Computational docking and solution x-ray scattering predict a membrane-interacting role for the histone domain of the Ras activator son of sevenless

  1. Holger Sondermann*,†,‡,
  2. Bhushan Nagar*,†,§,
  3. Dafna Bar-Sagi, and
  4. John Kuriyan*,†,∥
  1. *Howard Hughes Medical Institute and Departments of Molecular and Cell Biology and Chemistry, University of California, Berkeley, CA 94720;†Physical Biosciences Division, Lawrence Berkeley National Laboratory, Berkeley, CA 94720; and Department of Molecular Genetics and Microbiology, State University of New York, Stony Brook, NY 11794

  1. Contributed by John Kuriyan, September 22, 2005

 
Next Section

Abstract

The Ras-specific nucleotide exchange factor son of sevenless (SOS) is a large, multidomain protein with complex regulation, including a Ras-dependent allosteric mechanism. The N-terminal segment of SOS, the histone domain, contains two histone folds, which is highly unusual for a cytoplasmic protein. Using a combination of computational docking, small-angle x-ray scattering, mutagenesis, and calorimetry, we show that the histone domain folds into the rest of SOS and docks onto a helical linker that connects the pleckstrin-homology (PH) and Dbl-homology (DH) domains of SOS to the catalytic domain. In this model, a positively charged surface region on the histone domain is positioned so as to provide a fourth potential anchorage site on the membrane for SOS in addition to the PH domain, the allosteric Ras molecule, and the C-terminal adapter-binding site. The histone domain in SOS interacts with the helical linker, using a region of the surface that in nucleosomes is involved in histone tetramerization. Adjacent surface elements on the histone domain that correspond to the DNA-binding surface of nucleosomes form the predicted interaction site with the membrane. The orientation and position of the histone domain in the SOS model implicates it as a potential mediator of membrane-dependent activation signals.

The small G protein Ras is a cellular switch that cycles between an active GTP-bound state and an inactive GDP-bound state. GTP and GDP are both slow to dissociate from Ras, and the action of nucleotide exchange factors is critical for converting inactive Ras·GDP to the active Ras·GTP form (1). The activation of growth factor receptors in animal cell signaling results in the recruitment of the Ras-specific nucleotide exchange factor son of sevenless (SOS) to the plasma membrane, where it engages Ras and causes dissociation of bound nucleotide (2). The unregulated activation of Ras is a frequent hallmark of many cancers, a fact that underscores the importance of strict control of the nature of the nucleotide bound to Ras (3).

SOS is a large protein (≈1,300 residues; see Fig. 1 for a schematic diagram of its domain structure) that promotes nucleotide release by binding to Ras and opening up its active site (4). The Rem (Ras exchanger motif) and Cdc25 (named for the Ras activator protein in yeast) domains, located in the C-terminal half of SOS, are the minimal elements required for nucleotide exchange activity. The catalytic site of SOS, where nucleotide release from Ras occurs, is located entirely within the Cdc25 domain (4).

Fig. 1.
View larger version:
Fig. 1.

Domain organization of SOS. The structures of SOScat (bound to two Ras molecules) and SOSDH-PH-cat are shown schematically.


We recently made the surprising discovery that Ras, which is the substrate of SOS, is itself required for SOS activity (5, 6). The binding of Ras·GTP to a distal site on the Cdc25 domain, bracketed by the Rem domain, stimulates nucleotide exchange activity. Perhaps more surprising was the realization that the binding of Ras·GDP to the distal allosteric site is required even for basal SOS activity (6). Disruption of the distal site, located ≈42 Å from the SOS active site, by mutation results in essentially complete shutdown of SOS activity because Ras can no longer bind allosterically.

The logic behind this dependence of SOS on Ras became partly clear upon analysis of a structure of a large fragment of SOS that contains roughly two-thirds of the N-terminal half of SOS in addition to the Rem and Cdc25 domains. This construct, referred to as SOSDH-PH-cat, contains a Dbl-homology (DH) domain and a pleckstrin-homology (PH) domain in addition to the catalytic Rem and Cdc25 domains (Fig. 1; see also Fig. 7B, which is published as supporting information on the PNAS web site). The DH-PH unit is located so as to block the allosteric Ras-binding site in the catalytic domain (6). The fact that the catalytic domain is inactive without Ras bound to the allosteric site, combined with the blockage of the allosteric site by the DH-PH unit, explains the autoinhibition of constructs of SOS that include the DH-PH unit (7–9). These results do not, however, provide an explanation for how the blockage of the critical allosteric Ras-binding site by the DH-PH unit is alleviated during the activation of SOS.

