Phospholamban (PLB) associates with the Ca2+-ATPase in sarcoplasmic reticulum (SR) membranes to permit the modulation of contraction in response to β-adrenergic signaling. To understand how coordinated changes in the abundance and intracellular trafficking of PLB and the Ca2+-ATPase contribute to the maturation of functional muscle, we measured changes in abundance, location, and turnover of endogenous and tagged proteins in myoblasts and during their differentiation. We found that PLB is constitutively expressed in both myoblasts and differentiated myotubes, whereas abundance increases of the Ca2+-ATPase coincide with the formation of differentiated myotubes. We observed that PLB is primarily present in highly mobile vesicular structures outside the endoplasmic reticulum, irrespective of the expression of the Ca2+-ATPase, indicating that PLB targeting is regulated through vesicle trafficking. Moreover, using pulse-chase methods, we observed that in myoblasts, PLB is trafficked through directed transport through the Golgi to the plasma membrane before endosome-mediated internalization. The observed trafficking of PLB to the plasma membrane suggests an important role for PLB during muscle differentiation, which is distinct from its previously recognized role in the regulation of the Ca2+-ATPase.
- sarco(endo)plasmic reticulum calcium-adenosine triphosphatase
- C2C12 myocytes
- vesicle trafficking
the sarcoplasmic reticulum (SR) Ca2+-ATPase mediates the rate-limiting resequestration of calcium ions from the cytosol to the SR lumen after each contractile event in muscle. Its association with the inhibitory protein phospholamban (PLB) provides an additional level of regulation whereby calcium transport rates, and the force of contraction, are directly modulated by β-adrenergic stimulation by means of the phosphorylation of PLB through either cAMP- or CaM-dependent kinases (3, 18, 26, 27, 31). PLB and the Ca2+-ATPase reside as an integral membrane protein complex in the SR, where maximal inhibition of the Ca2+-ATPase requires an equal stoichiometry of PLB, as is present in coexpressed or reconstituted vesicles, as well as in vivo in SR isolated from slow-twitch skeletal muscle (8, 20). However, in the heart, PLB is present in a three- to fourfold molar excess relative to the Ca2+-ATPase, suggesting that this reserve pool of free PLB provides a graded response of calcium transport rates to cellular kinases. The functional importance of maintaining optimal expression levels of PLB relative to that of the Ca2+-ATPase is highlighted by variations in PLB-to-Ca2+-ATPase ratios that correlate with defective calcium regulation in hyper- and hypothyroidism, aging, and the failing heart (16, 17). Furthermore, genetic manipulations have demonstrated that cellular changes in PLB-to-Ca2+-ATPase molar ratios can act to compensate for altered calcium affinities of Ca2+-ATPase isoforms (33).
The initial formation of PLB-Ca2+-ATPase interactions in the developing heart is coincident with progressively decreasing stoichiometries of total expressed PLB relative to Ca2+-ATPase as differentiation to mature myocytes occurs (2, 24). PLB is expressed early in the undifferentiated myocyte, whereas upregulation of the Ca2+-ATPase occurs coincidentally with differentiation and the formation of the SR membrane, which is a specialized region of the endoplasmic reticulum (ER). In addition, differentiation requires new trafficking pathways for these proteins from their site of synthesis in the ER to the newly forming SR. However, little is known regarding the routing of these or any SR resident proteins from ER to SR, or how the crucial interactions between PLB and the Ca2+-ATPase are formed in either immature or adult myocytes.
Therefore, to understand the cellular mechanisms associated with targeting PLB to the SR in developing muscle, we investigated the coordinated expression of PLB during myocyte differentiation using the C2C12 muscle cell line, which is amenable to transfection with a variety of fusion proteins with fluorescent labels. By monitoring the localization and cellular trafficking of PLB and the Ca2+-ATPase, we found that both endogenous and transfected Ca2+-ATPase exhibit a reticular localization that is consistent with an ER/SR localization, whereas PLB is present as punctate vesicles in myoblasts that undergo directed trafficking to the plasma membrane before their degradation. Following differentiation, PLB colocalizes with the Ca2+-ATPase in the SR of myotubes. These results indicate an important role for cellular trafficking in mediating the formation of a functional SR associated with cellular signaling and suggest a possible role for PLB in the plasma membrane associated with muscle development.
