- Email: [email protected]
Targeting of Calsequestrin to the Sarcoplasmic Reticulum of Skeletal Muscle upon Deletion of Its Glycosylation Site
Experimental Cell Research 265, 104 –113 (2001) doi:10.1006/excr.2001.5172, available online at http://www.idealibrary.com on
Targeting of Calsequestrin to the Sarcoplasmic Reticulum of Skeletal Muscle upon Deletion of Its Glycosylation Site Alessandra Nori, Giorgia Valle, Maria Lina Massimino, and Pompeo Volpe 1 Centro di Studio per la Biologia e la Fisiopatologia Muscolare del CNR, Dipartimento di Scienze Biomediche Sperimentali, Universita` di Padova, viale G. Colombo 3, 35121 Padova, Italy
INTRODUCTION The glycoprotein calsequestrin (CS) is segregated to the junctional sarcoplasmic reticulum (jSR) and is responsible for intraluminal Ca 2ⴙ binding. A chimeric CS– hemoagglutinin 1 (HA1), obtained by adding the nine amino acid viral epitope hemoagglutinin to the carboxy terminal of CS and shown to be correctly segregated to skeletal muscle jSR [A. Nori, K. A. Nadalini, A. Martini, R. Rizzuto, A. Villa, and P. Volpe (1997). Chimeric calsequestrin and its targeting to the junctional sarcoplasmic reticulum of skeletal muscle. Am. J. Physiol. 272, C1420 –C1428] lends itself as a molecular tool to investigate the targeting domains of CS. A putative targeting mechanism of CS to jSR implies glycosylation-dependent steps in the endoplasmic reticulum (ER) and Golgi complex. To test this hypothesis, CS-HA1⌬Gly, a mutant in which the unique Nglycosylation site Asn316 was changed to Ile, was engineered by site-directed mutagenesis. The mutant cDNA was transiently transfected in either HeLa cells, myoblasts of rat skeletal muscle primary cultures, or regenerating soleus muscle fibers of adult rats. The expression and intracellular localization of CSHA1⌬Gly was studied by double-labeling epifluorescence by means of antibodies against either CS, HA1, or the ryanodine receptor calcium release channel. CS-HA1⌬Gly was expressed and retained to ER and ER/sarcoplasmic reticulum of HeLa cells and myotubes, respectively, and expressed, sorted, and correctly segregated to jSR of regenerating soleus muscle fibers. Thus, the targeting mechanism of CS in vivo appears not to be affected by glycosylation—that is, the sorting, docking, and segregation of CS are independent of cotranslational and posttranslational glycosylation or glycosylations. © 2001 Academic Press Key Words: calsequestrin; protein targeting; sarcoplasmic reticulum; skeletal muscle; site-directed mutagenesis.
The sarcoplasmic reticulum (SR) 2 of skeletal muscle, a network of tubules and cisternae devoted to intracellular Ca 2⫹ homeostasis [1, 2], is composed of two continuous compartments, the nonjunctional SR, enriched in Ca 2⫹ pump, and the junctional sarcoplasmic reticulum (jSR), juxtaposed to the transverse tubules and enriched in Ca 2⫹ release channel (also referred to as ryanodine receptors, or RYRs) and calsequestrin (CS). CS is a glycoprotein endowed with low-affinity (K d around 1 mM), high-capacity (40 –50 mol/mol) Ca 2⫹ binding properties [3, 4]. CS has been originally detected as electron-dense material confined to the jSR lumen (see Franzini-Armstrong et al. [5] and Jorgensen et al. [6] and references therein) and plays a crucial role in the storage of Ca 2⫹ between uptake and release cycles [1]. Binding and unbinding of Ca 2⫹ to CS (see Kawasaki and Kasai [7] and references therein), as well as changes of the phosphorylation state of CS [8] influence the activity of the RYR-Ca 2⫹ release channel both directly [9] and indirectly. The SR is a subcompartment of the endoplasmic reticulum (ER), as indicated by the coexistence of ER general markers, such as immunoglobulin binding protein (BiP), calnexin, and protein disulfide isomerase, with specific molecular components of Ca 2⫹ stores (e.g., CS, Ca 2⫹ pump, and RYR [2]). The molecular and morphological transition from ER to SR occurs from a wide-mesh membrane network, entails an early and focal concentration of CS, and rapidly evolves into the establishment of triadic couplings between terminal cisternae (TC) and transverse tubules [10, 11]. Because the structure of the N-linked oligosaccharide of mammalian skeletal muscle CS seems to be GlcNAc 2–Man 5GlcNAc [1, 3, 6], and because it seems to be Endo-H insensitive [12], processing through the 2
1 To whom reprint requests should be addressed. Fax: 39-498276040. E-mail: [email protected].
0014-4827/01 $35.00 Copyright © 2001 by Academic Press All rights of reproduction in any form reserved.
Abbreviations used: BiP, immunoglobulin binding protein; CS, calsequestrin; DMEM, Dulbecco’s modified Eagle’s medium; ER, endoplasmic reticulum; HA1, hemoagglutinin 1; jSR, junctional SR; RYR, ryanodine receptor; SDS, sodium dodecyl sulfate; SR, sarcoplasmic reticulum; TC, terminal cisternae.
104
GLYCOSYLATION AND CALSEQUESTRIN TARGETING
Golgi complex appears obvious. A stage through the Golgi complex has indeed been described for CS in developing skeletal myotubes [13]. Gatti et al. [14] recently observed CS in ER subdomains of trasfected L6 myoblasts, found CS to be completely Endo-H sensitive, and proposed that CS never cycles through the Golgi complex. Assuming that CS is indeed processed through the Golgi complex [12, 13], it is not known whether CS reaches SR via cis-medial, Golgi-derived, clathrin-coated vesicles [13] or whether it travels back to the ER, diffuses intraluminally to SR via ER/SR continuities [2, 11], and segregates to the jSR [5, 6]. Whether and how CS cycles through the Golgi complex is at present still an unresolved question. Targeting mechanisms to jSR are not yet known. CS lacks the characteristic carboxy-terminal tetra-peptide KDEL [3] that allows luminal retention, without segregation, to several ER proteins (BiP, protein disulfide isomerase, and calreticulin) shuttling from pre-Golgi and Golgi compartments [15, 16]. One of the leading hypotheses put forth to account for CS segregation to the jSR is based on interactions between integral proteins, restricted in their expression to the jSR [17–20] and able to bind CS (CS-binding protein or proteins) with their luminal domains: putative CS-binding proteins are triadin, junctin, both integral membrane proteins (see Jones et al. [18] and Knudson et al. [19] and references therein), and RYR, shown to form complexes with CS (see Zhang et al. [20] and Murray and Ohlendieck [21] and references therein). This hypothesis also suggests that CS segregation to the jSR is based on complementary recognition sites on CS. Given the primary sequences of CS, mostly acidic, and given the luminal domains of triadin and junctin, mostly basic, electrostatic interactions are likely. However, it is not known which and where such putative CS domains might be and whether cotranslational modifications of CS, posttranslational modifications of CS, or both (e.g., glycosylation or phosphorylation) determine or influence both routing and segregation. There are no specific data concerning the role of N-linked glycosylation in both routing and segregation of CS in skeletal muscle, whereas it is known, for instance, that specific N-linked oligosaccharides determine sorting of lysosomal enzymes [22] and targeting of some secretory proteins [23]. Epitope-tagged CS has been developed to investigate the targeting mechanisms of CS—that is, sorting, routing, retention, docking, and segregation. A DNA fragment coding for the nine amino acids of the influenza virus hemoagglutinin 1 (HA1) was added at the 3⬘ end of the rabbit skeletal muscle CS cDNA. The chimeric cDNA was transiently transfected in HeLa cells, myoblasts of newborn rat, and regenerating skeletal muscles of adult rat. Expression and intracellular localization of CS-HA1 were monitored with both anti-CS and
105
anti-HA1 antibodies, and revealed, in particular, that CS-HA1 segregates to the jSR of regenerating skeletal muscle fibers of adult rats after in vivo transfection of CS-HA1 cDNA [24]. In an attempt to identify CS targeting domains, we have thus engineered CS-HA1⌬Gly, a mutant in which the unique N-glycosylation site of CS (Asn316) was changed to Ile. Such a mutation was intended to abolish the initial ER glycosylation—that is, transfer of the common Glc 3Man 9GlcNAc 2 to the target Asn of the growing polypeptide, and thus to disrupt both cotranslational and posttranslational glycosylations of CS. The present results strongly indicate that glycosylation of CS appears not to be needed for the sorting, routing, retention, docking, and segregation of CS to the jSR in vivo. MATERIALS AND METHODS Generation of CS-HA1 cDNA and CS-HA1⌬Gly cDNA. The cDNA corresponding to rabbit skeletal muscle CS was used [3]. The CSHA1 cDNA was developed by means of a cloning strategy already described [24, 25] with the following PCR primers: forward primer: 5⬘-GAATTCTTAGAGATTCTCAAGTCT-3⬘; and reverse primer 5⬘CTA GAGGCTAGCATAATCAGGAACATCATAGTCATCGTCATC GTCGTCATCGTCT-3⬘. Underlined nucleotides indicate the coding sequence of the HA1 tag [26], and the STOP codon is represented by characters in small caps. The CS-HA1⌬Gly cDNA was generated by site-directed mutagenesis [27] of the 339-bp EcoRI/CS-HA1 cDNA fragment ligated to the pBS⫹ vector. Two single substitutions (T155 3 A and A 158 3 T) were introduced by means of the primer 5⬘-CGCATCAGTAACAA 155TGT 158CGACTCC-3⬘. The generated fragment was initially inserted with sticky ends into EcoRI-cut pBSK⫹CS-HA1 BamHI–EcoRI to create the complete pBSK⫹CS-HA1Gly plasmid. For expression in eukaryotic systems, the chimeric cDNA was isolated and inserted into NotI–XhoI sites of the expression vector pcDNA3 (InVitrogen) downstream of the CMV promoter; the final construct was called pCS-HA1⌬Gly. Orientation and correct sequence of chimeric mutants were checked by restriction assays, and sequence of the synthetic region was obtained by the dideoxy chain termination method [28] by means of modified T7 DNA polymerase. Protein modeling of CS-HA1⌬Gly. Predictions on the tertiary structure of CS-HA1⌬Gly were provided by the SWISS-MODEL Protein Modelling Server [29 –31] and viewed with RasMol 2.6. Three amino acid residues (Gly 3, domain I; Glu 158, domain II; and Val 346, domain III) were chosen to measure relative distances and angles because they are placed in the outermost point of each topological domain [32]. When compared to native CS, distances are as follows: Gly 3–Glu 158, 57.014 vs. 58.831 Å; Gly 3–Val 346, 61.279 vs. 59.523 Å; and Val 346–Glu 158, 51.469 vs. 50.279 Å; and angles are as follows: Gly–Glu–Val, 68.3 vs. 69.8°; Glu–Gly–Val, 51.6 vs. 52.1°; and Gly–Val–Glu, 60.1 vs. 62.2°. Cell cultures. HeLa cells were grown in Dulbecco’s modified Eagle’s medium (DMEM) containing 2 mM glutamine and 10% fetal calf serum. Primary myoblasts were isolated as previously described [25]. Differentiation was obtained by changing the medium to DMEM with 10 to 20% horse serum and then DMEM with 2% horse serum. Bupivacaine-induced necrosis and regeneration of adult rat skeletal muscle. Male adult Wistar rats (⬃220 g body weight) were anesthetized with ketamine (1.5 mg/100 g body weight). Soleus muscles were exposed and injected with 0.5 ml of 0.5% bupivacaine, as
106
NORI ET AL.