The primary anchorage of SOS to the membrane occurs by means of the docking of the C-terminal segment of SOS to adapter proteins that bind to activated receptors (10). Membrane targeting by and signals received via the N-terminal domains of SOS have also been reported to play a role in the activation of SOS (11, 12). It is therefore likely that both the N-terminal and C-terminal segments of SOS interact with the membrane at various points during the functional cycle of the molecule. The PH domain of SOS, located in the N-terminal region, binds to phosphoinositol phosphates (13, 14), although the precise identity of the PH ligand that results in membrane localization is unclear (11).

The very N-terminal segment of SOS contains clear sequence similarity to histones (sequence identity between the second histone fold of SOS and human histone H2A is 32%), a surprising feature in a nonnuclear protein (15). We have determined the structure of the ≈200-residue N-terminal segment and shown that it consists of two histone folds that form a pseudodimer with striking similarity to histone dimers (rms deviation in Cα positions of 1.2 Å over 144 residues) (16, 17). This instance appears to be the only known occurrence of a nonnuclear protein segment with sequence and structural similarity to histones (16). The sequence of the histone domain is conserved in SOS across metazoan species, suggesting an important functional role for the domain, but there are few clues as to what this function might be. The obvious possibility that the histone domain might serve to oligomerize SOS seems to be ruled out (16).

In this report, we show that the isolated histone domain of SOS binds to SOSDH-PH-cat with significant affinity (K d ≈ 2 μM) when added in trans. The interface within the intact molecule is likely to be tight, given the strength of the interaction in trans, but we have failed to obtain crystals of SOS constructs containing the histone domain (SOSHistone-DH-PH-cat). Computational docking of the histone domain onto SOSDH-PH-cat using the ClusPro docking server (18) yielded a top-scoring solution with intriguing attributes. The histone domain in this docked model bridges the DH, PH, and Rem domains. A key element of the interface is the tight interaction between an invariant arginine residue (Arg-552) in the helical linker between the PH and Rem domains and two invariant acidic residues (Asp-140 and Asp-169) in the histone domain. We screened alternative docking solutions against small-angle x-ray scattering (SAXS) data for SOSHistone-DH-PH-cat and show that the top-scoring model by the ClusPro criteria is the one that best fits the SAXS data. Mutation of Arg-552 or Asp-140 completely abolishes the interaction between the isolated histone domain and SOSDH-PH-cat, indicating that the docked model captures essential elements of the localization of the histone domain within SOS. The docked model places a prominent patch of positive electrostatic potential in the histone domain in an orientation where it can interact with lipid headgroups while allowing the simultaneous engagement of the membrane by the PH domain of SOS and two Ras molecules bound to it.

Previous SectionNext Section

Materials and Methods

Protein Expression and Purification. SOSHistone-DH-PH-cat (residues 1–1049), SOSDH-PH-cat (residues 198-1049), and SOSHistone (residues 1–198) of human SOS1 were cloned into the bacterial expression vector pProEx HTb (Invitrogen) with an N-terminal His-6 tag. Escherichia coli cells (BL21DE3, Novagen) were transformed with expression constructs and grown in Terrific Broth medium supplemented with 100 mg/ml ampicillin. Protein production was induced by addition of 1 mM IPTG at a cell density corresponding to an absorbance of 1 at 600 nm, and the protein was expressed at 18°C for 16 h. Cells were collected by centrifugation at 4,000 × g for1h, resuspended in Ni-NTA (nickel-nitrilotriacetic acid) buffer A (25 mM Tris·Cl, pH 7.5/500 mM NaCl/20 mM imidazole) containing protease inhibitors. Cell suspensions were lysed by French press (EmulsiFlex-C5, Avestin, Ottawa). Cell debris was removed by ultracentrifugation at 100,000 × g for1hat4°C.Clearsupernatants were loaded onto a Ni-NTA column (Qiagen, Valencia, CA) equilibrated in Ni-NTA buffer A. The resin was washed with 20 column volumes of the same buffer, and proteins were eluted in Ni-NTA buffer B (Ni-NTA buffer A supplemented with 500 mM imidazole). Buffers were exchanged by using a fast desalting column (Amersham Pharmacia) into tobacco etch virus (TEV) buffer (25 mM Tris·Cl, pH 8.3/50 mM NaCl/5 mM 2-mercaptoethanol). His-6 tags were cleaved by incubation with TEV protease. For SOSDH-PH-cat and SOSHistone-DH-PH-cat, the solution was loaded onto a MonoQ column (Amersham Pharmacia) equilibrated in MonoQ buffer A (25 mM Tris·Cl, pH 8.3/1 mM DTT). Proteins were eluted on a gradient from 0 to 500 mM NaCl over 20 column volumes. All proteins were further subjected to size exclusion chromatography on a Superdex200 column (Amersham Pharmacia) equilibrated in gel filtration buffer (25 mM Tris·Cl, pH 7.5/100 mM NaCl/1 mM DTT). Fractions containing protein were pooled and concentrated on a Centricon ultrafiltration device (Millipore) to a final concentration of ≈50 mg/ml. Protein aliquots were frozen in liquid nitrogen and stored at –80°C.

