ABSTRACT
Control of enzymatic function by peptide hormones can occur at a number of different levels and can involve diverse pathways that regulate cleavage, intracellular trafficking, and protein degradation. Gastrin is a peptide hormone that binds to the cholecystokinin B-gastrin receptor and regulates the activity ofl-histidine decarboxylase (HDC), the enzyme that produces histamine. Here we show that gastrin can increase the steady-state levels of at least six HDC isoforms without affecting HDC mRNA levels. Pulse-chase experiments indicated that HDC isoforms are rapidly degraded and that gastrin-dependent increases are due to enhanced isoform stability. Deletion analysis identified two PEST domains (PEST1 and PEST2) and an intracellular targeting domain (ER2) which regulate HDC protein expression levels. Experiments with PEST domain fusion proteins demonstrated that PEST1 and PEST2 are strong and portable degradation-promoting elements which are positively regulated by both gastrin stimulation and proteasome inhibition. A chimeric protein containing the PEST domain of ornithine decarboxylase was similarly affected, indicating that gastrin can regulate the stability of other PEST domain-containing proteins and does so independently of antizyme/antizyme inhibitor regulation. At the same time, endoplasmic reticulum localization of a fluorescent chimera containing the ER2 domain of HDC was unaltered by gastrin stimulation. We conclude that gastrin stabilization of HDC isoforms is dependent upon two transferable and sequentially unrelated PEST domains that regulate degradation. These experiments revealed a novel regulatory mechanism by which a peptide hormone such as gastrin can disrupt the degradation function of multiple PEST-domain-containing proteins.
There is increasing consensus that protein degradation can play a key role in the control of protein function, with parameters such as covalent modification (36), intracellular localization (50), and regulated cleavage (7, 31) all capable of influencing turnover rates. In many cases this degradation is mediated by a multicatalytic proteinase referred to as the proteasome. For the majority of proteasome substrates so far described, degradation is regulated by ubiquitin-protein ligases that recognize specific sequences in proteins targeted for degradation. Subsequent addition of multimeric ubiquitin units represents the first stage in a process that ultimately leads to degradation of the substrate (5, 24, 26). There are a few noted examples where degradation by the proteasome does not involve ubiquitination (22, 25), including that of ornithine decarboxylase (ODC), the first enzyme in the catalytic conversion of ornithine to polyamines (20). A negative feedback loop is generated where polyamines induce translation of the protein antizyme, which binds to sequences in the amino terminus of ODC. Ensuing conformational changes expose a hydrophilic PEST domain in the carboxy-terminal region that targets ODC for degradation directly by the proteasome (20, 28).
Numerous studies have identified factors that promote degradation of specific proteins by the proteasome; however, there are fewer reports of stimulants that inhibit degradation, either specifically or generally (24). In a rare example, it has been shown that internalized insulin can bind a proteasome activator called insulin-degrading enzyme and, in so doing, inhibit general protein metabolism (13, 19). However, there is no evidence to suggest that insulin specifically promotes the transcription or translation of factors that are capable of inhibiting the degradation of a small group of related proteins or that it could do so through activation of recognized signal transduction pathways.
Histamine is a biogenic amine that serves a number of important biological functions, including stimulation of acid secretion within the stomach (39, 45). It is generated through the action of the enzyme l-histidine decarboxylase (HDC; EC 4.1.1.22 ), which is initially translated as a 74-kDa protein but is subsequently cleaved to a number of smaller isoforms (8, 23). Recombinant experiments showed that a 54-kDa carboxy-terminally truncated HDC isoform had much greater activity than the primary translation product, leading to the general belief that such a regulated step in vivo would facilitate the production of an enzymatically active homodimer of 100 to 110 kDa (8, 42, 46, 47). However, other groups were unable to confirm that the 54-kDa isoform had a higher specific activity (48), and the presence of other isoforms has to some extent been ignored.
While the physiological relevance of multiple cleavage steps has not yet been fully deduced, studies on the rat protein sequence suggest the presence of a number of functional domains within the enzyme. This includes a putative intracellular targeting domain, located somewhere near the carboxy-terminal tail (44, 47, 48). Computer analysis has also identified two putative PEST domains at either end of the protein, which have been hypothesized to regulate degradation (15, 29, 43). While one or more of these regions could be cleaved off during posttranslational processing, specific cleavage sites have not yet been identified, and only carboxy-terminal cleavages have previously been considered. Recent studies have shown that the 74-kDa form of HDC can be degraded through the ubiquitin-proteasome pathway (43). While the contribution of this pathway to HDC activation has not yet been considered seriously, it is noteworthy that partial degradation of proteins by this pathway has, in some cases, been shown to regulate activation (30, 31).
Gastrin stimulation of enterochromaffin-like (ECL) cells in the gastric mucosa leads to activation of cholecystokinin B-gastrin receptor signaling shortly after feeding. This is followed by rapid release of histamine, which is then replenished through increased expression of HDC (3, 10, 12, 18, 32, 39). This upregulation of HDC occurs in part through increases in HDC mRNA abundance (3, 10, 11). However, recent studies additionally suggest some degree of posttranscriptional regulation, due both to the rapidity of the response and to the absence of inhibition by actinomycin D (3, 4, 21). These studies also showed that cycloheximide was able to block enzyme activation, suggesting that gastrin-dependent increases represent a primarily translational response (4, 9, 21). Indirect support for these proposals has come from studies which suggest that 5′ untranslated region (5′UTR) sequences can promote HDC translation (23). Indirect support has also come from studies which showed that gastrin is able to stimulate translation of ODC mRNA through 5′UTR sequences and an eIF4E-dependent pathway (33).