previously described [24, 25, 33]. Muscles were removed 10 to 15 days later and either fixed for immunocytochemistry or frozen in liquid nitrogen for biochemical analysis. In agreement with previous reports [34], the local anesthetic bupivacaine induced complete necrosis of the whole soleus by day 3. Regeneration started by day 3 and was completed by day 10 to day 15. Generation of transient transfectants. Twenty-four hours before transfection, either HeLa cells or primary myoblasts were seeded onto 25-mm-diameter wells of a 24-well Corning plate containing a 13-mm-diameter round coverslip with a cell density suitable to obtain 50% confluence at the moment of transfection. Either pCSHA1⌬Gly or the control pcDNA3 vector (4 g/well) were transfected by the calcium phosphate precipitation method, as previously described [24, 25, 33]. Forty-eight hours after transfection, cells were fixed for immunofluorescence; incubation of myoblasts was prolonged and the medium changed for differentiation. Adult rat soleus muscles were exposed 3 days after bupivacaine injection under ketamine anesthesia and injected with about 100 g of plasmid DNA in 20% sucrose. Rats were killed 7 to 12 days later, and both transfected and mock-transfected, contralateral muscles were excised and processed for either immunocytochemistry or biochemical analysis. Immunofluorescence. Cells, myotubes, and skeletal muscles were fixed in paraformaldehyde, as previously described [24, 25, 33]. Both transversal 9-m and longitudinal 6-m sections were obtained from soleus skeletal muscles. Incubation with either polyclonal anti-CS [2] and monoclonal anti-CS (Affinity Bioreagents), monoclonal and polyclonal anti-HA1 antibodies (BabCo and Santa Cruz, respectively), monoclonal anti-RYR1 (Affinity Bioreagents), or polyclonal anti-calreticulin antibodies (Affinity Bioreagents) was performed at room temperature. After washing, either cells or muscle sections were incubated with either rhodamine isothiocyanate or fluorescein-conjugated anti-mouse or anti-rabbit antibodies (DAKO). Images were obtained with a Leica DMRB microscope. Preparation of homogenates from HeLa cells and rat soleus muscle, and of SR vesicles from rabbit skeletal muscle. HeLa cells were cultured as described above, transiently transfected with plasmid DNAs for 2 days, harvested in phosphate-buffered saline, rinsed and lysed in 1 ml of 150 mM NaCl, 15 mM MgCl 2, 1 mM EGTA, 1 mM PMSF, 50 mM HEPES, pH 7.5, 10% glycerol, 1% Triton X-100, 2% sodium dodecyl sulfate (SDS), and 20% mercaptoethanol (buffer B), for 30 min at 0°C while shaking. Homogenates of transfected, regenerating soleus muscle of the rat were obtained as described [33]. Homogenates were kept at ⫺20°C until use. Protein concentration was determined according to Lowry et al. [35]. Purified SR vesicles, referable to CS-enriched TC, were prepared as described [2] from predominantly fast-twitch skeletal muscles of the rabbit. SDS–polyacrylamide gel electrophoresis and Western blot test. SDS–polyacrylamide linear gradient (5–15%) gels, transfer to nitrocellulose sheets, and immunoblot with either anti-CS or anti-HA1 antibodies were carried out essentially as previously described [2, 25]. Materials. DNA modification and restriction enzymes were purchased from Promega except T7 DNA Polymerase, which were purchased from Pharmacia. DMEM, complements, and all other chemicals were purchased from Sigma.
RESULTS
Construction of CS-HA1⌬Gly cDNA To obtain the CS-HA1⌬Gly cDNA, a mutant chimeric CS cDNA encoding for a CS immunologically distinguishable from endogenous CS and lacking the N-glycosylation site (Asn 316), the coding region of rabbit
skeletal muscle CS cDNA was modified by: 1) substitution of a single nucleotide (A 948 3 T) so that the mutated triplet codes for Ile instead of Asn 316; 2) substitution of a single nucleotide (T 945 3 A) so that the mutated triplet codes for Asp instead of Val 315; and 3) addition of a 27-bp fragment coding for the nine amino acids of the influenza virus HA1. Conservative amino acid substitutions were implemented to preserved the 4-sheet configuration of the involved microdomain (residues 312–320), as implied by Fliegel et al. [3] and as shown by the crystal structure of Wang et al. [32]. The tertiary structure of CS-HA1⌬Gly was determined to be only slightly modified in comparison to that of wild-type CS [32] on the basis of computer simulations of protein modeling, as detailed in Materials and Methods. Orientation, sequence, and restriction map of the mutant cDNA were determined by sequencing, and indicated that the synthetic CS cDNA fragments were correctly fused to the EcoRI site and the whole cDNA was an uninterrupted open reading frame; two substitutions were generated in positions 945 and 948; and the T 945 3 A mutation introduced a new restriction site for SalI, valuable in the cloning strategy; the remaining sequence of the CS cDNA corresponded to that of the wild-type CS cDNA; and the 3⬘ end of the CS cDNA had been modified by addition of the tag (HA1) coding sequence. The complete cDNA was transferred by directional cloning downstream of the CMV promoter of the eukaryotic expression vector pcDNA3. Thus, the construct was suitable for transfection and expression of a mutant chimeric CS, immunologically distinguishable from endogenous CS, and suitable for probing the role of the glycosylation site in the targeting mechanism of CS. CS-HA1⌬Gly Expression in Transiently Transfected HeLa Cells Expression of the mutant chimeric CS-HA1⌬Gly in transfected HeLa cells was studied in immunofluorescence experiments with either anti-HA1 antibodies or anti-CS antibodies (Figs. 1A, 1B, respectively). About 20% of transfected cells were strongly CS positive, and the reactivity patterns with the two antibodies were identical; thus, identification by anti-HA1 antibodies was not affected by the overall steric conformation of the mutant protein that could indeed hide the epitope itself. On the other hand, HeLa cells transfected with the empty pcDNA3 vector (mock-transfected cells) were CS negative, as expected from the lack of expression of endogenous CS in HeLa cells, nor did they immunostain with anti-HA1 antibodies (data not shown). Moreover, control, untransfected HeLa cells
GLYCOSYLATION AND CALSEQUESTRIN TARGETING
107
FIG. 1. Immunofluorescence of double-labeled HeLa cells transiently transfected with CS-HA1⌬Gly cDNA. Two days after transfection with either pCS-HA1⌬Gly or pcDNA3 (not shown), cells were fixed and decorated with polyclonal anti-HA1 (A), or with monoclonal anti-CS (B) and polyclonal anti-calreticulin (C) antibodies. Panel D represents the merging of panels B and C. Four different preparations of HeLa cells were analyzed. Panels E–G represent control, untransfected HeLa cells decorated with either anti-HA1 and appropriate secondary antibodies (E), or anti-CS antibodies and fluorescent secondary antibodies followed by excitation at 492 nm (F) and 549 nm (G). Images in panels A:E, B:F, and C:G were obtained under comparable exposure times. Bar, 17 m.
were negative if processed with both anti-HA1 and appropriate secondary antibodies (Fig. 1E). The epifluorescence pattern obtained with both antibodies demonstrated that CS-HA1⌬Gly was expressed and retained to the internal membrane system of HeLa cells and did not have a cytoplasmic
distribution. Moreover, the colocalization signals referable to CS (panel B in Fig. 1) and to endogenous calreticulin (panel C in Fig. 1), a specific ER marker and molecular chaperone [36, 37], allows us to conclude that CS-HA1⌬Gly was within the ER lumen, as judged from labeling overlap (merge image
108
NORI ET AL.
celerated or altered turnover that in turn may lead to misleading interpretation of immunofluorescence data. CS-HA1⌬Gly Expression in SR/ER of Rat Myotubes in Primary Culture
FIG. 2. Western blot of purified SR vesicles (lanes a, b) and of homogenates from HeLa cells (lane c) and rat soleus muscle (lane d) transfected with CS-HA1⌬Gly cDNA (lanes c, d). Purified SR vesicles (5 g) and homogenates (80 g) were separated on polyacrylamide linear gradient (5–15%) gels and transferred to nitrocellulose, as described in Materials and Methods. Immunostaining was carried out with either monoclonal anti-CS (lane a) or polyclonal anti-HA1 (lanes b, c) antibodies. The apparent molecular weight of skeletal muscle CS (lane a) and of CS-HA1⌬Gly (lanes c, d) is about 63 kDa and was determined from a graph of relative mobilities vs. the log of Bio-Rad molecular weight standards, indicated in kilodaltons on the left-hand side. Subtle differences in electrophoretic mobility of CSHA1⌬Gly, due to lack of the N-linked carbohydrate chain, are likely compensated by the HA1 epitope.
