help button home button Endocrine Society Endocrinology
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Endocrinology, doi:10.1210/en.2007-0789
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Supplemental Data
Right arrow All Versions of this Article:
148/12/5831    most recent
Author Manuscript (PDF)
Right arrow Purchase Article
Right arrow View Shopping Cart
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow Request Copyright Permission
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Merz, W. E.
Right arrow Articles by Berger, P.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Merz, W. E.
Right arrow Articles by Berger, P.
Endocrinology Vol. 148, No. 12 5831-5841
Copyright © 2007 by The Endocrine Society

Nonassembled Human Chorionic Gonadotropin Subunits and {alpha}{alpha}-Homodimers Use Fast-Track Processing in the Secretory Pathway in Contrast to {alpha}ß-Heterodimers

Wolfgang E. Merz1, Jean-Michel Krause1, Jordi Roig, Vinod Singh2 and Peter Berger

Heidelberg University Biochemistry Center (W.E.M., J.-M.K., J.R., V.S.), 69120 Heidelberg, Federal Republic of Germany; Institute for Biomedical Aging Research (P.B.), Austrian Academy of Sciences, Innsbruck A6020, Austria; and Hormone Biochemistry Laboratory, Institute of Self Organizing Systems and Biophysics (V.S.), North-Eastern Hill University, Permanent Campus, Shillong-793022, Meghalaya, India

Address all correspondence and requests for reprints to: Wolfgang E. Merz, Ph.D., Heidelberg University Biochemistry Center, Im Neuenheimer Feld 328, 69120 Heidelberg, Federal Republic of Germany. E-mail: wolfgang.merz{at}bzh.uni-heidelberg.de.


   Abstract
Top
 Abstract
Introduction
 Materials and Methods
 Results
 Discussion
 References
 
In multimeric glycoproteins, like glycoprotein hormones, mutual subunit interactions are required for correct folding, assembly, and transport in the secretory pathway. However, character and time course of these interactions need further elucidation. The influence of the glycoprotein hormone {alpha}-subunit (GPH{alpha}) on the folding of the human chorionic gonadotropin (hCG) ß-subunit (hCGß) in hCG {alpha}ß-heterodimers was investigated in [35S]Met/Cys-labeled JEG-3 cells. Completeness of disulfide bridge formation during the time course of folding was estimated by labeling with [3H]N-ethylmaleinimide of free thiol groups not yet consumed. Subunit association took place between immature hCGß (high 3H/35S ratio) and almost completely folded GPH{alpha}. Analysis revealed a highly dynamic maturation process comprising of at least eight main hCGß folding intermediates (molecular masses from 107 to 28 kDa) that could be micro-preparatively isolated and characterized. These hCGß variants developed while being associated with GPH{alpha}. The 107-kDa variant was identified as a complex with calnexin. In contrast to hCG {alpha}ß-heterodimers, free nonassociated hCGß, free large GPH{alpha}, and GPH{alpha}{alpha} homodimers showed a fast-track-like processing in the secretory pathway. At 10 min before hCG secretion, sialylation of these variants had already been completed in the late Golgi, whereas hCG {alpha}ß-heterodimers had still not arrived medial Golgi. This shows that the GPH{alpha} in the hCG {alpha}ß-heterodimers decelerates the maturation of the hCGß portion in the heterodimer complex. This results in a postponed approval of hCG {alpha}ß-heterodimers by the endoplasmic reticulum quality control unlike GPH{alpha}{alpha} homodimers, free hCGß, and GPH{alpha} subunits.


   Introduction
Top
 Abstract
 Introduction
Materials and Methods
 Results
 Discussion
 References
 
THE GLYCOPROTEIN hormone human chorionic gonadotropin (hCG) is one of the most intensely investigated oligomeric glycoproteins. It represents a heterodimer of an {alpha}-subunit common to all glycoprotein hormones (referred to as GPH{alpha}) and a hormone-specific ß-subunit (hCGß) (1). The crystal structure of deglycosylated hCG (2, 3) as well as the carbohydrate structures (for review, see Refs. 1 and 4) have been resolved. The hCG is mainly synthesized and secreted by the placenta, by various tumors (1), and in small amounts by other organs, including the pituitary (5). Mature GPH{alpha} normally has two N-linked carbohydrate chains attached to Asn52 and Asn78. The hCGß has two N-glycosylated sites, at Asn13 and Asn30, and at its carboxy-terminal domain 4 O-linked glycans, linked to Ser 121, Ser127, Ser132, and Ser138. The Asn13 glycan seems to have little effect on secretion in contrast to the Asn30 glycan. The impact of the glycan residues on the steroidogenic activity of hCG is the opposite (6). Both hCG subunits show a considerable variation in the composition of the carbohydrate parts concerning presence and completeness of the complex, hybrid, multiantennary structures, fucosylation of the N-linked glycans, and with respect to sialylation (for a review of the N-glycan structures, see Ref. 4).

Correct folding, disulfide arrangements, and proper glycosylation are mandatory for the assembly of the subunits and the biological functions of hCG. Both hCG subunits belong to the cystine knot growth factor superfamily (7). The formation of disulfide bridges is favored in the endoplasmic reticulum (ER) due to the oxidative milieu in this compartment and occurs in part with the aid of protein disulfide isomerases (PDIs). The disulfide-folding pathway is of particular interest because disulfide bridges are not only elements that stabilize the conformation of a protein but may also act as a functional or structural switch (8, 9). A redox-regulated disulfide switch also seems to be critical for the assembly of the hCG {alpha}ß-heterodimer (10). N-glycosylation and disulfide pathways are linked by quality control machinery (11, 12, 13) that prevents misfolded or premature proteins from leaving the ER.

Folding of proteins follows definite pathways that have been studied for very few proteins, among which is hCGß (14, 15, 16, 17). The first folding steps of a protein take place in nsec or µsec time ranges (18). The generation of tertiary and quaternary structures in more complex multidomain proteins, in particular those composed of subunits, takes much more time, developing from a few minutes up to hours. This complex process is linked to the formation of disulfide bridges, the attachment and processing of carbohydrate parts, and subunit assembly. The folding of hCGß is characterized by the presence of definite folding intermediates and seems to require the assistance of chaperones, whereas GPH{alpha} seems to fold very rapidly. Finally, the biological properties of hCG require the presence of both subunits, which makes the correct association of the subunits an important step in hCG biogenesis. During subunit association a carboxy-terminal portion of hCGß wraps around the GPH{alpha} like a seat belt (2, 3). It has been shown that during subunit association, the glycosylated loop L2 of the GPH{alpha} has to thread through the cavity formed by the seat-belt structure of hCGß (19). Subunit association seems to be facilitated by the formation, disruption, and reformation of a "tensing" loop within the seat-belt structure (10), which serves to stabilize the heterodimer. The association of GPH{alpha} and hCGß was intensively investigated in vitro and in vivo, and two models have been proposed. The first model is based upon the presence of a prefolded hCGß having four of six disulfide bridges already established as a mandatory requirement for the association (20, 21, 22). The second model assumes that the association occurs much earlier and does not depend directly on the presence of preformed disulfide bridges (17). The number and quality of structural and functional details already known in the case of the hCG subunits concerning conformation, disulfide bridge arrangements, folding pathway, and subunit association render hCG a very good model for the study of the biogenesis of complex multidomain glycoproteins.

