the side-chain-side-chain interaction between Arg 68 of p21 and Glu 1002 of SOS, are maintained in both complexes. 3.6. Significance of the Conformation of the Switch 2 Region Based on our calculations of the average structures of activated wild-type and oncogenic ras-p2l proteins we found that oncogenic amino acid substitutions either at position 12 or 61 induce stereotypical changes in the dispositions and conformations of specific segments, i.e., 10-16, 32-47, 55-67 (from switch 2), 81-93, 96-110, and 115-126 (Monaco et al., I995a, b; Liwo et al., 1994). We found that changes in the conformation and disposition of the switch 2 region resulted in correlated changes in the conformation and disposition of the other segments of the protein (Liwo et al., 1994). From the results presented here, it is clear that SOS recognizes two different dispositions of the switch 2 domain, one for wild-type and the other for oncogenic p21. Binding of SOS to the oncogenic conformation of the switch 2 region may stabilize this conformation and, as a consequence, the
Fig. 5. Stereoview of the contacts between oncogenic ras-p2l switch 2 region (residues 54-74) shown in red with those of SOS (blue) where the contacting residues of SOS are labeled in gold. Asp 54 and Thr 74 are labeled in red in p21. View is from the average structure of oncogenic p21 bound to SOS.
Effector Domains of SOS Bound to ras-p21 activated conformations of the other domains of p21. These domains may be involved in continuous mitogenic signaling. In addition, binding of p21 to SOS in the oncogenic conformation may result in induced signal transduction by SOS, evidence for which is presented in an accompanying paper.
3.7. Residues 809-815 of SOS Make Multiple Domain Contacts with ras-p21 From Table II, it can be seen that one region of SOS, residues 809-815, makes multiple contacts with almost all of the critical effector domains of p21, including the regions around Val 12, Gln 61, Arg 102, and Cys 118-Ala 121, but not including the switch 1 region, residues 25-40. These contacts are shown for the oncogenic p21-SOS complex in Fig. 6. In this figure, the side chains of the SOS 809-815 segment are shown to demonstrate the closeness of the residues of this region to the disparate domains of p21. Since no other region of SOS makes such extensive contacts with p21, it appears that this domain of SOS may trigger conformational changes in bound p21 such as to allow correlated motions in these different domains. In this way, SOS may, in addition to its role in promoting nucleotide exchange, act as a catalyst for inducing activating conformational changes in ras-p21. Importantly, SOS residues 809-815 adopt the same conformation in the average structure for both complexes, as can be seen from Fig. 3, in which the rms deviation for this segment from the superposition of the two complexes is low. This result suggests that the SOS 809-815 segment acts as a coordinating region which allows different domains to change conforma-
Fig. 6. Stereoview of the contacts between SOS segment 809-815 (blue), including side chains, and oncogenic ras-p21 (red) from the average structure of oncogenic p21 bound to SOS. Representative residues from critical domains of oncogenic p21 contacting the SOS segment are labeled in red.
873
tion in a concerted fashion. The fact that it does not interact with the switch 1 region suggests that it anchors the different effector domains with which it does interact, while the switch 1 region undergoes large structural changes.
3.8. Identification of Possible Effector Segments of SOS From Fig. 3, there are four distinct segments (excluding Pro 653, which undergoes an isolated large end-residue motion) whose structures differ markedly from one another, namely residues 631-641, 676-691, 718-729, and 994-1004. Of these segments, only residues 994-1004 occur in a ras-binding domain (residues 1001-1010) of SOS. The other three lie in the structural region of SOS. The structures of these segments in both the wildtype and oncogenic p21 complexes with SOS are shown from the superimposed complexes (Fig. 2) in Fig. 7 (structures from oncogenic and wild-type p21bound SOS are blue and red, respectively). In this figure it may be noted that corresponding segments differ mainly in disposition; their internal conformations are similar to one another. For example, residues 676-691 are both largely A-helical in both complexes.
Fig. 7. Stereoview of the four putative effector domains of SOS from the superimposed average structure of oncogenic p21-SOS on that for wild-type p21-SOS as in Fig. 2. The SOS segments are blue from the oncogenic p21-SOS complex and purple from the wildtype p21-SOS complex.
