A C 2 Muscle Cell Variant Defective in Transport of the Acetylcholine Receptor to the Cell Surface*

We have previously reported the isolation of variants of the C2 mouse muscle cell line that express reduced amounts of acetylcholine receptors (AChRs) on their surface (Black, R. A., and Hall, Z. W. (1985) Proc. Natl. Acad. Sci. U. S. A. 82, 124-128). One of the variants, T-, makes an approximately normal amount of the AChR but accumulates most of it in an intracel- lular pool. This pool is stable and does not serve as precursor for surface AChR. Surface levels of insulin receptor and transferrin receptor are normal in T-cells, and a normal proportion of total hemagglutinin is expressed on the surface after infection of the T- variant with influenza virus. Pulse-chase experiments and kinetic analysis show: 1) that T- cells synthesize a normal amount of the a subunit but degrade it much more slowly than do wild-type cells; and 2) that newly synthesized a subunit is assembled into the AChR at a normal rate. A small fraction of the assembled AChR in T- cells is transported to the surface with normal kinetics, but most of it remains in an internal pool. This variant may provide an important tool for inves-tigation of the factors that regulate AChR assembly and transport to the surface membrane.

The factors regulating assembly and transport to the cell surface of multimeric membrane proteins are poorly understood (Carlin and Merlie, 1987). One of the proteins whose synthesis and assembly has been studied most intensively is the nicotinic acetylcholine receptor (AChR),' a pentameric transmembrane ion channel consisting of four different but highly homologous subunits that are assembled in the stoichiometric ratio az&6 (McCarthy et al., 1986). The a subunit carries the binding site for acetylcholine and also binds a-* This work was supported in part by postdoctoral fellowships from the Damon Runyon-Walter Winchell Cancer Fund and the National Institutes of Health (to R. A. B.) and by research grants from the National Institutes of Health and the Muscular Dystrophy Association (to 2. W. H). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
II To whom correspondence should be addressed.
Each subunit of the AChR is synthesized as a separate polypeptide chain and is cotranslationally inserted into the endoplasmic reticulum (ER), where the signal sequence is removed and the polypeptide N-glycosylated (Anderson and Blobel, 1981;Merlie et al., 1981Merlie et al., , 1982. The subunits are then assembled in the ER to form the oligomeric AChR (Smith et al., 1987). Assembled AChRs are transported to the Golgi apparatus (Fambrough and Devreotes, 1978) where the Nlinked oligosaccharides on the y and 6 subunits are processed to complex forms (Gu and Hall, 1988). The receptor is then transported to the cell surface. Assembled AChR is first detected in the ER about 15 min after the synthesis of a-and 8-polypeptide chains (Merlie and Lindstrom, 1983;Smith et al., 1987), and most of the assembled AChR appears on the surface between 50 and 160 min after synthesis (Devreotes et al., 1977;Smith et al., 1987).
The a subunit undergoes an early maturation step before assembly occurs. Shortly after synthesis, the primary translation product of the a subunit undergoes a change in its conformation that is reflected by the acquisition of a-BuTxbinding activity and by changes in its immunological properties. This transition occurs in the ER, after the addition of N-linked oligosaccharides and the early trimming reactions, but before receptor assembly (Merlie et al., 1982;Smith et al., 1986). Only a small portion (30%) of the synthesized a subunit is assembled into oligomeric AChR the remainder is rapidly degraded (Merlie and Lindstrom, 1983;Smith et al., 1987).
We have recently described genetic variants of the C2 mouse muscle cell line that are defective in expression of the AChR on the cell surface (Black and Hall, 1985). One of these variants, T-, appears to make a normal amount of the AChR but accumulates most of it in an intracellular pool. We report here experiments whose purpose is to explore the functional defect in this variant by characterizing AChR assembly and transport to the cell surface.

DISCUSSION
The reduced binding of a-BuTx to intact Tmyotubes (Black and Hall, 1985) is apparently the result of a decreased number of functionally normal AChRs in the surface membrane. Both the amount of surface a subunit detected by a monoclonal antibody and the rate of carbachol-stimulated Portions of this paper (including "Materials and Methods," "Results," Tables I-IV,   are presented in miniprint at the end of this paper. Miniprint is easily read with the aid of a standard magnifying glass. Full size photocopies are included in the microfilm edition of the Journal that is available from Waverly Press.

