Misfolded growth hormone causes fragmentation of the Golgi apparatus and disrupts endoplasmic reticulum-to-Golgi traffic.

In some individuals with autosomal dominant isolated growth hormone deficiency, one copy of growth hormone lacks amino acids 32-71 and is severely misfolded. We transfected COS7 cells with either wild-type human growth hormone or Delta 32-71 growth hormone and investigated subcellular localization of growth hormone and other proteins. Delta 32-71 growth hormone was retained in the endoplasmic reticulum, whereas wild-type hormone accumulated in the Golgi apparatus. When cells transfected with wild-type or Delta 32-71 growth hormone were dually stained for growth hormone and the Golgi markers beta-COP, membrin or 58K, wild-type growth hormone was colocalized with the Golgi markers, but beta-COP, membrin and 58K immunoreactivity was highly dispersed or undetectable in cells expressing Delta 32-71 growth hormone. Examination of alpha-tubulin immunostaining showed that the cytoplasmic microtubular arrangement was normal in cells expressing wild-type growth hormone, but microtubule-organizing centers were absent in nearly all cells expressing Delta 32-71 growth hormone. To determine whether Delta 32-71 growth hormone would alter trafficking of a plasma membrane protein, we cotransfected the cells with the thyrotropin-releasing hormone (TRH) receptor and either wild-type or Delta 32-71 growth hormone. Cells expressing Delta 32-71 growth hormone, unlike those expressing wild-type growth hormone, failed to show normal TRH receptor localization or binding. Expression of Delta 32-71 growth hormone also disrupted the trafficking of two secretory proteins, prolactin and secreted alkaline phosphatase. Delta 32-71 growth hormone only weakly elicited the unfolded protein response as indicated by induction of BiP mRNA. Pharmacological induction of the unfolded protein response partially prevented deletion mutant-induced Golgi fragmentation and partially restored normal TRH receptor trafficking. The ability of some misfolded proteins to block endoplasmic reticulum-to-Golgi traffic may explain their toxic effects on host cells and suggests possible strategies for therapeutic interventions.

A serum prolactin-binding protein: implications for growth hormone.

Many membrane receptors have truncated soluble forms that circulate in the blood, and a protein in serum with characteristics of the extracellular domain of the human prolactin receptor has recently been identified. Because the extracellular domain of the prolactin receptor binds human growth hormone, does the prolactin-binding protein in serum affect the amount of bound circulating growth hormone? A more general question is whether this serum protein has a minor or a major effect on the biological activities of growth hormone and prolactin.

Concentrating hormones into secretory granules: layers of control.

Soluble protein hormones are concentrated and stored in secretory granules The cisternal maturation model for transport of proteins through the Golgi complex allows a major role for formation of reversible aggregates as a means of both concentrating and sorting hormones, since soluble proteins will be removed in small vesicles, leaving behind the aggregated hormones. The storage of secretory granule proteins, however, is more selective than would be expected if passive aggregation were the only process involved. Aggregation of hormones in the secretory pathway may not be completely passive, but may be controlled by the cells. In addition to aggregation, other layers of sorting must exist, because there is selective retention of proteins after aggregation or packaging into granules.

Acquisition of Lubrol insolubility, a common step for growth hormone and prolactin in the secretory pathway of neuroendocrine cells.

Rat prolactin in the dense cores of secretory granules of the pituitary gland is a Lubrol-insoluble aggregate. In GH(4)C(1) cells, newly synthesized rat prolactin and growth hormone were soluble, but after 30 min about 40% converted to a Lubrol-insoluble form. Transport from the endoplasmic reticulum is necessary for conversion to Lubrol insolubility, since incubating cells with brefeldin A or at 15 degrees C reduced formation of insoluble rat (35)S-prolactin. Formation of Lubrol-insoluble aggregates has protein and cell specificity; newly synthesized human growth hormone expressed in AtT20 cells underwent a 40% conversion to Lubrol insolubility with time, but albumin did not, and human growth hormone expressed in COS cells underwent less than 10% conversion to Lubrol insolubility. del32-46 growth hormone, a naturally occurring form of growth hormone, and P89L growth hormone underwent conversion, although they were secreted more slowly, indicating that there is some tolerance in structural requirements for aggregation. An intracellular compartment with an acidic pH is not necessary for conversion to Lubrol insolubility, because incubation with chloroquine or bafilomycin slowed, but did not prevent, the conversion. GH(4)C(1) cells treated with estradiol, insulin, and epidermal growth factor accumulate more secretory granules and store more prolactin, but not more growth hormone, than untreated cells; Lubrol-insoluble aggregates of prolactin and growth hormone formed to the same extent in hormone-treated or untreated GH(4)C(1) cells, but prolactin was retained longer in hormone-treated cells. These findings indicate that aggregation alone is not sufficient to cause retention of secretory granule proteins, and there is an additional selective process.

