The interaction of RGSZ1 with SCG10 attenuates the ability of SCG10 to promote microtubule disassembly.

RGS proteins (regulators of G protein signaling) are a diverse family of proteins that act to negatively regulate signaling by heterotrimeric G proteins. Initially characterized as GTPase-activating proteins for Galpha subunits, recent data have implied additional functions for RGS proteins. We previously identified an RGS protein (termed RGSZ1) whose expression is quite specific to neuronal tissue (Glick, J. L., Meigs, T. E., Miron, A., and Casey, P. J. (1998) J. Biol. Chem. 273, 26008-26013). In a continuing effort to understand the role of RGSZ1 in cellular signaling, the yeast two-hybrid system was employed to identify potential effector proteins of RGSZ1. The microtubule-destabilizing protein SCG10 (superior cervical ganglia, neural specific 10) was found to directly interact with RGSZ1 in the yeast system, and this interaction was further verified using direct binding assays. Treatment of PC12 cells with nerve growth factor resulted in Golgi-specific distribution of SCG10. A green fluorescent protein-tagged variant of RGSZ1 translocated to the Golgi complex upon treatment of PC12 cells with nerve growth factor, providing evidence that RGSZ1 and SCG10 interact in cells as well as in vitro. Analysis of in vitro microtubule polymerization/depolymerization showed that binding of RGSZ1 to SCG10 effectively blocked the ability of SCG10 to induce microtubule disassembly as determined by both turbidimetric and microscopy-based assays. These results identify a novel connection between RGS proteins and the cytoskeletal network that points to a broader role than previously envisioned for RGS proteins in regulating biological processes.

RGS proteins (regulators of G protein signaling) are a diverse family of proteins that act to negatively regulate signaling by heterotrimeric G proteins. Initially characterized as GTPase-activating proteins for G␣ subunits, recent data have implied additional functions for RGS proteins. We previously identified an RGS protein (termed RGSZ1) whose expression is quite specific to neuronal tissue (Glick, J. L., Meigs, T. E., Miron, A., and Casey, P. J. (1998) J. Biol. Chem. 273, 26008 -26013). In a continuing effort to understand the role of RGSZ1 in cellular signaling, the yeast two-hybrid system was employed to identify potential effector proteins of RGSZ1. The microtubule-destabilizing protein SCG10 (superior cervical ganglia, neural specific 10) was found to directly interact with RGSZ1 in the yeast system, and this interaction was further verified using direct binding assays. Treatment of PC12 cells with nerve growth factor resulted in Golgi-specific distribution of SCG10. A green fluorescent protein-tagged variant of RGSZ1 translocated to the Golgi complex upon treatment of PC12 cells with nerve growth factor, providing evidence that RGSZ1 and SCG10 interact in cells as well as in vitro. Analysis of in vitro microtubule polymerization/ depolymerization showed that binding of RGSZ1 to SCG10 effectively blocked the ability of SCG10 to induce microtubule disassembly as determined by both turbidimetric and microscopy-based assays. These results identify a novel connection between RGS proteins and the cytoskeletal network that points to a broader role than previously envisioned for RGS proteins in regulating biological processes.
The mechanisms by which G proteins transduce signals from the external milieu to internal targets are complex, and multiple control points exist where this transduction process may be regulated (1,2). In resting cells, G proteins associate with specific receptors on the intracellular face of the plasma membrane. In this state, the intact heterotrimer exists as a membrane-tethered inactive complex of ␣ and ␤␥ subunits. Upon binding ligand, receptors induce ␣ subunits of associated G proteins to release GDP and to bind GTP. The GTP-bound G␣ subunit is then released from the ␤␥ complex and is competent to act on targeted effector molecule(s). At the same time, however, an intrinsic GTPase activity associated with the G␣ subunit catalyzes the hydrolysis of GTP to GDP and subsequent reassociation of G␣ with the ␤␥ complex. This rate of GTP hydrolysis is crucial to the persistence of G protein responses; hence, this reaction can be considered a molecular timer, controlling the duration of both ␣ and ␤␥ signaling.
