Advertisement

Stimulation of phospholipase C‐β2 by the Rho GTPases Cdc42Hs and Rac1

Daria Illenberger, Frieder Schwald, Dominik Pimmer, Wolfhard Binder, Gernot Maier, Alexander Dietrich, Peter Gierschik

Author Affiliations

  1. Daria Illenberger1,
  2. Frieder Schwald1,
  3. Dominik Pimmer1,
  4. Wolfhard Binder1,
  5. Gernot Maier2,
  6. Alexander Dietrich1 and
  7. Peter Gierschik*,1
  1. 1 Department of Pharmacology and Toxicology, University of Ulm, 89069, Ulm, Germany
  2. 2 German Cancer Research Center, 69009, Heidelberg, Germany
  1. *Corresponding author. E-mail: peter.gierschik{at}medizin.uni-ulm.de
View Full Text

Abstract

Neutrophils contain a soluble guanine‐nucleotidebinding protein, made up of two components with molecular masses of 23 and 26 kDa, that mediates stimulation of phospholipase C‐β2 (PLCβ2). We have identified the two components of the stimulatory heterodimer by amino acid sequencing as a Rho GTPase and the Rho guanine nucleotide dissociation inhibitor LyGDI. Using recombinant Rho GTPases and LyGDI, we demonstrate that PLCβ2 is stimulated by guanosine 5′‐O‐(3‐thiotriphosphate) (GTP[S])‐activated Cdc42Hs×LyGDI, but not by RhoA×LyGDI. Stimulation of PLCβ2, which was also observed for GTP[S]‐activated recombinant Rac1, was independent of LyGDI, but required C‐terminal processing of Cdc42Hs/Rac1. Cdc42Hs/Rac1 also stimulated PLCβ2 in a system made up of purified recombinant proteins, suggesting that this function is mediated by direct protein–protein interaction. The Cdc42Hs mutants F37A and Y40C failed to stimulate PLCβ2, indicating that the Cdc42Hs effector site is involved in this interaction. The results identify PLCβ2 as a novel effector of the Rho GTPases Cdc42Hs and Rac1, and as the first mammalian effector directly regulated by both heterotrimeric and low‐molecular‐mass GTP‐binding proteins.

Introduction

The signal transduction pathways responsible for activation of neutrophils by chemoattractants have been the subject of intense investigation over recent years. The binding of chemoattractants to their receptors on the neutrophil plasma membrane generates multiple intracellular second messengers and signalling molecules through the activation of phospholipases C, A2 and D, as well as through the activation of lipid and protein kinases (Bokoch, 1995; Ye and Boulay, 1997; Baggiolini, 1998). Phospholipase C (PLC) hydrolyses phosphatidylinositol 4,5‐bisphosphate (PtdInsP2) to produce two second messengers, d‐myo‐inositol‐1,4,5‐trisphosphate and sn‐1,2‐diacylglycerol. d‐myo‐inositol‐1,4,5‐trisphosphate causes the release of Ca2+ from intracellular stores, whereas sn‐1,2‐diacylglycerol activates protein kinase C. Molecular cloning has revealed the existence of three major families of PLC: β, γ and δ. PLCβ and PLCγ enzymes are activated by G protein‐coupled receptors (GCRs) and tyrosine‐kinase‐linked receptors, respectively. Whether PLCδ enzymes are regulated by cell‐surface receptors is unknown. PLCβ exists in four isoforms, PLCβ1‐4, which are activated by members of the αq subfamily of the G protein α subunits and, except for PLCβ4, by G protein βγ dimers (Rhee and Bae, 1997; Singer et al., 1997).

We have previously shown that cytosolic preparations of myeloid differentiated HL‐60 cells and of bovine neutrophils contain a soluble phospholipase C, which is stimulated by the poorly hydrolyzable GTP analogue guanosine 5′‐O‐(3‐thiotriphosphate) (GTP[S]) (Camps et al., 1990; Illenberger et al., 1997). A factor mediating GTP[S]‐dependent stimulation of PLC in neutrophils was resolved from endogenous PLC, and functionally reconstituted with exogenous recombinant PLCβ2 (Illenberger et al., 1997). The resolved protein mediated GTP[S]‐dependent activation of PLCβ2, but not PLCβ1 and PLCδ1. Stimulation of phosphatidylinositol‐4‐phosphate 5‐kinase was not involved in this activation. Purification of the stimulatory factor revealed that it consisted of two proteins of apparent molecular masses of 23 and 26 kDa. The purified heterodimer stimulated a deletion mutant of PLCβ2, PLCβ2Δ, lacking a C‐terminal region necessary for stimulation by G protein αq subunits. These results led us to suggest that a GTP‐binding protein distinct from αq subunits, most likely a low‐molecular‐mass GTP‐binding protein associated with a regulatory protein, is involved in activation of PLCβ2.

Herein, we report the identification of the PLCβ2‐stimulating factor of bovine neutrophils as a member of the Rho subfamily of low‐molecular‐mass GTP‐binding proteins associated with the Rho GDP dissociation inhibitor LyGDI. We demonstrate that GTP[S]‐activated recombinant Cdc42Hs and Rac1, but not RhoA, stimulate the activity of PLCβ2. While the presence of LyGDI is not necessary for PLCβ2 activation, both C‐terminal processing of Cdc42Hs and Rac1 and integrity of their effector‐regulating domains are required for this effect. These results identify PLCβ2 as a novel effector of the Rho GTPases Cdc42Hs and Rac1.