Point mutations (SOSHistone:R153A, SOSHistone:D140A, and SOSDH-PH-cat:R552A) were introduced by using the QuikChange XL mutagenesis kit (Stratagene) following the manufacturer's instructions and confirmed by DNA sequencing. Mutant proteins were expressed and purified as for wild-type proteins. All mutant proteins expressed to comparable levels (data not shown).

Isothermal Titration Calorimetry. Apparent dissociation constants (K d) and stoichiometry of interactions were measured by isothermal titration calorimetry (ITC) using a VP calorimeter (Microcal, Amherst, MA). Calorimetric titrations of SOSHistone (330–500 μM in the syringe) and SOSDH-PH-cat (33–50 μM in the cuvette) were carried out at 25°C in 25 mM Tris·Cl, pH 7.5/100 mM NaCl. SOSHistone:R153A or SOSHistone:D140A was titrated into SOSDH-PH-cat or SOSDH-PH-cat:R552A in 15-μl injections with a spacing between injections of 300 s.

ITC data were analyzed by integrating heat effects normalized to the amount of injected protein and curve-fitting based on a 1:1 binding model using the origin software package (Microcal, Northampton, MA). The dissociation constant was derived from the data by using standard procedures.

Molecular Docking. SOSDH-PH-cat (molecule B in Protein Data Bank entry 1XD4) and SOSHistone (molecule E in PDB entry 1Q9C) were docked by using the ClusPro docking server (18, 19) with the dot algorithm [without electrostatics (20)] and the present default parameters. We followed the recommended default procedure in which 20,000 models were retained after initial docking and ranked based on desolvation energies and electrostatics. The top 2,000 solutions were then clustered, and the clusters were ranked by using the ClusPro algorithm (19). In this procedure, the rank itself is the only metric provided for the quality of a solution because the clustering procedure removes a direct connection between a model and a scoring energy function. The procedure essentially treats the proteins as rigid bodies, except for an energy minimization step at the end using only van der Waals energy, which results in no change in backbone structure and minimal changes in side-chain positions.

SAXS. SAXS data were collected at beamline 12-IDC (BESSRCCAT beamline) at the Advanced Photon Source (Argonne, IL), essentially as described for an analysis of the AAA+ protein p97 (21). Scattering data were collected at protein concentrations of 1–10 mg/ml. Background scattering from the buffer alone was subtracted, and data were scaled by using the program primus (22). Scattering data were analyzed by using the programs gnom and crysol (23, 24). Radii of gyration (R g) were computed from the Guinier plot by using primus and also gnom and crysol. The pair–distance distribution function [P(r)] was calculated by using the program gnom. Calculated scattering curves and goodness of fit (χ) based on atomic models were obtained by using crysol. The χ parameter is defined as Formula where N is the number of data points, s is the magnitude of the scattering vector, c is a scale factor, I e is the experimental intensity, I c is the calculated intensity, and σ is the experimental error (24).

Previous SectionNext Section

Results and Discussion

Rationale for Docking the Isolated Histone Domain onto SOSDH-PH-cat.By using pull-down assays, an interaction between the histone domain and the PH domain of SOS has been demonstrated previously (25). We therefore investigated the interaction between the histone domain and various constructs of SOS by analytical gel filtration and found that the histone domain coelutes with SOSDH-PH-cat but not the DH-PH domain or SOSHistone-DH-PH-cat (data not shown).

We used isothermal titration calorimetry to measure the apparent binding constants for the interaction between SOSHistone and SOSDH-PH-cat. We have noted previously that the histone domains in the crystal form an open oligomeric interaction in which Arg-153 of one molecule forms hydrogen bonds with Asp-140 and Asp-169 of a neighboring molecule (Fig. 7A and ref. 16). Although the histone domains are monomeric at concentrations <≈0.2 mM, we wished to exclude potential complications from this interaction in the studies that follow. We therefore introduced a mutation Arg-153-Ala in SOSHistone. This mutation prevents the observed aggregation of SOSHistone at concentrations of >≈0.5 mM (data not shown). Arg-153 is not an invariant residue in SOS (e.g., the corresponding residue in the Drosophila sequence is an alanine).

Calorimetric titrations of SOSHistone:R153A into SOSDH-PH-cat yield an apparent K d of 2.5 μM for complex formation, with a stoichiometry of 1:0.91 (Fig. 2A). This interaction is a rather strong one, given that it is likely to occur intramolecularly in the context of full-length SOS (there is no evidence that SOSHistone-DH-PH-cat forms dimers or higher order assemblies; data not shown).