Marked differences exist in the dynamics of GPH{alpha} and hCGß folding and maturation. Association occurs between an almost completely folded GPH{alpha} and an immature hCGß. We show here that while being associated with GPH{alpha}, hCGß matures via a broad spectrum of folding intermediates. The presence of GPH{alpha} in the hCG {alpha}ß-heterodimers seems to result in a much slower hCGß folding, compared with free hCGß. This seems to postpone the approval of hCG {alpha}ß-heterodimer complexes by the ER quality control machinery. In contrast, the formation of GPH{alpha}{alpha} homodimers, by the association of free GPH{alpha}, does not cause any delay in export from the ER and movement along the secretory pathway. Likewise, free GPH{alpha} and free nonassociated hCGß seem to pass the quality control of the ER and the processing in the Golgi in a fast-track manner, thus becoming sialylated much earlier than the hCG {alpha}ß-heterodimers. The latter reached the sites of sialylation just before the secretion of newly synthesized compounds began. To our knowledge this is the first report of a direct, probably unidirectional negative effect, of one subunit on the folding of another in an oligomeric protein complex. This seems to indicate that the process of attaining a proper quaternary structure of the hCG {alpha}ß-heterodimer seems to interfere with protein folding and the maturation of the individual subunit members in the complex.


   Materials and Methods
Top
 Abstract
 Introduction
 Materials and Methods
Results
 Discussion
 References
 
All experiments were reproduced several times.

Cell culture, radioactive labeling, and identification procedures
Cultivation of JEG-3 choriocarcinoma cells (ATCC, Rockville, MD), radioactive labeling with [35S]Met/Cys in pulse-chase experiments, preparation of lysates, immunopurification with hCG subunit-specific antibodies, SDS-PAGE analysis, Western blotting, the isolation of individual bands in Whole Gel Eluter (Bio-Rad Laboratories, Munich, Germany), two-dimensional (2D)-electrophoretic analysis, and the analysis of subunit microheterogeneity by enzymatic deglycosylation were performed as described in the supplemental data, which are published on The Endocrine Society’s Journals Online web site at http://endo.endojournals.org.

Semipermeabilization of JEG-3 cells, blockade, and labeling of free thiol groups
Free thiol groups present on subunit intermediates during the maturation process were radioactively labeled and blocked at cell lysis by reaction with [3H]N-ethylmaleinimide (1.48–2.2 TBq/mmol; PerkinElmer Inc., Boston, MA). This procedure required the prior semipermeabilization of the cells to remove the bulk of accessible thiol-containing cytoplasmic metabolites and proteins. The conditions had to warrant that only the plasma membrane was disrupted, leaving the other cellular membranes intact. For that reason, following the described procedures, we investigated the use of saponin, digitonin, streptolysin-O, and {alpha}-toxin (23), and in addition, the hypotonic shock method was applied (24). The efficacy of the permeabilization was determined on the basis of the release of the cytoplasmic enzyme lactate dehydrogenase (EC 1.1.1.27), the maintenance of the integrity of the labeled secretory proteins, and their resistance against digestion with exogenous trypsin added after semipermeabilization. Finally, the protocol that follows gave the best results. After the application of the labeling with [35S]Met/Cys (supplemental data), cells were placed on ice. Medium was discarded, and cells were rinsed twice with ice-cold permeabilization buffer consisting of 10 mM HEPES (pH 7.3) containing 125 mM potassium chloride, 19 mM sodium chloride, and 1 mM EGTA. A solution of 160 µg digitonin/ml permeabilization buffer (prepared from a stock solution of 100 mg digitonin/ml dimethylsulfoxide) was applied to the cell cultures for 30 sec, after which, the cells were washed three times with permeabilization buffer. The entire permeabilization procedure, including the washing steps, was performed with ice-cold solutions and completed within 90 sec. To achieve this short handling time, the solutions required were pipetted into small glass beakers, placed on ice, and poured on the cell cultures with caution. After gentle shaking on ice, the solutions were removed by carefully inverting the cultures.

After semipermeabilization, free thiol groups were labeled by applying [3H]N-ethylmaleinimide (NEM) (15 µCi/100 µl total volume), in pentane, directly to the cell cultures with modified lysis buffer (supplemental data) and shaking gently. The cultures were incubated for 5 min on ice, after which, a large excess of nonradioactive New England Nuclear Corp. Life Science Products (Chestertown, MD) (30 µl of a 260 mM NEM solution) was added.


   Results
Top
 Abstract
 Introduction
 Materials and Methods
 Results
Discussion
 References
 
Differences in the maturation of free subunits and hCG {alpha}ß-heterodimers
The most prominent intracellular hCGß subunit variants visible in SDS-PAGE consisted of the early hCGß variants with apparent molecular masses (Mr app) of 28, 30 (double band), and a late 35-kDa variant. In addition, a GPH{alpha} (nonreduced Mr app = 22 kDa, reduced Mr app = 21 kDa) was coprecipitated with hCGß-specific monoclonal and polyclonal antibodies (Fig. 1Go). This GPH{alpha} originated from hCG {alpha}ß-heterodimers, immunoprecipitated by hCGß-specific antibodies. During the subsequent desorption of the immune complexes from protein-A-agarose, the hCG {alpha}ß-heterodimers were completely dissociated into the subunits. The 30-kDa hCGß variant appeared earlier than the 28-kDa isoform (compare the band intensities at 2 vs. the 40-min chase). Both early isoforms collapsed into a single band with a Mr app of 25 kDa when the sample was reduced before electrophoresis, whereas the late 35-kDa variant was resistant (Fig. 1AGo, lower panel). This indicates that the double band of hCGß variants resulted from the status of disulfide bridge formation and did not represent carbohydrate variants. As shown below, the 30-kDa hCGß variant seems to be more immature because it contains significantly more free nonlinked thiol groups than the 28-kDa variant. The 35-kDa hCGß appeared 10 min before the onset of the secretion. It was sensitive to neuraminidase digestion, whereas the 28 and 30-kDa variants were not digested by the enzyme (Fig. 1BGo). The 35-kDa hCGß band most likely represents the mature secretion-competent free hCGß and contains small quantities of the hCGß part of mature hCG {alpha}ß-heterodimers that were already formed under the conditions applied (see Fig. 2Go and corresponding text).


Figure 1
View larger version (46K):
[in this window]
[in a new window]

  		FIG. 1. The most prominent hCG variants in JEG-3 cells and evidence for the sialylation of the 35-kDa hCGß variants. A, The cells were labeled for 30-min pulse with [35S]Met/Cys and chased 10 min. Immunopurification with a mAb directed against the adjacent tips of loops 1 and 3 of hCGß (code INN-hCG-22), epitope ß2 (50 ). Separation in the SDS-PAGE was performed under nonreducing (upper part) and reducing conditions (lower part). The 40-min chase condition represents the time point of approximately 10 min before the onset of the secretion. B, JEG-3 cells were radioactively labeled (45-min pulse [35S]Met/Cys, 30-min chase); the proteins were immunopurified from the lysate with purified polyclonal anti-hCGß rabbit immunoglobulin and digested with neuraminidase as described in the supplemental data. It is shown that only the mature 35-kDa free hCGß is sialylated. Visualization of the radioactive signal was detected by fluorography.