MATERIALS AND METHODS
The green fluorescent protein (GFP)-SERCA1a vector, coding for the chicken SERCA1a sequence (4), was a kind gift from Norman J. Karin. GFP-PLB was made by isolating a BamHI/EcoRI fragment from the PLB-pGEX2T vector (36) containing the porcine PLB sequence and inserting it into pEGFP-C1 (Clontech, Mountain View, CA) digested with BglII/EcoRI. Cyanofluorescent protein (CFP) and yellow fluorescent protein (YFP) versions of these vectors were made by replacing the AgeI/BsrG1 fragment from the GFP vectors with the corresponding fragments from the pECFP-C1 and pEYFP-C1 vectors (Clontech). FlAsH·EDT2 was synthesized as previously described (1).
Antibodies to SERCA2a, SERCA1, and PLB were obtained from Affinity BioReagents (Golden, CO), and antibodies to golgin-97 and Alexa Fluor 594 goat anti-mouse IgG were obtained from Invitrogen (Carlsbad, CA). The H2 kinesin heavy chain antibody has been described previously (25). The monoclonal antibody (KLC-2B) recognizing a subset of kinesin light chain-1 alternatively spliced isoforms (KLC-B and KLC-C; Ref. 6) was generated by immunizing mice with a peptide (MRKMKLGLVK) found exclusively in these kinesin light chain-1 isoforms and producing antibodies as described previously (25). These antibodies have been extensively characterized and recognize a subset of native kinesin light chains and bacterially expressed kinesin light chains B and C, but not A (28). The MHC monoclonal antibody (MF-20) developed by Donald A. Fischman was obtained from the Developmental Studies Hybridoma Bank developed under the auspices of the National Institute of Child Health and Human Development and maintained by The University of Iowa (Department of Biological Sciences, Iowa City, IA 52242). Mitotracker red dye was obtained from Invitrogen, and mitochondria in live cells were labeled according to the manufacturer's instructions. BG-505, BG-DAF, and BG-TMR were obtained from Open Biosystems (Huntsville, AL). All cell culture media, unless otherwise stated, were obtained from GIBCO Laboratories (Grand Island, NY).
O6-alkylguanine-DNA alkyltransferase (AGT)-tagged proteins were expressed in myocytes from AGT utility vectors created using PCR to amplify the AGT coding sequence from the pSS26m vector (Covalys, Witterswil, Switzerland) with primers containing 5′ Age1 and NotI sites and a 3′ BsrG1 site. The enhanced GFP (EGFP) coding sequence was replaced with the PCR-amplified AGT sequence by using the Age1 and BsrG1 restriction sites to create the vector pAGT-C1 containing the same multicloning site as the original pEGFP-C1 vector. PLB and SERCA were directly subcloned from the corresponding pEGFP-based vectors into the pAGT-based vectors. Retroviral vectors containing AGT-PLB and AGT-SERCA were created by inserting the Not1/BamH1 digest into the pBM retroviral vector (originally from the Garry Nolan Lab, Stanford University, Palo Alto, CA) modified by expansion and inversion of the multiple cloning site. C2C12 stable cell lines expressing tagged PLB and SERCA1a were generated by selecting transduced cells in puromycin.
The plasmid pTC-C1 was made by synthesizing two oligonucleotides containing 5′ AgeI and NotI sites and a 3′ BsrG1 site, separated by a DNA sequence encoding the amino acids MAEAAAREACCPGCCARAR. The two oligos were annealed and digested with AgeI and BsrG1, and this fragment was inserted into the Age1 and BsrG1 sites in pEGFP-C1. The PLB sequence was inserted as described above for the AGT vector. As described for AGT-tagged PLB, a retroviral vector coding for this tetracysteine (TC)-tagged PLB was created and stably expressed in C2C12 cell lines.
C2C12 skeletal myocytes (CRL-1722; ATCC, Rockville, MD) were grown at 37°C and 5% CO2 in DMEM supplemented with 10% fetal bovine serum (FBS; GIBCO), penicillin/streptomycin, 1 mM sodium pyruvate, and 4 mM l-glutamine. For differentiation experiments, this growth medium was replaced with differentiation medium consisting of DMEM, 2% horse serum, penicillin/streptomycin, 1 mM sodium pyruvate, and 4 mM l-glutamine after the cells reached confluency. Additional myocyte cell lines tested (Sol8, L6, and H2C9) were obtained from ATCC and grown under identical conditions.
Immunoblotting and immunofluorescence.