Fig. 2.
View larger version:
Fig. 2.

Isothermal titration calorimetry data for SOSHistone binding to SOSDH-PH-cat. (A) Binding of SOSHistone:R153A to SOSDH-PH-cat. Calorimetric titration for SOSHistone:R153A(500μM) titrated into SOSDH-PH-cat (50μM) is shown. Derived values for K d and stoichiometry (N) are shown. (B) Binding of SOSHistone:D140A to SOSDH-PH-cat. Calorimetric titration for SOSHistone:D140A (500 μM) titrated into SOSDH-PH-cat (50μM) is shown. (C) Binding of SOSHistone:R153A to SOSDH-PH-cat:R552A. Calorimetric titration for SOSHistone:R153A (500 μM) titrated into SOSDH-PH-cat:R552A (50 μM) is shown. Values for K d and stoichiometry (N) were not determined in B and C because of weak binding.


Given the evidence for a direct interaction between the histone domain and SOSDH-PH-cat, we looked for possible modes of interaction by visual inspection of molecular shapes and electrostatic potentials. Our attention was drawn to an invariant arginine residue (Arg-552) in the helical linker connecting the PH domain to the Rem domain (Figs. 3A and 7B). This residue serves no obvious structural or mechanistic function and is completely accessible. It was immediately obvious from manual docking that the side chain of Arg-552 could interact with the side chains of the invariant residues Asp-140 and Asp-169 in the histone domain in a manner analogous to that seen between Arg-153 and the same acidic residues in the crystal lattice of the histone domain (Fig. 7A and ref. 16). Such an interaction positions the histone domain between the PH, DH, and Rem domains with no steric clashes. This docking exercise took no account of the energetics of the interaction, and we therefore turned to a computational docking procedure for a more objective analysis.

Fig. 3.
View larger version:
Fig. 3.

Docking of SOSHistone onto SOSDH-PH-cat. (A) Crystal structure of SOSDH-PH-cat. The structure of SOSDH-PH-cat (molecule B in PDB entry 1XD4) is shown with coloring according to the diagram shown in Fig. 1 (6). A surface-exposed, strictly conserved arginine (Arg-552) located in the helical linker between the PH and Rem domains is shown. (B) Crystal structure of the histone domain of SOS. The structure of SOSHistone (molecule E in PDB entry 1Q9C) is shown (16). The histone folds are shown in green and yellow. (C) Model of SOSHistone-DH-PH-cat. The structure of SOSHistone was docked onto the structure of SOSDH-PH-cat as described in the text (18). A close-up view of the SOSHistone:SOSDH-PH-cat interface is shown. Domains are shown colored according to the diagram in Fig. 1.


Computational Docking of the Histone Domain onto SOSDH-PH-cat. We submitted the crystal structures of SOSDH-PH-cat (Fig. 3A) and SOSHistone (Fig. 3B) obtained by x-ray crystallography to a docking server (ClusPro) (see Materials and Methods and ref. 18). We placed no restrictions on the possible docking modes and did not bias the calculations to consider Arg-552, Asp-140, or Asp-169 in any particular way. The docking procedure, which relies on complementarity of shape, does not use information regarding the conservation of residues in the proteins. We ran the docking search twice with the same input structures, recovering 10 solutions in one and 30 in the other. The top-scoring solution by the docking criteria was essentially the same in both runs (rms deviation of 0.8 Å over all atoms) and was also the top-ranked solution by SAXS (see below). The coordinates of the top-ranked docked model are provided as Dataset 1, which is published as supporting information on the PNAS web site. (Note that the residue numbering does not correspond to the sequence of human SOS.)

Most strikingly, the top-scoring model exhibits an arrangement as proposed above, with the histone domain packing against the helical linker between the PH and Rem domains (Fig. 3C). In particular, the detailed interactions seen in this docked model include clamping of Arg-552 in the helical linker by the two invariant aspartate residues (Asp-140 and Asp-169) of the histone fold domain. There is general electrostatic complementarity between the histone domain and a concave surface formed by the DH, PH, and Rem domains (Fig. 4; see also Fig. 8, which is published as supporting information on the PNAS web site).

Fig. 4.
View larger version:
Fig. 4.

Surface of the docked histone domain. The docked model for SOSHistone-DH-PH-cat is shown, with the histone domain shown in surface representation. The electrostatic potential of the histone domain is mapped onto its molecular surface, with red representing negative potential and blue representing positive potential (–5 to +5 k B T). Electrostatic potentials were calculated by using the program grasp (27), with an ionic strength corresponding to 100 mM KCl.