11952
22Na influx into Tmyotubes are reduced in proportion to the decrease in a-BuTx binding. In addition, the kinetics of the ~-B u T x binding reaction and the half-life of surface AChR are similar in wild-type and Tcells.
Examination of detergent extracts of Tcells demonstrate that the cells contain toxin-binding activity in an internal compartment that is inaccessible to externally added cu-B~Tx. This internal pool apparently consists of fully assembled 9 S AChR that is accumulated to levels that are three to five times those in wild-type myotubes. Two Functionally Distinct Pools of Internal AChR-Intracellular AChR was first detected by Devreotes and Fambrough (1975) in primary cultures of chick myotubes, who found that it constituted about 30% of the total cellular AChR. They also showed that the internal AChR consists of two pools. The larger pool contains precursor for surface AChR, in the absence of protein synthesis continued transport of AChR to the surface depletes this pool within several hours. The second pool, which they termed the "hidden pool," constitutes about 25% of the total intracellular AChR, and is not depleted when protein synthesis is inhibited. Our experiments show that C2 cells behave similarly. In this case, approximately 80% of the internal AChR serves as precursor for surface AChR which is transported to the surface over a period of a few hours. This scheme offers two possible explanations for the increased proportion of internal receptor in the Tvariant. One possibility is that a kinetic block slows the movement of the AChR to the cell surface, causing AChR to accumulate in the precursor pool. If this were the case, inhibition of protein synthesis should have little effect on the rate of AChR appearance on the surface as the large precursor pool slowly depletes over many hours. A second possibility is that the accumulated AChR is not in the precursor pool, but in another pool, perhaps resembling the hidden pool in its properties. Our experiments clearly indicate that most of the accumulated AChR in Tmyotubes does not serve as precursor for surface AChR. Tcells have a precursor pool, which is smaller than in normal cells, but which, in the absence of protein synthesis, is depleted by transport to the surface at normal rates. The remaining AChR, approximately 70% of the total internal AChR, remains in an internal pool that is unaffected by protein synthesis inhibitors.
Pulse-chase experiments confirmed the results obtained with cycloheximide. There was no difference between wildtype and Tcells in the kinetics of transport of newly assembled AChR to the surface. A striking difference was observed, however, between the proportion of newly assembled AChR appearing on the surface in the two cell types. Whereas almost all of the AChR in wild-type cells is transported to the surface, only a small proportion of the AChR that is made by Tcells is destined for the cell surface; the rest remains internal for many hours.
The decreased efficiency of transport of the AChR to the surface does not appear to be a general defect. Thus, Tmyotubes contain a normal amount of insulin receptor and transferrin receptor on their surface and transport hemagglutinin to the surface with normal efficiency after infection with influenza virus.
AChR Assembly in T -Cells-Because of the availability of a monoclonal antibody that recognizes all forms of the CY subunit (Smith et al., 1986(Smith et al., , 1987, we were able to use pulsechase experiments to follow its synthesis and assembly into intact receptor. These experiments showed that synthesis of a subunit, conversion to a toxin-binding form, and assembly into intact AChR occur at similar rates in wild-type and Tcells. A major difference, however, is the slower degradation of subunit in Tcells. Thus, a larger amount of a subunit was found in all three forms: as ao, the primary translation product, as aT, the 5 S toxin-binding form, and as assembled AChR. The Tcells are not completely lacking the ability to rapidly degrade CY subunit, however, as a subunit lacking Nlinked oligosaccharides, or with abnormal N-linked sugars, is quickly broken down (Fig. 7).
These experiments lead to two important insights related to AChR synthesis and assembly. First, although earlier experiments in BCSH-1 cells had shown that most newly synthesized CY subunit was degraded, it was unclear whether extensive degradation of the a. form, the aT form, or both, occurred. The accumulation of a. to a greater extent than either of the other forms in Tcells suggests that this is the form that is normally degraded.
Second, the accumulation of a subunit in Tcells to a level that is approximately %fold that found in normal cells results in increased formation of both aT and assembled AChR. The proportion of total a that is converted to assembled AChR is in fact almost the same in wild-type and Tcells. These results suggest that the availability of a subunit is at least one limiting factor in the assembly of the AChR and that the degradation of the CY subunit does not occur because it is in excess. Other subunits are probably also degraded in wildtype cells.