Cell lysates were prepared using 50 mM Tris (pH 7.4), 150 mM NaCl, 1 mM EDTA, 1% (wt/vol) Nonidet P-40, 5% glycerol, 10 μM leupeptin, 2 μg/ml aprotinin, 1 μM pepstatin, 2 mM PMSF, 50 mM NaF, and 1 mM Na3VO4. These lysed cells were centrifuged (30 min at 16,000 g), and the protein concentration in the lysates was determined before electrophoretic separation. SDS-PAGE was performed using 7.5, 15, or 4–12% polyacrylamide gels. Proteins were transferred to a polyvinylidene difluoride membrane using a semidry transfer cell (Bio-Rad) for 30 min at 25 V for SERCA1 and SERCA2 identification and for 15 min at 20 V for PLB identification, in buffer containing 25 mM Tris, 192 mM glycine, and 20% methanol (for SERCA) or 10% methanol (for PLB). For detection of SERCA2a, blots were probed with anti-SERCA2a monoclonal antibody 2A7-A1, which recognizes residues 386–396 of canine SERCA2a (12), with dilution 1:10,000. SERCA1 isoform was detected using anti-SERCA1 monoclonal antibody IIH11 with dilution 1:20,000. PLB was detected with anti-PLB monoclonal antibody 2D12 with dilution 1:2,000. Secondary antibodies conjugated with horseradish peroxidase were used, and proteins were detected using enhanced chemiluminescence (ECL; Amersham, Piscataway, NJ). To detect endogenous proteins in myocytes, we carried out immunolabeling as previously described (29) using the above antibodies. Primary antibodies were detected using Alexa Fluor 594 goat anti-mouse IgG from Invitrogen.
Live cell imaging and fluorescence recovery after photobleaching.
Live cell microscopy was performed on cells transiently transfected with GFP-, CFP-, or YFP-tagged PLB or SERCA1a as indicated. Cells were either fixed for colabeling with antibodies or imaged live using a Nikon TE2000 inverted microscope equipped with stage and lens heaters (Bioptechs, Butler, PA). To label AGT-tagged proteins, we incubated cells in medium containing 4 μM benzyl guanine (BG)-505, BG-diacetylfluorescein (DAF), or BG-tetramethylrhodamine (BG-TMR) star for 30 min, followed by a 15-min wash in medium before live cell analysis or fixation. For dual-color pulse-chase experiments, cells were labeled with BG-TMR, washed, and incubated for 5 h in ligand-free cell medium before being relabeled with BG-505. BG-505 was visualized using fluorescein filter sets, whereas BG-TMR was imaged using rhodamine filter sets. Labeling TC-PLB-expressing cells involved incubation of cells for 30 min in HEPES-buffered saline (HBS) containing 1 μM FlAsH·EDT2 reagent. Cells were washed with HBS, incubated for 5 min in HBS containing 5 mM EDT to reduce nonspecific background, and washed twice in HBS before replacement of tissue culture medium and imaging of live cells. The supplemental movie showing movement of AGT-PLB was acquired using Metamorph software (Universal Imaging, Downington, PA) to acquire a stack of time-lapse images and saved in Quicktime format. (The online version of this article contains supplemental data.)
Fluorescence recovery after photobleaching (FRAP) was performed using a Leica confocal microscope equipped with an argon laser and appropriate filters for imaging GFP-tagged proteins. For photobleaching, a rectangular region of the desired size was photobleached using 100% laser power for a total duration of ∼1 s. Images were collected before, immediately after, and at set intervals following the bleach for a time sufficient to allow complete recovery of each protein. Original images were exported as tagged image files, and final illustrations were assembled using Adobe Photoshop.
Red fluorescent protein endosomal marker.
As a marker of endosomes, the RFP-Endo vector was created by inserting the RhoB GTPase (from pEYFP-Endo; Clontech) into a pBM retroviral vector containing a red fluorescent protein (RFP) tag using PCR. A double stable cell line expressing AGT-tagged PLB and RFP-Endo was created using two retroviral vectors containing different selection markers (puromycin for AGT-PLB and blasticidin for RFP-Endo).
Differential abundance of PLB and Ca2+-ATPase proteins during myocyte differentiation.