In our previous analysis of the structure of the histone dimer, we had noted the presence of a highly conserved patch of surface residues (16). This region is precisely where the docking procedure places the helical linker from SOSDH-PH-cat. We aligned the histone domain of SOS onto histones H3/H4 (subunits E and F) of the nucleosome core (17). In such an alignment, an interfacial α-helix from the other H3/H4 dimer (subunit H) in the nucleosome packs onto the region corresponding to the conserved patch on the histone domain of SOS. The alignment results in overlap between the helical linker in SOSDH-PH-cat and the interfacial helix from the H3/H4 tetramer, although the two helices are oriented differently (not shown).

The docking procedure was not constrained in terms of the polypeptide connectivity in full-length SOS. Nevertheless, the separation between the C-terminal residue of the histone domain in the docked model and the N terminus of the DH domain is ≈25 Å (Fig. 9 and Table 1, which are published as supporting information on the PNAS web site). This gap can be bridged readily by the 14 residues in the linker between them, which is not present in crystal structures and is presumed to be flexible.

Validation of the Docking Model by SAXS. The x-ray scattering pattern of a monodisperse protein solution can be used to calculate the radius of gyration (R g) and the maximum dimensions of the molecule (D max) (26). The scattering curve is also sensitive to the overall shape of a molecule or complex, and structural models can be scored against the SAXS data. SAXS data can also be used to obtain molecular shape reconstructions (26), but because of the large size and highly asymmetric shape of SOS, we have not used such procedures here.

We collected x-ray scattering data from solutions of SOSDH-PH-cat and SOSHistone-DH-PH-cat (Fig. 5A). Both proteins are monomeric in solution with low polydispersity as revealed by static multiangle light scattering (data not shown). Despite significant differences in the scattering profiles, the values of R g and D max for SOSDH-PH-cat and SOSHistone-DH-PH-cat are similar (see below), suggesting a similar degree of compactness and overall shape of the two proteins. The distance distribution [P(r)] functions for the two molecules, calculated from the primary scattering data, are shown in Fig. 5B.

Fig. 5.
View larger version:
Fig. 5.

SAXS data for SOSDH-PH-cat and SOSHistone-DH-PH-cat. (A) Scattering profile from solutions of SOSDH-PH-cat and SOSHistone-DH-PH-cat. X-ray scattering curves of SOSDH-PH-cat (green) and SOSHistone-DH-PH-cat (black) are shown. Values for the radius of gyration (R g) obtained from a Guinier plot using the program primus (22, 28) are shown. As a comparison, the value of R g calculated from the crystal structure of SOSDH-PH-cat by using the program crysol (24) is also shown. (B) Distance distribution function [P(r)] for SOSDH-PH-cat and SOSHistone-DH-PH-cat. P(r) for SOSDH-PH-cat (green) and SOSHistone-DH-PH-cat (black) calculated from the scattering data shown in A are shown. Values for R g and D max computed by using the program gnom are listed (23). (C) Comparison of experimental and calculated scattering curves for SOSDH-PH-cat. The experimental scattering curve from SOSDH-PH-cat (green) and that obtained from the atomic structure (orange), computed by using crysol, are shown. R g calculated by using crysol and the goodness of fit (χ) are indicated. (D) Evaluation of docking results using the solution scattering profile from SOSHistone-DH-PH-cat. Three models obtained by docking (ClusPro) of the structure SOSHistone onto SOSDH-PH-cat are shown. Model 1 is identical to the one shown in Fig. 2. Models were analyzed as described in C. s = 2π(1/d), where d is the Bragg spacing.


To validate our x-ray scattering data, we evaluated the fit between the experimental scattering data for SOSDH-PH-cat and the known crystal structure (Fig. 5C). The observed scattering curve is in good agreement with the curve calculated from the crystal structure of SOSDH-PH-cat (the value of the goodness of fit, χ, is 4.76). The χ value is a normalized residual coefficient (see Materials and Methods) calculated by the program crysol (24). Experiments with small proteins, such as lysozyme, yield values of χ that are close to 1 (24). For large protein molecules or assemblies, such as the nonreceptor tyrosine kinase Abl or the association domain of the Ca2+/calmodulin-dependent kinase II, it is our experience that the values of χ are comparable to the value obtained here for SOSDH-PH-cat, presumably because of segmental flexibility (B.N., O. Hantschel, J. M. Davies, W. I. Weis, G. Superti-Furga, and J.K., unpublished data; O. S. Rosenberg, S. Deindl, R.-J. Sung, A. C. Nairn, and J.K. unpublished data).

We evaluated the fits of the theoretical scattering profiles computed from the best models obtained by computational docking to the experimental scattering pattern (Fig. 5D; see also Fig. 10, which is published as supporting information on the PNAS web site). Although the calculated R g values for the models cluster in a narrow range (between 40.8 and 41.0 Å for the solvated docking models fitted to the experimental scattering data), the fit of the models to the scattering data varies significantly, as given by the value of the residual χ. The model that is most compatible with the experimental data is the top-scoring docking model shown in Fig. 3 (χ = 4.78). The goodness of the fit is very similar to the one computed for SOSDH-PH-cat, for which an accurate atomic model is available.

