G Protein β1γ2 Subunits Promote Microtubule Assembly*

α and βγ subunits of G proteins are thought to transduce signals from cell surface receptors to intracellular effector molecules. Gα and Gβγ have also been implicated in cell growth and differentiation, perhaps due to their association with cytoskeletal components. In this report Gβγ is shown to modulate the cytoskeleton by regulation of microtubule assembly. Specificity among βγ species exists, as β1γ2 stimulates microtubule assembly, and β1γ1 is without any effect. Furthermore, a mutant β1γ2, β1γ2(C68S), which does not undergo prenylation and subsequent carboxyl-terminal processing on the γ subunit, does not stimulate the formation of microtubules. β immunoreactivity was detected exclusively in the microtubule fraction after assembly in the presence of β1γ2, suggesting a preferential association with microtubules rather than soluble tubulin. Crude microtubule fractions from ovine brain contain Gβγ, and electron microscopy reveals a specific association with microtubules. The decoration of microtubules by Gβγ appears to be strikingly similar to the periodic pattern observed for microtubule-associated proteins, suggesting a similar site of activation of microtubule assembly by both agents. It is suggested that reformation of the cytoskeleton represents an additional cellular process mediated by Gβγ.

␣ and ␤␥ subunits of G proteins are thought to transduce signals from cell surface receptors to intracellular effector molecules. G ␣ and G ␤␥ have also been implicated in cell growth and differentiation, perhaps due to their association with cytoskeletal components. In this report G ␤␥ is shown to modulate the cytoskeleton by regulation of microtubule assembly. Specificity among ␤␥ species exists, as ␤1␥2 stimulates microtubule assembly, and ␤1␥1 is without any effect. Furthermore, a mutant ␤1␥2, ␤1␥2(C68S), which does not undergo prenylation and subsequent carboxyl-terminal processing on the ␥ subunit, does not stimulate the formation of microtubules. ␤ immunoreactivity was detected exclusively in the microtubule fraction after assembly in the presence of ␤1␥2, suggesting a preferential association with microtubules rather than soluble tubulin. Crude microtubule fractions from ovine brain contain G ␤␥ , and electron microscopy reveals a specific association with microtubules. The decoration of microtubules by G ␤␥ appears to be strikingly similar to the periodic pattern observed for microtubule-associated proteins, suggesting a similar site of activation of microtubule assembly by both agents. It is suggested that reformation of the cytoskeleton represents an additional cellular process mediated by G ␤␥ .
G proteins play important roles in signal transduction by transferring signals from cell surface receptors to intracellular effector molecules. Although receptor-G protein-effector complexes can reconstitute hormone-sensitive signaling systems in vitro, it is likely that the regulation of receptor-G protein signaling is substantially more complex in the cell. Many studies have implicated the participation of the cytoskeleton in neurotransmitter signaling pathways (1)(2)(3)(4)(5)(6)(7). Although G proteins are likely to be membrane-bound when coupled to receptors, recent results from several laboratories suggest their association with several subcellular compartments. In Caenorhabditis elegans embryos, G ␤ is required for proper spindle orientation and transiently associates with the region of asters (the array of microtubules emanating from the centrosomes) just before and during cell division (8). Astral localization of G o␣ and G ␤ has also been observed in mammalian cells (9). In addition, G ␤␥ , microtubules, and phosphatidyl inositol 4,5-biphosphate may participate in synaptic vesicle recycling by regulating the GTPase activity of dynamin 1 (10). These studies suggest a link between microtubules and G protein signaling.
Association of the signal-transducing G proteins, G s , G i1 , and G q , with the synaptic membrane tubulin has been observed previously (2)(3)(4)(5)11). Tubulin appears to activate G proteins directly, and complexes between tubulin and G ␣ have been isolated from plasma membranes. While some interaction between tubulin and G ␤␥ has been observed previously, the role of such interaction remains unclear. Modification of microtubule cytoskeleton by G ␤␥ might provide an explanation for the association of G ␤␥ with the mitotic spindle and its role in cell growth and differentiation. The present study was undertaken to explore the role of G ␤␥ in microtubule assembly and dynamics.

