Insulin Induces the Translocation of GLUT4 From a Unique Intracellular Organelle to Transverse Tubules in Rat Skeletal Muscle ANDRE MARETTE, ELENA BURDETT, ANDRE DOUEN, MLADEN VRANIC, AND AMIRA KLIP
Skeletal muscle surface membrane is constituted by the PM domain and its specialized deep invaginations known as TTs. We have shown previously that insulin induces a rapid translocation of GLUT4s from an IM pool to the PM in rat skeletal muscle (6). In this study, we have investigated the possibility that insulin also stimulates the translocation of GLUT4 proteins to TTs, which constitute the largest area of the cell surface envelope. PM, TTs, and IM components of control and insulinized skeletal muscle were isolated by subcellular fractionation. The TTs then were purified further by removing vesicles of SR origin by using a Ca-loading procedure. Ca-loading resulted in a five- to sevenfold increase in the purification of TTs in the unloaded fraction relative to the loaded fraction, assessed by immunoblotting with an anti-OHP-receptor monoclonal antibody. In contrast, estimation of the content of Ca 2+ -ATPase protein (a marker of SR) with a specific polyclonal antibody revealed that most, if not all, SR vesicles were recovered in the Ca-loaded fraction. Western blotting with an anti-COOH-terminal GLUT4 protein polyclonal antibody revealed that acute insulin injection in vivo (30 min) increased the content of GLUT4 (by 90%) in isolated PMs and markedly enhanced (by 180%) GLUT4 content in purified TTs. Importantly, these insulin-dependent changes in GLUT4 content of PM and purified TTs were seen in the absence of changes
From the Division of Cell Biology, The Hospital for Sick Children, Toronto, Ontario; and the Department of Physiology, University of Toronto, Toronto, Ontario, Canada. Address correspondence and reprint requests to Amira Klip, PhD, Division of Cell Biology, The Hospital for Sick Children, 555 University Avenue, Toronto, Ontario M5G 1X8, Canada. Received for publication 31 January 1992 and accepted in revised form 4 June 1992. GLUT, glucose transporter; TT, transverse tubule; cTT, crude transverse tubule; PM, plasma membrane; IM, intracellular membrane; CM, crude membrane; SR, sarcoplasmic reticulum; LSR, light sarcoplasmic reticulum; HSR, heavy sarcoplasmic reticulum; DHP, dihydropyridine; BSA, bovine serum albumin; PMSF, phenylmethylsulfonyl fluoride; CLV, calcium-loaded vesicle; SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis; PVDF, polyvinylidene difluoride; type II diabetes, non-insulin-dependent diabetes mellitus.
1562
in the o^-subunit of the Na + -K + -ATPase, a surface membrane marker. Isolated IM components such as LSR, HSR, and triads (terminal cisternae plus junctional TT) contained low or barely detectable amounts of GLUT4; furthermore, insulin treatment did not change the distribution of the transporter protein within these fractions. In contrast, a unique IM fraction that was not associated with either SR or triad markers, contained significant amounts of GLUT4 and showed an insulin-dependent decrease (40%) in GLUT4 protein content. These results show that acute insulin treatment induces the translocation of GLUT4s to both the PM and TTs from a unique intracellular organelle not associated with the SR. Diabetes 41:1562-69, 1992
I
t is generally accepted that glucose transport across the cell surface is the rate-limiting step for glucose utilization in skeletal muscle (1,2). This metabolic process is markedly stimulated by insulin, and, in the postabsorptive state, skeletal muscle represents the predominant site of whole-body glucose disposal. Glucose transport into muscle cells is mediated by membrane surface GLUT proteins. Two isoforms of the GLUT family have been identified so far in adult rat skeletal muscle: the ubiquitously expressed GLUT1, and the GLUT4 which is present exclusively in insulin-sensitive tissues such as skeletal muscle, heart, and white and brown adipose tissues (3,4). Previous studies in adipose cells have shown that acute insulin exposure stimulates the translocation of GLUT4s from an IM pool to the PM, resulting in an increase in glucose transport activity (5). Other studies also showed an insulin-dependent mobilization of GLUT4s from an intracellular pool to the PM in rat skeletal muscle (6,7). Although insulin also translocates GLUT1 in adipose cells (5), this effect is marginal and could not be observed in insulinized skeletal muscles (6). Therefore, it is recognized that recruitment of GLUT4 proteins is the predominant mechanism by which
DIABETES, VOL. 41, DECEMBER 1992
A. MARETTE AND ASSOCIATES
insulin increases glucose transport activity in these tissues, together with a possible activation of GLUT intrinsic activity by the hormone (8). The insulin-induced membrane redistribution of GLUT4s in muscle is in many respects similar to that of adipose cells. However, important functional and morphological differences have been noted between skeletal muscle and fat. Because of their contractile properties, skeletal muscle cells have developed highly specialized organelles that are involved in excitation-contraction coupling. In addition to the PM, muscle fibers are characterized by the presence of TTs (deep invaginations of the PM that interact with the terminal cistemae of the SR). The TTs are involved in propagating the action potential within the muscle cells where depolarization triggers the release of Ca 2+ from SR, leading to fiber contraction. These tubular membranes are exposed to the extracellular milieu and therefore may play an important role in the transport of ions and nutrients such as glucose into the muscle fiber. Moreover, TT membranes possess insulin receptors (9), suggesting that these organelles are involved in the regulation of glucose transport and metabolism in muscle. We have reported previously that an isolated membrane fraction enriched in TTs contained five times more cytochalasin-B binding sites (which measure total GLUT content irrespective of isoform) compared with isolated PMs of rat skeletal muscle. This preferential segregation of the transporter protein may have potential consequences for the functional response of glucose transport to insulin in skeletal muscle (9). Recently, Friedman et al. (10) reported that GLUT4s are concentrated in triad membranes (where SR apposes the TTs), as detected by immunoelectron microscopy of human muscle sections. They further suggested that insulin induces the movement of GLUT4 from internal sites (possibly SR) to the triad-TT membranes but not to the PM. In contrast, we have reported previously the isolation of PM and IM fractions from rat skeletal muscle that were devoid of TT, triad, or SR markers (11). Yet, we observed an insulindependent translocation of GLUT4 from the IMs to the PMs, demonstrating that GLUT4 can translocate from non-SR and nontriad membranes to surface membranes not containing TTs. Biochemical evidence for an insulininduced mobilization of GLUT4 to TT membranes in muscle is, however, still lacking. The objectives of this study were to determine the following: 1) the effect of acute insulin treatment on the GLUT4 protein content in isolated TTs; and 2) the relative amounts of GLUT4 in various skeletal muscle intracellular organelles suspected to be the site of GLUT4 storage for insulin-dependent mobilization in muscle. The results indicate that insulin induces the translocation of GLUT4s to both the PM and the TTs of skeletal muscle. The original site of the transporters is neither the SR nor the triad elements, but rather a unique insulin-sensitive GLUT4-enriched intracellular organelle. RESEARCH DESIGN AND METHODS Fatty-acid free BSA and gradient grade sucrose were obtained from Sigma (St. Louis, MO). The polyclonal
DIABETES, VOL. 41, DECEMBER 1992
GLUT4-specific antibody and the polyclonal GLUT1specific antibody were obtained from East Acres Biologicals (Southbridge, MA). A monoclonal antibody against the a r subunit of the Na+-K+-ATPase was a gift from Dr. K. Sweadner (Harvard, Boston, MA). The anticalsequestrin monoclonal antibody and the anti-Ca2+-ATPase polyclonal antibody were gifts of Dr. D. MacLennan (Banting and Best Department of Medical Research, University of Toronto). Antibodies to the DHP receptor (IIC12) and to the ryanodine receptor were supplied by Dr. K. Campbell (University of Iowa, Aimes) and Dr. Judith Airey (University of Nevada), respectively. In vivo insulin injection. Male Sprague-Dawley rats weighing 280-320 g were fasted overnight before insulin injection. Rats were anesthetized by intraperitoneal injection of Ketamine-Acepromazin. Insulin (Humulin R, 1.5 U/rat) or saline then was given by tail-vein injection. After 30 min, the rats were killed, and back and hindlimb white muscles were rapidly excised. The muscles were either used immediately or rapidly frozen and kept at -80°C until processed for membrane preparation. Blood was collected before and after insulin injection. The decrease in glycemia (>50%) after insulin administration was used as a positive index of insulin action. Membrane preparation and characterization. PMs and IMs were isolated as described previously and characterized (11-13). Briefly, 10-15 g of frozen muscle first was cleaned of all visible fat, nerve, and vessels. Then they were minced in buffer I (10 mM NaHCO3, 0.25 M sucrose), containing 5 mM NaN3 and 100 jxM PMSF, and homogenized in a Polytron (setting 5) for 5 s. The resulting homogenate was subjected to a series of differential centrifugation steps to yield a CM pellet. The CMs were applied on discontinuous sucrose gradients (25, 30, 35%, w/w) and centrifuged for 16 h at 150,000 g (12). Membranes were collected from each sucrose layer and stored at -80°C until used for Western blots. Extensive characterization of the isolated membranes was given elsewhere (11-13). Briefly, the 25% sucrose fraction contains measurable amounts of GLUTs (GLUT1 and GLUT4) and is enriched in PM markers, such as 5'nucleotidase and phosphodiesterase activity, and content of Na+-K+-ATPase c^-subunit, all of which are increased five- to sixfold relative to nonfractionated CMs. Moreover, the Na+-K+-ATPase activity of the 25% sucrose fraction is 15-fold higher than that of muscle homogenates (14). Conversely, the 35% sucrose fraction is enriched in GLUT4s only, but is depleted of the PM markers discussed above. This fraction therefore is referred to as a GLUT4-enriched IM fraction. A TTs-enriched fraction and a triad-enriched fraction (i.e., terminal cistemae and junctional tubules) were prepared from fresh leg and back muscles as described previously (9,11). When indicated, TT membranes were further purified by using a modification of the Ca-loading procedure previously reported for the isolation of rabbit TTps (15,16). This method allows the removal of contaminating SR vesicles from the TT fractions, based on the SR selective permeability to phosphate ions. Briefly, isolated TTs were incubated at a protein concentration of —0.1-0.2 mg/ml in a solution containing 50 mM K-phos-
1563
GLUT4 IN MUSCLE TRANSVERSE TUBULES
106 kda 80kda 49.5 kda 32.5 kda 27.5 kda Std LSR HSR TT CM PM IM FIG. 1. Protein composition of various membrane components isolated from rat skeletal muscle. Membrane proteins (50 y.g) were subjected to SDS-PAGE on 9% polyacrylamlde mini-gels. The gels were stained with 0.1% Coomassie brilliant blue, destalned, and dried as described In METHODS. Arrowheads point to protein bands that were observed In PM but absent or barely detected in TT membranes.