of panel D). Control, untransfected HeLa cells showed no labeling with anti-CS antibodies and relevant fluorescent secondary antibodies (cf. Fig. 1, panels F and G). Study of CS-HA1⌬Gly in Homogenates Derived from Transfected HeLa Cells and Soleus Muscle The expression of recombinant CS was also assayed by SDS–polyacrylamide gel electrophoresis and Western blots (Fig. 2). Homogenates from both HeLa cells (lane c) and regenerating soleus muscles (lane d), transfected with CS-HA1⌬Gly cDNA (Fig. 2, lanes c, d), were obtained as described in Materials and Methods. In Western blots, anti-HA1 antibodies recognized CS-HA1⌬Gly as a single protein band having an apparent molecular weight of about 63 kDa (Fig. 2, lanes c, d), virtually identical to that of wild-type CS from rabbit skeletal muscle SR detected by anti-CS antibodies (Fig. 2, lane a). Thus, the epitope recognized by anti-HA1 antibodies was incorporated into the recombinant protein, and no proteolytic breakdown products were detected. The results rule out the possibility that chimeric proteins, as it may happen [38], undergo ac-
The possible consequences of deletion of the glycosylation site were also studied upon transfection of CSHA1⌬Gly cDNA in myoblasts. Myoblasts from 0- to 3-day-old rat hindlimb skeletal muscles were cultured in vitro, induced to differentiate and fuse into multinucleated myotubes (for details, see the Materials and Methods of Nori et al. [24]), and examined 4 days after induction. Expression of CS-HA1⌬Gly was observed in about 15% of rat myotubes by anti-HA1 antibodies, whereas all myotubes were obviously CS-positive, as judged by reactivity with anti-CS antibodies. Figure 3 shows that CS-HA1⌬Gly was arranged in fluorescent strands running parallel to the longitudinal axis of the myotube and had a widespread subcellular distribution, coincident with that of endogenous CS (cf. Fig. 3, panel B and panel A). The merge image of panel C further supports this contention: although a limited green area could be detected in panel C, thus indicating the prevalent expression of endogenous CS in one myotube domain, CS-HA1⌬Gly appears to be localized in the ER/SR of myotubes. CS-HA1⌬Gly Is Targeted to jSR in Regenerating Skeletal Muscle Fibers of Adult Rats The in vivo transfection approach relies on knowledge that bupivacaine injected in the soleus muscle of adult rats [34] causes complete necrosis within 3 days and regeneration in the next 7 to 12 days; the knowledge that satellite cells of regenerating muscles display a higher efficiency of transfection [24]. Thus, transfection of CS-HA1⌬Gly cDNA in the soleus muscle allows us to determine whether and where the mutant chimeric CS-HA1⌬Gly is targeted in vivo at the end of the regeneration phase [24, 25, 33]—in particular, whether it segregates to the jSR. Ten days after bupivacaine treatment, all skeletal muscle fibers were labeled with anti-CS antibodies, whereas only a few fibers were labeled with anti-HA1 antibodies, as judged by immunofluorescence of transverse sections (data not shown) [24, 25]. Thus, a few fibers express the recombinant CS-HA1⌬Gly. Localization of CS-HA1⌬Gly was investigated by immunofluorescence of double-labeled longitudinal sections of soleus muscle fibers. Fig. 4A (anti-CS antibodies) shows the typical CS pattern—that is, parallel rows of bright spots, each resulting from clusters of TC filled up with CS (see Franzini-Armstrong [5] and references therein). Fig. 4B (anti-HA1 antibodies) shows that CSHA1⌬Gly was indeed localized at the A-I interface, as
GLYCOSYLATION AND CALSEQUESTRIN TARGETING
indicated by the regular banding pattern of punctate fluorescence—that is, two rows of triads on either side of the Z line. Merge images (Fig. 4C) unambiguously show colocalization of endogenous CS and of CSHA1⌬Gly at the TC level. Colocalization of CS-HA1⌬Gly and RYR to jSR in Regenerating Skeletal Muscle Fibers of Adult Rats Additional evidence for CS-HA1⌬Gly targeting to jSR was sought by double-labeling experiments with anti-HA1 and anti-RYR antibodies. Fig. 5 shows that in muscle fibers expressing the recombinant CSHA1⌬Gly (panel A), the reactivity with anti-RYR antibodies (panel B) was completely superimposable, as indicated by the merge image of panel C. The regularly repeating bands of staining indicate that CS-HA1⌬Gly and RYR are codistributed at regularly spaced intervals in structures that run transversely across the fiber. The results further suggest that CS-HA1⌬Gly and RYR are located in the jSR. DISCUSSION
Several steps and phases can be envisioned to account for targeting of CS to the jSR of skeletal muscle (i.e., sorting, routing, retention, docking, and segregation). Any single step and phase, in possibly different subcellular compartments (i.e., ER, Golgi complex, and SR) may be conceivably determined by cotranslational glycosylation, posttranslational glycosylation, or both. The mutant CS-HA1⌬Gly was designed so that removal of the unique Asn 316 [3] abolished cotranslational and posttranslational glycosylation while preserving the 4 sheet of CS domain III [32]. Here we show that the mutant CS-HA1⌬Gly is retained to ER compartments in HeLa cells (Fig. 1); sorted and retained to ER/SR of differentiating rat myotubes (Fig. 3); and segregated to jSR of skeletal muscle fibers (Figs. 4, 5) after in vivo transfection of
109
the recombinant cDNA. Strong indications are drawn by experiments concerning in vivo expression of CSHA1⌬Gly cDNA in regenerating skeletal muscle of adult rats: CS-HA1⌬Gly is not only retained to SR but also segregated to jSR. Thus, expression of CSHA1⌬Gly during SR biogenesis, TC development, and triad formation, indicates that deletion of the glycosylation site does not interfere with CS targeting to jSR. Surprisingly, at first glance, analysis of our data indicates that sorting, routing, retention, docking, and finally segregation of CS to jSR are totally independent of glycosylation or glycosylations. In particular, the consolidated issue pertaining to CS cycling through the Golgi complex needs to be reconsidered. To begin with, proper protein folding during translocation into the ER, regulated by several and distinct chaperones [36, 37], might be blocked in CS-HA1⌬Gly. We do not know whether CS-HA1⌬Gly is indeed misfolded in any of our experimental systems. If CSHA1⌬Gly were misfolded and thus destined to degradative compartments, we would have found some evidence of CS-HA1⌬Gly aggregation and coprecipitation with ER chaperonines. Instead, our experiments did not show any conspicuous CS-HA1⌬Gly-containing vacuoles in HeLa cells (Fig. 1; cf. Raichman et al. [38]), and the chimeric CS-HA1⌬Gly, obtained from both HeLa cells and soleus muscles, was found to be a single protein without proteolytic fragments (Fig. 2). Mechanisms for protein folding are so redundant that lectinbased modes [15] might be substituted for by those exerted by BiP and related chaperonines localized in the lumen of both nonjunctional SR and jSR [2, 11, 33]. In fact, knowing the primary sequence of CS [3], knowing the location of N-linked glycans [3], and on the basis of recent evidence about chaperone selection [37], it is likely that folding is ensured by chaperonines of the BiP family through binding to hydrophobic sequences exposed on the surface of incompletely folded CS.
FIG. 3. Immunofluorescence of double-labeled rat skeletal muscle myotubes transiently transfected with CS-HA1⌬Gly cDNA. Transfection was carried out with either pCS-HA1⌬Gly (A, B, C) or pcDNA3 (not shown), and fixed myotubes were sequentially decorated with either monoclonal anti-CS (A) or polyclonal anti-HA1 (B) antibodies. Distribution of the HA1 epitope (red) was detected with anti-HA1 antibodies and rhodamine-conjugated anti-rabbit antibodies; distribution of CS (green) was detected with anti-CS antibodies and fluorescein-conjugated anti-mouse antibodies. Panel C represents the merging of panels A and B. Three different preparations of myotubes were analyzed. Bar, 7 m. FIG. 4. Immunofluorescence of double-labeled soleus muscle fibers after bupivacaine treatment and transfection with CS-HA1⌬Gly cDNA. All sections were obtained 10 days after bupivacaine treatment and 7 days after transfection with CS-HA1⌬Gly cDNA. Double-labeled longitudinal sections were stained sequentially with monoclonal anti-CS antibodies (green in panel A) and polyclonal anti-HA1 antibodies (red in panel B). Note two rows of punctate labeling on either side of Z lines corresponding to TC. Panel C represents the merge of panels A and B. The “green” fiber in the upper left corner of panels A and C is a nontransfected fiber that expresses endogenous CS only. Four different muscle preparations were examined. Bar, 8 m. FIG. 5. Immunofluorescence of double-labeled soleus muscle fibers after bupivacaine treatment and transfection with CS-HA1⌬Gly cDNA. Sections were obtained 15 days after bupivacaine treatment and 12 days after transfection with CS-HA1⌬Gly cDNA. Double-labeled, longitudinal sections were stained sequentially with polyclonal anti-HA1 antibodies (red in panel A) and monoclonal anti-RYR antibodies (green in panel B). Panel C represents the merging of panels A and B. Bar, 6 m.
110
NORI ET AL.
GLYCOSYLATION AND CALSEQUESTRIN TARGETING
It has been established that some misfolded proteins are retained as aggregates that are too large or insoluble to be exported from the ER (see Molinari and Helenius [37] and references therein): if such were the case for CS-HA1⌬Gly, we could have improperly defined ER retention as what is indeed predegradative aggregation in both HeLa cells and rat myotubes. Data on soleus muscle fibers, however, provide evidence for specific segregation of CS-HA1⌬Gly to jSR (Figs. 4, 5): putative aggregates of misfolded CS-HA1⌬Gly would have been retained to the entire SR lumen without segregation to jSR. If glycosylation does not influence CS targeting, routing through the Golgi complex and recycling to either ER or SR appear to be plethoric. This contention does not fit with the reported structure of the N-oligosaccharide, whose terminal N-acetylglucosamine should be transferred in the medial Golgi [39], and it does not fit with reports on the Endo-H insensitivity of CS [12], but it agrees with recent data by Gatti et al. [14], who reported that CS, overexpressed in L6 myoblasts and trapped in ER subdomains, is Endo-H sensitive. By this criterion, it may be excluded from trafficking through the Golgi complex. Further experimental investigation is required to clarify the issue of whether and how CS cycles between ER, the cis-Golgi network, and SR [15, 16]. It can be safely stated, however, that CS-HA1⌬Gly targeting is not influenced by the Golgi complex: either CS-HA1⌬Gly cannot be processed in the Golgi complex if it followed by cargo flow the same path of wild-type CS (ER 3 ER port of exit 3 Golgi complex 3 SR), or it must bypass the Golgi complex, if N-glycosylation were a specific sorting signal for ER export (as shown for other glycoproteins; cf. Martina et al. [40] and Verhoeven et al. [41]), and outrightly be targeted to jSR. Overall, the present results indicate that glycosylation is not necessary for CS sorting and does not interfere with subsequent targeting steps. Segregation of CS to jSR includes multimerization of CS (i.e., homologous protein–protein interactions) and docking (i.e., heterologous protein–protein interactions). Because CS-HA1⌬Gly segregates to jSR of regenerating soleus muscle fibers (Figs. 4, 5), multimerization and docking to jSR appear to be unaffected by the mutagenized glycosylation site. Because Asn 316 is located in the 4 sheet of domain III of CS, a region not directly involved in calcium-dependent CS dimerization [32], CS-HA1⌬Gly multimerization may proceed unabated: kin recognition and multimerization into complexes kinetically and physically excluded from ER export may be the prerequisite for CS retention to ER/SR. We have recently shown that deletion of three distinct phosphorylation sites on CS (Thr 189, Thr 229, and Thr 353) does not affect targeting to jSR in regenerating
111
soleus muscle fibers [33]. Thus, it would appear that both homologous and heterologous protein–protein interactions are not regulated by both cotranslational and posttranslational modifications and are essentially electrostatic. We already approached the “pure” electrostatic hypothesis by means of two CS-HA1 deletion mutants, CS-HA1⌬Glu-Asp and CS-HA1⌬49 COOH, in which the carboxy-terminal acidic tail is made of 14 amino acids (Glu 354-Asp 367), and the mostly acidic (42% Glu ⫹ Asp [3]) 49 residues at the carboxy terminal (Asp 319-Asp 367), respectively, were removed. We found that the mostly acidic COOH terminal of CS, irrespective of length of deletion, is not involved in targeting to jSR [25]. It is plausible that other CS domains are involved—for instance, acidic stretches placed on the surface of each of the three topological domains [32] and the switch points of domain II and III (around residues 228 –229), rich with acidic amino acids [32], as well as the NH 2-terminus [17, 20, 32]. Very recent data concerning the effects of mutation of N-linked glycosylation sites in a variety of proteins— for example, receptors [42], plasma membrane channels [43, 44], ER proteins [45], and lysosomal enzymes [46] have offered a wide spectrum of disparate results. Specific function and targeting are reported to be normal [43, 45], targeting can be altered whereas function is unchanged [42], and both function and targeting can be disrupted [44]. Our results are congruous with complex and diversified roles of N-linked glycosylation because normal targeting for CS-HA1⌬Gly was observed. Which is the role of the N-linked oligosaccharide and of glycosylation if CS-HA1⌬Gly, the mutagenized, carbohydrate-deprived form of CS, still segregates to jSR? CS glycosylation could play distinct roles during SR differentiation. In the early stages of SR biogenesis, the N-linked oligosaccharide could help ER chaperons in “trapping” CS into the ER/SR compartments and enhance CS oligomerization; once SR differentiation is achieved, glycosylation may very well influence Ca 2⫹ binding, phosphorylation, or both, as well as functional interactions with the RYR-Ca 2⫹ release channel [7, 9, 20, 21]. This remains to be assessed in specific experiments that use the purified recombinant CS-HA1⌬Gly. In conclusion, taking into account the present results and those by Gatti et al. [14], along with the observation that CS is found both in the ER lumen and nonjunctional SR during early postnatal development [11], it appears that CS targeting includes events occurring mostly in the ER/SR compartment. This work was supported by Telethon, Grant 1274, and by funds from the Ministero dell’Universita` e della Ricerca Scientifica e Tecnologica (1999 –2001, Programma di ricerca di rilevante interesse nazionale on “Biopatologia della fibra muscolare scheletrica”). We thank G. A. Tobaldin for skillful assistance in surgical procedures, S. Furlan for performing Western blot experiments, and the reviewers for useful comments.