Results

The purification of the PLCβ2‐stimulating guanine nucleotide‐binding protein from soluble fractions of bovine neutrophils revealed that it consists of two polypeptides of ∼23 and 26 kDa (Illenberger et al., 1997). The two proteins were subjected to protease digestion and proteolytic fragments were obtained for amino acid sequence analysis. The sequences of two fragments of the 23 kDa protein (DQFPEVYVPTVFENY and DRLRPLSYP) were identical as partial amino acid sequences of bovine RhoA (Ogorochi et al., 1989) and human RhoC (Chardin et al., 1988). Three peptides of the 26 kDa fragments were highly related in amino acid sequence to internal peptide fragments of human LyGDI (Scherle et al., 1993) and D4 (Lelias et al., 1993), but were clearly different from the corresponding fragments of bovine RhoGDI (Fukumoto et al., 1990) (DESLTKYTKTLLG, DPTAPNVKVTRLTLVXESAP, DKATFMVGSYGPRPE; italics, residues not present in human LyGDI/D4; underlined, residues not present in bovine RhoGDI). LyGDI and D4 are homologues of the Rho guanine nucleotide dissociation inhibitor RhoGDI, and are highly expressed in hematopoietic cells (Lelias et al., 1993; Scherle et al., 1993). Their amino acid sequences differ by only two out of 202 residues. Taken together, these results led us to conclude that the soluble PLCβ2‐stimulating guanine nucleotide‐binding protein of bovine neutrophils is a heterodimer of a Rho GTPase and LyGDI/D4.

To obtain further functional support of this notion, recombinant glutathione S‐transferase fusion protein of human LyGDI (GST–LyGDI) was produced in Escherichia coli in order to deplete the Rho GTPases from the PLCβ2‐stimulating factor purified from bovine neutrophils. To this end, matrix‐immobilized GST–LyGDI was incubated with the PLCβ2‐stimulating factor. Following incubation, matrix and supernatant were separated by centrifugation, and bound proteins were eluted from the matrix with glutathione. Figure 1A (left) shows that the treatment of the PLCβ2‐stimulating factor with matrix‐immobilized GST–LyGDI led to an almost complete loss of the GTP[S]‐dependent stimulation of PLCβ2Δ. Importantly, the stimulatory activity was recovered from the matrix by treatment with glutathione (Figure 1B, right). Thus, an almost 2‐fold stimulation of inositol phosphate formation by GTP[S] was observed upon reconstitution of the eluate with PLCβ2Δ. These results provide direct evidence that the PLCβ2‐stimulating factor specifically interacts with LyGDI and thus confirm its identity as a member of the Rho subfamily of Ras‐related proteins.

Figure 1.

Depletion of the PLCβ2‐stimulating activity from the purified soluble stimulatory factor with a GST fusion protein of LyGDI. The PLCβ2‐stimulating factor was incubated as indicated on the abscissa with GST or GST–LyGDI bound to Glutathione Sepharose 4B. (A) After removal of the matrix by centrifugation, 10 μl aliquots of the supernatants were supplemented with soluble proteins of insect cells expressing PLCβ2Δ (0.2 μg protein/sample) and incubated in the absence (open bars) or presence (hatched bars) of 100 μM GTP[S] with phospholipid vesicles containing PtdInsP2 (left panel). The proteins bound to the affinity matrix were eluted with glutathione and concentrated. Aliquots (10 μl) of the concentrated solutions were supplemented with soluble proteins of insect cells expressing PLCβ2Δ (0.2 μg protein/sample) and incubated in the absence (open bars) or presence (hatched bars) of 100 μM GTP[S] with phospholipid vesicles containing PtdInsP2 (right panel). The reaction was terminated by the addition of chloroform/methanol/concentrated HCl, and the mixture analysed for inositol phosphates. See Materials and methods for experimental details. (B) Aliquots of the sample containing the PLCβ2‐stimulating factor (lanes 1 and 6), of the supernatants obtained after incubation of the factor with matrix‐immobilized GST (lanes 2 and 7) or GST–LyGDI (lanes 3 and 8), or of the concentrated solutions containing proteins bound to matrix‐immobilized GST (lanes 4 and 9) or GST–LyGDI (lanes 5 and 10) were subjected to SDS–PAGE and immunoblotting was performed using antibodies reactive against either RhoA (lanes 1–5) or Cdc42Hs (lanes 6–10). The data are representative of three independent experiments with similar results.

LyGDI/D4 has been reported to interact with RhoA, Cdc42Hs and Rac1, but not with Ras and Rap (Adra et al., 1993; Scherle et al., 1993). In order to identify the Rho GTPase(s) associated with recombinant GST–LyGDI upon incubation with the PLCβ2‐stimulating factor from bovine neutrophils, immunostaining of the samples was performed using polyclonal antisera specifically reactive against human RhoA, Cdc42Hs and Rac1. Figure 1B shows that the purified PLCβ2‐stimulating factor was highly reactive with antisera raised against RhoA and Cdc42Hs. No immunoreactivity was observed with the serum raised against human Rac1 (results not shown). The specificity of the antisera was verified using bacterially expressed human RhoA, Cdc42Hs and Rac1 (not shown). Analysis of the fractions obtained after incubation of the factor with matrix‐immobilized GST–LyGDI revealed a reduction of both the RhoA and the Cdc42Hs immunoreactivity. Most importantly, both immunoreactivities were recovered upon glutathione elution of the matrix containing GST–LyGDI (Figure 1B). No reactivity was detected in the eluate with serum raised against human Rac1 (not shown). These data demonstrate that at least two Rho GTPases, RhoA and Cdc42Hs, are present in the soluble PLCβ2‐stimulating factor from bovine neutrophils.

To determine whether LyGDI×Cdc42Hs and/or LyGDI×RhoA mediate(s) the GTP[S]‐dependent stimulation of PLCβ2, the proteins were produced as recombinant heterodimers for functional characterization. To this end, the heterodimers were prepared by solubilization of recombinant isoprenylated human Cdc42Hs or RhoA from membranes of baculovirus‐infected insect cells by bacterially expressed, purified GST–LyGDI. The soluble fractions obtained after incubation of insect cell membranes with GST or GST–LyGDI were analysed by Western blotting. Figure 2A shows that GST–LyGDI was capable of extracting both Cdc42Hs and RhoA into the supernatant. The fact that only minor quantities of Cdc42Hs were observed in the supernatant after treatment of membranes with GST demonstrates that the extraction of the Rho GTPases was due to specific interaction with LyGDI. Figure 2B shows that the heterodimer made up of GST–LyGDI and Cdc42Hs led to a marked (∼3‐fold) stimulation of the activity of recombinant PLCβ2Δ. In contrast, no stimulation of PLCβ2Δ was observed for heterodimeric GST–LyGDI×RhoA or free GST–LyGDI. These results demonstrate that LyGDI×Cdc42Hs, but not LyGDI×RhoA or LyGDI alone, mediates the GTP[S]‐dependent stimulation of PLCβ2.

