Immunogold Electron Microscopic Demonstration of Distinct Submembranous Localization of the Activated γPKC Depending on the Stimulation

Abstract

We examined the precise intracellular translocation of γ subtype of protein kinase C (γPKC) after various extracellular stimuli using confocal laser-scanning fluorescent microscopy (CLSM) and immunogold electron microscopy. By CLSM, treatment with 12-O-tetradecanoylphorbol-13-acetate (TPA) resulted in a slow and irreversible accumulation of green fluorescent protein (GFP)-tagged γPKC (γPKC–GFP) on the plasma membrane. In contrast, treatment with Ca²⁺ ionophore and activation of purinergic or NMDA receptors induced a rapid and transient membrane translocation of γPKC–GFP. Although each stimulus resulted in PKC localization at the plasma membrane, electron microscopy revealed that γPKC showed a subtle but significantly different localization depending on stimulation. Whereas TPA and UTP induced a sustained localization of γPKC–GFP on the plasma membrane, Ca²⁺ ionophore and NMDA rapidly translocated γPKC–GFP to the plasma membrane and then restricted γPKC–GFP in submembranous area (<500 nm from the plasma membrane). These results suggest that Ca²⁺ influx alone induced the association of γPKC with the plasma membrane for only a moment and then located this enzyme at a proper distance in a touch-and-go manner, whereas diacylglycerol or TPA tightly anchored this enzyme on the plasma membrane. The distinct subcellular targeting of γPKC in response to various stimuli suggests a novel mechanism for PKC activation. (J Histochem Cytochem 56:253–265, 2008)

Protein kinase c (PKC) consists of a family of serine/threonine kinases that plays crucial roles in the signal transduction pathways of virtually all cell types (Nishizuka 1988,1992). More than 10 different PKC subtypes have been identified, which can be classified into three groups based on their activation requirements: calcium- and diacylglycerol (DG)-dependent “conventional PKC” including α, βI, βII, and γ; calcium-independent “novel PKC” including δ, ɛ, η, and θ; and the calcium- and phorbol ester-independent “atypical PKC” including ζ and λ/ι subtypes (Nishizuka 1988). However, little is known about in situ associations of each subtype with its specific substrates.

The fact that all cells express multiple PKC subtypes suggests that each subtype has unique functions. However, it is difficult to assign subtype-specific functions by conventional enzymological analysis because of their low substrate specificity. Therefore, we studied the intracellular localization of PKC subtypes in the tissues (Tanaka and Saito 1992; Tanaka and Nishizuka 1993) and also investigated the translocation of PKC subtypes in living cells using green fluorescent protein (GFP)-tagged PKCs under a confocal laser-scanning microscope (CLSM) (Sakai et al. 1997). We have reported that the spatial and temporal translocation of PKC is dependent on the PKC subtype and on the extracellular stimulus (Sakai et al. 1997; Shirai et al. 1998; Ohmori et al. 2000; Kashiwagi et al. 2002) and that the live imaging studies were efficient for investigating the intracellular signaling system.

Although it has been reported that different stimuli lead PKC translocation to the different sites (Sakai et al. 1997; Oancea and Meyer 1998; Shirai et al. 1998; Ohmori et al. 2000; Kashiwagi et al. 2002), some different stimuli induced similar translocation, for example, to the plasma membrane (Sakai et al. 1997). Because the plasma membrane is the site where DG, the PKC activator, is produced, PKC translocation to the plasma membrane is considered to be critical for the signaling system.

Under physiological conditions, receptor-activated hydrolysis of phosphatidylinositol (4,5)-bisphosphate (PIP2) by phospholipase C (PLC) generates DG to activate cPKCs and nPKCs. The other product, inositol (1,4,5)-triphosphate (IP3), induces Ca²⁺ mobilization from intracellular Ca²⁺ stores (Berridge 1984). As one of the cPKCs, γPKC has C1 and C2 domains in the regulatory domain; the C1 domain interacts with DG or phorbol esters and the C2 domain binds Ca²⁺ (Nishizuka 1988,1992; Ono et al. 1989; Ananthanarayanan et al. 2003; Dries et al. 2007). As visualized by confocal microscopy, activation of either C1 or C2 domain independently, or the simultaneous stimulation of both domains by UTP receptor-activation, translocates γPKC to the plasma membrane (Sakai et al. 1997; Oancea and Meyer 1998; Oancea et al. 1998). With respect to translocation to the plasma membrane, further biochemical study revealed that the affinity of C1 or C2 domain for membrane interaction is different. It was reported that C2 domain is responsible for the initial Ca²⁺- and phosphatidylserine-dependent electrostatic membrane binding of cPKC, whereas the C1 domain is involved in subsequent membrane penetration and DG binding for the activation of the enzyme (Medkova and Cho 1999). In this study we investigated the precise submembranous translocation of γPKC using immunogold electron microscopy to understand how the biochemical difference in membrane binding between C1 and C2 domain could be detected morphologically, and we suggested the molecular mechanism of temporal and spatial sequences of PKC activation.