Initially, the regulation of GTP hydrolysis was thought to be independent of external proteins; rather, it was the intrinsic hydrolytic rate of GTP by the G␣ subunit that controlled the overall duration of a response. However, a variety of studies demonstrated that the rates of in vitro GTP hydrolysis appeared to be much slower than recovery rates in specific systems (1,2). Such was the case for visual signal transduction, as inconsistencies were clearly exhibited between the rates of GTP hydrolysis and the deactivation of light responses (3). It was these differences that led to the search for proteins that would accelerate the hydrolysis of G␣-bound GTP. Indeed, such proteins have been recently identified and shown to play major roles in G protein-mediated GTP hydrolysis (4 -6). RGS 1 proteins (regulators of G protein signaling) bind to activated G␣ subunits and act as GTPase-activating proteins (GAPs) by hastening the hydrolysis of GTP to GDP. Since the initial discovery of an RGS protein in Saccharomyces cerevisiae (7), Ͼ20 unique mammalian RGS proteins have been identified (1,6,8). All identified RGS proteins contain a conserved region of ϳ120 residues referred to as the RGS domain. Although RGS proteins all contain this core domain, the family members differ greatly in size, sequence, cellular localization, and tissue distribution (1,6).
The hypothesis that RGS proteins can bind cellular proteins in addition to G␣ subunits has been widely investigated. Recent findings indicate that some RGS proteins do in fact contain a variety of motifs that are important for protein-protein interactions as well as membrane attachment. For example, the R7 family of RGS proteins (consisting of RGS6, RGS7, RGS9, and RGS11) contains a DEP (dishevelled/Egl-10/pleckstrin) domain and a novel GGL (G protein ␥ subunit-like) domain. The DEP domain is found in a variety of proteins and has been hypothesized to play a role in targeting DEP domaincontaining proteins to G protein-coupled receptors (9). GGL domains have been shown to be responsible for selective binding to G␤ 5 , a G␤ isoform quite divergent from G␤ [1][2][3][4] . RGS7 and RGS11 have both been shown to bind G␤ 5 in vitro (10,11), and further studies have indicated that other members of the R7 family copurify with G␤ 5 from both brain (12) and retina (10). GAIP and RGS12 contain PDZ domain-binding motifs (13), whereas RGS12 and RGS14 contain both a Rap-binding motif and a Loco homology domain (14,15), each possibly contributing to novel protein-protein interactions involving RGS proteins. Additionally, the RZ family members (RGSZ1, RGSZ2, GAIP, and RET-RGS1) contain a cysteine string that is thought to play a role in membrane attachment (16 -18). Taken together, these data suggest that portions outside the RGS domain play key roles in determining G␣ specificity, membrane attachment, cellular localization, and other important functions.
In this study, we employed the yeast two-hybrid system to identify a microtubule-destabilizing protein, termed SCG10 (superior cervical ganglia, neural specific 10), as a novel binding partner for RGSZ1. SCG10 is a member of the stathmin family of proteins (19,20), all of which bind to tubulin to act as sequestering agents as well as to promote microtubule catastrophes (21-24). SCG10 is different from stathmin in two regards: its expression is limited to neuronal tissue, and it also possesses a unique N-terminal domain that is critical for membrane binding (20,25). The interaction between RGSZ1 and SCG10 promoted overall microtubule stability, as RGSZ1 blocked the ability of SCG10 to induce microtubule depolymerization in vitro. These findings provide a unique link between heterotrimeric G protein signaling and the cytoskeleton that may play significant roles in neuronal development.

EXPERIMENTAL PROCEDURES
Miscellaneous Materials and Methods-Rabbit polyclonal anti-SCG10 and mouse monoclonal anti-SCG10 antibodies were generated as previously described (26,27). Rabbit polyclonal anti-RGSZ1 antibody was a generous gift of Elliott Ross (University of Texas Southwestern Medical Center, Dallas, TX). RGS10 protein was provided by Patrick Burgon (Harvard Medical Center, Boston, MA). Rabbit polyclonal antihexahistidine antibody was purchased from QIAGEN Inc. (Valencia, CA). Protein concentrations were determined according to the method of Bradford (28) using bovine serum albumin as a standard.