112
NORI ET AL. 16.
Pelham, H. R. B. (1991). Recycling of proteins between the endoplasmic reticulum and Golgi complex. Curr. Opin. Cell Biol. 3, 585–591.
17.
Guo, W., and Campbell, K. P. (1995). Association of triadin with the ryanodine receptor and calsequestrin in the lumen of the sarcoplasmic–reticulum. J. Biol. Chem. 270, 9027–9030.
18.
Jones, L. R., Zhang, L., Sanborn, K., Jorgensen, A. O., and Kelley, J. (1995). Purification, primary structure, and immunological characterization of the 26-kDa calsequestrin binding protein (junctin). from cardiac junctional sarcoplasmic reticulum. J. Biol. Chem. 270, 30787–30796.
19.
Knudson, C. M., Stang, K. J., Moomaw, C. R., Slaughter, C. R., and Campbell, K. P. (1993). Primary structure and topological analysis of a skeletal muscle–specific junctional sarcoplasmic reticulum glycoprotein (triadin). J. Biol. Chem. 268, 12646 – 12654.
20.
Zhang, L., Kelley, J. Schmeisser, G., Kobayashi, Y. M., and Jones, L. R. (1997). Complex formation between junctin, triadin, calsequestrin, and the ryanodine receptor. J. Biol. Chem. 272, 23389 –23397.
21.
Murray, B. E., and Ohlendieck, K. (1998). Complex formation between calsequestrin and the ryanodine receptor in fast- and slow-twitch rabbit skeletal muscle. FEBS Lett. 429, 317–322.
22.
Kornfeld, S. (1987). Trafficking of lysosomal enzymes. FASEB J. 1, 462– 468.
23.
Scheiffele, P., Peranen, J., and Simons, K. (1995). N-glycans as apical sorting signals in epithelial cells. Nature 378, 96 –98.
24.
Nori, A., Nadalini, K. A., Martini, A., Rizzuto, R., Villa, A., and Volpe, P. (1997). Chimeric calsequestrin and its targeting to the junctional sarcoplasmic reticulum of skeletal muscle. Am. J. Physiol. 272, C1420 –C1428.
25.
Nori, A., Gola, E., Tosato, S., Cantini, M., and Volpe, P. (1999). Targeting of calsequestrin to sarcoplasmic reticulum after deletions of its acidic carboxy terminus Am. J. Physiol. 277, C974 –C981.
26.
Flucher, B. E. (1992). Structural analysis of muscle development: Transverse tubules, sarcoplasmic reticulum and the triad. Dev. Biol. 154, 245–260.
Wilson, I. A., Niman, H. L., Houghten, R. A., Cherenson, A. R., Connolly, M. L., and Lerner, R. A. (1984). The structure of an antigenic determinant in a protein. Cell 37, 767–778.
27.
Villa, A., Podini, P., Nori, A., Panzeri, C., Martini, A., Meldolesi, J., and Volpe, P. (1993). The endoplasmic reticulum–sarcoplasmic reticulum connection II. Postnatal differentiation of the sarcoplasmic reticulum in skeletal muscle fibers. Exp. Cell Res. 209, 140 –148.
Kunkel, T. A., Monaco, A. P., Middlesworth, W., Ochs, H. D., and Latt, S. A. (1985). Specific cloning of DNA fragments absent from the DNA of a male patient with an X chromosome deletion. Proc. Natl. Acad. Sci. USA 82, 488 – 492.
28.
Sanger, F., Nicklen, F., and Coulson, A. R. (1977). DNA sequencing with chain-terminating inhibitors. Proc. Natl. Acad. Sci. USA 74, 5463–5467.
Campbell, K. P., MacLennan, D. H., Jorgensen, A. O., and Mintzer, M. C. (1983). Purification and characterization of calsequestrin from canine cardiac sarcoplasmic reticulum and identification of the 53,000 dalton glycoprotein. J. Biol. Chem. 258, 1197–1204.
29.
Guex, N., and Peitsch, M. C. (1997). SWISS-MODEL and the Swiss-PdbViewer: An environment for comparative protein modelling. Electrophoresis 18, 2714 –2723.
30.
Peitsch, M. C. (1995). Protein modelling by e-mail. BioTechnology 13, 658 – 660.
Thomas, K., Navarro, L. R., Benson, R. J. J., Campbell, K. P., Rotundo, R. L., and Fine, R. E. (1989). Newly synthesized calsequestrin, destined for the sarcoplasmic reticulum, is contained in early/intermediate Golgi-derived clathrin-coated vesicles. J. Biol. Chem. 264, 3140 –3145.
31.
Peitsch, M. C. (1996). ProMod and Swiss-Model: Internet-based tools for automated comparative protein modelling. Biochem. Soc. Trans. 24, 274 –279.
32.
Wang, S., Trumble, W. R., Liao, H., Wesson, C. R., Dunker, A. K., and Kang, C. H. (1998). Crystal structure of calsequestrin from rabbit skeletal muscle sarcoplasmic reticulum. Nat. Struct. Biol. 5, 476 – 483.
33.
Nori, A., Furlan, S., Patiri, F., Cantini, M., and Volpe, P. (2000). Site-directed mutagenesis and deletion of three phosphorylation sites of calsequestrin of skeletal muscle sarcoplasmic reticulum: Effects on intracellular targeting. Exp. Cell Res. 260, 40 – 49.
REFERENCES 1.
2.
3.
4.
5.
6.
7.
Lytton, J., and MacLennan, D. H. (1992). In “The Heart and Cardiovascular System” (H. A. Fozzard et al., Eds.), pp. 1203– 1222. New York: Raven Press. Volpe, P., Villa, A., Podini, P., Martini, A., Nori, A., Panzeri, M. C., and Meldolesi, J. (1992). The endoplasmic reticulum– sarcoplasmic reticulum connection: Distribution of endoplasmic reticulum markers in the sarcoplasmic reticulum of skeletal muscle fibers. Proc. Natl. Acad. Sci. USA 89, 6142– 6146. Fliegel, L., Ohnishi, M., Carpenter, M. R., Khanna, V. K., Reithmeier, R. A. F., and MacLennan, D. H. (1987). Amino acid sequence of rabbit fast-twitch skeletal muscle calsequestrin deduced from cDNA and peptide sequence. Proc. Natl. Acad. Sci. USA 84, 1167–1171. Scott, B. J., Simmermann, H. K. B., Collins, J. H., NadalGinard, B., and Jones, L. R. (1988). Complete amino acid sequence of canine cardiac calsequestrin deduced by cDNA cloning. J. Biol. Chem. 263, 8958 – 8964. Franzini-Armstrong, C., Kenney, L. J., and Varriano-Marston, M. (1987). The structure of calsequestrin in triads of vertebrate skeletal muscle: A deep-etch study. J. Cell Biol. 105, 49 –56. Jorgensen, A. O., Kalnins, V. I., Zubrzycka, E., and MacLennan, D. H. (1977). Assembly of the sarcoplasmic reticulum proteins in differentiating rat skeletal muscle cell cultures. J. Cell Biol. 74, 287–298. Kawasaki, T., and Kasai, M. (1994). Regulation of calcium channel of sarcoplasmic reticulum by calsequestrin. Biochem. Biophys. Res. Commun. 199, 1120 –1127.
8.
Szegedi, C., Sarkozi, S., Herzog, A., Jona, I., and Varsanyi, M. (1999). Calsequestrin: More than “only” a luminal Ca2⫹ buffer inside the sarcoplasmic reticulum. Biochem. J. 337, 19 –22.
9.
Herzog. A., Szegedi, C., Jona, I., Herberg, F. W., and Varsanyi, M. (2000). Surface plasmon resonance studies prove the interaction of skeletal muscle sarcoplasmic reticular Ca 2⫹ release channel/ryanodine receptor with calsequestrin. FEBS Lett. 472, 73–77.
10.
11.
12.
13.
14.
15.
Gatti, G., Podini, P., and Meldolesi, J. (1997). Overexpression of calsequestrin in L6 myoblasts: Formation of endoplasmic reticulum subdomains and their evolution into discrete vacuoles where aggregates of the protein are specifically accumulated. Mol. Biol. Cell 8, 1789 –1803. Mellman, I., and Simons, K. (1992). The Golgi complex: In vitro veritas? Cell 68, 829 – 840.
GLYCOSYLATION AND CALSEQUESTRIN TARGETING 34.
Hall-Graggs, E. C. B. (1974). Rapid degeneration and regeneration of a whole skeletal muscle following treatment with bupivacaine (Marcaine). Exp. Neurol. 43, 349 –358.
35.
Lowry, O. H., Rosebrough, N. J., Farr, A. L., and Randall, R. J. (1951). Protein measurement with the Folin phenol reagent. J. Biol. Chem. 226, 9453–9459.
36.
Krause, K. H., and Michalak, M. (1997). Calreticulin. Cell 88, 439 – 443.
37.
Molinari, M., and Helenius, A. (2000). Chaperone selection during glycoprotein translocation into the endoplasmic reticulum. Science 288, 331–333.
38.
Raichman, M., Panzeri, M. C., Clementi, E., Papazafiri, P., Eckley, M., Clegg, D. O., Villa, A., and Meldolesi, J. (1995). Differential localization and functional role of calsequestrin in growing and differentiated myoblasts. J. Cell. Biol. 128, 341– 354, 1995.
39.
Dunphy, W. G., Brands, R., and Rothman, J. E. (1985). Attachment of terminal N-acetylglucosamine to asparagin-linked oligosaccharides occurs in central cisternae of the Golgi stack. Cell 40, 463– 472.
40.
Martina, J. A., Daniotti, J. L., and Maccioni, H. J. (1998). Influence of N-glycosylation and N-glycan trimming on the
Received September 20, 2000 Revised version received January 9, 2001 Published online March 16, 2001
41.
42.
43.
44.
45.
46.
113
activity and intracellular traffic of GD3 synthase. J. Biol. Chem. 273, 3725–3731. Verhoeven, A. J., Neve, B. P., and Jansen, H. (1999). Secretion and apparent activation of human hepatic lipase requires proper oligosaccharide processing in the endoplasmic reticulum. Biochem. J. 337, 133–140. Jayadev, S., Smith, R. D., Jagadeesh, G., Baukal, A. G., Hunydady, L., and Catt, K. J. (1999). N-linked glycosylation is required for optimal AT1a angiotensin receptor expression in COS-7 cells. Endocrinology 140, 2010 –2017. McCormick, K. A., Isom, L. L., Ragsdale, D., Smith, D., Scheuer, T., and Catterall, W. A. (1998). Molecular determinants of Na⫹ channel function in the extracellular domain of the beta1 subunit. J. Biol. Chem. 273, 3954 –3962. Petrecca, K., Atanasiu, R., Akhavan, A., and Shrier, A. (1999). N-linked glycosylation sites determine HERG channel surface membrane expression. J. Physiol. 515, 41– 48. Pan, C. J., Lei, K. J., and Chou, J. K. (1998). Asparagine-linked oligosaccharide are localized to a luminal hydrophilic loop in human glucose-6-phosphatase. J. Biol. Chem. 273, 21658 –21662. Park, H., Vettese-Dadey, M., and Aronson, N. N., Jr. (1996). Glycosylation and phosphorylation of lysosomal glycosylasparaginase. Arch. Biochem. Biophys. 328, 73–77.