Materials and Methods

Materials

The following reagents were used in this study: A23187 (Biomol Research Laboratories; Plymouth Meeting, PA); 1,2-dioctanoylglycerol (DiC8; Calbiochem, La Jolla, CA); N-methyl-d-asparate (NMDA), 12-O-tetradecanoylphorbol-13-acetate (TPA), uridine triphosphate (UTP), and bovine serum albumin (BSA, Fraction V; Sigma, St Louis, MO); anti-γPKC polyclonal antibody (cat. #sc-211; Santa Cruz Biotechnology, Santa Cruz, CA); and glycine (Nacalai Tesque; Kyoto, Japan). All other chemicals used were analytical grade.

Methods

Cell Culture

CHO-K1 cells were purchased from Health Science Research Resources Bank (Osaka, Japan). CHO-K1 cells were cultured in Ham's F12 medium (Gibco; Grand Island, NY) at 37C in a humidified atmosphere containing 5% CO₂. The medium contained 25 mM glucose and was buffered with 44 mM NaHCO₃. It was then supplemented with 10% FBS, penicillin (100 U/ml), and streptomycin (100 μg/ml) (Gibco). FBS used was not heat inactivated.

CHO-K1 cells were seeded and cultured on glass-bottomed culture dishes (35-mm diameter; MatTech Corp., Ashland, MA) for CLSM study and on collagen-coated membranes (10-mm diameter; Cellgen, Koken, Tokyo, Japan) for electron microscopic study.

Preparation of γPKC–GFP Adenovirus

The adenoviral plasmid (pAdEasy-1) and the shuttle vector (pShuttle-CMV) were gifts from Dr. B. Vogelstein (Johns Hopkins University Oncology Center; Baltimore, MD) (He et al. 1998). Generation of an adenoviral vector including γPKC–GFP cDNA was performed according to the methods described on the web site of Johns Hopkins University Oncology Center (http//:www.coloncancer.org/adeasy/protocol.htm). Briefly, the plasmid encoding γPKC–GFP (BS 338) was digested with XhoI and XbaI, the digested products were subcloned into the XhoI and XbaI sites of the expression vector pShuttle-CMV, and the new plasmid was designated BS Ad011. Production of recombinant virus (γPKC–GFP in pAdEasy-1) was performed as described previously (Kajimoto et al. 2001).

Infection of the γPKC–GFP

CHO-K1 cells were infected with the adenoviral vectors with multiplicity of infection 10 (MOI = 10). Infected cells were cultured with virus at 37C for 20–24 hr for optimal fluorescence.

Coexpression of NMDA Receptor and γPKC–GFP in CHO-K1 Cells

For coexpression of NMDR receptor (NMDAζ1 and NMDAɛ1 subunits) and γPKC–GFP (Sakai et al. 1997), transfection into CHO-K1 cells was operated instead of the infection by adenovirus. CHO-K1 cells were transfected using 3 μl of FuGene 6 Transfection Reagent (Roche Applied Science; Indianapolis, IN) and 1 μg each of DNA (NMDAζ1, NMDAɛ1, and γPKC–GFP) according to the manufacturer's protocol for each single glass-bottomed culture dish. Transfected cells were cultured at 37C for 16–24 hr prior to imaging.

Confocal Microscopy

Before observation, culture medium was replaced with HEPES buffer (135 mM NaCl, 5.4 mM KCl, 1 mM MgCl₂, 1.8 mM CaCl₂, 5 mM HEPES, and 10 mM glucose, pH 7.3). Translocation of γPKC–GFP was induced by the addition of a concentrated solution of the activator to the HEPES to achieve the appropriate final concentration. To observe NMDA-induced translocation, NMDA was applied into the dish in the presence of 10 μM glycine and absence of MgCl₂. All experiments were done at room temperature. Fluorescence of GFP was monitored by fluorescent CLSM (Carl Zeiss; Jena, Germany) using 488-nm argon excitation with a 515-nm-long-pass barrier filter.

