Aggregation and Lack of Secretion of Most Newly Synthesized

0013-7227/04/$15.00/0 Printed in U.S.A.

Endocrinology 145(8):3840 –3849 Copyright © 2004 by The Endocrine Society doi: 10.1210/en.2003-1512

Aggregation and Lack of Secretion of Most Newly Synthesized Proinsulin in Non-␤-Cell Lines YONG LIAN ZHU, ALEXANDER ABDO, JOAN F. GESMONDE, KATHLEEN C. ZAWALICH, WALTER ZAWALICH, AND PRISCILLA S. DANNIES Department of Pharmacology, Yale School of Medicine and Yale School of Nursing, New Haven, Connecticut 06520 Myoblasts transfected with HB10D insulin secrete more hormone than those transfected with wild-type insulin, as published previously, indicating that production of wild-type insulin is not efficient in these cells. The ability of non-␤-cells to produce insulin was examined in several cell lines. In clones of neuroendocrine GH4C1 cells stably transfected with proinsulin, two thirds of 35S-proinsulin was degraded within 3 h of synthesis, whereas 35S-prolactin was stable. In transiently transfected neuroendocrine AtT20 cells, half of 35S-proinsulin was degraded within 3 h after synthesis, whereas 35S-GH was stable. In transiently transfected fibroblast COS cells, 35Sproinsulin was stable for longer, but less than 10% was secreted 8 h after synthesis. Proinsulin formed a concentrated patch detected by immunofluorescence in transfected cells

that did not colocalize with calreticulin or BiP, markers for the endoplasmic reticulum, but did colocalize with membrin, a marker for the cis-medial Golgi complex. Proinsulin formed a Lubrol-insoluble aggregate within 30 min after synthesis in non-␤-cells but not in INS-1E cells, a ␤-cell line that normally produces insulin. More than 45% of 35S-HB10D proinsulin was secreted from COS cells 3 h after synthesis, and this mutant formed less Lubrol-insoluble aggregate in the cells than did wild-type hormone. These results indicate that proinsulin production from these non-␤-cells is not efficient and that proinsulin aggregates in their secretory pathways. Factors in the environment of the secretory pathway of ␤-cells may prevent aggregation of proinsulin to allow efficient production. (Endocrinology 145: 3840 –3849, 2004)

I

insulin is produced in non-␤-cells, either proinsulin has been modified so it is cleaved by other enzymes or the appropriate processing enzymes have been introduced into non-␤-cells. Apart from proteolytic processing, the extent to which cells that are not ␤-cells efficiently transport proinsulin through the secretory pathway has not been extensively investigated; in general, secretion into the blood or culture medium has been measured. Myocytes transfected with the mutant HB10D proinsulin secrete more insulin than those transfected with wild-type hormone (22). The HB10D mutant has sometimes been used in engineering replacements for ␤-cells because of the enhanced secretion (12, 22, 23). The increased secretion of the mutant from myoblasts differs from its behavior in ␤-cells, in which less of the HB10D mutant is secreted than wild-type hormone, and some HB10D mutant is degraded (26). The increased production of HB10D proinsulin from myoblasts indicates that these cells must not produce wild-type hormone efficiently. Efficient production of insulin would be desirable in cells designed to replace ␤-cells. We followed the stability and secretion of newly synthesized proinsulin in non-␤-cells to determine whether proinsulin is efficiently transported through the secretory pathway of these cells.

NSULIN IS SYNTHESIZED in ␤-cells of pancreatic islets as the precursor preproinsulin, which is transported into the lumen of the endoplasmic reticulum during its synthesis (1). The signal sequence is removed by proteolytic cleavage to yield proinsulin, which folds, and the correct disulfide bonds form. Monomeric proinsulin has hydrophobic surfaces that are not exposed after proinsulin self-associates to form a hexamer (1). In the endoplasmic reticulum, proinsulin is bound to the chaperone BiP, and this association prevents exposure of the hydrophobic surfaces of the monomer (2). The formation of dimers and hexamers may begin in the endoplasmic reticulum and occurs during transport to the Golgi complex (2, 3). Zinc may facilitate hexamer formation and is present in relatively high concentrations in the lumen of the Golgi complex of islet cells (4). C-peptide sequences connect the A and B chains in proinsulin and are located on the surface of the hexamer, accessible to proteolytic enzymes (1). The enzymes that cleave the C-peptide are activated in immature secretory granules that form from the trans layer of the Golgi complex (5). After proinsulin is converted to insulin in the secretory granule, insulin forms an insoluble crystalline-like core in most species (6). It is stored in this form in secretory granules until it is released from the ␤-cells. Proinsulin has been expressed in many different cell types in attempts to develop cells producing insulin to replace ␤-cells that are not functioning properly (7–25). Enzymes that process proinsulin are not present in all cells, so that when Abbreviations: BiP, Chaperone in the endoplasmic reticulum; SDS, sodium dodecyl sulfate. Endocrinology is published monthly by The Endocrine Society (http:// www.endo-society.org), the foremost professional society serving the endocrine community.

Materials and Methods Plasmids and transfections pcDNA3 containing sequences coding for rat preproinsulin 2 was created using clone pBC12B1 (American Type Culture Collection, Manassas, VA) (27); Not1 and XbaI restriction sites were added using primers 5⬘-aaagcggccgctacagtcggaaaccat-3⬘ and 5⬘-tgttctagatggacagggcagtggt-3⬘. The mutant VA3L proinsulin was created by the technique of Ho et al. (28) using as primers 5⬘-ggcatcttggatcagtgctgcacc-3⬘ and 5⬘-gcactgatccaagatgccgcgctt-3⬘. The mutant HB10D proinsulin was created using

3840

Zhu et al. • Proinsulin Aggregation in Non-␤-Cells

primers 5⬘-gtggttctgacttggtggaagct-3⬘ and 5⬘-ccaccaagtcagaaccacaaag3⬘. All constructions were sequenced in both directions by the WM Keck Foundation Biotechnology Resource Laboratory (Yale University, New Haven, CT) and the sequences determined to be correct, with no ambiguous base assignments. The plasmid containing wild-type preproinsulin 2 was sequenced again after performing the experiments described here and was still correct with no ambiguous base assignments, so that no mutations had occurred in the coding sequence during replication in bacteria. GH4C1 cells were transfected using Superfect (Qiagen, Santa Clarita, CA) with pcDNA3 containing the preproinsulin sequences and with Bluescript SK⫹ containing the neomycin resistance gene under the control of the Rous sarcoma virus promoter. The transfected cells were incubated with 0.5 mg/ml G418 and surviving colonies separately isolated after 2–3 wk and each screened for insulin immunoreactivity by RIA. Clones producing insulin were expanded further. AtT20 and COS cells were plated at a density of 2.3 ⫻ 105 cells per 60-mm plate and transiently transfected 1 d after plating using 30 ␮l Superfect and 5 ␮g pcDNA3 containing the preproinsulin or human GH sequences as previously described, unless otherwise stated (29).