The Defect in T -Cells-The experiments reported here demonstrate that the Tvariant of C2 muscle cells has two apparently paradoxical defects in AChR assembly and transport. The first is decreased degradation of newly synthesized a subunit, resulting in increased a levels and, consequently, higher levels of newly assembled AChR. In spite of the increased amount of newly assembled AChR, however, Tcells transport less AChR to the surface than do wild-type cells.
Recent experiments indicate that the AChR that is not transported to the surface accumulates in a pre-Golgi compartment that is probably the endoplasmic reticulum (Gu et aL, 1989).
Are these two defects linked? The best evidence for such a relationship would be the demonstration that phenotypic reversion of the Tstrain results in the loss of both defects.
In the absence of such evidence, the relationship of these properties must remain speculative. One possibility is that degradation normally removes a subunit polypeptides that have folded incorrectly and thus prevents their incorporation into the oligomeric AChR. A defect in degradation might then lead to the production of AChRs that are not competent for transport to the surface. It should be noted, however, that primary myotubes do not break down unassembled a subunit and have no apparent defect in transporting AChR to the surface (Carlin et al., 1986). Because the defect in transport appears to be specific for the AChR, it most likely arises from a mutation in the structural gene for one of the subunits. If this proves to be the case, it will be of interest to know what distinguishes the two pools of AChR, one that is competent for transport to the surface and one that is not.  For precipitation of 01 subunit urth toxin-binding activity. the extracts were incubated with OI-BUTX coupled to Sepharose. Alpha Subunits assembled into intact AChR were determined by lmmunapreclpiCation vlth Sepharose coupled to wAb 88R. an antlbody t h a t reCagnlze5 I and 6 subunits I i r o e h n e i et a l . , 1983; Gu and ,Hall,19U8) In experiments in uhlch surface and lnternal RChRS were to be determined, the receptors. C e l l e x t r a c t s were then preclprtafed e i t h e r *ith rabbit-anti-toxin rnrdcr cells were incubated w i t h non-radioactive t o x i n to label Surface sntibodres co yield s u r f a c e receptor or w i t h toxin-Sepharose to yield internal (I activity, and precipitates bath the 5 . 5 species that binds toxin and the assembled suhunlt.

SvpplemenLal Material to
Toxin-Sepharose rccognrzes a l l OL subunit species with toxin-bindinq R C h R . X 1 1 preciplrates were washed 3 times ai room temperature with 1.0 M NaCI, 1% Triton x-100. 50 mM Trrs-HC1, pH ' 1 . 4 and dlssalved i n SDS sample buffer for SDS-PAGE analyeis. protocols was monitored by using 1251-m-BUTX-llheled surface AChR as a standard.
i n some experiments. the recovery of AChRs with dlffezent precipitation The recovery of radioactivity 10 the precipitate indicated the efficiency O f each preciplration protocol a n d was used to normalize data from different preclpltations.
The efficiency of precipltation with a-BuTx-Sepharose Lld5 estimated by determining AChR concentration before and a f t e r precipitation ulth the filtration assay. Other methods --Carbachol-induced 22Na uptake into C2 myotubes was performed of radioactivity from the cell9 a s described in Gu et a l . ( 1 9 8 5 ) . The was determined by labeling Surface AChR with 1 2 5 1 -~-B u T~ and follouinq the loss association rate constant of toxin binding for Surface RChR was determined using t h e lntegrated equations for bimolecular reactions a s described in Gu et a l .
Velocity sedimentation analysis of taxin-hinding activity i n C2 nyorubes was performed as described 19 Berg and Hall I19751 with 5.20% 5ucrose gradients.