Figure 2.

Stimulation of PLCβ2Δ by recombinant heterodimeric GST–LyGDI×Cdc42Hs. (A) Membranes of insect cells infected with baculovirus encoding Cdc42Hs (lanes 1 and 2), RhoA (lane 3) or β‐galactosidase (lane 4) were incubated with 45 μM purified, bacterially expressed GST (lane 1) or GST–LyGDI (lanes 2–4). The membranes were sedimented by centrifugation and the supernatants were concentrated. Aliquots of the concentrates were subjected to SDS–PAGE and immunoblotting was performed using antibodies reactive against (a) Cdc42Hs, (b) RhoA, (c) LyGDI or (d) GST. (B) Aliquots (10 μl) of the concentrates were supplemented with soluble proteins of insect cells expressing PLCβ2Δ (0.2 μg protein/sample). The samples were incubated in the absence (open bars) or presence (hatched bars) of 100 μM GTP[S] with phospholipid vesicles containing PtdInsP2.

The next experiments were designed to examine whether both constituents of the Cdc42×LyGDI heterodimer are necessary for PLCβ2 stimulation or whether monomeric Cdc42Hs is sufficient to elicit this effect. Recombinant Cdc42Hs was solubilized from membranes of baculovirus‐infected insect cells by extraction with detergent‐containing buffer. As shown in Figure 3A, addition of GTP[S] greatly (∼3.6‐fold) enhanced the PLCβ2Δ activity in the presence of extracts containing Cdc42Hs. In contrast, no stimulation by GTP[S] was observed in the presence of extracts from control cells expressing recombinant β‐galactosidase rather than Cdc42Hs. Thus, monomeric Cdc42Hs is sufficient to mediate GTP[S]‐dependent stimulation of PLCβ2.

Figure 3.

Specificity of PLCβ2Δ stimulation by Rho GTPases. (A) Membranes of insect cells infected with baculovirus encoding β‐galactosidase (βGal) or Cdc42Hs were extracted with buffer containing sodium cholate. Aliquots (10 μl) of the detergent extracts were supplemented with soluble proteins of insect cells expressing PLCβ2Δ (0.4 μg protein/sample) and incubated in the absence (open bars) or presence (hatched bars) of 100 μM GTP[S] with phospholipid vesicles containing PtdInsP2. The activities of the detergent extracts and the preparation containing soluble PLCβ2Δ alone in the absence or presence of 100 μM GTP[S] is shown for comparison. In (B), membranes of insect cells infected with baculovirus encoding RhoA, Cdc42Hs or Rac1 were extracted with buffer containing sodium cholate. The abundance of the recombinant Rho GTPases in the detergent extracts was estimated by 35S‐GTP[S] binding. Aliquots of the detergent extracts containing equal amounts of the three Rho GTPases (12 pmol) were supplemented with soluble proteins of insect cells expressing PLCβ2Δ (0.5 μg protein/sample) and incubated in the absence (open bars) or presence (hatched bars) of 100 μM GTP[S] with phospholipid vesicles containing PtdInsP2. The activity of the preparation containing soluble PLCβ2Δ alone in the absence or presence of 100 μM GTP[S] is shown for comparison. The activities of the detergent extracts containing RhoA, Cdc42Hs or Rac1 were 0.39 ± 0.13 pmol/min, not detectable, and 0.13 ± 0.02 pmol/min, respectively, in the absence of PLCβ2Δ, and were independent of the presence of GTP[S] (results not shown). The inset of (B) shows the results of the immunochemical analysis of the three detergent extracts. Aliquots of the three samples were subjected to SDS–PAGE and immunoblotting was performed using antibodies reactive against RhoA (lane 1), Cdc42Hs (lane 2) or Rac1 (lane 3).

Since it is well‐established that a wide range of effectors is regulated by both Cdc42Hs and Rac1, but not by RhoA (Van Aelst and D'Souza‐Schorey, 1997), the specificity of PLCβ2 stimulation by Rho family members was investigated. The recombinant GTPases Cdc42Hs, Rac1 and RhoA were extracted from membranes of baculovirus‐infected insect cells with detergent‐containing buffer and reconstituted with recombinant PLCβ2Δ. In agreement with the results shown in Figure 2, the activity of PLCβ2Δ was not stimulated by GTP[S]‐activated RhoA (Figure 3B). Additional experiments showed that RhoA failed to mediate GTP[S]‐dependent stimulation of PLCβ2Δ, even at concentrations up to 800 nM (not shown). In contrast, addition of both Cdc42Hs and Rac1 to PLCβ2Δ led to marked GTP[S]‐dependent increases in inositol phosphate formation. Both GTPases, Cdc42Hs and Rac1, stimulated PLCβ2Δ activity to a similar extent (6.2‐ and 4.8‐fold, respectively), when tested at the same concentration (200 nM). Thus, Rac1, although not detected by immunochemical means in the purified preparation from bovine neutrophils, may represent an additional Rho GTPase capable of stimulating PLCβ2.

In order to determine whether stimulation of PLCβ2 by Cdc42Hs or Rac1 is mediated by direct protein–protein interaction or whether additional proteins are required to mediate this effect, the protein components were purified from insect cells for functional reconstitution. PLCβ2Δ and LyGDI were purified from soluble fractions, and Cdc42Hs and Rac1 were purified from membranes of baculovirus‐infected insect cells. Figure 4A shows the result of the analysis of the purified proteins by SDS–PAGE. The proteins were reconstituted and the inositol phosphate formation was measured in the absence and presence of GTP[S]. As shown in Figure 4B, purified Cdc42Hs and Rac1 indeed caused a marked (∼3.4‐ and 4.3‐fold, respectively) stimulation of purified PLCβ2Δ.

Figure 4.