C2C12 myocytes provide an accurate model of developing muscle, because they can be readily induced to differentiate. PLB expression is essentially constitutive throughout differentiation, whereas that of the Ca2+-ATPase (the SERCA1 and SERCA2 isoforms) is upregulated coincidentally with differentiation into myotubes. This differential expression is illustrated by Western blot analysis (Fig. 1A). PLB was already present in undifferentiated myoblasts (day 0), whereas both fast-twitch (SERCA1) and slow-twitch (SERCA2) isoforms of the Ca2+-ATPase were present at low or undetectable levels in these cells. Of note, SERCA2b has been observed to be present at low levels in the ER of myoblasts, whereas its splice isoform, SERCA2a, is predominant in differentiated myocytes; both SERCA2 isoforms are detected by this antibody (2, 7). To determine how differentiation affects the relative expression of these proteins, we allowed C2C12 cells to reach ∼90% confluency in growth medium (DMEM with 10% FBS) before its replacement with differentiation medium (DMEM with 2% horse serum) for 1–7 days. As shown in Fig. 1B, differentiation had no significant effect on total PLB expression levels, but both SERCA1a and SERCA2 showed a linear increase in abundance as cells differentiated. This increase in SERCA abundance coincided with increased proteins levels of myosin heavy chain (MHC), a classic marker of muscle differentiation (data not shown). We note that this differential expression pattern was also observed in other myocyte cell lines tested, i.e., the L6 and Sol8 skeletal myocytes and the cardiac H2C9 cell line (data not shown), indicating that the early appearance of PLB in myoblasts in the absence of the Ca2+-ATPase is a general feature of cultured myocytes, and this pattern mirrors the same changes in relative abundances of PLB and the Ca2+-ATPase observed in the developing heart (2, 24).
Cellular localization of endogenous PLB and SERCA2a.
In adult myocytes, the PLB-SERCA complex of integral membrane proteins resides in the SR membrane, which develops during differentiation as a specialized and finally extensive region of the ER (19). For determination of the cellular distribution of PLB as myocytes differentiate, C2C12 myocytes were fixed and immunostained for PLB and compared with SERCA1. Myoblasts immunostained for PLB did not show a reticular distribution but, rather, a punctate staining, indicative of vesicles that are distinct from the ER (Fig. 1C, top left). Only in myotubes did PLB staining show a reticular staining consistent with ER or SR localization (Fig. 1C, top right). No detectable SERCA1 immunostaining was observed in undifferentiated myoblasts (Fig. 1C, bottom left), consistent with the absence of total SERCA protein assayed by Western blot (Fig. 1B). However, in myocytes in which the differentiation program had been activated, SERCA1 was detected and exhibited a reticular distribution consistent with SR localization (Fig. 1C, bottom right). These observations suggest the requirement for altered trafficking of PLB from its localization in vesicles to the SR membrane for the formation of PLB-SERCA interactions during differentiation.
Localization of GFP-PLB and GFP-SERCA1a.
To better address intracellular trafficking of PLB and SERCA in living myocytes, we transiently transfected C2C12 myoblasts with GFP-PLB or GFP-SERCA1a, essentially as described by Biehn et al. (4). The distribution of GFP-PLB (Fig. 2, top left), like that of endogenous PLB (Fig. 1), exhibited a punctate pattern that extended to the edge of the cell and localized to regions of the plasma membrane and cellular extensions in many cells. On the other hand, in myoblasts, GFP-SERCA1a (Fig. 2, bottom left) was constrained within, but evenly distributed throughout, the endoplasm and was observed to be reticular in many regions. This reticular intracellular distribution is similar to that expected for wild-type SERCA in differentiated muscle, i.e., in the SR, and mirrors that observed when exogenous GFP-SERCA1a is expressed in fibroblasts, demonstrating that in the absence of an SR membrane, SERCA1 is targeted to the ER (4).
Immunolabeling of GFP-transfected cells with anti-PLB or anti-SERCA1a antibodies showed that the GFP localization pattern coincided with the immunolocalization pattern (data not shown). In the case of GFP-SERCA1a, the expressed sequence was from chicken, permitting the use of an antibody specific for avian SERCA1a to specifically detect only the transfected protein (13). Thus the identical patterns of localization of endogenous and GFP-tagged PLB indicate both that the GFP tag does not affect normal protein distribution in the myoblast and that the fixation required for immunolabeling of endogenous PLB does not produce artifactual staining patterns.
