Rac/Cdc42 and p65PAK regulate the microtubule-destabilizing protein stathmin through phosphorylation at serine 16.

We have identified a rapid protein phosphorylation event at residue serine 16 of stathmin using two-dimensional gel electrophoresis coupled to matrix-assisted laser desorption/ionization mass spectrometry in combination with post-source decay analysis, which is induced by the epidermal growth factor receptor. Phosphorylation is specifically mediated by the small GTPases Rac and Cdc42 and their common downstream target, the serine/threonine kinase p65PAK. Both GTPases have previously been shown to regulate the dynamics of actin polymerization. Because stathmin destabilizes microtubules, and this process is inhibited by phosphorylation at residue 16, Rac and Cdc42 can potentially regulate both F-actin and microtubule dynamics.

Members of the Rho GTPase family, Rho, Rac, and Cdc42, control the assembly of filamentous actin structures in all mammalian cells (1). Their ability to link extracellular signals to the reorganization of the actin cytoskeleton suggests that they are likely to be important regulators of actin-driven cell processes, and Rac, for example, is crucial for growth cone guidance and cell migration both in tissue culture cells and in vivo in Drosophila and Caenorhabditis elegans (2)(3)(4). We report here that activation of Rac and to a lesser extent Cdc42 by EGF 1 leads to the phosphorylation of stathmin at residue 16. Phosphorylation at this site has been shown to inhibit stathmin-induced destabilization of microtubules, and our results suggest, therefore, that Rac and Cdc42 can regulate the dynamics of both the actin and the microtubule cytoskeletons (5,6).
Cell Culture and Transfections-HEp-2 cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum. 1 ϫ 10 5 cells per well were seeded into 12-well dishes and serum-starved 24 h later. For preparative purposes, cells were seeded at 1 ϫ 10 6 cells per 10-cm dish and were serum-starved 2 days later for another 20 h prior to lysis. For [ 32 P] or [ 33 P]orthophosphate labeling, cells were incubated in phosphate-free medium in the presence of 12.5 Ci/ml 32 P i for preparative or 25 Ci/ml 32 P i or 100 Ci/ml 33 P i for analytical purposes for 3 h prior to lysis. For transfection experiments, HEp-2 cells were seeded at 2 ϫ 10 5 per well into 6-well dishes 20 h before transfection. Cells were incubated for 4 h in 1.0 ml of serum-free medium containing 6 l of LipofectAMINE (Life Technologies, Inc.) and 1.5 to 1.7 g of total DNA per well. The transfection mixture was supplemented with 1 ml of medium containing 20% fetal bovine serum, and cells were lysed 20 h later.
Protein Purification-Fourteen 10-cm dishes of serum-starved and 32 P i -labeled HEp-2 cells were stimulated with 100 ng/ml EGF for 7 min and lysed (350 l per dish). Lysates were precleared, supplemented with 80 l of 20% SDS and 120 l of 87% glycerol per ml, and the sample containing about 30 mg of total protein was loaded on the Model 491 Prep Cell (Bio-Rad). Preparative gel electrophoresis was performed utilizing a 11% separating gel. Aliquots containing proteins in the 19,000-to 23,000-dalton range were precipitated and analyzed by twodimensional gel electrophoresis and autoradiography of silver-stained gels (10). The 21,000-dalton phosphoprotein of interest, which was found in two fractions, was dried to one-tenth of its original volume, precipitated, and subjected to preparative two-dimensional gel electrophoresis.
MALDI-MS Analysis-The protein spot of interest was visualized by Coomassie staining and excised, and the gel fragment was washed in water. The gel piece was allowed to shrink in 100 l of acetonitrile/ water (1:1) (both Baker HPLC Analyzed, Mallinckrodt Baker B.V., § § To whom correspondence should be addressed. E-mail: alan. hall@ucl.ac.uk. 1 The abbreviations used are: EGF, epidermal growth factor; MALDI-TOF-MS, matrix-assisted laser desorption/ionization time-of-flight mass spectrometry; PSD, post-source decay analysis; MAP, mitogenactivated protein; RP-HPLC, reversed phase-high performance liquid chromatography; ERK, extracellular receptor kinase. Deventer, The Netherlands) for 30 min and dried in a centrifugal vacuum concentrator. 10 l of 50 mM ammonium bicarbonate buffer containing 0.05 g of sequencing grade modified trypsin (Promega) was added followed by ammonium bicarbonate buffer to submerge the gel piece. Digestion proceeded overnight at 37°C and was stopped by FIG. 2. Effects of GTPase-modifying toxins on EGF-stimulated stathmin phosphorylation. A, serum-starved HEp-2 cells were incubated with toxin B1470 for 3 h and stimulated with 100 ng/ml EGF for 7 min. Total cell lysates were analyzed by immunoblotting with antisera specific for stathmin phosphorylated at serines 16, 25, or 38 or with an anti-phospho-ERK antibody. B, serum-starved HEp-2 cells were treated with 2 g/ml CNF-1 for the indicated times, and stathmin phosphorylation was assessed as described above. Fig. 1. EGF-induced and GTPase-dependent phosphorylation of stathmin. A, HEp-2 cells were serum-starved for 18 h and labeled with 100 Ci/ml 33 P i for 3 h. Cells were pretreated with 160 ng/ml toxin B1470 or 2 g/ml toxin B10463 for 3 h prior to stimulation with 100 ng/ml EGF for 5 min. After cell lysis and two-dimensional gel electrophoresis, phosphoproteins were visualized by autoradiography. B, HEp-2 cells were serum-starved for 18 h and labeled with 25 Ci/ml 32 P i for 3 h before EGF stimulation and cell lysis. After two-dimensional gel electrophoresis and transfer to nitrocellulose membrane, phosphoproteins were visualized by autoradiography (upper panels) and protein by anti-stathmin antiserum C (lower panels). A and B, the position of the phosphoprotein purified for MALDI-MS analysis is indicated by arrows. Rho GTPase Regulation of Microtubule Dynamics 1678 acidification. The supernatant was removed from the gel pieces and a fraction (10%) was purified/concentrated on added Poros® 50 R2 beads (Roche Molecular Biochemicals GmbH, Mannheim, Germany) and used for direct MALDI-MS peptide mass fingerprint analysis (11). The remainder of the digestion mixture was separated on a RP-HPLC column; eluting peptides were automatically collected on Poros R2 beads and used for MALDI-PSD measurements (12). Measurements were performed on a Bruker Reflex III MALDI-TOF-MS (Bruker Daltonik GmbH, Bremen, Germany) operating in the reflectron mode. RP-HPLC fractions were first scanned in the reflectron mode, and candidate peptides were selected and their PSD spectra recorded (12). The information present in the PSD spectra was used by the SEQUEST algorithm (13) to identify the protein in a public nonredundant protein database.

