Structural Insights into Formation of an Active Signaling Complex ...

Molecular Cell

Article Structural Insights into Formation of an Active Signaling Complex between Rac and Phospholipase C Gamma 2 Tom D. Bunney,1,5,* Olaniyi Opaleye,2,5,6 S. Mark Roe,2 Petra Vatter,3 Rhona W. Baxendale,1 Claudia Walliser,3 Katy L. Everett,1 Michelle B. Josephs,1 Carolin Christow,3 Fernando Rodrigues-Lima,4 Peter Gierschik,3 Laurence H. Pearl,2 and Matilda Katan1,* 1Section

of Cell and Molecular Biology of Structural Biology, Chester Beatty Laboratories The Institute of Cancer Research, London SW3 6JB, UK 3Institute of Pharmacology and Toxicology, University of Ulm Medical Center, 89070 Ulm, Germany 4Universite ´ Paris Diderot-Paris 7, Unit of Functional and Adaptive Biology (BFA), CNRS, Laboratory of Molecular and Cellular Responses to Xenobiotics, 75013 Paris, France 5These authors contributed equally to this work 6Present address: Department of Biochemistry, University of Cambridge, Cambridge CB2 1GA, UK *Correspondence: [email protected] (M.K.), [email protected] (T.D.B.) DOI 10.1016/j.molcel.2009.02.023 2Section

SUMMARY

Rho family GTPases are important cellular switches and control a number of physiological functions. Understanding the molecular basis of interaction of these GTPases with their effectors is crucial in understanding their functions in the cell. Here we present the crystal structure of the complex of Rac2 bound to the split pleckstrin homology (spPH) domain of phospholipase C-g2 (PLCg2). Based on this structure, we illustrate distinct requirements for PLCg2 activation by Rac and EGF and generate Rac effector mutants that specifically block activation of PLCg2, but not the related PLCb2 isoform. Furthermore, in addition to the complex, we report the crystal structures of free spPH and Rac2 bound to GDP and GTPgS. These structures illustrate a mechanism of conformational switches that accompany formation of signaling active complexes and highlight the role of effector binding as a common feature of Rac and Cdc42 interactions with a variety of effectors. INTRODUCTION Members of the Rho family of Ras-related small G proteins are pivotal regulators of several aspects of cell behavior, including movement, polarity, morphogenesis, and cell division (Bustelo et al., 2007; Jaffe and Hall, 2005; Ridley, 2006; Sahai and Marshall, 2002). So far, about 20 members have been identified, and these can be further subdivided according to their sequence and function. One distinct subgroup—the Rac/Cdc42 branch, which has been best studied for the role in control of actin dynamics—includes Rac1, Rac2, Rac3, Cdc42, and RhoG (Sahai and Marshall, 2002). As with other GTPases, these proteins undergo a GTP-binding/GTP-hydrolysis cycle that enables

them to act as molecular switches in cells; this cycle is regulated by guanine nucleotide exchange factors (GEFs) and GTPaseactivating proteins (GAPs) (Bos et al., 2007). The function of these GTPases is also regulated by interaction with GDP dissociation inhibitors (GDIs) that influence membrane localization and the nature of the bound guanine nucleotide. Given the involvement of Rho family GTPases in a wide variety of important cellular processes, a great deal of effort has been put not only into understanding regulation of GTP/GDPbound states, but also into identifying their cellular targets (Bishop and Hall, 2000). For the Rac/Cdc42 branch, the effectors include several protein kinases (most notably, PAK kinases and ACK), protein complexes containing Wiskott-Aldrich syndrome protein (WASP) family members, and scaffold proteins such as IQGAP. Several lipid-modifying enzymes, including phosphoinositide-specific phospholipases, PLCb2 (Illenberger et al., 1998; Illenberger et al., 2003), PLCg2 (Piechulek et al., 2005; Walliser et al., 2008), and type Ia phosphatidylinositol-4-phosphate 5-kinase (Tolias et al., 1995), are also implicated as effectors. Among the various effectors, Rac and Cdc42 have some proteins in common, whereas they bind to others in a specific way. Structural studies of Cdc42 and Rac1 have defined the molecular basis for several examples of these interactions (Bishop and Hall, 2000; Dvorsky and Ahmadian, 2004). Based on these studies, it transpired that the interacting domains in different effectors contain different folds that can make important contacts, not only with switch I and switch II regions but also with other sites on Rac/Cdc42 GTPases. The structural diversity of effector-binding surfaces makes it difficult to predict interacting domains and their binding properties without performing further structural analysis. Furthermore, there is only a limited understanding of the type of changes underlying the basis for GTP-dependent signal propagation that, as recently suggested (Phillips et al., 2008), may be different from mechanisms illustrated for the a subunits of heterotrimeric G proteins (Noel et al., 1993) and several members of Ras and Rho family studied so far (Dvorsky and Ahmadian, 2004; Herrmann, 2003).

Molecular Cell 34, 223–233, April 24, 2009 ª2009 Elsevier Inc. 223

Molecular Cell Phospholipase C-g2 and Rac

In our previous studies, we focused on structural and mechanistic aspects of regulation of one class of GTPase effectors, the PLC isoforms (reviewed in Bunney et al., 2006). Members of the Rac/Cdc42 branch—Rac1, Rac2, and Rac3 proteins—interact with and stimulate the activity of PLCb2 and PLCg2 isoforms; Cdc42 also regulates PLCb2 (Illenberger et al., 1998, 2003; Piechulek et al., 2005). For both PLC isoforms, it has been shown that interaction and activation are GTP dependent and that, in a cellular setting, the regulatory mechanism involves translocation to membrane compartments (Illenberger et al., 2003; Piechulek et al., 2005). Functional implications of PLCg2 regulation by Rac have been suggested by studies of B cells (Walmsley et al., 2003) and, more recently, it has been shown that Rac is essential for activation of PLCg2 in platelets and subsequent aggregation and thrombus formation in vivo (Pleines et al., 2008). As a major advance in our studies, we describe here the characterization of PLCg2 regulation by Rac at the molecular level and present the high-resolution crystal structures of Rac2 (bound to GDP and to GTPgS), PLCg2 spPH, and of their complex, that, together with subsequent structure-based functional analyses, provide further understanding of Rac/effector interactions. In particular, we dissect the specificity of Rac and PLCg2 regulation and suggest a mechanism for the formation of active signaling complexes with general relevance for the Rac/Cdc42 branch. RESULTS Crystal Structure of the Rac2-PLCg2spPH Complex We have previously shown that, of the two PLCg isoforms, only PLCg2 was directly activated by Rac (Piechulek et al., 2005). Based on this finding and reconstitution experiments with PLCg variants and Rac2, we have found that an unusual pleckstrin homology domain, the split PH (spPH) domain, is both necessary and sufficient to effect activation of PLCg2 by Rac (Walliser et al., 2008). We have also demonstrated that Rac2 directly binds to PLCg2 as well as to the isolated spPH from this isoform (Walliser et al., 2008). As these experiments established the spPH domain as the site of interaction with Rac, our further studies aimed to understand the structural basis of this interaction using X-ray crystallography. To identify suitable protein variants for structural analysis of the complex, we tested spPH domain proteins with different point mutations that have affinities for GTP-loaded Rac2G12V similar to those of wild-type spPH (Table 1). We were able to cocrystallize spPHY495F with Rac2G12V bound to GTPgS and determine the structure of the complex at 2.3 A˚ resolution (Figure 1 and Table 2). The interaction surface involves predominantly residues from the b5-strand and a-helix of the spPH domain, and residues present in both the switch I (amino acids 27 to 40) and switch II (amino acids 56 to 71) regions of Rac2 (Figures 1B and 1C). The total surface area buried at the interaction interface is approximately 900 A˚2. The interaction is predominantly hydrophobic, with Phe37 and Val36 from Rac2 forming a two-pronged plug fitting into a double hydrophobic cleft on the surface of the spPH domain (Figure 1D). Rac2 residues Ile33, Trp56, Tyr64, Leu67, and Leu70 intercalate with spPH residues Phe872, Val893, Leu896, Phe897, and Phe900 to