Pulse chase procedure Clones of GH4C1 cells expressing wild-type proinsulin or the mutants, 1.5 ⫻ 105 cells per 60-mm plate, were cultured in a 1:1 mixture of DMEM and Ham’s F10 nutrient mixture plus 15% gelding serum (Central Biomedia Inc., Irwin, MO) with hormone treatment (1 nm estradiol and 5 nm epidermal growth factor), unless otherwise indicated and used 4 d after plating. INS-1E cells were cultured as described (30, 31) and used 1 d after plating. AtT20 and COS cells were cultured in DMEM with 10% fetal bovine serum and 2 mm [scap]l-glutamine. They were transiently transfected with Superfect as described above and used the day after transfection. For incorporation of 35S-amino acids, cells were incubated with 200 ␮C Express 35S-Protein labeling mix (DuPont-NEN Life Science Products, Boston, MA) in cysteine and methionine-free DMEM with 10 mm 2-(N-morpholine) ethane sulfonic acid, 10 mm HEPES, 4 mm NaHC03, and 5% serum. During the chase period, cells were incubated with DMEM plus 2 mm cysteine, 2 mm methionine, 10 mm HEPES, 10 mm 2-(N-morpholine) ethane sulfonic acid, 4 mm NaHC03, and 15% serum. Both insulin and proinsulin are stable when incubated in this medium with or without cells (32). Both pulse and chase medium were equilibrated in an atmosphere of 5% CO2, 95% air before use. Excess methionine and cysteine were in the chase medium during this equilibration. The 35S-amino acids were added after equilibration immediately before the pulse period. When indicated, cells were incubated with 35S-methionine alone; in these conditions cystine was present at concentrations of 0.2 mm in the pulse and chase medium, the normal concentrations in the medium. For analysis at each time point, medium and cells were collected separately. Medium was centrifuged at 5000 ⫻ g for 15 sec, and the pellet, containing any cells that had detached from the plate, was combined with the cells from the plate. Cells were dissolved in lysis buffer [50 mm Tris-HCl (pH 8.0), 150 mm NaCl, 0.2% sodium azide, 0.1% sodium dodecyl sulfate (SDS), 2% Triton X-100, 0.5% sodium deoxycholate, 100 ␮g/ml phenylmethylsulfonylfluoride, 1 ␮g/ml aprotinin, 2.5 ␮g/ml leupeptin, and 2.5 ␮g/ml pepstatin]. Protein A Sepharose beads (5 mg/sample) that had been first incubated with antiserum were added to each sample of medium and lysate and incubated at room temperature for 18 h with continual mixing of each sample. The beads were washed once with 50 mm Tris HCl (pH 7.5), 150 mm NaCl, 0.2% Triton X-100, 1 mm EDTA, 0.2% sodium azide, 0.25% BSA, 2.5 ␮g/ml leupeptin, and 2.5 ␮/ml pepstatin. Proteins were eluted from the beads by heating at 100 C for 5 min and 60 C for 15 min in 125 mm Tris HCl (pH 6.8), 4% SDS, 0.02% bromophenol blue, 10% glycerol, 10% ␤-mercaptoethanol, and 200 mm dithiothreitol. After electrophoresis on 18% acrylamide gels with SDS, gels were dried at 75 C for 2 h under vacuum and then quantification and visualization of 35S-protein bands performed using a molecular imager (Bio-Rad Laboratories, Hercules, CA). The Tricine buffer system was not routinely used, although it gives sharper bands of insulin chains because low-molecular-weight bands may be lost when the gel is fixed to remove the urea before drying (33). There was no centrifugation before immunoprecipitation, a step sometimes used by others to eliminate background. When solubility of 35S-proinsulin in

Endocrinology, August 2004, 145(8):3840 –3849 3841

Lubrol was investigated, cells were dissolved in 1.5% Lubrol, 0.32 m sucrose, 0.5% BSA, 10 mm sodium phosphate (pH 7), 10 ␮g/ml phenylmethylsulfonylfluoride, 1 ␮g/ml aprotenin, 2.5 mg/ml leupeptin, and 2.5 ␮g/ml pepstatin. The lysate centrifuged at 50,000 ⫻ g and immunoprecipitation performed on the supernatant and pellet after adding lysis buffer (29). Insulin was precipitated using guinea pig antiporcine insulin (Sigma, St. Louis, MO), 2 ␮l per sample; this amount caused the maximum amount of 35S-proinsulin and 35S-insulin to be precipitated from the samples, determined by repeating the immunoprecipitation or by using more antiserum. Two microliters of this antiserum precipitated 90 ng insulin. The total maximum amount that accumulated in transfected cultures was 30 ng/1.5 ⫻ 105 cells in 24 h in initial screening experiments, determined by RIA. That antiserum plus one made to the peptide sequence ALEVARQKRGILDQC by Affinity Bioreagents were used to precipitate the mutant proinsulins. The antiserum to human GH was provided by Dr. A. F. Parlow (National Hormone and Pituitary Program, Torrance, CA).

Immunocytochemistry Cells were grown on glass coverslips and transfected as described above for the transiently transfected cells; the transfection efficiency was lower than when we used the same protocol in plastic culture dishes. Two days after plating and 1 d after transient transfections, cells were fixed in 2% formaldehyde and 120 mm sodium phosphate buffer (pH 7.4). To stain insulin, the primary antiserum was guinea pig antiporcine insulin, and the secondary antiserum goat antiguinea pig conjugated to Texas Red (Sigma) used for single staining, and donkey antiguinea pig Ig conjugated to Texas Red [Affinipure quality from Jackson Laboratories (Bar Harbor, ME), absorbed with immunoglobulins from several different species including guinea pig] used for double staining. Other primary antisera were rabbit anticalreticulin and rabbit anti-GRP78 (BiP) (Affinity BioReagents, Golden, CO), mouse anti-Vti1b (BD Biosciences, Franklin Lakes, NJ), and mouse antimembrin (StressGen Biotechnologies Corp., San Diego, CA). Fluorescein isothiocyanate-conjugated antimouse or rabbit Ig were Affinipure quality (Jackson Laboratories). Mouse anti-TGN38 was a gift from Dr. Barbara Reaves. Confocal microscopy was performed on a LSM 510 microscope (Carl Zeiss, Go˜ ttingen, Germany) at the Yale University Center for Cell Imaging.

Results

GH4C1 cells are a clone of rat pituitary tumor cells that make prolactin and GH. These cells accumulate secretory granules with dense cores, especially when cultured with insulin, estradiol, and epidermal growth factor (34). The patterns of prolactin release from GH4C1 cells and prolactin release from primary cultures of lactotrophs are similar when release is stimulated by several different mechanisms (35), indicating that the regulated pathway of secretory granules in these cells behaves as it does in normal lactotrophs. When these cells are transfected with proinsulin, the patterns of stimulated release of prolactin and proinsulin are similar (32). GH4C1 cells lack prohormone convertases 1 and 2 (36) and do not process proinsulin to insulin when they express it (32, 37). In clones of GH4C1 cells stably transfected with proinsulin sequences, newly synthesized proinsulin was assayed in cells after incubation with 35S-amino acids for 10 min. The immunoprecipitate contained a band of 35S-protein that migrated to the position of proinsulin and was not present in untransfected cells (Fig. 1A). Three hours after synthesis, there was no conversion to 35S-insulin (Fig. 1, B and C). We confirmed the lack of conversion by assaying insulin immunoreactivity of an acid extract of the cell lysate by RIA after chromatography of cell lysates on Biogel P-30; all the immunoreactivity migrated to the position of proinsulin (not shown), as found previously (32).