Targeting of Calsequestrin to the Sarcoplasmic Reticulum of Skeletal Muscle upon Deletion of Its Glycosylation Site Alessandra Nori, Giorgia Valle, Maria Lina Massimino, and Pompeo Volpe 1 Centro di Studio per la Biologia e la Fisiopatologia Muscolare del CNR, Dipartimento di Scienze Biomediche Sperimentali, Universita` di Padova, viale G. Colombo 3, 35121 Padova, Italy
INTRODUCTION The glycoprotein calsequestrin (CS) is segregated to the junctional sarcoplasmic reticulum (jSR) and is responsible for intraluminal Ca 2ⴙ binding. A chimeric CS– hemoagglutinin 1 (HA1), obtained by adding the nine amino acid viral epitope hemoagglutinin to the carboxy terminal of CS and shown to be correctly segregated to skeletal muscle jSR [A. Nori, K. A. Nadalini, A. Martini, R. Rizzuto, A. Villa, and P. Volpe (1997). Chimeric calsequestrin and its targeting to the junctional sarcoplasmic reticulum of skeletal muscle. Am. J. Physiol. 272, C1420 –C1428] lends itself as a molecular tool to investigate the targeting domains of CS. A putative targeting mechanism of CS to jSR implies glycosylation-dependent steps in the endoplasmic reticulum (ER) and Golgi complex. To test this hypothesis, CS-HA1⌬Gly, a mutant in which the unique Nglycosylation site Asn316 was changed to Ile, was engineered by site-directed mutagenesis. The mutant cDNA was transiently transfected in either HeLa cells, myoblasts of rat skeletal muscle primary cultures, or regenerating soleus muscle fibers of adult rats. The expression and intracellular localization of CSHA1⌬Gly was studied by double-labeling epifluorescence by means of antibodies against either CS, HA1, or the ryanodine receptor calcium release channel. CS-HA1⌬Gly was expressed and retained to ER and ER/sarcoplasmic reticulum of HeLa cells and myotubes, respectively, and expressed, sorted, and correctly segregated to jSR of regenerating soleus muscle fibers. Thus, the targeting mechanism of CS in vivo appears not to be affected by glycosylation—that is, the sorting, docking, and segregation of CS are independent of cotranslational and posttranslational glycosylation or glycosylations. © 2001 Academic Press Key Words: calsequestrin; protein targeting; sarcoplasmic reticulum; skeletal muscle; site-directed mutagenesis.
The sarcoplasmic reticulum (SR) 2 of skeletal muscle, a network of tubules and cisternae devoted to intracellular Ca 2⫹ homeostasis [1, 2], is composed of two continuous compartments, the nonjunctional SR, enriched in Ca 2⫹ pump, and the junctional sarcoplasmic reticulum (jSR), juxtaposed to the transverse tubules and enriched in Ca 2⫹ release channel (also referred to as ryanodine receptors, or RYRs) and calsequestrin (CS). CS is a glycoprotein endowed with low-affinity (K d around 1 mM), high-capacity (40 –50 mol/mol) Ca 2⫹ binding properties [3, 4]. CS has been originally detected as electron-dense material confined to the jSR lumen (see Franzini-Armstrong et al. [5] and Jorgensen et al. [6] and references therein) and plays a crucial role in the storage of Ca 2⫹ between uptake and release cycles [1]. Binding and unbinding of Ca 2⫹ to CS (see Kawasaki and Kasai [7] and references therein), as well as changes of the phosphorylation state of CS [8] influence the activity of the RYR-Ca 2⫹ release channel both directly [9] and indirectly. The SR is a subcompartment of the endoplasmic reticulum (ER), as indicated by the coexistence of ER general markers, such as immunoglobulin binding protein (BiP), calnexin, and protein disulfide isomerase, with specific molecular components of Ca 2⫹ stores (e.g., CS, Ca 2⫹ pump, and RYR [2]). The molecular and morphological transition from ER to SR occurs from a wide-mesh membrane network, entails an early and focal concentration of CS, and rapidly evolves into the establishment of triadic couplings between terminal cisternae (TC) and transverse tubules [10, 11]. Because the structure of the N-linked oligosaccharide of mammalian skeletal muscle CS seems to be GlcNAc 2–Man 5GlcNAc [1, 3, 6], and because it seems to be Endo-H insensitive [12], processing through the 2
1 To whom reprint requests should be addressed. Fax: 39-498276040. E-mail: [email protected].
0014-4827/01 $35.00 Copyright © 2001 by Academic Press All rights of reproduction in any form reserved.
Abbreviations used: BiP, immunoglobulin binding protein; CS, calsequestrin; DMEM, Dulbecco’s modified Eagle’s medium; ER, endoplasmic reticulum; HA1, hemoagglutinin 1; jSR, junctional SR; RYR, ryanodine receptor; SDS, sodium dodecyl sulfate; SR, sarcoplasmic reticulum; TC, terminal cisternae.
104
GLYCOSYLATION AND CALSEQUESTRIN TARGETING
Golgi complex appears obvious. A stage through the Golgi complex has indeed been described for CS in developing skeletal myotubes [13]. Gatti et al. [14] recently observed CS in ER subdomains of trasfected L6 myoblasts, found CS to be completely Endo-H sensitive, and proposed that CS never cycles through the Golgi complex. Assuming that CS is indeed processed through the Golgi complex [12, 13], it is not known whether CS reaches SR via cis-medial, Golgi-derived, clathrin-coated vesicles [13] or whether it travels back to the ER, diffuses intraluminally to SR via ER/SR continuities [2, 11], and segregates to the jSR [5, 6]. Whether and how CS cycles through the Golgi complex is at present still an unresolved question. Targeting mechanisms to jSR are not yet known. CS lacks the characteristic carboxy-terminal tetra-peptide KDEL [3] that allows luminal retention, without segregation, to several ER proteins (BiP, protein disulfide isomerase, and calreticulin) shuttling from pre-Golgi and Golgi compartments [15, 16]. One of the leading hypotheses put forth to account for CS segregation to the jSR is based on interactions between integral proteins, restricted in their expression to the jSR [17–20] and able to bind CS (CS-binding protein or proteins) with their luminal domains: putative CS-binding proteins are triadin, junctin, both integral membrane proteins (see Jones et al. [18] and Knudson et al. [19] and references therein), and RYR, shown to form complexes with CS (see Zhang et al. [20] and Murray and Ohlendieck [21] and references therein). This hypothesis also suggests that CS segregation to the jSR is based on complementary recognition sites on CS. Given the primary sequences of CS, mostly acidic, and given the luminal domains of triadin and junctin, mostly basic, electrostatic interactions are likely. However, it is not known which and where such putative CS domains might be and whether cotranslational modifications of CS, posttranslational modifications of CS, or both (e.g., glycosylation or phosphorylation) determine or influence both routing and segregation. There are no specific data concerning the role of N-linked glycosylation in both routing and segregation of CS in skeletal muscle, whereas it is known, for instance, that specific N-linked oligosaccharides determine sorting of lysosomal enzymes [22] and targeting of some secretory proteins [23]. Epitope-tagged CS has been developed to investigate the targeting mechanisms of CS—that is, sorting, routing, retention, docking, and segregation. A DNA fragment coding for the nine amino acids of the influenza virus hemoagglutinin 1 (HA1) was added at the 3⬘ end of the rabbit skeletal muscle CS cDNA. The chimeric cDNA was transiently transfected in HeLa cells, myoblasts of newborn rat, and regenerating skeletal muscles of adult rat. Expression and intracellular localization of CS-HA1 were monitored with both anti-CS and
105
anti-HA1 antibodies, and revealed, in particular, that CS-HA1 segregates to the jSR of regenerating skeletal muscle fibers of adult rats after in vivo transfection of CS-HA1 cDNA [24]. In an attempt to identify CS targeting domains, we have thus engineered CS-HA1⌬Gly, a mutant in which the unique N-glycosylation site of CS (Asn316) was changed to Ile. Such a mutation was intended to abolish the initial ER glycosylation—that is, transfer of the common Glc 3Man 9GlcNAc 2 to the target Asn of the growing polypeptide, and thus to disrupt both cotranslational and posttranslational glycosylations of CS. The present results strongly indicate that glycosylation of CS appears not to be needed for the sorting, routing, retention, docking, and segregation of CS to the jSR in vivo. MATERIALS AND METHODS Generation of CS-HA1 cDNA and CS-HA1⌬Gly cDNA. The cDNA corresponding to rabbit skeletal muscle CS was used [3]. The CSHA1 cDNA was developed by means of a cloning strategy already described [24, 25] with the following PCR primers: forward primer: 5⬘-GAATTCTTAGAGATTCTCAAGTCT-3⬘; and reverse primer 5⬘CTA GAGGCTAGCATAATCAGGAACATCATAGTCATCGTCATC GTCGTCATCGTCT-3⬘. Underlined nucleotides indicate the coding sequence of the HA1 tag [26], and the STOP codon is represented by characters in small caps. The CS-HA1⌬Gly cDNA was generated by site-directed mutagenesis [27] of the 339-bp EcoRI/CS-HA1 cDNA fragment ligated to the pBS⫹ vector. Two single substitutions (T155 3 A and A 158 3 T) were introduced by means of the primer 5⬘-CGCATCAGTAACAA 155TGT 158CGACTCC-3⬘. The generated fragment was initially inserted with sticky ends into EcoRI-cut pBSK⫹CS-HA1 BamHI–EcoRI to create the complete pBSK⫹CS-HA1Gly plasmid. For expression in eukaryotic systems, the chimeric cDNA was isolated and inserted into NotI–XhoI sites of the expression vector pcDNA3 (InVitrogen) downstream of the CMV promoter; the final construct was called pCS-HA1⌬Gly. Orientation and correct sequence of chimeric mutants were checked by restriction assays, and sequence of the synthetic region was obtained by the dideoxy chain termination method [28] by means of modified T7 DNA polymerase. Protein modeling of CS-HA1⌬Gly. Predictions on the tertiary structure of CS-HA1⌬Gly were provided by the SWISS-MODEL Protein Modelling Server [29 –31] and viewed with RasMol 2.6. Three amino acid residues (Gly 3, domain I; Glu 158, domain II; and Val 346, domain III) were chosen to measure relative distances and angles because they are placed in the outermost point of each topological domain [32]. When compared to native CS, distances are as follows: Gly 3–Glu 158, 57.014 vs. 58.831 Å; Gly 3–Val 346, 61.279 vs. 59.523 Å; and Val 346–Glu 158, 51.469 vs. 50.279 Å; and angles are as follows: Gly–Glu–Val, 68.3 vs. 69.8°; Glu–Gly–Val, 51.6 vs. 52.1°; and Gly–Val–Glu, 60.1 vs. 62.2°. Cell cultures. HeLa cells were grown in Dulbecco’s modified Eagle’s medium (DMEM) containing 2 mM glutamine and 10% fetal calf serum. Primary myoblasts were isolated as previously described [25]. Differentiation was obtained by changing the medium to DMEM with 10 to 20% horse serum and then DMEM with 2% horse serum. Bupivacaine-induced necrosis and regeneration of adult rat skeletal muscle. Male adult Wistar rats (⬃220 g body weight) were anesthetized with ketamine (1.5 mg/100 g body weight). Soleus muscles were exposed and injected with 0.5 ml of 0.5% bupivacaine, as
106
NORI ET AL.