Immunoelectron Microscopy (Post-embedding Method)

Cells were seeded onto collagen-coated membranes at a density of 8 × 10³ cells/10 mm and cultured 18–24 hr before infection. Two days after infection, culture medium was replaced with HEPES buffer. Translocation of γPKC–GFP was induced by the addition of activators to the HEPES buffer to obtain the appropriate final concentration. All experiments were done at room temperature. After stimulation, cells on the collagen-coated membranes were directly immersed in ice-cold periodate–lysine–paraformaldehyde (4%) and fixed for 12 hr at 4C. Fixed cells were dehydrated and then flat embedded on siliconized slides in LR-Gold resin (London Resin Company Ltd.; Berkshire, UK). After UV polymerization for 3 hr at −20C, areas of interest were cut off, and 60-nm ultrathin sections were made by an ultramicrotome (Ultracut E, Reichert-Jung; Vienna, Austria). Ultrathin sections were mounted on 200-mesh Bioden Meshcement (Oken Shoji; Tokyo, Japan) -coated nickel grids (TAAB; Berkshire, UK). One side or both sides of the grids were then treated with 5% BSA for 30 min at room temperature and incubated with anti-GFP polyclonal antibody (1:40, cat. #A-11122; Molecular Probes, Eugene, OR) for 18 hr at 4C and then incubated with 10-nm colloidal gold protein A (AuroProbe EM; Amersham Life Science, Buckinghamshire, UK) for 4.5 hr at room temperature. Grids were stained with 2% uranyl acetate and Reynold's solution and then observed and photographed under a Hitachi 7100 electron microscope (Tokyo, Japan).

Quantitative Analysis of the Subcellular Localization

Subcellular distribution of γPKC–GFP immunoreactivity was quantified by counting the gold particles in defined areas of electron micrographs of the sections, the bottoms of which were immunostained (Figure 1A). Image-Pro Plus (version 4.5; MediaCybernetics, Bethesda, MD) was used for this quantitation.

Figure 1.

Quantitative analysis of PKC localization. (A) Visualization of the gap between top and bottom surface of the section. Thickness of ultrathin section (60 nm) resulted in localization of gold particles beneath the plasma membrane or along the outer edge of the plasma membrane when antibody was applied to the top or bottom surface, respectively. (B) Left: subcellular distribution of γPKC–GFP as determined by gold particles. A rectangular area (0.5 × 2.5 μm) placed vertical to and attached to the plasma membrane was selected as shown in the model. Each rectangle was divided into five squares (0.5 × 0.5 μm). The square was further derived into five rectangular areas (0.1 × 0.5 μm). Gold particles in each square were counted and results were expressed as percentage of total counts in each rectangle. Right: submembrane distribution of γPKC–GFP as determined by gold particles. Step 1: The areas were selected within 200 nm from the cell margin. Step 2: Twenty gold particles were counted and each adjacent gold particle was joined. Measurement of the joined lines was calculated together. Length of the plasma membrane was defined to respond to the distribution of the 20 gold particles. Results were expressed as ratio of the measurement/length.

For subcellular distribution, we selected several rectangular areas (0.5 μm wide and 2.5 μm long) that extended from the plasma membrane inward, perpendicular to the membrane on the micrographs. The rectangle was divided into five squares (0.5 × 0.5 μm) (Figure 1B, left). The number of gold particles in each square was determined, and the results from each square are presented as the percentage of total number of particles in the rectangular area ± standard error (SE).

For submembrane distribution, we counted 20 gold particles within the area of 200 nm from the plasma membrane on the micrographs (Figure 1, bottom). The 20 gold particles were counted, and each adjacent gold particle was linked to be a jagged line (red), and a curved line along the cell margin was made starting from the first particle to the twentieth particle (green). The length of the jagged line was divided by that of the curved line. Values (ratio of jagged line/curved line in Figure 1B) were presented as the flux of the gold particles beneath the plasma membrane ± SE. That is, if the gold particles are localized on the plasma membrane, the length of the jagged line should be almost the same as that of the curved line (cell margin).