Accumulation of synaptosomal-associated protein of 25 kDa (SNAP-25) and other proteins associated with the secretory pathway in GH4C1 cells upon treatment with estradiol, insulin, and epidermal growth factor.

Treatment of rat pituitary GH4C1 cells with estradiol, insulin, and epidermal growth factor induces secretory granule accumulation, PRL storage, and stabilization of ICA512, a membrane protein associated with secretory granules. In these investigations we found that the same treatment induced accumulation over 2-fold of other proteins in the secretory pathway, including synaptosomal-associated protein of 25 kDa (SNAP-25), synaptotagmin III, synaptobrevin, synaptophysin, and cyclophilin B, and did not affect accumulation of others, including synaptotagmin I, calnexin, and glucose-regulated protein 94. The induction of proteins was not a coordinate event, because epidermal growth factor alone maximally stimulated SNAP-25 accumulation, but not that of synaptotagmin III. Induction of SNAP-25 accumulation occurred without an increase in its synthesis, and induction of cyclophilin B occurred without an increase in its messenger RNA accumulation, suggesting that accumulation may be caused by stabilization of the proteins. SNAP-25 immunofluorescence was located in the cytoplasm and on the plasma membrane and sometimes was heavily concentrated in protrusions from the cell surface, especially in hormone-treated cells. Frequenin immunofluorescence was also sometimes concentrated in intense patches, but did not colocalize with SNAP-25. Growth hormone and prolactin immunofluorescence was not found in the protrusions and sometimes did not colocalize with each other when they were present in the same cell. Hormone treatment of GH4C1 cells therefore induces accumulation of specific proteins in all parts of the secretory pathway and causes morphological changes in addition to accumulating secretory granules.

Autosomal dominant growth hormone (GH) deficiency type II: the Del32-71-GH deletion mutant suppresses secretion of wild-type GH.

Familial isolated GH deficiency type II is an autosomal dominant form of short stature, associated in some families with mutations that result in missplicing to produce del32-71-GH, a protein that cannot fold normally. The mechanism by which this mutant suppresses the secretion of wild-type GH encoded by the normal allele is not known. Coexpression of del32-71-GH with wild-type human GH in transient transfections of the neuroendocrine cell lines GH4C1 and AtT20 suppressed accumulation of wild-type GH. The suppression of wild-type GH accumulation by del32-71-GH was a posttranslational effect on wild-type GH caused by decreased stability, rather than decreased synthesis, of wild-type GH. Coexpression of del32-71-GH with human PRL did not suppress accumulation of PRL, indicating that there was not a general suppression of secretory pathway function. Accumulation of del32-71-GH protein was not necessary for the suppression of wild-type GH, because del32-71-GH did not accumulate in the neuroendocrine cell lines in which suppression of accumulation of wild-type GH was observed. Del32-71-GH did accumulate in transfected COS and CHO cells, but did not suppress the accumulation of wild-type GH in these cells. These studies suggest that del32-71-GH may cause GH deficiency in somatotropes of heterozygotes expressing both wild-type and del32-71-GH by decreasing the intracellular stability of wild-type GH.

Protein folding and deficiencies caused by dominant-negative mutants of hormones.

Protein folding and transport in the secretory pathway of cells is a controlled process, facilitated by chaperones. Proteins that do not fold well elicit several different programmed responses from the cells. A comparison of mutants of growth hormone that result in growth hormone deficiency suggests that cells do not respond in the same way to all growth hormone mutants that cannot fold, because some mutants are dominant and some are recessive. Causes for autosomal dominant hormone deficiencies include accumulation of toxic or dysfunctional forms, competition for chaperones important for folding or transport, induction of protein degradation in the endoplasmic reticulum, or long-term responses of the cells to synthesis of proteins that do not fold that decrease hormone synthesis or cell viability.