previously described [24, 25, 33]. Muscles were removed 10 to 15 days later and either fixed for immunocytochemistry or frozen in liquid nitrogen for biochemical analysis. In agreement with previous reports [34], the local anesthetic bupivacaine induced complete necrosis of the whole soleus by day 3. Regeneration started by day 3 and was completed by day 10 to day 15. Generation of transient transfectants. Twenty-four hours before transfection, either HeLa cells or primary myoblasts were seeded onto 25-mm-diameter wells of a 24-well Corning plate containing a 13-mm-diameter round coverslip with a cell density suitable to obtain 50% confluence at the moment of transfection. Either pCSHA1⌬Gly or the control pcDNA3 vector (4 g/well) were transfected by the calcium phosphate precipitation method, as previously described [24, 25, 33]. Forty-eight hours after transfection, cells were fixed for immunofluorescence; incubation of myoblasts was prolonged and the medium changed for differentiation. Adult rat soleus muscles were exposed 3 days after bupivacaine injection under ketamine anesthesia and injected with about 100 g of plasmid DNA in 20% sucrose. Rats were killed 7 to 12 days later, and both transfected and mock-transfected, contralateral muscles were excised and processed for either immunocytochemistry or biochemical analysis. Immunofluorescence. Cells, myotubes, and skeletal muscles were fixed in paraformaldehyde, as previously described [24, 25, 33]. Both transversal 9-m and longitudinal 6-m sections were obtained from soleus skeletal muscles. Incubation with either polyclonal anti-CS [2] and monoclonal anti-CS (Affinity Bioreagents), monoclonal and polyclonal anti-HA1 antibodies (BabCo and Santa Cruz, respectively), monoclonal anti-RYR1 (Affinity Bioreagents), or polyclonal anti-calreticulin antibodies (Affinity Bioreagents) was performed at room temperature. After washing, either cells or muscle sections were incubated with either rhodamine isothiocyanate or fluorescein-conjugated anti-mouse or anti-rabbit antibodies (DAKO). Images were obtained with a Leica DMRB microscope. Preparation of homogenates from HeLa cells and rat soleus muscle, and of SR vesicles from rabbit skeletal muscle. HeLa cells were cultured as described above, transiently transfected with plasmid DNAs for 2 days, harvested in phosphate-buffered saline, rinsed and lysed in 1 ml of 150 mM NaCl, 15 mM MgCl 2, 1 mM EGTA, 1 mM PMSF, 50 mM HEPES, pH 7.5, 10% glycerol, 1% Triton X-100, 2% sodium dodecyl sulfate (SDS), and 20% mercaptoethanol (buffer B), for 30 min at 0°C while shaking. Homogenates of transfected, regenerating soleus muscle of the rat were obtained as described [33]. Homogenates were kept at ⫺20°C until use. Protein concentration was determined according to Lowry et al. [35]. Purified SR vesicles, referable to CS-enriched TC, were prepared as described [2] from predominantly fast-twitch skeletal muscles of the rabbit. SDS–polyacrylamide gel electrophoresis and Western blot test. SDS–polyacrylamide linear gradient (5–15%) gels, transfer to nitrocellulose sheets, and immunoblot with either anti-CS or anti-HA1 antibodies were carried out essentially as previously described [2, 25]. Materials. DNA modification and restriction enzymes were purchased from Promega except T7 DNA Polymerase, which were purchased from Pharmacia. DMEM, complements, and all other chemicals were purchased from Sigma.
RESULTS
Construction of CS-HA1⌬Gly cDNA To obtain the CS-HA1⌬Gly cDNA, a mutant chimeric CS cDNA encoding for a CS immunologically distinguishable from endogenous CS and lacking the N-glycosylation site (Asn 316), the coding region of rabbit
skeletal muscle CS cDNA was modified by: 1) substitution of a single nucleotide (A 948 3 T) so that the mutated triplet codes for Ile instead of Asn 316; 2) substitution of a single nucleotide (T 945 3 A) so that the mutated triplet codes for Asp instead of Val 315; and 3) addition of a 27-bp fragment coding for the nine amino acids of the influenza virus HA1. Conservative amino acid substitutions were implemented to preserved the 4-sheet configuration of the involved microdomain (residues 312–320), as implied by Fliegel et al. [3] and as shown by the crystal structure of Wang et al. [32]. The tertiary structure of CS-HA1⌬Gly was determined to be only slightly modified in comparison to that of wild-type CS [32] on the basis of computer simulations of protein modeling, as detailed in Materials and Methods. Orientation, sequence, and restriction map of the mutant cDNA were determined by sequencing, and indicated that the synthetic CS cDNA fragments were correctly fused to the EcoRI site and the whole cDNA was an uninterrupted open reading frame; two substitutions were generated in positions 945 and 948; and the T 945 3 A mutation introduced a new restriction site for SalI, valuable in the cloning strategy; the remaining sequence of the CS cDNA corresponded to that of the wild-type CS cDNA; and the 3⬘ end of the CS cDNA had been modified by addition of the tag (HA1) coding sequence. The complete cDNA was transferred by directional cloning downstream of the CMV promoter of the eukaryotic expression vector pcDNA3. Thus, the construct was suitable for transfection and expression of a mutant chimeric CS, immunologically distinguishable from endogenous CS, and suitable for probing the role of the glycosylation site in the targeting mechanism of CS. CS-HA1⌬Gly Expression in Transiently Transfected HeLa Cells Expression of the mutant chimeric CS-HA1⌬Gly in transfected HeLa cells was studied in immunofluorescence experiments with either anti-HA1 antibodies or anti-CS antibodies (Figs. 1A, 1B, respectively). About 20% of transfected cells were strongly CS positive, and the reactivity patterns with the two antibodies were identical; thus, identification by anti-HA1 antibodies was not affected by the overall steric conformation of the mutant protein that could indeed hide the epitope itself. On the other hand, HeLa cells transfected with the empty pcDNA3 vector (mock-transfected cells) were CS negative, as expected from the lack of expression of endogenous CS in HeLa cells, nor did they immunostain with anti-HA1 antibodies (data not shown). Moreover, control, untransfected HeLa cells
GLYCOSYLATION AND CALSEQUESTRIN TARGETING
107
FIG. 1. Immunofluorescence of double-labeled HeLa cells transiently transfected with CS-HA1⌬Gly cDNA. Two days after transfection with either pCS-HA1⌬Gly or pcDNA3 (not shown), cells were fixed and decorated with polyclonal anti-HA1 (A), or with monoclonal anti-CS (B) and polyclonal anti-calreticulin (C) antibodies. Panel D represents the merging of panels B and C. Four different preparations of HeLa cells were analyzed. Panels E–G represent control, untransfected HeLa cells decorated with either anti-HA1 and appropriate secondary antibodies (E), or anti-CS antibodies and fluorescent secondary antibodies followed by excitation at 492 nm (F) and 549 nm (G). Images in panels A:E, B:F, and C:G were obtained under comparable exposure times. Bar, 17 m.
were negative if processed with both anti-HA1 and appropriate secondary antibodies (Fig. 1E). The epifluorescence pattern obtained with both antibodies demonstrated that CS-HA1⌬Gly was expressed and retained to the internal membrane system of HeLa cells and did not have a cytoplasmic
distribution. Moreover, the colocalization signals referable to CS (panel B in Fig. 1) and to endogenous calreticulin (panel C in Fig. 1), a specific ER marker and molecular chaperone [36, 37], allows us to conclude that CS-HA1⌬Gly was within the ER lumen, as judged from labeling overlap (merge image
108
NORI ET AL.
celerated or altered turnover that in turn may lead to misleading interpretation of immunofluorescence data. CS-HA1⌬Gly Expression in SR/ER of Rat Myotubes in Primary Culture
FIG. 2. Western blot of purified SR vesicles (lanes a, b) and of homogenates from HeLa cells (lane c) and rat soleus muscle (lane d) transfected with CS-HA1⌬Gly cDNA (lanes c, d). Purified SR vesicles (5 g) and homogenates (80 g) were separated on polyacrylamide linear gradient (5–15%) gels and transferred to nitrocellulose, as described in Materials and Methods. Immunostaining was carried out with either monoclonal anti-CS (lane a) or polyclonal anti-HA1 (lanes b, c) antibodies. The apparent molecular weight of skeletal muscle CS (lane a) and of CS-HA1⌬Gly (lanes c, d) is about 63 kDa and was determined from a graph of relative mobilities vs. the log of Bio-Rad molecular weight standards, indicated in kilodaltons on the left-hand side. Subtle differences in electrophoretic mobility of CSHA1⌬Gly, due to lack of the N-linked carbohydrate chain, are likely compensated by the HA1 epitope.
