Ras regulation by GEFs

Identities of Ras-specific GEFs

There are several guanosine exchange factors (GEFs) that are specific to Ras subfamily proteins.

[PMID9585556]: Several genes have been isolated from different organisms encoding proteins that have a GEF activity specific for Ras (for which we use the general name RasGEFs throughout this paper): SOS1 and SOS2 ([PMID9585556_1] [PMID9585556_2] [PMID9585556_3] [PMID9585556_4]) ; Cdc25Mm, also called RasGrf ([PMID9585556_5] [PMID9585556_6] [PMID9585556_7]); and mRas-GRF2 ([PMID9585556_8]).

RasGEFs contain a specific domain responsible for activating Ras proteins.

[PMID9585556]: The RasGEFs are proteins of considerable length, 120 - 160 kDa, and contain several regions which are generally accepted to represent structural domains (12). A region of 200 - 300 amino acids, the RasGEF domain, is shared by all GEFs which act on members of the Ras subfamily, and their activity is specific toward either Ras, Ral, or Rap. The fact that truncated versions of various lengths, containing this RasGEF domain, have been shown to be active RasGEFs in vivo and in vitro (4 , 13 - 16) confirms that this region indeed represents the Ras-specific guanine nucleotide exchange domain.

def rasgef_monomers():
    # Declare a list of RasGEFs along with their site structure.
    # The names in the list below are HGNC standard names.
    # (note: Cdc25Mm = RASGRF1)
    ras_gef_names = ['SOS1', 'SOS2', 'RASGRF1', 'RASGRF2']
    for ras_gef_name in ras_gef_names:
        Monomer(ras_gef_name, ['rasgef'])

Mechanism of GEFs

Some key features of the mechanism:

  1. Ras binds GEFs in the absence of nucleotides

  2. GEF binding causes Ras to release GTP/GDP.

  3. Rebinding of nucleotides causes Ras to release the GEF.

  4. Reloading of Ras with GTP vs. GDP is not determined by the GEF, but rather by the relative cellular concentrations of the nucleotides.

    [PMID9690470]: Biochemical studies of Ras exchange factors have shown that the complex of Ras with these proteins is stable in the absence of nucleotides and is dissociated by the rebinding of either GDP or GTP ([PMID9585556] [PMID9690470_17] [PMID9690470_18] [PMID9690470_21] [PMID9690470_22]). The principal role for the exchange factor is to facilitate nucleotide release, and it does not seem to control significantly the preferential rebinding of GTP over GDP ([PMID9585556], [PMID9690470_22], [PMID7548002]). Cellular concentrations of GTP are 10-fold higher than GDP, which results in the loading of GTP onto Ras.

The fact that GTP/GDP can displace GEFs, while GEFs can also displace GTP/GDP, leads to a paradox that is resolved by the fact that Ras undergoes a conformational change that retains the necessary “state”. The structural basis of this conformational change is described as follows:

[PMID9690470]: As a nucleotide-exchange factor, Sos functions under two apparently conflicting imperatives. The interaction between Sos and Ras must be strong enough to dislodge the tightly bound nucleotide, but the Ras – Sos complex must also be poised for subsequent displacement by incoming nucleotides. The structure of the Ras – Sos complex shows that Ras and Sos meet these demands by forming a tight complex that is anchored at one end of the nucleotide- binding site, where phosphate and magnesium are normally bound. The interface between Sos and Ras is mainly hydrophilic, suggesting a ready unzippering through water-mediated displacements of the coordinating side chains. The main interacting elements of Sos avoid direct occlusion of the nucleotide-binding site, except the region where the terminal phosphate groups and the magnesium ion are bound. This feature allows incoming nucleotides to reverse the process by competing for the groups that ligate the phosphate and metal ion.

This conformational state change has been analyzed kinetically and identified as the process of the nucleotide being first loosely and then tightly bound, (see also Ras proteins are GTPases):

[PMID9690470]: Kinetic analysis of nucleotide association shows that the reaction proceeds by the formation of a ternary complex of a loosely bound nucleotide and Ras – Cdc25Mm followed by conversion to a form in which the nucleotide is tightly bound to Ras [PMID9585556]. In light of the structure of the Ras–Sos complex, the first step can be interpreted as the interaction of the base and the ribose of the nucleotide with the part of the Ras binding site that is not occluded by Sos. The second step would involve a conformational change in the Switch 2 segment and release of Switch 1, resulting in the restructuring of a competent binding site for phosphate and magnesium, and the subsequent dissociation of Sos.