The scattering profiles for the other docking models diverge significantly from the experimental scattering data (Figs. 5D and 10 and Table 1). The details of the deviations of the computed scattering profile from the experimental data are distinct for each model, indicating the sensitivity of the measurements to conformational differences.

Mutation of Conserved Residues Predicted to Be at the Interface. As noted above, a key feature of the top-scoring docking model is the formation of ion pairing interactions between the invariant residues Arg-552 (in the PH-Rem helical linker) and Asp-140 and Asp-169 (both in the histone domain). Asp-169 is a helix capping residue (see Figs. 3C and 7A), and mutation of this residue leads to loss of protein expression (data not shown). Arg-552 and Asp-140, in contrast, are surface-exposed and make no significant interactions in SOSDH-PH-cat and SOSHistone, respectively. We mutated both residues to alanine and tested the effects of the mutations on binding. SOSHistone:D140A and SOSDH-PH-cat:R552A both behaved normally during purification and concentration, indicating no adverse effects on stability. As shown in Fig. 2, mutation of either Arg-552 or Asp-140 completely abolishes the interaction between SOSHistone and SOSDH-PH-cat, as measured by calorimetry.

Taken together, the SAXS data and the binding analysis suggest that the docked model for SOSHistone-DH-PH-cat is correct in terms of its general orientation.

Implications for Membrane Interaction. We examined the surface electrostatic potential of SOSHistone-DH-PH-cat, which reveals a region with markedly positive electrostatic potential at the surface of the histone domain. The patch of positive potential extends from the phosphatidylinositol phosphate-binding site on the PH domain (13, 14) to the most distal tip of the histone domain (Fig. 6 A and B). In contrast, the opposite face of the molecule shows a markedly negative electrostatic potential (Fig. 11, which is published as supporting information on the PNAS web site).

Fig. 6.
View larger version:
Fig. 6.

Model for SOS localization at the membrane. (A) Model for SOS localization at the membrane. The model for SOSHistone-DH-PH-cat is shown oriented at the membrane. The phosphatidylinositol phosphate-binding site [with inositol phosphate modeled from the structure of the Dapp1 PH domain (29)] is indicated. Note that the histone domain presents a conspicuous region of positive electrostatic potential on the surface, oriented toward the hypothetical location of the membrane. The Ras-binding sites are indicated based on a structural superposition of the SOSHistone-DH-PH-cat model with a ternary Ras:SOScat crystal structure (ref. 5; PDB ID code 1NVV). (B) Electrostatic potential of SOSHistone-DH-PH-cat. Electrostatic potentials were calculated as described in the legend of Fig. 4. (C) Superposition of the nucleosome with SOSHistone-DH-PH-cat. SOSHistone-DH-PH-cat is shown with the electrostatic potential mapped onto its molecular surface. The nucleosome (17) is shown in gray, with DNA in green. For the superpositioning, one nucleosomal H3/H4 dimer was aligned with the histone domain of SOS (see the text).


The docked model can be aligned so that the positively charged region on the surface of the histone domain faces the membrane while allowing the simultaneous interaction of the PH domain with the membrane (Fig. 6A). This alignment might provide an extensive interface for the interaction of SOS with negatively charged phospholipids such as phosphatidylserine, thereby increasing the local concentration or residence time of SOS at the membrane. Such an orientation of SOS at the membrane also enables it to interact with membrane-tethered Ras at both the distal (allosteric) and active sites (Fig. 6A).

Previous SectionNext Section

Concluding Remarks

The realization that the N-terminal segment of SOS contains histone folds raised the possibility that this region of SOS is involved in the formation of higher order structure. Such a role for the histone folds seems unlikely, at least for homotypic interactions, because SOS is monomeric (16). Evidence from computational docking, SAXS, and mutagenesis support a model in which the histone domain folds into the main body of SOS and docks onto a helical linker connecting the PH domain to the Rem domain while straddling the DH domain. Strikingly, this model places the helical linker on the histone domain at a location that corresponds to the docking site on one nucleosomal histone dimer (e.g., H3/H4) for another histone dimer (e.g., the other H3/H4 dimer). This observation suggests that the transposition of the pseudohistone dimer in SOS from an ancestral nucleosomal-like protein to its present location has preserved the histone tetramerization interface for utilization in intramolecular rather than intermolecular interactions.