To further compare the intracellular trafficking and turnover of PLB and SERCA, we created stable cell lines expressing PLB and SERCA, each as fusion proteins with AGT. This genetically engineered version of AGT is a 21-kDa protein that cleaves cell-permeable para-substituted benzyl guanines (BG) to create a stable thioether intermediate that provides a fluorescent group covalently bound to PLB or SERCA and is suitable for live cell imaging (14, 15). These stable myocyte cell lines were amenable to differentiation into myotubes in which the levels of expressed AGT-PLB and AGT-SERCA protein were ∼50% of that of endogenous (wild type) PLB and SERCA (data not shown). As shown in Fig. 2, middle, C2C12 myoblasts expressing AGT-PLB and labeled with a BG fluorescein derivative, BG-505, showed the same punctate distribution as endogenous or GFP-PLB, which is distinct from the reticular distribution of SERCA. Finally, as an additional control for the possibility that the large size of either the GFP (30 kDa) or AGT (21 kDa) tags relative to the smaller PLB (6 kDa) might produce artifactual aggregation, a PLB construct was genetically engineered with an NH2-terminal sequence containing a small TC motif suitable for labeling with the cell-permeable FlAsH (<700 Da; Ref. 9). This TC-tagged PLB (C4-PLB) expressed in C2C12 myoblasts and labeled with FlAsH also exhibited a punctate fluorescence pattern (Fig. 2, right). Thus the punctate distribution of PLB in undifferentiated myoblasts consistently observed for wild-type PLB in fixed and immunostained cells as well as PLB expressed with both large and small tags for live cell imaging provides a strong level of confidence that PLB localization in myoblasts is distinct from an ER localization or its reticular localization in the mature myotube.
Differential mobilities of PLB and SERCA in myoblasts.
As an alternative to cell imaging, we addressed the differential localization of PLB and SERCA by biophysical measurements of FRAP, since rates of lateral protein mobility can be altered by cell locale. As shown in Fig. 3A, when a region of a myoblast expressing GFP-SERCA1a was photobleached, a gradual, uniform recovery of fluorescence was observed. The half-life of recovery was ∼20 s, significantly slower than the half-life of freely diffusing GFP (∼0.3 s; Ref. 30). Thus the mobility of SERCA1a is constrained, consistent with its lateral diffusion within the ER membrane. In contrast, when FRAP was performed on GFP-PLB, a much different recovery pattern was observed (Fig. 3B). First, the recovery occurred more slowly, requiring several minutes for complete recovery of the fluorescence. Second, the GFP-PLB entered the bleached area as discrete, subcellular structures, which suggests recovery is due to transport of GFP-PLB-containing vesicles into the bleached zone (see magnified images in Fig. 3C). Thus these very different rates and modes of diffusion of SERCA compared with PLB are incompatible with colocalization, indicating their localization to distinct compartments, i.e., ER and vesicles, respectively, in undifferentiated myocytes.
PLB-containing vesicle movements are characteristic of microtubule-based transport.
To track the movement of individual PLB-containing vesicles, we performed live cell imaging of both GFP- and AGT-PLB, both of which gave equivalent results. Figure 3D shows a magnified portion of a live cell stably expressing AGT-PLB and labeled with the fluorescein dye BG-505; many of the PLB-containing vesicles undergo rapid bidirectional movement in the cell. Further expansion of a portion of this cell in the subsequent panels shows one AGT-PLB-containing vesicle (large arrow) that underwent bidirectional transport over time, on what appears to be a linear track, relative to two other vesicles that remained more stationary (small arrow) and therefore serve as landmarks. The movement in both directions appears to occur by pulling as the vesicle is stretched in the direction of the movement, followed by a rapid movement of the bulk of the vesicle in the same direction. The movement can be better appreciated in the movie included as supplemental data (Fig. S1). This bidirectional movement is characteristic of microtubule-based transport mediated by kinesins and cytoplasmic dynein and further suggests that the directed transport of PLB in vesicles occurs in myoblasts (5).
Differential turnover of PLB and SERCA.