The kinetic analysis described in [PMID9585556] resulted in the following reaction scheme for the interactions between Ras, GTP/GDP, and GEFs:


Note that the upper equilibria for Ras-nucleotide binding, K1a and K1b, were implemented in the section Ras proteins are GTPases, along with corresponding rates. Here we implement only the equilibria involving GEFs: K2, K3, K4a and K4b.

def ras_gef_exchange_cycle(ras, rasgef, gxp,
                           k2_list, k3_list, k4a_list, k4b_list):
    # Alias for Ras bound to GXP
    rasgxp = ras(gef=None, gtp=99) % gxp(p=99)

    # Binding of RasGEF to nucleotide-free Ras (K2)
    bind(ras(gtp=None, s1s2='closed'), 'gef', rasgef(), 'rasgef', k2_list)

    # Binding of RasGEF to RasGXP (K3)
    bind(rasgxp(s1s2='open'), 'gef', rasgef(), 'rasgef', k3_list)

    # Binding of GXP to Ras/RasGEF complex
    bind(ras(s1s2='closed', gef=1) % rasgef(rasgef=1), 'gtp',
         gxp(), 'p', k4a_list)

    # Isomerization of Ras-RasGEF-GXP from loose to tight
    equilibrate(rasgxp(gef=1, s1s2='closed') % rasgef(rasgef=1),
                rasgxp(gef=1, s1s2='open') % rasgef(rasgef=1), k4b_list)

Rates of GEF activation

# Binding of RasGEF to nucleotide-free Ras
kf2 = 0.33e6        # M^-1 s^-1
kr2 = 1e-3          # s^-1

# Binding of RasGEF to RasGXP
KD3 = 0.6e-3        # M
kf3 = 3.4e4         # M^-1 s^-1 (lower limit)
kr3 = KD3 * kf3     # s^-1

# Binding of GXP to Ras/RasGEF complex
KD4a = 8.6e-6       # M
kf4a = 1e7          # M^-1 s^-1
kr4a = KD4a * kf4a  # s^-1

# = kf1a, i.e., on rate is insensitive to presence of GEF

# Isomerization of Ras-RasGEF-GXP from loose to tight
kf4b = 20.4         # s^-1
kr4b = 3.9          # s^-1

The following study used purified HRAS and mouse RASGRF1:

[PMID9690470]: The mechanism of nucleotide release by the catalytic domain of murine Cdc25 (Cdc25Mm) has been investigated recently using fluorescently labelled nucleotides [PMID9585556]. The affinity of Cdc25Mm for nucleotide-free Ras (Kd = 4.6 nM) is found to be several orders of magnitude higher than that for nucleotide-bound Ras, and the maximal acceleration by Cdc25Mm of the rate of dissociation of nucleotide is more than 10^5.

[PMID9585556]: The best fit of our data resulted in similar quantum yields and a value of 4.6 nM for KD2 (NOTE: Kd between nucleotide-free H-Ras and RasGRF1). A variation in the value for KD2 of approximately 2-fold resulted in fits of comparable quality.

The activity of GEF (RASGRF1 in this case) does not depend on whether Ras (HRAS) is loaded with GTP or GDP.

[PMID9585556]: However, since the intrinsic dissociation rate of Ras for GTP (1 × 10-5 s-1) is 2-fold lower than that for GDP (2 × 10-5 s-1), the stimulatory action of Cdc25Mm285 is practically independent of the nature of the bound nucleotide.

[PMID9585556]: Although we did not reach complete saturation at 600 μM Ras‚nucleotide, the data could be fitted to obtain a maximal rate of 3′mdGDP release from Ras of 3.9 s-1 and an apparent Km value of 386 μM. Since the intrinsic dissociation rate of 3′mdGDP is 2 × 10-5 s-1 (Table 1), the acceleration of GDP dissociation from Ras by this GEF is approximately 2 × 10^5-fold. An apparent Km of approximately 300 μM was obtained for the triphosphate-bound form of Ras, confirming that there is no pronounced specificity toward the nature of the Ras-bound nucleotide (data not shown).


GEF binding to GTP bound Ras?

Can GEFs bind to Ras and cause ejection of nucleotide before the GTP/GDP conversion is complete? Moreover, if GEF binds to Ras-GTP, can the hydrolysis to GDP proceed while GEF is bound?