224 Molecular Cell 34, 223–233, April 24, 2009 ª2009 Elsevier Inc.

extend this hydrophobic pocket (Figure 1D). The spPH residue Lys862 makes multiple interfacial contacts with Rac2. A side chain-side chain salt bridge is formed to Rac2 residue Asp38, a side chain-main chain hydrogen bond to Phe37, and a cation-p interaction to Phe37 (Figure 1E). Other polar interactions involve spPH residue Gln901 in a side chain-side chain hydrogen bond to Rac2 residue Asn39 and spPH residue Arg904 in a cation-p interaction to Trp56. Finally, a number of p-s interactions are visible in the complex between Phe37 and Trp56 of Rac2 with Phe897 and Phe900 of spPH. As will be discussed in more detail below, in this complex, Rac2 switch regions are positioned so that the residues Thr35 and Tyr32 of switch I and Gly60 of switch II coordinate the g-phosphate of GTPgS and the Mg2+ ion. Functional Analysis of the Rac2-PLCg2spPH Domain Interface To further explore the nature of the complex interface and to identify crucial amino acid residues, a number of biochemical analyses were performed. Binding affinities between wild-type and mutant Rac2G12V (GTPgS-bound) and PLCg2spPH were determined using isothermal titration calorimetry (ITC) (Table 1; see Figure S1 available online). The functional importance of these mutated residues was furthermore determined by cotransfection of COS-7 cells and the subsequent measurement of Rac2 stimulated phosphatidylinositol 4,5-bisphosphate (PtdInsP2) hydrolysis by PLCg2 (Figure 2); the data from these experiments in intact cells are in general agreement with reconstitution of activation of PLCg2 by a number of Rac2 mutants in vitro (Figure S2). A comparison of the data in Table 1 and Figure 2 illustrate that there is a general trend for the association equilibrium constant (Ka) to be positively correlated with the effectual activation for the various Rac2 and spPH mutants. This result supports further our previous findings suggesting that Rac stimulates PLCg2 activity through direct binding and that the spPH domain is both necessary and solely sufficient for the activation of PLCg2 mediated by Rac (Walliser et al., 2008). On closer inspection of the ITC data, it becomes apparent that all residues involved in the interface (see above) are indeed important for the binding affinity. In particular, Rac2 residue Phe37 is a key element, since its mutation to alanine abolishes binding and nearly all activation (Table 1 and Figure 2A). In the spPH domain, residue Phe897 plays an equally important role (Table 1 and Figure 2C). We have also established important roles for Rac2 residues Asp38, Trp56, Tyr64, Leu67, and Leu70 (Table 1 and Figures 2A and 2B), and spPH domain residues Lys862, Val893, Phe900, and Gln901 (Table 1 and Figures 2C and 2E). To exclude the possibility that the PLCg2 mutants are otherwise deficient in their ability to hydrolyze PtdInsP2, we also assessed their ability to be activated by EGF in the COS-7 cell assay. All the mutants that were poorly activated by Rac2G12V behaved like wild-type with respect to the EGF response (Figures 2D and 2F). Furthermore, these data define the pathways that activate PLCg2 through Rac and EGF as having separate structural requirements and provide tools to test further relative importance of these different input signals in PLCg2 activation.


Table 1. Thermodynamic Quantities for the Binding of RacG12V, Cdc42G12V GTPases, and Their Mutants to the PLCg2spPH Domain and Mutants Thereof Protein in Syringe

n

KA (M

1

31000)

KD (mM)

-DH (kcal mol 1)

DS (cal mol

1

K 1)

Cell

GTP Dependence of Binding PLCg2spPHWT

Rac2(GTPgS)

0.7

54.1 ± 0.4

18.5

11480 ± 16

PLCg2spPHWT

Rac2(GDP)

—

NB

NB

NB

PLCg2spPHWT

Rac2I33A(GTPgS)

0.6

20.6 ± 0.4

48.7

8545 ± 90

PLCg2spPHWT

Rac2V36A(GTPgS)

0.7

11.6 ± 0.2

86.1

6568 ± 69

PLCg2spPHWT

Rac2F37A(GTPgS)

—

NB

NB

NB

PLCg2spPHWT

Rac2D38A(GTPgS)

1.0

4 ± 0.3

251

4012 ± 190

3.0

WT

Rac2N39A(GTPgS)

0.5

136.8 ± 12.0

7.3

6112 ± 84

3.0

PLCg2spPHWT

Rac2Y40C(GTPgS)

0.5

8.3 ± 0.3

120.5

5992 ± 164

1.2

WT

Rac2N43A(GTPgS)

0.8

61.4 ± 1.2

16.3

11090 ± 44

PLCg2spPHWT

Rac2W56A(GTPgS)

0.7

5.9 ± 0.3

168.2

2830 ± 97

7.8

PLCg2spPHWT

Rac2Y64A(GTPgS)

0.9

8.2 ± 0.3

122.3

4630 ± 69

2.4

WT

Rac2R66A(GTPgS)

0.7

59.2 ± 1.4

16.9

13550 ± 70

PLCg2spPHWT

Rac2L67A(GTPgS)

0.9

6.6 ± 0.3

151.7

2404 ± 55

9.41

PLCg2spPHWT

Rac2L67E(GTPgS)

—

NB

NB

NB

NB

WT

Rac2L67F(GTPgS)

0.7

49.1 ± 3.3

20.4

10460 ± 164

PLCg2spPHWT

Rac2L70A(GTPgS)

1.0

6.3 ± 0.1

158.8

5801 ± 74

2.07

PLCg2spPHWT

Rac2S71A(GTPgS)

0.7

72.0 ± 0.8

13.9

15810 ± 34

30.8

PLCg2spPHWT

Rac2A27K(GTPgS)

0.8

57.1 ± 0.8

17.5

11640 ± 29

17.3

WT

Rac2G30S(GTPgS)

0.8

57.3 ± 0.4

17.4

11770 ± 16

17.7

PLCg2spPHWT

Rac2I33V(GTPgS)

0.8

46.3 ± 0.5

21.6

12220 ± 29

19.6

PLCg2spPHWT

Rac2W56F(GTPgS)

0.8

6.6 ± 0.1

152.6

7564 ± 81

7.9

Rac2QM(GTPgS)

0.8

6.7 ± 0.2

148.8

7040 ± 126

6.1

PLCg2spPHWT

Cdc42(GTPgS)

0.7

3.7 ± 0.3

270.3

3984 ± 256

PLCg2spPHWT

Cdc42K27A(GTPgS)

0.8

3.5 ± 0.2

288.4

4068 ± 208

2.6

PLCg2spPHWT

Cdc42S30G(GTPgS)

0.8

7.3 ± 0.4

136.1

2508 ± 75

9.3

PLCg2spPHWT

Cdc42V33I(GTPgS)

0.7

7.2 ± 0.3

138.0

3245 ± 108

6.8

PLCg2spPHWT

Cdc42F56W(GTPgS)

