| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
,1
Departments of
* Physiology,
* Biophysics and
*, Psychiatry, Department of
Pharmacology, Department of
Microbiology and Immunology, University of Illinois at Chicago, College of Medicine, Chicago, Illinois, USA
1Correspondence: Department of Physiology and Biophysics, University of Illinois at Chicago, 835 S. Wolcott M/C 901, Chicago, IL 60612-7342, USA. E-mail: RAZ{at}uic.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
s, Gi1, and Go from cell bodies to cellular processes where they appear to localize with tubulin-containing structures. This relocation appeared to depend on the integrity of microtubules, as it was blocked and reversed by nocodazole. Latrunculin, which promotes actin filament depolymerization, had no effect. Both deconvolution microscopy and immunoprecipitation showed a significant increase of G association with microtubules that was coincident with the extension of "neurites." There were distinctions among the G subtypes, with Gs showing the most profound NGF-induced colocalization with tubulin. Translocation of G was blocked by agents that inhibit the MAP kinases required for neuronal differentiation, suggesting that G-protein relocation is triggered by the intracellular signals for differentiation. Consistent with this, G in Neuro-2A cells, which spontaneously differentiate, showed a similar translocation coincident with differentiation. Thus, diverse signals that promote neuronal differentiation and changes in cell morphology may use specific G-proteins to evoke cytoskeletal rearrangement.—Sarma, T., Voyno-Yasenetskaya, T., Hope, T. J., Rasenick, M. M. Heterotrimeric G-proteins associate with microtubules during differentiation in PC12 pheochromocytoma cells.
Key Words: tubulin • G-protein • NGF • cytoskeleton • growth cone
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
subunit. Subsequently the activated G and Gß
are able to activate intracellular effector molecules, such as adenylyl cyclase or phospholipase C. There is a convergence of G-protein-coupled receptors and tyrosine kinase receptors on mitogenic signaling pathways (1
2
3
4
5)
.
Nerve growth factor (NGF) acts on receptors with tyrosine kinase activity to differentiate PC12 cells into a "neuronal" phenotype (6
7
8
9
10)
. Binding of NGF to the tyrosine kinase receptors in PC12 cells is known to stimulate three main signaling pathways: MAP kinase (ERK1/2), PI3-kinase/Akt, and PLC-1 (11
, 12)
.
Purinergic P2Y2 receptors, which are G-protein-coupled, stimulate PC12 cell MAP kinase activity through a pathway distinct from the classical RTK-Ras-Raf-MAPK cascade (13
, 14)
. Related adhesion focal tyrosine kinase (RAFTK, also called PYK2 and CAKß) and protein kinase C 
are involved in mediating the MAP kinase activation by UTP. UTP causes increased tyrosine phosphorylation of multiple proteins in PC12 cells that are common for growth factor and G-protein receptor-mediated signaling. This is sensitive to pertussis toxin and thought to be mediated by Gi (14)
.
Although the generation of intracellular signals by tyrosine kinase receptors has been investigated intensively, regulation of signaling and trafficking events is still not well understood. Many studies have revealed a functional relationship between tubulin and various signaling molecules. A series of studies has demonstrated direct binding between various G and Gß subunits and tubulin (15
16
17
18)
, suggesting new modes of regulation for microtubule assembly. Other studies provide examples of tubulin or microtubules regulating either the activity or the localization of signaling molecules such as the 1 adrenergic receptors (19)
, phospholipase C-ß1 (20)
, and Ki-Ras (21)
.
It has become increasingly clear that G-proteins are found in cellular compartments other than the plasma membrane. Heterotrimeric G-proteins have been implemented in membrane trafficking (22)
and have been seen in the nucleus (23)
. Some have been seen in the cytosol, and this appears to be increased upon G-protein activation (24
25
26
27)
. The relationship of heterotrimeric G-proteins to cell growth and differentiation is not well established.