phate, pH 7.4, 5 mM MgCI2, 0.15 M KCI, 0.3 mM CaCI2) and 2 mM Tris-ATP (Ca-loading solution). The membranes were incubated for 30 min at 22°C and Ca loading was optimized by further additions of 0.3 mM CaCI2 and 2 mM ATP at 5,10, and 20 min of incubation. Membranes then were collected by centrifugation at 150,000 gffor 1 h and resuspended in 2 ml of Ca-loading solution. The Ca-treated membranes were layered on top of discontinuous density gradients of 35 and 50% (w/w) sucrose in Ca-loading solution and centrifuged for 90 min at 150,000 g in an SW 40 Ti rotor. Ca- phosphate complexes precipitate in SR vesicles because they are permeable to phosphate and possess high Ca-uptake rates (in contrast to TT vesicles). Ca loading resulted in the separation of two protein fractions: a Ca-loaded fraction (at the 35/50% sucrose layer interphase) that was found to contain significant amounts of SR vesicles (CLV fraction) and a lighter fraction (top of the 35% sucrose layer) that contained purified TT fraction (RESULTS). LSR and HSR vesicles were prepared from rat skeletal muscle as described previously for rabbit skeletal muscle (17). Protein was determined by the method of Lowry et al. (18). SDS-PAGE. The protein composition of the various fractions was studied by SDS-PAGE in 9% mini-gel slabs as described by Laemmli (19). Gels were stained with 0.1% Coomassie brilliant blue in 5:4:1 methanol/H2O/acetic acid, destained in the same methanol/H2O/acetic acid solution, and dried for photography. Western blot analysis. Membranes (20 \xg of protein) were subjected to SDS-PAGE on 9-10% polyacrylamide gels according to Laemmli (19) and electrophoretically transferred to PVDF filter membranes for 2 h. PVDF
1564
membranes were incubated for 1 h at room temperature with Tris-saline (50 mM Tris-HCI, pH 7.4, 150 mM NaCI) containing 0.04% of the nonionic surfactant Nonidet P 40 and 3% BSA (buffer A), followed by overnight incubation at 4°C with antibodies in buffer A. The PVDF membranes then were washed three times in buffer A (without BSA) followed by a 1-h incubation with either 125l-labeled protein A (2 |xCi/10 ml) or 125l-labeled sheep anti-mouse immunoglobulin G (2 |xCi/10 ml) in buffer A (Fig. legends), and washed three times in buffer A (without BSA), air-dried, and exposed to XAR-5 Kodak film for 12-48 h. Autoradiographs were quantitated by laser-scanning densitometry with an LKB Bromma ultrascan XL enhancer laser scanner with on-line analysis by a Packard Bell computer. The colorimetric alkaline phosphatase procedure was used for detection of the ryanodinereceptor antibody.
RESULTS
The protein composition of the different membranous organelles isolated from rat skeletal muscles is shown in Fig. 1. Equal protein amounts were loaded for each fraction. LSR and HSR fractions share essentially the SDS-PAGE pattern previously reported for SR (20). A band corresponding to a molecular size of -106,000 Mr (106 kDa) represented 80-90% of the protein content of both the LSR and HSR and is likely to correspond to the Ca2+-ATPase of SR (Figs. 3 and 4). In the HSR fraction, a protein of -65,000 Mr (65kDa) was also abundant and probably corresponds to calsequestrin, a Ca-binding protein known to be located in the cisternal portion of SR (21). The TT vesicles were characterized by a more
DIABETES, VOL. 41, DECEMBER 1992
A. MARETTE AND ASSOCIATES
CM C
PM
I C
IM
I C I
106 kd
a - Na7 K+-ATPase
49.5 kd
GLUT4 1 2
3 4 5 6
FIG. 2. Effects of acute insulin treatment on the distribution of GLUT4 In CM, PM, and IM isolated from rat white hindiimb muscles. « 1 -Na + -K + -ATPase (-100,000 Mr [100 kDa] protein) content In the same membrane samples Is also shown as a control. Membrane proteins (20 jig) were subjected to SDS-PAGE on 10% polyacrylamide mini-gels and transferred to PVDF membranes. PVDF membranes were incubated for 1 h at room temperature with Trls-saline (50 mM TrIs-HCI, pH 7.4,150 mM NaCI) containing 0.04% of the nonlonic surfactant Nonldet P 40 and 3% BSA, followed by overnight Incubation at 4°C with either anti-GLUT4 (1:500 dilution) or antl-« 1 -Na + -K + -ATPase (1:100 dilution) antibodies In buffer A. The PVDF membranes then were washed three times In buffer A (without BSA), followed by a 1-h Incubation with either 125l-labeled protein A (2 p.Ci/10 ml) or 125l-labeled sheep anti-mouse immunoglobulin G (2 JJLCI/10 ml) in buffer A, washed three times in buffer A (without BSA), air-dried, and exposed for radloautography.
complex protein composition also typical of a previously reported pattern (15), with prominent polypeptides of 106,000; 68,000; 53,000; 39,000; and 30,000 Mr (106, 68, 53, 39, and 30 kDa). A large number of proteins was seen in CMs, PMs, and IMs of skeletal muscles. These
fractions contained multiple proteins in the 30,000 to 80,000 Mr (30-80 kDa) range, most of which were only marginally represented in the LSR, HSR, and TT fractions. Importantly, the protein composition of the PMs and the IMs was significantly different. Furthermore, several
TT LSR HSR IM C I C I C I C I
205 kda DHP receptor
116 kda
2+
Ca ATPase 49.5 kda
GLUT4 12
3
4
5
6
7 8
FIG. 3. Effects of insulin on the distribution of GLUT4 In cTT, LSR, HSR, and IM Isolated from rat white hindiimb muscles. The specific protein markers for TT (DHP receptor) and SR (Ca2+-ATPase) also are depicted. Membrane proteins (20 fig) were subjected to SDS-PAGE on 9% polyacrylamide mini-gels and transferred to PVDF membranes as described in Fig. 2, except that PVDF filters also were incubated overnight with antl-DHP receptor (1:300 dilution) and anti-Ca 2+ -ATPase (1:100 dilution) antibodies.