 

Figure 2
View larger version (43K):
[in this window]
[in a new window]

  		FIG. 2. Microheterogeneity of free hCG subunits and hCG {alpha}ß-heterodimers. Labeling of JEG-3 cells with [35S]Met/Cys and chase was performed as indicated. In the first dimension, the protein samples were separated by isoelectric focusing in the pH range between three and 10 under nonreducing conditions, followed by a SDS-PAGE into the second dimension (linear gradient gels 8–16%). Samples depicted in F, G, and H were analyzed under reducing conditions. The pH scale on top of the figure depicted was obtained by marker proteins; the kDa values marked at E–H refer to standard molecular mass markers (for details, see the supplemental data). The arrowheads indicate the position of the electrodes in the first dimension. Spots seen outside the space between the arrows should be ignored because they represent nonresolved starting material. Visualization of the radioactively labeled proteins was detected by fluorography. A, The subunit variants were precipitated with a mAb recognizing all GPH{alpha} variants [code INN-FSH-179, epitope {alpha}3* (50 )], as shown in one (panel A, small figure, right-hand side) and 2D electrophoresis (left). B, Microheterogeneity of the subunits is completely lost in response to desialylation (same mAb used as in A; nonreducing conditions). C and D, Under short-time labeling conditions, the population of microheterogeneous subunit variants is almost completely missing. E, Intracellular free hCGß show the mature native sialylation pattern in contrast to the subunits of hCG {alpha}ß-heterodimers (immunoprecipitation with affinity purified polyclonal anti-hCGß antibody to isolate all variants of free hCGß and hCG {alpha}ß-heterodimers). F, Highly sialylated microheterogeneous hCG isoforms were secreted into the cell culture medium (immunoprecipitation with the same antibody as in E; for details see the supplemental data). G, Enzymatic release of the N-linked glycan residues from the material shown in F by treatment with PNGase F. H, Treatment of secreted hCG and free hCGß after isolation with the immunoaffinity purified polyclonal anti-hCGß antibody with PNGase F, followed by release of sialic acid by digestion with neuraminidase. Microheterogeneity of hCGß is completely abolished by neuraminidase treatment.

 
Microheterogeneous isoforms
Analysis of the dynamics of the microheterogeneity in 2D electrophoreses offered insights into the subunit maturation and transport in the secretory pathway. For immunoprecipitation, a mAb recognizing all GPH{alpha} variants was used (code INN-hFSH-179; directed against epitope {alpha}3*, presumably located on peptide loop L2). It precipitated large amounts of free GPH{alpha} (Mr app = 22 kDa; also containing GPH{alpha} dissociated from hCG {alpha}ß-heterodimers), small quantities of monoglycosylated GPH{alpha} (Mr app = 18 kDa), and early GPH{alpha}{alpha} homodimers (Mr app = 35 kDa). Furthermore, the hCGß portion of mature hCG {alpha}ß-heterodimers was coprecipitated (Fig. 2AGo). The different isoelectric points of intracellular subunit isoforms originate from the degree of sialylation. This became evident when the sialic acid residues were removed by neuraminidase treatment. In response to that digestion, the microheterogeneous pattern collapsed into hCGß and GPH{alpha} variants, with isoelectric points above 8.5 (Fig. 2BGo). This also shows that at least a proportion of the subunits has already left the ER at cell lysis (in that case, after 30-min pulse, 10-min chase) and has been sialylated in the Golgi apparatus.

When the pulse time was lowered to 5 min (no chase), the number of microheterogeneous isoforms was decreased considerably (Fig. 2CGo). In that case only a fraction of GPH{alpha} (most probably free GPH{alpha}) seems to have moved to the Golgi and become sialylated. The {alpha}ß-heterodimers were not sialylated under these conditions, which was evident from isoforms of the subunits, after the dissociation of the heterodimer complexes (Fig. 2DGo). From the cell lysate, an anti-pan hCGß mAb precipitated free hCGß in addition to hCG {alpha}ß-heterodimers. The latter were not sialylated, which is in contrast to free subunits (Fig. 2EGo). However, the final processing of the {alpha}ß-heterodimers and their secretion into the culture medium was associated with strong sialylation, and this generated a characteristic pattern of acidic subunit isoforms (Fig. 2FGo) and was coupled with immediate secretion. In response to the release of the entire N-linked glycan residues of secreted hCG, by digestion with PNGase F, both subunits showed a shift toward lower molecular masses and more basic variants (Fig. 2GGo). The microheterogeneity of the GPH{alpha} part was completely lost by this procedure, whereas the hCGß subunit still showed a group of nine acidic variants due to the presence of sialic acid residues on the O-linked glycans. When these sialic acid residues were removed with neuraminidase, the microheterogeneity of hCGß was completely abolished (Fig. 2HGo). These observations indicate a sequestration of free subunits as well as GPH{alpha}{alpha} homodimers from the hCG {alpha}ß-heterodimers in the secretory pathway, leading to the rapid transport of free subunits and GPH{alpha}{alpha} homodimers. In contrast, the hCG {alpha}ß-heterodimers were restrained from sialylation for a much longer period.

Dynamics of hCGß maturation
In all pulse-chase experiments performed here, the maturation of hCGß seemed to be slower and more differentiated than in the case of GPH{alpha}. This led us to study the maturation process of the ß-subunit in more detail by performing pulse-chase experiments at 25 C in comparison to 37 C. The rationale behind this was to decrease the velocity of the maturation process by lowering the temperature and thereby increasing the chance of monitoring the dynamic development of hCGß variants in greater detail. In addition, the pulse time was lowered to 5 min to facilitate the observation of early variants in the maturation process, before the translocation of the bulk of the proteins to the Golgi apparatus had occurred, as shown in Fig. 2Go, C and D. Furthermore, the resultant samples were submitted to micro-preparative isolation of single bands by a Whole Gel Eluter procedure followed by concentration using ultrafiltration (Fig. 3BGo). The isolation of the bands resulted in a more detailed picture; compared with the original SDS-PAGE, a higher concentration was present. Multiple variants of hCGß appeared during the maturation process showing Mr app between 28 and 107 kDa (Fig. 3BGo). In general, the subunit variants present in the samples from 25 and 37 C cultures were very similar; however, time point and duration of their occurrence differed (Fig. 3Go, B and C). At 37 C the 107-kDa hCGß-variant was only present during the 5 min of the pulse (0-min chase; Fig. 3BGo, 37 C, lane 1), suggesting that it occurs very early and transiently. When labeling was performed at 25 C, the 107-kDa band was observed with a delay of 5 min (Fig. 3BGo, 25 C, lane 5), indicating that it is an unlikely candidate of the original precursor for the other hCGß variants, but an intermediate that appears to develop from molecules of lower molecular masses. At 25 C (Fig. 3CGo, 60-min chase), a 70-kDa hCGß variant was observed that was not present at 37 C. It was most probably formed due to the transport blockade at this temperature (see Discussion). The free 35-kDa hCGß was present at 60-min chase under both experimental conditions. As already mentioned previously, the 30-kDa part of the double band was observed at 0-min chase, and the 28-kDa band followed later. GPH{alpha} was coprecipitated under all conditions from the beginning, indicating that association of the subunits has been completed during the 5-min pulse time. Furthermore, it shows that the hCGß intermediates developed while being associated as hCG {alpha}ß-heterodimers.


Figure 3
View larger version (38K):
[in this window]
[in a new window]

  		FIG. 3. Dynamic development of hCG subunit variants during maturation. JEG-3 cells were [35S]Met/Cys labeled (5-min pulse) at 25 or 37 C and chased for the given time. Thereafter, cells were semipermeabilized at 0 C and treated with [3H]N-ethylmaleinimide (Fig. 5Go) to block and label the free thiol groups at cell lysis. Radioactive signals were detected by fluorography. A, Proteins were purified by immunoprecipitation with an immunoaffinity purified polyclonal anti-hCGß rabbit antibody and separated by SDS-PAGE applying nonreducing conditions. B, Micro-preparative isolation of bands by the Whole Gel Eluter (SDS-PAGE, 1 mm polyacrylamide gel, sample slot of 50-mm width). Individual fractions obtained were concentrated by ultrafiltration and reapplied to an SDS-PAGE (nonreducing conditions). C, Schematic presentation of time-dependent appearance of subunit variants in the course of the maturation process.