This study compares the effects of activation of receptor tyrosine kinases with G-protein-coupled receptors with regard to altered distribution of G subunits and the association of those G-proteins with tubulin or microtubules. The role of G-protein relocalization to the newly forming processes and growth cone-like structures was investigated. Stimulation of tyrosine kinase and P2Y2 receptors and the interaction between G-proteins and microtubules may serve to direct the formation of cellular processes. The G-protein association with the microtubule cytoskeleton may regulate localization of those signaling proteins, thereby providing a putative link between receptor tyrosine kinase and G-protein-mediated signaling pathways.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Immunocytochemistry
PC12 cells were plated on 12 mm coverslips in 12-well tissue culture plates. They were washed twice with PBS (180 mM NaCl, 10 mM sodium phosphate buffer, pH 7.4) before and after permeabilization. For cell surface staining, NGF-treated cells were first surface-labeled for 30 min at 4°C by addition of tetramethylrhodamine isothiocyanate (TRITC) -conjugated wheat germ agglutinin (WGA) (1:100, Molecular Probes, Eugene, OR, USA) in PBS containing 0.2% gelatin). Free WGA was removed by five washes with ice-cold buffer. Cells were fixed in methanol for 3 min at -20°C. After washing 3 x 10 min in PBS with 0.1% Triton X-100 cells were incubated for 20 min with blocking buffer (5% nonfat dry milk in PBS), followed by 3 x 10 min washes in PBS. The coverslips were incubated for 40 min with monoclonal anti--tubulin, clone DM 1A, or phalloidin-TRITC (Sigma), washed 3 x 10 min in PBS with 0.1% Triton X-100, and incubated for 30 min with 10 µg/mL goat anti-mouse rhodamine-conjugated antibody (Pierce, Rockford, IL, USA) in PBS. Unbound antibody was washed out with 0.1% Triton X-100 in PBS 3 x 10 min. To demonstrate colocalization of -tubulin or F-actin with G-protein subunits, anti-Gs, Gi1, and Go rabbit polyclonal antibodies (Santa Cruz Biotechnology, Santa Cruz, CA, USA) were used. After the above -tubulin/anti-mouse rhodamine incubations, the cells were incubated for 1 h with the second primary antibody at the appropriate concentration (15 µg/mL, 15 µg/mL, 10 µg/mL) in PBS. After 3 x 10 min washes with 0.1% Triton X-100 in PBS, the cells were incubated for 30 min with 10 µg/mL donkey anti-rabbit fluorescein-conjugated antibody (Pierce), washed three times with 0.1% Triton X-100 in PBS, and finally once in PBS. Stained cells were mounted onto slides with Vectashield (Vector Laboratories, Burlingame, CA, USA) and viewed under a Zeiss LSM510 laser-scanning confocal microscope (Zeiss, Oberkochen, Germany).
Deconvolution microscopy
Images were captured with the Applied Precision (Seattle, WA, USA) DeltaVision system built on an Olympus IX-70 base. Z-stacks were deconvolved using the softworx software. Sections were captured every 200 nm. Typically, 15 iterations based on a measured point spread function, calculated from 1 µM fluorescent beads, were used; 24 images from control and NGF-stimulated cells were counted by three different observers blind to the experimental conditions. The total content of G was measured by counting the number of green "dots" in each randomly selected image on a monitor. The number of G clusters associated with microtubules was determined in a separate count. G immunofluorescence "touching" microtubules was carefully revealed by computer-driven rotation of the images. Agreement among the three observers varied by 6% or less.