EXPERIMENTAL PROCEDURES
Purification of Proteins-PC-tubulin 1 (tubulin free of microtubuleassociated proteins) was purified from ovine brain by two cycles of assembly and disassembly (12) followed by phosphocellulose chromatography (13). The tubulin preparation made by two assembly-disassembly cycles (microtubule proteins) contains microtubule-associated proteins (MAPs). These MAPs were removed by phosphocellulose chromatography (13). ␤1␥2, purified from SF9 cell membranes (14) was kindly provided by Drs. T. Kosaza  Crude Microtubule Protein Preparation-Fresh ovine brains were obtained from a local abbatoir and homogenized in PEM buffer (100 mM PIPES, PH 6.9, 2 MM EGTA, 1 mM MgCl 2 ) containing 0.1 mM GTP, followed by centrifugation at 100,000 ϫ g for 60 min at 4°C. This initial pellet (P1), the crude membrane preparation, was resuspended in Hepes buffer (0.1 mM Hepes, pH 8.0, 1 mM EDTA, 1 mM dithiothreitol, and 0.3 mM phenylmethylsulfonyl fluoride) and saved for immunoblotting. Tubuluin from the initial supernatant (S1, cytosolic fraction) was allowed to assemble at 37°C in PEM buffer containing 30% glycerol and 0.5 mM GTP for 60 min (12). Suspended microtubules were divided into two tubes, and both were centrifuged at 150,000 ϫ g for 30 min at 37°C. The resulting pellet (P2) represented the crude microtubule preparation. One crude microtubule pellet was fixed in 1% glutaraldehyde for 1 h, embedded in LR white (medium grade, Ernest F. Fullam Inc.), and subjected to thin sectioning. The other crude microtubule pellet was resuspended in PEM buffer, depolymerized on ice, and examined for the presence of G ␤␥ .
Microtubule Assembly-PC-tubulin in PEM buffer and 1 mM GTP was incubated with or without ␤1␥2 (0.05 mg/ml) or ␤1␥1 (0.1 mg/ml) at 0°C for 10 min followed by assembly at 37°C for 45 min to 1 h. Assembly was quantitated by centrifuging the polymer at 150,000 ϫ g for 20 min at 37°C, and pellets and supernatants were separated. Pellets were resuspended in cold PEM buffer and protein concentrations were determined both in pellets and supernatant fractions. Protein was estimated by Coomassie Blue binding according to the method of Bradford (15) using bovine serum albumin (BSA) as a standard. All ␤␥ combinations were exchanged in PEM buffer through a rapid spin * This work was supported by National Institute of Mental Health Grants MH 39595 and MH 00669. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
§ To whom correspondence should be addressed: Dept. of Physiology and Biophysics, m/c 901, University of Illinois, College of Medicine, 835 South Wolcott Ave., Chicago, IL 60612-7342; Fax: 312-996 -1414; E-mail: raz@uic.edu or sukla@uic.edu. column (Bio-Gel P6DG) before testing its effect on microtubule assembly. Alternately, to avoid reduction in protein concentration by gel filtration, a buffer control was performed in preparations where ␤␥ concentrations were low.
Electrophoresis and Immunoblotting-Samples for electrophoresis were dissolved in 3% SDS Laemmli sample buffer with 50 mM dithiothreitol and subjected to SDS-polyacrylamide gel (10% acrylamide and 0.133% bisacrylamide) electrophoresis. For Western blotting, samples after electrophoresis, were transferred to nitrocellulose membrane using a semi-dry transfer apparatus (Bio-Rad). The nitrocellulose membranes were incubated in 3% BSA in TBS (10 mM Tris-HCl and 150 mM NaCl, pH 7.4) for 2 h at room temperature followed by an overnight incubation with anti-G ␤ antibody. Detection of antibody binding was with an alkaline phosphatase-conjugated goat anti-rabbit antibody (Pierce, Rockford IL). The membrane was washed with 0.02% Tween 20 in TBS, and the substrate solutions were added for color development.