0.7

25.3 ± 0.6

39.6

9887 ± 82

13.0

PLCg2spPHWT

Cdc42QM(GTPgS)

0.6

25.1 ± 0.4

39.9

9987 ± 62

13.3

Rac2(GTPgS)

PLCg2spPHWT

1.2

54.2 ± 2.9

18.5

10130 ± 99

12.3

Rac2(GTPgS)

PLCg2spPHY495F

1.3

45.7 ± 3.4

21.9

10300 ± 141

Rac2(GTPgS)

PLCg2spPHK862E

—

NB

NB

NB

NB

Rac2(GTPgS)

PLCg2spPHK862I

1.5

6.4 ± 0.7

155.9

4047 ± 198

3.8

Rac2(GTPgS)

PLCg2spPHV893Q

1.3

1.5 ± 0.1

>300

6369 ± 291

Rac2(GTPgS)

PLCg2spPHF897Q

—

NB

NB

NB

Rac2(GTPgS)

PLCg2spPHF900V

1.4

12.3 ± 0.6

81.6

9639 ± 152

Rac2(GTPgS)

PLCg2spPHY495C

1.1

17.8 ± 2.8

56.2

13860 ± 1470

Rac2(GTPgS)

PLCg2spPHW899A

—

NB

NB

NB

16.8 NB

Rac Switch Region Mutations

PLCg2spPH PLCg2spPH

PLCg2spPH

PLCg2spPH

8.9 3.4 NB

15.3

23.6

13.6

Rac and Cdc42 Swap Mutations PLCg2spPH

PLCg2spPH

WT

3.0

spPH Domain Mutations

The PH Domain as a Versatile Interaction Surface in Protein-Protein Interactions The PH domain is one of the most thermodynamically stable and structurally adaptable folds that is commonly involved in interactions with cellular membranes; there are a number of examples where PH domains make direct interactions with membrane

13.2

6.8 NB 13.6 27 NB

polyphosphoinositides, but there are also examples of protein/ protein interactions with membrane-tethered proteins (reviewed in Lemmon, 2004). The structural requirement for the recognition of polyphosphoinositides by PH domains has been well characterized and mainly involves amino acid residues in the b1- and b2-strands and the b3/b4-loop. However, the structural basis



A

B

C

D

E

Figure 1. Crystal Structure of the Rac2G12V(GTPgS)/PLCg2spPH Complex Domain organization of PLCg2 (A) shows domains common to all families (nPH, EF-hands, catalytic, and C2) and domains specific to PLCg family (nSH2, cSH2, SH3, and spPH); spPH is shown in orange. Overview of the Rac2G12V(2-177) GTPgS/PLCg2spPH complex structure (B). Surface topology (shown as translucent surfaces) and ribbon representation of the complex show the spPH domain in orange and the Rac2 molecule in blue. The Rac2 switch I region is shown in green, and the switch II region is shown in magenta. The GTPgS molecule bound to Rac2 is in ball-and-stick representation. Closeup of the interaction interface is viewed from the side with important amino acid residues of both Rac2 and spPH represented as ball and sticks (C). Also shown is a close-up from above with important Rac2 amino acid residues (ball-and-stick representation) and the spPH surface (D). The hydrophobic cleft into which Rac2 residues Val36 and Phe37 protrude can be clearly seen. The summary of interactions at the interface is illustrated schematically highlighting the predominant types of bonding between the two species in the complex (E).

for protein recognition is highly variable and can involve virtually any interface of the PH fold, as illustrated by the structures of GTPases bound to PH domain folds, PLCb2/Rac1 (Jezyk et al., 226 Molecular Cell 34, 223–233, April 24, 2009 ª2009 Elsevier Inc.

2006) (pdb:2fju), Exo84/RalA (Jin et al., 2005) (pdb:1zc3), and RanBP1/Ran (Vetter et al., 1999) (pdb:1rrp) (Figure S3). Interestingly, the PLCg2spPH contact region is completely different from the PH domain region in PLCb2 interacting with the closely related Rac1 (Figures 3A and 3B). As highlighted in the secondary structure alignment of the two PH domains (Figure 3C), although the PLCg2 spPH contacts Rac predominantly by residues from the b5-strand and a-helix, the main contact region in the PLCb2 PH domain is located in b strand 1 and loop regions that align with this strand. Despite these differences, the Rac switch region residues that form contacts with the two PH domains are broadly similar, and, furthermore, the type of main interactions is conserved. Mutational analyses (Figure 2; Jezyk et al., 2006) show that Phe37, Trp56, and Leu70 are equally crucial for both complexes, whereas alterations of Asp38, Tyr64, and Leu67 appear to have a less deleterious effect on activation of PLCb2 than of PLCg2. In particular, a Rac effector mutant possessing the Phe37 mutation (as a result of its central role in both complexes) would be ineffective at activating either PLC isoform in vivo. For PLCb2, this has already been illustrated to be a useful tool in delineating the role of Rac3 in activating different downstream effector pathways (Keller et al., 2005). We thought that the development of a Rac effector mutant that could differentiate between PLCg2 and PLCb2 would be a similarly useful tool in further studies in cells, where the roles of these PLC isoforms downstream of Rac could be established. From a comparison of the two Rac-PLC complexes, it was apparent that a mutation of Rac Leu67 to a more bulky and/or negatively charged residue could better be accommodated by PLCb2 than by PLCg2 (Figures 3A and 3B). We tested this hypothesis by mutating the Leu67 to Ala, Phe, or Glu and testing the ability of these Rac2G12V mutants to activate PLCg2 or PLCb2 in vitro (Figure 3D) and in COS-7 cells (Figure 3E). We also measured the affinity of these Rac mutants for the PLCg2spPH domain using ITC (Table 1). As predicted from the structural comparison, PLCb2 activation by Rac was not significantly affected by either the Phe or Glu mutation; the activation by the Ala mutation was somewhat reduced. In contrast, all three mutations caused a significant reduction in the PLCg2 activation by Rac, with the Glu mutation preventing any activation and abolishing binding. The mutant RacL67E would therefore be a useful tool in differentiating between PLCg2 or PLCb2 activation downstream of Rac. Rac Specificity of the spPH Domain Interaction The activation of PLCg2 by Rac GTPases is highly specific, and the closely related GTPases Rho and Cdc42 do not activate PLCg2 in the COS-7 cell assay or in an in vitro reconstitution assay (Piechulek et al., 2005). This is in some contrast to the less specific activation of PLCb2, since Cdc42 can also activate this PLC isoform, albeit to a lesser extent than Rac (Illenberger et al., 1998; Piechulek et al., 2005). We wanted to investigate the molecular basis for the Rac specificity in activating PLCg2. Since we have shown that Rac engages the spPH domain of PLCg2 exclusively through its switch regions, we investigated the amino acid residue differences in these regions on Rac and Cdc42. We identified three amino acid differences in the switch I region—Ala27Lys, Gly30Ser, and Ile33Val—and a further single


Table 2. Data Collection and Refinement Statistics: Molecular Replacement Rac2G12V(GTPgS)/PLCg2spPHY495F

PLCg2spPHY495F

Rac2G12V(GTPgS)

Rac2G12V(GDP)

Space group

C2

P212121

P21

C2221

Cell dimensions a, b, c (A˚)

131.57, 84.46, 74.20

92.25, 105.91, 194.02

42.19, 94.33, 107.22

42.21, 98.11, 108.06

90.0, 112.21, 90.0

90.0, 90.0, 90.0

90.0, 89.96. 90.0

90.0, 90.0, 90.0

Resolution (A˚)