3842

Endocrinology, August 2004, 145(8):3840 –3849


estradiol and epidermal growth factor (34), but newly synthesized 35S-proinsulin was not stable regardless of whether cells were treated to increase prolactin storage (Fig. 1, B and C). There was more 35S-prolactin in treated cells because estradiol and epidermal growth factor increase prolactin synthesis (32). The experimental conditions did not cause instability of all newly synthesized proteins because 35S-prolactin was stable. The majority of newly synthesized proinsulin was not stable, whereas newly synthesized prolactin was stable in all clones examined (Fig. 1D). Newly synthesized proinsulin was also not stable in transiently transfected AtT20 cells. AtT20 cells are mouse pituitary cells that produce endorphins and ACTH and have secretory granules (38, 39). In these cells, some loss of 35Sproinsulin with time after synthesis was caused by conversion to insulin. Ten percent of the 35S-proinsulin present

FIG. 1. Stability of newly synthesized prolactin and proinsulin in clones of GH4C1 cells transfected with proinsulin. Panel A, Incorporation of 35S-amino acids into proinsulin in a stably transfected clone of GH4C1 cells. Immunoprecipitation was performed after a 10-min incubation with 35S-amino acids. Untransfected GH4C1 cells are also shown. The positions to which proinsulin (PINS) and the insulin chains (INS) migrate are shown. Panel B, Incorporation of 35S-amino acids into prolactin and proinsulin in a clone that had been cultured in the absence of estradiol and epidermal growth factor. Pulse, Immunoprecipitation performed after a 10-min incubation with 35Samino acids. Three-hour chase, Immunoprecipitation performed 3 h after the end of the 10-min incubation with 35S-amino acids. C, Cell lysate; M, medium; PRL, prolactin. Panel C, Incorporation of 35Samino acids into prolactin and proinsulin in the same clone as B but cultured with 1 nM estradiol and 5 nM epidermal growth factor. Panel D, Summary of experiments performed with three clones of GH4C1 cells transfected with proinsulin. The amount of 35S-prolactin and 35 S-proinsulin remaining 3 h after the incubation with 35S-amino acids is shown as a percent of the amounts that were present at the end of the 10-min pulse. The means ⫾ SE are given.

Total 35S-proinsulin, in the cells and medium, was reduced 3 h after synthesis compared with that present immediately after synthesis (Fig. 1). We found a similar loss of 35S-proinsulin in six clones, three each from two independent transfections. We have previously shown proinsulin is stable in the medium (32). Serum in the medium did not interfere with immunoprecipitation of proinsulin because when cell lysates containing 35S-proinsulin were combined with medium, recovery of 35S-proinsulin was 98 ⫾ 10% of that recovered when immunoprecipitation was performed directly from cell lysates. Therefore, newly synthesized proinsulin appears unstable in the cells. GH4C1 cells have more secretory granules and a larger capacity to store prolactin when treated with

FIG. 2. Stability of newly synthesized proinsulin and GH in transiently transfected AtT20 cells. Panel A, Incorporation of 35S-amino acids into proinsulin and insulin in AtT20 cells transfected with proinsulin sequences, indicated by ⫹, or not transfected, indicated by ⫺. Pulse, Immunoprecipitation performed after a 20-min incubation with 35S-amino acids. Three-hour chase, Immunoprecipitation performed 3 h after the end of the 20-min incubation with 35S-amino acids. C, Cell lysate; M, medium; PINS, proinsulin; INS, insulin chains. Panel B, Incorporation of 35S-amino acids into human GH in cells transfected with GH sequences. Panel C, Summary of experiments performed with AtT20 cells transiently transfected with proinsulin or GH. The total amounts of 35S-proinsulin and insulin or 35 S-GH in the lysates and medium remaining 3 h after the incubation with 35S-amino acids is shown as a percent of the amounts that were present at the end of the 20-min pulse. The means ⫾ SE are given. PINS, Proinsulin plus insulin.



immediately after synthesis migrated to the position of insulin chains 3 h later; these bands, although faint, were not present in untransfected cells (Fig. 2). Longer incubations with 35S-amino acids, which label stable components to a proportionally greater extent (40), resulted in a greater amount of 35S-insulin as a percent of total 35S-hormone (not shown), consistent with results of others using longer labeling times (37). The total amount of 35S-insulin and 35S-proinsulin present in cells and medium 3 h after synthesis was reduced, compared with that present immediately after incubation with 35S-amino acids (Fig. 2A). Experimental conditions did not cause degradation of all secretory proteins in AtT20 cells because newly synthesized human GH was about 130%, but total proinsulin and insulin about 55%, of that present at the end of the incubation with 35S-amino acids. Incorporation of 35S-amino acids into protein must continue during the chase period in AtT20 cells accounting for the greater than 100% recovery of GH; continued incorporation most likely occurs because an internal pool of amino acids derived from protein breakdown does not readily equilibrate with amino acids transported from outside the cell (40 – 42). Newly synthesized human prolactin also is stable in transiently transfected AtT20 cells (43). The loss of proinsulin occurred to the same extent when 35S-methionine alone was used; the amount remaining after 3 h was 53 ⫾ 4%, n ⫽ 2, of that present at the end of the incubation with 35S-methionine. COS cells are fibroblast cells that do not have secretory FIG. 4. Stability of newly synthesized proinsulin and proinsulin mutants in transfected clones of GH4C1 cells. Panel A, Incorporation of 35 S-amino acids into proinsulin mutants after a 10-min incubation with 35S-amino acids or 3 h after the end of the 10-min incubation. VAL, Incubation with a clone of GH4C1 cells expressing VA3L proinsulin; VALCAS, incubation with a clone of GH4C1 cells expressing VA3L,CA7S proinsulin. C, Cell lysate; M, medium. Panel B, Summary of experiments with clones of GH4C1 cells expressing proinsulin or proinsulin mutants. The amount of 35S-proinsulin or mutant proinsulin remaining with time after the end of the 10-min incubation with 35 S-amino acids is shown as a percent of the amounts that were present at the end of the 10-min pulse. The values are the mean ⫾ range or SE from one or more experiments with each clone; where no bars are shown, they fell within the symbol. Triangles, Results from three clones of GH4C1 cells transfected with proinsulin; circles, results from two clones of GH4C1 cells transfected with VA3L proinsulin; squares, results from two clones of GH4C1 cells transfected with VA3L,CA7S proinsulin.