Results

Translocation of γPKC in Living Cells Under CLSM

CLSM was used to study the localization and translocation of γPKC–GFP in CHO-K1 cells that express purinergic P2Y receptors (Iredale and Hill 1993). In resting cells, γPKC–GFP was homogeneously distributed throughout the cytoplasm with weak fluorescence in the nucleus (Figure 2, before). TPA (1 μM) induced translocation of γPKC–GFP to the plasma membrane, which was slow and irreversible (Figure 2, top row). Translocation was evident 10 min after TPA stimulation and γPKC–GFP remained on the plasma membrane for at least 30 min (data not shown). A rapid and reversible translocation of γPKC–GFP was observed when cells were stimulated with Ca²⁺ ionophore, A23187, at 5 μM (Figure 2, second row). UTP (1 mM) also induced membrane translocation of γPKC–GFP within 20 sec (Figure 2, third row) with retranslocation to the cytoplasm after 1 min. When CHO-K1 cells expressing both NMDA receptor subunits and γPKC–GFP were treated with NMDA (1 mM), similar membrane translocation of γPKC–GFP was observed and retranslocation to the cytoplasm seemed relatively earlier than in the case of A23187 or UTP (Figure 2, bottom row).

Figure 2.

Plasma membrane translocation of γPKC–GFP induced by TPA, A23187, UTP, or NMDA was followed by confocal laser-scanning microscopy. TPA treatment induced the slow and irreversible translocation of γPKC–GFP from the cytoplasm to the plasma membrane (top row). A23187 treatment induced the rapid and reversible translocation from the cytoplasm to the plasma membrane (second row). UTP induced the rapid and reversible translocation (third row). NMDA treatment to the CHO-K1 cells coexpressing NMDA receptors induced the rapid and reversible translocation of γPKC–GFP from the cytoplasm to the plasma membrane (fourth row). Bar = 10 μm.

Immunoelectron Microscopic Localization of γPKC and Its Quantitative Evaluation

Localization of γPKC–GFP was further analyzed by immunogold electron microscopic study. Distribution of the gold particles representing γPKC–GFP after those different stimuli were compared at the time when the membrane translocation was most evident: at 10 min after the stimulation for TPA and at 10 sec after the stimulation for other stimuli. Before stimulation, gold particles were evenly distributed throughout the cytoplasm and nucleus but not associated with any specific organelles (Figure 3). These results are in agreement with those obtained by CLSM. Here we used anti-GFP antibody for the immunogold electron microscopic study to trace γPKC–GFP localization. This is to prove the correlation with the GFP signals from CLSM study, but we also used anti-γPKC antibody. Localizations of the gold particles were in good agreement with those when anti-GFP antibody was used (data not shown). For the control study, we used control serum or PBS instead of the primary antibody, and no gold particles were detected in the control sections (data not shown).

Figure 3.

Immunoelectron microscopic localization of γPKC–GFP in resting CHO-K1 cells. γPKC–GFP immunoreactivity was visualized with immunogold anti-GFP antibodies. Gold particles were distributed throughout the cytoplasm and nucleus and not associated with specific organelles. Two cell samples are shown. Bar = 200 nm.

TPA treatment significantly concentrated the gold particles along the plasma membrane. When the top of the sections on a nickel grid were treated with anti-GFP antibody (Figure 1A, “top”), the gold particles lined up along the cell border and were largely localized just beneath the plasma membrane (Figure 4, “top”). In contrast, when the bottom was treated with the antibody (Figure 1A, “bottom”), gold particles outlined the cell profile (Figure 4, “bottom”). As the antibody reacts only with either surface of the section but not deep inside the 50-nm-thick sections as shown in the diagram in Figure 1A, the gap in the localization of the gold particles between the top side and bottom side is due to the thickness of the section. This finding strongly suggests that gold particles are localized at the fringe of the cells, namely, the plasma membrane. Staining both sides of the sections showing double lines of gold particles confirmed that TPA translocated γPKC–GFP onto the plasma membrane (Figure 4, “both”). We often found gold particles on apparently intracellular structures, which may be invaginations of the membrane with the connecting membrane out of the plane of the section (Figure 4, “both”). γPKC–GFP distribution in the nucleus was not altered by TPA (data not shown).

Figure 4.

Immunoelectron microscopic localization of γPKC–GFP in response to TPA. γPKC–GFP immunoreactivity 10 min after TPA stimulation (1 μM) was observed on the plasma membrane. Grids with ultrathin sections were developed with the solution of antibody with top side (top), bottom side (bottom), and both sides (both). Two cell samples are shown. Bar = 200 nm.