The connection to the functional aspects of histone domains in nucleosomes extends beyond the histone–helical linker interface. A notable feature in the docked model is the presence of a patch of positive electrostatic potential on the surface of the histone fold. Particularly suggestive is the fact that if the model is oriented so as to position this patch toward the membrane, then the three other anchor points to the membrane (the PH domain and the two Ras-binding sites) are also aligned appropriately. Fig. 6C shows the results of aligning the nucleosome onto the docked SOS model, using one H3/H4 dimer and the histone domain of SOS as guides. The DNA from the nucleosome tracks along the exposed patch of positive electrostatic potential on the histone domain (Fig. 6C). SOS is clearly not a DNA-binding protein (note the predominance of negative electrostatic potential on the surface in Figs. 6B and 9). Nevertheless, it seems that the region we have identified as a potential interaction surface for negatively charged lipid headgroups is a remnant of the DNA-binding capabilities of nucleosomes.

It is possible that the structure of intact SOS will continue to elude crystallographic analysis because of the segmental nature of the molecular assembly. Our work points to the growing power of combining computational methods with SAXS measurements to analyze supramolecular complexes in solution. The SAXS data have helped validate a structural model that provides a unique twist on the functional role of histone domains in proteins. Although these results leave unanswered the question of how the N-terminal autoinhibition of SOS is alleviated, the availability of our model for SOS structure should spur experiments aimed at elucidating the role of the histone domain in controlling SOS activity at the membrane.

Previous SectionNext Section

Acknowledgments

We are grateful to the scientists at beamline 12-IDC (Advanced Photon Source, Argonne, IL) for assistance with synchrotron data collection and Jason Davies and William Weis for helping to initiate the SAXS studies. We thank Oren Rosenberg, Tanya Weitze, and Jodi Gureasko for valuable discussions and Patricia Pellicena for critically reading the manuscript. This work was supported by the Leukemia and Lymphoma Society (H.S.) and the National Institutes of Health (D.B.-S.).

Previous SectionNext Section

Footnotes

  • ↵ ∥ To whom correspondence should be addressed. E-mail: kuriyan{at}berkeley.edu.

  • ↵ ‡ Present address: Department of Molecular Medicine, College of Veterinary Medicine, Cornell University, Ithaca, NY 14853.

  • ↵ § Present address: Department of Biochemistry, McGill University, Montreal, QC, Canada H3A 2T5.

  • Author contributions: H.S. and J.K. designed research; H.S. performed research; H.S., B.N., D.B.-S., and J.K. analyzed data; and H.S. and J.K. wrote the paper.

  • Conflict of interest statement: No conflicts declared.

  • Abbreviations: SOS, son of sevenless; SAXS, small-angle x-ray scattering; DH, Dbl homology; PH, pleckstrin homology; Rem, Ras exchanger motif; NTA, nitrilotriacetic acid.

  • Freely available online through the PNAS open access option.