Electron Microscopy-For electron microscopic analysis, about 15 l of the sample, after polymerization, were removed and placed on a carbon-coated nickel grid (Electron Microscopy Sciences Inc.). After 10 -15 s, the grids were rinsed with 10 drops of 2% uranyl acetate for negative staining, blotted dry with a filter paper, and viewed in a JEOL 100S electron microscope. For immunoelectron microscopy, microtubules were fixed by adding 0.5% glutaraldehyde in warm PEM buffer. After 15 min at room temperature, 15-l aliquots were removed and placed on carbon coated nickel grids. After 15-20 s, the grids were rinsed with 10 drops of warm PEM buffer and placed into wells of a grid box containing 0.1% BSA in TBS (17-20 l). Subsequently the grids were transferred into 5% normal goat serum (NGS) in BSA-TBS and incubated for 30 min at room temperature. The grids were then subjected to overnight incubation at 4°C with polyclonal anti-G ␤ diluted in 1% NGS-BSA-TBS. Following this, grids were transferred into a drilled grid box immersed in BSA-TBS in a square Petri dish. The grid box was then suspended in a beaker containing BSA-TBS and subjected to gentle agitation with three buffer changes at 10-min intervals. This was followed by a subsequent incubation (3 h) with a goat anti-rabbit IgG antibody (1:5 dilution) conjugated with 10-nm colloidal gold (Amersham Corp.). Grids were washed as before and were counterstained with 2% uranyl acetate and viewed, blind to experimental conditions, in a JEOL 100S electron microscope.
For thin sectioning, pellets were fixed in 1% glutaraldehyde for 1 h and kept overnight in PEM buffer at 4°C. They were post-fixed with 1% osmium tetroxide in 0.1 M phosphate buffer, pH 7.2, for 10 min, dehydrated in alcohol and embedded in LR white. Thin sections were cut according to standard procedures, placed on Formvar-coated nickel grids, and subjected to immunocytochemistry using polyclonal anti-G ␤ antibody and colloidal gold-conjugated secondary antibody as described above.

RESULTS AND DISCUSSION
Several species of G protein ␤ and ␥ subunits have been shown to exist (16), and they appear to show preferences for various forms of G ␣ as well as for specific receptors (17)(18)(19)(20). ␤ and ␥ subunits of G proteins function as a tightly bound complex, which is disrupted only by denaturation. Coexpression of both subunits are required to detect the activity of the complex and functional properties of the complex are dependent on contributions from both proteins (21). Although transducin ␤␥ primarily consists of ␤1 and ␥1 isotypes, ␤␥ subunits purified from bovine brain are heterogenous in composition.
␤1␥2 Promotes Microtubule Assembly-The effect of different combinations of ␤␥ on microtubule assembly was tested. Tubulin purified free of microtubule-associated proteins was incubated at 37°C in the presence of ␤1␥2 or ␤1␥1 (transducin ␤␥). Assembly was monitored by negative staining electron microscopy and measuring protein in polymers collected by centrifugation. Under these buffer conditions, tubulin does not assemble well unless glycerol (25-30% by volume) is present in the buffer. However, microtubule assembly was stimulated markedly, when ␤1␥2 was present at ϳ1:20 molar ratio with tubulin ( Figs. 1, 2, and 3). In contrast, ␤1␥1 had no effect on microtubule assembly. Electron microscopic analysis indicated very few microtubules formed by tubulin alone or in the presence of ␤1␥1 (Fig. 1). In the presence of ␤1␥2, however, robust microtubule polymerization occurred. To quantify this response, eight random fields were examined by electron microscopy on a total of five grids from three separate experiments. The result is shown in Fig. 2. In a field of 5 m in an average grid, 1-4 microtubules were found in the control samples (mean ϭ 1.7) or samples incubated with ␤1␥1 (mean ϭ 2.7), while 49 -65 (mean ϭ 54.7) isolated microtubules were detected in samples incubated with ␤1␥2. Protein estimation in the pellets also indicated a 71% increase in the presence of ␤1␥2 (Fig. 3A). No detectable change in pellet protein concentration (compared with controls) was observed in the presence of ␤1␥1. The maximal extent to which added ␤1␥2 if incorporated into the microtubule pellet could increase the pellet protein is calculated to be 13%. This suggests that the increase in pellet protein seen in Fig. 1 is indicative of ␤1␥2-induced enhancement of microtubule assembly. SDS-PAGE of the samples further confirmed FIG. 1. Electron micrograph of microtubules assembled in presence of ␤1␥2, ␤1␥1, or ␤1␥2(C68S); stimulation of assembly by ␤1␥2. PC-tubulin (1.75 mg/ml) in PEM buffer and 1 mM GTP was incubated with or without ␤1␥2 (0.05 mg/ml) or ␤1␥1 (0.1 mg/ml) at 0°C for 10 min followed by assembly at 37°C for 1 h. Samples were then divided into two aliquots. One aliquot was used for estimating polymer mass (see Fig. 3), and the other was processed for electron microscopy and viewed in a JEOL 100S electron microscope. Samples from the experiment described in Fig. 4 were used for electron microscopic analysis of microtubule assembly in the presence of ␤1␥2(C68S). The bar represents 0.1 m.