Direct activation of PLCβ2Δ by Cdc42Hs and Rac1. Recombinant PLCβ2Δ and LyGDI were purified from soluble fractions of baculovirus‐infected insect cells. Recombinant Cdc42Hs and Rac1 were purified from membranes of baculovirus‐infected insect cells. (A) Aliquots of purified PLCβ2Δ (5 μg of protein) (lane 1), purified LyGDI (1 μg of protein) (lane 2), Cdc42Hs (0.5 μg of protein) (lane 3) and Rac1 (1 μg of protein) (lane 4) were subjected to SDS–PAGE. Proteins were visualized by staining with silver. The positions of PLCβ2Δ, LyGDI, Rac1, Cdc42Hs and the molecular weight standards are indicated. We noticed a degradation of purified PLCβ2Δ upon storage and/or freezing and thawing, resulting in fragments migrating at ∼90, 45 and 27 kDa. Attempts to resolve these fragments from full‐length PLCβ2Δ by several chromatographic procedures were unsuccessful (results not shown). (B) Aliquots of purified LyGDI, Cdc42Hs and Rac1 were supplemented with purified PLCβ2Δ (150 ng protein/sample) and incubated in the absence (open bars) or presence (hatched bars) of 100 μM GTP[S] with phospholipid vesicles containing PtdInsP2. The final concentrations of LyGDI, Cdc42Hs and Rac1 were 10, 140 and 220 nM, respectively.

Even at a concentration of 10 μM, purified LyGDI had little, if any, effect on the activity of PLCβ2Δ, regardless of whether GTP[S] was absent or present in the incubation medium. These data suggest that stimulation of PLCβ2 by Cdc42Hs and Rac1 is due to direct interaction between the GTPases and PLCβ2. In additional experiments (results not shown), we found that purified LyGDI had no effect on the ability of GTP[S] to activate PLCβ2Δ in the presence of purified Cdc42Hs when present in up to 100‐fold excess of the Rho GTPase. Only at very high concentrations of LyGDI (≥200‐fold excess of Cdc42Hs or Rac1) was GTP[S]‐dependent activation of the enzyme reduced.

The observation that LyGDI is apparently not required for PLCβ2 stimulation by GTP[S]‐activated Cdc42Hs raises the question of whether the posttranslational modifications of Cdc42Hs, which are required for its interaction with GDIs (Leonard et al., 1992), are necessary for PLCβ2 stimulation. To address this question, C‐terminally modified Cdc42Hs extracted with detergent‐containing buffer from membranes and soluble, unmodified Cdc42Hs purified from cytosol of baculovirus‐infected insect cells were compared in their abilities to stimulate recombinant PLCβ2Δ. A marked ∼3.6‐fold stimulation of PLCβ2 activity by GTP[S] was observed in the presence of modified Cdc42Hs, but not in the presence of a similar quantity of soluble Cdc42Hs (Figure 5A). These results suggest that C‐terminal processing of Cdc42Hs is a prerequisite for the ability of the protein to stimulate PLCβ2.

Figure 5.

Structural elements of Cdc42Hs involved in PLCβ2Δ stimulation. (A) C terminus. Cdc42Hs was either purified from cytosol or extracted with buffer containing sodium cholate from membranes of insect cells infected with baculovirus encoding Cdc42Hs. The abundance of recombinant Cdc42Hs in the two preparations was estimated by 35S‐GTP[S] binding. Aliquots of the two preparations containing 15 pmol soluble Cdc42Hs (sCdc42Hs) or 22 pmol membrane‐bound Cdc42Hs (mCdc42Hs) were supplemented with soluble proteins of insect cells expressing PLCβ2Δ (0.5 μg protein/sample) and incubated in the absence (open bars) or presence (hatched bars) of 100 μM GTP[S] with phospholipid vesicles containing PtdInsP2. Inset: aliquots of the two preparations containing 15 pmol soluble Cdc42Hs (lane 1) or 22 pmol membrane‐bound Cdc42Hs (lane 2) were subjected to SDS–PAGE and immunoblotting was performed using antibodies reactive against Cdc42Hs. (B) Effector‐regulating domain. Membranes of insect cells infected with baculovirus encoding wild‐type Cdc42Hs (Cdc42Hswt), or the mutants Cdc42HsF37A and Cdc42HsY40C, were extracted with buffer containing sodium cholate. The abundance of the recombinant Cdc42Hs proteins in the detergent extracts was estimated by 35S‐GTP[S] binding. Aliquots of the detergent extracts containing equal amounts of the three GTPases (30 pmol) were supplemented with soluble proteins of insect cells expressing PLCβ2Δ (0.5 μg protein/sample) and incubated as indicated at the abscissa in the absence or presence of 100 μM GTP[S] with phospholipid vesicles containing PtdInsP2. The activities of the detergent extracts in the absence of PLCβ2Δ are shown for comparison. Inset: aliquots of the detergent extracts containing equal amounts (10 pmol) of Cdc42Hswt (lane 1), Cdc42HsF37A (lane 2) or Cdc42HsY40C (lane 3) were subjected to SDS–PAGE and immunoblotting was performed using antibodies reactive against Cdc42Hs.

Substitution of amino acids in a region (residues 26–45) of Cdc42Hs and Rac homologous to the effector region in Ras has been shown to interfere with their binding to and regulation of downstream effectors (Diekmann et al., 1994; Lamarche et al., 1996; Westwick et al., 1997). To investigate whether the effector region in Cdc42Hs is involved in the stimulation of PLCβ2, two single‐amino‐acid substitutions were introduced into the effector domain of Cdc42Hs. Both mutants, Cdc42HsF37A and Cdc42HsY40C, were expressed in baculovirus‐infected insect cells, and their ability to bind 35S‐GTP[S] was determined. Cdc42HsF37A and Cdc42HsY40C were indistinguishable from wild‐type Cdc42Hs in terms of their 35S‐GTP[S] binding parameters (not shown). However, in marked contrast to wild‐type Cdc42Hs, which caused an ∼6‐fold stimulation of recombinant PLCβ2, neither Cdc42HsF37A nor Cdc42HsY40C affected the activity of PLCβ2Δ in the absence or presence of GTP[S] (Figure 5B). Therefore, the integrity of the effector‐regulating region of Cdc42Hs is required for PLCβ2 stimulation.