FIG. 3. Stability of newly synthesized proinsulin in transiently transfected COS cells. Panel A, Incorporation of 35S-amino acids into proinsulin and insulin in COS cells. Pulse, Immunoprecipitation performed after a 20-min incubation with 35S-amino acids. Eight-hour chase, Immunoprecipitation performed 8 h after the end of the 20-min incubation. C, Cell lysate; M, medium; PINS, proinsulin. Panel B, Summary of experiments with COS cells expressing proinsulin. The total amounts of 35S-proinsulin remaining at times after the incubation with 35S-amino acids is shown as a percent of the amounts that were present at the end of the 20-min pulse.

granules and rapidly secrete human prolactin and GH (29, 43). When COS cells were transiently transfected with proinsulin, however, less than 10% of newly synthesized proinsulin was secreted 8 h after synthesis (Fig. 3A), so proinsulin was also not efficiently secreted from COS cells. Proinsulin, however, was stable for longer times in COS cells; 80% was still present 8 h after synthesis (Fig. 3B). Proinsulin was also retained and less than 10% secreted 8 h after synthesis in CHO cells (not shown). Steiner and colleagues (26) demonstrated that 20% of the human HB10D mutant of proinsulin was degraded, and 15% was rapidly secreted when it was expressed in mouse ␤-cells. HB10D insulin binds more tightly to insulin receptors than wild-type hormone (44). Steiner et al. (45) proposed that a likely explanation is that enhanced binding of the HB10D

3844



FIG. 6. Immunostaining of a clone of GH4C1 cells stably expressing proinsulin. Top row, Calreticulin (green) and proinsulin (red). Middle row, Membrin (green) and proinsulin (red). Bottom row, TGN38 (green) and proinsulin (red). The second antibody used to stain proinsulin was donkey antiguinea pig Ig, absorbed with Igs from other species.

FIG. 5. Immunostaining of cells transfected and with proinsulin (PINS) or VA3L,CA7S proinsulin (VALCAS), AtT20 cells and COS cells were transiently transfected, and GH4C1 cells were stably transfected. The second antibody is goat antiguinea pig Ig.

mutant to newly synthesized or recycling insulin receptors in ␤-cells diverts the mutant to pathways other than storage in secretory granules. Substituting a leucine for valine 3 in the A chain reduces receptor binding by two orders of magnitude; the combined mutant VA3L, HB10D proinsulin was efficiently stored in secretory granules in mouse ␤-cells, suggesting that binding to the insulin receptor affects transport of the HB10D mutant through the secretory pathway (45). We tested whether the mutation that reduced binding to the receptor affected transport in the transfected cells. In stably transfected clones of GH4C1 cells expressing VA3L-proinsulin, the mutant behaved as wild-type insulin (Fig. 4), an indication that binding to the insulin receptor is not the cause of inefficient secretion in these cells. A mutant of insulin in which there is a substitution for cysteine 7 in the A chain cannot form disulfide bonds correctly and is degraded in ␤-cells without transport beyond the endoplasmic reticulum (46). In clones of GH4C1 cells expressing VA3L,CA7S-proinsulin, this mutant disappeared more rapidly than wild-type proinsulin (Fig. 4), so the behavior of a protein that cannot fold normally differs from that of wild-type proinsulin. Immunofluorescent staining of GH4C1 and COS cells for proinsulin demonstrates that VA3L,CA7S-proinsulin shows a pattern of staining consis-

FIG. 7. Immunostaining of transiently transfected COS cells (top row) and AtT20 cells (bottom three rows). Top two rows, calreticulin (green) and proinsulin (red). Third row down, BiP (green) and proinsulin (red). Bottom row, Vti1b (green) and proinsulin (red). The second antibody used to stain proinsulin was donkey antiguinea pig Ig, absorbed with Igs from other species.

tent with retention in the endoplasmic reticulum in both cell types. The pattern of staining for wild-type proinsulin differs in that there are bright perinuclear patches in both COS and GH4C1 cells, consistent with accumulation in the Golgi com-


FIG. 8. Effect of chloroquine and bafilomycin on stability of proinsulin in clones of GH4C1 cells. The amount of 35S-proinsulin or mutant proinsulin remaining 2 h after the incubation with 35S-amino acids is shown as a percent of the amounts that were present at the end of the 10-min incorporation period. A, Effect of 30 ␮M chloroquine. PINS, Proinsulin; VAL,CAS, VA3L,CA7S proinsulin; c, control; chl, chloroquine added to cells during the 2-h chase. B, Effect of 1.6 ␮M bafilomycin. c, Control; baf, bafilomycin added to cells during the 2-h chase.

plex as well as punctate staining in GH4C1 cells, consistent with its presence in secretory granules (Fig. 5). The pattern of staining of AtT20 cells resembled that of GH4C1 cells, with punctate staining consistent with accumulation in secretory granules. We found similar patterns of staining in AtT20 cells transfected with human GH, a protein that is stored in secretory granules in these cells (47). To determine whether the perinuclear patches of proinsulin were localized in the Golgi complex, we costained with other marker proteins. For this double staining, we used a second antibody absorbed with antibodies from guinea pig and other species, and with this antibody, detection of the proinsulin patch was less intense and primarily at the perinuclear patch for these three cell types. Proinsulin in the perinuclear patch in GH4C1 cells generally did not colocalize with calreticulin (Fig. 6, top row). Membrin is an integral membrane protein on the surface of cis and medial Golgi layers that is involved in transport from the endoplasmic reticulum to the Golgi complex (48). In a transfected GH4C1 clone, the proinsulin patch was colocalized with membrin (Fig. 6, middle row). TGN38 is a protein primarily associated with the trans-Golgi network (49, 50). There was usually little colocalization of the proinsulin patch with TGN38 in GH4C1 cells, although the two stains were in close proximity (Fig. 6, bottom row). The pattern of staining was similar in transiently transfected cells. Proinsulin also accumulated primarily in a patch that did not colocalize with calreticulin in either COS or AtT20 cells (Fig. 7). In transfected AtT20 cells, proinsulin also showed little colocalization with BiP, a chaperone in the endoplasmic reticulum. The patch of proinsulin was located near, but not usually with, staining for Vti1b, a protein located on tubules and vesicles in the trans-Golgi network and endosomes (51). Thus, in stably and transiently transfected cells, the concentrated patches of intracellular proinsulin did


FIG. 9. Solubility of newly synthesized proinsulin in Lubrol, when expressed transiently in COS and AtT20 cells and endogenously in INS-1E cells. A, Appearance of insoluble proinsulin in AtT20 and COS cells after a 30⬘ chase. Pulse, Immunoprecipitation performed after a 20-min incubation with 35S-amino acids. 30⬘ chase, Immunoprecipitation performed 30 min after the end of the 20-min incubation with 35 S-amino acids. Cells were dissolved in buffer containing Lubrol and centrifuged before immunoprecipitation. S, Supernatant; P, pellet; M, medium. B, Lack of appearance of insoluble proinsulin in INS-1E cells after a 2-h chase. INS, Insulin chains.

