Supplementary Information

Transcription

1 Gureaso et al., Supplementary Information Supplementary Information Membrane-dependent Signal Integration by the Ras ctivator Son of Sevenless Jodi Gureaso, William J. Galush, Sean Boyevisch, Holger Sondermann, Dafna Bar-Sagi, Jay T. Groves and John Kuriyan

2 Gureaso et al., Supplementary Information Supplementary Figure Lipid attachment does not affect the intrinsic nucleotide release rates from Ras. To ensure that membrane anchorage of Ras does not alter the intrinsic rate of nucleotide release, we too advantage of an isolated dc5 domain construct of the SOS homolog Ras Guanine Nucleotide Releasing Factor (RasGRF), which is active in the absence of an allosteric activator. The rate of fluorescently labeled mant-dgdp release from Ras in solution and membrane-bound Ras (~, molecules µm ) in the presence of SOS cat and RasGRF dc5 is compared (mant-dgdp echanged for GDP). The bul volume concentration of Ras, RasGRF dc5, and SOS cat is µm (when present). Note that membrane attachment does not alter the intrinsic nucleotide release rate from Ras alone or the RasGRF dc5 -catalyzed rate.

3 Gureaso et al., Supplementary Information Supplementary Figure The enhanced affinity of Ras-GTP for the allosteric site of SOS underlies the positive feedbac activation of Ras by SOS. The rates for SOS cat - catalyzed nucleotide echange for Ras coupled to lipid vesicles when fluorescently labeled mant-dgdp is echanged for either unlabeled GDP or GTP are compared for a given surface density of Ras. lthough the rates are shown for three Ras surface densities, the bul volume concentrations of Ras and SOS are the same for each measurement ( µm). Note that the replacement of GDP for GTP on Ras results in a ~- fold increase in the rate of SOS cat -catalyzed nucleotide release for an equal surface density of membrane-bound Ras.

4 Gureaso et al., Supplementary Information Supplementary Figure 3 PH domain-dependent binding of SOS to PIP. The binding of SOS proteins to lipid vesicles composed of a miture of phosphatidylcholine (P), phosphatidylserine (PS) and the relevant phospholipid species at ratios of :: is assayed using a sucrose density gradient flotation assay (see Supplementary Methods). Binding of the PH domain containing SOS constructs, SOS DP and SOS HDP, to vesicles containing PIP is observed. No binding to vesicles containing PIP is observed for a mutant form of SOS DP, SOS DP(K5E R59E), in which the two basic residues that are critical for PIP binding are replaced by glutamate. The results shown are representative of five independent eperiments.

5 Gureaso et al., Supplementary Information Supplementary Figure Lateral mobility of Ras-coupled supported lipid bilayers. Membrane-lined Ras mobility is detected by fluorescence recovery after photobleaching (FRP) eperiments. High intensity illumination is used to bleach a small region of the membrane, bleaching most of the BODIPY-GDP fluorophores. fter 9 minutes the bleached fluorophores have laterally diffused out of the spot, as indicated by intensity traces along the yellow line. [White bar = µm]

6 Gureaso et al., Supplementary Information Supplementary Figure 5 Schematic of supported lipid bilayer eperimental setup. planar bilayer in a µl well presents mobile, lipid-lined Ras to the solution (left). small region of bilayer is removed by scratching with a needle (center, bar = µm). Upon addition of SOS, signal from the bilayer decays with time, while the scratched region controls for signal from fluorescent nucleotide in solution (right).