A comparison of the effects of increasing concentrations of purified βγt and GTP[S]‐activated recombinant Cdc42Hs on purified PLCβ2Δ is presented in Figure 6. The results show that βγt caused a higher maximal enzyme stimulation (∼6.4‐fold) than did GTP[S]‐activated Cdc42Hs (∼3.6‐fold), but was required at ∼5‐fold higher concentrations to elicit this effect. Thus, half‐maximal stimulation of PLCβ2Δ required ∼400 nM βγt, but only ∼70 nM Cdc42Hs. In additional experiments (not shown), we found that the potency of GTP[S]‐activated Rac1 to stimulate PLCβ2Δ was ∼20‐fold higher than the potency of GTP[S]‐activated Cdc42Hs. Consistent with earlier results (Illenberger et al., 1997), we observed that wild‐type recombinant PLCβ2 was stimulated to a lower maximal extent by Cdc42Hs or Rac1 (up to ∼2.5‐fold) than PLCβ2Δ. No evidence was obtained for synergism between either Cdc42Hs or Rac1 and βγt upon co‐stimulation of PLCβ2Δ or wild‐type PLCβ2.

Figure 6.

Stimulation of PLCβ2Δ by βγt versus Cdc42Hs. Purified recombinant PLCβ2Δ (120 ng protein/sample) was incubated at increasing concentrations of βγt purified from bovine retinal rod outer segments (diamonds) or recombinant wild‐type Cdc42Hs extracted with buffer containing sodium cholate from membranes of baculovirus‐infected insect cells (circles) with phospholipid vesicles containing PtdInsP2. The incubation was performed in the presence of 1.9 mM sodium cholate and 100 μM GTP[S]. Cdc42Hs did not stimulate PLCβ2Δ in the absence of GTP[S] (not shown).

Discussion

We have previously shown that cytosolic preparations of bovine neutrophils contain a soluble, heterodimeric guanine nucleotide‐binding protein that mediates stimulation of PLCβ2. The molecular masses of the two components of the purified heterodimer, 23 and 26 kDa, suggested that a low‐molecular‐mass GTP‐binding protein might mediate PLCβ2 activation. This notion was supported by the finding that the C‐terminal region of PLCβ2, which is necessary for stimulation by members of the αq subfamily of the heterotrimeric G protein α subunits, was not required for PLCβ2 stimulation by the purified heterodimer (Illenberger et al., 1997). In this study, we identified the constituents of the purified PLCβ2‐stimulating protein from bovine neutrophils as the Rho GTPase Cdc42 and its regulatory protein LyGDI. This result is consistent with our previous observations and demonstrates, for the first time, that PLCβ2 is regulated not only by heterotrimeric G proteins, but also by low molecular mass GTP‐binding proteins of the Rho family. In addition, our experiments demonstrate that a recombinant heterodimer made up of human LyGDI and Cdc42Hs substituted for the heterodimer purified from bovine neutrophils in terms of its ability to stimulate PLCβ2. Interestingly, the GTPase Cdc42Hs alone was sufficient to mediate GTP[S]‐dependent stimulation of PLCβ2. Analysis of the Rho family member specificity of PLCβ2 stimulation revealed that this stimulation was not limited to activated Cdc42Hs, but was also observed with activated Rac1. Importantly, activated RhoA was without effect, adding PLCβ2 to the growing list of targets regulated by Cdc42Hs and Rac1, but not by RhoA.

An important outcome of this study is the observation that stimulation of PLCβ2 by Cdc42Hs required C‐terminal processing of Cdc42Hs (Glomset and Farnsworth, 1994). It should be noted that the inability of soluble unmodified Cdc42Hs to stimulate PLCβ2 was not due to its inability to bind GTP[S]. Both modified and unmodified Cdc42Hs possessed similar 35S‐GTP[S] binding activities when tested under a variety of conditions, including those used for measurement of their abilities to stimulate PLCβ2 (cf. legend to Figure 5A). Rho GTPases need to be C‐terminally modified in order to interact with guanine nucleotide dissociation inhibitors (GDIs) and, at least in some cases, with guanine nucleotide exchange factors (GEFs) and effectors (reviewed in Takai et al., 1995). The results reported here demonstrate that C‐terminal modification of Cdc42Hs is not required for guanine nucleotide exchange, at least under the conditions used in this study, but is a prerequisite to its ability to mediate GTP[S]‐dependent stimulation of PLCβ2. The effector site substitutions Cdc42HsF37A and Cdc42HsY40C resulted in mutants which failed to stimulate PLCβ2, indicating that both C‐terminal modification(s) and the integrity of the effector‐regulating region of Cdc42Hs are required for PLCβ2 activation. In addition, our results obtained with purified proteins suggest that Cdc42Hs and Rac1 directly interact with PLCβ2. Taken together, these data indicate that PLCβ2 serves as an effector rather than a regulator of Cdc42Hs and Rac1.

It is likely that regulation of PLCβ2 activity by Cdc42Hs and Rac1 is subject to control of the GTPases by regulatory proteins, e.g. Rho family GDIs, GEFs or GAPs (Van Aelst and D'Souza‐Schorey, 1997). Interestingly, GTP[S] is shown in this study to promote the Cdc42Hs‐mediated stimulation of PLCβ2 activity, even in the presence of LyGDI and in the absence of any GEF. This implies spontaneous guanine nucleotide exchange under the conditions of our experiments, which is consistent with previous findings which indicate that such an exchange is feasible at low magnesium concentrations and/or in the presence of PtdInsP2 (Self and Hall, 1995a; Zheng et al., 1996). The existence of a specific GEF in neutrophils, which might enhance the Cdc42Hs/Rac1‐mediated PLCβ2 stimulation, and the functional role of LyGDI will be investigated in future studies.