Immunoblot analysis
Tissue culture medium was removed and nonadherent cells were washed off with PBS. Cells were solubilized by scraping with ice-cold solubilization buffer (1 mM EDTA, pH 7.4, 20 mM HEPES, pH 7.4, 2 mM MgCl2) containing protease inhibitors. Cells were passed through a 26-gauge needle syringe 15 times on ice to lyse and homogenize the cells. The lysate was centrifuged at 1000 x g for 5 min at 4°C to pellet unbroken cells and nuclei. Protein concentration was determined by the method of Bradford (Bio-Rad Laboratories protein assay, Hercules, CA, U SA). Whole cell extracts containing equal proteins were separated by SDS-PAGE (PAGErTM Gold Precast Gels, BioWhittaker Molecular Applications, Rockland, ME, USA) using 10% gels and transferred to PVDF membranes (Millipore; Bedford, MA, USA) as described previously (28)
. Filters were blocked for 1 h at room temperature with 5% nonfat dry milk in Tris-buffered saline (50 mM Tris-HCl, pH 7.4, 200 mM NaCl) and incubated overnight with primary antibodies (polyclonal rabbit antisera against the various G-protein subunits from Santa Cruz Biotechnology) Gs, Gi1, and Go or ßIII-tubulin (Sigma) at a 1:1000 dilution. The PVDF membranes were washed five times for 5 min with Tris-buffered saline supplemented with 0.1% Tween-20, followed by an 1 h incubation with the appropriate peroxidase-conjugated secondary antibody (1:10,000), (Jackson ImmunoResearch, West Grove, PA, USA). The filters were washed five times for 5 min with Tris-buffered saline supplemented with the same detergent, and developed by chemiluminescence using ECL (Amersham Pharmacia Biotech, Uppsala, Sweden). Bands on X-ray film were quantified by laser densitometry. Filters were occasionally subjected to stripping and reprobing according to the manufacturer’s instructions.
Coimmunoprecipitation of G subunits and -tubulin from PC12 cells
PC12 cells were treated with 10 µM nocodazole for 30 min and with 50 ng/mL NGF for 3 days in DMEM supplemented with 4.5 g/L glucose, 10% bovine calf serum, and 100 µg/mL penicillin/streptomycin. Cells were washed twice before lysis in 500 µL of lysis buffer (50 mM HEPES, pH 8.0, 50 mM NaCl, 0.5% lubrol, 1 mM EDTA, 5 mM MgCl2, 1 mM DTT) containing protease inhibitors. Cells were passed through a 27-gauge needle 10 times. After centrifugation for 10 min at 14,000 rpm in a bench-top centrifuge, the supernatant was mixed with 5 µL of rabbit polyclonal anti-G subunit/antibody and 40 µL of 50% slurry of protein A/agarose (Gibco BRL, Life Technologies, Gaithersburg, MD, USA), followed by an overnight incubation at 4°C. The next day agarose was pelleted and washed three times with 500 µL of lysis buffer each time. The agarose was then resuspended in 30 µL of 2x sample buffer and separated by SDS-PAGE, followed by immunoblot using monoclonal anti--tubulin antibody (Sigma) with 1:1000 dilution.
Online supplemental material
Representative examples of Gi1 association with microtubules and the induction of this interaction to NGF stimulation are viewable as Quicktime movies. Videos 1 and 2 accompany Fig. 3Aa
and Fig. 3Ba
. Microtubules are shown in red, Gi1 in green; 3D rotational analysis was performed with Deltavision system software.
|
Figure 1
shows the distribution of a surface-labeled TRITC-conjugated wheat germ agglutinin that does not associate with tubulin. NGF-treated cells were costained using TRITC-conjugated wheat germ agglutinin (red) and anti--tubulin antibody (green) and processed for deconvolution microscopy. Insert indicates higher magnification of the cell process highlighted by square to show no colocalization of the membrane and fluorescein-labeled tubulin. Scale bar represents 15 µm.
|
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
i1, Gs, and Go subunits in resting PC12 cells
). In untreated PC12 cells, Gi1, Gs, and Go appeared to be distributed throughout the cytoplasm (Fig. 1Ab, e, h
). Gs and Go showed an intense plasma membrane localization, with uniform staining across the cell surface (Fig. 1Ae-f, h-I
, arrowhead). G protein staining was most pronounced in the intracellular perinuclear region and coincided with regions of high microtubule density (Fig. 1B
). Gi1 and Gs, but not Go, were present in nuclei of the resting PC12 cells. Z-stack analysis of the fluorescence pattern (Fig. 1B
) confirmed this.