Protein hormone storage in secretory granules: mechanisms for concentration and sorting.

Recent findings in cell biology have demonstrated there are several kinds of active sorting from the trans-Golgi network in all cells. The presence of several sorting pathways, using more than one sorting signal, in neuroendocrine cells means that mutations that direct a hormone to a constitutive pathway instead of a regulated one may not simply be interpreted as a signal for sorting to a regulated pathway. The use of three-dimensional electron microscopy of lactotrophs and the possibility that the trans-Golgi network is consumed during sorting has suggested a major role for hormone aggregation, not only as a sorting mechanism, but also as a mechanism for granule formation, in that other transport vesicles may bud from the trans-Golgi network, leaving the aggregated protein as the dense core granule. If aggregation plays such a role, it is unclear how it works in cases where the prohormone must be processed one or more times; does a porous aggregate or colloid form? Obtaining information about the kinds of aggregates that occur in cells is difficult, because at this time there is not a definitive way of knowing whether an aggregate that occurs in solution also occurs in cells. Although secretory granule proteins tend to aggregate relatively easily in solution, the separate treatment of different secretory granule proteins in the same cell make it unlikely that aggregation is a purely passive process, but suggests that the process of aggregation of each hormone is actively controlled in cells. Even if the ability to aggregate accounts for most of the sorting of cargo-secretory granule proteins into granules, other sorting must still occur to get correct membrane proteins necessary for transport and exocytosis into secretory granule membranes. Possible recognition sites for these secretory granule membrane proteins include the cargo itself in an aggregated form, membrane lipids in some unrecognized way, or the proteins and factors that specifically control aggregation of the cargo.

Stabilization of the receptor protein tyrosine phosphatase-like protein ICA512 in GH4C1 cells upon treatment with estradiol, insulin, and epidermal growth factor.

Treatment with 1 nM estradiol, 300 nM insulin, and 5 nM epidermal growth factor induces secretory granule accumulation and prolactin storage in GH4C1 rat pituitary tumor cells. The same triple treatment induced more than 6-fold accumulation of both the precursor (100 kDa, pro-ICA512) and the mature forms (60-70 kDa, ICA512 transmembrane fragment) of ICA512, a receptor protein tyrosine phosphatase-like protein that is preferentially localized in secretory granule membranes. Accumulation of ICA512 resembles that of prolactin storage, for the combination of all three, estradiol, insulin, and epidermal growth factor, gave the greatest induction, which was maximal at 4 days. This effect was specific, as the levels of the small GTP-binding protein Rab3, which is also associated with secretory granule membranes, were unaffected by the triple hormone/growth factor treatment. Increased transcription and translation of ICA512 could only partially account for its 6-fold accumulation, as ICA512 messenger RNA and ICA512 synthesis levels were 1.8 +/- 0.35- and 1.6 +/- 0.17-fold in triple treated GH4C1 cells compared with those in untreated cells, respectively. Pulse-chase procedures showed that pro-ICA512 was more stable in treated cells. These results indicate that the enlargement of the secretory granule compartment results in the stabilization of ICA512 and raise the possibility that trafficking of secretory granules may affect ICA512's function and vice versa.

Inefficient secretion of human H27A-prolactin, a mutant that does not bind Zn2+.

Human PRL binds Zn2+, but the function of the binding is not known. We investigated the effect on PRL production in pituitary cells by obtaining clones of GH4C1 cells stably transfected with human H27A-PRL, a mutant that does not bind Zn2+. Unexpectedly, clones transfected with the mutant human PRL made little rat PRL. Untransfected GH4C1 cells made between 0.5 to 10 microg rat PRL/10(5) cells in 24 h. Clones transfected with vector alone (four of four), wild type human PRL (six of six), or with human K69A-PRL (two of two) made amounts of rat PRL in the same range. Clones transfected with human H27A-PRL (five of five) made 0.003-0.1 microg rat PRL/10(5) cells in 24 h, and the production of rat PRL mRNA was reduced. Human H27A-PRL was not efficiently secreted; 20-40% newly synthesized H27A-PRL was degraded by 60 min, and there was usually a delay in release of newly synthesized H27A-PRL. Reduction of rat PRL production is not mediated through the PRL receptor, because no sequences for the receptor in GH4C1 cells were detected by RT-PCR. Proteins involved in folding, such as BiP, were not specifically elevated in the H27A-PRL clones. In transient transfections, in which cells have not undergone selection, we found no evidence for disulfide-bonded aggregates of the mutant protein. The results indicate that Zn2+ binding stabilizes PRL in the secretory pathway; the instablility of the mutant protein may trigger effects that suppress rat PRL production directly or that indirectly result in selection of clones with low rat PRL production.