 
Disulfide bridge formation and subunit maturation
The process of disulfide bridge formation in the subunit folding intermediates was studied by the blocking and labeling of free thiol groups that had not yet reacted to disulfide bridges at different time points by reaction with [3H]NEM. For that purpose, we optimized a short-time semipermeabilization method that allowed us to wash out the cytoplasm yet maintain the integrity of the compartments of the secretory pathway.

The amount as well as the integrity of 35S-labeled GPH{alpha} and hCGß variants being protected in the compartments of the secretory pathway remained unchanged by the semipermeabilization procedure (Fig. 4AGo). The integrity of the membranes along the secretory pathway was proven using experiments in which the aforementioned semipermeabilized cells were incubated at 37 C for 1 h before the entire cell lysis. The band pattern as well as the intensity of the radiolabeled bands appeared to be completely maintained (Fig. 4CGo), indicating the integrity of the compartmental membranes of the secretory pathway. Furthermore, the 35S-labeled secretory proteins in the semipermeabilized cells were protected from trypsin digestion, demonstrating again the integrity of the inner cellular membranes (data not shown). We performed semipermeabilization experiments with cells that were incubated during the pulse-chase interval at 25 and 37 C, respectively (Fig. 4Go, A and B), to decrease the maturation velocity of the subunits. At both temperatures a nearly identical band pattern was observed. However, in the case of the 25 C samples, a clear delay in the development of the band pattern was visible (Fig. 4BGo). Additional experiments showed that in fact the 5-min band pattern of the 37 C condition was observed at 20-min chase when the labeling was performed at 25 C (data not shown).


Figure 4
View larger version (48K):
[in this window]
[in a new window]

  		FIG. 4. Effect of semipermeabilization and blockade of the free thiol groups with [3H]NEM. A, JEG-3 cells [35S]Met/Cys labeled at 25 and 37 C (A: 5-min pulse, 5-min chased; B: 5-min pulse, chase as indicated). Semipermeabilization was performed at 0 C by incubation of the cells with 160 µg digitonin/ml for 30 sec. Cytoplasm was washed out within 60 sec, blockade of free thiol groups with [3H]NEM (5 min at 0 C) and cell lysis as described in Materials and Methods. Proteins isolated with the affinity purified polyclonal anti-hCGß rabbit antibody. SDS-PAGE at nonreducing conditions. A, Semipermeabilization changed neither band pattern nor concentration of the hCG subunit variants. B, Incubation of cells at 25 C delayed the maturation of the subunits. C, Integrity of GPH{alpha} and hCGß subunits is maintained even when the semipermeabilized cells were incubated at 37 C for 1 h (instead of keeping on ice for 5 min, as performed in the regular protocol). [35S]Met/Cys labeling 5 min at 37 C (no chase), semipermeabilization, [3H]NEM treatment, and cell lysis (control) or incubation as indicated before cell lysis. Immunoprecipitation with mAb INN-hCG-22 (epitope ß2), SDS-PAGE under reducing conditions. Visualization was detected by fluorography.

 
The isolation of the subunit intermediates by the Whole Gel Eluter as shown in Figs. 3BGo and 5AGo, enabled us to determine the individual 3H/35S ratios of these isolated fractions. The method provides a rough estimate of the progress of disulfide bridge formation. A high ratio was expected when large numbers of free thiol groups were present in a subunit variant, and a low ratio should indicate that a great number of disulfide bonds had already been formed. Moreover, the maturation process of the subunits should cause a decrease in the 3H/35S ratio during the pulse-chase experiment. The disulfide bridges of the coprecipitated GPH{alpha} seemed to be formed much more rapidly than in the case of hCGß, as indicated by the low 3H/35S ratios in the case of the GPH{alpha} variants (Fig. 5AGo). The 3H/35S ratios of coprecipitated GPH{alpha} variants were initially high (25 C, 0 min, 5-min chase; 37 C, 0-min chase) and decreased rapidly later to a level indicating completeness of disulfide bridge formation.


Figure 5
View larger version (61K):
[in this window]
[in a new window]

  		FIG. 5. Progress of disulfide bridge formation in the individual subunit variants. A, Subunit variants were double labeled by the incorporation of [35S]Met/Cys into the polypeptide chains and by reaction of the free thiol groups with [3H]NEM. In the Whole Gel Eluter fractions (Fig. 3BGo), the 3H/35S ratios were determined by radioactivity counting to estimate roughly the amount of free thiol groups in the individual fractions. High 3H/35S ratios indicate a high number of free reactive thiol groups present at a given time point; low ratios suggest advanced disulfide bridge formation. B, Association of [35S]Met/Cys-labeled proteins of JEG-3 cells (30-min pulse, 15-min chase) with CNX and GRP78 (BiP), detected by immunoprecipitation with anti-CNX and anti-BiP antibodies. C, The immature hCGß variants collapsed into a single 25-kDa band when separated in the SDS-PAGE under reducing conditions, except for the 35-kDa mature secretion-competent hCGß. D, The [3H]NEM reacted almost exclusively with immature hCGß variants, and not in the GPH{alpha}. SDS-PAGE gels of experiments shown in this figure and Fig. 3Go were stored until the 35S-label had decayed below the detection limit.3H-labeled variants were detected in the following exposure to the x-ray film for 132 d in the presence of enhancer. The fact that the strongly 35S-labeled GPH{alpha} is missing in D provides an internal control that the 35S-label has decayed to an undetectable limit. This indicates that the signal of the hCGß in D represents only the 3H-label. Radioactive signals were visualized by fluorography.

 
In comparison to GPH{alpha}, the hCGß variants seem to have more free thiol groups for a longer time, at both temperature conditions. As expected, the number of free thiol groups of hCGß decreased with time. The 3H/35S ratios did not correlate with the apparent molecular masses of the hCGß variants. The early 107-kDa hCGß variant showed a lower 3H/35S ratio than hCGß variants that were formed later. The variants with higher molecular masses, in particular the 107-kDa variant, were most likely complexes with chaperones. Antisera directed against the ER-resident lectin chaperone calnexin (CNX) and the GRP78 chaperone (BiP) precipitated proteins from the JEG-3 lysates (Fig. 5BGo) that could be identified in Western blot experiments as the 22 and 18-kDa variants of GPH{alpha} and variants of hCGß (data not shown). Similar results were obtained with anticalreticulin antisera. The optimum reaction of the hCG subunits with CNX was found to be at 15 min (data not shown). This supports the hypothesis that chaperones form transient complexes with the hCG subunits.

All these different hCGß variants collapsed into a single 25-kDa band when reduced before electrophoresis, except the secretion-competent 35-kDa variant of hCGß (Fig. 5CGo). This indicates that the hCGß folding intermediates represent molecules that differ in the established disulfide bridges at the various time points investigated. This was supported by the fact that the [3H]NEM reacted almost exclusively with hCGß and only to a very small extent with GPH{alpha}, as demonstrated by the radioactive signal of the tritium label, monitored after the 35S-label had decayed under the detection limit (Fig. 5DGo).