Similarly, distribution of gold particles was observed exactly on the plasma membrane at 10 sec after A23187 (Figures 5A and 5B), UTP (Figures 5C and 5D), and NMDA (Figures 5E and 5F). Ten sec after the treatment with A23187 or UTP, the gold particle outlined the cell surface when the bottom of the ultrathin sections was immunostained with the antibody (Figures 5A–5D). Double lines of gold particles were evident when both surfaces were stained in the sections that were fixed 10 sec after the stimulation with NMDA (Figures 5E and 5F). These results indicate that γPKC–GFP is localized on the plasma membrane 10 sec after this stimulation as seen in the case of TPA treatment. However, γPKC–GFP changed its subcellular localization at 20 sec after A23187 treatment. Although TPA induced translocation at 10 min or A23187 induced one at 10 sec that showed the localization of gold particles on the membrane, gold particles were observed in the area beneath the plasma membrane 20 sec after A23187 (5 μM) stimulation (Figure 6). In response to A23187, the particles first targeted to the plasma membrane at an earlier time (Figures 5A and 5B) and then randomly distributed within that area beneath the plasma membrane. Those distribution patterns suggested that γPKC–GFP was repeating the “touch-and-go” to the plasma membrane in the later distribution. Expression of γPKC–GFP in the nucleus was not altered by A23187 (data not shown).

Figure 5.

(A,B) Immunoelectron microscopic localization of γPKC–GFP after A23187 treatment. γPKC–GFP immunoreactivity 10 sec after A23187 stimulation (5 μM) was observed onto the plasma membrane. Grids with ultrathin sections were developed with the solution of antibody with both sides. Two cell samples are shown. (C,D) Immunoelectron microscopic localization of γPKC–GFP after UTP treatment. γPKC–GFP immunoreactivity 10 sec after UTP treatment (1 mM) was observed on the plasma membrane. Grids with ultrathin sections were developed with the solution of antibody with bottom side. (E,F) Immunoelectron microscopic localization of γPKC–GFP in CHO-K1 cells coexpressing NMDA receptors after NMDA treatment. γPKC–GFP immunoreactivity 10 sec after NMDA treatment (1 mM) was observed onto the plasma membrane. Grids with ultrathin sections were developed with the solution of antibody with both sides. Bar: A,B,E,F = 200 nm; C,D = 500 nm.

Figure 6.

Immunoelectron microscopic localization of γPKC–GFP after A23187 treatment. γPKC–GFP immunoreactivity 20 sec after A23187 stimulation (5 μM) was observed on the plasma membrane and also just beneath the plasma membrane. Grids with ultrathin sections were developed with the solution of antibody with top side or both sides. Two cell samples are shown. Bar = 200 nm.

Subcellular distribution of γPKC–GFP immunoreactivity was quantified by counting the gold particles as described in Materials and Methods. All of the quantitative analysis in the present studies was performed using sections stained at the bottom surface (Figure 1A). Quantification of subcellular localization revealed that γPKC–GFP showed a homogeneous distribution throughout all areas examined without stimulation (Figure 7A, none). After TPA treatment (1 μM) for 10 min, >70% of the gold particles were found in the square closest to the plasma membrane (0–0.5 μm from the margin of the cell) (Figure 7A, TPA), with the gold particles lined up along the edge of the cells (Figure 4). At 20 sec after A23187 treatment (5 μM), quantitative analysis showed that ∼50% of the gold particles were present within 0.5 μm of the plasma membrane (Figure 7A, A23187); this is similar to the quantitative pattern induced by TPA. This indicates that the quantitative analysis using 0.5 × 0.5-μm windows failed to show the detailed difference between localizations of the gold particles after TPA (10 min) and A23187 (20 sec) treatment. Therefore, we used smaller windows to analyze the spatial difference of γPKC–GFP concentration along the membrane after the two stimuli, and the square adjacent to the plasma membrane (0–0.5 μm) was subdivided into five rectangular areas (0.125 × 0.5 μm). Figure 7B shows that the γPKC–GFP-immunoreactive gold particles concentrated in the area closest to the plasma membrane (0–125 nm from the margin of the cells) after TPA treatment and at 10 sec after A23187 treatment, whereas at 20 sec after A23187, the particles were homogeneously distributed throughout the 0- to 500-nm rectangle. This significant quantitative difference between TPA- and A23187-induced localization corresponded with the observation under electron microscopy.