the increase in tubulin concentration in the pellet formed in the presence of ␤1␥2 (Fig. 3B). This result clearly indicates that ␤1␥2 promotes microtubule assembly, perhaps by reducing the critical concentration of tubulin required for assembly. In the presence of ␤1␥2, PC-tubulin, at a concentration as low as 0.8 mg/ml, assembled readily into microtubules (15-20 microtubules/5-m field in a grid).
Prenylation of ␤1␥2 Is Required for Promotion of Microtubule Assembly-The finding that ␤1␥2 stimulated microtubule assembly, while ␤1␥1 was ineffective, suggests that the ␥ subunit is the determining factor for the observed difference. One major difference between ␥1 and ␥2 is the modification at the carboxyl terminus of ␥; while ␥1 is farnesylated, ␥2 is geranylgeranylated (22)(23)(24)(25). Furthermore, the amino acid sequences of ␥1 and ␥2 are substantially different. Thus, differential effect of ␤1␥1 and ␤1␥2 in microtubule assembly might be attributed to the ␥ subunit and the type of isoprenoid moiety attached to it.
Although prenylation and/or further carboxyl-terminal processing of G ␥ is not required for assembly of ␤␥ complexes, such modifications are indispensable for the function of ␤␥. High affinity interactions of ␤␥ with either G protein ␣ subunits or effector molecules require that the ␥ subunit be prenylated (21,26,27). To examine the role of the ␥ subunit carboxyl-terminal processing, a mutant ␤1␥2, ␤1␥2(C68S), which cannot undergo prenylation and further carboxyl-terminal processing was used. This mutant ␤␥ was inactive in all functional assays (21). As shown in Figs. 1 and 2, ␤1␥2(C68S) did not stimulate microtubule assembly. Analysis of microtubules from electron microscopic data (Fig. 2) indicated no significant difference from control (3 Ϯ 1.0 versus 1.66 Ϯ 0.49 per 5-m field in an average grid). However, larger tubulin aggregates were seen frequently in samples assembled from ␤1␥2(C68S) (Fig. 1). ␤1␥2(C68S) did not alter total polymer mass over controls as determined by protein estimation in the pellet (data not shown).
␤1␥2 Shows Specific Binding to Microtubules-SDS-PAGE and immunoblot analysis of pellet and supernatant fractions indicate ␤ immunoreactivity in microtubule pellets after assembly in the presence of ␤1␥2 (Fig. 4), suggesting a preferential association with microtubules rather than soluble tubulin.
However, the microtubule and tubulin aggregates formed in the presence of ␤1␥1 or ␤1␥2(C68S) did not bind these ␤␥ species, as all ␤ immunoreactivity remained in the supernatant (Fig. 4). To rule out the possibility that the observed association of ␤1␥2 to microtubules is not due to aggregation and subse-  Fig. 1) were pelleted by centrifugation at 150,000 ϫ g for 20 min at 37°C, and pellets and supernatants were separated. Pellets were resuspended in cold PEM buffer. Assembly was quantitated by estimating the protein in resuspended pellets and supernatants and plotted as indicated (A). B, tubulin staining on resuspended pellets when subjected to SDS-PAGE. One of three similar experiments is shown. ␤1␥2 increased polymer mass by 50 -70%, while no significant change in polymer mass was observed in presence of ␤1␥1. Note that both microtubules and tubulin aggregates appear in the pellets, rendering polymer mass in the pellets underestimated for microtubule formation. Since microtubule assembly follows a nucleationelongation phenomenon requiring a critical tubulin concentration for assembly, protein estimates in the pellet do not accurately reflect microtubule assembly, particularly when rate and extent of assembly is suboptimal. When assembly proceeds efficiently, the correlation is much higher.