100–2.30 (2.42–2.30)

100–3.3 (3.48–3.3)

60.0–2.0 (2.11–2.0)

60.0–1.95 (2.06–1.95)

Rsym or Rmerge

0.063 (0.499)

0.091 (0.560)

0.088 (0.495)

0.082 (0.524)

I/sI

12.8 (1.9)

11.3 (2.8)

7.2 (1.9)

12.9 (2.3)

Completeness (%)

86.9 (52.7)

98.4 (95.2)

94.3 (92.4)

98.7 (96.1)

Redundancy

3.0 (2.7)

3.7 (3.6)

2.0 (2.0)

4.8 (4.5)

Data Collection

a, b, g ( )

Refinement Resolution (A˚)

2.3

3.3

2.0

1.95

No. reflections

54867

28821

92910

30557

Rwork/Rfree

0.234/0.294

0.303/0.352

0.280/0.345

0.210/0.254

Protein

4277

9250

5644

1345

Ligand/ion

66

0

64

29

Water

38

0

547

138

Protein

64.9

94.9

31.1

40.9

Ligand/ion

61.7

—

25.0

33.7

Water

48.4

—

34.4

44.6

Bond lengths (A˚)

0.010

0.010

0.005

0.023

Bond angles ( )

1.406

1.226

1.041

2.048

No. atoms

B factors

Rmsds

One crystal was used for each data collection. Values in parentheses are for highest-resolution shell.

difference in the switch II region, Trp56Phe (Figures 4A and 4B). We created several mutants of Rac2G12V and Cdc42G12V where these ‘‘switch swap’’ mutations were introduced singly or all four simultaneously (thereby creating a Cdc42 switch region within Rac and vice versa) and tested these various mutants for their ability to activate PLCg2 in the COS-7 cell assay and their binding to PLCg2spPH determined by ITC (Figure 4C and Table 1). All the Rac2 mutants bound to spPH with similar affinity (KD around 18 mM) except the Trp56Phe and the quadruple mutants (QM) that have significantly impaired binding (KD around 150 mM). The activation correlated well with these binding data, with the weaker binders also being poorer activators. As expected, all the Cdc42 mutants bound to spPH with low affinity (KD around 200 mM) except the Phe56Trp and the quadruple mutants that bound significantly tighter (KD around 40 mM). From the crystal structure of the Rac2-spPH complex, it is apparent that a Phe at position 56 would not support the cation-p or the p-s interactions that are seen with Trp56, since the aromatic ring of Phe does not occupy the same position or amount of space as that of Trp (Figure 4B). This reduced space filling would also affect the hydrophobic interactions at the interface. Taken together, these factors likely account for the reduced affinity of spPH for RacW56F. Although the binding experiments identify Phe56 in Rac as a critical determinant for preferential binding over Cdc42,

a correlation between binding affinities and the ability to activate PLCg2 by different Cdc42 variants appears to be more complex. Notably, there was no activation of PLCg2 by any of the Cdc42G12V variants incorporating residues from Rac switch regions (Figure 4C). In addition to the experiments in COS cells, the experiments in vitro using equal amounts of purified proteins also demonstrate that replacement of critical Trp56 of Rac2 by Phe blocks stimulation of PLCg2 by the protein and that reverse replacement does not impart on Cdc42 the ability to cause this stimulation (Figure 4D, right panel). Considering that the binding affinity of the Cdc42 F56W or the quadruple mutant (Table 1) is within the range where we would expect it to be able to activate PLCg2 (see RacI33A in Table 1), this raises further questions about the relevant differences between Rac and Cdc42. Based on our findings, it is likely that the Cdc42 F56W and the quadruple mutant bind PLCg2 nonproductively and that there are domains outside the switch regions that are relevant for the formation of a correct binding interface and/or activation. Further analysis is however required to clarify a structural perspective of these findings. We also compared different codon 56 variants of Rac and Cdc42 in their ability to activate the two PLC isoforms. Interestingly, we found that Trp56, a critical determinant for Rac binding selectivity by PLCg2 spPH (Table 1), is not likely to affect activation of PLCb2 or have a prominent role in discriminating between Molecular Cell 34, 223–233, April 24, 2009 ª2009 Elsevier Inc. 227


(G12V) variant incorporating this mutation described in the previous study.

Figure 2. Functional Analysis of the Rac2/spPH Interface Using PLCg2 Activation by Rac2 and Their Mutants in COS-7 cells A number of Rac2 and PLCg2 variants were introduced into COS cells and analyzed for their impact on PLC activity. The mutations were introduced into Rac2 switch I and switch II regions (A and B) and spPH domain of PLCg2 (C–F). The indicated mutations in the spPH domain were analyzed either after cotransfection with Rac2 G12V (C and E) or after stimulation of cell with EGF (D and F). Stimulation of the control, mock transfected COS-7 cells with EGF resulted only in a slight (20%) increase of basal activity. For these experiments, COS-7 cells were transfected with 1 mg of pMT2hPLCg2 vector or mutants thereof and 0.5 mg of pcDNADEST40Rac2G12V vector or mutants thereof. PLC activity was measured as production of inositol phosphates. SD is represented by error bars. Inserts show the specific detection of protein expression of the transfected components in each individual experiment. The order of loading for the western blot corresponds to the order of components on the x axis of the bar chart.

the two GTPases by PLCb2 PH (Figure 4D; Figure S4, upper panels). This contradicts previous findings that G12V variant of Rac1W56A had impaired ability to activate PLCb2 (Jezyk et al., 2006), and our further experiments (Figure S4, lower panels) suggest a likely reason for this discrepancy. We have observed suppression of basal PLCb2 activity by the signaling-inactive form of Rac1W56A (i.e., a nonspecific inhibitory effect) that could account for the decrease in PLCb2 activation by the constitutive 228 Molecular Cell 34, 223–233, April 24, 2009 ª2009 Elsevier Inc.

Insights into the Mechanism of Conformational Changes that Accompany Formation of the Active Signaling Complex Because we obtained structures of Rac2G12V(GDP), Rac2G12V(GTPgS), and spPHY495F as the free proteins, as well as in the complex, it was possible to identify conformational changes that accompany complex formation (Table 2; Figures 5 and 6). The crystal structure of the free form of the PLCg2 spPH contained 12 subunits in the asymmetric unit. We observed no significant difference in the conformation of the Ca backbone chain between the free and complexed forms (Figure 5A). Interestingly, the different subunits within the asymmetric unit of the free PLCg2spPH show some heterogeneity, particularly with respect to the side chain orientation of Phe897, the key residue involved in Rac binding (Figure 5, inset). In half the subunits, this orientation was the same as that in the complex with Rac2, which permits insertion of Rac2 residue Phe37. However, in the other molecules, the side chain of spPH residue Phe897 is rotated by 90 , blocking the hydrophobic socket. This suggests that a ‘‘soft switch’’ could exist on the surface of the spPH that opens or closes the hydrophobic cleft that ultimately accommodates the Rac switch I residue, Phe37. Comparison of the Rac2 structures shows some differences in the orientation of the Ca backbone of the switch II region and highlights an obvious change in the switch I effector loop (Figure 5B). The GDP-bound and GTPgS-bound free forms of Rac superimpose well, but upon binding the spPH domain, the Rac switch regions are clearly repositioned. Strikingly, the positioning of the side chain of the key effector-interacting residue Phe37 is shifted upon binding, whereas the nature of the bound nucleotide in the free form of Rac does not influence the position of this residue (Figure 5C). Effector binding also influences the Rac amino acid residues that coordinate the g-phosphate on GTPgS. In stark contrast to the situation with crystal structures of several members of Ras family bound to GTP-analogs (Milburn et al., 1990; Vetter and Wittinghofer, 2001), where the g-phosphate is coordinated by Thr35 and Gly60 when the protein is free or bound to effectors, in the case of Rac2, this coordination is seen only in the complex with spPH. The absence of this interaction in free Rac2(GTPgS) would imply that the structures of GDP- and GTPgS -bound Rac2 should be similar. Indeed, the comparison with Rac2GDP shows that both Thr35 and Gly60 are similarly positioned (Figure 5B). Although coordination of the g-phosphate by Thr35 is facilitated by changes in conformation of switch I in the Rac2/spPH complex, the direct interaction with Gly60 from switch II results from a movement of the GTPgS g-phosphate that is shifted to the vicinity of Gly60. Together, these data suggest that the signaling-active conformation of Rac2(GTP), resembling that of noncomplexed Ras (GTP), is greatly stabilized through interaction with its effector, PLCg2 (Figure 6). An important feature of its mechanism is likely to be the formation of a hydrophobic pocket that buries Phe37 and Val36 of the switch I region of Rac. This, in turn, reverses