Figure 7.

Quantification of γPKC–GFP immunoreactivity in the 2.5-μm (A) or 0.5-μm (B) region from the plasma membrane. Cells were stimulated with TPA (1 μM, 10 min) and A23187 stimulation (5 μM, 10 sec, or 20 sec). Results are expressed as percentage ± SEM of the total number of gold particles in the 2.5-μm rectangle (n=5). *p<0.01 vs the distribution of γPKC–GFP immunoreactivity without stimulation.

Electron microscopy revealed that the γPKC–GFP-immunoreactive gold particles were localized exactly on the plasma membrane at 10 sec after UTP treatment (Figures 5C and 5D), also at 20 sec (Figures 8A and 8B), and even at 30 sec (data not shown). UTP stimulation never produced the pattern of gold accumulation elicited by A23187 at 20 sec at any time points (data not shown). At 10 sec after NMDA stimulation, which leads extracellular Ca²⁺ influx via the Ca²⁺-permeable ionotropic receptor, gold particles representing γPKC–GFP were observed exactly on the plasma membrane (Figures 5E and 5F; note the distinction between the outer line and the inner line of the gold particles). At 20 sec after NMDA stimulation, however, gold particles were localized in the area beneath the plasma membrane (Figures 8C and 8D) as observed 20 sec after A23187 stimulation.

Figure 8.

(A,B) Immunoelectron microscopic localization of γPKC–GFP after UTP treatment. γPKC–GFP immunoreactivity 20 sec after UTP treatment (1 mM) was observed on the plasma membrane. Grids with ultrathin sections were developed with the solution of antibody with bottom side. Two cell samples are shown. (C,D) Immunoelectron microscopic localization of γPKC–GFP in CHO-K1 cells coexpressing NMDA receptors after NMDA treatment. γPKC–GFP immunoreactivity 20 sec after NMDA treatment (1 mM) was observed on the plasma membrane and also just beneath the plasma membrane. Grids with ultrathin sections were developed with the solution of antibody with bottom side. Two cell samples are shown. Bar = 200 nm.

We further examined the submembrane distribution of the gold particles by quantifying the dispersion within the restricted submembrane area as described in Materials and Methods (Figure 1B, right). Quantitative analysis showed that the distribution patterns of the gold particles after UTP treatment at 10 sec and at 20 sec were similar to those after TPA treatment at 10 min (Figure 9), whereas the membrane localization of γPKC–GFP of “NMDA 20 sec” or “A23187 20 sec” was significantly different from that of “TPA 10 min.” However, localization of γPKC–GFP of “NMDA 20 sec” was closer to the plasma membrane than “A23187 20 sec.”

Figure 9.

Quantification of γPKC–GFP immunoreactivity in the 200-nm region from the plasma membrane. Cells were stimulated with TPA (1 μM, 10 min), A23187 (5 μM, 10 sec, and 20 sec, respectively), UTP (1 mM, 10 sec and 20 sec, respectively), and NMDA (1 mM, 10 sec, and 20 sec, respectively). Results are expressed as a ratio ± SEM of the total number of measurement/length (Figure 1B; Step 2) (n=5). Each ratio was compared by best-subset selection procedure. *p<0.01 vs the ratio at TPA stimulation.

Discussion

Live imaging of PKC translocation using GFP conjugates has provided new insight for studying PKC functions in signal transduction (Sakai et al. 1997; Oancea and Meyer 1998; Oancea et al. 1998). There are numerous reports visualizing translocations of PKC subtypes in response to various stimuli (Almholt et al. 1999; Medkova and Cho 1999; Wang et al. 1999; Feng et al. 2000; Maasch et al. 2000; Schaefer et al. 2001; Pinton et al. 2002; Gartsbein et al. 2006; Warwar et al. 2006). These studies demonstrated that PKC localizes to many intracellular compartments including the plasma membrane, Golgi complex, and microtubule (Sakai et al. 1997; Shirai et al. 1998; Kajimoto et al. 2001; Nakhost et al. 2002). Among these translocations, the translocation between cytoplasm and plasma membrane is most extensively studied by the activation of G-protein-coupled receptors, which leads to PIP2 hydrolysis as well as direct activation with phorbol ester (Oancea et al. 1998; Shirai et al. 2000a; Teruel and Meyer 2000). Ohmori et al. (1998) reported that the translocation of PKC to the membrane was in good agreement with the increase in its kinase activity, suggesting that the translocation could be used for monitoring the kinase activation of this enzyme. Monitoring GFP fluorescence under fluorescent microscopy is feasible to trace the movement of PKC in living cells, but we cannot yet determine the precise subcellular compartment where PKC is translocated. To examine the spatio-temporal association of activated PKC with the plasma membrane, it is necessary to show its precise intracellular localization. In the present study we used immunogold electron microscopy and showed two distinct translocation patterns of PKC.