␤1␥1 and ␤1␥2(C68S) Do Not Inhibit Microtubule Assembly and Fail to Bind to Microtubules-Even though significant microtubule formation occurred at subcritical tubulin concentrations in the presence of ␤1␥2, it was necessary to determine if ␤1␥1 or ␤1␥2(C68S) inhibited microtubule assembly (as opposed to failing to promote assembly). To address this issue, microtubule assembly was induced by adding 30% glycerol to the incubation buffer, which reduces the critical concentration for assembly of purified tubulin (12,28). Utracentrifugation studies and electron microscopic analysis demonstrated normal microtubule formation. Addition of ␤1␥2(C68S) or ␤1␥1 did not alter the level of assembly as detected by protein assays of the pellets as well as by electron microscopic observation. Since the degree of microtubule assembly was similar with all ␤␥ species in the presence of glycerol, association of G ␤␥ with microtubules was tested under this condition. Consistent with the data in Fig. 4, only ␤1␥2 was associated with microtubules. The other two ␤␥ species (␤1␥1 and ␤1␥2C68S) remained exclusively in the supernatant fractions (data not shown).
The results in Figs. 1-4 suggest that the functional interactions of ␤␥ subunit of G proteins with tubulin/microtubule systems require a similar structural specificity of ␤␥ to those which determine G ␤␥ interactions with G ␣ or effector molecules such as phospholipases or ion channels. In this context, tubulin/microtubules may represent a new class of effector system for ␤␥ subunits of G proteins. As a corollary to this, ␤␥ subunits might represent a new class of microtubule associated protein.
It is noteworthy that the major microtubule-associated protein (MAP) of sea urchins and several other echinoderms is a 77-kDa protein (EMAP) with no sequence homology with any other identified MAPs. However, EMAP exhibits a significant homology with ␤ subunits of heterotrimeric G proteins (29).
G ␤␥ Binds to Microtubules at Regular Intervals, when Used to Promote Assembly-The suggestion that G ␤␥ might be behaving as a MAP for the promotion of microtubule polymerization elicited the question whether the binding of G ␤␥ to microtubules resembled that seen with MAPs. A polyclonal antibody directed against the ␤ subunit was used to determine the nature of association of ␤/␤␥ with microtubules assembled in the presence of ␤1␥2. ␤ subunits were detected upon short projections which extended laterally from microtubules at regular intervals (Fig. 5). In addition, the ␤ subunit was bound to some of the oligomeric structures. Similar structures, consisting of high molecular weight MAPs and tubulin, are frequently found in microtubule preparations and are suggested to be intermediate structures for microtubule formation (30). The decoration of microtubules by G ␤ appears to be strikingly similar to that observed by MAPs (31), suggesting perhaps, a similar site of activation of microtubule assembly by both agents.
Crude Microtubule Fractions Contain G ␤␥ -To assess the binding of ␤␥ to microtubules in situ, ␤ immunoreactivity was tested in crude microtubule fractions prepared from ovine brain homogenate. As shown in Fig. 6A, ␤␥ was detected clearly in the crude microtubule preparation. In addition, ␤␥ was detected in the unpolymerized cytosolic fraction (S2), although in lesser amount. Since particulate structures were removed prior to microtubule polymerization, the ␤␥ associated with microtubules was present in the cytosol. The bulk of this ␤␥ was able to associate with the crude microtubule fraction during the polymerization step (Fig. 6A).
G ␤ immunoreactivity in the S2 fraction may represent ␤␥ species that do not bind to microtubules. Alternately, one or several MAP(s) may compete with ␤1␥2 for microtubule binding, releasing some ␤1␥2 into the S2 fraction. This is supported by the observed association of ␤␥ with the crude microtubule pellet. As shown in Fig. 6B, the ␤ subunit is detected laterally on microtubule walls as a short projection. Note that the periodic association of G ␤␥ seen when G ␤␥ -induced microtubule polymerizatin (Fig. 5) was not observed. Further, ␤␥ was not detected in the two cycle purified microtubule fraction (microtubule protein), suggesting the loss of association of ␤␥ with microtubules during the purification cycle.