DIABETES, VOL. 41, DECEMBER 1992
1565
GLUT4 IN MUSCLE TRANSVERSE TUBULES
proteins of the isolated PMs (Fig. 1, -35,000-40,000 and 80,000 Mr [35-40 and 80 kDa]) were either absent or barely detectable in TTs, suggesting that the tubular elements did not contain vesicles of PM origin. We next investigated the effects of insulin on the membrane distribution of immunodetectable GLUT4s in equal protein samples of the different membranous organelles isolated. The GLUT4 is a protein that migrates with a size corresponding to -45,000 Mr (45 kDa) on polyacrylamide gels. In three individual experiments, insulin did not change the distribution of GLUT4 in unfractionated CMs isolated from white skeletal muscles (Fig. 2, lower band, lanes 1 and 2). However, insulin increased the GLUT4 protein content by 86 ± 12% in purified PMs (lanes 3 and 4) and induced a concomitant 40 ± 12% decrease in the transporter content in the IMs of the same muscle samples (lanes 5 and 6). The a r subunit of the Na+-K+-ATPase (Fig. 2, upper band), a PM marker, was detected in the CMs and was enriched five- to sixfold in the PMs. This protein, however, was absent from the IM pool. When several blots were scanned for densitometry, insulin treatment was found not to alter the content of the enzyme subunit in any membrane fractions, highlighting the GLUT4 proteinspecific translocation in response to the hormone. Note that antibodies to both the DHP receptor (a TT marker) and the ryanodine receptor (a triad marker) did not react with CMs, PMs, or IMs, confirming a previous report that these cellular components were not contaminated with TTs (Fig. 4) (11). The TT membranes were found to contain high amounts of the GLUT4s (Fig. 3, lower band, lanes 1 and 2). As expected, these vesicles were characterized by the presence of the DHP receptor (Fig. 3, upper band, -180,000 H [180 kDa]). However, the TT membranes also contained significant amounts of SR Ca2+-ATPase, a protein migrating at -106,000 Mr (106 kDa) on polyacrylamide gels (Fig. 3, middle band), indicating contamination of the tubular elements with vesicles of SR origin. In these TTs, insulin had no consistent effect on GLUT4 content. The absence of reproducibility of insulin action on TT GLUT4 protein content may have reflected the degree of contamination with SR because the Ca 2+ ATPase content varied from one preparation to another. Low amounts of the GLUT4 protein were detected in the LSR and HSR fractions (Fig. 3, lanes 3-6). The identity of these membranes was confirmed by the abundance of the SR Ca2+-ATPase and by the low amounts of the TT-DHP receptor. In two membrane SR vesicle preparations, insulin had no effect on the distribution of the GLUT4 protein. Conversely, insulin significantly reduced the GLUT4 content of IMs (Fig. 3, lanes 7 and 8) isolated from the same muscles. These vesicles were free of TT-DHP receptor and relatively depleted of SR Ca 2+ ATPase, supporting the tenet that these membranous organelles represent a unique intracellular pool of GLUT4s occluded to the extracellular milieu. It must be emphasized that all fractions analyzed in Fig. 3 were loaded with equal protein values and were run on the same gel. GLUT4 protein was barely detectable in isolated triad
1566
fractions (Fig. 4). When autoradiograms were exposed for longer periods, no significant changes in GLUT4 contents were seen in triads prepared from insulintreated rats (data not shown). The triad membranes, however, were found to contain significant amounts of calsequestrin and ryanodine receptor (Fig. 4), as expected from membranes originating from a TT-HSR complex. Conversely, the levels of SR Ca2+-ATPase were extremely low in this fraction (data not shown), in agreement with a previous study showing a predominant localization of this protein in the LSR (21). From the Western blot depicted in Fig. 3, it is clear that the TTs were significantly contaminated with vesicles of SR origin. Although the relative concentration of GLUT4 is small in SR membranes, large amounts of the intracellular organelle may still mask insulin-induced changes in GLUT4 distribution in the tubular elements, especially if the level of contamination by SR is different in control and insulin-treated muscle vesicles. Ca loading, therefore, was used to further purify TTs from SR vesicles. Figure 5 shows that vesicles that could not be loaded with Ca
Tr
TT
49.5 kda GLUT-4
Tr
TT
Ryanodine Receptor
Tr
Calsequestrin
m
TT
80 kda
49.5 kda
FIG. 4. Comparison of GLUT4 content In TT and triads Isolated from rat white hlndllmb muscles. Markers for triads (ryanodine receptor) and HSR (calsequestrin) also are shown. Membrane proteins (50 n.g) were subjected to SDS-PAGE on 10% (GLUT4 and calsequestrin) or 5% (ryanodine receptor) polyacrylamide mlnl-gels and transferred to PVDF membranes as described In Fig. 2, except that PVDF filters also were Incubated overnight with antlryanodine receptor and anticalsequestrin antibodies. Mr standards are shown on the right. The M, of the reacting doublet of the ryanodine receptor was higher than 500,000.
DIABETES, VOL. 41, DECEMBER 1992
AND ASSOCIAILS
cTT CLV ITp C
I C I C I
205 kda DHP receptor 116 kda
Ca2+ATPa®e
49.5 kda 1
2 3
4
5
6
FIG. 5. Purification of TT by loading the contaminating SR vesicles with Ca phosphate. Control (C) and Insuilnlzed (I) muscle membranes were Isolated and purified as described In METHODS. Membranes proteins (10 p.g) then were analyzed by Western blotting as described In the legend of Fig. 3. The blot Is representative of four Western blots from four separated loading experiments. Mr markers are shown on the left
(TTp, lanes 5 and 6) contained significant amounts of DHP receptor, whereas the CLV (lanes 3 and 4) were depleted of the TT marker. Integration of scans from four gels revealed that the DHP receptor protein was enriched five to seven times in the TTp fraction as compared to the CLV fraction (Fig. 6). In contrast, marginal amounts of the SR Ca2+-ATPase (Fig. 5, middle band; and Fig. 6, middle panel) were detected in TTp, but high levels of the enzyme were found in the Ca2+-loaded CLV. A four- to sixfold reduction in the Ca2+-ATPase content was observed in the TTp fraction as compared to the CLV fraction (Fig. 6). Thus, the TTp represent vesicles of TT origin cleared from SR vesicles, the latter being recovered in the loaded CLV fraction. Note that TTp and CLV vesicles could not be separated when cTT were incubated in the absence of ATP during the loading procedure, underlining the necessity of the ATP-dependent loading for TT purification. Insulin caused a marked increase in GLUT4 content in TTp membranes (Fig. 5, lower band, lane 6 vs. lane 5). Scanning of multiple gels by densitometry revealed that insulin treatment increased GLUT4 protein by almost threefold (180%) in TTp (Fig. 6). This increase was observed despite an absence of change in the TT content of the a r subunit of the Na+-K+-ATPase after insulin treatment (data not shown). No consistent effect of insulin was seen in the CLV vesicles. In four separate Ca2+-loading experiments, the distribution of GLUT4 in the CLV was found to be either augmented (Fig. 5) or diminished, the overall averaged effect of insulin being negligible (Fig. 6). Note that the CLV fraction probably contained GLUT4 of non-SR origin (possibly of GLUT4-
DIABETES, VOL. 41, DECEMBER 1992
enriched intracellular organelle origin), because its GLUT4 content was found to be higher than that observed in isolated LSR and HSR (Fig. 3) when compared with gels analyzed in parallel. One explanation for that observation is that the GLUT4-containing organelle (possibly the GLUT4-enriched IM discussed above) contains a different form of Ca 2+ - pumping activity (perhaps a specific Ca2+-ATPase isoform), explaining its recovery in the CLV fraction after Ca 2+ loading. This possibility, however, remains to be tested experimentally.
DISCUSSION The subcellular localization of GLUTs in skeletal muscle has been the subject of intense recent investigation. Because of the unique intracellular sequestration of the GLUT4 isoform and the known insulin sensitivity of this transporter in muscles, knowledge of its membrane distribution is of particular importance. Furthermore, considering the studies reporting that the total number of GLUTs is not reduced in skeletal muscle homogenates of type II diabetic subjects (22,23), it has become crucial to analyze the GLUT4 content of fractionated subcellular surface membranes and I Ms with the hope that this will lead to unraveling the nature of the defect in glucose transport of diabetic muscle. This study identified both the plasmalemmal and TT membranes as surface components containing GLUT4s that gain transporters upon insulin stimulation. These results suggest that both domains of the muscle cell surface are involved in the transport of glucose into muscle, and that this occurs, at least in part, through
1567
GLUT4 IN MUSCLE TRANSVERSE TUBULES
GLUT1 is localized to the plasma membrane of skeletal) (29). This supports the notion that insulin induces the recruitment of GLUT4 to these two specific cell surface components. Third, translocation of GLUT4 to heart TTs has been recently observed using immunoelectron microscopy (24). Clearly, a surface compartmentation of GLUT4s in muscle would be of physiological importance. We have previously reported that both D-glucose-protectable binding of cytochalasin-B and GLUT4 protein content are decreased in PMs isolated from streptozocin-induced diabetic rat muscle (25,26). Whether downregulation of GLUT4 also occurs in diabetic rat TT membranes is still unknown. It will be of great interest to assess whether the reduced GLUT4 content in diabetes is manifested in both regions of the cell surface or, conversely, whether it is confined to a specific domain. Hence, considering that TTs represent the major fraction (-70%) of total surface membranes in frog skeletal muscle (28), a preferential depletion of GLUT4s in that compartment would be of obvious metabolic consequence. Another important aspect of this study was to determine the intracellular locations of the occluded GLUT4s that are translocated by insulin. These transporters are not recruited from membranes associated with LSR, HSR, or triads. In fact, the latter three intracellular organelles were found to be either poor or devoid of the GLUT4 protein. These data are consistent with our previous observation of very low cytochalasin-B binding in isolated SR vesicles (9). They are, however, in contrast with the immunoelectron microscopy findings of Friedman et al. (10), who previously reported GLUT4 immulolabeling within the triad region and in an intracellular compartment, possibly SR. However, as stated by the authors of the latter study, the fixation protocol used to preserve GLUT4 antigenicity precluded to ascribe definitively the transporter protein to the SR. The fractionation experiments of our study indicate that intracellular GLUT4s are localized in an intracellular compartment that is associated neither with SR nor with triads in skeletal muscle. Rather, GLUT4s appear to be predominantly localized in CLV TTp cTT a unique, insulin-sensitive membranous intracellular orFRACTIONS ganelle. As shown in Table 1, this membrane pool is devoid of typical PM markers (5'-nucleotidase and a-,FIG. 6. Densttometry data showing purification of TT by the Na+-K+-ATPase) and is depleted of markers of both the Ca-loading experiments represented in Fig. 5. Asterisk represents a significant difference (P < 0.02) from control purified (TTp) compared SR (calsequestrin, ryanodine receptor, Ca2+-ATPase) by Student's f test for unpaired data. Scanning data are expressed and triads (ryanodine receptor), underlying the unique relative to control cTT protein contents and represent means ± SE of 4 nature of its origin. Taken together with the recent immuWestern blots from four separate experiments. noelectron microscopy observation of an intracellular GLUT4-enriched tubulo-vesicular network in rat adipoinsulin-dependent regulation of GLUT4 protein content. cytes and heart (24,27), these findings strongly supThe latter suggestion is supported by the following ob- port the existence of a distinct, as yet uncharacterized servations: First, we have observed previously colocal- organelle endowed with GLUT4s. A more complete charization of GLUTs and insulin receptors in TTs (9). Second, acterization of this GLUT4-enriched intracellular orthe isolated PMs are not contaminated with TT mem- ganelle is underway. Preliminary observations (A.M., branes or vice versa, as depicted in Table 1. Regarding A.K., unpublished observations) suggest that this pool is this observation, note that the TTs are totally devoid of the not associated with the endoplasmic reticulum because it GLUT1 isoform (Table 1), whereas this protein is rela- is devoid of calreticulin, a Ca-binding protein known to tively abundant in PMs, strongly suggesting that the two be associated with this cellular organelle. An attractive surface compartments are significantly separated by our possibility is that these IMs are of endosomal origin, as fractionation procedures (We recently observed that suggested by the elegant studies of Slot et al. (24,27),
• •
1568
CONTROL INSULIN
DIABETES, VOL 41, DECEMBER 1992
A. MARETTE AND ASSOCIATES
TABLE 1 Characterization of membrane fractions isolated from rat hindlimb skeletal muscle 25% sucrose (PM)
35% sucrose (IM)
Purified TTp
Longitudinal SR
Triads
Markers GLUT4 GLUT1 a r Na + -K + -ATPase (PM marker) DHP receptor (TT marker) Ryanodine receptor (cisternal SR marker) Calsequestrin (cisternal SR marker) Ca2+-ATPase (longitudinal SR marker)
Symbols (+ and -) denote relative enrichments for equal amounts of protein in the indicated membrane fractions.