Figure 3. Comparison of the spPH Domain of PLCg2 and the N-terminal PH Domain of PLCb2 in Their Interaction with Rac

A

B

C

D

E

orientation of the key residues within the switch I region so that the coordination of the g-phosphate is achieved. As depicted in Figure 6, a similar mechanism is likely to underlie formation of a PLCb2/Rac complex and complexes formed by Rac/ Cdc42 proteins and several other effectors. As discussed further below, these structural insights provide a basis to consider more generally dynamic conformational states of small GTPases and the implications for their interaction with and specific recognition of effector proteins. DISCUSSION The ability of Rho family GTPases to control a wide range of signaling pathways is attributed to association with their cellular targets, the effector proteins (Bishop and Hall, 2000; Dvorsky and Ahmadian, 2004). The PLC enzymes are one class of effectors that mediate signal transduction from a variety of cell surface receptors; other signaling molecules that directly interact with different PLC isoforms include tyrosine kinases

Overview of the Rac2G12V(GTPgS)/PLCg2spPH (A) and Rac1(GTPgS)/PLCb2N-PH (from pdb: 2fju) (B) complexes showing ribbon representations of the complexes (left panels) and surface representations of the PH domains (right panels). In the right panels, Rac switch regions interacting with the surfaces of the PLCg2spPH (orange) and N-terminal PH of PLCb2 (brown) are shown as an Ca chain, with Rac residue 67 represented as ball-and-stick. For both complexes, Rac switch I is shown in green and switch II in magenta. Structure-based sequence alignment of the PLCg2spPH with the PLCb2N-PH (C) shows the residues that interact with Rac in each case (underlined); identical amino acids are in red. Secondary structure elements are indicated above, in yellow for b strands and magenta for a helices. Different variants of Rac2, designed from structural comparisons of the complexes, were tested for the ability to activate PLC isoforms in vitro and in COS-7 cells. Reconstitution system in vitro was used to evaluate effect of L67A mutation (D). Purified wild-type Rac2 (wt) and Rac2L67A (L67A) were geranylgeranylated in vitro with purified type I gerany-lgeranyltransferase and geranyl-geranylpyrophosphate and then reconstituted in vitro with purified PLCb2D or PLCg2 in the presence of 100 mM GDP or GTPgS as indicated. A larger panel of Rac2 variants was analyzed in COS-7 cells (E). These experiments were performed by co-transfection of pMT2hPLCg2 or pMT2hPLCb2 with pcDNADEST40Rac2G12V and various L67 mutations and PLC activity measured as production of inositol phosphates. SD is represented by error bars.

and subunits of heterotrimeric G proteins, as well as members of the Ras-related superfamily of small GTPases (Katan, 2005; Rebecchi and Pentyala, 2000; Rhee, 2001). The structure and subsequent structure-based functional analysis presented in this work demonstrates the molecular basis for regulation of PLCg2 by Rac proteins and signaling interconnectivity with other regulators of this PLC isoform. Together with previously reported structural studies of regulation of PLCb2 by Rac (Jezyk et al., 2006), this work also expands our insights into the diversity of signaling interactions that converge on PLC isoforms and the diversity of Rac/effector interactions. Although, in common with PLCb2, the site of interaction with Rac in PLCg2 is the PH domain fold, the contact region involved is completely different from the PH domain surface in PLCb2 (Figures 1 and 2). Consistent with the versatility of interactions between PH domain folds and their protein ligands, the PLCg2 spPH/Rac interface is also unique (Lemmon, 2004; Figure S3). One distinct feature of this interface is a relatively small area (about 900 A˚2) of contact, compared with areas typically occupied by GTP-ase/effector interactions (about 1400 A˚2) (Dvorsky and Ahmadian, 2004; Jezyk et al., 2006; Kiel et al., 2008). Nevertheless, the experimentally determined binding affinities between Rac and the two PLC isoforms are similar and within the range of reversible regulatory interactions (Walliser et al., 2008). Both PH domains interact only with switch I and switch II of Rac, and, in this respect, this mode of binding is different



Figure 4. Molecular Basis for the Specificity of PLCg2 Binding to Rac over Cdc42 Sequence alignment of Rac2 and Cdc42 highlights the amino acid differences (red) in their switch regions (A). Illustration of the different space filling properties of a tryptophan at position 56 in switch II of Rac2 and a phenylalanine in Cdc42 is shown in (B). (C) shows COS-7 cell PLC activation experiments with cotransfection of pMT2hPLCg2 with pcDNADEST40Rac2G12V or pcDNADEST40Cdc42G12V incorporating indicated mutations in switch regions. PLC activity is expressed as formation of inositol phosphates. Functional role of Trp56 of Rac2 and Phe56 of Cdc42 for stimulation of PLCb2D and PLCg2 was further tested upon reconstitution of purified proteins (D). Purified wild-type Rac2 (wt), Rac2W56F (W56F), wild-type Cdc42 (wt), and Cdc42F56W (F56W) were geranyl-geranylated in vitro with purified type I geranyl-geranyltransferase and geranyl-geranylpyrophosphate and then reconstituted in vitro with purified PLCb2D (left panel) or PLCg2 (right panel) in the presence of 100 mM GDP or GTPgS as indicated. SD is represented by error bars.