7 Gureaso et al., Supplementary Information Supplementary Table alculation of Ras molecules per vesicle and Ras surface density (molecules µm ) mol% Ras-coupled maleimide-lipids Number of Ras molecules per vesicle Surface density per vesicle (Ras µm ). 3 mol% Ras-coupled maleimide-lipids is calculated using final protein and lipid concentrations after Ras-conjugation (see Methods). The number of Ras molecules per vesicle and the surface density of Ras (Ras molecules per µm ) are calculated based on the following assumptions: Radius of vesicle 5 nm, surface area of vesicle = πr 3 nm (.3 µm ) Surface area of lipid. nm, number of surface lipids per vesicle 3.9 Number of Ras molecules per vesicle = (mol% Ras-coupled lipids/) (surface lipids/vesicle) (.) (3.9 ) 39 Ras molecules/vesicle Surface density of Ras molecules per µm = (Ras molecules/vesicle)/(surface area of vesicle) (39 Ras/vesicle)/(.3 µm /vesicle) 3 Ras/µm

8 Gureaso et al., Supplementary Information Supplementary Methods Liposome Binding ssays The binding of SOS proteins to phospholipid vesicles composed of a miture of phosphatidylcholine (P), phosphatidylserine (PS) and the relevant phospholipid species at ratios of :: was assayed using flotation through a sucrose density gradient to separate protein-associated vesicles from free protein. Phospholipid vesicles ( µg total lipid, corresponding to ~ µm lipid of interest) were incubated with SOS proteins ( ng) at room temperature for min. Binding reactions were mied with 5 µl of.5 M sucrose in binding buffer (5 mm HEPES-NaOH [ph.] and mm Nal) to mae a final concentration of. M sucrose. µl sample of this miture was layered onto the bottom of a polycarbonate ( mm) centrifuge tube (Becman) and overlaid with µl of.5 M sucrose in binding buffer followed by µl of binding buffer. entrifugation was carried out at, g for min at room temperature in a TL rotor (Becman). Phospholipid vesicles were recovered by flotation through the gradient and phospholipid-associated proteins were analyzed using SDS PGE followed by silver staining (SilverQuest, Invitrogen) after normalizing for lipid recovery.

9 Gureaso et al., Supplementary Information Supplementary Discussion Ras/SOS reaction networ model model was constructed to simulate the interaction of membrane-bound Ras with SOS, where SOS is able to bind two Ras molecules, but only catalyzes nucleotide echange at the catalytic site. We use a convention where Ras SOS indicates Ras bound in the allosteric site, SOS Ras indicates Ras bound in the catalytic site, and Ras SOS Ras indicates Ras in both allosteric and catalytic sites. The Ras-bound nucleotide is indicated by a * or, where Ras* is a protein loaded with fluorescent nucleotide, and Ras is a protein with non-fluorescent nucleotide. Our model contains the following reactions: Ras*SOS - Ras* SOS SOS Ras* - SOS Ras* Ras* SOS Ras* - Ras* SOS Ras* Ras*SOS Ras* - Ras* SOS Ras* Ras SOS - Ras SOS SOS Ras SOS Ras - Ras SOS Ras Ras SOS Ras - Ras SOS Ras Ras SOS Ras - Ras* SOS Ras Ras* SOS Ras - Ras SOS Ras* - Ras SOS Ras* Ras SOS Ras* - Ras SOS Ras*

10 Gureaso et al., Supplementary Information Ras*SOS Ras Ras* SOS Ras - fast Ras* SOS Ras* Ras* SOS Ras SOS Ras* slow SOS Ras fast Ras SOS Ras* Ras SOS Ras where the rate constants are: = binding to allosteric site (empty catalytic site) = unbinding from allosteric site (empty catalytic site) = binding to catalytic site (empty allosteric site) = unbinding from catalytic site (empty allosteric site) = binding to allosteric site (filled catalytic site) = unbinding from allosteric site (filled catalytic site) = binding to catalytic site (filled allosteric site) = unbinding from catalytic site (empty allosteric site) slow = nucleotide echange rate constant with no allosteric Ras bound fast = nucleotide echange rate constant with allosteric Ras bound and the or subscripts indicate the absence or presence, respectively, of Ras in the nonechanging, allosteric binding site. Thus, is the forward rate constant to bind Ras to the allosteric site of SOS, while is the forward rate constant for Ras binding to the allosteric site of SOS Ras (i.e. SOS with a Ras already in the catalytic site). ll rate constants have units of s (discussed below). Variables fast and slow are pseudo-first order constants for an ecess of unlabeled nucleotide. In order to have a unified scheme that deals with both solution and membrane-bound components, we consider all species and rates in terms of the absolute number of molecules rather than their volume or surface concentrations. Thus, for unimolecular reactions, rates are epressed in the form d i = i () l where i is number of molecules i and l is the rate constant for the reaction in units of s. For bimolecular interactions, the rate of reaction of species i is simply proportional to: i) the number i, ii) the probability of encountering a molecule of type j, and iii) a rate constant establishing the probability of the reaction occurring. This leads to bimolecular reaction rates of the form