which showed that GLUT4 undergoes recycling through the coated pits-early endosome pathway in adipose cells and heart. In summary, the results of this study show that insulin increases the content of GLUT4s in both the PMs and TT membranes isolated from rat skeletal muscle. The GLUT4s appear to be recruited from an intracellular storage site not associated with either SR or triad but rather from a unique, insulin-sensitive intracellular organelle. We suggest that TTs play an important role in the regulation of glucose transport and metabolism in skeletal muscle. ACKNOWLEDGMENTS
This work was supported by a grant to A.K. from the Medical Research Council of Canada. A.M. is the recipient of a postdoctoral fellowship from the Medical Research Council of Canada. We thank Toolsie Ramlal for excellent technical assistance in some of the experiments described in this study. REFERENCES 1. Simpson IA, Cushman SW: Hormonal regulation of mammalian glucose transport. Annu Rev Biochem 55:1059-89, 1986 2. Klip A, Paquet M: Glucose transport and glucose transporters in muscle and their metabolic regulation. Diabetes Care 13:228-43, 1990 3. Bell Gl, Kayano T, Buse JB, Burant CF, Takeda J, Lin D, Fukumoto H, Seino S: Molecular biology of mammalian glucose transporters. Diabetes Care 13:198-208, 1990 4. Mueckler M: Family of glucose transporter genes: implications for glucose homeostasis and diabetes. Diabetes 39:6-11, 1990 5. Zorzano A, Wilkinson W, Kotliar N, Thoidis G, Wadzinksi BE, Ruoho A, Pilch PF: Insulin-regulated glucose uptake in rat adipocytes is mediated by two transporter isoforms present in a least two vesicle populations. J Biol Chem 264:12358-63, 1989 6. Douen AG, Ramlal T, Rastogi S, Bilan PJ, Cartee GD, Vranic M, Holloszy JO, Klip A: Exercise induces recruitment of the "insulinresponsive glucose transporter". Evidence for distinct intracellular insulin- and exercise-recruitable transporter pools in skeletal muscle. J Biol Chem 265:13427-30, 1990 7. Hirshman MF, Goodyear LJ, Wardzala LJ, Horton ED, Horton ES: Identification of an intracellular pool of glucose transporters from basal and insulin-stimulated rat skeletal muscle. J Biol Chem 265:987-91, 1990 8. Sternlicht E, Barnard RJ, Grimditch GK: Mechanism of insulin action on glucose transport in rat skeletal muscle. Am JPhysiol254:E63338, 1988 9. Burdett E, Beeler T, Klip A: Distribution of glucose transporters and insulin receptors in the plasma membrane and transverse tubules of skeletal muscle. Arch Biochem Biophys 253:279-86, 1987 10. Friedman JE, Dudek RW, Whitehead DS, Downes DL, Frisell WR, Caro JF, Dohm GL: Immunolocalization of glucose transporter
DIABETES, VOL. 41, DECEMBER 1992
GLUT4 within human skeletal muscle. Diabetes 40:150-54, 1991 11. Douen AG, Burdett E, Ramlal T, Rastogi S, Vranic M, Klip A: Characterization of glucose transporter-enriched membranes from rat skeletal muscle: assessment of endothelial cell contamination and presence of sarcoplasmic reticulum and transverse tubules. Endocrinology 128:611 -16, 1991 12. Klip A, Ramlal T, Young DA, Holloszy, JO: Insulin-induced translocation of glucose transporters in rat hindlimb muscles. FEBS Lett 224:224, 1987 13. Marette A, Hundal HS, Klip A: Regulation of glucose transporter proteins in skeletal muscle. In: Diabetes Mellitus and Exercise. Devlin J, Horton E, Vranic M, Eds. London, Smith and Johnson, 1992, p. 27-43 14. Ahmed A, Taylor PM, Rennie MJ: Characteristics of glutamine transport in sarcolemmal vesicles from rat skeletal muscle. Am J Physiol 259 (Endocrinol Metab 22):E284-91, 1990 15. Hidalgo C, Gonzalez ME, Lagos R: Characterization of the Ca- or Mg-ATPase of transverse tubule membranes isolated fom rabbit skeletal muscle. J Biol Chem 258:13937-45, 1983 16. Rosemblatt M, Hidalgo C, Vergara C, Ikemoto N: Immunological and biochemical properties of transverse tubule membranes isolated from rabbit skeletal muscle. J Biol Chem 256:8140-48, 1981 17. MacLennan DH: Purification and properties of adenosine triphosphatase from sarcoplasmic reticulum. J Biol Chem 245:4508-18, 1970 18. Lowry OH, Rosebrough OH, Farr AL, Randall RJ: Protein measurement with the Folin phenol reagent J Biol Chem 193:265-75, 1951 19. Laemmli UK: Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature (Lond) 227:680-85, 1970 20. MacLennan DH, Campbell KP: Structure, function and biosynthesis of sarcoplasmic reticulum proteins. Trends Biochem Sci 5:148-51, 1979 21. Jorgensen AO, Kalnins V, Maclennan DH: Localization of sarcoplasmic reticulum proteins in rat skeletal muscle by immunofluorescence. J Cell Biol 80:372-84, 1979 22. Handberg A, Vaag A, Damsbo P, Beck-Nielsen H, Vinten J: Expression of insulin-regulatable glucose transporters in skeletal muscle from Type II (noninsulin-dependent) diabetic patients. Diabetologia 33:625-27, 1990 23. Pederson O, Bak JF, Andersen PH, Lund S, Moller DE, Flier JS, Kahn BB: Evidence against altered expression of GLUT1 or GLUT4 in skeletal muscle of patients with obesity or NIDDM. Diabetes 39:865-70, 1990 24. Slot JW, Geuze HJ, Gigendack S, James DE, Lienhard GE: Translocation of the glucose transporter GLUT4 in cardiac myocytes of the rat. Proc Natl Acad Sci USA 88:7815-19, 1991 25. Ramlal T, Rastogi S, Vranic M, Klip A: Decrease in glucose transporter number in skeletal muscle of midly diabetic (streptozotocintreated) rats. Endocrinology 125:890-97, 1989 26. Klip A, Ramlal T, Bilan PJ, Cartee GD, Gulve EA, Holloszy JO: Recruitment of GLUT4 glucose transporters by insulin in diabetic rat skeletal muscle. Biochem Biophys Res Comm 172:728-36, 1990 27. Slot JW, Geuze HJ, Gigengack S, Lienhard GE, James DE: Immunolocalization of the insulin regulatable glucose transporter in brown adipose tissue of the rat. J Cell Biol 113:123-35, 1991 28. Venosa RA, Horowicz P: Density and apparent location of the sodium pump in frog sartorius muscle. J Membr Biol 59:225-32, 1981 29. Marette A, Richardson JM, Ramlal T, Balon TW, Vranic M, Pesson JE, Klip A: Abundance, localization, and insulin-induced translocation of glucose transporters in red and white muscle. Am J Physiol 263 (Cell Physiol 32):C443-52, 1992
1569
Skeletal muscle surface membrane is constituted by the PM domain and its specialized deep invaginations known as TTs. We have shown previously that insulin induces a rapid translocation of GLUT4s from an IM pool to the PM in rat skeletal muscle (6). In this study, we have investigated the possibility that insulin also stimulates the translocation of GLUT4 proteins to TTs, which constitute the largest area of the cell surface envelope. PM, TTs, and IM components of control and insulinized skeletal muscle were isolated by subcellular fractionation. The TTs then were purified further by removing vesicles of SR origin by using a Ca-loading procedure. Ca-loading resulted in a five- to sevenfold increase in the purification of TTs in the unloaded fraction relative to the loaded fraction, assessed by immunoblotting with an anti-OHP-receptor monoclonal antibody. In contrast, estimation of the content of Ca 2+ -ATPase protein (a marker of SR) with a specific polyclonal antibody revealed that most, if not all, SR vesicles were recovered in the Ca-loaded fraction. Western blotting with an anti-COOH-terminal GLUT4 protein polyclonal antibody revealed that acute insulin injection in vivo (30 min) increased the content of GLUT4 (by 90%) in isolated PMs and markedly enhanced (by 180%) GLUT4 content in purified TTs. Importantly, these insulin-dependent changes in GLUT4 content of PM and purified TTs were seen in the absence of changes
From the Division of Cell Biology, The Hospital for Sick Children, Toronto, Ontario; and the Department of Physiology, University of Toronto, Toronto, Ontario, Canada. Address correspondence and reprint requests to Amira Klip, PhD, Division of Cell Biology, The Hospital for Sick Children, 555 University Avenue, Toronto, Ontario M5G 1X8, Canada. Received for publication 31 January 1992 and accepted in revised form 4 June 1992. GLUT, glucose transporter; TT, transverse tubule; cTT, crude transverse tubule; PM, plasma membrane; IM, intracellular membrane; CM, crude membrane; SR, sarcoplasmic reticulum; LSR, light sarcoplasmic reticulum; HSR, heavy sarcoplasmic reticulum; DHP, dihydropyridine; BSA, bovine serum albumin; PMSF, phenylmethylsulfonyl fluoride; CLV, calcium-loaded vesicle; SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis; PVDF, polyvinylidene difluoride; type II diabetes, non-insulin-dependent diabetes mellitus.