Formation of {alpha}ß-heterodimers
Formation of {alpha}ß-heterodimers occurred very early, most likely immediately after the completion of transcription. This became evident when the pulse time was lowered to 30 sec, followed immediately by cell lysis (data not shown). To exclude the possibility that the radioactive labeling was in part caused by the formation of mixed disulfides between thiol groups of the protein and the cysteine contained in the [35S]Met/Cys mixture, we incubated immunoprecipitated proteins obtained from nonlabeled JEG-3 cells (lysis in absence of NEM) with the radioactive label for several hours under a nitrogen atmosphere. No labeling potentially caused by mixed disulfide bridge formation was observed (data not shown).

The elution of the immune complexes from the protein A agarose during the isolation procedure of the radiolabeled proteins caused the complete dissociation of the {alpha}ß-heterodimers. To ensure that the GPH{alpha} coprecipitated by the anti-hCGß antibodies originated from {alpha}ß-heterodimers, a cell lysate of a pulse-chase experiment was applied to the Sephadex G-150 column, for separation of free subunits and {alpha}ß-heterodimers before immunoprecipitation (the results can be found in supplemental Fig. 1Go). The hCG {alpha}ß-heterodimers and free GPH{alpha} were clearly separated and identified. In addition, attempts were made to trap the expected immature {alpha}ß-heterodimers present at cell lysis before immunoprecipitation, by reversible cross-linking with dithiobis succinimidyl propionate (Lomant’s reagent). After micro-preparative isolation of the individual bands of the cross-linked nonreduced samples, it was shown that a cross-linked 39-kDa band (nonreduced) corresponded to the early, immature {alpha}ß-heterodimers (the results can be found in supplemental Fig. 2Go).


   Discussion
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
References
 
For the studies presented here, the investigation of folding, association, and transport in the secretory pathway with the natural nonmutated subunits was preferred. It cannot be excluded that even the introduction of point mutations could result in a different behavior in this context, e.g. as shown by the more stable interactions of mutated proteins with the chaperone BiP (25).

A number of differences between the subunits in maturation and association was obvious. In contrast to GPH{alpha}, hCGß as part of hCG {alpha}ß-heterodimers folds and matures much slower, thereby passing through a series of distinct reduction-sensitive folding intermediates. Different folding velocities of the two subunits have also been described in JAR cells (26). Whereas a monoglycosylated GPH{alpha} was observed, hCGß seemed to be N-glycosylated cotranslationally at both N-glycosylation consensus sites, by default. GPH{alpha} showed the formation of GPH{alpha}{alpha} homodimers. Homodimerization of hCGß was not found in JEG-3 cells. In purified preparations of recombinant hCGß as well as in standard preparations, hCGßß homodimers were found (27). However, it is not clear up until now whether hCGßß homodimer formation takes place in nontransformed cells, or if the hCGßß homodimers are artificially formed during isolation and purification of hCGß preparations. On the other hand, it cannot be excluded that the lack of the hCGßß homodimers in JEG-3 cells is due to the low concentration of the free monomeric subunit in these cells.

hCGß maturation
Folding of hCGß is tightly connected to the disulfide bridge formation, which was already shown (21). Both subunits of hCG belong to the family of cysteine knot proteins and growth factors (7). The cysteines, as well as the noncysteine residues, of the cystine knot consensus sequence seem to play a crucial role for folding and efficient subunit assembly (28). The experiments described here extend the spectrum of known hCGß disulfide-bridge dependant folding intermediates beyond what was known so far. The highly dynamic maturation of hCGß developed, while being associated with the GPH{alpha} in hCG {alpha}ß-heterodimers. Apparent molecular masses ranged from 28 kDa up to 107 kDa (Fig. 3Go). An overall estimation of the progress of the disulfide bridge formation was achieved by the reaction with [3H]NEM of free thiol groups, of cysteine residues, that had not yet formed disulfide bridges. A similar technique was used previously to demonstrate reversible disulfide bridge formation between unassembled immunoglobulin light chains and ER matrix proteins (29). The low 3H/35S ratio of GPH{alpha} assembled in hCG {alpha}ß-heterodimers (Fig. 5Go) shows the advanced maturation of GPH{alpha}. In contrast, the hCGß variants reacted for a much longer period and to a wider extent with the [3H]NEM, indicating a much higher number of free cysteine residues. This was confirmed by the analysis of the binding of [3H]NEM after the decay of [35S]. It transpired that the assembled GPH{alpha} portion contained almost no nonlinked disulfide bridges. The [3H]NEM label was predominantly incorporated into the 30-kDa variant of hCGß, which seems to be the more immature form in comparison to the 28 kDa (Fig. 5DGo).

At least some of the variants with higher molecular mass are supposed to be transient complexes with chaperones like CNX and GRP78 (BiP) (Fig. 5BGo). A complex formation of the hCGß with BiP, and possibly also with ERp72 and ERp94, was previously suggested after Western blot experiments with transfected Chinese hamster ovary cells (30). Binding of BiP seems to occur preferentially when the proteins fold slowly or unstably (31). This also appears to apply to hCGß. The relatively low 3H/35S ratio of the 107-kDa variant (Fig. 5Go), in comparison to preceding and following subunit variants, may be explained by the inaccessibility of cysteine residues, of the subunit to [3H]NEM, due to sterical hindrance and/or transient formation of mixed disulfides with ERp57 or other thiol-oxidoreductases. A very similar spectrum of hCGß intermediate variants was also observed in Western blots after separation in Whole Gel Eluter fractions (data not shown). However, in Western blots of the cell lysate, the entire spectrum of intracellular hCGß early, intermediate, and late variants can be visualized, preventing the observation of the dynamics and the concentration changes during biogenesis. It is remarkable that all the different hCGß intermediates collapsed into a single band of Mr app equal to 25 kDa, except the 35-kDa variant, which was first noticed at approximately 40-min chase (32) and seems to be the fully mature secretion-competent hCGß. The transport velocities through the secretory pathway of free GPH{alpha} and the GPH{alpha}{alpha} homodimers seem to be different because a nonsialylated 35-kDa GPH{alpha}{alpha} homodimer variant was detectable, but not a nonsialylated precursor of the free large GPH{alpha}. This is probably due to its rapid transport and sialylation.

Besides 37 C, incubation at 25 C was also applied to obtain a more detailed picture of the maturation process. A similar strategy was used in the in vitro measurement of the translation rates of the hemoglobin {alpha}- and ß-subunits (33). Using spectrin domains as a model for cooperative folding, it was shown that the cooperativity and the stabilizing effects on the protein were much greater at 25 C than at higher temperatures (34). In the case of hCG, the same folding intermediates were observed at both temperatures except the 70-kDa hCGß variant. At 25 C, a blockade of the terminal steps of the secretory pathway was observed when the condition was applied for a longer period without indication of degradation of the accumulated proteins (32).

Subunit association
The hCG subunits associated very early. Even when the pulse time was lowered to 1 min or below, hCG {alpha}ß-heterodimers were detected (32). Given a translation velocity in the order of 100 amino acids per minute (35), hCG ß-subunit (145 amino acid residues) should be completed in roughly 60–90 sec. The presence of {alpha}ß-heterodimers at that time point seems to indicate that the subunits associate very early, shortly after the completion of translation and first folding steps. GPH{alpha} was coprecipitated by hCG-specific antibodies together with all the various very immature hCGß intermediates (Fig. 3Go). Completely unfolded subunits would not have been detected because the immunological epitopes detected by monoclonal and polyclonal antibodies used are strictly conformation dependent. However, hCGß at subunit association seems to be very immature, in contrast to the almost completely folded GPH{alpha}. In JAR cells the folding of GPH{alpha} showed a t1/2 of approximately 90 sec, whereas in the case of the hCGß, a t1/2 of 5 min was found (28). As shown here, the folding of the hCGß seems to take even longer. It was previously reported that the hCGß is more assembly competent in the partially unfolded state than in the completely folded state (22). However, in that report it was assumed that four of six disulfide bridges of hCGß had already formed before subunit association, one together with subunit association and the last thereafter. This model is in clear contrast to our present and previous findings (17, 32).