Yeast Two-hybrid Screen-The cDNA encoding the C-terminal RGS domain of RGSZ1 (denoted ⌬RGSZ1, comprising residues 87-217 of the RGSZ1 open reading frame) was obtained by PCR amplification of the region from full-length RGSZ1 in the pCMV sport vector (17) with the following primers: 5Ј-CCGCATGCCATGGAGATGGGATCAGAGCGG-ATGGAG-3Ј (forward) and 5Ј-GGCGGATCCTCATGCTTCAATAGATT-TCTCGGA-3Ј (reverse). PCRs were performed under standard conditions using Vent ® polymerase (New England Biolabs Inc., Beverly, MA). The resulting PCR fragment was directionally ligated into the "bait" vector pGBKT7 (CLONTECH, Palo Alto, CA) using the 5Ј-NcoI and 3Ј-BamHI sites of the vector, producing the construct pGBKT7-⌬RG-SZ1. The frame of the insert as well as the ligation junctions was confirmed by sequence analysis (Duke University Sequencing Facility).
The two-hybrid screen was conducted using the Matchmaker 3 System ® (CLONTECH) according to the manufacturer's recommendations. Briefly, pGBKT7-⌬RGSZ1 was transformed into yeast strain PJ69 -2a, which was then mated with a pre-transformed human fetal brain library. Clones that survived the quadruple dropout selection (ϪHis, ϪAde, ϪLeu, ϪTrp) and that stained positive for ␤-galactosidase activity were isolated. The plasmids recovered from each yeast clone were introduced into JM109 cells by electroporation, and the subsequent cDNA isolated from the JM109 cells was sequenced.
Protein Expression and Purification-Tubulin was isolated from porcine brain by two cycles of polymerization and depolymerization as described previously (29). A portion of the MAP-rich tubulin obtained from this preparation was subsequently passed over a phosphocellulose column (Amersham Biosciences); the purity of the tubulin obtained from this procedure was Ͼ99%. Both MAP-rich and purified tubulin preparations were stored in 80 mM PIPES (pH 6.9), 0.5 mM EGTA, and 0.5 mM MgCl 2 at Ϫ80°C until used.
Full-length SCG10 fused to GST was insoluble as produced in bacteria. To purify GST-SCG10, the 20,000 ϫ g pellet obtained after centrifugation of the lysate was resuspended in 8 M urea, incubated for 2 h at 4°C, and subsequently centrifuged at 30,000 ϫ g for 30 min at 4°C to obtain solubilized protein. The protein contained within the supernatant was then renatured by dialysis in Buffer A for 12 h at 4°C. The precipitated protein was removed by centrifugation at 20,000 ϫ g for 30 min at 4°C, and GST-SCG10 was purified using glutathione-Sepharose beads as described above.
RGSZ1 and an N-terminal truncation mutant (denoted ⌬RGSZ1, comprising residues 87-217) were purified as His-tagged proteins as previously described (17). The cDNA encoding GAIP was obtained from the Guthrie cDNA Resource Center (Sayre, PA) and subcloned into pRSET-B (Invitrogen). The corresponding protein was subsequently purified as a His-tagged protein following a protocol similar to that described for the purification of RGSZ1. Purified RGS proteins were analyzed for GAP activity for G␣ z (RGSZ1) and G␣ o (GAIP and RGS10) and in all cases were found to be functionally active (data not shown).
Protein-Protein Interaction Studies-For pull-down experiments using purified proteins, His-tagged RGS proteins (5 g) were incubated with GST or GST-tagged proteins (GST-⌬SCG10 or GST-SCG10) each at 5 g in 50 mM Tris-HCl (pH 7.7), 1 mM EDTA, 1 mM DTT, 150 mM NaCl, 10 mM imidazole, and 0.1% Lubrol for 4 h at 4°C with shaking. Glutathione-Sepharose beads were added, and the suspension was allowed to incubate for an additional 1 h. Bound proteins were precipitated by centrifugation and washed three times with the same buffer. Samples were subjected to SDS-PAGE on 12% polyacrylamide gels and transferred to nitrocellulose, and proteins were visualized by immunoblot analysis using alkaline phosphatase-based detection methods (Promega, Madison, WI).