RESULTS
SurCace AChR --The Tvarlant was Orlglnally Isolated from rnutdgenlzed CZ c e l l s on the b a s , $ Of reduced bindlng o f o-bungarotoxin ln-Butxi to the surface of differentiated myotubes 18lack and H a l l , 1 9 8 5 1 . When measured quantitatively, the amount of o-BuTx hvund to intact T myotubes was about one-fifth that bound t o wlld-type myotubes [ Table 11. That the reduction ~n toxln-binding acrlviry was due to d reduced amount of Burface AChR was conflrrned by the Observation that i n t a c t myotubes in T-c u l t u 1 e 5 a l s o hound reduced amounts of mRb 210 (Table I] monoclonal antibody that recoqniies a n exferndl domain of the 0 subunit of the Table I  InEeznal dChR --TO measure internal AChR, Surface receptors were blocked by Incubation Of intact myotuhes with Unlabeled o-BuTx before extraction and assay with 1251-a-BuT~ in a f i l t r a t i o n assay IBIOCkeS and Hall, 1 9 1 5 1 In Contrast to surface AChR, the amount of internal receptor in T-myotubes was 3-5 times higher than that Of wild-type cell5 (Table X I . To determine if the internal toxinbinding activity represented fully assembled AChR, extracts obtained from Tcells were subjected to velocity sedimentation analysis. Toxin binding activity "as found as a single peak whose sedimentation coefricient 19s) was lndlstinquishable from that of wild-type AChR (Figure 1 1 . The accumulated toxinbinding activity in Tmyotubeo thus appears to be fully assembled AChR. surface Of T-myotubes? The amounts Of Surface insulin receptor and fransferrln other surface proteins --Is there a general deficit Of proteins on the receptor as determined by binding assays were not different i n wild-type and Tmyotube CUltureS (Table 1111. Although Other, unassayed Surface proteins may have altered levels, it 1s unlikely from the5e reSultS that the deficlt is a general one. The fact that T-cell5 grow and fuse normally (Slack and Hall, 1 9 8 5 1 also argues against a general defect that reduces the l e v e l s Of cell SUrfdCe proteins.  i t a l l , Unpublished observatiansl. When the Surface and lnternal AChR were assayed. t h e p r o p a r t l m of AChR that was internal l n Co-Cultures was found t o be inteimediate t o the Yalues found I n wlld-type or T-cells cultured alone ( Table   1 " ) .
Moreover, when t h e ratio Of wild-type and variant cells was varied. the amount of surface AChR expressed was approxlmately proportional to the fraction of vlld-type cells I" the culture. These experiments suggest that neither the wild-type nor varlant phenotype is d o r n l n a n t , and that t h e products O f each nucleus behave Independently.  In the case of Tcultures, the ZeSultS were similar 2" that there was a the appearance pf new receptor an the surface. The proportion of internal AChR that Was depleted under these conditions, however, wa5 clearly different from that found ln uild-type cells.
whereas over 80% of the internal AChR was deplete0 in wlld-type cells, only about 30% was lost in Tcells. The ocher 70% rernalned ~n an internal Pool whose concentration dld not chanqe over the course of the experiment I F l g . 3 a j . When the result5 Of this experiment w e~e plotted on d semi-iagarlthmic plot IFiy. 3bl. the half-life O f the rapidly depleted portion o f the internal pool was abouc 2 hr for both T-and wild-type cells.
in she Internal mol and on the surface of invotubes was determined by Dulse-chase Pulse-ChdSe klnetics --The klnerics of appearance of newly synthesized AChR experiments. Wild-cype or T-myotubes were incubated with 35S-merhionine and cysteine for 10 min. followed by unlabeled amlnO acids for various peiiods of tlrne.
Unlabeled o-BuTx was present throughout the Incubation to label Surface AChRa. for each time point, the total amount of newly synthesized 01 SUbUnir ~8 5 determined by ~rnmunoprecipltation with mAb 61, a monDclOnal antibody that recognlres both the prmary trdnsiation product of o. and the assembled AChR ;pm/mg for v i l e cype and Tcells, respectlveiy. Each data paint is the average of 4 determinations. Arrows Inaicate the trme Of cyclohexlmide addition.
lbj The rapidly deQleted portion of the internal AChR 180% and 30% Of fatal for Wlld by mAb 6 1 *a9 a t a maximum rmmediately after the pulse with labeled amino acid3 In wlld-type myotubcs, the amount Of radloactive 01 subunit imunoprecipitatea approximately 301 of the initial value six hr after the pulse. These results are ( fig. 4 A l . It initially declined rapidly, then more ~lorly, t o a l e v e l of similar to those of Merlie and colleagues, who fovnd in BC3H-1 cells chat approximately 10% of the e subunit does not assemble info RChR, but is rapidly degraded IMerlie, et d l . , 1983; Olson eC al., 19841. In Chis experiment the amount of labeled a subunit synthesized by Tmyorubes during the pulse Sllghtly 10wez than that fovnd with wild-type myotubes ! Fig. 4 A l . It declined very little after the pulse, however. 5 0 Chat at the end of 6 ht chase the mount of labeled LT Subvnit remaining vas substantially higher than in wild-type cells.