As a means to monitor the localization of PLB and SERCA over time after protein synthesis, we took advantage of the covalent labeling of AGT-tagged proteins with BG-derivatized fluorophores, which provides a means to pulse label a population of proteins. After labeling, changes in the location and intensity of the fluorescence signal associated with either PLB or SERCA are attributable to either trafficking or protein turnover and are insensitive to the biosynthesis of new proteins. To determine the relative turnover rates of PLB and SERCA1a, AGT-PLB- or AGT-SERCA-expressing myoblasts were labeled with a pulse of BG-fluorescein, which labels all expressed protein. After a wash step to remove unbound fluorophore, both the levels of labeled protein and their cellular distribution were monitored over time (Fig. 4A). As illustrated in Fig. 4A, top, the fluorescence associated with labeled PLB disappeared within 24 h and was accompanied by a noticeable time-dependent change in the distribution of the labeled PLB within the cell. Specifically, fluorescence at the plasma membrane disappeared relatively early on, and over time the fluorescence became less intense but more centrally localized around the nucleus. In contrast, fluorescently labeled AGT-SERCA exhibited a relatively stable reticular distribution over the time course of the experiment given that its fluorescence intensity diminished, consistent with the synthesis and retention of AGT-SERCA as an integral ER membrane protein. Moreover, fluorescence associated with SERCA persisted for over 24 h (Fig. 4A), consistent with the 30-h half-life for SERCA1 previously measured with radioisotopes in chick myocytes (11). The more rapid loss of fluorescence associated with AGT-PLB compared with that of AGT-SERCA suggests its shorter protein half-life than that of SERCA; to our knowledge, protein turnover of PLB has not been previously measured.
Additional kinetic resolution of cellular trafficking was obtained through a dual-color pulse-chase experiment, permitting the measurement of the cellular localization of newly synthesized proteins (Fig. 4B). In this case, existing AGT-PLB or AGT-SERCA1a was labeled with BG-TMR (red); after an additional 5 h in culture, newly synthesized proteins were labeled with BG-505 (green). In the case of PLB (Fig. 4B, top), both the preexisting protein (red) and newly synthesized protein (green) had similar cellular distributions, with the exception that there were elevated levels of newly synthesized protein, as evidenced by the concentration of green label, in a perinuclear region (arrow). There was significant overlap in the distribution of the red and green labeled PLB in other regions of the cell. Similar experiments performed with AGT-SERCA showed that both existing and newly synthesized SERCA were predominantly localized in the ER (Fig. 4B). Thus, in contrast to the redistribution of PLB observed as a function of its time after synthesis in the myoblast, SERCA appears to be synthesized rapidly with a short residence time in the Golgi and longer residence in ER within the time window used in these studies. These data indicate that SERCA is trafficked independently of PLB.
Altered PLB distribution occurs upon myocyte differentiation.
The increased expression of SERCA in differentiating myocytes or developing muscle correlates with proliferation of newly formed SR. Therefore, the change in PLB localization as myocytes differentiate could be due to recruitment of PLB to SERCA as it becomes available in the SR. To test whether the expression of SERCA in myoblasts affects PLB localization, we cotransfected myoblasts with bioluminescent forms of PLB and either SERCA1a or SERCA2a and examined their intracellular distribution in individual cells. As shown in Fig. 5A, YFP-PLB retained its vesicular distribution even in the presence of coexpressed CFP-SERCA1a. Coexpression of SERCA2a, which directly interacts with PLB in vivo, also did not affect the localization of PLB in myoblasts. In this case, untagged SERCA2a was utilized to avoid any potential interference of a tag with PLB-SERCA2a interactions (Fig. 5B). Thus the presence of SERCA2a or SERCA1a in the ER of the myoblast is not sufficient to stimulate the redistribution of PLB into the reticular compartment and to co-localize with SERCA.
We next addressed whether changes in PLB localization correlate with the initiation of differentiation for individual myocytes. Since differentiated myocytes are not amenable to transfection, the initial analysis involved transient transfection with YFP-PLB and CFP-SERCA1a and analysis after 2 days in differentiation medium, resulting in a mixed population of both myoblasts and myocytes in the early stages of differentiation. Under these conditions, most of the transfected cells remained as myoblasts, which do not express MHC, whereas both PLB and SERCA retained their unique distributions (Fig. 5C). However, in the small number of cells that had begun to differentiate as judged by both MHC expression and multinucleation, YFP-PLB developed a localization pattern similar to that of CFP-SERCA1a (Fig. 5D). Using our stable cell lines expressing AGT-PLB and AGT-SERCA1a, we were able to extend our analysis to examine later stages of differentiation. After 4 days of differentiation, AGT-PLB lost its punctuate distribution (Fig. 5E) and adopted a distribution similar to that of AGT-SERCA1a (Fig. 5F). Together these results suggest that differentiation must occur before PLB colocalizes with SERCA isoforms in SR membranes.