Free subunits and GPH{alpha}{alpha} homodimers are sequestered in the secretory pathway from hCG {alpha}ß-heterodimers
Maturation and processing of secretory glycoproteins are controlled mainly by N-glycosylation in the ER, glycosylation reactions in the Golgi apparatus, and disulfide bridge formation. As shown here, the free GPH{alpha}, nonassociated hCGß and GPH{alpha}{alpha} homodimers are sequestered from the hCG {alpha}ß-heterodimers and move along the secretory pathway much faster. This was concluded from the microheterogeneous pattern, glycosylation analysis, apparent molecular mass data of the various molecules, and results of the hCG {alpha}ß-heterodimer maturation studies. Whereas free GPH{alpha} variants and free hCGß (Fig. 1BGo) were almost completely sialylated 10 min before the beginning of secretion (in JEG-3 cells ~40 min after translation), the bulk of the hCG {alpha}ß-heterodimers were not sialylated at all at that time. To accomplish sialylation the glycoproteins have to reach parts of the Golgi with sialylation capacity. Activated sialic acid and sialyltransferases are present from cis to trans Golgi cisternae in an increasing gradient (36). The highest concentration of sialyltransferases was found in the trans Golgi and the trans-Golgi network (37, 38, 39). At 30 min after the start of the pulse, the majority of the subunits of hCG {alpha}ß-heterodimers were still endoglycosidase H sensitive (data not shown). Resistance against endoglycosidase H is acquired when the glycoproteins have reached the medial Golgi, where the mannosidase II is located. The conclusion of these results is that the free subunits (free large GPH{alpha}, free nonassociated hCGß) and GPH{alpha}{alpha} homodimers had already reached the trans Golgi and the trans-Golgi network, whereas the hCG {alpha}ß-heterodimers have not yet arrived in the medial Golgi (Fig. 6Go).


Figure 6
View larger version (46K):
[in this window]
[in a new window]

  		FIG. 6. Biosynthesis and maturation of hCG subunits in the secretory pathway and sequestration of free subunits and hCG {alpha}ß-heterodimers following the cisternal maturation model. A snapshot approximately 10 min before starting the secretion of the hCG {alpha}ß-heterodimer is schematically depicted. GPH{alpha} (blue) and hCGß (green) are synthesized into the lumen of the ER. The pathway only shown for GPH{alpha} also refers to hCGß. A large fraction of the subunits associates immediately after completion of translation, forming hCG {alpha}ß heterodimers ("6"). Excessive, unassociated GPH{alpha} variants combine to form immature 35-kDa GPH{alpha}{alpha} homodimers ("7") or proceed as GPH{alpha}, as does free hCGß. Folding and maturation of the free subunits and of GPH{alpha}{alpha} homodimers seem to proceed much faster than in the case of hCG {alpha}ß-heterodimers. This sequesters the free subunits and the GPH{alpha}{alpha} homodimers from the hCG {alpha}ß-heterodimers, most likely at the level of the ER. Glycoproteins leave the ER possibly aided by the ER-Golgi intermediate compartment (ERGIC) cargo transporter ERGIC-53 ("9"). GalNAc residues are attached to the serine residues of the carboxy-terminal part of hCGß (structure "e"). In the medial Golgi ("12"), GlcNAc-residues are added to N- and O-linked glycans (structures "f–h"), and further branches of the N-glycans are introduced in part (for further details and an overview of additional glycan variants, see Ref. 4 ). From free nonassociated GPH{alpha}, the large free 24-kDa GPH{alpha} variants are formed. When the free subunits and GPH{alpha}{alpha} homodimers have already reached Golgi cisternae with highest sialylation activity ["13" and trans-Golgi network (TGN) "14"], the majority of the hCG {alpha}ß-heterodimers have not arrived in the medial Golgi, and the hCGß portion is not yet completely folded, as indicated by its sensitivity to reduction and accessibility for NEM. The subunit variants indicated without a Mr app represent intermediates that were almost unobservable, probably because their intracellular concentration or their half-lives were too low. PM, Plasma membrane.

 
Secretory glycoproteins have to pass a tight quality control system before they are allowed to leave the ER. This machinery comprises components that interact with the glycan residues, and those that are involved in folding and disulfide bridge formation (Fig. 6Go, "2" and "8"). In most cases the attachment of the N-linked glycan residues occur cotranslationally by attachment of a dolichol pyrophosphate-anchored core oligosaccharide precursor, Glc3Man9(GlcNAc)2, (Fig. 6Go, compound a) to the Asn-X-Ser/Thr sequon, catalyzed by the oligosaccharyltransferase. In the case of GPH{alpha}, to a minor fraction, only one N-glycan is transferred cotranslationally, contributing to the association-incompetent 18-kDa GPH{alpha} variant (Fig. 6Go, "3"). The missing N-glycan is supposed to be attached posttranslationally (Fig. 6Go, "4"). In the case of hCGß, both N-glycans are linked cotranslationally (Fig. 6Go, "5"). The release of two terminal glucose residues prepares the glycoprotein for entry into the CNX/calreticulin (CRT) cycle (for review, see Refs. 40, 41, 42). PDIs, in particular ERp57, a glycoprotein-specific PDI homolog, establish mixed disulfides with the glycoprotein (43, 44). It is recruited to glycoprotein/CNX/CRT complexes, and acts as a chaperone that helps to establish the proper disulfide bridges and folding of glycoproteins. A PDI seems to support effectively folding of hCGß, even in vitro (22). When the folding has not yet completed, reglucosylation of the glycan residue is performed by the UDP-Glc:glycoprotein glucosyltransferase. The glycoprotein is again processed in the CNX/CRT/ERp57 cycle (42) until either a native transport-competent form is achieved that is allowed to leave the ER, or it is submitted to the ERAD degradation system (36, 45).

In JEG-3 cells and other choriocarcinoma cells, a large excess of GPH{alpha} variants are synthesized in relation to hCG and hCGß (46). Free GPH{alpha} (Mr app = 22 kDa) has the possibility to associate with hCGß to yield hCG {alpha}ß heterodimers. Alternatively, it may dimerize to GPH{alpha}{alpha} homodimers (47) or move to the Golgi, where the free, nonassociated GPH{alpha} is converted into large free GPH{alpha} (Mr app = 24 kDa) by additional glycosylation. Based on the results shown here and elsewhere (32, 47), it may be roughly estimated that in JEG-3 cells, the amount of the monomeric GPH{alpha} variants is 6.5- to 7.5-fold higher than of the GPH{alpha}{alpha} homodimers. The GPH{alpha} released from hCG {alpha}ß-heterodimers by dissociation of the complex accounts for one fifth of the total amount of free, nonassociated GPH{alpha} variants.