874
Since these regions differ most between the two structures, and since three of them occur in a region of the protein that is remote from the binding domains, these regions may be important either in controlling the activity of SOS or in interacting with other signal transduction target proteins. The one ras-contacting region, 994-1004, containing contact residues 1001-1004, is involved in contacting residues 54-74 (switch 2 region) and 85-107 of p21. The latter region includes most of the residues in the 96-110 segment of p21 which constitutes an effector domain of oncogenic p21 that interacts with JNK and jun target proteins (Adler et al., 1995, 1996; Amar et al., 1997; Pincus et al., 2000). In an accompanying paper, we present the results of experiments in which we have synthesized peptides corresponding to the four SOS domains that differ most in the superimposed ras-p21-SOS complexes and have microinjected them into oocytes that are induced to mature either by microinjection of Val 12-p21 protein or by insulin-stimulated activated and overexpressed normal p21 protein. ACKNOWLEDGMENTS
This work was supported in part by NIH Grant CA 42500 and by a Merit Award from the Veterans Administration to M.R.P. REFERENCES Abola, E. E., Bernstein, F. C., Bryant, S. H., Koetzle, T. F., and Weng, J. (1987). In Crystallographic Data Bases—Information Content, Software Systems, Scientific Applications (Allen, F. H., Bergerhoff, G., and Sievers, R., eds.), International Union of Crystallography, Bonn, pp. 107-132. Adler, V., Pincus, M. R., Brandt-Rauf, P. W., and Ronai, Z. (1995). Proc. Natl. Acad. Sci. USA 92, 10585-10589. Adler, V., Pincus, M. R., Polotskaya, A., Montano, X., Friedman, F., and Ronai, Z. (1996). J. Biol. Chem. 271, 23304-23309.
Chen et al. Amar, S., Glozman, A., Chung, D. L., Alder, V., Ronai, Z., Friedman, F. K., Robinson, R., Brandt-Rauf, P. W., Yamaizumi, Z., and Pincus, M. R. (1997). Cancer Chemother. Pharmacol. 41, 79-85. Barbacid, M. (1987). Annu. Rev. Biochem. 56, 779-827. Birchmeier, C., Broek, D., and Wigler, M. (1985). Cell43, 615-621. Boriak-Sjodin, P. A., Margarit, S. M., Bar-Sagi, D., and Kuriyan, J. (1998). Nature 394, 337-343. Chardin, P., Camonis, J. H., Gale, N. W., Van Aelst, L., Schlessinger, J., Wigler, M. H., and Bar-Sagi, D. (1993). Science 260, 1338-1343. Chen, J. M., Manolatos, S., Brandt-Rauf, P. W., Monaco, R., and Pincus, M. R. (1996), J. Protein Chem. 15, 511-518. Chen, J. M., Monaco, R., Manolatos, S., Brandt-Rauf, P. W., Friedman, F. K., and Pincus, M. R. (1997). J. Protein Chem. 16, 631-635. Chung, D. L., Brandt-Rauf, P. W., Weinstein, I. B., Nishimura, S., Yamaizumi, Z., Murphy, R. B., and Pincus, M. R. (1992). Proc. Natl. Acad. Sci. USA 89, 1993-1996. Chung, D., Amar, S., Glozman, A., Chen, J. M., Friedman, F. K., Robinson, R., Monaco, R., Brandt-Rauf, P. W., Yamaizumi, Z., and Pincus, M. R. (1997). Inhibition of oncogenic and activated wild-type ras—p21 protein-induced oocyte maturation by peptides from the ras binding domain of the raf-p74 protein, identified from molecular dynamics calculations, J. Protein Chem. 16, 619-629. Liwo, A., Gibson, K. D., Scheraga, H. A., Brandt-Rauf, P. W., Monaco, R., and Pincus, M. R. (1994). J. Protein Chem. 13, 237-251. Monaco, R., Chen, J. M., Chung, D. L., Brandt-Rauf, P. W., and Pincus, M. R. (1995a). J. Protein Chem. 14, 457-466. Monaco, R., Chen, J. M., Friedman, F. K., Brandt-Rauf, P. W., and Pincus, M. R. (1995b). J. Protein Chem. 14, 721-730. Nassar, N., Horn, G., Herrmann, C., Scherer, A., McCormick, F., and Wittinghofer, A. (1995). Nature 375, 554-560. Nemethy, G., Pottle, M. S., and Scheraga, H. A. (1983). J. Phys. Chem. 87, 1883-1887. Pincus, M. R., Brandt-Rauf, P. W., Gibson, K., Carty, R. P., Lubowsky, J., Avitable, M., and Scheraga, H. A. (1987). Proc. Natl. Acad. Sci. USA 89, 8375-8379. Pincus, M. R., Brandt-Rauf, P. W., and Nostro, D. (1996). Cell biology and early tumor detection, in Clinical Diagnosis and Management by Laboratory Methods, 19th ed., (Henry, J. B., ed.), Saunders, Philadelphia, pp. 310-321. Pincus, M. R., Brandt-Rauf, P. W., Michl, J., and Friedman, F. K. (2000). Cancer Invest. 46, 35-53. Scheffzek, K., Ahmadian, M. R., Kabsch, W., Weismuller, L., Lautwein-Schmitz, F., and A. Wittinghofer, (1997). Science 227, 333-338. Weiner, S. J., Kollman, P. A., Case, D. A., Singh, V. C., Ghio, C., Alagona, G., Profeta, S., and Weiner, P. K. (1986). J. Am. Chem. Soc. 106, 765-784.