Turbidimetric Evaluation of Microtubule Assembly and Disassembly-The assembly/disassembly of tubulin was measured using a light scattering assay as previously described (31). For the microtubule assembly assays, MAP-rich tubulin (3.25 mg/ml final concentration) in PEM buffer containing 1 mM GTP at 4°C was mixed with SCG10, RGSZ1, or a combination of these proteins. Microtubule assembly was initiated by raising the temperature to 37°C, and the absorbance at 350 nm was monitored over 15 min in a Hewlett-Packard Model 8453 spectrophotometer. For the microtubule disassembly experiments, microtubules were prepared as described above, and the temperature was maintained at 37°C. SCG10, RGSZ1, or a combination of the two proteins was added, and the absorbance at 350 nm was monitored over 10 min. In assembly/disassembly experiments, in which the effects of SCG10 and RGSZ1 together were tested, the two proteins were preincubated at 4°C for 1 h prior to use.
Fluorescence Analysis of Rhodamine-labeled Tubulin Disassembly-MAP-free tubulin and rhodamine-labeled tubulin (Cytoskeleton Inc., Denver, CO) were premixed at a ratio of 4:1 (unlabeled/labeled) to a final concentration of 3.25 mg/ml in PEM buffer containing 2 mM GTP at 37°C for 30 min to form microtubules. GST-SCG10 (5 M) alone, GST-SCG10 (5 M) and RGSZ1 (10 M), or buffer was then added to the mixture and incubated at 37°C for 10 min. Reactions (10 l) were diluted in 50 l of 60% glycerol and gently mixed. A 4-l aliquot was then placed on a glass slide with a coverslip, and microtubule patterns were observed by fluorescence microscopy using a Nikon Eclipse TE300 inverted microscope.
Cell Culture and Immunofluorescence Analysis-PC12 cells were cultured as previously described (32). Transient transfections of PC12 cells were performed using LipofectAMINE ® reagent (Invitrogen) according to the manufacturer's recommendations. Cells destined for transfection with GFP constructs were plated on chambered coverslips coated with poly-D-lysine. Cells were treated with nerve growth factor (NGF) (Promega) for the indicated times and placed in phosphatebuffered saline, and live cells were viewed using the Nikon Eclipse TE300 inverted microscope. Cells destined for immunofluorescence analyses were plated on poly-D-lysine-coated glass coverslips and treated with NGF as described above prior to processing.
For immunofluorescence analysis, cells were fixed in 4% paraformaldehyde in phosphate-buffered saline for 20 min and rinsed two times with phosphate-buffered saline. Cells were blocked in phosphate-buffered saline containing 10% goat serum, 2% bovine serum albumin, and 0.3% Tween 20 for 30 min and subsequently incubated with anti-SCG10 primary antibodies overnight at 4°C in the same buffer. After three rinses with the same buffer, the cells were incubated with Cy3-conjugated goat anti-rabbit secondary antibodies (Jackson ImmunoResearch Laboratories, Inc., West Grove, PA) for 1 h at 25°C. Cells were mounted using Prolong ® antifade mounting medium (Molecular Probes, Inc., Eugene, OR) and viewed using the Nikon Eclipse TE300 inverted microscope.

RESULTS
Yeast Two-hybrid Screen of RGSZ1-In our ongoing efforts to elucidate the roles of RGSZ1 in cellular signaling, a yeast two-hybrid screen was performed. We initially attempted to conduct screens using full-length RGSZ1 as the bait, but it alone acted as a transcriptional activator in our system, rendering it unsuitable for screening purposes (data not shown). However, we were successful using the C-terminal 130 residues of RGSZ1 (⌬RGSZ1) as bait. This truncation mutant lacked the cysteine string and contained only the RGS domain of RGSZ1. After construction of the bait plasmid, yeast strain PJ69 -2a was transformed and evaluated for bait protein expression, toxic effects of the protein, and transcriptional activation in the absence of a binding partner. These initial control studies suggested that ⌬RGSZ1 would be an effective bait molecule in the screen (data not shown). The actual screen employed a pretransformed cDNA library derived from fetal human brain, as RGSZ1 is predominantly found in brain (17,18). The library contained a total of 3.5 ϫ 10 6 clones, and the screen yielded ϳ5.8 ϫ 10 6 total transformants, representing a 1.7-fold overscreen of the library.