of panel D). Control, untransfected HeLa cells showed no labeling with anti-CS antibodies and relevant fluorescent secondary antibodies (cf. Fig. 1, panels F and G). Study of CS-HA1⌬Gly in Homogenates Derived from Transfected HeLa Cells and Soleus Muscle The expression of recombinant CS was also assayed by SDS–polyacrylamide gel electrophoresis and Western blots (Fig. 2). Homogenates from both HeLa cells (lane c) and regenerating soleus muscles (lane d), transfected with CS-HA1⌬Gly cDNA (Fig. 2, lanes c, d), were obtained as described in Materials and Methods. In Western blots, anti-HA1 antibodies recognized CS-HA1⌬Gly as a single protein band having an apparent molecular weight of about 63 kDa (Fig. 2, lanes c, d), virtually identical to that of wild-type CS from rabbit skeletal muscle SR detected by anti-CS antibodies (Fig. 2, lane a). Thus, the epitope recognized by anti-HA1 antibodies was incorporated into the recombinant protein, and no proteolytic breakdown products were detected. The results rule out the possibility that chimeric proteins, as it may happen [38], undergo ac-
The possible consequences of deletion of the glycosylation site were also studied upon transfection of CSHA1⌬Gly cDNA in myoblasts. Myoblasts from 0- to 3-day-old rat hindlimb skeletal muscles were cultured in vitro, induced to differentiate and fuse into multinucleated myotubes (for details, see the Materials and Methods of Nori et al. [24]), and examined 4 days after induction. Expression of CS-HA1⌬Gly was observed in about 15% of rat myotubes by anti-HA1 antibodies, whereas all myotubes were obviously CS-positive, as judged by reactivity with anti-CS antibodies. Figure 3 shows that CS-HA1⌬Gly was arranged in fluorescent strands running parallel to the longitudinal axis of the myotube and had a widespread subcellular distribution, coincident with that of endogenous CS (cf. Fig. 3, panel B and panel A). The merge image of panel C further supports this contention: although a limited green area could be detected in panel C, thus indicating the prevalent expression of endogenous CS in one myotube domain, CS-HA1⌬Gly appears to be localized in the ER/SR of myotubes. CS-HA1⌬Gly Is Targeted to jSR in Regenerating Skeletal Muscle Fibers of Adult Rats The in vivo transfection approach relies on knowledge that bupivacaine injected in the soleus muscle of adult rats [34] causes complete necrosis within 3 days and regeneration in the next 7 to 12 days; the knowledge that satellite cells of regenerating muscles display a higher efficiency of transfection [24]. Thus, transfection of CS-HA1⌬Gly cDNA in the soleus muscle allows us to determine whether and where the mutant chimeric CS-HA1⌬Gly is targeted in vivo at the end of the regeneration phase [24, 25, 33]—in particular, whether it segregates to the jSR. Ten days after bupivacaine treatment, all skeletal muscle fibers were labeled with anti-CS antibodies, whereas only a few fibers were labeled with anti-HA1 antibodies, as judged by immunofluorescence of transverse sections (data not shown) [24, 25]. Thus, a few fibers express the recombinant CS-HA1⌬Gly. Localization of CS-HA1⌬Gly was investigated by immunofluorescence of double-labeled longitudinal sections of soleus muscle fibers. Fig. 4A (anti-CS antibodies) shows the typical CS pattern—that is, parallel rows of bright spots, each resulting from clusters of TC filled up with CS (see Franzini-Armstrong [5] and references therein). Fig. 4B (anti-HA1 antibodies) shows that CSHA1⌬Gly was indeed localized at the A-I interface, as
GLYCOSYLATION AND CALSEQUESTRIN TARGETING
indicated by the regular banding pattern of punctate fluorescence—that is, two rows of triads on either side of the Z line. Merge images (Fig. 4C) unambiguously show colocalization of endogenous CS and of CSHA1⌬Gly at the TC level. Colocalization of CS-HA1⌬Gly and RYR to jSR in Regenerating Skeletal Muscle Fibers of Adult Rats Additional evidence for CS-HA1⌬Gly targeting to jSR was sought by double-labeling experiments with anti-HA1 and anti-RYR antibodies. Fig. 5 shows that in muscle fibers expressing the recombinant CSHA1⌬Gly (panel A), the reactivity with anti-RYR antibodies (panel B) was completely superimposable, as indicated by the merge image of panel C. The regularly repeating bands of staining indicate that CS-HA1⌬Gly and RYR are codistributed at regularly spaced intervals in structures that run transversely across the fiber. The results further suggest that CS-HA1⌬Gly and RYR are located in the jSR. DISCUSSION
Several steps and phases can be envisioned to account for targeting of CS to the jSR of skeletal muscle (i.e., sorting, routing, retention, docking, and segregation). Any single step and phase, in possibly different subcellular compartments (i.e., ER, Golgi complex, and SR) may be conceivably determined by cotranslational glycosylation, posttranslational glycosylation, or both. The mutant CS-HA1⌬Gly was designed so that removal of the unique Asn 316 [3] abolished cotranslational and posttranslational glycosylation while preserving the 4 sheet of CS domain III [32]. Here we show that the mutant CS-HA1⌬Gly is retained to ER compartments in HeLa cells (Fig. 1); sorted and retained to ER/SR of differentiating rat myotubes (Fig. 3); and segregated to jSR of skeletal muscle fibers (Figs. 4, 5) after in vivo transfection of
109
the recombinant cDNA. Strong indications are drawn by experiments concerning in vivo expression of CSHA1⌬Gly cDNA in regenerating skeletal muscle of adult rats: CS-HA1⌬Gly is not only retained to SR but also segregated to jSR. Thus, expression of CSHA1⌬Gly during SR biogenesis, TC development, and triad formation, indicates that deletion of the glycosylation site does not interfere with CS targeting to jSR. Surprisingly, at first glance, analysis of our data indicates that sorting, routing, retention, docking, and finally segregation of CS to jSR are totally independent of glycosylation or glycosylations. In particular, the consolidated issue pertaining to CS cycling through the Golgi complex needs to be reconsidered. To begin with, proper protein folding during translocation into the ER, regulated by several and distinct chaperones [36, 37], might be blocked in CS-HA1⌬Gly. We do not know whether CS-HA1⌬Gly is indeed misfolded in any of our experimental systems. If CSHA1⌬Gly were misfolded and thus destined to degradative compartments, we would have found some evidence of CS-HA1⌬Gly aggregation and coprecipitation with ER chaperonines. Instead, our experiments did not show any conspicuous CS-HA1⌬Gly-containing vacuoles in HeLa cells (Fig. 1; cf. Raichman et al. [38]), and the chimeric CS-HA1⌬Gly, obtained from both HeLa cells and soleus muscles, was found to be a single protein without proteolytic fragments (Fig. 2). Mechanisms for protein folding are so redundant that lectinbased modes [15] might be substituted for by those exerted by BiP and related chaperonines localized in the lumen of both nonjunctional SR and jSR [2, 11, 33]. In fact, knowing the primary sequence of CS [3], knowing the location of N-linked glycans [3], and on the basis of recent evidence about chaperone selection [37], it is likely that folding is ensured by chaperonines of the BiP family through binding to hydrophobic sequences exposed on the surface of incompletely folded CS.
FIG. 3. Immunofluorescence of double-labeled rat skeletal muscle myotubes transiently transfected with CS-HA1⌬Gly cDNA. Transfection was carried out with either pCS-HA1⌬Gly (A, B, C) or pcDNA3 (not shown), and fixed myotubes were sequentially decorated with either monoclonal anti-CS (A) or polyclonal anti-HA1 (B) antibodies. Distribution of the HA1 epitope (red) was detected with anti-HA1 antibodies and rhodamine-conjugated anti-rabbit antibodies; distribution of CS (green) was detected with anti-CS antibodies and fluorescein-conjugated anti-mouse antibodies. Panel C represents the merging of panels A and B. Three different preparations of myotubes were analyzed. Bar, 7 m. FIG. 4. Immunofluorescence of double-labeled soleus muscle fibers after bupivacaine treatment and transfection with CS-HA1⌬Gly cDNA. All sections were obtained 10 days after bupivacaine treatment and 7 days after transfection with CS-HA1⌬Gly cDNA. Double-labeled longitudinal sections were stained sequentially with monoclonal anti-CS antibodies (green in panel A) and polyclonal anti-HA1 antibodies (red in panel B). Note two rows of punctate labeling on either side of Z lines corresponding to TC. Panel C represents the merge of panels A and B. The “green” fiber in the upper left corner of panels A and C is a nontransfected fiber that expresses endogenous CS only. Four different muscle preparations were examined. Bar, 8 m. FIG. 5. Immunofluorescence of double-labeled soleus muscle fibers after bupivacaine treatment and transfection with CS-HA1⌬Gly cDNA. Sections were obtained 15 days after bupivacaine treatment and 12 days after transfection with CS-HA1⌬Gly cDNA. Double-labeled, longitudinal sections were stained sequentially with polyclonal anti-HA1 antibodies (red in panel A) and monoclonal anti-RYR antibodies (green in panel B). Panel C represents the merging of panels A and B. Bar, 6 m.
110
NORI ET AL.
GLYCOSYLATION AND CALSEQUESTRIN TARGETING
It has been established that some misfolded proteins are retained as aggregates that are too large or insoluble to be exported from the ER (see Molinari and Helenius [37] and references therein): if such were the case for CS-HA1⌬Gly, we could have improperly defined ER retention as what is indeed predegradative aggregation in both HeLa cells and rat myotubes. Data on soleus muscle fibers, however, provide evidence for specific segregation of CS-HA1⌬Gly to jSR (Figs. 4, 5): putative aggregates of misfolded CS-HA1⌬Gly would have been retained to the entire SR lumen without segregation to jSR. If glycosylation does not influence CS targeting, routing through the Golgi complex and recycling to either ER or SR appear to be plethoric. This contention does not fit with the reported structure of the N-oligosaccharide, whose terminal N-acetylglucosamine should be transferred in the medial Golgi [39], and it does not fit with reports on the Endo-H insensitivity of CS [12], but it agrees with recent data by Gatti et al. [14], who reported that CS, overexpressed in L6 myoblasts and trapped in ER subdomains, is Endo-H sensitive. By this criterion, it may be excluded from trafficking through the Golgi complex. Further experimental investigation is required to clarify the issue of whether and how CS cycles between ER, the cis-Golgi network, and SR [15, 16]. It can be safely stated, however, that CS-HA1⌬Gly targeting is not influenced by the Golgi complex: either CS-HA1⌬Gly cannot be processed in the Golgi complex if it followed by cargo flow the same path of wild-type CS (ER 3 ER port of exit 3 Golgi complex 3 SR), or it must bypass the Golgi complex, if N-glycosylation were a specific sorting signal for ER export (as shown for other glycoproteins; cf. Martina et al. [40] and Verhoeven et al. [41]), and outrightly be targeted to jSR. Overall, the present results indicate that glycosylation is not necessary for CS sorting and does not interfere with subsequent targeting steps. Segregation of CS to jSR includes multimerization of CS (i.e., homologous protein–protein interactions) and docking (i.e., heterologous protein–protein interactions). Because CS-HA1⌬Gly segregates to jSR of regenerating soleus muscle fibers (Figs. 4, 5), multimerization and docking to jSR appear to be unaffected by the mutagenized glycosylation site. Because Asn 316 is located in the 4 sheet of domain III of CS, a region not directly involved in calcium-dependent CS dimerization [32], CS-HA1⌬Gly multimerization may proceed unabated: kin recognition and multimerization into complexes kinetically and physically excluded from ER export may be the prerequisite for CS retention to ER/SR. We have recently shown that deletion of three distinct phosphorylation sites on CS (Thr 189, Thr 229, and Thr 353) does not affect targeting to jSR in regenerating
111
soleus muscle fibers [33]. Thus, it would appear that both homologous and heterologous protein–protein interactions are not regulated by both cotranslational and posttranslational modifications and are essentially electrostatic. We already approached the “pure” electrostatic hypothesis by means of two CS-HA1 deletion mutants, CS-HA1⌬Glu-Asp and CS-HA1⌬49 COOH, in which the carboxy-terminal acidic tail is made of 14 amino acids (Glu 354-Asp 367), and the mostly acidic (42% Glu ⫹ Asp [3]) 49 residues at the carboxy terminal (Asp 319-Asp 367), respectively, were removed. We found that the mostly acidic COOH terminal of CS, irrespective of length of deletion, is not involved in targeting to jSR [25]. It is plausible that other CS domains are involved—for instance, acidic stretches placed on the surface of each of the three topological domains [32] and the switch points of domain II and III (around residues 228 –229), rich with acidic amino acids [32], as well as the NH 2-terminus [17, 20, 32]. Very recent data concerning the effects of mutation of N-linked glycosylation sites in a variety of proteins— for example, receptors [42], plasma membrane channels [43, 44], ER proteins [45], and lysosomal enzymes [46] have offered a wide spectrum of disparate results. Specific function and targeting are reported to be normal [43, 45], targeting can be altered whereas function is unchanged [42], and both function and targeting can be disrupted [44]. Our results are congruous with complex and diversified roles of N-linked glycosylation because normal targeting for CS-HA1⌬Gly was observed. Which is the role of the N-linked oligosaccharide and of glycosylation if CS-HA1⌬Gly, the mutagenized, carbohydrate-deprived form of CS, still segregates to jSR? CS glycosylation could play distinct roles during SR differentiation. In the early stages of SR biogenesis, the N-linked oligosaccharide could help ER chaperons in “trapping” CS into the ER/SR compartments and enhance CS oligomerization; once SR differentiation is achieved, glycosylation may very well influence Ca 2⫹ binding, phosphorylation, or both, as well as functional interactions with the RYR-Ca 2⫹ release channel [7, 9, 20, 21]. This remains to be assessed in specific experiments that use the purified recombinant CS-HA1⌬Gly. In conclusion, taking into account the present results and those by Gatti et al. [14], along with the observation that CS is found both in the ER lumen and nonjunctional SR during early postnatal development [11], it appears that CS targeting includes events occurring mostly in the ER/SR compartment. This work was supported by Telethon, Grant 1274, and by funds from the Ministero dell’Universita` e della Ricerca Scientifica e Tecnologica (1999 –2001, Programma di ricerca di rilevante interesse nazionale on “Biopatologia della fibra muscolare scheletrica”). We thank G. A. Tobaldin for skillful assistance in surgical procedures, S. Furlan for performing Western blot experiments, and the reviewers for useful comments.