Colocalization of AGT-PLB with cellular reference markers in live cells.
Since PLB does not show substantial localization in the ER in undifferentiated cells, we performed colocalization experiments with cell compartment-specific dyes and antibodies to identify the cellular compartment(s) associated with PLB trafficking. As shown in Fig. 6A, a concentration of AGT-PLB was observed in the region of the Golgi, as evidenced by colocalization of a subset of PLB in the perinuclear region stained with an antibody to the Golgi resident protein golgin. In contrast, there was no significant colocalization of PLB with mitochondria when this compartment was labeled with Mitotracker red (data not shown), indicating that PLB is excluded from mitochondrial membranes. Additional resolution regarding the possible role of endosomes in mediating PLB trafficking was addressed through an assessment of the possible colocalization of AGT-PLB (shown labeled with BG-505 in green) and an RFP-endosome marker, i.e., the RhoB GTPase (Fig. 6, B and C). We found extensive colocalization of PLB with the labeled endosomes, as highlighted in the enlarged image that shows many vesicles containing both AGT-PLB and the RFP-Endo (Fig. 6C).
As a further demonstration that PLB localizes to transported vesicles, we performed colocalization experiments with antibodies recognizing the vesicle transport motor kinesin. Shown in Fig. 6D is an AGT-PLB-expressing cell labeled with an antibody (KLC-2B) recognizing a subset of kinesin light chain-1 isoforms generated by alternative splicing (6). This image shows that regions near the plasma membrane containing AGT-PLB also contained elevated levels of kinesin light chains, suggesting that these membrane domains are targeted by actively transported vesicles. Similar results were obtained with an antibody (H2; Ref. 25) recognizing kinesin heavy chains (Fig. 6E). These colocalization results support the directed transport of PLB-containing vesicles in undifferentiated myoblasts, from the site of synthesis near the perinuclear region with movement to the plasma membrane through the Golgi before internalization and subsequent degradation of PLB by endosomes.
Changes in relative abundances of PLB and SR Ca2+-ATPase occurring in heart disease and other pathophysiological states suggest the importance of understanding the trafficking pathways associated with the biosynthesis of functional SR membranes. Using both wild-type and fluorescence-tagged protein-expressing C2C12 myocytes, we found striking differences in the abundance, cellular localization, and trafficking of PLB relative to the Ca2+-ATPase in myoblasts before their colocalization in the differentiated myotube. In particular, in myoblasts, a punctate staining pattern for PLB was consistently observed that was distinct from the reticular staining of expressed Ca2+-ATPase; this pattern suggests a vesicular localization for PLB (Figs. 1–6). The very different rates of fluorescence recovery after photobleaching for PLB and the Ca2+-ATPase in the myoblast lend additional support for their distinct cellular compartments (Fig. 3). From live cell imaging, the bidirectional character of movement of the PLB-associated points of fluorescence suggests the transport of vesicles along microtubules, as does the colocalization with the motor protein, kinesin (Figs. 3 and 6). Pulse-chase experiments that permitted the fluorescent labeling of a transient population of proteins indicate that recently synthesized PLB is more highly concentrated in a perinuclear locale consistent with synthesis within the rough ER, whereas at later times after its synthesis, the distribution of PLB extends to the periphery of the cell (Figs. 2 and 4). The early loss of fluorescence of labeled AGT-PLB at the plasma membrane and the colocalization of PLB with RhoB and golgin suggests movement of PLB vesicles outward toward the plasma membrane through the Golgi apparatus, where PLB becomes incorporated into endosomes and is finally degraded. In this regard, a truncation mutant of PLB (PLB1–39) that leads to a lethal cardiomyopathy in humans, exhibits accumulation at the plasma membrane when expressed in tissue culture and is less stable than the wild-type protein (10). This altered localization of PLB could be due to its failure to be endocytosed at the plasma membrane, thereby disrupting the normal trafficking cycle that is presently proposed.
Of note, although a large fraction of PLB is associated with vesicles, it is apparent from the low level of background fluorescence associated with both endogenous and tagged PLB in our images that a fraction of PLB is associated with ER (Figs. 1–6). This observation is consistent with the expression of SERCA2b in myoblasts and with prior observations regarding the ability of PLB to regulate SERCA2b when coexpressed in cells and in genetically altered mice (34, 35). Thus PLB may have a dual role in the myoblast, one to regulate SERCA2b in the ER, and a second, as yet unidentified, that involves extensive trafficking of PLB to the plasma membrane.