Given the evidence that tubulin and G-proteins, particularly Gi1 and Gs, have a functional interaction and the observation that G-protein subunits affect the regulation of microtubule assembly (16
, 18)
, we looked for colocalization of all three G subunits with tubulin in PC12 cells. Filamentous cytoplasmic tubulin distributed throughout the untreated PC12 cell cytoplasm was observed (Fig. 1Aa, d, g
). Images in Fig. 1A
represent 1 µ optical sections. Double labeling revealed a significant overlap in the distribution of Gi1 subunit with microtubules, whereas Gs and Go showed a spotty regional codistribution with -tubulin (Fig. 1A
, compare c with f, i). Note the intense colocalization at a perinuclear ring-like structure (Fig. 1B
) and no colocalization at nucleus and membrane.
Stimulation with nerve growth factor results in a differential redistribution of G proteins
PC12 cells were treated with 50 ng/mL NGF for 3 days. Cytoplasmic tubulin redistribution was observed throughout the cells (Fig. 2
a, d, g). Treatment with NGF for 3 days translocated each of the G subunits to the newly formed cellular processes and their terminals. G subunits appeared to be more concentrated in multiple filopodia as well as at the very tip of the growth cone-like extensions (Fig. 2b-c, e-f, h-i
, arrow). -Tubulin and G proteins revealed an increased colocalization primarily in the processes along the microtubule network (Fig. 2c, f, i
). There was no noticeable change in G distribution after 1 day of treatment with NGF.
|
NGF promotes association of G subunits with microtubules in PC12 cells
Although confocal observations and results of immunoprecipitation indicated that G-proteins are associated with tubulin in PC12 cells, it was not clear whether G-proteins were bound to microtubules. To test this, we examined deconvolved images. Three-dimensional (3D) deconvolution microscopy improves experiment performance and image appearance by mathematically removing the out-of-focus effects common to optical light microscopy. We examined large, well-spread PC12 cells for possible colocalization of G-protein with microtubules (Fig. 3
A). As expected, G proteins were localized in the cell body and on the cell surface or at the border of microtubules in control cells. However, some G-proteins were randomly distributed along the length of microtubules. We focused our attention on those optical cell sections where microtubule networks can be neatly visualized. We tested the localization of G-proteins after 3 days of NGF exposure (Fig. 3B
). Colocalization between tubulin and G was diminished in the perinuclear region, since G translocated substantially to the growth cone tip (Fig. 2
, Fig. 3D
). However, G proteins showed an apparently high degree of colocalization with tubulin in the cellular processes induced after NGF treatment (Fig. 3)
. This appeared to be similar for each G-protein examined. Analysis of the microtubules at higher magnification from the box region in Fig. 3B
showed direct association of G-proteins with microtubules in the cellular processes. Images of 0.2 µm optical, planar sections containing microtubule networks taken from four randomly selected control and NGF-treated intact cells were quantified. Total G-protein was measured by counting the number of "dots" in the specified region. The percentage of G protein juxtaposed with microtubules of control vs. NGF-treated cells was compared. Gi1 showed the greatest degree of microtubule association in the resting state and NGF increased this by 73%. This is a marked contrast to Gs and Go, for which NGF increased colocalization by 140% and 163% compared with the respective controls (Fig. 3C
); 3D rotational analysis was used to confirm an increase in association of G with microtubules subsequent to NGF treatment (Fig. 3
video1.mov; Fig. 3
video2.mov).