Instantiate the RasGEF cycle for HRAS and RASGRF1:

def rasgef_exchange_hras_rasgrf1(model):
    HRAS = model.monomers['HRAS']
    RASGRF1 = model.monomers['RASGRF1']
    GTP = model.monomers['GTP']
    GDP = model.monomers['GDP']
    ras_gef_exchange_cycle(HRAS, RASGRF1, GTP, GDP)

[PMID9585556]: Therefore, we tested the nucleotide specificity of the interaction of Cdc25Mm285 (CdcMm285 is the fragment of CdcMm/RasGRF1 containing the RasGEF domain) with Ras. Figure 1 shows the release of Ras-bound 3′mdGDP or 3′mdGTP (4 μM), in the presence of an excess of unlabeled nucleotide and in the presence or absence of 1 μM Cdc25Mm285. The Cdc25Mm285-stimulated dissociation rate of Ras-3′mdGDP is approximately twice that of Ras-3′mdGTP, with values of 0.0098 and 0.0046 s-1, respectively. However, since the intrinsic dissociation rate of Ras for GTP (1 × 10-5 s-1) is 2-fold lower than that for GDP (2 × 10-5 s-1), the stimulatory action of Cdc25Mm285 is practically independent of the nature of the bound nucleotide. The difference in stimulated dissociation rates is somewhat smaller than the results of Jacquet et al. (16) but is similar to the results with the yeast proteins CDC25 and RAS2 obtained by Haney and Broach (28).

[PMID9690470]: The overall shape of the catalytic domain of Sos is that of an oblong bowl (Fig. 2), with Ras bound at the centre of the bowl. The regions of Ras that interact most closely with Sos include the phosphate-binding P-loop (residues 10 – 17) and surrounding segments (including strand 􏰧1 and helix 􏰦1), the Switch 1 region (defined here as residues 25–40) and the Switch 2 region (defined here as residues 57 – 75). Additional interactions are seen with helix 3 (residues 95–105; Fig. 3a, b). The interface between Ras and Sos is primarily hydrophilic and very extensive, with 3,600 A^2 of surface area buried in the complex.

[PMID9690470]: The most obvious effect of Sos binding to Ras is the opening of the nucleotide binding site as a result of the displacement of Switch 1 of Ras by the insertion of the helical hairpin formed by aH and aI of Sos (Fig. 5)

Switch 1 and Switch 2 are the only regions of Ras in which structural changes are directly induced by Sos.

The change in the Switch 1 region of Ras when bound to Sos is drastic...Switch 1 is completely removed from the nucleotide-binding site.

One important aspect of the insertion of the helical hairpin of Sos into the Switch 1 region is that it does not result in a significant occlusion of the guanine and ribose binding sites (Fig. 5d). Instead, this structural distortion breaks the network of direct and water-mediated interactions between Switch 1 and the nucleotide. For example, in the nucleotide-bound forms of Ras, Phe 28 interacts with the guanine base through a perpendicular aromatic – aromatic interaction (Fig. 5a). Mutation of Phe28 to leucine results in a significant increase in the intrinsic rate of dissociation of nucleotide from Ras18. In the Sos complex, the Calpha of Phe 28 moves 9.6 A and the side chain no longer interacts with the nucleotide-binding site (Fig. 5b).

The Switch 2 region of Ras makes important interactions with GTP and not with GDP (19,46). Nevertheless, structural changes that are induced in Switch 2 by Sos result in the exclusion of both GDP and GTP, because they affect magnesium binding as well as the conformation of Lys 16 in the P-loop, a crucial phosphate ligand.

Specificity of RASGRF1 for Ras isoforms

[PMID9585556]: Three mammalian isoforms of Ras, H-, K-, and N-Ras, have been identified which are highly conserved intheirprimarysequence. Thesignificanceofhavingmore than one isoform is not understood at present, although the isoforms may have different functions in different tissues, since certain types of tumors have a preference for a particular activated Ras gene, such as K-Ras for lung, colon and pancreas cancers and N-Ras for myeloid leukemias (25). To see whether Cdc25Mm285 acts differently on the three isoforms, we tested the GEF activity of Cdc25Mm285 on these proteins. As summarized in Table 1, Cdc25Mm285 is active on all isoforms, being somewhat more active on N-Ras, in accordance with the results of Leonardsen et al. (26).