The yeast two-hybrid screen yielded a total of 415 positives, the inserts of which were isolated and sequenced. Evaluation of the cDNA sequences revealed that 47 inserts encoded the same molecule, termed SCG10. This membrane-associated, developmentally regulated protein is a member of the stathmin family of proteins, all of which act to destabilize microtubules. Neither null bait (empty vector) nor false bait (the N-terminal region of protein farnesyltransferase) demonstrated any binding to SCG10 in counter-screening experiments using the yeast twohybrid system, verifying the specificity of the RGSZ1-SCG10 interaction (data not shown).
Biochemical Analysis of the RGSZ1-SCG10 Interaction-Although the yeast two-hybrid system clearly indicated that RGSZ1 and SCG10 could interact in a cellular context, it was necessary to confirm this interaction by direct biochemical studies. Native SCG10 is a difficult protein to purify and even more difficult to obtain in a concentrated form (33). We constructed N-terminal GST-tagged forms of both SCG10 and ⌬SCG10 to facilitate the purification of SCG10. The N-terminal region of SCG10 (amino acids 1-34) is thought to be important for membrane localization, and removal of this region results in a cytosolic protein (⌬SCG10) that is closely homologous to conventional stathmin (27). Although bacterially produced GST-⌬SCG10 was indeed a cytosolic protein that was easily purified by affinity chromatography, GST-SCG10 was recovered in the insoluble pellet. After denaturation with urea and renaturation by dialysis, GST-SCG10 could be purified by affinity chromatography (see "Experimental Procedures"). Purifying SCG10 in this manner allowed us to obtain a concentration of 2.5 mg/ml, which is Ͼ10-fold more concentrated than previously reported (33).
Binding of GST-SCG10 to RGSZ1 was examined using glutathione-Sepharose pull-down experiments. As shown in Fig.  1A, GST-SCG10 readily bound RGSZ1 and, to a much lesser extent, the N-terminal truncation mutant ⌬RGSZ1. The amount of RGSZ1 bound to SCG10 represents ϳ10% of the total protein input. Although GST-⌬SCG10 did not interact with RGSZ1, it did appear to weakly interact with ⌬RGSZ1. Control experiments showed that GST alone did not bind RGSZ1, but did appear to weakly bind ⌬RGSZ1, suggesting that the interaction between ⌬SCG10 and ⌬RGSZ1 is largely nonspecific. Taken together, these data indicate that the Nterminal domain of SCG10 is important for the interaction with RGSZ1.
To examine specificity of the RGSZ1-SCG10 interaction, RGS10 (a member outside the RZ family of RGS proteins) and GAIP (a member within the RZ family) were tested for their abilities to bind SCG10. Pull-down experiments similar to those used to show RGSZ1 binding revealed that GAIP also bound to SCG10, whereas RGS10 did not (Fig. 1B). Both   FIG. 1. Binding of RGS proteins to SCG10 in vitro. A, binding of RGSZ1. Purified GST, GST-⌬SCG10, or GST-SCG10 (5 g) was incubated with either RGSZ1 or ⌬RGSZ1 (5 g) for 4 h at 4°C. Glutathione-Sepharose beads (20 l) were added, and reactions were incubated for an additional 1 h. Beads were processed, and bound proteins were analyzed for RGSZ1 by the procedure described under "Experimental Procedures." The data presented are from a single experiment and are representative of five separate experiments. B, comparison of RGSZ1, RGS10, and GAIP binding. GST, GST-⌬SCG10, or GST-SCG10 (5 g) was incubated with RGSZ1, GAIP, or RGS10 (each at 5 g) for 4 h at 4°C. Glutathione-Sepharose beads (20 l) were added, and reactions were incubated for an additional 1 h. Beads were processed, and bound proteins were analyzed by the immunoblot procedures described under "Experimental Procedures" using an antibody that recognizes the hexahistidine sequence inserted into all RGS constructs. The data presented are from a single experiment and are representative of five separate experiments.
RGSZ1 and GAIP contain a region rich in cysteines referred to as a cysteine string. This region serves as a site for palmitoylation, a post-translational modification involved in membrane anchoring, trafficking, and protein-protein interactions (16). Although the analysis of RGS proteins was not totally inclusive, it appears that SCG10 may specifically interact with the RZ family of RGS proteins.