The findings of this study have important implications for the role of Cdc42Hs‐mediated PLCβ2 stimulation in pathways controlled by Cdc42Hs in intact cells. Rho GTPases control the assembly and organization of the actin cytoskeleton in eukaryotic cells and regulate intracellular protein kinase cascades and gene transcription (Van Aelst and D'Souza‐Schorey, 1997; Hall, 1998). Cdc42 and Rac1 mutants carrying effector site substitutions have previously been used to dissect the biochemical pathways involved in these cellular responses (Joneson et al., 1996; Lamarche et al., 1996; Westwick et al., 1997). Results from these studies have provided evidence for bifurcating pathways controlling cytoskeletal changes and MAP kinase activation. Thus, amino acid substitution at position 40 of Cdc42 blocked c‐Jun N‐terminal kinase (JNK) activation without affecting activation of Rac1, while changes at codon 37 had the opposite effect. Both mutants, however, still induced DNA synthesis, and caused formation of filopodia and focal complexes (Lamarche et al., 1996). Our own observation that both Cdc42HsY40C and Cdc42HsF37A failed to stimulate PLCβ2 suggests that PLCβ2 may represent an additional independent effector moiety regulated by Cdc42Hs. Furthermore, PLCβ2 stimulation by Cdc42Hs is unlikely to be involved in Cdc42Hs‐mediated activation of JNK, activation of Rac1, and formation of filopodia and focal complexes, at least in fibroblasts, since these pathways are activated by either Cdc42Y40C or Cdc42F37A.

An important issue concerning the role of Cdc42Hs and Rac1 in the regulation of PLCβ2 is how these low‐molecular‐mass GTP‐binding proteins fit into the scheme by which GCRs control this enzyme in intact cells. Previous results provide strong evidence that GCRs activate PLCβ2 in neutrophils via βγ dimers released from pertussis toxin‐sensitive G proteins (Camps et al., 1992; Jiang et al., 1996). However, chemoattractant receptors have also been shown to cause the activation of Rho GTPases in intact leukocytes (Laudanna et al., 1996). The mechanisms by which this activation takes place are largely unknown. Several scenarios are possible. First, GCRs could interact directly with Rho GTPases to catalyse their activation (Polakis et al., 1989; Wieland et al., 1990; Mitchell et al., 1998). Secondly, Rho GTPases could be regulated through GCRs via direct interaction with free G protein βγ dimers (Harhammer et al., 1996; Alberts et al., 1998). Thirdly, subunits of receptor‐activated heterotrimeric G proteins could interact with and activate GEFs of Rho GTPases (Zhao et al., 1995; Hart et al., 1998; Nern and Arkowitz, 1998). Finally, leukocyte GCRs have been shown to cause alterations in the levels and/or activities of multiple intracellular second messengers and signalling molecules, some of which are known to modulate, either directly or indirectly, the activity of Rho GTPases (Hawkins et al., 1995; Han et al., 1998). Importantly, Cdc42 and/or Rac1 may also be activated by the latter mechanism(s) via receptors not coupled to G proteins, e.g. receptor tyrosine kinases. Signals emerging from both G protein‐coupled receptors and tyrosine kinases may converge at the level of PLCβ2. Of interest, signalling through CD5, a co‐receptor and tyrosine kinase substrate of T lymphocytes, has recently been shown to cause an increase in intracellular Ca2+ through a process requiring phosphatidylinositol 3‐kinase, Vav and Rac‐1 (Gringhuis et al., 1998).

In conclusion, we have shown that a heterodimer made up of Cdc42 and LyGDI mediates GTP[S] stimulation of PLCβ2 in soluble fractions of neutrophils. Our results identify PLCβ2 as a new target of the Rho GTPases Cdc42Hs and Rac1, but not RhoA. PLCβ2 may, therefore, represent an integral part of signal transduction pathways activated by Cdc42Hs and/or Rac1 in neutrophils and, possibly, other cells expressing this effector enzyme.

Materials and methods

Amino acid sequence analysis of the PLCβ2Δ‐stimulating factor purified from bovine neutrophils

The PLCβ2‐stimulating factor present in soluble fractions of bovine neutrophils was purified as described previously (Illenberger et al., 1997). For amino acid sequence analysis, the 23 and 26 kDa constituents of the purified protein were separated by SDS–PAGE and electroblotted onto ProBlott membranes (Applied Biosystems). The bands were excised and subjected to digestion with endoproteinase Asp‐N (Boehringer Mannheim). The proteolytic fragments were fractionated by reverse‐phase high‐performance liquid chromatography prior to sequence analysis on an Applied Biosystems Model 470A automated gas phase sequenator.

Plasmids

The cDNAs of human Rho GTPases were kindly provided by Drs Martijn F.B.G.Gebbink, Wouter H.Moolenaar, Rainer Müller and Martin Bähler. The cDNA of human LyGDI (Scherle et al., 1993) was amplified from poly(A)+ RNA of undifferentiated HL‐60 cells by reverse‐transcriptase–PCR (RT–PCR) using synthetic oligonucleotides complementary to the ends of the coding region and carrying recognition sites for BamHI and EcoRI on their 5′ and 3′ ends, respectively. The PCR product was ligated into the EcoRI/BamHI sites of the prokaryotic expression vector pGEX‐2TK (Pharmacia) and the baculovirus transfer vector pVL1393 (Invitrogen). Point mutations were introduced into the cDNA of Cdc42Hs by the PCR overlap extension method (Horton and Pease, 1991) to generate the cDNAs of Cdc42HsF37A and Cdc42HsY40C. The cDNAs encoding the Cdc42Hs mutants were ligated into the BamHI/EcoRI site of pVL1393. All cloned PCR products were sequenced using the T7 Sequenase version 2.0 DNA sequencing kit (Amersham).

Production of recombinant proteins in E.coli

GST fusion proteins of RhoA, Cdc42Hs, Rac1 and LyGDI were produced in E.coli DH5α and purified by chromatography on Glutathione Sepharose 4B (Pharmacia) according to the method described previously (Self and Hall, 1995b). The purified GST fusion proteins were dialysed extensively against buffer containing 20 mM Tris–HCl pH 7.6, 2 mM MgCl2, 1 mM dithiothreitol, 0.1 mM phenylmethylsulfonyl fluoride and 1 μM GDP (GST–RhoA, GST–Cdc42Hs and GST–Rac1) or 20 mM Tris–HCl pH 7.5, 1 mM EDTA, 1 mM dithiothreitol and 0.1 mM phenylmethylsulfonyl fluoride (GST–LyGDI). The purity of the proteins was at least 90%, as judged by analysis on SDS–polyacrylamide gels stained with Coomassie Blue.