not colocalize with markers of the endoplasmic reticulum but were found in proximity to markers for the trans-Golgi network. The disappearance of proinsulin in transfected GH4C1 cells was not affected by neutralizing the pH of intracellular compartments, an action that reduces the activity of lysosomal enzymes. Incubation with 30 ␮m chloroquine after synthesis of proinsulin had no effect on the stability of 35S-proinsulin 2 h after synthesis in clones of GH4C1 cells expressing proinsulin or the VA3L,CA7S mutant (Fig. 8A). We obtained similar results in AtT20 cells (not shown). The concentration of 30 ␮m is sufficient to neutralize acidic intracellular compartments of GH4C1 and AtT20 cells (29, 52). Incubation of a clone of GH4C1 cells stably expressing proinsulin with 1.6 ␮m bafilomycin A1, which inhibits vacuolar ATPases, also had little or no effect on the amount of 35S-proinsulin remaining 2 h after synthesis (Fig. 8B). The pattern of staining indicates that at least some proinsulin is transported beyond the endoplasmic reticulum. It seemed unusual for a soluble protein in the Golgi complex not to be secreted from the cells, especially in COS cells in which proinsulin is stable for long periods of time. We examined the solubility of proinsulin in Lubrol with time after synthesis. Lubrol is a nonionic detergent that dissolves membranes of cells but does not disrupt all protein-protein interactions. In transfected COS and AtT20 cells, about 50% of the 35S-proinsulin was insoluble in Lubrol 30 min after synthesis (Fig. 9A). Proinsulin, however, remained soluble in INS-1E cells, a ␤-cell line that normally synthesizes proinsulin (30, 31). ␤-Mercaptoethanol, 50 ␮m, is present in the medium of INS-1E cells. Adding this concentration to the medium of AtT20 cells did not prevent the acquisition of Lubrol insolubility by proinsulin. Proinsulin in INS-1E cells

3846


did not form Lubrol-insoluble aggregates at longer times after synthesis; we examined up to 2 h (Fig. 9B). If proinsulin aggregation by self-association prevents secretion from COS cells, then reducing the amount of proinsulin expressed should reduce aggregation, and secretion would be more efficient. COS cells transiently transfected with 1 ␮g (1/5 the normal amount) still secreted less than 10% of the 35S-proinsulin 3 h after synthesis (not shown). To detect synthesis of lower amounts of proinsulin, we incubated with 35S-amino acids for longer times. COS cells transiently transfected with 5 ␮g proinsulin plasmid and incubated with 35S-amino acids for 1 h secreted less than 10% of the 35S-proinsulin 90 min later, but cells transfected with 0.1 ␮g proinsulin plasmid secrete over 40% 90 min later (Fig. 10), so that when less proinsulin was made, proportionately more was secreted. HB10D proinsulin may be secreted to a greater extent than proinsulin (22) because it has less tendency to aggregate in cells. The HB10D mutant was secreted more efficiently from transfected COS cells than wild-type proinsulin (Fig. 11A); in three experiments examining secretion, an average of 47 ⫾ 4% was secreted 3 h after synthesis, whereas only 6 ⫾ 2.3% wild-type proinsulin was secreted. Less newly synthesized HB10D proinsulin became insoluble in Lubrol than wildtype hormone 30 min after synthesis (Fig. 11, B and C), and even at that short time, more HB10D mutant was secreted than wild-type hormone (Fig. 11, B and C).


FIG. 11. Secretion and solubility of wild-type proinsulin (PINS) and HB10D proinsulin (HB10D) in transiently transfected COS cells. Panel A, Secretion into the medium. Pulse, Immunoprecipitation performed after a 20-min incubation with 35S-amino acids; 3-h chase, immunoprecipitation performed 3 h after the end of the 20-min incubation; C, cell lysate; M, medium. Panel B, Solubility after synthesis. Pulse, Immunoprecipitation performed after a 20-min incubation with 35S-amino acids; 30⬘ chase, immunoprecipitation performed 30 min after the end of the 20-min incubation. Cells were dissolved in Lubrol-containing buffer and centrifuged before immunoprecipitation. S, Supernatant; P, pellet; M, medium. Panel C, Summary of solubility and secretion of wild-type and HB10D proinsulin in three experiments after a 30-min chase.

Discussion

FIG. 10. Secretion of proinsulin from transiently transfected COS cells. Cells were transfected with 5 ␮g proinsulin in pcDNA3 or 0.1 ␮g proinsulin in pcDNA3 plus 4.9 ␮g pcDNA3 to keep the total amounts of DNA in each transfection equal. Panel A, COS cells transfected with 5 or 0.1 ␮g proinsulin plasmid; 60⬘ pulse, immunoprecipitation performed after a 60-min incubation with 35S-amino acids; 90⬘ chase, immunoprecipitation performed 90 min after the incubation with 35 S-amino acids; C, cell lysate; M, medium. Samples from 0.1 ␮g proinsulin (PINS) were exposed 160 h and from 5 ␮g 16 h. Panel B, Summary of three experiments measuring secretion. 1/1, 5 ␮g proinsulin plasmid; 1/50, 0.1 ␮g proinsulin plasmid.

By following the fate of newly synthesized proinsulin using pulse chase techniques, we found that proinsulin was not secreted efficiently from COS, CHO, AtT20, and GH4C1 cells and that a portion was degraded. Within 30 min after synthesis, proinsulin had formed an aggregate that was insoluble in the detergent Lubrol and could be separated from the lysate by centrifugation. There are several possible explanations why inefficient secretion of proinsulin from non-␤-cells has not been reported by others. First, we use relatively short labeling periods of up to 20 min when possible. The longer the labeling time, the more the labeled protein will consist of stable components and the more difficult it will be to detect a labile fraction (40). Second, we do not centrifuge the cell lysate before immunoprecipitation, except in the cases in which we are assessing the solubility of labeled proinsulin. Some other protocols have a filtration or centrifugation step before assay of the cell lysate, which may eliminate aggregated proinsulin from the assay (37, 46, 53– 60). Third, we briefly centrifuge the medium when collected to remove floating cells without breaking them and combine these with the


lysate. Some other groups do not state whether this step is performed (46, 54 –57, 59, 60); if it is not, proinsulin in the medium may come in part from cells. Fourth, we compare the amounts present during the chase to the amounts present at the start of the chase period to assess degradation. Investigations that were primarily interested in proinsulin processing reported the amounts as a percent of the total present in each culture at the time the sample was collected (54, 55, 58, 60). A loss of labeled hormone with time is not detectable by this analysis and what is secreted may be a small portion of what was initially synthesized. Fifth, if transfected cells are synthesizing less than those described here, then those with lower rates of synthesis would be predicted to secrete what is made more efficiently, based on the results with 50-fold less plasmid. The amounts of proinsulin synthesized per cell may vary with the vector used. There are some results consistent with our findings. Wang et al. (46) used transfected CHO cells and found little secretion of 35S-proinsulin up to 24 h after synthesis, and most of the 35S-proinsulin disappeared from the cultures. In transfected AtT20 cells, 70% of 35S-proinsulin and 35S-insulin present at the end of a 16-h incubation were recovered 9 h later (56) The long labeling period means the loss of hormone is likely to be underestimated. In transfected FAO cells, hepatoma cells that do not store protein, 40% of newly synthesized proinsulin was secreted in 90 min (59), which is more secretion than we found from COS cells but still relatively inefficient secretion. More than 80% of newly synthesized prolactin and GH are secreted within 60 min after synthesis in COS cells (29, 43). Finally, the increased production of the HB10D mutant, compared with wild-type from myoblasts (22), is consistent with inefficient secretion of wild-type hormone, even though pulse chase techniques were not used in those investigations. We found less processing of proinsulin in our AtT20 cell line than some others have seen (58), but the amount we saw is comparable with that found by Quinn et al. (56), so the AtT20 cells must vary in their ability to process proinsulin. We did not increase stability of proinsulin or its processing in AtT20 cells by cotransfecting the processing enzyme PC2 with proinsulin (not shown). The variation in processing may be caused by unidentified differences in culture conditions from one laboratory to another because the INS-1E cells we obtained from Drs. Claes Wolheim and Pierre Machler (University Medical Center, Geneva, Switzerland) processed more proinsulin to insulin when we first received them than they did as we continued to culture the cells. In the cells that we examined, the most concentrated area of proinsulin is near the trans-Golgi network and associated with membrin, present in the cis and medial Golgi layers. These results are consistent with transport of at least some proinsulin from the endoplasmic reticulum to the Golgi complex, but a lack of efficient transport through the Golgi complex to the trans-Golgi network. Half or more of the 35Sproinsulin disappears from GH4C1 and AtT20 cells within 3 h after synthesis. The lack of accumulation in the trans-Golgi network and the lack of effect of lysosomal inhibitors are consistent with no lysosomal involvement in degradation. A possibility is that proinsulin is carried from the Golgi com-