Free nonassociated hCG subunits and GPH{alpha}{alpha}-homodimers folded and matured much faster than the subunits in hCG {alpha}ß-heterodimers, and, thus, were allowed to leave the ER before the hCG {alpha}ß-heterodimers. This suggests that the mutual intersubunit interactions required to accomplish a stable quaternary structure of the hCG {alpha}ß-heterodimers slows down the maturation of the hCGß portion. The distinct sequence of hCGß folding intermediates are probably the reason why the hCG {alpha}ß-heterodimers are prevented from leaving the ER by the repeated interactions of the CNX/CRT/ERp57 complexes or other ER resident PDIs.

However, this does not explain why free hCGß, unlike the hCGß assembled in the hCG {alpha}ß-heterodimers, seems to be allowed to leave the ER much earlier. Several possibilities may account for this sequestration. Free hCGß may escape the quality control of the ER without having completed most of the disulfide bridges. Intracellular 35-kDa free hCGß and hCGß assembled in secreted hCG were not sensitive to reducing conditions, which argues against that possibility. Furthermore, we have tested the stability of the intracellular variants against reducing agents in vivo by incubation of the cells with dithiothreitol (0.1 mM up to 20 mM). Intracellular free 35-kDa hCGß remained resistant to reduction up to 20 mM, whereas immature hCGß (double-band 28, 30 kDa) was reduced in vivo at 2 mM dithiothreitol (data not shown). This also argues against an escape of the free hCGß from the ER-based quality control system. Alternatively, free subunits and GPH{alpha}{alpha} may use a "fast track" route in the secretory pathway, whereas hCG {alpha}ß-heterodimers move on the regular route. The "fast track" anterograde transport of cargo by percolating coat protein complex I vesicles opposite to the "slow track" through the Golgi stack was recently suggested (48), providing an elegant suggestion for different transport velocities through the Golgi apparatus. As shown here, the sequestration of free subunits and GPH{alpha}{alpha} homodimers from hCG {alpha}ß-heterodimers is probably not only a matter of different transport velocities, but also of different maturation progress. An indication in favor of a fast-track mechanism could be the rapid development of the sialylation-based microheterogeneity of free GPH{alpha} that starts approximately 5 min after a short pulse (Fig. 2CGo).

In conclusion, GPH{alpha} and hCGß seem to associate immediately after the translation attended by the first rapid folding steps. Nonassociated subunit variants like free hCGß and GPH{alpha}, as well as GPH{alpha}{alpha} homodimers, seem to fold and to complete disulfide bridge formation much faster than hCG {alpha}ß-heterodimers. This seems to indicate that the intersubunit interactions in the hCG {alpha}ß-heterodimers appear to decelerate the folding speed, particularly of the hCGß part. This apparently increases the retention time in the quality control system of the ER, leading to a sequestration of free subunits and hCG {alpha}ß-heterodimers in the secretory pathway. To our knowledge a negative impact of subunit assembly of oligomeric glycoproteins on the folding has not been previously reported. The opposite applies for immunoglobulin biosynthesis. In that case, assembly of heavy and light chains seems to drive the complete folding of protein subunits (49).


   Acknowledgments
 
We thank U. Doyle for language editing.


   Footnotes
 
This work was supported by grants from the Deutsche Forschungsgemeinschaft, Bonn-Bad Godesberg (Me 545/12-1, 2) (to W.E.M.), the Austrian Science Fund (FWF; NRN, S9307B05) (to P.B.), and a fellowship of the Alexander von Humboldt Foundation, Bonn-Bad Godesberg (to V.S.).

Disclosure Statement: The authors have nothing to disclose.

First Published Online August 30, 2007

1 W.E.M. and J.-M.K. are equally contributing authors. Back

2 In memory of V.S., who has passed away. Back

Abbreviations: CNX, Calnexin; CRT, calreticulin; ER, endoplasmic reticulum; GPH{alpha}, glycoprotein hormone {alpha}-subunit; hCG, human chorionic gonadotropin; hCGß, human chorionic gonadotropin ß-subunit; NEM, N-ethylmaleinimide; PDI, protein disulfide isomerase; 2D, two-dimensional.

Received June 13, 2007.

Accepted for publication August 22, 2007.