112
NORI ET AL. 16.
Pelham, H. R. B. (1991). Recycling of proteins between the endoplasmic reticulum and Golgi complex. Curr. Opin. Cell Biol. 3, 585–591.
17.
Guo, W., and Campbell, K. P. (1995). Association of triadin with the ryanodine receptor and calsequestrin in the lumen of the sarcoplasmic–reticulum. J. Biol. Chem. 270, 9027–9030.
18.
Jones, L. R., Zhang, L., Sanborn, K., Jorgensen, A. O., and Kelley, J. (1995). Purification, primary structure, and immunological characterization of the 26-kDa calsequestrin binding protein (junctin). from cardiac junctional sarcoplasmic reticulum. J. Biol. Chem. 270, 30787–30796.
19.
Knudson, C. M., Stang, K. J., Moomaw, C. R., Slaughter, C. R., and Campbell, K. P. (1993). Primary structure and topological analysis of a skeletal muscle–specific junctional sarcoplasmic reticulum glycoprotein (triadin). J. Biol. Chem. 268, 12646 – 12654.
20.
Zhang, L., Kelley, J. Schmeisser, G., Kobayashi, Y. M., and Jones, L. R. (1997). Complex formation between junctin, triadin, calsequestrin, and the ryanodine receptor. J. Biol. Chem. 272, 23389 –23397.
21.
Murray, B. E., and Ohlendieck, K. (1998). Complex formation between calsequestrin and the ryanodine receptor in fast- and slow-twitch rabbit skeletal muscle. FEBS Lett. 429, 317–322.
22.
Kornfeld, S. (1987). Trafficking of lysosomal enzymes. FASEB J. 1, 462– 468.
23.
Scheiffele, P., Peranen, J., and Simons, K. (1995). N-glycans as apical sorting signals in epithelial cells. Nature 378, 96 –98.
24.
Nori, A., Nadalini, K. A., Martini, A., Rizzuto, R., Villa, A., and Volpe, P. (1997). Chimeric calsequestrin and its targeting to the junctional sarcoplasmic reticulum of skeletal muscle. Am. J. Physiol. 272, C1420 –C1428.
25.
Nori, A., Gola, E., Tosato, S., Cantini, M., and Volpe, P. (1999). Targeting of calsequestrin to sarcoplasmic reticulum after deletions of its acidic carboxy terminus Am. J. Physiol. 277, C974 –C981.
26.
Flucher, B. E. (1992). Structural analysis of muscle development: Transverse tubules, sarcoplasmic reticulum and the triad. Dev. Biol. 154, 245–260.
Wilson, I. A., Niman, H. L., Houghten, R. A., Cherenson, A. R., Connolly, M. L., and Lerner, R. A. (1984). The structure of an antigenic determinant in a protein. Cell 37, 767–778.
27.
Villa, A., Podini, P., Nori, A., Panzeri, C., Martini, A., Meldolesi, J., and Volpe, P. (1993). The endoplasmic reticulum–sarcoplasmic reticulum connection II. Postnatal differentiation of the sarcoplasmic reticulum in skeletal muscle fibers. Exp. Cell Res. 209, 140 –148.
Kunkel, T. A., Monaco, A. P., Middlesworth, W., Ochs, H. D., and Latt, S. A. (1985). Specific cloning of DNA fragments absent from the DNA of a male patient with an X chromosome deletion. Proc. Natl. Acad. Sci. USA 82, 488 – 492.
28.
Sanger, F., Nicklen, F., and Coulson, A. R. (1977). DNA sequencing with chain-terminating inhibitors. Proc. Natl. Acad. Sci. USA 74, 5463–5467.
Campbell, K. P., MacLennan, D. H., Jorgensen, A. O., and Mintzer, M. C. (1983). Purification and characterization of calsequestrin from canine cardiac sarcoplasmic reticulum and identification of the 53,000 dalton glycoprotein. J. Biol. Chem. 258, 1197–1204.
29.
Guex, N., and Peitsch, M. C. (1997). SWISS-MODEL and the Swiss-PdbViewer: An environment for comparative protein modelling. Electrophoresis 18, 2714 –2723.
30.
Peitsch, M. C. (1995). Protein modelling by e-mail. BioTechnology 13, 658 – 660.
Thomas, K., Navarro, L. R., Benson, R. J. J., Campbell, K. P., Rotundo, R. L., and Fine, R. E. (1989). Newly synthesized calsequestrin, destined for the sarcoplasmic reticulum, is contained in early/intermediate Golgi-derived clathrin-coated vesicles. J. Biol. Chem. 264, 3140 –3145.
31.
Peitsch, M. C. (1996). ProMod and Swiss-Model: Internet-based tools for automated comparative protein modelling. Biochem. Soc. Trans. 24, 274 –279.
32.
Wang, S., Trumble, W. R., Liao, H., Wesson, C. R., Dunker, A. K., and Kang, C. H. (1998). Crystal structure of calsequestrin from rabbit skeletal muscle sarcoplasmic reticulum. Nat. Struct. Biol. 5, 476 – 483.
33.
Nori, A., Furlan, S., Patiri, F., Cantini, M., and Volpe, P. (2000). Site-directed mutagenesis and deletion of three phosphorylation sites of calsequestrin of skeletal muscle sarcoplasmic reticulum: Effects on intracellular targeting. Exp. Cell Res. 260, 40 – 49.
REFERENCES 1.
2.
3.
4.
5.
6.
7.
Lytton, J., and MacLennan, D. H. (1992). In “The Heart and Cardiovascular System” (H. A. Fozzard et al., Eds.), pp. 1203– 1222. New York: Raven Press. Volpe, P., Villa, A., Podini, P., Martini, A., Nori, A., Panzeri, M. C., and Meldolesi, J. (1992). The endoplasmic reticulum– sarcoplasmic reticulum connection: Distribution of endoplasmic reticulum markers in the sarcoplasmic reticulum of skeletal muscle fibers. Proc. Natl. Acad. Sci. USA 89, 6142– 6146. Fliegel, L., Ohnishi, M., Carpenter, M. R., Khanna, V. K., Reithmeier, R. A. F., and MacLennan, D. H. (1987). Amino acid sequence of rabbit fast-twitch skeletal muscle calsequestrin deduced from cDNA and peptide sequence. Proc. Natl. Acad. Sci. USA 84, 1167–1171. Scott, B. J., Simmermann, H. K. B., Collins, J. H., NadalGinard, B., and Jones, L. R. (1988). Complete amino acid sequence of canine cardiac calsequestrin deduced by cDNA cloning. J. Biol. Chem. 263, 8958 – 8964. Franzini-Armstrong, C., Kenney, L. J., and Varriano-Marston, M. (1987). The structure of calsequestrin in triads of vertebrate skeletal muscle: A deep-etch study. J. Cell Biol. 105, 49 –56. Jorgensen, A. O., Kalnins, V. I., Zubrzycka, E., and MacLennan, D. H. (1977). Assembly of the sarcoplasmic reticulum proteins in differentiating rat skeletal muscle cell cultures. J. Cell Biol. 74, 287–298. Kawasaki, T., and Kasai, M. (1994). Regulation of calcium channel of sarcoplasmic reticulum by calsequestrin. Biochem. Biophys. Res. Commun. 199, 1120 –1127.
8.
Szegedi, C., Sarkozi, S., Herzog, A., Jona, I., and Varsanyi, M. (1999). Calsequestrin: More than “only” a luminal Ca2⫹ buffer inside the sarcoplasmic reticulum. Biochem. J. 337, 19 –22.
9.
Herzog. A., Szegedi, C., Jona, I., Herberg, F. W., and Varsanyi, M. (2000). Surface plasmon resonance studies prove the interaction of skeletal muscle sarcoplasmic reticular Ca 2⫹ release channel/ryanodine receptor with calsequestrin. FEBS Lett. 472, 73–77.
10.
11.
12.
13.
14.
15.
Gatti, G., Podini, P., and Meldolesi, J. (1997). Overexpression of calsequestrin in L6 myoblasts: Formation of endoplasmic reticulum subdomains and their evolution into discrete vacuoles where aggregates of the protein are specifically accumulated. Mol. Biol. Cell 8, 1789 –1803. Mellman, I., and Simons, K. (1992). The Golgi complex: In vitro veritas? Cell 68, 829 – 840.
GLYCOSYLATION AND CALSEQUESTRIN TARGETING 34.
Hall-Graggs, E. C. B. (1974). Rapid degeneration and regeneration of a whole skeletal muscle following treatment with bupivacaine (Marcaine). Exp. Neurol. 43, 349 –358.
35.
Lowry, O. H., Rosebrough, N. J., Farr, A. L., and Randall, R. J. (1951). Protein measurement with the Folin phenol reagent. J. Biol. Chem. 226, 9453–9459.
36.
Krause, K. H., and Michalak, M. (1997). Calreticulin. Cell 88, 439 – 443.
37.
Molinari, M., and Helenius, A. (2000). Chaperone selection during glycoprotein translocation into the endoplasmic reticulum. Science 288, 331–333.
38.
Raichman, M., Panzeri, M. C., Clementi, E., Papazafiri, P., Eckley, M., Clegg, D. O., Villa, A., and Meldolesi, J. (1995). Differential localization and functional role of calsequestrin in growing and differentiated myoblasts. J. Cell. Biol. 128, 341– 354, 1995.
39.
Dunphy, W. G., Brands, R., and Rothman, J. E. (1985). Attachment of terminal N-acetylglucosamine to asparagin-linked oligosaccharides occurs in central cisternae of the Golgi stack. Cell 40, 463– 472.
40.
Martina, J. A., Daniotti, J. L., and Maccioni, H. J. (1998). Influence of N-glycosylation and N-glycan trimming on the
Received September 20, 2000 Revised version received January 9, 2001 Published online March 16, 2001
41.
42.
43.
44.
45.
46.
113
activity and intracellular traffic of GD3 synthase. J. Biol. Chem. 273, 3725–3731. Verhoeven, A. J., Neve, B. P., and Jansen, H. (1999). Secretion and apparent activation of human hepatic lipase requires proper oligosaccharide processing in the endoplasmic reticulum. Biochem. J. 337, 133–140. Jayadev, S., Smith, R. D., Jagadeesh, G., Baukal, A. G., Hunydady, L., and Catt, K. J. (1999). N-linked glycosylation is required for optimal AT1a angiotensin receptor expression in COS-7 cells. Endocrinology 140, 2010 –2017. McCormick, K. A., Isom, L. L., Ragsdale, D., Smith, D., Scheuer, T., and Catterall, W. A. (1998). Molecular determinants of Na⫹ channel function in the extracellular domain of the beta1 subunit. J. Biol. Chem. 273, 3954 –3962. Petrecca, K., Atanasiu, R., Akhavan, A., and Shrier, A. (1999). N-linked glycosylation sites determine HERG channel surface membrane expression. J. Physiol. 515, 41– 48. Pan, C. J., Lei, K. J., and Chou, J. K. (1998). Asparagine-linked oligosaccharide are localized to a luminal hydrophilic loop in human glucose-6-phosphatase. J. Biol. Chem. 273, 21658 –21662. Park, H., Vettese-Dadey, M., and Aronson, N. N., Jr. (1996). Glycosylation and phosphorylation of lysosomal glycosylasparaginase. Arch. Biochem. Biophys. 328, 73–77.