11 Gureaso et al., Supplementary Information di = P () i j l where i is number of molecules i, P j is the probability of encountering a molecule of species j (given by j s volume or area fraction of the maimum close-paced density), and the rate constant l in units of s. It is important to note that this analysis avoids the need to consider rate constants of mied dimensionality that relate interactions between solution and surface species. dditionally, all rate constants are directly intercomparable, regardless of what chemical components are involved. s a matter of clarification, the forward bimolecular reaction i j Product is often epressed as d[ i ] = i j molar (3) where i is the usual molar concentration of i, and molar is the usual molar bimolecular rate constant. Note that the left-hand side of (3) can be re-epressed as d[ i ] = d i () N V = d i N V (5) where i is the number of molecules of i, N is vogadro s number and V is the volume (assumed to be liter). Our molecular-based formalism applied to the same bimolecular reaction given by equation () can be written as d i = i j j ma l () where jma is the maimum close paced number of j in liter and l is once again the forward rate constant in units of s. ombining equations (5) and (), d[ i ] so, = N V i j l = i j molar () j ma N V N V

12 Gureaso et al., Supplementary Information l N V = molar () j ma s an eample, converting the value of below into molar concentration terms, one finds that the dissociation constant for Ras binding to the allosteric site ( / ) is in the micromolar range. We chose rate constants for the simulation that qualitatively describe the rate phenomena present in both vesicle and supported bilayer eperiments. We have observed that BODIPY-GDP echanges several fold faster than mant-dgdp, thus we set the rate constant of nucleotide echange on supported bilayers to 3X the value used for vesicle simulations. We chose this ratio to best describe the behavior of both systems. The specific rate constants chosen were: = s = s = s = s = s = s = s = s slow = s fast = 5 s (vesicle); 5 s (supported bilayer) The above model leads to the following eleven differential equations, which were solved using Matlab (MathWors) for the real eperimental conditions for both the vesicle and supported bilayer systems. We analyzed the behavior of the sum of the fluorescent Ras species with respect to time to simulate the eperimental observable. The term P j (the probability of molecule i encountering molecule j) is the volume fraction of j, given by j divided by jma, where jma is the maimum close paced number of j in the volume or area. d d = = ma ma ma 3 ma ma 3 3ma 5 ma 3 ma ma ma 5 d3 = ma ma 3ma

13 Gureaso et al., Supplementary Information slow d ma 5 ma ma = fast d 5 5 ma 5 3ma 3 5 = ma 3ma 3 9 ma 9 ma ma ma d = ma 9 ma ma d = slow d ma 9 ma ma = fast d 9 ma 9 ma 9 = fast d 5 ma 3ma 3 = fast d ma ma = where = Ras* = SOS 3 = Ras* SOS = SOS Ras* 5 = Ras* SOS Ras* = Ras = Ras SOS = SOS Ras 9 = Ras SOS Ras* = Ras* SOS Ras = Ras SOS Ras

14 Gureaso et al., Supplementary Information References. Miller, E., ntonny, B., Hamamoto, S. & Scheman, R. argo selection into OPII vesicles is driven by the Secp subunit. Embo J, 5-3 ().. Freedman, T.S. et al. Ras-induced conformational switch in the Ras activator Son of sevenless. Proc Natl cad Sci U S 3, 9- ().