Microtubule depolymerization decreases G–tubulin association in resting PC12 cells
Microtubules are important for the extension of cellular processes (29
30
31)
; thus, we wondered whether microtubule depolymerization would effect G protein localization and its association with microtubules. Cells were treated with nocodazole for 30 and 90 min and costained with monoclonal anti--tubulin and G subunits before processing for confocal or deconvolution microscopy. Nocodazole caused profound changes in the distribution of cytoplasmic G subunits and microtubules. After 30 min of nocodazole treatment, the G subunits were dispersed in the cytoplasm and redistributed to specific sites around the nucleus. After removal of the drug, scattered G-proteins moved back to the cytoplasmic region (data not shown) and reunited into a continuous system similar to the images presented in Fig. 1Aa, d, g
. When cells were treated for 90 min, G subunits became more tightly clustered near the nucleus (Fig. 4
Ab, f). One or multiple bright immunofluorescent foci staining were observed around the nucleus in some cells (Fig. 4Aj
). The localization of tubulin was seen to change from a filamentous pattern to punctate structures by 90 min (cf. Fig. 1A
and Fig. 4A
). Nocodazole treatment led to a strong accumulation of G proteins in both the perinuclear and bleb regions, where the depolymerized tubulin was localized (Fig. 4Ac-d, g, h, l
). However, the nuclear presence of Gi1 and Gs was unaffected by nocodazole. There was diminished staining of all three G proteins subsequent to nocodazole treatment.
|
Microtubule depolymerization reverses NGF-induced G translocation to microtubules in cellular processes
As these results support the hypothesis that NGF induces association of G-proteins and microtubules in cellular processes, we wondered whether microtubule depolymerization would reverse this process in differentiated cells. NGF-treated cells were exposed to 10 µM nocodazole, and cell morphology was analyzed at different times after addition. Before nocodazole treatment, >95% of the differentiated cells showed long processes. Neurite retraction was time dependent, and the extended processes were almost completely retracted 1–2 h after nocodazole addition. At that time, 80–90% of cells were rounded and showed only blunt cytoplasmic extensions. Analysis of the microtubule network during the growth cone collapse and neurite retraction process showed that the polymerized tubulin was disorganized 30 min after nocodazole addition (Fig. 4Ba, e, i
); 90 min after nocodazole addition, tubulin was no longer in the axon-like extension but had accumulated in the base of the retracted processes, correlating with the morphological retraction shown in Fig. 4A
. Thus, microtubule depolymerization effectively eliminated the formation of filamentous outgrowths in PC12 cells after NGF pretreatment. G subunits moved back toward the perinuclear region. Confocal immunofluorescence examination of all three G proteins suggested a colocalization with -tubulin in the region around the nucleus and at the tips of the growth cones (Fig. 4Bc, g, k
). Analysis by deconvolution microscopy of the cells from five different experiments confirmed the association of G-proteins with tubulin in these regions (Fig. 4Bd, h, l
). Thus, depolymerization of microtubules in NGF-treated cells dispersed G in a manner similar to that of control cells.
G subunits show minimal interaction with actin
To confirm the specificity of microtubules and G-protein interaction, we performed immunostaining for fluorescent phalloidin. The actin cytoskeleton visualized at the cell bottom formed well-organized, parallel filaments that extended into the cell cortex with uniform staining across the cell surface (Fig. 5
Ad). At higher planar sections, F-actin filaments were markedly reduced, being more diffuse and decreasing in both thickness and length. G-proteins were mostly diffusely distributed in the cell body at the higher sections of the cell and did not localize to actin (Fig. 5)
. Unlike the situation with G subunits and tubulin, minimal colocalization between three of the G subunits and actin was observed in F-actin-rich lamellopodia (Fig. 5A
, arrow). As these observations suggested that heterotrimeric G-proteins have little interaction with the actin cytoskeleton, we treated the cells with latrunculin B, a compound altering F-actin polymerization. PC12 cells treated with latrunculin B for 30 min acquired an elongated, polarized shape with narrow, finger-like projections along the cell edge. G-proteins were often enriched in those cell-edge projections (Fig. 5B
). Changes in overall cell morphology induced by latrunculin B were accompanied by a decrease in the intensity of actin staining. Actin fibers became small patches of punctate actin clustered at certain focal adhesion-like sites. Treatment with latrunculin had no noticeable effects on both overall distribution of G-protein and its association with microtubules (Fig. 5B
).