1562
in the o^-subunit of the Na + -K + -ATPase, a surface membrane marker. Isolated IM components such as LSR, HSR, and triads (terminal cisternae plus junctional TT) contained low or barely detectable amounts of GLUT4; furthermore, insulin treatment did not change the distribution of the transporter protein within these fractions. In contrast, a unique IM fraction that was not associated with either SR or triad markers, contained significant amounts of GLUT4 and showed an insulin-dependent decrease (40%) in GLUT4 protein content. These results show that acute insulin treatment induces the translocation of GLUT4s to both the PM and TTs from a unique intracellular organelle not associated with the SR. Diabetes 41:1562-69, 1992
I
t is generally accepted that glucose transport across the cell surface is the rate-limiting step for glucose utilization in skeletal muscle (1,2). This metabolic process is markedly stimulated by insulin, and, in the postabsorptive state, skeletal muscle represents the predominant site of whole-body glucose disposal. Glucose transport into muscle cells is mediated by membrane surface GLUT proteins. Two isoforms of the GLUT family have been identified so far in adult rat skeletal muscle: the ubiquitously expressed GLUT1, and the GLUT4 which is present exclusively in insulin-sensitive tissues such as skeletal muscle, heart, and white and brown adipose tissues (3,4). Previous studies in adipose cells have shown that acute insulin exposure stimulates the translocation of GLUT4s from an IM pool to the PM, resulting in an increase in glucose transport activity (5). Other studies also showed an insulin-dependent mobilization of GLUT4s from an intracellular pool to the PM in rat skeletal muscle (6,7). Although insulin also translocates GLUT1 in adipose cells (5), this effect is marginal and could not be observed in insulinized skeletal muscles (6). Therefore, it is recognized that recruitment of GLUT4 proteins is the predominant mechanism by which
DIABETES, VOL. 41, DECEMBER 1992
A. MARETTE AND ASSOCIATES
insulin increases glucose transport activity in these tissues, together with a possible activation of GLUT intrinsic activity by the hormone (8). The insulin-induced membrane redistribution of GLUT4s in muscle is in many respects similar to that of adipose cells. However, important functional and morphological differences have been noted between skeletal muscle and fat. Because of their contractile properties, skeletal muscle cells have developed highly specialized organelles that are involved in excitation-contraction coupling. In addition to the PM, muscle fibers are characterized by the presence of TTs (deep invaginations of the PM that interact with the terminal cistemae of the SR). The TTs are involved in propagating the action potential within the muscle cells where depolarization triggers the release of Ca 2+ from SR, leading to fiber contraction. These tubular membranes are exposed to the extracellular milieu and therefore may play an important role in the transport of ions and nutrients such as glucose into the muscle fiber. Moreover, TT membranes possess insulin receptors (9), suggesting that these organelles are involved in the regulation of glucose transport and metabolism in muscle. We have reported previously that an isolated membrane fraction enriched in TTs contained five times more cytochalasin-B binding sites (which measure total GLUT content irrespective of isoform) compared with isolated PMs of rat skeletal muscle. This preferential segregation of the transporter protein may have potential consequences for the functional response of glucose transport to insulin in skeletal muscle (9). Recently, Friedman et al. (10) reported that GLUT4s are concentrated in triad membranes (where SR apposes the TTs), as detected by immunoelectron microscopy of human muscle sections. They further suggested that insulin induces the movement of GLUT4 from internal sites (possibly SR) to the triad-TT membranes but not to the PM. In contrast, we have reported previously the isolation of PM and IM fractions from rat skeletal muscle that were devoid of TT, triad, or SR markers (11). Yet, we observed an insulindependent translocation of GLUT4 from the IMs to the PMs, demonstrating that GLUT4 can translocate from non-SR and nontriad membranes to surface membranes not containing TTs. Biochemical evidence for an insulininduced mobilization of GLUT4 to TT membranes in muscle is, however, still lacking. The objectives of this study were to determine the following: 1) the effect of acute insulin treatment on the GLUT4 protein content in isolated TTs; and 2) the relative amounts of GLUT4 in various skeletal muscle intracellular organelles suspected to be the site of GLUT4 storage for insulin-dependent mobilization in muscle. The results indicate that insulin induces the translocation of GLUT4s to both the PM and the TTs of skeletal muscle. The original site of the transporters is neither the SR nor the triad elements, but rather a unique insulin-sensitive GLUT4-enriched intracellular organelle. RESEARCH DESIGN AND METHODS Fatty-acid free BSA and gradient grade sucrose were obtained from Sigma (St. Louis, MO). The polyclonal
DIABETES, VOL. 41, DECEMBER 1992
GLUT4-specific antibody and the polyclonal GLUT1specific antibody were obtained from East Acres Biologicals (Southbridge, MA). A monoclonal antibody against the a r subunit of the Na+-K+-ATPase was a gift from Dr. K. Sweadner (Harvard, Boston, MA). The anticalsequestrin monoclonal antibody and the anti-Ca2+-ATPase polyclonal antibody were gifts of Dr. D. MacLennan (Banting and Best Department of Medical Research, University of Toronto). Antibodies to the DHP receptor (IIC12) and to the ryanodine receptor were supplied by Dr. K. Campbell (University of Iowa, Aimes) and Dr. Judith Airey (University of Nevada), respectively. In vivo insulin injection. Male Sprague-Dawley rats weighing 280-320 g were fasted overnight before insulin injection. Rats were anesthetized by intraperitoneal injection of Ketamine-Acepromazin. Insulin (Humulin R, 1.5 U/rat) or saline then was given by tail-vein injection. After 30 min, the rats were killed, and back and hindlimb white muscles were rapidly excised. The muscles were either used immediately or rapidly frozen and kept at -80°C until processed for membrane preparation. Blood was collected before and after insulin injection. The decrease in glycemia (>50%) after insulin administration was used as a positive index of insulin action. Membrane preparation and characterization. PMs and IMs were isolated as described previously and characterized (11-13). Briefly, 10-15 g of frozen muscle first was cleaned of all visible fat, nerve, and vessels. Then they were minced in buffer I (10 mM NaHCO3, 0.25 M sucrose), containing 5 mM NaN3 and 100 jxM PMSF, and homogenized in a Polytron (setting 5) for 5 s. The resulting homogenate was subjected to a series of differential centrifugation steps to yield a CM pellet. The CMs were applied on discontinuous sucrose gradients (25, 30, 35%, w/w) and centrifuged for 16 h at 150,000 g (12). Membranes were collected from each sucrose layer and stored at -80°C until used for Western blots. Extensive characterization of the isolated membranes was given elsewhere (11-13). Briefly, the 25% sucrose fraction contains measurable amounts of GLUTs (GLUT1 and GLUT4) and is enriched in PM markers, such as 5'nucleotidase and phosphodiesterase activity, and content of Na+-K+-ATPase c^-subunit, all of which are increased five- to sixfold relative to nonfractionated CMs. Moreover, the Na+-K+-ATPase activity of the 25% sucrose fraction is 15-fold higher than that of muscle homogenates (14). Conversely, the 35% sucrose fraction is enriched in GLUT4s only, but is depleted of the PM markers discussed above. This fraction therefore is referred to as a GLUT4-enriched IM fraction. A TTs-enriched fraction and a triad-enriched fraction (i.e., terminal cistemae and junctional tubules) were prepared from fresh leg and back muscles as described previously (9,11). When indicated, TT membranes were further purified by using a modification of the Ca-loading procedure previously reported for the isolation of rabbit TTps (15,16). This method allows the removal of contaminating SR vesicles from the TT fractions, based on the SR selective permeability to phosphate ions. Briefly, isolated TTs were incubated at a protein concentration of —0.1-0.2 mg/ml in a solution containing 50 mM K-phos-
1563
GLUT4 IN MUSCLE TRANSVERSE TUBULES
106 kda 80kda 49.5 kda 32.5 kda 27.5 kda Std LSR HSR TT CM PM IM FIG. 1. Protein composition of various membrane components isolated from rat skeletal muscle. Membrane proteins (50 y.g) were subjected to SDS-PAGE on 9% polyacrylamlde mini-gels. The gels were stained with 0.1% Coomassie brilliant blue, destalned, and dried as described In METHODS. Arrowheads point to protein bands that were observed In PM but absent or barely detected in TT membranes.