from the interactions with several other effectors (e.g., WASP, PAK, and ACK) that contain a common binding structure incorporating the CRIB (Cdc42/Rac-interactive binding) motif (Bishop and Hall, 2000; Dvorsky and Ahmadian, 2004). Binding domains of CRIB-containing effectors make extensive contacts with Cdc42; the domain binds with its b-hairpin and C-terminal a-helix to a1, switch I, and switch II regions and wraps around the a5 and b2 regions of GTPase with its extended N terminus that encompasses the CRIB motif. Further studies based on the PLCg2 spPH/Rac structure revealed that mutations in this binding interface do not affect activation of PLCg2 by EGF, mediated through engagement of SH2 domains and phosphorylation of specific tyrosine residues (Figure 2). Similarly, activation of PLCb2 by Rac can be abolished without affecting regulation through Gbg interactions (Jezyk et al., 2006) and is also likely to be different from the regulation of PLCb isozymes by Gaq (Rhee, 2001). Therefore, these structural studies resulted in generation of PLC variants that will greatly facilitate dissection of their physiological roles in different cellular settings and assess relative contribution of Rac in their activation. Furthermore, a comparison of PLCg2 spPH/Rac2 and PLCb2 PH/Rac1 structures led to the generation of Rac variants that can selectively activate the PLCg2 isoform (Figure 3E), providing important tools to assess signaling components that link Rac activation and cellular responses mediated by PLCgenerated second messengers. In addition to insights into the regulation of PLC isoforms and into the diversity of Rac/effector interactions, the structural studies described here (Figures 5 and 6) also provide important


insights into mechanisms of conformational changes that accompany formation of active signaling complexes between small GTPases and their effectors. The switch mechanism proposed for GTP-binding proteins suggests that the trigger for the conformational change following GDPGTP exchange is likely to be universal (reviewed in Vetter and Wittinghofer, 2001). A crucial component of this mechanism is the formation of a hydrogen bond between the g-phosphate of GTP and the invariant switch I Thr residue (Thr35 in Ras and Rac), resulting in the GTP-bound conformation proposed to specifically interact with effector proteins. In-depth studies of this general scheme for Ras GTPases revealed a more complex picture with different conformational states of the nucleotide-bound GTPase and effector complexes in dynamic equilibrium. In particular, NMR studies of GTP-bound Ras in solution revealed two main conformational states (Spoerner et al., 2001): one corresponding to ‘‘active’’ RasGTP, with Thr35 interacting with the g-phosphate of GTP (state 2), and the other conformation where this interaction is lacking (state 1) with much lower affinity for effectors. Several crystal structures of Ras family members (H-Ras, Rap1A, and Rap2A) correspond to state 2 conformation (Cherfils et al., 1997; Nassar et al., 1995), whereas one structure (GTP-analog-bound M-Ras) could correspond to the state 1 conformation, suggesting that it is more similar to the GDP-bound state (Ye et al., 2005). It is expected that similar conformational interconversions occur for Rho family members. In the structure of Rac2(GTPgS) shown here, the interactions between Thr35 and g-phosphate are lacking, and the conformation of this protein bound to the nonhydrolyzable-GTP-analog is virtually superimposable with the structure of Rac2(GDP), suggesting that Rac2(GTPgS) corresponds to state 1 described for Ras (Figure 5). Recent structural data for Cdc42 bound to a GTP-analog—Cdc42(GMP.PCP) (pdb: 2qrz)—have shown that the same applies to Cdc42 (Phillips et al., 2008). Furthermore, two other Rac proteins, Rac1 (pdb: 1i4t and 1i4d) (Tarricone et al., 2001) and Rac3 (pdb:


Figure 5. Different Conformations of Free PLCg2spPH and Free Rac2 from Their Conformation in the Rac2/PLCg2spPH Complex Ribbon diagram of superimposed spPH domains shows two forms of free spPH, free form I (blue) and free form II (red), and spPH domain from the Rac2(GTPgS)/PLCg2spPH complex (yellow) (A). Close-up of the region within the spPH domain a-helix shows position of Phe897 (represented as ball-and-stick) in the structures of free form I, free form II, and from the complex (inset). Conformational variability of flexible Phe897 among the 12 copies of the free sPH domain is influenced by crystal packing. Ribbon diagram of superimposed Rac2(GDP) (purple), Rac2(GTPgS) free form (yellow), and Rac2(GTPgS) from the Rac2(GTPgS)/PLCg2spPH complex (blue) (B). The two switch regions have been boxed to allow the differences to be easily visualized. Close-up examination of the nucleotide binding site of the above three structures highlights important residues of switches I and II (C). Indicated specific residues (represented as ball-and-stick) show different positions of Phe37, Val36, and Thr35 and interactions with GDP or GTPgS by Thr17, Thr35, Tyr32, and Gly60.

2g0n and 2ic5) (unpublished) are likely to be similar in this respect. Although it is possible that for Rac/Cdc42 GTPases only the state 1 conformation is amenable for crystallization or stabilized in a particular crystal form, further experimental evidence provided for Cdc42 by NMR analysis in solution suggests that this state represents a major conformation (Phillips et al., 2008). Furthermore, using the intrinsic fluorescence of a mutated switch I residue (Trp replacing Tyr32), it has also been shown that the GDP-GTP exchange in Cdc42 does not significantly alter the position of switch I (Phillips et al., 2008). Very little change was observed following GDP-GTP exchange on Cdc42, contrasting greatly with fluorescence measurements obtained using H-Ras. However, the addition of effector protein to Cdc42(GTP) resulted in a marked change in fluorescent emission from the switch I tryptophan. Together, these data suggest that for Rac/Cdc42 proteins, the uncomplexed GTP-bound conformation mainly corresponds to state 1 and that interconversion to state 2, found in all complexes with their effectors, is induced by interaction with effector proteins. Interestingly, analyses of the differences between uncomplexed components and the spPH/Rac2 complex (Figure 5),

together with the comparison with several Cdc42 and Rac1 complexes with their effectors, suggest a common molecular basis for effector-facilitated conformational switches, despite the overall diversity of binding surfaces (Figure 6). This analysis also suggests how effectors proteins could distinguish between structurally similar GTP-bound state 1 and GDP-bound Cdc42/Rac proteins. As described here for Rac2 interacting with PLCg2 (Figure 1), interactions with other effectors also center on residues Phe37 and Val36, that become buried in a hydrophobic pocket, formed in part by the residues from the effector (pdb: 2ov2, 2qme, 1nf3, 1e0a). The residue in PLCg2 spPH that forms the core of the hydrophobic pocket (Phe897) also displays conformational flexibility that could be stabilized by the complex formation (Figure 5A). As suggested for several Cdc42/effector interactions (Phillips et al., 2008), the potential role of the hydrophobic pocket is to enable Phe37 to act as a lever to flip and stabilize the switch I loop, so that Thr35 and Tyr32 coordinate the g-phosphate and Mg2+ ion, resulting in a conversion from state 1 to state 2 conformation. These interactions, dependent on the g-phosphate (i.e., GTP), could result in complexes with considerably higher affinity and could provide the basis for selectivity between GDP- and GTP-bound Cdc42/Rac proteins. This analysis included effectors with quite diverse folds and interaction surfaces, such as PLCg2 spPH (Figure 1), PLCb2 PH (Jezyk et al., 2006), PAKbinding domain (with CRIB motif) (Morreale et al., 2000), and Par6-binding domain (Garrard et al., 2003). The only feature shared by the different domains and folds interacting with



and purified as described (Moomaw et al., 1995). Nonprenylated wild-type and mutant human Rac2 and Cdc42 (isoform 1; acc. no. NP_001034891) as well as human PLCb2D (Illenberger et al., 1998, 2003), a deletion mutant of PLCb2 lacking a C-terminal region necessary for stimulation by Gaq subunits (F819–1166) and carrying a serine to alanine replacement in position 2, were purified from the soluble fraction of baculovirus-infected insect cells (Illenberger et al., 1998). The sources of all other proteins used for in vitro reconstitution have been described (Piechulek et al., 2005; Walliser et al., 2008). Crystallization, data collection, structure solution, and refinement are described in detail in the Supplemental Experimental Procedures. Data collection and refinement statistics (molecular replacement) are presented in Table 2. Isothermal Titration Calorimetry Heats of interaction were measured on a MSC system (Microcal) with a cell volume of 1.458 ml. GTPases were loaded with the nonhydrolyzable GTP analog, GTPgS, following previously published methods (Herrmann et al., 1996). Proteins were dialyzed for 16 hr in ITC buffer (25 mM Tris.Cl, 50 mM NaCl, 5 mM MgCl2, and 1 mM TCEP [pH 8.0]). Proteins were loaded in the sample cell at 300 mM and titrated with the binding partner in the syringe (2 mM). The titrations were performed while samples were being stirred at 260 rpm. at 25 C. A total of 29 injections was carried out, with 10 ml injected each time and a 4-min interval between each injection to allow the baseline to stabilize. The data were fitted with a single site model to calculate the number of binding sites (n), the binding constant (Ka), the change in enthalpy (DHo), and change in entropy (DS) using Origin software (Microcal).