|
NFG increases tubulin-G complex formation in PC12 cells
To confirm that NGF increased the interaction between G and tubulin, biochemical evidence was also needed. The expression of G was not altered by NGF treatment and the amount of tubulin remained constant (Fig. 6
A, B). Anti-G antibodies coprecipitated G and -tubulin from PC12 cell extracts treated with either NGF, nocodazole or both agents (Fig. 6C
). Tubulin coimmunoprecipitates with all three G subunits in cells whether or not they are NGF treated (Fig. 6C
, left). Nocodazole did not disrupt the association of tubulin with G proteins (Fig. 6C
, lanes 2 and 4). Actin was not coimmunoprecipitated with G antibodies under these same conditions. These results were consistent with the colocalization data (Figs. 1
, 2)
. NGF treatment increased coimmunoprecipitation of tubulin with G-proteins, suggesting stabilization of G–tubulin interaction. The NGF-induced increase in coimmunoprecipitation of tubulin and G ranges from 31 to 69%.
|
UTP stimulation alters the subcellular localization of G-proteins and increases its association with microtubules in the cell
Treatment with UTP for 5 days enhanced the proportion of cells displaying neurite-like processes and, in contrast to NGF, evoked slowly developing processes with multiple branching at the tips (Fig. 7
). The shapes shown were those that represent the transition in form from flattened to a monopolar morphology. Continuous stimulation of P2Y2 receptor exerted a strong influence on cytoskeletal organization and dynamics. PC12 cells displayed large numbers of microtubules extending into the outgrowths formed by the cells (Fig. 7A
). G-proteins were preferentially localized on the surface or edges of microtubule filaments (Fig. 7Ad-f
). An increased association of G-proteins similar to effects of NGF was observed with microtubules in the cellular processes as well as in the cell body (Fig. 7B
). The percentage of punctate G "dots" touching microtubules of control vs. UTP-stimulated cells was compared. UTP increased Gs and Go association with microtubules by 124% and 146%, in contrast to Gi1, where G–microtubule association increased by 73% (not statistically significant). These values were quite similar to those seen for NGF stimulation. Pretreatment with pertussis toxin blocked this association with microtubules, suggesting that the effects of UTP are mediated by Gi or Go rather than Gq. Pertussis toxin blocked the phenotypic changes of the differentiated cells. About 50% of cells treated with UTP lost processes in response to pertussis toxin.
|
G-protein redistribution is linked to the neuronal differentiation process
To test the relationship between G-protein redistribution and neuronal differentiation, we used a protein kinase inhibitor that blocks the MAP kinase pathway. Cells were incubated with NGF and PD98059 for 3 days, immunostained with Gs and microtubule antibodies, and processed for deconvolution microscopy. As expected, most of the cells did not develop processes under such conditions. Gs was enriched at the perinuclear region and was closely associated with the microtubules. The tyrosine kinase inhibitor PD98059 blocked the redistribution of G-protein association with microtubules in the cell process (Fig. 8
). Whereas NGF induced a substantial increase in ßIII-tubulin (an indicator of neuronal phenotype), PD98059 blocked this ßIII-tubulin increase in NGF-treated PC12 cells (Fig. 8C
).
|
Migration of G protein with tubulin occurs during the development and formation of NGF-induced processes
To explore the universality of the migration of tubulin and G-protein into cellular processes, we examined the distribution and association of G proteins and tubulin in neuroblastoma SK-N-SH and Neuro-2A cell lines. SK-N-SH cells send out processes spontaneously shortly after they become adherent. Double labeling revealed a substantial overlap in the distribution of Gs and tubulin in the developing processes of both cell types (Fig. 9
f, i). Process formation in Neuro-2A cells is spontaneous but does not begin until 
48 h after plating. Although there is substantial colocalization in process-bearing Neuro-2A cells, those without processes showed little or no colocalization of G proteins and tubulin (cf. Fig. 9f, c
). This suggests that G protein association with microtubules was likely to occur during the formation and development of processes (Fig. 9c
) regardless of the stimulus for process formation.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, 32
33
34)
, and have been implicated in intracellular vesicle trafficking (35)
. In PC12 cells, Gi1, Gs, and Go display cytoplasmic staining (Fig. 1A
). An intense Gs and Go plasma membrane localization with uniform staining across the cell surface was observed (Fig. 1A
). Optical sectioning of PC12 cells shows all three G subunits to be prominent at the perinuclear region (Fig. 1B
). Gi1 and Gs, but not Go, are present within the nucleus as well, indicating that Gi1 and Gs might be selectively imported through the nuclear pores (cf. Fig. 1Ba, b
with c).