Properties of human prolactin (PRL) and H27A-PRL, a mutant that does not bind Zn++.

The ability of protein hormones to self-associate is likely to play an important role in concentrating hormones into secretory granules; therefore, the aggregation properties of human PRL and H27A mutant were investigated. Human PRL bound (65)Zn++; the Scatchard analysis was convex up, and limited by the solubility of PRL, but at least 0.7 mole Zn++ bound per mole of PRL. Binding of (65)Zn++ to H27A-PRL was greatly reduced. The biological activity in an Nb2 cell assay and the circular dichroism spectrum of wild type and H27A-PRL were similar, indicating the H27A mutant folded correctly, and the binding of Zn++ to the high affinity site is not essential for biological activity. Dynamic light scattering measurements indicated 10 and 20 mu M Zn++ caused some aggregation of both wild type and H27A-PRL. Sedimentation equilibrium analysis indicated that PRL is primarily a monomer in the absence of Zn++ and that there is increasing self-association in the presence of 5 and 10 mu M Zn++. The mutant H27A exhibited a greater tendency to aggregate without changing detectably the mode of association. Although human PRL binds Zn++ as human GH does, it differs in that the ability to bind Zn++ and to self-associate were decoupled in PRL. Human PRL must have two types of interactions with Zn++; one is binding to a site involving histidine 27, and the other is weaker interactions that induce self-association of PRL.

KN-62, a calcium/calmodulin-dependent protein kinase II inhibitor, inhibits high potassium-stimulated prolactin secretion and intracellular calcium increases in anterior pituitary cells.

In isolated rat anterior pituitary cells, KN-62 (10 microM), an isoquinoline sulfonamide inhibitor of calcium/calmodulin-dependent protein kinase II, inhibited high KCl(50 milliM)-stimulated prolactin secretion almost completely, with an IC50 of 95 nM KN-62 inhibited TRH-induced prolactin secretion less effectively. KN-04, a compound that is over 100-fold less active in inhibiting purified calcium/calmodulin-dependent protein kinase II, also inhibited high KCl-stimulated prolactin secretion with an IC50 of 500 nM. KN-62 and KN-04 (10 microM) both inhibited high KCl-stimulated increases in intracellular Ca2+ concentrations. We conclude that KN-62 and KN-04 inhibit activation of voltage-dependent calcium channels in anterior pituitary cells either directly or indirectly.

Biological activity and immunological reactivity of human prolactin mutants.

We examined the biological activity and immunological reactivity of four mutants of human PRL. Two were mutants that changed the ability of human PRL to inhibit rat PRL storage when transfected into a rat pituitary cell line:mutations S34A and N31T. Two mutations were in regions of PRL that are highly conserved. One, des(3-11)-PRL, removed the N-terminal cystine loop that most PRLs, except those from certain fish, have, and no GHs have. The other, S90A, mutated a serine that is present in all PRLs but those from some fish and in all GHs. The immunological properties of des(3-11)-PRL were reduced 10-fold compared to those of wildtype human PRL in a RIA using NIH antihuman PRL-3, AFP C11580; the others were similar to those of wild-type PRL. The biological activity of des(3-11)-PRL was the most affected; activity was reduced about 8-fold compared to that of wild-type PRL in the Nb2 cell assay. The activities of the others were similar to that of the wild type. Serine 90 may be partially buried by loops connecting the alpha-helixes. The mutation of serine 90 did not affect the stability of the molecule in vitro, determined by comparing the red shift in tryptophan fluorescence that occurs with increasing concentrations of urea in S90A and wild-type PRL. The activity of S34A and N31T mutations indicates there is no correlation between biological activity and ability to affect storage. The N-terminal cystine loop may be conserved because it is needed for biological activity, but the conservation of serine 90 in GH and PRL must be determined by other properties, such as spacial requirements.