Figure 6. Effector-Facilitated Conformational Changes in Rac/ Cdc42 Proteins Comparison of switch I regions of uncomplexed, GTP analog-bound Ras (pdb: 121P) (A) with an overlay of Rac2 and Cdc42 uncomplexed, GTP analogbound structures (pdb: 2w2v and 2QRZ) ([B], left) and Rac2 and Cdc42 GTP analog-bound structures (overlay) from complexes with PLCg2 (pdb: 2w2x) and Par6 (pdb: 1NF3), respectively ([B], right). Conformations of switch I region and orientation of indicated key residues in free Rac/CDC42-(GTP) proteins and in complexes with their effectors are clearly different. Binding of several effectors to Rac/Cdc42 proteins involves interaction with switch I residues Val36 and Phe37 that stabilizes signaling-active (state 2) conformation where other switch I residues, Thr35 and Tyr32, can coordinate g-phosphate of GTP.

Cdc42/Rac proteins is that they provide hydrophobic binding surfaces that can engage Phe37 and/or Val36 in a specific orientation that facilitates coordination of the g-phosphate. EXPERIMENTAL PROCEDURES Constructs, Antibodies, Protein Purification, and Structure Determination Constructs used in this study have been described previously in Piechulek et al. (2005) and Walliser et al. (2008). Various mutants are described in the text and were generated using the Quikchange PCR mutagenesis protocol (Stratagene). All mutated constructs were sequenced to confirm the correct mutant had been created. All proteins that were expressed in Escherichia coli and in baculovirus-infected insect cells for crystallization, ITC, and functional reconstitution were prepared as described in (Walliser et al., 2008). All the Rac2 and Cdc42 mutants used in the COS-7-cell-based PLC activation assay were cloned as untagged ORFs into pDONR207 (Invitrogen) using Gateway technology or as untagged or N-terminally HA-epitope-tagged ORFs into pcDNA3.1. For expression in COS-7 cells, the former inserts were transferred to pcDNA-DEST40 using the LR reaction (Invitrogen). The following antibodies, all from Santa Cruz, were used in this study: a-Rac2 (sc-96), a-PLCg2 (sc-407), a-PLCb2 (sc-206), and a-Cdc42 (sc-6083). Rat type I geranylgeranyltransferase was produced in baculovirus-infected insect cells


Measurements of PLC Activity Determination of PLC activity in transfected COS-7 cells and visualization of expressed proteins were carried out as previously described (Sorli et al., 2005; Walliser et al., 2008). The concentrations of functionally competent wild-type and mutant Rac2 and Cdc42 GTPases were measured in detergent extracts and purified preparations by quantitative [35S]GTPgS binding, as detailed before (Piechulek et al., 2005). In vitro reconstitution experiments were done as described (Piechulek et al., 2005). For the reconstitution of purified Rac2 and Cdc42 with purified PLCs, the soluble GTPases were first incubated for 15 min at 30 C in a volume of 20 ml containing 8 mg type I geranylgeranyl-transferase, 60 mM geranylgeranylpyrophosphate, 65 mM Tris/maleate (pH 7.3), 0.7 mM EDTA, 70 mM NaCl, 0.7 mM MgCl2, 0.7 mM DTT, 70 mM PMSF, 0.7 mM GDP, and 10 mM ZnCl2. At the end of the isoprenylation reaction, the samples were reconstituted with purified PLCb2D or PLCg essentially as described in Piechulek et al. (2005). The incubation was performed for 30 min at 30 C with phospholipid vesicles containing PtdInsP2 in the presence of 100 mM GDP or GTPgS, in a final volume of 60 ml. The incubation was in the presence of 150 nM free Ca2+ (for PLCb2D) or 30 nM free Ca2+ and 1 mM sodium deoxycholate (for PLCg2). For the comparison of Rac2 (wt) and Rac2L67A (L67A), the amounts were 75 pmol/reconstitution assay; for comparisons of Rac2 and Cdc42 variants, 30 pmol/reconstitution assay of wild-type and mutant Rac2 and Cdc42 protein was used. The amounts were determined by quantitative [35S]GTPgS binding. ACCESSION NUMBERS The PDB deposition codes for these crystal structure are as follows: PLCg2spPHY495F/Rac2G12V complex is 2w2x, spPHY495F is 2w2w, Rac2G12V(GTPgS-bound) is 2w2v, and Rac2G12V(GDP-bound) is 2w2t. SUPPLEMENTAL DATA The Supplemental Data include Supplemental Experimental Procedures and four figures and can be found with this article online at http://www.cell.com/ molecular-cell/supplemental/S1097-2765(09)00139-7. ACKNOWLEDGMENTS We thank A. Smith, C. Weiss, and N. Zanker for excellent technical assistance. Work in M.K.’s and L.H.P.’s laboratories is supported by Cancer Research UK.


Work in P.G.’s laboratory is supported by the Deutsche Forschungsgemeinschaft (DFG; SFB 497), and work in F.R.L.’s laboratory is supported by the Association pour la Recherche sur le Cancer (ARC) and Association Franc¸aise contre les Myopathies (AFM). Received: December 18, 2008 Revised: January 28, 2009 Accepted: February 20, 2009 Published: April 23, 2009 REFERENCES Bishop, A.L., and Hall, A. (2000). Rho GTPases and their effector proteins. Biochem. J. 348, 241–255. Bos, J.L., Rehmann, H., and Wittinghofer, A. (2007). GEFs and GAPs: critical elements in the control of small G proteins. Cell 129, 865–877. Bunney, T.D., Harris, R., Gandarillas, N.L., Josephs, M.B., Roe, S.M., Sorli, S.C., Paterson, H.F., Rodrigues-Lima, F., Esposito, D., Ponting, C.P., et al. (2006). Structural and mechanistic insights into ras association domains of phospholipase C epsilon. Mol. Cell 21, 495–507.

Milburn, M.V., Tong, L., deVos, A.M., Brunger, A., Yamaizumi, Z., Nishimura, S., and Kim, S.H. (1990). Molecular switch for signal transduction: structural differences between active and inactive forms of protooncogenic ras proteins. Science 247, 939–945. Moomaw, J.F., Zhang, F.L., and Casey, P.J. (1995). Isolation of protein prenyltransferases from bovine brain and baculovirus expression system. Methods Enzymol. 250, 12–21. Morreale, A., Venkatesan, M., Mott, H.R., Owen, D., Nietlispach, D., Lowe, P.N., and Laue, E.D. (2000). Structure of Cdc42 bound to the GTPase binding domain of PAK. Nat. Struct. Biol. 7, 384–388. Nassar, N., Horn, G., Herrmann, C., Scherer, A., McCormick, F., and Wittinghofer, A. (1995). The 2.2 A crystal structure of the Ras-binding domain of the serine/threonine kinase c-Raf1 in complex with Rap1A and a GTP analogue. Nature 375, 554–560. Noel, J.P., Hamm, H.E., and Sigler, P.B. (1993). The 2.2 A crystal structure of transducin-alpha complexed with GTP gamma S. Nature 366, 654–663. Phillips, M.J., Calero, G., Chan, B., Ramachandran, S., and Cerione, R.A. (2008). Effector proteins exert an important influence on the signaling-active state of the small GTPase Cdc42. J. Biol. Chem. 283, 14153–14164.