One surprising result of this study is that it is not the availability of SERCA2a that is responsible for the localization of PLB to the ER as evidenced by the observation that the distribution of PLB in vesicles was retained in the presence of exogenous SERCA2a expressed in myoblasts at levels comparable to those of differentiated myotubes. Only upon the extensive reorganization associated with myocyte differentiation and the formation of SR membranes did the bulk of PLB colocalize with the Ca2+-ATPase, suggesting an important role for intracellular signaling processes in mediating the development of functional SR membranes that contain optimal amounts of PLB and the Ca2+-ATPase. Although the possibility that the same myoblast trafficking pathway of PLB-containing vesicles also occurs in differentiated myotubes has not been explicitly addressed in the present study, the inability to detect PLB-containing vesicles in myotubes implies that PLB, like the Ca2+-ATPase, resides primarily in the SR after differentiation. However, the possibility cannot be ruled out that the extensive proliferation of SR in myotubes masks ready detection of a minor vesicular population of PLB that trafficks from ER to plasma membrane (Fig. 7).
Based on the extensive knowledge gained in recent years regarding trafficking pathways for plasma membrane and secreted proteins, it is known that these proteins are synthesized in the ER, exported at ER exit sites via COPII (coat protein complex II) vesicles, and transported by microtubules to the ER-Golgi intermediate complex (ERGIC), where sorting occurs before trafficking to their final cellular destination (22). Thus the colocalization of PLB vesicles with the motor protein kinesin and the vesicle movements observed by live cell imaging that are characteristic of microtubule transport strongly support the directed trafficking of PLB (Figs. 2 and 6). Moreover, this result rules out a significant involvement of the alternative pathway for ER exit, i.e., the retrograde transport of misfolded proteins that are routed to the cytoplasm for proteasomal degradation. The proteasome is located at the cytoplasmic face of the ER membrane for retrograde protein exit, and directed transport mechanisms are not involved (21). Instead, the colocalization of PLB with endosomes (Fig. 6) suggests that PLB degradation involves lysosomal degradation and is reminiscent of the mechanisms by which receptor abundance at the cell surface is regulated. Resident ER proteins that do escape from the ER in COPII vesicles are retrieved and recycled via specific carboxy-terminal ER retention sequences such as the dilysine (-KKXX) motif (in type I integral membrane proteins) and KDEL (in ER lumenal proteins); these motifs are not found within the PLB sequence (22). Alternatively, ER proteins may be selectively retained through binding to other ER proteins; however, such a mechanism involving binding to SERCA2a, the normal binding partner of PLB, can be excluded. Thus the trafficking of PLB in myoblasts would appear to be consistent with classic secretory routes and suggests a novel function for PLB in the myoblast.
Differentiation requires that both PLB and SERCA transit to the SR from their site of synthesis in the ER. The associated mechanisms for these or other SR resident proteins are not well understood, but current work has suggested that this relocation does not involve direct lateral diffusion through the membrane bilayer as previously thought. For example, the SR lumenal protein calsequestrin has been shown to be routed to the SR via COPII vesicles budding at ER exit sites, but without Golgi involvement (23). Similarly, SERCA in the heart has been suggested to be routed to SR via vesicles by an ankyrin-B-assisted protein sorting mechanism (32). The route that PLB takes has not yet been identified, but the excess stoichiometry of PLB to SERCA in the heart and its faster protein turnover rate suggests that PLB is transported independently from SERCA.
In summary, a combination of new and established imaging tools has revealed the developmentally regulated trafficking of PLB in the myoblast that is distinct from that associated with the Ca2+-ATPase, its interaction partner in the mature myocyte. These observations suggest possible new roles for PLB in the myoblast and highlight the important cellular regulation associated with the trafficking of PLB in the formation of fully functional SR membranes in which the Ca2+-ATPase is under β-adrenergic control through its association with PLB. Future measurements should focus on understanding the relationship between the trafficking of PLB and the maturation and maintenance of cardiac function.
The research described was conducted under the LDRD Program at the Pacific Northwest National Laboratory: a multiprogram national laboratory operated by Battelle for the U.S. Department of Energy under Contract DE-AC06-76RL01830. Portions of the work were also supported by National Institutes of Health Grants HL-064031 (to T. C. Squier) and NS-23868 (S. T. Brady).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2007 the American Physiological Society