Thus there 19 a clear difference i n the acCUmulafion of labeled OL Subunit by wild-type and varlant c e l l s . At various times after the chase, the cells were lysed and equal aliquots Of the lysate precipitated with wAb 61 (AI, a-BUT%-Sepharose (BI or mAb 88U-Sepharose IC1 as described under Methods. Samples were analyzed on SDS gels and fluoroqiaphed. The Lop panels Show autoradiographs of the gels that had been exposed for 10 hr ( A I or 21 hr 10 and Cl.
The positions of (1 and P subunits on The gels were marked. Appropriately exposed gels Were Open Symbols, wild type cells; closed symbols.

Tvariants.
was complete within about one hr after synthesis. Little or no degradation In wild-type myotubes, the acquisition of toxin-binding activity by D Subunit occurred after this point (Fig. 5Dl. A similar time-course was seen in Tmyotubes, with the exception that a ma11 amount of breakdown at later rimes may have occurred (Fig. 5 E l . In any case. the absolute amount of labeled a Subunit with toxin-binding activity was greater in T-than in wild-type myotubes. When the gels from these experiments are examined. 0 subunit isolated by O-BYTx-

more
Subunit relative to D at later timer. indicating that the toxin-binding SepharOSe can be seen to be aSSOCiaZed with P subunit (Fig. 5A and 581 rig. 6 . Efficiency Of AChR assembly in wild type (AI and T-(E) cells. The data in Figure 5 were used to calculate the amount Of (I subunit precipitated by a-BuTX-SepharOse ISqUdreSl and miib 888-Sepharose Itrianqlesl expressed as a percentage of total n Subunit determined by mlib 6 1 precipitation. Symbols used are the same as in Figure 5 . measured by isolating ihe AChR with mAb 888-Sepharose. Quantitation of the The assembly Of newly-synthesized a and p Sumnits was more quantitatively labeled a subunir assembled into AChR in wild-type cells Shoved that assembly was complete by about two hr (Fig. 5DI. Similar TeSYltS Were Seen for T-myotubes greater, however, in T-myotubes than in wild-type myotubes. The proportion of ( Fig. 5 E I . The total amount of neuly-synthesized ( I subunit in assembled AChR was total labeled (I subunit that acquired toxin-binding activity End that was assembled into intact AChR was about the same in T-and wild-type c e l l s (Fig. 6). Thus the Steps from synthesis of the 0 Subunit to assembly Of the AChR all appear LO occur normally in T-cells with One exception: D Subunit is broken down less rapidly.
a Subunit degradation in the absence O f qlycosylation --Merlie e t al.
(19821 have shown that (1 subunit Synthesized in the presence Of tunicsmycin (TU) does not acqvire the ability to bind Ct-BuTx with high affinity and is rapidly degraded. Inhibition of trimming Of glucose residues on N-linked o l~g o s a~~h a r i d e~ with deoxynojirimycin IdNMI also increases the degradation of the D subunit (Smith et al., 19861. TO learn If T-cells lack the capacity to break down 0 Subunit under all conditions, we incubated T-myotube Cultures with and analyzed the newly synthesized n . Subunit immediately after the pulse or 2 hr 1 mM dNM or with Zltg/ml TM for 2 hr, then gave a 10-min pulse Of fsS-amino acid, l a t e r . Comparable results were obtained with both T-and wild-type cells. In Chase period (Fig. 7, l a n e s 3 and 4 1 , compared to about 308 degradation in the the dNM treated T-cell3 Over 80% Of the D SubUnlt was deqraded durlng the two-hr dbs?nCe Of the drug (Fiq. 7, lanes 1 and 21. In TM-treated Cells, over 90% was dcqradcd (Fig. 7 , l a n e s 5 and 6 1 .
The altrred mobility of the n subunit ~nrl>cated Char N-llnkcd <ilyro?ylerion O r its early processing had been blocked. T-cclls thus a r~ capablo of drgradinl n ;,>bunit rapidly under These conditlons.