plex back to the endoplasmic reticulum to be degraded there. Inhibitors of proteasome activity did not reduce proinsulin disappearance (not shown), but obtaining sufficient inhibition of proteasome activity for prolonged periods in intact cells may not be possible. As an alternative possibility, there are proteolytic enzymes that are in the lumen of the endoplasmic reticulum that may degrade the aggregated proinsulin if it is returned there (61– 64). Aggregation was reduced and secretion enhanced in COS cells when histidine in position 10 in the B chain of proinsulin was mutated to aspartate. This change will eliminate coordination of zinc by histidine 10, which stabilizes formation of proinsulin hexamers in solution (65). If hexamer formation is a step in the aggregation process, then reducing hexamer accumulation may enhance secretion, but there are other possible explanations. The native structure of the HB10D mutant is more stable than wild-type hormone, and the mutant is less susceptible to formation of disulfide-linked multimers in solution (66, 67). Increased stability may prevent aggregation, either with or without intermolecular disulfide bonds, if denaturation is a step in proinsulin aggregation in cells. We were unable to detect proinsulin or proinsulin oligomers in nonreducing gels, but disulfide-bonded oligomers may not have been resolved from other highmolecular-weight proteins nonspecifically precipitated because we could not reduce the background by steps such as centrifugation. If transfected cells overexpress proinsulin, normal mechanisms that prevent protein aggregation in the secretory pathway may be overwhelmed, resulting in aggregation and inefficient production. The stably transfected GH4C1 clones synthesize somewhat less proinsulin than INS-1E cells, based on counts incorporated during a 10-min pulse, so proinsulin is not overexpressed in GH4C1 cells, compared with INS-1E cells. Aggregation of proinsulin in AtT20 and COS cells suggests that the environment of their secretory pathways differ from that in INS-1E cells, so that the ␤-cell line has factors that prevent aggregation. One marked difference is that ␤-cells have high Zn2⫹ concentrations in their secretory pathway (68). Culturing COS and AtT20 cells in Zn2⫹ concentrations up to 100 ␮m did not improve the efficiency of proinsulin production (not shown), but Zn2⫹ concentrations in cells are regulated by several transporters (69, 70), and higher extracellular Zn2⫹ concentrations may not affect concentrations available in the secretory pathway sufficiently. A second difference occurs because INS-1E cells produce both rat proinsulin I and II, whereas we have transfected cells only with rat proinsulin II. Two indirect pieces of evidence suggest that lack of proinsulin I may not be an important factor. One is that mice in which the insulin I gene has been knocked out still have normal levels of insulin and normal islet morphology (71). The second is that the human H10BD mutant is produced from non-␤-cells in larger amounts than wild-type hormone (22), so that human proinsulin must be produced inefficiently, as well as rat proinsulin II. There are other features that differ in the secretory pathway of ␤-cells, including the presence of a chaperone, cpn 60 (72). Elucidating what differences are important in causing efficient insulin production may be useful in engineering cells to replace ␤-cells.

3848


Acknowledgments We thank Drs. Claes Wolheim and Pierre Machler for INS-E1 cells, Dr. Richard Mains for the expression vector for proinsulin-converting enzyme 2, and Dr. Barbara Reaves for antibody to TGN38. Received November 7, 2003. Accepted April 23, 2004. Address all correspondence and requests for reprints to: Priscilla S. Dannies, Yale University School of Medicine, Department of Pharmacology, 333 Cedar Street, New Haven, Connecticut 06520-8066. E-mail: [email protected]. This work was supported by grants from the Juvenile Diabetes Fund, Charles E. Culpepper Biomedical Pilot Initiative, American Diabetes Association (to P.S.D.), and Cell Biology Core of the Diabetes Endocrinology Research and National Institutes of Health Grant DK-41230 (to W.Z.).