We also examined whether the binding of SCG10 affects the ability of RGSZ1 to accelerate GTP hydrolysis of G␣ z . To date, the only known biological function of RGSZ1 is to act as a GAP for G␣ z , so it seemed plausible that the interaction with SCG10 may have some effect on the ability of RGSZ1 to act as a GAP. However, even using concentrations of SCG10 that were 10fold greater than those of RGSZ1, no effect on the GAP activity of RGSZ1 was observed (data not shown). This result suggests that the binding of these two proteins occurs in such a way that the critical residues responsible for the ability of RGSZ1 to act as a GAP are unaffected, thus allowing unperturbed interaction with the G␣ protein.
Localization of GFP-RGSZ1 and SCG10 in PC12 Cells-As noted in the Introduction, expression of SCG10, like that of RGSZ1, is predominantly confined to neuronal tissue. Expression of SCG10 has also been detected in certain neuron-like cell lines such as PC12 cells. In addition, it has been shown that treatment of PC12 cells with NGF induces the expression of SCG10 (20). Using antibodies directed against SCG10 in immunoblot and immunofluorescence protocols, we confirmed that expression of the protein is indeed up-regulated by NGF treatment (Fig. 2, A-C). Consistent with previous reports, endogenous SCG10 was localized to the Golgi region of the cells as well as to the growth cones of the extending neurites (Fig. 2, B and C). Taking advantage of these findings, we transfected PC12 cells with either GFP-RGSZ1 or GFP-⌬RGSZ1 and examined their localization before and after NGF stimulation. Although RGSZ1 has been reported to localize to the Golgi complex in certain cells (18), expression of the GFP-RGSZ1 fusion protein in PC12 cells resulted in a predominantly diffuse cytosolic expression pattern. GFP-⌬RGSZ1, which lacks the cysteine string motif thought to direct the membrane localiza-tion of RGSZ1, also exhibited cytosolic localization in untreated PC12 cells. However, treatment of PC12 cells with NGF resulted in a striking redistribution of GFP-RGSZ1 to a punctate staining pattern that closely resembles Golgi staining patterns in PC12 cells (Fig. 2G). It must be noted that the transfection efficiency of PC12 cells was extremely low, normally achieving only 3-5% of the total cell population. However, it was observed that in all experiments performed, 50 -75% of the cells that expressed GFP-RGSZ1 exhibited the redistribution pattern presented in Fig. 2G. Importantly, although GFP-RGSZ1 translocated in response to NGF treatment, the localization of GFP-⌬RGSZ1, which interacted only poorly with SCG10, was unaffected by NGF treatment (Fig. 2, D and E). Taken together, these results suggest that production of Golgi-associated SCG10 causes GFP-RGSZ1 to translocate from the cytosol to the Golgi because of an interaction between the two proteins.
Influence of RGSZ1 Binding on the Ability of SCG10 to Modulate Microtubule Assembly and Disassembly-A well characterized activity of SCG10 is its ability to prevent microtubule assembly (34). To evaluate the effect of RGSZ1 binding on this activity of SCG10, an in vitro assay of microtubule assembly was employed. Incubation of soluble tubulin at 37°C in the presence of GTP results in its polymerization into microtubules, and this process can be followed using a turbidimetric assay (31). Consistent with previous results (34), addition of SCG10 resulted in a dose-dependent decrease in tubulin polymerization under the assay conditions (Fig. 3A). To evaluate the effect of RGSZ1 in this system, an intermediate concentration of SCG10 (5 M) was chosen, and the assays were repeated under varying concentrations of added RGSZ1. However, even at concentrations as high as 40 M, RGSZ1 had no discernible effect on the ability of SCG10 to attenuate microtubule assembly (Fig. 3B).