   References
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 References
 

  1. Pierce JG, Parsons TF 1981 Glycoprotein hormones: structure and function. Annu Rev Biochem 50:465–495[CrossRef][Medline]
  2. Lapthorn AJ, Harris DC, Littlejohn A, Lustbader JW, Canfield RE, Machin KJ, Morgan FJ, Isaacs NW 1994 Crystal structure of human chorionic gonadotropin. Nature 369:455–461[CrossRef][Medline]
  3. Wu H, Lustbader JW, Liu Y, Canfield RE, Hendrickson WA 1994 Structure of human chorionic gonadotropin at 2.6 A resolution from MAD analysis of the selenomethionyl protein. Structure 2:545–558[Medline]
  4. Kobata A, Takeuchi M 1999 Structure, pathology and function of the N-linked sugar chains of human chorionic gonadotropin. Biochim Biophys Acta 1455:315–326[Medline]
  5. Dirnhofer S, Hermann M, Hittmair A, Hoermann R, Kapelari K, Berger P 1996 Expression of the human chorionic gonadotropin-ß gene cluster in human pituitaries and alternate use of exon 1. J Clin Endocrinol Metab 81:4212–4217[Abstract]
  6. Matzuk MM, Keene JL, Boime I 1989 Site specificity of the chorionic gonadotropin N-linked oligosaccharides in signal transduction. J Biol Chem 264:2409–2414[Abstract/Free Full Text]
  7. Vitt UA, Hsu SY, Hsueh AJ 2001 Evolution and classification of cystine knot-containing hormones and related extracellular signaling molecules. Mol Endocrinol 15:681–694[Abstract/Free Full Text]
  8. Hogg PJ 2003 Disulfide bonds as switches for protein function. Trends Biochem Sci 28:210–214[CrossRef][Medline]
  9. Linke K, Jakob U 2003 Not every disulfide lasts forever: disulfide bond formation as a redox switch. Antioxid Redox Signal 5:425–434[CrossRef][Medline]
  10. Xing Y, Myers R V, Cao D, Lin W, Jiang M, Bernard MP, Moyle WR 2004 Glycoprotein hormone assembly in the endoplasmic reticulum: II. Multiple roles of a redox sensitive beta-subunit disulfide switch. J Biol Chem 279:35437–35448[Abstract/Free Full Text]
  11. Hammond C, Helenius A 1995 Quality control in the secretory pathway. Curr Opin Cell Biol 7:523–529[CrossRef][Medline]
  12. Kamhi-Nesher S, Shenkman M, Tolchinsky S, Fromm SV, Ehrlich R, Lederkremer GZ 2001 A novel quality control compartment derived from the endoplasmic reticulum. Mol Biol Cell 12:1711–1723[Abstract/Free Full Text]
  13. Hebert DN, Garman SC, Molinari M 2005 The glycan code of the endoplasmic reticulum: asparagine-linked carbohydrates as protein maturation and quality-control tags. Trends Cell Biol 15:364–370[CrossRef][Medline]
  14. Schlumberger HD, Blobel R 1969 [Reoxidation of the disulfide bridges of human chorionic gonadotropin (HCG)]. Z Naturforsch B 24:54–57 (German)[Medline]
  15. Huth JR, Mountjoy K, Perini F, Bedows E, Ruddon RW 1992 Domain-dependent protein folding is indicated by the intracellular kinetics of disulfide bond formation of human chorionic gonadotropin ß subunit. J Biol Chem 267:21396–21403[Abstract/Free Full Text]
  16. Ruddon RW, Bedows E 1997 Assisted protein folding. J Biol Chem 272:3125–3128[Free Full Text]
  17. Singh V, Merz WE 2000 Disulfide bond formation is not required for human chorionic gonadotropin subunit association. Studies with dithiothreitol in JEG-3 cells. J Biol Chem 275:11765–11770[Abstract/Free Full Text]
  18. Mayor U, Guydosh NR, Johnson CM, Grossmann JG, Sato S, Jas GS, Freund SM, Alonso DO, Daggett V, Fersht AR 2003 The complete folding pathway of a protein from nanoseconds to microseconds. Nature 421:863–867[CrossRef][Medline]
  19. Xing Y, Williams C, Campbell RK, Cook S, Knoppers M, Addona T, Altarocca V, Moyle WR 2001 Threading of a glycosylated protein loop through a protein hole: implications for combination of human chorionic gonadotropin subunits. Protein Sci 10:226–235[Abstract/Free Full Text]
  20. Beebe JS, Mountjoy K, Krzesicki RF, Perini F, Ruddon RW 1990 Role of disulfide bond formation in the folding of human chorionic gonadotropin ß subunit into an {alpha} ß dimer assembly-competent form. J Biol Chem 265:312–317[Abstract/Free Full Text]
  21. Bedows E, Huth JR, Suganuma N, Bartels CF, Boime I, Ruddon RW 1993 Disulfide bond mutations affect the folding of the human chorionic gonadotropin-ß subunit in transfected Chinese hamster ovary cells. J Biol Chem 268:11655–11662[Abstract/Free Full Text]
  22. Huth JR, Perini F, Lockridge O, Bedows E, Ruddon RW 1993 Protein folding and assembly in vitro parallel intracellular folding and assembly. Catalysis of folding and assembly of the human chorionic gonadotropin {alpha} ß dimer by protein disulfide isomerase. J Biol Chem 268:16472–16482[Abstract/Free Full Text]
  23. Schulz I 1990 Permeabilizing cells: some methods and applications for the study of intracellular processes. Methods Enzymol 192:280–300[Medline]
  24. Beckers CJ, Keller DS, Balch WE 1987 Semi-intact cells permeable to macromolecules: use in reconstitution of protein transport from the endoplasmic reticulum to the Golgi complex. Cell 50:523–534[CrossRef][Medline]
  25. Gething MJ, Sambrook J 1992 Protein folding in the cell. Nature 355:33–45[CrossRef][Medline]
  26. Ruddon RW, Krzesicki RF, Norton SE, Beebe JS, Peters BP, Perini F 1987 Detection of a glycosylated, incompletely folded form of chorionic gonadotropin ß subunit that is a precursor of hormone assembly in trophoblastic cells. J Biol Chem 262:12533–12540[Abstract/Free Full Text]
  27. Butler SA, Laidler P, Porter JR, Kicman AT, Chard T, Cowan DA, Iles RK 1999 The ß-subunit of human chorionic gonadotrophin exists as a homodimer. J Mol Endocrinol 22:185–192[Abstract]
  28. Darling RJ, Wilken JA, Miller-Lindholm AK, Urlacher TM, Ruddon RW, Sherman SA, Bedows E 2001 Functional contributions of noncysteine residues within the cystine knots of human chorionic gonadotropin subunits. J Biol Chem 276:10692–10699[Abstract/Free Full Text]
  29. Reddy P, Sparvoli A, Fagioli C, Fassina G, Sitia R 1996 Formation of reversible disulfide bonds with the protein matrix of the endoplasmic reticulum correlates with the retention of unassembled Ig light chains. EMBO J 15:2077–2085[Medline]
  30. Feng W, Bedows E, Norton SE, Ruddon RW 1996 Novel covalent chaperone complexes associated with human chorionic gonadotropin ß subunit folding intermediates. J Biol Chem 271:18543–18548[Abstract/Free Full Text]
  31. Hellman R, Vanhove M, Lejeune A, Stevens FJ, Hendershot LM 1999 The in vivo association of BiP with newly synthesized proteins is dependent on the rate and stability of folding and not simply on the presence of sequences that can bind to BiP. J Cell Biol 144:21–30[Abstract/Free Full Text]
  32. Roig J, Krause JM, Berger P, Merz WE 2007 Time-dependant folding of immunological epitopes of the human chorionic gonadotropin ß-subunit. Mol Cell Endocrinol 260- 262:12–22[CrossRef]
  33. Lodish HF, Jacobsen M 1972 Regulation of hemoglobin synthesis. Equal rates of translation and termination of {alpha}- and ß-globin chains. J Biol Chem 247:3622–3629[Abstract/Free Full Text]
  34. Batey S, Randles LG, Steward A, Clarke J 2005 Cooperative folding in a multi-domain protein. J Mol Biol 349:1045–1059[CrossRef][Medline]
  35. Bergmann JE, Lodish HF 1979 A kinetic model of protein synthesis. Application to hemoglobin synthesis and translational control. J Biol Chem 254:11927–11937[Abstract/Free Full Text]
  36. Roth J 2002 Protein N-glycosylation along the secretory pathway: relationship to organelle topography and function, protein quality control, and cell interactions. Chem Rev 102:285–303[CrossRef][Medline]
  37. Berger EG 1985 How Golgi-associated glycosylation works. Cell Biol Int Rep 9:407–417[CrossRef][Medline]
  38. Rabouille C, Hui N, Hunte F, Kieckbusch R, Berger EG, Warren G, Nilsson T 1995 Mapping the distribution of Golgi enzymes involved in the construction of complex oligosaccharides. J Cell Sci 108:1617–1627[Abstract]
  39. Fenteany FH, Colley KJ 2005 Multiple signals are required for alpha2,6-sialyltransferase (ST6Gal I) oligomerization and Golgi localization. J Biol Chem 280:5423–5429[Abstract/Free Full Text]
  40. Leitzgen K, Haas IG 1998 Protein maturation in the endoplasmic reticulum. Chemtracts-Biochem Mol Biol 11:423–445
  41. Schrag JD, Procopio DO, Cygler M, Thomas DY, Bergeron JJ 2003 Lectin control of protein folding and sorting in the secretory pathway. Trends Biochem Sci 28:49–57[CrossRef][Medline]
  42. Helenius A, Aebi M 2004 Roles of N-linked glycans in the endoplasmic reticulum. Annu Rev Biochem 73:1019–1049[CrossRef][Medline]
  43. Molinari M, Helenius A 1999 Glycoproteins form mixed disulphides with oxidoreductases during folding in living cells. Nature 402:90–93[CrossRef][Medline]
  44. Frickel EM, Frei P, Bouvier M, Stafford WF, Helenius A, Glockshuber R, Ellgaard L 2004 ERp57 is a multifunctional thiol-disulfide oxidoreductase. J Biol Chem 279:18277–18287[Abstract/Free Full Text]
  45. Roth J, Zuber C, Guhl B, Fan JY, Ziak M 2002 The importance of trimming reactions on asparagine-linked oligosaccharides for protein quality control. Histochem Cell Biol 117:159–169[CrossRef][Medline]
  46. Cosgrove DE, Campain JA, Cox GS 1989 Chorionic gonadotropin synthesis by human tumor cell lines: examination of subunit accumulation, steady-state levels of mRNA, and gene structure. Biochem Biophys Acta 1007:44–54[Medline]
  47. Krause J-M, Berger P, Roig J, Singh V, Merz WE 2007 Rapid maturation of glycoprotein hormone free {alpha}-subunit (GPH{alpha}) and GPH{alpha}{alpha} homodimers. Mol Endocrinol 21:2551–2564[Abstract/Free Full Text]