References 1. Dodson G, Steiner D 1998 The role of assembly in insulin’s biosynthesis. Curr Opin Struct Biol 8:189 –194 2. Schmitz A, Maintz M, Kehle T, Herzog V 1995 In vivo iodination of a misfolded proinsulin reveals colocalized signals for Bip binding and for degradation in the ER. EMBO J 14:1091–1098 3. Huang XF, Arvan P 1995 Intracellular transport of proinsulin in pancreatic beta-cells. Structural maturation probed by disulfide accessibility. J Biol Chem 270:20417–20423 4. Howell SL, Tyhurst M, Duvefelt H, Andersson A, Hellerstrom C 1978 Role of zinc and calcium in the formation and storage of insulin in the pancreatic ␤-cell. Cell Tissue Res 188:107–118 5. Orci L, Ravazzola M, Amherdt M, Madsen O, Perrelet A, Vassalli JD, Anderson RG 1986 Conversion of proinsulin to insulin occurs coordinately with acidification of maturing secretory vesicles. J Cell Biol 103:2273–2281 6. Michael J, Carroll R, Swift HH, Steiner DF 1987 Studies on the molecular organization of rat insulin secretory granules. J Biol Chem 262:16531–16535 7. Taniguchi H, Fukao K, Nakauchi H 1997 Constant delivery of proinsulin by encapsulation of transfected cells. J Surg Res 70:41– 45 8. Muzzin P, Eisensmith RC, Copeland KC, Woo SL 1997 Hepatic insulin gene expression as treatment for type 1 diabetes mellitus in rats. Mol Endocrinol 11:833– 837 9. Stewart C, Taylor NA, Green IC, Docherty K, Bailey CJ 1994 Insulin-releasing pituitary cells as a model for somatic cell gene therapy in diabetes mellitus. J Endocrinol 142:339 –343 10. McClenaghan NH, Flatt PR 1999 Engineering cultured insulin-secreting pancreatic B-cell lines. J Mol Med 77:235–243 11. Yamasaki K, Sasaki T, Nemoto M, Eto Y, Tajima N 1999 Differentiationinduced insulin secretion from nonendocrine cells with engineered human proinsulin cDNA. Biochem Biophys Res Commun 265:361–365 12. Rivera VM, Wang X, Wardwell S, Courage NL, Volchuk A, Keenan T, Holt DA, Gilman M, Orci L, Cerasoli Jr F, Rothman JE, Clackson T 2000 Regulation of protein secretion through controlled aggregation in the endoplasmic reticulum. Science 287:826 – 830 13. Stewart C, Taylor NA, Docherty K, Bailey CJ 1993 Insulin delivery by somatic cell gene therapy. J Mol Endocrinol 11:335–341 14. Yang YW, Hsieh YC 2001 Regulated secretion of proinsulin/insulin from human hepatoma cells transduced by recombinant adeno-associated virus. Biotechnol Appl Biochem 33:133–140 15. Shah R, Jindal RM 1999 Stable transfection of rat preproinsulin II gene into rat hematopoietic stem cells via recombinant adeno-associated virus. Life Sci 65:2041–2047 16. Short DK, Okada S, Yamauchi K, Pessin JE 1998 Adenovirus-mediated transfer of a modified human proinsulin gene reverses hyperglycemia in diabetic mice. Am J Physiol 275:E748 –E756 17. Kolodka TM, Finegold M, Moss L, Woo SL 1995 Gene therapy for diabetes mellitus in rats by hepatic expression of insulin. Proc Natl Acad Sci USA 92:3293–3297 18. Meoni C, Bertuzzi F, Pontiroli AE, Falqui L, Monaco L, Soria M, Arcelloni C, Paroni R, Foglieni C, Polastri L, Galbiati F, Folli F, Davalli AM 2000 Development and characterization of pituitary GH3 cell clones stably transfected with a human proinsulin cDNA. Cell Transplant 9:829 – 840 19. Davalli AM, Galbiati F, Bertuzzi F, Polastri L, Pontiroli AE, Perego L, Freschi M, Pozza G, Folli F, Meoni C 2000 Insulin-secreting pituitary GH3 cells: a potential ␤-cell surrogate for diabetes cell therapy. Cell Transplant 9:841– 851 20. Arcelloni C, Falqui L, Martinenghi S, Stabilini A, Pontiroli AE, Paroni R 2000 Processing and release of human proinsulin-cleavage products into culture media by different engineered non-endocrine cells: a specific assessment by capillary electrophoresis. J Endocrinol 166:437– 445 21. Sato Y, Nio Y, Omori H, Inoue Y, Hirahara N, Sasaki S, Tamura K 1999 Human proinsulin gene-transfected Chinese hamster ovary cells secrete immunoreactive insulin. In Vivo 13:535–540

Zhu et al. • Proinsulin Aggregation in Non-␤-Cells 22. Simonson GD, Groskreutz DJ, Gorman CM, MacDonald MJ 1996 Synthesis and processing of genetically modified human proinsulin by rat myoblast primary cultures. Hum Gene Ther 7:71–78 23. Abai AM, Hobart PM, Barnhart KM 1999 Insulin delivery with plasmid DNA. Hum Gene Ther 10:2637–2649 24. Halban PA, Kahn SE, Lernmark A, Rhodes CJ 2001 Gene and cell-replacement therapy in the treatment of type 1 diabetes: how high must the standards be set? Diabetes 50:2181–2191 25. Chen X, Patil JG, Lok SH, Kon OL 2002 Human liver-derived cells stably modified for regulated proinsulin secretion function as bioimplants in vivo. J Gene Med 4:447– 458 26. Carroll RJ, Hammer RE, Chan SJ, Swift HH, Rubenstein AH, Steiner DF 1988 A mutant human proinsulin is secreted from islets of Langerhans in increased amounts via an unregulated pathway. Proc Natl Acad Sci USA 85:8943– 8947 27. Cullen BR 1987 Use of eukaryotic expression technology in the functional analysis of cloned genes. Methods Enzymol 152:684 –704 28. Ho SN, Hunt HD, Horton RM, Pullen JK, Pease LR 1989 Site-directed mutagenesis by overlap extension using the polymerase chain reaction. Gene 77:51–59 29. Lee MS, Zhu YL, Chang JE, Dannies PS 2001 Acquisition of Lubrol insolubility, a common step for growth hormone and prolactin in the secretory pathway of neuroendocrine cells. J Biol Chem 276:715–721 30. Janjic D, Maechler P, Sekine N, Bartley C, Annen AS, Wollheim CB 1999 Free radical modulation of insulin release in INS-1 cells exposed to alloxan. Biochem Pharmacol 57:639 – 648 31. Maechler P, Wollheim CB 1999 Mitochondrial glutamate acts as a messenger in glucose-induced insulin exocytosis. Nature 402:685– 689 32. Reaves BJ, Van Itallie CM, Moore HH, Dannies PS 1990 Prolactin and insulin are targeted to the regulated pathway in GH4C1 cells, but their storage is differentially regulated. Mol Endocrinol 4:1017–1026 33. Schagger H, von Jagow G 1987 Tricine-sodium dodecyl sulfate-polyacrylamide gel electrophoresis for the separation of proteins in the range from 1 to 100 kDa. Anal Biochem 166:368 –379 34. Scammell JG, Burrage TG, Dannies PS 1986 Hormonal induction of secretory granules in a pituitary tumor cell line. Endocrinology 119:1543–1548 35. Delbeke D, Kojima I, Dannies PS 1985 Comparison of patterns of prolactin release in GH4C1 cells and primary pituitary cultures. Mol Cell Endocrinol 43:15–22 36. Seidah NG, Chretien M, Day R 1994 The family of subtilisin/kexin like pro-protein and pro-hormone convertases: divergent or shared functions. Biochimie (Paris) 76:197–209 37. Kuliawat R, Prabakaran D, Arvan P 2000 Proinsulin endoproteolysis confers enhanced targeting of processed insulin to the regulated secretory pathway. Mol Biol Cell 11:1959 –1972 38. Moore HP, Gumbiner B, Kelly RB 1983 A subclass of proteins and sulfated macromolecules secreted by AtT-20 (mouse pituitary tumor) cells is sorted with adrenocorticotropin into dense secretory granules. J Cell Biol 97:810 – 817 39. Mains RE, Glembotski CC, Eipper BA 1984 Peptide ␣-amidation activity in mouse anterior pituitary AtT-20 cell granules: properties and secretion. Endocrinology 114:1522–1530 40. Schimke RT, Doyle D 1970 Control of enzyme levels in animal tissues. Annu Rev Biochem 39:929 –976 41. Mortimore GE, Woodside KH, Henry JE 1972 Compartmentation of free valine and its relation to protein turnover in perfused rat liver. J Biol Chem 247:2776 –2784 42. Righetti P, Little EP, Wolf G 1971 Reutilization of amino acids in protein synthesis in HeLa cells. J Biol Chem 246:5724 –5732 43. Sankoorikal BJ, Zhu YL, Hodsdon ME, Lolis E, Dannies PS 2002 Aggregation of human wild-type and H27A-prolactin in cells and in solution: roles of Zn2⫹, Cu2⫹, and pH. Endocrinology 143:1302–1309 44. Schwartz GP, Burke GT, Katsoyannis PG 1987 A superactive insulin: [B10aspartic acid]insulin(human). Proc Natl Acad Sci USA 84:6408 – 6411 45. Steiner DF, Tager HS, Nanjo K, Chan SJ, Rubenstein AH 1995 Familial syndrome of hyperproinsulinemia and hyperinsulinemia with mild diabetes. In: Scriver CR, Beaudet AL, Sly WS, Valle D, eds. The metabolic and molecular bases of inherited disease. Vol 1. New York: McGraw-Hill, Inc.; 897–904 46. Wang J, Takeuchi T, Tanaka S, Kubo SK, Kayo T, Lu D, Takata K, Koizumi A, Izumi T 1999 A mutation in the insulin 2 gene induces diabetes with severe pancreatic ␤-cell dysfunction in the Mody mouse. J Clin Invest 103:27–37 47. Zhu YL, Conway-Campbell BL, Waters MJ, Dannies PS 2002 Prolonged retention after aggregation into secretory granules of R183H-growth hormone, a mutant that causes autosomal growth hormone deficiency type II. Endocrinology 1443:4243– 4248 48. Xu D, Joglekar AP, Williams AL, Hay JC 2000 Subunit structure of a mammalian ER/Golgi SNARE complex. J Biol Chem 275:39631–39639 49. Reaves BJ, Roquemore EP, Luzio JP, Banting G 1998 TGN38 cycles via the basolateral membrane of polarized Caco-2 cells. Mol Membr Biol 15:133–139 50. Reaves BJ, Banting G, Luzio JP 1998 Lumenal and transmembrane domains play a role in sorting type I membrane proteins on endocytic pathways. Mol Biol Cell 9:1107–1122 51. Kreykenbohm V, Wenzel D, Antonin W, Atlachkine V, von Mollard GF 2002