Lack of correlation of distribution of prolactin (PRL) charge isoforms with induction of PRL storage.

GH4C1 cells, rat pituitary tumor cells, produce PRL, but store little relative to normal lactotrophs; treating cells with estradiol, insulin, and epidermal growth factor increases both PRL storage and accumulation of dense core granules. Normal lactotrophs contain several differently charged forms of PRL. We investigated whether inducing PRL storage in GH4C1 cells altered the production of these forms. Four forms of PRL that differed by charge were found by immunoblots of two-dimensional gels of extracts of female rat pituitary glands and of secretory granules isolated from the glands. Four forms were also secreted by GH4C1 cells. The relative abundance of the four forms in the medium of GH4C1 cells, determined by [35S]amino acid incorporation for 24h, was 1.2 +/- 0.58%, 91.3 +/- 1.09%, 6.3 +/- 0.72%, and 1.2 +/- 0.39% of total PRL (mean +/- SEM), from the most basic to the most acidic, respectively. Treatment with 1 nM estradiol, 300 nM insulin, and 10 nM epidermal growth factor did not significantly change the relative abundance of the forms. All four forms also were found in GH4C1 cells after 2 h of incubation with [35S]amino acids, although no incorporation of 32PO4 was detectable over the same incubation time. We conclude that GH4C1 cells produce four forms of PRL that differ by charge, as normal lactotrophs do. The increase in storage of PRL caused by insulin, estrogen, and epidermal growth factor does not result in or is caused by increased secretion of a specific form.

Inhibition of rat prolactin (PRL) storage by coexpression of human PRL.

GH4C1 cells increase storage of rat PRL and accumulation of dense core secretory granules when they are treated with 300 nM insulin, 1 nM estradiol, and 10 nM epidermal growth factor. In the experiments reported here, intracellular rat PRL increased from 4% of the total amount produced in 24 h in control cultures to 15% in treated cultures. When GH4C1 cells were transfected with DNA sequences so that they coexpressed human PRL, the storage of rat PRL was no longer induced by hormone treatment; transfected clones contained less than 5% of the total produced in 24 h in both control and treated cultures. The sum of rat and human PRL produced by these clones was not more than rat PRL produced by the untransfected cells, so the storage capacity of the cells was not exceeded. The transfected clones made between 1 to 40 times more rat than human PRL. Release of human and rat PRL was stimulated by depolarizing the cells, indicating both still were in a regulated pathway, even though storage could not be induced. Mutations of human PRL with threonine substituted for asn31 or alanine substituted for ser34 did not block induction of rat PRL storage when coexpressed in GH4C1 clones. We conclude the ability to increase rat PRL storage is a process with marked specificity because it is inhibited by relatively low amounts of human PRL and inhibition requires asn31 and ser34 in human PRL. Such specificity is consistent with a receptor-mediated mechanism that concentrates PRL into dense cores of secretory granules.

Calcium/calmodulin-dependent protein kinase-II activation in rat pituitary cells in the presence of thyrotropin-releasing hormone and dopamine.

PRL release from rat lactotrophs in response to TRH is Ca2+ dependent. TRH-induced PRL release is inhibited either after repeated pulses of TRH or in the presence of dopamine. TRH, however, generates increases in intracellular Ca2+ concentrations ([Ca2+]i) in both conditions. Calcium/calmodulin-dependent protein kinase-II (CaM kinase-II) is a ubiquitous enzyme implicated in secretion. To determine whether down-regulation of CaM kinase-II activity caused the lack of responsiveness to increases in [Ca2+]i, we measured the generation of calcium/calmodulin-independent kinase activity. Anterior pituitary cells contain a 50-kilodalton form of CaM kinase-II, determined by immunoblot, and the enzyme is in lactotrophs, determined by immunocytochemistry. TRH rapidly and transiently increased calcium/calmodulin-independent kinase activity; the increase was maximal by 15 sec and returned to basal by 2 min. When TRH pulses (1 microM, 15 sec) were applied every 10 min, each pulse caused an increase in calcium/calmodulin-independent kinase activity of similar magnitude, and the activity returned to basal values between pulses. Pretreatment of cells with dopamine (1 microM; 30 min) inhibited PRL release, but did not prevent the increase in calcium/calmodulin-independent kinase activity. These results indicate that TRH still activates CaM kinase-II when PRL release is inhibited. Dopamine and repeated pulses of TRH must inhibit PRL release at a site after the TRH-induced increase in [Ca2+]i and at a site other than CaM kinase-II.