Production of recombinant proteins in baculovirus‐infected insect cells

Recombinant baculoviruses were produced as described (Dietrich et al., 1992). Recombinant proteins were produced by infecting Spodoptera frugiperda cells (Sf9 cells, Invitrogen) with recombinant baculoviruses as described (Dietrich et al., 1994). PLCβ2Δ, a deletion mutant of human PLCβ2Δ, lacking a C‐terminal region necessary for stimulation by αq subunits (F819–E1166) and carrying a serine to alanine replacement in position 2, was produced in baculovirus‐infected Sf9 or Trichoplusia ni cells (High Five™ cells, Invitrogen). PLCβ2Δ is indistinguishable from wild‐type recombinant PLCβ2 in terms of its interaction with PtdInsP2, Ca2+ and G protein βγ dimers.

Purification of recombinant Rho GTPases and LyGDI from baculovirus‐infected insect cells

Baculovirus‐infected Sf9 cells were harvested and homogenized, and membranes and soluble proteins were prepared from the homogenate as described previously (Cerione et al., 1995). Soluble Cdc42Hs was purified from the soluble fraction of baculovirus‐infected insect cells by gel permeation chromatography using a HiLoad 26/60 Superdex 75 prep grade column (Pharmacia) (Illenberger et al., 1997). Membrane‐bound Cdc42Hs and Rac1 was solubilized by extracting the membranes with buffer containing 23 mM sodium cholate and the detergent‐solubilized Rho GTPases were purified according to Cerione (1995), except that the proteins obtained by chromatography on Mono Q were diluted ∼5‐fold with buffer A (20 mM Tris–HCl pH 8.0, 23 mM sodium cholate) and applied to a column (0.5×0.5 cm) of hydroxylapatite (Calbiochem, HPLC grade) connected to a SMART micropurification chromatography system (Pharmacia) which had been equilibrated with buffer A. The flow rate was 50 μl/min. After application of the sample, the resin was washed with 1 ml of buffer A and eluted with a linear gradient (1 ml) starting with buffer A and ending with buffer B containing 20 mM Tris–HCl pH 8.0, 1 mM EDTA, 1 mM dithiothreitol, 100 mM K2HPO4/H3PO4, 40% (v/v) glycerol and 23 mM sodium cholate. Recombinant Rho GTPases eluted in single peaks at ∼50 mM K2HPO4/H3PO4. LyGDI was purified from the soluble fraction of baculovirus‐infected insect cells by sequential chromatography on Superdex 75 prep grade, and Mono Q as described for the purification of the soluble PLCβ2Δ‐stimulating factor from bovine neutrophils (Illenberger et al., 1997).

Purification of recombinant PLCβ2Δ

The soluble fraction of Trichoplusia ni cells infected with baculovirus encoding PLCβ2Δ (10 ml, 200 mg of protein) was applied to a column (1.6×8 cm) of Heparin Sepharose CL‐6B (Pharmacia) that had been equilibrated with buffer A [20 mM Tris–HCl pH 8.0, 1 mM EDTA, 1 mM dithiothreitol, 0.2 mM phenylmethylsulfonyl fluoride, 3 mM benzamidine, 10 μM leupeptin, 2 μM pepstatin, 1 μg/ml aprotinin, 16 μg/ml l‐1‐chloro‐3‐(4‐tosylamido)‐4‐phenyl‐2‐butanone (TPCK), 16 μg/ml l‐1‐chloro‐3‐(4‐tosylamido)‐7‐amino‐2‐heptanone hydrochloride (TLCK), 2 μg/ml soybean trypsin inhibitor]. The flow rate was 1 ml/min. After application of the sample, the column was washed with 20 ml of buffer A and eluted with a linear gradient (60 ml) of NaCl (0–900 mM) in buffer A, followed by 10 ml of buffer A containing 900 mM NaCl. Fractions of 1 ml were collected and analysed by SDS–PAGE and immunoblotting using antibodies reactive against PLCβ2Δ (Camps et al., 1992). Fractions containing maximal levels of PLCβ2Δ were pooled (15 ml, 40 mg of protein), concentrated ∼10‐fold by pressure filtration in a stirred cell equipped with an Amicon PM10 membrane, diluted 4‐fold with buffer A and applied to a Mono Q HR 5/5 column (Pharmacia), which had been equilibrated with buffer A containing 100 mM NaCl. The flow rate was 0.5 ml/min. After application of the sample, the column was washed with 5 ml of buffer A containing 100 mM NaCl and eluted with a linear gradient (10 ml) of NaCl (100–900 mM) in buffer A, followed by 5 ml of buffer A containing 900 mM NaCl. Fractions of 500 μl were collected into tubes containing 50 μl of 10‐fold concentrated protease inhibitors as listed above. Fractions containing PLCβ2Δ (1 ml, 10 mg of protein) were diluted 4‐fold with buffer B (20 mM Tris–HCl pH 7.5, 100 μM EDTA, 1 mM dithiothreitol, 0.1 mM phenylmethylsulfonyl fluoride, 3 mM benzamidine, 16 μg/ml TPCK, 16 μg/ml TLCK). The sample (2 ml, 5 mg of protein) was applied to a column (0.5×0.5 cm) of hydroxylapatite (Calbiochem, HPLC grade) connected to a SMART micropurification chromatography system (Pharmacia) which had been equilibrated with buffer B. The flow rate was 100 μl/min. After application of the sample, the resin was washed with 0.5 ml of buffer B and eluted with a linear gradient (5 ml) of K2HPO4/H3PO4 pH 7.5, (0–500 mM) in buffer B. PLCβ2Δ eluted in a single peak at ∼110 mM K2HPO4/H3PO4.