52.

53. 54.

55. 56.

57.

58.

59.

60.

61.

62.

The SNAREs vti1a and vti1b have distinct localization and SNARE complex partners. Eur J Cell Biol 81:273–280 Mains RE, May V 1988 The role of a low pH intracellular compartment in the processing, storage, and secretion of ACTH and endorphin. J Biol Chem 263:7887–7894 Liu M, Ramos-Castaneda J, Arvan P 2003 Role of the connecting peptide in insulin biosynthesis. J Biol Chem 278:14798 –14805 Kaufmann JE, Irminger JC, Mungall J, Halban PA 1997 Proinsulin conversion in GH3 cells after coexpression of human proinsulin with the endoproteases PC2 and/or PC3. Diabetes 46:978 –982 Kaufmann JE, Irminger JC, Halban PA 1995 Sequence requirements for proinsulin processing at the B-chain/C-peptide junction. Biochem J 310:869 – 874 Quinn D, Orci L, Ravazzola M, Moore HP 1991 Intracellular transport and sorting of mutant human proinsulins that fail to form hexamers. J Cell Biol 113:987–996 Shakur Y, Shennan KI, Taylor NA, Docherty K 1989 A major C-peptide deletion prevents secretion of a mutant human proinsulin from transfected monkey kidney cells. J Mol Endocrinol 3:155–162 Irminger JC, Vollenweider FM, Neerman-Arbez M, Halban PA 1994 Human proinsulin conversion in the regulated and the constitutive pathways of transfected AtT20 cells. J Biol Chem 269:1756 –1762 Vollenweider F, Irminger JC, Gross DJ, Villa-Komaroff L, Halban PA 1992 Processing of proinsulin by transfected hepatoma (FAO) cells. J Biol Chem 267:14629 –14636 Gross DJ, Halban PA, Kahn CR, Weir GC, Villa-Komaroff L 1989 Partial diversion of a mutant proinsulin (B10 aspartic acid) from the regulated to the constitutive secretory pathway in transfected AtT-20 cells. Proc Natl Acad Sci USA 86:4107– 4111 Okamoto T, Toyooka K, Minamikawa T 2001 Identification of a membraneassociated cysteine protease with possible dual roles in the endoplasmic reticulum and protein storage vacuole. J Biol Chem 276:742–751 Urade R, Yasunishi A, Okudo H, Moriyama T, Kito M 1999 Autodegradation of protein disulfide isomerase. Biosci Biotechnol Biochem 63:610 – 613


63. Urade R, Takenaka Y, Kito M 1993 Protein degradation by ERp72 from rat and mouse liver endoplasmic reticulum. J Biol Chem 268:22004 –22009 64. Cardozo C, Wu X, Pan M, Wang H, Fisher EA 2002 The inhibition of microsomal triglyceride transfer protein activity in rat hepatoma cells promotes proteasomal and nonproteasomal degradation of apoprotein b100. Biochemistry 41:10105–10114 65. Brange J, Ribel U, Hansen JF, Dodson G, Hansen MT, Havelund S, Melberg SG, Norris F, Norris K, Snel L, Sorensen AR, Voight HO 1988 Monomeric insulins obtained by protein engineering and their medical implications. Nature 333:679 – 682 66. Nielsen L, Frokjaer S, Brange J, Uversky VN, Fink AL 2001 Probing the mechanism of insulin fibril formation with insulin mutants. Biochemistry 40:8397– 8409 67. Brems DN, Brown PL, Bryant C, Chance RE, Green LK, Long HB, Miller AA, Millican R, Shields JE, Frank BH 1992 Improved insulin stability through amino acid substitution. Protein Eng 5:519 –525 68. Hutton JC, Penn EJ, Peshavaria M 1983 Low-molecular-weight constituents of isolated insulin-secretory granules. Bivalent cations, adenine nucleotides and inorganic phosphate. Biochem J 210:297–305 69. Clifford KS, MacDonald MJ 2000 Survey of mRNAs encoding zinc transporters and other metal complexing proteins in pancreatic islets of rats from birth to adulthood: similar patterns in the Sprague-Dawley and Wistar BB strains. Diabetes Res Clin Pract 49:77– 85 70. Kambe T, Narita H, Yamaguchi-Iwai Y, Hirose J, Amano T, Sugiura N, Sasaki R, Mori K, Iwanaga T, Nagao M 2002 Cloning and characterization of a novel mammalian zinc transporter, zinc transporter 5, abundantly expressed in pancreatic ␤-cells. J Biol Chem 277:19049 –19055 71. Leroux L, Desbois P, Lamotte L, Duvillie B, Cordonnier N, Jackerott M, Jami J, Bucchini D, Joshi RL 2001 Compensatory responses in mice carrying a null mutation for Ins1 or Ins2. Diabetes 50(Suppl 1):S150 –S153 72. Arias AE, Velez-Granell CS, Mayer G, Bendayan M 2000 Colocalization of chaperone cpn60, proinsulin and convertase PC1 within immature secretory granules of insulin-secreting cells suggests a role for cpn60 in insulin processing. J Cell Sci 113:2075–2083

Endocrinology is published monthly by The Endocrine Society (http://www.endo-society.org), the foremost professional society serving the endocrine community.