In addition to regulating effector molecules, ␤␥ subunits may play a role in macromolecular assembly (32). They enhance the association of ␣ subunits with receptors. ␤␥ binding to the pleckstrin homology domains of ␤-adrenergic receptor kinase (␤ARK), permits the translocation of ␤ARK from cytosol to membrane and facilitates its association with target receptors (33,34). The observed stimulation of microtubule assembly by ␤1␥2 further supports its role in the macromolecular organization of the cell. Although the assembly properties of ␤␥ subunits are partly attributable to the WD-40 repeating units (a 40 -43-amino acid tandem repeat usually punctuated with tryptophan-aspartate) present in the ␤ subunit of the ␤␥ complex (32), prenylation of ␥ subunits may play a crucial role in mediating the association between ␤␥ and other proteins (35,36). Furthermore, it appears from this study that prenylation FIG. 5. Immunoelectron micrograph of association of ␤ subunit of G proteins in microtubules assembled in presence of ␤1␥2. PC-tubulin in PEM buffer and 0.5 mM GTP was allowed to assemble at 37°C for 30 min in the presence of ␤1␥2. The samples were then subjected to immunocytochemistry using polyclonal G ␤ antibody and 10-nm colloidal gold-conjugated secondary antibody as described under "Experimental Procedures." After the final wash, grids were counterstained with 2% uranyl acetate and viewed in a JEOL 100S electron microscope. G ␤ subunits are shown to be associated with both microtubules (arrows) and oligomeric structures (star). In addition, ␤ immunoreactivity on microtubules appears to be at regular intervals (multiple arrows). The bar represents 0.1 m. This labeling was specific for G ␤ , since no obvious gold labeling was observed in control microtubules or ␤1␥2 microtubules incubated with normal rabbit serum. of the ␥ subunit is critical for the interaction of ␤␥ with tubulin for stimulation of microtubule assembly.
In mammalian cells, G ␤ immunoreactivity appears consistently associated with both microtubules in mitotic spindles and with the plasma membrane (9). Spindle morphogenesis is associated with changes in the level of microtubule polymers and microtubule dynamics, and these two processes are tightly coupled (37). Thus, the association of G ␤␥ with microtubules in mitotic cells might represent a vehicle for coordination of mitotic centers by extracellular signals. It is also noteworthy, in this regard, that G ␤ has been shown essential for Dictyostelium development (38) and oocyte maturation in starfish (39). Although several suggestions have been made, the target for ␤/␤␥ action is unknown. More recently, in C. elegans embryos, G ␤ was shown to be localized to the region of asters just before and during cell division and to be required for proper orientation of the early cell division axis (8). G ␤ , which was found in cell membranes, from the two-cell stage onwards appeared to be concentrated at the contact between cells. Should ␤␥ become activated subsequent to cell-cell contact, the resulting regulation of microtubule assembly and dynamics could orchestrate certain developmental programs.
In addition to ␤␥, G protein ␣ subunits have been suggested to be associated with several subcellular compartments including microtubules (9, 40, 41) and the binding of G ␣ to tubulin has been well documented (2)(3)(4)(5). The family of tubulin proteins (␣, ␤, and ␥) which binds and hydrolyzes GTP, but fails to show G protein consensus motifs, may be included in a broad spectrum of a GTPase superfamily (42,43). The ␣ subunit of several G proteins have been shown to modulate microtubule assembly (7,41). Recently, G ␣ has been shown to activate tubulin GTPase and inhibit microtubule assembly in a GTP-dependent fashion. 2 Inhibition and stimulation of microtubule assembly by activated G ␣ and G ␤␥ , in a concerted manner, might lead to microtubule reorganization in vivo. Such a process could induce localized changes in cell shape and might represent an important element in hormone or neurotransmitter signaling.
Cellular microtubules exhibit a highly dynamic behavior and are involved in multiple functions including chromosome movement in cell division, the organelle transport within the cell, and the maintenance of cell morphology (44 -46). In general, the biological function of microtubules are based on the ability of tubulin to polymerize and depolymerize. Although results from different laboratories suggest the involvement of Ca 2ϩ , cAMP, Mg 2ϩ , GTP, and several MAPs in the regulation of microtubule dynamics, precise spatial and temporal control of the process is unknown. The data presented here point toward G proteins as a physiological regulator for microtubule assembly and dynamics. These data further suggest a role for membrane-associated ␤␥, perhaps as a focal point for microtubulemembrane interaction. Extracellular signals to G proteincoupled receptors could modify a variety of microtubuledependent events, using G proteins themselves as a second messenger. Regulation of microtubule assembly by G ␤␥ may provide an alternative pathway by which G proteins function in cellular signaling.