phate, pH 7.4, 5 mM MgCI2, 0.15 M KCI, 0.3 mM CaCI2) and 2 mM Tris-ATP (Ca-loading solution). The membranes were incubated for 30 min at 22°C and Ca loading was optimized by further additions of 0.3 mM CaCI2 and 2 mM ATP at 5,10, and 20 min of incubation. Membranes then were collected by centrifugation at 150,000 gffor 1 h and resuspended in 2 ml of Ca-loading solution. The Ca-treated membranes were layered on top of discontinuous density gradients of 35 and 50% (w/w) sucrose in Ca-loading solution and centrifuged for 90 min at 150,000 g in an SW 40 Ti rotor. Ca- phosphate complexes precipitate in SR vesicles because they are permeable to phosphate and possess high Ca-uptake rates (in contrast to TT vesicles). Ca loading resulted in the separation of two protein fractions: a Ca-loaded fraction (at the 35/50% sucrose layer interphase) that was found to contain significant amounts of SR vesicles (CLV fraction) and a lighter fraction (top of the 35% sucrose layer) that contained purified TT fraction (RESULTS). LSR and HSR vesicles were prepared from rat skeletal muscle as described previously for rabbit skeletal muscle (17). Protein was determined by the method of Lowry et al. (18). SDS-PAGE. The protein composition of the various fractions was studied by SDS-PAGE in 9% mini-gel slabs as described by Laemmli (19). Gels were stained with 0.1% Coomassie brilliant blue in 5:4:1 methanol/H2O/acetic acid, destained in the same methanol/H2O/acetic acid solution, and dried for photography. Western blot analysis. Membranes (20 \xg of protein) were subjected to SDS-PAGE on 9-10% polyacrylamide gels according to Laemmli (19) and electrophoretically transferred to PVDF filter membranes for 2 h. PVDF
1564
membranes were incubated for 1 h at room temperature with Tris-saline (50 mM Tris-HCI, pH 7.4, 150 mM NaCI) containing 0.04% of the nonionic surfactant Nonidet P 40 and 3% BSA (buffer A), followed by overnight incubation at 4°C with antibodies in buffer A. The PVDF membranes then were washed three times in buffer A (without BSA) followed by a 1-h incubation with either 125l-labeled protein A (2 |xCi/10 ml) or 125l-labeled sheep anti-mouse immunoglobulin G (2 |xCi/10 ml) in buffer A (Fig. legends), and washed three times in buffer A (without BSA), air-dried, and exposed to XAR-5 Kodak film for 12-48 h. Autoradiographs were quantitated by laser-scanning densitometry with an LKB Bromma ultrascan XL enhancer laser scanner with on-line analysis by a Packard Bell computer. The colorimetric alkaline phosphatase procedure was used for detection of the ryanodinereceptor antibody.
RESULTS
The protein composition of the different membranous organelles isolated from rat skeletal muscles is shown in Fig. 1. Equal protein amounts were loaded for each fraction. LSR and HSR fractions share essentially the SDS-PAGE pattern previously reported for SR (20). A band corresponding to a molecular size of -106,000 Mr (106 kDa) represented 80-90% of the protein content of both the LSR and HSR and is likely to correspond to the Ca2+-ATPase of SR (Figs. 3 and 4). In the HSR fraction, a protein of -65,000 Mr (65kDa) was also abundant and probably corresponds to calsequestrin, a Ca-binding protein known to be located in the cisternal portion of SR (21). The TT vesicles were characterized by a more
DIABETES, VOL. 41, DECEMBER 1992
A. MARETTE AND ASSOCIATES
CM C
PM
I C
IM
I C I
106 kd
a - Na7 K+-ATPase
49.5 kd
GLUT4 1 2
3 4 5 6
FIG. 2. Effects of acute insulin treatment on the distribution of GLUT4 In CM, PM, and IM isolated from rat white hindiimb muscles. « 1 -Na + -K + -ATPase (-100,000 Mr [100 kDa] protein) content In the same membrane samples Is also shown as a control. Membrane proteins (20 jig) were subjected to SDS-PAGE on 10% polyacrylamide mini-gels and transferred to PVDF membranes. PVDF membranes were incubated for 1 h at room temperature with Trls-saline (50 mM TrIs-HCI, pH 7.4,150 mM NaCI) containing 0.04% of the nonlonic surfactant Nonldet P 40 and 3% BSA, followed by overnight Incubation at 4°C with either anti-GLUT4 (1:500 dilution) or antl-« 1 -Na + -K + -ATPase (1:100 dilution) antibodies In buffer A. The PVDF membranes then were washed three times In buffer A (without BSA), followed by a 1-h Incubation with either 125l-labeled protein A (2 p.Ci/10 ml) or 125l-labeled sheep anti-mouse immunoglobulin G (2 JJLCI/10 ml) in buffer A, washed three times in buffer A (without BSA), air-dried, and exposed for radloautography.
complex protein composition also typical of a previously reported pattern (15), with prominent polypeptides of 106,000; 68,000; 53,000; 39,000; and 30,000 Mr (106, 68, 53, 39, and 30 kDa). A large number of proteins was seen in CMs, PMs, and IMs of skeletal muscles. These
fractions contained multiple proteins in the 30,000 to 80,000 Mr (30-80 kDa) range, most of which were only marginally represented in the LSR, HSR, and TT fractions. Importantly, the protein composition of the PMs and the IMs was significantly different. Furthermore, several
TT LSR HSR IM C I C I C I C I
205 kda DHP receptor
116 kda
2+
Ca ATPase 49.5 kda
GLUT4 12
3
4
5
6
7 8
FIG. 3. Effects of insulin on the distribution of GLUT4 In cTT, LSR, HSR, and IM Isolated from rat white hindiimb muscles. The specific protein markers for TT (DHP receptor) and SR (Ca2+-ATPase) also are depicted. Membrane proteins (20 fig) were subjected to SDS-PAGE on 9% polyacrylamide mini-gels and transferred to PVDF membranes as described in Fig. 2, except that PVDF filters also were incubated overnight with antl-DHP receptor (1:300 dilution) and anti-Ca 2+ -ATPase (1:100 dilution) antibodies.