Depletion of Rho GTPases from the PLCβ2‐stimulating factor with matrix‐immobilized GST–LyGDI

Two hundred μl of PLCβ2‐stimulating factor partially purified from bovine neutrophils by gel filtration [Illenberger et al. (1997) fraction 11 of the preparation shown in figure 2 therein] containing 1.2 mg of protein were supplemented with 200 μl of a 50% slurry of Glutathione Sepharose 4B loaded with GST–LyGDI or GST as described in the GST Gene Fusion System Manual (Pharmacia) and rotated end‐over‐end for 15 h at 4°C. The beads were sedimented by centrifugation at 600 g for 5 min at 4°C and washed three times with 1 ml each of ice‐cold buffer containing 20 mM Tris–HCl pH 7.5, 1 mM EDTA, 100 mM NaCl, 1 mM MgCl2 and 1 μM GDP. Proteins were eluted from the matrix by adding to the wet gel cake 400 μl of 10 mM reduced glutathione in 40 mM Tris–HCl pH 7.5, 1 mM EDTA, 1 mM MgCl2 and 1 μM GDP. The eluate was concentrated ∼6‐fold using Nanosep‐10 microconcentrators (Pall Filtron, Dreieich).

Solubilization of membrane‐bound recombinant Rho GTPases with soluble GST–LyGDI

Membranes of insect cells expressing Cdc42Hs, RhoA, Rac1 or β‐galactosidase (3.5 mg of protein each) were rotated end‐over‐end for 15 h at 4°C in 200 μl of 20 mM Tris–HCl pH 7.5, 1 mM EDTA, 1 mM dithiothreitol and 0.1 mM phenylmethylsulfonyl fluoride containing either 40 μM GST or 40 μM GST–LyGDI. The samples were centrifuged at 100 000 g for 10 min at 4°C and the supernatants used for further analysis.

Phospholipase C assay

Phospholipase C activity was determined as described previously (Illenberger et al., 1997). In brief, 10 μl of either native or recombinant heterodimeric LyGDI×Rho were supplemented with 5 μl of the soluble fraction of PLCβ2Δ‐baculovirus‐infected insect cells and incubated for 1 h at 25°C in a volume of 70 μl containing 28 μM 3H‐PtdInsP2 (5 Ci/mol), 280 μM phosphatidylethanolamine, 0.9 mM sodium deoxycholate, 50 mM Tris–maleate pH 7.4, 150 mM NaCl, 3 mM EGTA, 10 mM LiCl and 10 mM 2,3‐bisphosphoglycerate. CaCl2 was added to adjust the final concentration of free Ca2+ to 100 nM. The abilities of detergent‐solubilized recombinant Cdc42Hs and Rac1 as well as soluble insect cell‐expressed Cdc42Hs to stimulate PLCβ2Δ were determined according to Hepler et al. (1993), with some modifications. Ten microlitres of either crude or purified Cdc42Hs or Rac1 were supplemented with 10 μl of either the soluble fraction of PLCβ2Δ‐baculovirus‐infected insect cells or purified PLCβ2Δ, and incubated for 1 h at 25°C in a volume of 60 μl containing 33 μM 3H‐PtdInsP2 (5 Ci/mol), 50 mM HEPES–NaOH pH 7.2, 70 mM KCl, 3 mM EGTA, 2 mM dithiothreitol and 150 nM free Ca2+. The final concentration of free Mg2+ was ⩽35 μM. The samples were analysed for inositol phosphates as described previously (Gierschik and Camps, 1994).

35S‐GTP[S] binding

Binding of 35S‐GTP[S] was assayed as described previously (Illenberger et al., 1997) with minor modifications. In brief, samples containing 0.06–0.2 μg of protein were incubated for 60 min at 30°C in 40 μl of 25 mM HEPES–NaOH pH 8.0, 1 mM EDTA, 1 mM dithiothreitol, 100 mM NaCl, 20 mM MgCl2, 0.1% (v/v) GENAPOL® C‐100 and 0.5 nM 35S‐GTPγS (1250 Ci/mmol). Unlabelled GTP[S] was added at concentrations ranging from 10 pM to 10 μM. The incubation was terminated by rapid filtration through nitrocellulose membranes with a pore size of 0.45 μm, and three washes with 2 ml each of ice‐cold buffer containing 50 mM Tris–HCl pH 8.0, 100 mM NaCl and 5 mM MgCl2. The radioactivity retained on the membranes was determined by liquid‐scintillation counting. Non‐specific binding was defined as the binding not competed for by 1 mM unlabelled GTP[S]. Homologous competitive binding curves were constructed to estimate the binding parameters KD and Bmax using the GraphPad Prism Version 2.01 software package as described previously (Motulsky, 1996). The KD values determined for binding of GTP[S] to Cdc42Hswt, Cdc42HsF37A and Cdc42HsY40C extracted with cholate from membranes of baculovirus‐infected insect cells were 8.3 ± 0.6 nM, 8.3 ± 0.4 nM and 6.5 ± 0.4 nM (means ± SEM), respectively. The Bmax values of these preparations were 610, 220 and 190 pmol/mg protein, respectively. By comparison, the Bmax values obtained for GTP[S] binding to detergent extracts of membranes from β‐galactosidase‐baculovirus‐infected cells were <20 pmol/mg protein. In some experiments, the number of GTP[S] binding sites was determined by using 35S‐GTP[S] at a single concentration (100 nM) in the binding assay. The number thus obtained corresponds to 85–90% of Bmax.

Miscellaneous

The method used to purify βγt from bovine retinal rod outer segment membranes is described in Gierschik and Camps (1994). Protein concentrations were determined according to Bradford (1976) using bovine IgG as standard. SDS–PAGE and immunoblotting were performed according to standard protocols (Harlow and Lane, 1988), except that immunoreactive proteins were visualized using the ECL Western blotting detection system (Amersham). The method used to stain the gels with silver is specified in Oakley et al. (1980). Polyclonal antibodies reactive against RhoA (sc‐179), Cdc42Hs (sc‐87), Rac1 (sc‐217), LyGDI (sc‐359) and GST (sc‐459) were from Santa Cruz Biotechnology. A baculovirus encoding E.coli β‐galactosidase was from Dr M.Ruffing. All other materials were from standard vendors or from sources described previously (Illenberger et al., 1997).

Acknowledgements

The expert technical assistance of Norbert Zanker and Sandra Beier is greatly appreciated. We wish to thank Drs Rainer Müller, Martin Bähler, Martijn F.B.G.Gebbink, Wouter H.Moolenaar and Michael Ruffing for their generous gifts of cDNAs and other reagents used in this study. This work was supported by a grant from the Deutsche Forschungsgemeinschaft.

References

View Abstract