Bustelo, X.R., Sauzeau, V., and Berenjeno, I.M. (2007). GTP-binding proteins of the Rho/Rac family: regulation, effectors and functions in vivo. Bioessays 29, 356–370.

Piechulek, T., Rehlen, T., Walliser, C., Vatter, P., Moepps, B., and Gierschik, P. (2005). Isozyme-specific stimulation of phospholipase C-g2 by Rac GTPases. J. Biol. Chem. 280, 38923–38931.

Cherfils, J., Menetrey, J., Le Bras, G., Janoueix-Lerosey, I., de Gunzburg, J., Garel, J.R., and Auzat, I. (1997). Crystal structures of the small G protein Rap2A in complex with its substrate GTP, with GDP and with GTPgammaS. EMBO J. 16, 5582–5591.

Pleines, I., Elvers, M., Strehl, A., Pozgajova, M., Varga-Szabo, D., May, F., Chrostek-Grashoff, A., Brakebusch, C., and Nieswandt, B. (2008). Rac1 is essential for phospholipase C-gamma2 activation in platelets. Pflugers Arch. 457, 1173–1185.

Dvorsky, R., and Ahmadian, M.R. (2004). Always look on the bright site of Rho: structural implications for a conserved intermolecular interface. EMBO Rep. 5, 1130–1136.

Rebecchi, M.J., and Pentyala, S.N. (2000). Structure, function, and control of phosphoinositide-specific phospholipase C. Physiol. Rev. 80, 1291–1335.

Garrard, S.M., Capaldo, C.T., Gao, L., Rosen, M.K., Macara, I.G., and Tomchick, D.R. (2003). Structure of Cdc42 in a complex with the GTPasebinding domain of the cell polarity protein, Par6. EMBO J. 22, 1125–1133.

Rhee, S.-G. (2001). Regulation of phophoinositide-specific phospholipase C. Annu. Rev. Biochem. 70, 281–312. Ridley, A.J. (2006). Rho GTPases and actin dynamics in membrane protrusions and vesicle trafficking. Trends Cell Biol. 16, 522–529.

Herrmann, C. (2003). Ras-effector interactions: after one decade. Curr. Opin. Struct. Biol. 13, 122–129.

Sahai, E., and Marshall, C.J. (2002). RHO-GTPases and cancer. Nat. Rev. Cancer 2, 133–142.

Herrmann, C., Horn, G., Spaargaren, M., and Wittinghofer, A. (1996). Differential interaction of the Ras family GTP-binding proteins H-Ras, Rap1A, and R-Ras with the putative effector molecules Raf kinase and Ral-guanine nucleotide exchange factor. J. Biol. Chem. 271, 6794–6800.

Sorli, S.C., Bunney, T.D., Sugden, P.H., Paterson, H.F., and Katan, M. (2005). Signaling properties and expression in normal and tumor tissues of two phospholipase C epsilon splice variants. Oncogene 24, 90–100.

Illenberger, D., Schwald, F., Pimmer, D., Binder, W., Maier, G., Dietrich, A., and Gierschik, P. (1998). Stimulation of phospholipase C-beta2 by the Rho GTPases Cdc42Hs and Rac1. EMBO J. 17, 6241–6249. Illenberger, D., Walliser, C., Nurnberg, B., Diaz Lorente, M., and Gierschik, P. (2003). Specificity and structural requirements of phospholipase C-beta stimulation by Rho GTPases versus G protein beta/gamma dimers. J. Biol. Chem. 278, 3006–3014. Jaffe, A.B., and Hall, A. (2005). Rho GTPases: biochemistry and biology. Annu. Rev. Cell Dev. Biol. 21, 247–269. Jezyk, M.R., Snyder, J.T., Gershberg, S., Worthylake, D.K., Harden, T.K., and Sondek, J. (2006). Crystal structure of Rac1 bound to its effector phospholipase C-b2. Nat. Struct. Mol. Biol. 13, 1135–1140. Jin, R., Junutula, J.R., Matern, H.T., Ervin, K.E., Scheller, R.H., and Brunger, A.T. (2005). Exo84 and Sec5 are competitive regulatory Sec6/8 effectors to the RalA GTPase. EMBO J. 24, 2064–2074. Katan, M. (2005). New insights into the families of PLC enzymes: looking back and going forward. Biochem. J. 391, e7–e9. Keller, P.J., Gable, C.M., Wing, M.R., and Cox, A.D. (2005). Rac3-mediated transformation requires multiple effector pathways. Cancer Res. 65, 9883–9890. Kiel, C., Beltrao, P., and Serrano, L. (2008). Analyzing protein interaction networks using structural information. Annu. Rev. Biochem. 77, 415–441. Lemmon, M.A. (2004). Pleckstrin homology domains: not just for phosphoinositides. Biochem. Soc. Trans. 32, 707–711.

Spoerner, M., Herrmann, C., Vetter, I.R., Kalbitzer, H.R., and Wittinghofer, A. (2001). Dynamic properties of the Ras switch I region and its importance for binding to effectors. Proc. Natl. Acad. Sci. USA 98, 4944–4949. Tarricone, C., Xiao, B., Justin, N., Walker, P.A., Rittinger, K., Gamblin, S.J., and Smerdon, S.J. (2001). The structural basis of Arfaptin-mediated cross-talk between Rac and Arf signalling pathways. Nature 411, 215–219. Tolias, K.F., Cantley, L.C., and Carpenter, C.L. (1995). Rho family GTPases bind to phosphoinositide kinases. J. Biol. Chem. 270, 17656–17659. Vetter, I.R., Nowak, C., Nishimoto, T., Kuhlmann, J., and Wittinghofer, A. (1999). Structure of a Ran-binding domain complexed with Ran bound to a GTP analogue: implications for nuclear transport. Nature 398, 39–46. Vetter, I.R., and Wittinghofer, A. (2001). The guanine nucleotide-binding switch in three dimensions. Science 294, 1299–1304. Walliser, C., Retlich, M., Harris, R., Everett, K.L., Josephs, M.B., Vatter, P., Esposito, D., Driscoll, P.C., Katan, M., Gierschik, P., and Bunney, T.D. (2008). Rac regulates its effector phospholipase C gamma 2 through interaction with a split PH domain. J. Biol. Chem. 283, 30351–30362. Walmsley, M.J., Ooi, S.K., Reynolds, L.F., Smith, S.H., Ruf, S., Mathiot, A., Vanes, L., Williams, D.A., Cancro, M.P., and Tybulewicz, V.L. (2003). Critical roles for Rac1 and Rac2 GTPases in B cell development and signaling. Science 302, 459–462. Ye, M., Shima, F., Muraoka, S., Liao, J., Okamoto, H., Yamamoto, M., Tamura, A., Yagi, N., Ueki, T., and Kataoka, T. (2005). Crystal structure of M-Ras reveals a GTP-bound ‘‘off’’ state conformation of Ras family small GTPases. J. Biol. Chem. 280, 31267–31275.