DIABETES, VOL. 41, DECEMBER 1992
1565
GLUT4 IN MUSCLE TRANSVERSE TUBULES
proteins of the isolated PMs (Fig. 1, -35,000-40,000 and 80,000 Mr [35-40 and 80 kDa]) were either absent or barely detectable in TTs, suggesting that the tubular elements did not contain vesicles of PM origin. We next investigated the effects of insulin on the membrane distribution of immunodetectable GLUT4s in equal protein samples of the different membranous organelles isolated. The GLUT4 is a protein that migrates with a size corresponding to -45,000 Mr (45 kDa) on polyacrylamide gels. In three individual experiments, insulin did not change the distribution of GLUT4 in unfractionated CMs isolated from white skeletal muscles (Fig. 2, lower band, lanes 1 and 2). However, insulin increased the GLUT4 protein content by 86 ± 12% in purified PMs (lanes 3 and 4) and induced a concomitant 40 ± 12% decrease in the transporter content in the IMs of the same muscle samples (lanes 5 and 6). The a r subunit of the Na+-K+-ATPase (Fig. 2, upper band), a PM marker, was detected in the CMs and was enriched five- to sixfold in the PMs. This protein, however, was absent from the IM pool. When several blots were scanned for densitometry, insulin treatment was found not to alter the content of the enzyme subunit in any membrane fractions, highlighting the GLUT4 proteinspecific translocation in response to the hormone. Note that antibodies to both the DHP receptor (a TT marker) and the ryanodine receptor (a triad marker) did not react with CMs, PMs, or IMs, confirming a previous report that these cellular components were not contaminated with TTs (Fig. 4) (11). The TT membranes were found to contain high amounts of the GLUT4s (Fig. 3, lower band, lanes 1 and 2). As expected, these vesicles were characterized by the presence of the DHP receptor (Fig. 3, upper band, -180,000 H [180 kDa]). However, the TT membranes also contained significant amounts of SR Ca2+-ATPase, a protein migrating at -106,000 Mr (106 kDa) on polyacrylamide gels (Fig. 3, middle band), indicating contamination of the tubular elements with vesicles of SR origin. In these TTs, insulin had no consistent effect on GLUT4 content. The absence of reproducibility of insulin action on TT GLUT4 protein content may have reflected the degree of contamination with SR because the Ca 2+ ATPase content varied from one preparation to another. Low amounts of the GLUT4 protein were detected in the LSR and HSR fractions (Fig. 3, lanes 3-6). The identity of these membranes was confirmed by the abundance of the SR Ca2+-ATPase and by the low amounts of the TT-DHP receptor. In two membrane SR vesicle preparations, insulin had no effect on the distribution of the GLUT4 protein. Conversely, insulin significantly reduced the GLUT4 content of IMs (Fig. 3, lanes 7 and 8) isolated from the same muscles. These vesicles were free of TT-DHP receptor and relatively depleted of SR Ca 2+ ATPase, supporting the tenet that these membranous organelles represent a unique intracellular pool of GLUT4s occluded to the extracellular milieu. It must be emphasized that all fractions analyzed in Fig. 3 were loaded with equal protein values and were run on the same gel. GLUT4 protein was barely detectable in isolated triad
1566
fractions (Fig. 4). When autoradiograms were exposed for longer periods, no significant changes in GLUT4 contents were seen in triads prepared from insulintreated rats (data not shown). The triad membranes, however, were found to contain significant amounts of calsequestrin and ryanodine receptor (Fig. 4), as expected from membranes originating from a TT-HSR complex. Conversely, the levels of SR Ca2+-ATPase were extremely low in this fraction (data not shown), in agreement with a previous study showing a predominant localization of this protein in the LSR (21). From the Western blot depicted in Fig. 3, it is clear that the TTs were significantly contaminated with vesicles of SR origin. Although the relative concentration of GLUT4 is small in SR membranes, large amounts of the intracellular organelle may still mask insulin-induced changes in GLUT4 distribution in the tubular elements, especially if the level of contamination by SR is different in control and insulin-treated muscle vesicles. Ca loading, therefore, was used to further purify TTs from SR vesicles. Figure 5 shows that vesicles that could not be loaded with Ca
Tr
TT
49.5 kda GLUT-4
Tr
TT
Ryanodine Receptor
Tr
Calsequestrin
m
TT
80 kda
49.5 kda
FIG. 4. Comparison of GLUT4 content In TT and triads Isolated from rat white hlndllmb muscles. Markers for triads (ryanodine receptor) and HSR (calsequestrin) also are shown. Membrane proteins (50 n.g) were subjected to SDS-PAGE on 10% (GLUT4 and calsequestrin) or 5% (ryanodine receptor) polyacrylamide mlnl-gels and transferred to PVDF membranes as described In Fig. 2, except that PVDF filters also were Incubated overnight with antlryanodine receptor and anticalsequestrin antibodies. Mr standards are shown on the right. The M, of the reacting doublet of the ryanodine receptor was higher than 500,000.
DIABETES, VOL. 41, DECEMBER 1992
AND ASSOCIAILS
cTT CLV ITp C
I C I C I
205 kda DHP receptor 116 kda
Ca2+ATPa®e
49.5 kda 1
2 3
4
5
6
FIG. 5. Purification of TT by loading the contaminating SR vesicles with Ca phosphate. Control (C) and Insuilnlzed (I) muscle membranes were Isolated and purified as described In METHODS. Membranes proteins (10 p.g) then were analyzed by Western blotting as described In the legend of Fig. 3. The blot Is representative of four Western blots from four separated loading experiments. Mr markers are shown on the left
(TTp, lanes 5 and 6) contained significant amounts of DHP receptor, whereas the CLV (lanes 3 and 4) were depleted of the TT marker. Integration of scans from four gels revealed that the DHP receptor protein was enriched five to seven times in the TTp fraction as compared to the CLV fraction (Fig. 6). In contrast, marginal amounts of the SR Ca2+-ATPase (Fig. 5, middle band; and Fig. 6, middle panel) were detected in TTp, but high levels of the enzyme were found in the Ca2+-loaded CLV. A four- to sixfold reduction in the Ca2+-ATPase content was observed in the TTp fraction as compared to the CLV fraction (Fig. 6). Thus, the TTp represent vesicles of TT origin cleared from SR vesicles, the latter being recovered in the loaded CLV fraction. Note that TTp and CLV vesicles could not be separated when cTT were incubated in the absence of ATP during the loading procedure, underlining the necessity of the ATP-dependent loading for TT purification. Insulin caused a marked increase in GLUT4 content in TTp membranes (Fig. 5, lower band, lane 6 vs. lane 5). Scanning of multiple gels by densitometry revealed that insulin treatment increased GLUT4 protein by almost threefold (180%) in TTp (Fig. 6). This increase was observed despite an absence of change in the TT content of the a r subunit of the Na+-K+-ATPase after insulin treatment (data not shown). No consistent effect of insulin was seen in the CLV vesicles. In four separate Ca2+-loading experiments, the distribution of GLUT4 in the CLV was found to be either augmented (Fig. 5) or diminished, the overall averaged effect of insulin being negligible (Fig. 6). Note that the CLV fraction probably contained GLUT4 of non-SR origin (possibly of GLUT4-
DIABETES, VOL. 41, DECEMBER 1992
enriched intracellular organelle origin), because its GLUT4 content was found to be higher than that observed in isolated LSR and HSR (Fig. 3) when compared with gels analyzed in parallel. One explanation for that observation is that the GLUT4-containing organelle (possibly the GLUT4-enriched IM discussed above) contains a different form of Ca 2+ - pumping activity (perhaps a specific Ca2+-ATPase isoform), explaining its recovery in the CLV fraction after Ca 2+ loading. This possibility, however, remains to be tested experimentally.
DISCUSSION The subcellular localization of GLUTs in skeletal muscle has been the subject of intense recent investigation. Because of the unique intracellular sequestration of the GLUT4 isoform and the known insulin sensitivity of this transporter in muscles, knowledge of its membrane distribution is of particular importance. Furthermore, considering the studies reporting that the total number of GLUTs is not reduced in skeletal muscle homogenates of type II diabetic subjects (22,23), it has become crucial to analyze the GLUT4 content of fractionated subcellular surface membranes and I Ms with the hope that this will lead to unraveling the nature of the defect in glucose transport of diabetic muscle. This study identified both the plasmalemmal and TT membranes as surface components containing GLUT4s that gain transporters upon insulin stimulation. These results suggest that both domains of the muscle cell surface are involved in the transport of glucose into muscle, and that this occurs, at least in part, through
1567
GLUT4 IN MUSCLE TRANSVERSE TUBULES
GLUT1 is localized to the plasma membrane of skeletal) (29). This supports the notion that insulin induces the recruitment of GLUT4 to these two specific cell surface components. Third, translocation of GLUT4 to heart TTs has been recently observed using immunoelectron microscopy (24). Clearly, a surface compartmentation of GLUT4s in muscle would be of physiological importance. We have previously reported that both D-glucose-protectable binding of cytochalasin-B and GLUT4 protein content are decreased in PMs isolated from streptozocin-induced diabetic rat muscle (25,26). Whether downregulation of GLUT4 also occurs in diabetic rat TT membranes is still unknown. It will be of great interest to assess whether the reduced GLUT4 content in diabetes is manifested in both regions of the cell surface or, conversely, whether it is confined to a specific domain. Hence, considering that TTs represent the major fraction (-70%) of total surface membranes in frog skeletal muscle (28), a preferential depletion of GLUT4s in that compartment would be of obvious metabolic consequence. Another important aspect of this study was to determine the intracellular locations of the occluded GLUT4s that are translocated by insulin. These transporters are not recruited from membranes associated with LSR, HSR, or triads. In fact, the latter three intracellular organelles were found to be either poor or devoid of the GLUT4 protein. These data are consistent with our previous observation of very low cytochalasin-B binding in isolated SR vesicles (9). They are, however, in contrast with the immunoelectron microscopy findings of Friedman et al. (10), who previously reported GLUT4 immulolabeling within the triad region and in an intracellular compartment, possibly SR. However, as stated by the authors of the latter study, the fixation protocol used to preserve GLUT4 antigenicity precluded to ascribe definitively the transporter protein to the SR. The fractionation experiments of our study indicate that intracellular GLUT4s are localized in an intracellular compartment that is associated neither with SR nor with triads in skeletal muscle. Rather, GLUT4s appear to be predominantly localized in CLV TTp cTT a unique, insulin-sensitive membranous intracellular orFRACTIONS ganelle. As shown in Table 1, this membrane pool is devoid of typical PM markers (5'-nucleotidase and a-,FIG. 6. Densttometry data showing purification of TT by the Na+-K+-ATPase) and is depleted of markers of both the Ca-loading experiments represented in Fig. 5. Asterisk represents a significant difference (P < 0.02) from control purified (TTp) compared SR (calsequestrin, ryanodine receptor, Ca2+-ATPase) by Student's f test for unpaired data. Scanning data are expressed and triads (ryanodine receptor), underlying the unique relative to control cTT protein contents and represent means ± SE of 4 nature of its origin. Taken together with the recent immuWestern blots from four separate experiments. noelectron microscopy observation of an intracellular GLUT4-enriched tubulo-vesicular network in rat adipoinsulin-dependent regulation of GLUT4 protein content. cytes and heart (24,27), these findings strongly supThe latter suggestion is supported by the following ob- port the existence of a distinct, as yet uncharacterized servations: First, we have observed previously colocal- organelle endowed with GLUT4s. A more complete charization of GLUTs and insulin receptors in TTs (9). Second, acterization of this GLUT4-enriched intracellular orthe isolated PMs are not contaminated with TT mem- ganelle is underway. Preliminary observations (A.M., branes or vice versa, as depicted in Table 1. Regarding A.K., unpublished observations) suggest that this pool is this observation, note that the TTs are totally devoid of the not associated with the endoplasmic reticulum because it GLUT1 isoform (Table 1), whereas this protein is rela- is devoid of calreticulin, a Ca-binding protein known to tively abundant in PMs, strongly suggesting that the two be associated with this cellular organelle. An attractive surface compartments are significantly separated by our possibility is that these IMs are of endosomal origin, as fractionation procedures (We recently observed that suggested by the elegant studies of Slot et al. (24,27),
• •
1568
CONTROL INSULIN
DIABETES, VOL 41, DECEMBER 1992
A. MARETTE AND ASSOCIATES
TABLE 1 Characterization of membrane fractions isolated from rat hindlimb skeletal muscle 25% sucrose (PM)
35% sucrose (IM)
Purified TTp
Longitudinal SR
Triads
Markers GLUT4 GLUT1 a r Na + -K + -ATPase (PM marker) DHP receptor (TT marker) Ryanodine receptor (cisternal SR marker) Calsequestrin (cisternal SR marker) Ca2+-ATPase (longitudinal SR marker)
Symbols (+ and -) denote relative enrichments for equal amounts of protein in the indicated membrane fractions.