G-proteins have been seen to associate with cytoskeletal structures such as microtubules (15
16
17
, 36
37
38
39)
and actin (40)
. The immunofluorescence images shown in Fig. 1A
clearly demonstrate that Gi1 is microtubule associated, whereas Gs and Go have a regional distribution with -tubulin in the perinuclear region of resting PC12 cells. Deconvolution microscopy confirmed that the close association of G subunits with tubulin in selected regions was indeed with microtubules (Fig. 3A
). Higher magnification showed that the percentage of Gi1 localized on the surface or at edges of microtubules was much greater than colocalization of microtubules and Gs or Go in untreated cells (Fig. 3C
). Studies with dimeric tubulin in rat cerebral cortex synaptic membrane showed that Go displayed a much lower affinity for tubulin compared with Gi1 and Gs (17)
. However, -tubulin coimmunoprecipitated with all three G subunits, raising the possibility that all three G subunits are associated with tubulin polymers (41)
. This association is specific for tubulin and microtubules as actin does not colocalize with these G subunits (Fig. 5A
). Colocalization between tubulin and G subunits has been observed by coimmunoprecipitation (42)
. Numerous studies have shown that signal transducing G-proteins can bind synaptic membrane tubulin (17
, 20
, 43
44
45)
. G appears to activate the GTPase activity of tubulin and Gi1, Gs, and Go subunits increase microtubule polymerization dynamics in vitro (16)
. Note that Gß12 binds to microtubules and promotes microtubule assembly in vitro (15)
.
These results provide the initial evidence for a relocation of G-protein subunits in response to cytoskeletal reorganization. Rat pheochromocytoma PC12 cells used in these studies have been a primary model for studying mechanisms underlying neuronal differentiation (46)
and signal transduction. Nerve growth factor acts on receptors with tyrosine kinase activity through a Ras-dependent activation of ERKs (47
48
49)
. Activation of ERKs causes differentiation in PC12 cells (9)
and is associated with enhanced microtubule dynamics, a state in which process outgrowth is facilitated (50)
. We found that NGF not only stimulates the redistribution of G subunits in the newly developing neurites along the microtubule network, but translocates Gi1, Gs, and Go to the tips of neurites, specifically to the growth cones (Fig. 3D
). Translocation of Go to the growth cones during neurite development has been reported (51)
. There is an increased colocalization with -tubulin in the newly developing neurites along the microtubule network and a concomitant reduction in colocalization at the perinuclear region.
A closer examination of G subunits in the cell process using deconvolution microscopy revealed that NGF induced G subunit association with components of cytoskeleton in cell body and especially in tubular extensions of the cellular processes (Fig. 3B
). The induction of tubulin–G association by NGF was much more dramatic for Gs than for Gi1 (Fig. 3C
). Curiously, 85% of Gs became associated with microtubules after NGF treatment (Fig. 3C
). In a previous study, coimmunoprecipitation experiments showed that 85% of the Gs in synaptic membranes from rat cerebral cortex was complexed with tubulin (45)
. Taken together, these results suggest that G subunits are probably enriched in the tubulin subdomains as well as intermediate structures involved in the delivery of G subunits to the growth cones. Thus, it is plausible that G-protein association with microtubules may play an important role in the regulation of microtubule formation in addition to its regulatory role in cellular signal transduction.