Regulation of parathyroid hormone-related peptide gene expression by estrogen in GH4C1 rat pituitary cells has the pattern of a primary response gene.

The parathyroid hormone-related peptide (PTHrP) gene has been reported to be subject to a wide variety of physiological and pharmacological controls. Two distinct patterns of PTHrP mRNA response have been recognized, one characterized by a prolonged or plateau response lasting many hours to days and the second characterized by rapid induction-deinduction kinetics and lasting 1 to several hours. The kinetics of the second pattern are similar to those displayed by primary response genes like nuclear protooncogenes, cytokines, and growth factors. In GH4C1 rat pituitary cells, 17 beta-estradiol induced a rapid and transient increase in PTHrP mRNA expression, with a peak response at 1-2 h. This response appeared to be due to a rapid and transient burst in gene transcription, which by runoff analysis was maximal at 20-40 min and declined thereafter. PTHrP mRNA half-life was 30 min in these cells and was unaltered by estradiol. Cycloheximide did not block the 17 beta-estradiol-induced response but rather prolonged it, and runoff analysis revealed that this effect was due to a prolongation or persistence of PTHrP gene transcription. These findings suggest that the transient nature of the native response reflects the effects of an estrogen-inducible repressor. All of these features are characteristic of a prototypical primary response gene.

Rapid stimulation of rhodamine 123 efflux from multidrug-resistant KB cells by progesterone.

Rhodamine 123 is a mitochondrial dye that is retained for prolonged periods by carcinoma cells. While investigating causes of retention of this dye, we found that 10 microM progesterone caused a rapid stimulation of efflux of rhodamine 123 within 15 min from KB V20C cells, which overexpress the multidrug resistance pump. Progesterone did not stimulate efflux from KB cells that do not overexpress the pump, and verapamil blocked rhodamine 123 efflux in the presence or absence of progesterone, indicating that rhodamine 123 is removed from KB V20C cells by the multidrug resistance pump. Progesterone, however, is unlikely to stimulate rhodamine 123 efflux by simply increasing pump activity for two reasons: (1) progesterone inhibited the efflux of daunomycin from KB V20C cells, so it did not stimulate efflux of all drugs, and (2) progesterone inhibited efflux of rhodamine 123 from L1210/VMDRC cells and had little effect on Adr MCF7 cells; both overexpress the multidrug resistance pump. In the experiments with KB V20C cells, progesterone was the most active steroid tested. At 10 microM, progesterone caused a 70-fold stimulation, desoxycorticosterone, testosterone, promegestone and estradiol about 20-fold, and others had little or no effect. Progesterone may act by a non-genomic mechanism to decrease intracellular binding of rhodamine 123, making the dye accessible to the multidrug resistance pump.

Estradiol induction of rhodamine 123 efflux and the multidrug resistance pump in rat pituitary tumor cells.

Rhodamine 123 is a fluorescent dye that localizes in mitochondria, is a substrate for the multidrug resistance pump, and is retained for long periods of time by carcinoma cells. 17 beta-Estradiol causes GH4C1 cells (rat pituitary tumor cells) to lose rhodamine 123 fluorescence faster than untreated cells. We found that estradiol induces accumulation of the mRNA for the multidrug resistance pump 3-5-fold, with maximum induction occurring within 1 day at 10(-9) M estradiol. Immunoblot analysis demonstrated that estradiol induces a protein of 150 kDa that reacts with an antibody to P-glycoprotein, the multidrug resistance pump. The reduced retention of rhodamine 123 caused by estradiol is prevented by verapamil and cyclosporin, inhibitors of the pump. A clone resistant to the effects of estradiol on rhodamine 123 has greatly reduced levels of mRNA for the pump. The effect of estradiol is more marked on rhodamine 123 retention than it is on that of rhodamine 110 or tetramethylrhodamine methyl ester. We conclude that estradiol enhances rhodamine 123 efflux by inducing the multidrug resistance gene. The specificity for rhodamine 123, compared with other analogs, may be caused by differences in accessibility to the pump.