Association of PKC with membrane by TPA treatment was shown by biochemical fractionation (Kraft et al. 1982) and light microscopic immunocytochemistry (Shoji et al. 1986; Ito et al. 1988). In this study we demonstrated that, after TPA treatment, gold particles of γPKC–GFP were lined along the entire plasma membrane and not seen in any specific organelles. We further examined whether the linear distribution of gold particles corresponded to the plasma membrane or to another subcellular compartment by applying anti-GFP antibody to both top and bottom surfaces of the ultrathin section. As shown in Figure 4, double-lined localization of gold particles strongly suggested that TPA treatment translocates γPKC–GFP onto the plasma membrane. This association of γPKC–GFP with the plasma membrane was also seen when treated with UTP (Figures 5C, 5D, 8A, and 8D) and in the early periods after A23187 (Figures 5A and 5B) and NMDA (Figures 5E and 5F) treatment.

Because we used anti-GFP antibody to detect γPKC–GFP, it is possible that gold particles may be localized to degradative products of γPKC–GFP. Therefore, we incubated the ultrathin sections with anti-γPKC antibody and observed the same results as seen with anti-GFP antibody (data not shown). With the use of immunoblotting analysis, anti-GFP antibody detected a single 110-kDa band corresponding to the size of GFP-tagged γPKC (Sakai et al. 1997). This result confirms that anti-GFP antibody recognizes γPKC–GFP but not degradative products.

No difference was evident under CLSM between the targeting sites by any stimulation, but electron microscopic observation enabled the distinct distribution patterns of γPKC–GFP to be demonstrated. It is noteworthy that Ca²⁺ ionophore rapidly translocated γPKC–GFP to the plasma membrane and then to the different sites from TPA; γPKC–GFP accumulated randomly in the area beneath the plasma membrane. Because the resolution of GFP fluorescence microscopy with objective lens of 1.4 numerical apertures is calculated to be 0.2 μm, quantitative analysis using 0.5 × 0.5-μm boxes mimicked the light microscopic observation, and this rough quantification failed to discriminate Ca²⁺-induced translocation from TPA-induced translocation. Detailed quantitative analysis at submembrane area (Figure 7B and Figure 9), however, demonstrated that the dispersion of gold particles at 20 sec after Ca²⁺ ionophore treatment is significantly larger than after TPA treatment (p<0.01) and also larger than at 10 sec after Ca²⁺ ionophore treatment (p<0.01). This finding suggests a temporal translocation of γPKC.

Micrographs shown in Figure 6 may show a moment when gold particles are diffusing from the plasma membrane. However, as the mean lateral displacement of single PKC molecules is as fast as 5.4–7.1 μm/s² (Schaefer et al. 2001,2004; Lenz et al. 2002; Sinnecker and Schaefer 2003), it is unlikely that gold particles accumulate accidentally within the restricted area. However, it is more reasonable to define that γPKC is repeating association and dissociation with the plasma membrane in the restricted area than to expect that there may be unknown molecules that anchor γPKC to the soluble area for one reason or another.

PKC first associates with the plasma membrane via its C2 domain with Ca²⁺ (Newton 1997; Oancea and Meyer 1998; Johnson et al. 2000; Cho 2001; Raghunath et al. 2003). This association can be moderately weak without the association of C1 domain with DG. Because the duration of the association of C2 domain with the plasma membrane varies depending on the free Ca²⁺ level (Nalefski and Newton 2001), this weak association of C2 domain with the plasma membrane may easily cause the association and dissociation of the C2 domain with/from the plasma membrane when the free Ca²⁺ level changes.