In contrast to receptor tyrosine kinases, the intermediate steps linking GPCRs to the activation of ERK are poorly understood, and significant heterogeneity and complexity exist in the signaling pathways used by various GPCRs (52)
. ERK activation elicited by UTP acting on a P2Y2 receptor in HEK-293 cells (53)
differs from the NGF-mediated ERK activation. The activation of G-protein-coupled purinergic receptors by UTP in PC12 cells stimulates phosphoinositide breakdown, release of intracellular calcium, and influx of external calcium but does not stimulate norepinephrine release (54)
. In addition, recent evidence demonstrates that purinergic receptor agonists activated Gq/11 and Gi3 in gastric and aortic smooth muscle and heart membranes, Gq/11, Gi1, and/or Gi2 in liver membranes and Go and Gi1–3 in brain membranes (55)
.
In this study, a comparison of UTP and NGF responses reveals that UTP induces a few unbranched processes within 2 days whereas NGF induces multiple, highly branched processes within 24 h (14)
, (T. Sarma and M. M. Rasenick, unpublished results). Purinergic receptor activation results in a different time course of neurite formation from NGF, and association of microtubules with G subunits follows that same time course (cf. Figs. 2
and 3
with Fig. 5
). This idea is supported by recent findings demonstrating an association of G-protein-coupled receptors and G-protein-coupled receptor kinases with microtubules (56
57
58
59)
. In contrast to NGF-induced association with microtubules, it is interesting that UTP increases this association significantly for Gs and Go but not for Gi1. This may be related to the fact that the P2Y2 receptors are activating Gi, but the nature of that relationship is not clear.
It is noteworthy, however, that even in the absence of activated receptor (spontaneous differentiation of Neuro-2A cells; Fig. 6
), the association of G with microtubules increases upon neurite formation. Thus, G-protein activation is not likely to be required in order to increase association between G and microtubules in differentiating neuronal cells.
Microtubules are required for the formation of PC12 cell outgrowths. For both UTP-induced and NGF-induced processes, the outgrowths are retracted after removal of the agonist within 30 h. Treatment with PD98059 and subsequent inhibition of the MAP kinases required for NGF-induced neuronal differentiation blocked the G-protein translocation. This suggests that redistribution of G-protein is linked to the NGF-induced cell signaling pathway. This reversibility suggests that regulation of process formation is closely related to the stimulation of signaling pathways associated with cellular differentiation. It is unclear, however, whether translocation is secondary or is triggered by NGF independently.
The G-protein translocation appears to depend on the integrity of microtubules, but not the actin cytoskeleton, as it was blocked and reversed by nocodazole but unaffected by latrunculin. Furthermore, immunoprecipitation experiments demonstrated a consistent physical interaction between G subunits and tubulin (45)
. Deconvolved images of NGF-treated cells suggest that G subunits are associated along the microtubule network in the cellular processes and that the formation of G-protein/microtubule complexes is not dependent upon F-actin.
Two explanations could account for the redistribution of G subunits. The first would involve local movement of G from cell body to the cellular processes. In fact, we recently observed that an activated Gs-GFP construct translocates from the membrane in living cells (27)
. The second possibly would preserve local cytoarchitecture but would involve a global folding of G protein–microtubule complexes. Although other explanations are possible, these data support the idea that microtubule organization influences both morphology and the distribution of signaling proteins in the cell. G-protein association with microtubules in neuroblastoma cells lines Neuro-2A and SK-N-SH cells and NGF-induced PC12 cells demonstrates that microtubules play a pivotal role during the neurite formation and development of cell processes.
The discovery of signaling molecules that interact with microtubules, as well as the multiple effects on signaling pathways of drugs that reorganize microtubules, indicate that microtubules are likely to be critical to the spatial organization of signal transduction and that heterotrimeric G-proteins may help dictate the shape and movement of cells.
|
ACKNOWLEDGMENTS |
|---|
Received for publication July 30, 2002. Accepted for publication January 10, 2003.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