In addition to its ability to prevent microtubule assembly, SCG10 is also capable of initiating microtubule disassembly both in vitro (34) and in intact cells (26). Although the precise mechanism by which SCG10 triggers microtubule disassembly is not clear, this process can also be monitored using the turbidimetric assay described above. Under conditions in which tubulin was initially polymerized and subsequently treated with SCG10, it was observed that SCG10 promoted the disassembly of assembled microtubules in a concentration-dependent manner (Fig. 4A). Strikingly, addition of RGSZ1 blocked SCG10-induced microtubule disassembly in a dose-dependent manner (Fig. 4B). In fact, microtubule disassembly induced by 5 M SCG10 was almost totally ablated upon addition of 10 M RGSZ1. Similar experiments were also conducted using ⌬SCG10. Although ⌬SCG10 was also capable of initiating microtubule disassembly, RGSZ1 had no effect on this response (data not shown). This finding corresponds well to the in vitro binding data indicating that RGSZ1 interacts only with fulllength SCG10.
In addition to the rather indirect method of monitoring microtubule formation using light scattering, methods have been developed to visualize this process directly using fluorescence microscopy. Using rhodamine-labeled tubulin in in vitro assembly systems, microtubule formation can be viewed directly by fluorescence microscopy (21). Incubation of rhodamine-labeled tubulin with GTP at 37°C induced the assembly of microtu-bules (Fig. 5A), whereas incubation in the absence of GTP at 4°C prevented microtubule assembly, as indicated by the diffuse staining pattern (Fig. 5D). Subsequent addition of 5 M SCG10 resulted in almost complete loss of microtubule structure, which coincided with an increase in the more homogeneous fluorescence pattern of soluble tubulin (Fig. 5B). However, preincubation of SCG10 with a 2-fold excess of RGSZ1 markedly attenuated the ability of SCG10 to induce microtubule disassembly (Fig. 5C). Together with the data obtained from the turbidimetric studies, these data confirm that the binding of RGSZ1 blocks the ability of SCG10 to induce microtubule disassembly. It is interesting to note that there appears to be an increased number of high density microtubule areas upon treatment with both SCG10 and RGSZ1. At present, it is unclear why this occurs, but it may be that the two proteins act together as a nucleating point for microtubule formation. DISCUSSION Our search for novel interactors of RGSZ1 led to the identification of SCG10 as directly interacting with RGSZ1. Although recent evidence points to RGS proteins as playing roles in addition to serving as GAPs for heterotrimeric G proteins (for review, see Refs. 4 and 6), this seems to be the first indication that an RGS protein may directly regulate cytoskeletal organization.
SCG10 binds quite selectively to full-length RGSZ1 compared with the truncated form ⌬RGSZ1, which lacks the N terminus of the protein. This is somewhat surprising because the two-hybrid screen was conducted using only the core domain (⌬RGSZ1) as bait. However, the ability of this system to detect very weak interactions provides a possible explanation as to why this was observed (35,36). The N terminus of SCG10 was also found to be essential for RGSZ1 binding. This region contains two cysteine residues (positions 22 and 24) that are key for directing SCG10 to Golgi membranes (27,37). Deletion of this region totally ablated the ability of RGSZ1 to interact with SCG10, suggesting either that this region directly interacts with RGSZ1 or that its removal causes a change in the overall conformation of SCG10 and thus prevents RGSZ1 interaction.
Once the interaction was established in vitro, we explored the functional consequences resulting from this interaction. As indicated under "Results," we were unable to show that SCG10 could block the ability of RGSZ1 to accelerate the GTP hydrolysis of G␣ z . Previous studies have indicated that certain RGS complexes do indeed maintain their GAP functionality. For instance, the R7 family of RGS proteins contains a GGL domain that resides outside of the RGS core domain and is responsible for binding G␤ 5 . Recent work has shown that both RGS6 and RGS7 (38) as well as RGS11 (13) contain this GGL domain and bind G␤ 5 , yet still maintain significant GAP activity selectively for G␣ o . Other studies suggest that interacting proteins that bind to RGS proteins within their core domain can directly block GAP activity. Benzing et al. (39) demon-strated that the binding of 14-3-3 to phosphorylated RGS7 inhibits its GAP activity for G␣ i1 . The importance of RGS interactors binding either outside or within the core domain with regard to modulating GAP activity remains to be determined.