It has been reported that influx of extracellular Ca²⁺ results in Ca²⁺ increase beneath the plasma membrane, which is known as “Ca²⁺ gradients” or “Ca²⁺ spark” in various cells (O'Sullivan et al. 1989; ZhuGe et al. 1999; Quesada et al. 2000; Berridge 2006). The Ca²⁺ increase in the localized area is considered to be induced by the opening of clustered ryanodine-sensitive Ca²⁺ release channels in the sarcoplasmic reticulum in smooth muscle cells (Bolton and Imaizumi 1996; ZhuGe et al. 1999; Wang et al. 2004) or by opening the IP3 channels and releasing Ca²⁺ from the intracellular Ca²⁺ store with the mechanism of Ca²⁺-induced Ca²⁺ release in epithelial cells (Endo et al. 1970; Berridge and Dupont 1994; Braiman and Priel 2001) and in mouse oocytes (Endo et al. 1970; Berridge and Dupont 1994; Braiman and Priel 2001). In pancreatic islet cells, a high concentration of glucose that leads insulin secretion induces a transient translocation of αPKC (Deeney et al. 1996) or βIIPKC (Pinton et al. 2002) to the plasma membrane. In these cells, a sustained Ca²⁺ gradient (∼6.74 μM concentration) restricted to the domains within 670 nm beneath the plasma membrane was observed. Mathematical approach revealed that Ca²⁺ concentration reached ∼8.18 μM at a distance of 250 nm from the plasma membrane and reached ∼10.05 μM at a distance of 0 nm from the plasma membrane (i.e., at the plasma membrane) (Quesada et al. 2000). It was also reported recently that in smooth muscle cells, Ca²⁺ microdomains are created ∼200 nm near the channel, i.e., plasma membrane (McCarron et al. 2006). This is in good agreement with the present finding that Ca²⁺ ionophore treatment resulted in accumulation of γPKC–GFP within the 200-nm area beneath the plasma membrane (Figures 4A and 4B). Because such a distance of 250 nm at submembrane area produces different concentrations of Ca²⁺ (Quesada et al. 2000), in the present study it can be considered that association of PKC with the plasma membrane can be altered with a minor change in Ca²⁺ in the 200-nm area. Single-molecule video-imaging technique may support the results obtained in the present study. When analyzed at the single-molecule level, one single γPKC molecule translocated to the plasma membrane by TPA and remained on the plasma membrane >10 sec. By Ca²⁺ ionophore treatment, one single γPKC localized on the plasma membrane in a moment (<1 sec) and released from the membrane (Saito and Kusumi, unpublished data), although CLSM showed sustained localization of γPKC on the membrane. This suggests that γPKC is not directly associated with the membrane but localized around the membrane, which cannot be distinguished by light microscopy.

Sakai et al. (1997) showed that TPA-induced slow translocation of γPKC–GFP was observed in the absence of Ca²⁺ and also that γPKC–GFP lacking C2 domain was translocated by TPA. These findings indicate that TPA-induced translocation does not require the Ca²⁺ binding to the C2 domain of γPKC–GFP. As the diffusion of GFP-tagged PKCs was estimated to be ∼6 μm/s² (Schaefer et al. 2001), γPKC–GFP should collide with TPA on the plasma membrane ∼1–3 times/sec. Therefore, it is suggested that the slow translocation by TPA is due to low affinity of the cytoplasmic γPKC–GFP to TPA. Due to the low affinity to TPA, γPKC–GFP in a resting state slowly binds TPA on the plasma membrane, but the increase in cytoplasmic Ca²⁺ changes γPKC–GFP into high-affinity form to TPA and then rapidly translocates γPKC–GFP onto the plasma membrane. These findings support the idea that γPKC requires Ca²⁺ priming and following translocation to become accessible to DG (Oancea and Meyer 1998; Bartlett et al. 2005; Craske et al. 2005; Codazzi et al. 2006). Despite the generation of DG on the plasma membrane, UTP induced reversible translocations in contrast to TPA-induced irreversible translocation. This may be due to the rapid metabolism of DG by endogenous DG kinase (DGK) (Topham and Prescott 1999), whereas TPA cannot be metabolized by the same enzymes. In fact, DGK inhibitor delayed the retranslocation of γPKC–GFP from the plasma membrane to the cytoplasm (Shirai et al. 2000b; Yamaguchi et al. 2006).

In conclusion, immunogold electron microscopic studies determined the precise translocation of γPKC at submembrane area, which cannot be observed under light microscopy, and first revealed the low-affinity binding of C2 domain with the plasma membrane morphologically.