Previous Section
 

References

  1. ↵
    Vetter, I. R. & Wittinghofer, A. (2001) Science 294 , 1299–1304.pmid:11701921
    Abstract/FREE Full Text
  2. ↵
    Hall, B. E., Yang, S. S. & Bar-Sagi, D. (2002) Front. Biosci. 7 , d288–d294.pmid:11779695
    Medline
  3. ↵
    Coleman, M. L., Marshall, C. J. & Olson, M. F. (2004) Nat. Rev. Mol. Cell. Biol. 5 , 355–366.pmid:15122349
    CrossRefMedlineISI
  4. ↵
    Boriack-Sjodin, P. A., Margarit, S. M., Bar-Sagi, D. & Kuriyan, J. (1998) Nature 394 , 337–343.pmid:9690470
    CrossRefMedline
  5. ↵
    Margarit, S. M., Sondermann, H., Hall, B. E., Nagar, B., Hoelz, A., Pirruccello, M., Bar-Sagi, D. & Kuriyan, J. (2003) Cell 112 , 685–695.pmid:12628188
    CrossRefMedlineISI
  6. ↵
    Sondermann, H., Soisson, S. M., Boykevisch, S., Yang, S. S., Bar-Sagi, D. & Kuriyan, J. (2004) Cell 119 , 393–405.pmid:15507210
    CrossRefMedlineISI
  7. ↵
    Corbalan-Garcia, S., Margarit, S. M., Galron, D., Yang, S. S. & Bar-Sagi, D. (1998) Mol. Cell. Biol. 18 , 880–886.pmid:9447984
    Abstract/FREE Full Text
  8. Kim, J. H., Shirouzu, M., Kataoka, T., Bowtell, D. & Yokoyama, S. (1998) Oncogene 16 , 2597–2607.pmid:9632136
    CrossRefMedlineISI
  9. ↵
    Qian, X., Vass, W. C., Papageorge, A. G., Anborgh, P. H. & Lowy, D. R. (1998) Mol. Cell. Biol. 18 , 771–778.pmid:9447973
    Abstract/FREE Full Text
  10. ↵
    Schlessinger, J. & Bar-Sagi, D. (1994) Cold Spring Harbor Symp. Quant. Biol. 59 , 173–179.pmid:7587067
    MedlineISI
  11. ↵
    Chen, R. H., Corbalan-Garcia, S. & Bar-Sagi, D. (1997) EMBO J. 16 , 1351–1359.pmid:9135150
    CrossRefMedlineISI
  12. ↵
    Byrne, J. L., Paterson, H. F. & Marshall, C. J. (1996) Oncogene 13 , 2055–2065.pmid:8950972
    Medline
  13. ↵
    Zheng, J., Chen, R. H., Corblan-Garcia, S., Cahill, S. M., Bar-Sagi, D. & Cowburn, D. (1997) J. Biol. Chem. 272 , 30340–30344.pmid:9374522
    Abstract/FREE Full Text
  14. ↵
    Koshiba, S., Kigawa, T., Kim, J. H., Shirouzu, M., Bowtell, D. & Yokoyama, S. (1997) J. Mol. Biol. 269 , 579–591.pmid:9217262
    CrossRefMedlineISI
  15. ↵
    Baxevanis, A. D., Arents, G., Moudrianakis, E. N. & Landsman, D. (1995) Nucleic Acids Res. 23 , 2685–2691.pmid:7651829
    Abstract/FREE Full Text
  16. ↵
    Sondermann, H., Soisson, S. M., Bar-Sagi, D. & Kuriyan, J. (2003) Structure (Cambridge, MA) 11 , 1583–1593.
  17. ↵
    Luger, K., Mader, A. W., Richmond, R. K., Sargent, D. F. & Richmond, T. J. (1997) Nature 389 , 251–260.pmid:9305837
    CrossRefMedline
  18. ↵
    Comeau, S. R., Gatchell, D. W., Vajda, S. & Camacho, C. J. (2004) Nucleic Acids Res. 32 , W96–W99.pmid:15215358
    Abstract/FREE Full Text
  19. ↵
    Comeau, S. R., Gatchell, D. W., Vajda, S. & Camacho, C. J. (2004) BioInformatics 20 , 45–50.pmid:14693807
    Abstract/FREE Full Text
  20. ↵
    Mandell, J. G., Roberts, V. A., Pique, M. E., Kotlovyi, V., Mitchell, J. C., Nelson, E., Tsigelny, I. & Ten Eyck, L. F. (2001) Protein Eng. 14 , 105–113.pmid:11297668
    Abstract/FREE Full Text
  21. ↵
    Davies, J. M., Tsuruta, H., May, A. P. & Weis, W. I. (2005) Structure (Cambridge, MA) 13 , 183–195.
  22. ↵
    Konarev, P. V., Volkov, V. V., Sokolova, A. V., Koch, M. H. J. & Svergun, D. I. (2003) J. Appl. Crystallogr. 36 , 1277–1282.
    CrossRefISI
  23. ↵
    Svergun, D. I. (1992) J. Appl. Crystallogr. 25 , 495–503.
    CrossRefISI
  24. ↵
    Svergun, D., Barberato, C. & Koch, M. H. J. (1995) J. Appl. Crystallogr. 28 , 768–773.
    CrossRefISI
  25. ↵
    Jorge, R., Zarich, N., Oliva, J. L., Azanedo, M., Martinez, N., de la Cruz, X. & Rojas, J. M. (2002) J. Biol. Chem. 277 , 44171–44179.pmid:12223473
    Abstract/FREE Full Text
  26. ↵
    Koch, M. H., Vachette, P. & Svergun, D. I. (2003) Q. Rev. Biophys. 36 , 147–227.pmid:14686102
    CrossRefMedlineISI
  27. ↵
    Nichols, A., Sharp, K. & Honig, B. (1991) Proteins 11 , 281–296.pmid:1758883
    CrossRefMedlineISI
  28. ↵
    Guinier, A. & Fournet, G. (1955) Small-Angle Scattering of X-Rays (Wiley, New York).
  29. ↵
    Ferguson, K. M., Kavran, J. M., Sankaran, V. G., Fournier, E., Isakoff, S. J., Skolnik, E. Y. & Lemmon, M. A. (2000) Mol. Cell 6 , 373–384.pmid:10983984
    CrossRefMedlineISI
« Previous | Next Article »Table of Contents

This Article

  1. Published online before print November 2, 2005, doi: 10.1073/pnas.0508315102 PNAS November 15, 2005 vol. 102 no. 46 16632-16637
  1. AbstractFree
  2. Figures Only
  3. » Full Text
  4. Full Text (PDF)
  5. Supporting Information

Classifications

Navigate This Article

About Our New Site Design >>
Link: Advanced Search

This Week's Issue

  1. August 19, 2008, 105 (33)
  1. Current Issue
  1. Alert me to new issues of PNAS

Most