which showed that GLUT4 undergoes recycling through the coated pits-early endosome pathway in adipose cells and heart. In summary, the results of this study show that insulin increases the content of GLUT4s in both the PMs and TT membranes isolated from rat skeletal muscle. The GLUT4s appear to be recruited from an intracellular storage site not associated with either SR or triad but rather from a unique, insulin-sensitive intracellular organelle. We suggest that TTs play an important role in the regulation of glucose transport and metabolism in skeletal muscle. ACKNOWLEDGMENTS
This work was supported by a grant to A.K. from the Medical Research Council of Canada. A.M. is the recipient of a postdoctoral fellowship from the Medical Research Council of Canada. We thank Toolsie Ramlal for excellent technical assistance in some of the experiments described in this study. REFERENCES 1. Simpson IA, Cushman SW: Hormonal regulation of mammalian glucose transport. Annu Rev Biochem 55:1059-89, 1986 2. Klip A, Paquet M: Glucose transport and glucose transporters in muscle and their metabolic regulation. Diabetes Care 13:228-43, 1990 3. Bell Gl, Kayano T, Buse JB, Burant CF, Takeda J, Lin D, Fukumoto H, Seino S: Molecular biology of mammalian glucose transporters. Diabetes Care 13:198-208, 1990 4. Mueckler M: Family of glucose transporter genes: implications for glucose homeostasis and diabetes. Diabetes 39:6-11, 1990 5. Zorzano A, Wilkinson W, Kotliar N, Thoidis G, Wadzinksi BE, Ruoho A, Pilch PF: Insulin-regulated glucose uptake in rat adipocytes is mediated by two transporter isoforms present in a least two vesicle populations. J Biol Chem 264:12358-63, 1989 6. Douen AG, Ramlal T, Rastogi S, Bilan PJ, Cartee GD, Vranic M, Holloszy JO, Klip A: Exercise induces recruitment of the "insulinresponsive glucose transporter". Evidence for distinct intracellular insulin- and exercise-recruitable transporter pools in skeletal muscle. J Biol Chem 265:13427-30, 1990 7. Hirshman MF, Goodyear LJ, Wardzala LJ, Horton ED, Horton ES: Identification of an intracellular pool of glucose transporters from basal and insulin-stimulated rat skeletal muscle. J Biol Chem 265:987-91, 1990 8. Sternlicht E, Barnard RJ, Grimditch GK: Mechanism of insulin action on glucose transport in rat skeletal muscle. Am JPhysiol254:E63338, 1988 9. Burdett E, Beeler T, Klip A: Distribution of glucose transporters and insulin receptors in the plasma membrane and transverse tubules of skeletal muscle. Arch Biochem Biophys 253:279-86, 1987 10. Friedman JE, Dudek RW, Whitehead DS, Downes DL, Frisell WR, Caro JF, Dohm GL: Immunolocalization of glucose transporter
DIABETES, VOL. 41, DECEMBER 1992
GLUT4 within human skeletal muscle. Diabetes 40:150-54, 1991 11. Douen AG, Burdett E, Ramlal T, Rastogi S, Vranic M, Klip A: Characterization of glucose transporter-enriched membranes from rat skeletal muscle: assessment of endothelial cell contamination and presence of sarcoplasmic reticulum and transverse tubules. Endocrinology 128:611 -16, 1991 12. Klip A, Ramlal T, Young DA, Holloszy, JO: Insulin-induced translocation of glucose transporters in rat hindlimb muscles. FEBS Lett 224:224, 1987 13. Marette A, Hundal HS, Klip A: Regulation of glucose transporter proteins in skeletal muscle. In: Diabetes Mellitus and Exercise. Devlin J, Horton E, Vranic M, Eds. London, Smith and Johnson, 1992, p. 27-43 14. Ahmed A, Taylor PM, Rennie MJ: Characteristics of glutamine transport in sarcolemmal vesicles from rat skeletal muscle. Am J Physiol 259 (Endocrinol Metab 22):E284-91, 1990 15. Hidalgo C, Gonzalez ME, Lagos R: Characterization of the Ca- or Mg-ATPase of transverse tubule membranes isolated fom rabbit skeletal muscle. J Biol Chem 258:13937-45, 1983 16. Rosemblatt M, Hidalgo C, Vergara C, Ikemoto N: Immunological and biochemical properties of transverse tubule membranes isolated from rabbit skeletal muscle. J Biol Chem 256:8140-48, 1981 17. MacLennan DH: Purification and properties of adenosine triphosphatase from sarcoplasmic reticulum. J Biol Chem 245:4508-18, 1970 18. Lowry OH, Rosebrough OH, Farr AL, Randall RJ: Protein measurement with the Folin phenol reagent J Biol Chem 193:265-75, 1951 19. Laemmli UK: Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature (Lond) 227:680-85, 1970 20. MacLennan DH, Campbell KP: Structure, function and biosynthesis of sarcoplasmic reticulum proteins. Trends Biochem Sci 5:148-51, 1979 21. Jorgensen AO, Kalnins V, Maclennan DH: Localization of sarcoplasmic reticulum proteins in rat skeletal muscle by immunofluorescence. J Cell Biol 80:372-84, 1979 22. Handberg A, Vaag A, Damsbo P, Beck-Nielsen H, Vinten J: Expression of insulin-regulatable glucose transporters in skeletal muscle from Type II (noninsulin-dependent) diabetic patients. Diabetologia 33:625-27, 1990 23. Pederson O, Bak JF, Andersen PH, Lund S, Moller DE, Flier JS, Kahn BB: Evidence against altered expression of GLUT1 or GLUT4 in skeletal muscle of patients with obesity or NIDDM. Diabetes 39:865-70, 1990 24. Slot JW, Geuze HJ, Gigendack S, James DE, Lienhard GE: Translocation of the glucose transporter GLUT4 in cardiac myocytes of the rat. Proc Natl Acad Sci USA 88:7815-19, 1991 25. Ramlal T, Rastogi S, Vranic M, Klip A: Decrease in glucose transporter number in skeletal muscle of midly diabetic (streptozotocintreated) rats. Endocrinology 125:890-97, 1989 26. Klip A, Ramlal T, Bilan PJ, Cartee GD, Gulve EA, Holloszy JO: Recruitment of GLUT4 glucose transporters by insulin in diabetic rat skeletal muscle. Biochem Biophys Res Comm 172:728-36, 1990 27. Slot JW, Geuze HJ, Gigengack S, Lienhard GE, James DE: Immunolocalization of the insulin regulatable glucose transporter in brown adipose tissue of the rat. J Cell Biol 113:123-35, 1991 28. Venosa RA, Horowicz P: Density and apparent location of the sodium pump in frog sartorius muscle. J Membr Biol 59:225-32, 1981 29. Marette A, Richardson JM, Ramlal T, Balon TW, Vranic M, Pesson JE, Klip A: Abundance, localization, and insulin-induced translocation of glucose transporters in red and white muscle. Am J Physiol 263 (Cell Physiol 32):C443-52, 1992
1569