RESULTS AND DISCUSSION
Rac interacts with numerous cellular targets including at least three families of Ser/Thr kinases, p65PAK, MLK, and p70S6kinase. To identify phosphorylation events controlled by Rac, metabolically labeled HEp-2 cells were pretreated with either clostridial toxin B1470, which inactivates Rac, R-ras, Rap1, Ral and, to a lesser extent, Cdc42, or with toxin B10463, which inactivates Rho, Rac, and Cdc42 (14,15). Cells were then stimulated with EGF, a known activator of Rac, for 7 min and cell lysates were analyzed by two-dimensional gel electrophoresis. The increase in intensity of only 2 of about 20 EGF-stimulated phosphoproteins was blocked by both toxins (Fig. 1A). One was isolated from a preparative two-dimensional gel and subjected to MALDI-MS post-source decay analysis. The excised protein was identified as stathmin.
As shown in Fig. 1B, EGF treatment resulted in the appearance of two new forms of stathmin and a significant increase in the intensity of at least one other form. Stathmin has been shown to be phosphorylated on four distinct serines: 16, 25, 38, and 63, resulting in a variety of migration patterns on twodimensional gels (Fig. 1C and Refs. 7 and 16). To identify which residues are phosphorylated in response to EGF, antibodies specific for three phosphorylated forms (serines 16, 25, and 38) were used on Western blots (7). As seen in Fig. 1D, EGF induces a dramatic increase in phosphorylation at Ser-16 and Ser-25. Ser-38 is phosphorylated in control cells and does not increase significantly upon EGF addition. The lack of a significant change in the P3 isoform at 19 kDa (Refs. 25, 38, 63 and Fig. 1D) coupled with the strong increase in the P3 isoform at 23 kDa (16,25,38) suggests that very little phosphorylation is induced at Ser-63 by EGF.
To identify which phosphorylation event is mediated by Rho GTPases, phosphospecific antibodies were used to analyze Western blots of one-dimensional gels of lysates from cells treated with the two clostridial toxins. As can be seen in Fig.  2A, 160 ng/ml toxin B1470 results in a complete inhibition of EGF-induced phosphorylation of Ser-16 although at even 10fold higher concentrations, the toxin does not inhibit EGFinduced phosphorylation at Ser-25. Ser-25 has been reported to be phosphorylated by ERK MAP kinases and the EGF-stimu- Rho GTPase Regulation of Microtubule Dynamics 1679 lated activation of ERK1 and ERK2 was not changed by toxin treatment (Fig. 2A, bottom left  To confirm that Rac and Cdc42 can induce phosphorylation at Ser-16, HEp-2 cells were transfected with a Myc-tagged stathmin expression vector either alone or with dominantnegative Rac (N17Rac1) or Cdc42 (N17Cdc42) and treated 24 h later with EGF. Inhibition of Rac almost completely prevented phosphorylation on Ser-16, whereas inhibition of Cdc42 inhibited phosphorylation by around 50% (Fig. 3A). Next, stathmin was cotransfected with constitutively activated GTPases. L61Rac and L61Cdc42, but not L63Rho, induced phosphorylation specifically at Ser-16 (Fig. 3B).
In conclusion, we have show that the addition of EGF to HEp-2 cells leads to a Rac/Cdc42 and p65PAK-dependent phosphorylation of stathmin at Ser-16. Whether p65PAK directly phosphorylates stathmin or whether it activates another downstream kinase is not currently known. Two kinases have been reported to phosphorylate stathmin at Ser-16, the cAMP-dependent kinase A and the Ca 2ϩ /calmodulin-dependent kinase isoforms, CaMK IV/Gr and CaMKII (5,16,20). The cAMP-dependent kinase A is unlikely to be involved because it preferentially phosphorylates Ser-63, and pretreatment of HEp-2 cells with EGTA to block Ca 2ϩ influx had no effect, suggesting that CaM kinases are not involved (data not shown).
Stathmin plays an important role in controlling microtubule dynamic either by sequestering ␣/␤ tubulin heterodimers or by increasing catastrophe frequency at the plus ends of microtubules or both (21)(22)(23). As a result, stathmin causes destabilization of growing microtubules and phosphorylation at Ser-16 appears to block this activity (5,6). Rac and Cdc42 regulate actin polymerization and form membrane protrusions at the leading edge of migrating cells and neuronal growth cones. The results described here suggest, therefore, that Rac and Cdc42 might control both F-actin and microtubule dynamics in localized regions associated with cell protrusions (24,25).