Analysis of microtubule polymerization and depolymerization revealed a clear functional consequence of the RGSZ1-SCG10 interaction. Although RGSZ1 had no effect on the ability of SCG10 to block microtubule assembly, there was a striking effect on SCG10-induced disassembly. The mechanism for SCG10-mediated destabilization of microtubules is not yet clear, and there has been considerable controversy in the field concerning this activity of stathmin proteins. Initially, Belmont and Mitchison (21) reported that stathmin interacts with tubulin dimers and increases the catastrophe frequency of microtubules by 3-6-fold in vitro. They concluded that direct stimulation of the catastrophe frequency is the mechanism by which stathmin operates. Subsequently, Curmi et al. (40) failed to reproduce the catastrophe promotion by stathmin and demonstrated that one molecule of stathmin can sequester two molecules of tubulin, forming a tight "T 2 S complex" that can alter the equilibrium conditions leading to depolymerization of microtubules (24). This mechanistic conflict was finally resolved when the discrepancy between these two studies was found to be the result of different pH conditions (41). At pH 7.5, it appears that stathmin interacts with microtubules and specifically induces catastrophes at the plus ends (41), whereas at pH 6.8, stathmin acts primarily by tubulin sequestration. It was further shown that these two distinct activities can be separated by mutational analysis; the N-terminal region was found to be responsible for catastrophe promotion at microtubule plus ends, whereas the C-terminal region was found to be necessary for tubulin sequestration (41).
It is intriguing that RGSZ1 blocks only microtubule disassembly, whereas microtubule assembly remains unaffected. This observation is considered quite remarkable, as it suggests that RGSZ1 binding to SCG10 may provide a mechanism for the separation of the two distinct activities of SCG10, i.e. tubulin sequestration versus initiation of microtubule catastrophes. Perhaps the interaction of RGSZ1 with SCG10 alters the microenvironment of SCG10 such that it more resembles the high pH (7.5) condition than the low pH (6.8) condition. It must be noted that this hypothesis is based on data obtained from studies of stathmin, not SCG10. Although high in sequence homology and closely related (42,43), differences exist between the two that may complicate these assumptions. Further work is needed to clarify whether SCG10 does indeed behave like stathmin with regard to the separation of its microtubuleregulating activities and to clarify the mechanism by which the affinity of SCG10 for tubulin is influenced by RGSZ1.
One of the more interesting questions surrounding the RGSZ1-SCG10 interaction is its relevance in neurons, where distinct populations of both stable and labile microtubules coexist. The growing tips of extending processes contain particularly high levels of very labile polymers, and a number of studies have shown that axonal growth and guidance strongly depend on the highly dynamic behavior of microtubules in these motile structures (44 -46). Specific targeting of SCG10 involves the amino terminus, which mediates its association with membranous vesicles that are apparently directed to the growth cones (27,37). Although this region also contains MAPs that act to stabilize microtubules, the relative contribution of individual proteins controlling stabilization and destabilization in regions of neuronal outgrowth is a matter of debate. For example, tau, a MAP found in developing neurons (47), appears to play significant roles in stabilizing microtubules in vitro (48,FIG. 5. Effect of RGSZ1 on SCG10-induced disassembly of rhodamine-labeled microtubule. Rhodamine-labeled microtubules (3.25 mg/ml) were formed as described under "Experimental Procedures." Microtubules were incubated with buffer (A), 5 M GST-SCG10 (B), or a combination of 5 M GST-SCG10 and 10 M RGSZ1 (C) for 10 min at 37°C. Microtubule disassembly was monitored by fluorescence microscopy using the same conditions for each image. The data are from a single experiment that is representative of four separate experiments. As a control (D), rhodamine-labeled tubulin (3.25 mg/ml) was incubated at 4°C without GTP for 30 min and viewed as described above. 49), but in vivo data suggest that it is not a principal determinant of microtubule stability (50). At present, it is not clear what prevents the depolymerization of microtubules when they need to be stabilized, e.g. in response to extrinsic guidance cues. Possibilities include the presence of stabilizing proteins, the phosphorylation of SCG10 (which has been shown to inhibit its depolymerizing activity) (26), or perhaps the inhibition of SCG10 by RGSZ1, all of which could provide an alternative mechanism for stabilization. The idea that RGSZ1 could contribute to axonal outgrowth is an interesting hypothesis that may provide a link between G protein signaling and neuronal development.