In order to investigate the posttranscriptional regulation of HDC by gastrin, we subcloned the rat HDC cDNA into a eukaryotic expression vector and, after transfection into Cos-7 cells expressing the CCK-B/gastrin receptor, examined responses to gastrin stimulation. Our results suggest that transcription-independent increases in enzyme activity are unlikely to involve increases in HDC translation, mediated by the 5′UTR. Instead, we propose that gastrin stabilizes HDC isoforms that are generated through a novel combination of amino- as well as carboxy-terminal cleavage steps. We demonstrate that two PEST domains located within the enzyme mediate this stabilization and confirm that PEST domains from other rapidly turned-over proteins share this type of gastrin regulation.
MATERIALS AND METHODS
Plasmid DNA constructs.Unless otherwise stated, all expression constructs were generated by PCR amplification usingpfu DNA polymerase (Stratagene) on a thermocycler (Perkin Elmer 9700) and using the CMV-HDC18 vector (23) as the template. Amplification parameters for 30 cycles were as follows: 94°C for 30 s, 53°C for 45 s, and 72°C for 2 min per kb of PCR product. PCR products were initially blunt-end cloned into the pCRscript vector (Stratagene). Inserts were subsequently subcloned into appropriate expression constructs.
For generation of the pEP parent vector, NotI linker DNA (New England Biolabs) was used to introduce a NotI digestion site at the BamHI site in pEGFP-N1 (Clontech). This allowed removal of the green fluorescent protein (GFP) cartridge followingNotI digestion and religation. A series of HDC cDNA expression constructs were generated by subcloning into theEcoRI and SalI sites of the pEP vector. All PCRs were done with a common sense primer, 5′-GACCGCGAATTCGCACAGACAATAGTG (bp15, where the 3′-most nucleotide in the primer corresponds to bp 15 of the rat cDNA sequence). The antisense primer used for the pEP-HDC2.4 insert was 5′-ATGTCGACAGATATAGGCAC (bp2349), that for the pEP-HDC2.0 insert was 5′-ATGTCGACCTACACCATGGCCTGC (bp2028), that for the pEP-HDC1.7 insert was 5′-ATGTCGACCTATACCGACAAGTAACTG (bp1764), that for the pEP-HDC16 insert was 5′-ATGTCGACCTACTCATTGACAGACTCCAGG (bp1601), and that for the pEP-HDC15 insert was 5′-ATGTCGACCTACGGCTGAGAAGTGCAG (bp1514). The pEP-GR vector was generated by subcloning CCK-B receptor cDNA from the pEF1a-CCK/B-R plasmid (gift of R. Xavier) into the HindIII and NotI sites of the pEP parent vector.
To generate the inserts for the pBF-PEST1 and pBF-PEST2 vectors, the regions between bp 75 and 436 and between bp 1601 and 1777 of the rat HDC cDNA sequence were PCR amplified. For pBF-PEST1, the sense and antisense primers used were 5′-TCGAATTCTATGATGGAGCCC (bp87) and 5′-ATGTCGACCTACATCTCCAGCTCTGTGC (bp415), respectively. For pBF-PEST2, the sense and antisense primers were 5′-TCGAATTCTGGAGGAGATGACCCAGTACAGG (bp1620) and 5′-ATGTCGACCTATACCGACAAGTAACTG (bp1764), respectively. Amplified regions were subcloned in frame into the EcoRI andSalI sites of the pEBFP-C1 vector (Clontech).
The pGF-ER1, pGF-ER2, and pGF-antiER2 constructs used in intracellular localization experiments were generated by PCR-amplifying the regions between bp 75 and 196 and between bp 1777 and 2043 of the rat HDC cDNA sequence. For pGF-ER1, the sense and antisense primers used were 5′-TCGAATTCTATGATGGAGCCC (bp87) and 5′-ATGTCGACCTACAGGTACCCAGGCTTCAC (bp183), respectively. The antisense primer had an engineered stop codon. For pGF-ER2, the sense and antisense primers were 5′-TCGAATTCTCAGAACAAGAAGAAGACAATGCGG (bp1800) and 5′-ATGTCGACCTACACCATGGCCTGC (bp2028), respectively. For pGF-antiER2, the sense and antisense primers were 5′-TCGAATTCTCAGAACAAGAAGAAGACAATGCGG (bp1800) and 5′-ATGAATTCCTACACCATGGCCTGC (bp2028), respectively. Amplified regions for pGF-ER1 and pGF-ER2 were subcloned in frame into the EcoRI and SalI sites of the pEGFP-C1 vector (Clontech). The amplified region for pGF-antiER2 was subcloned in the antisense direction into the EcoRI site of pEGFP-C1. The pHDC2.0-GF vector was generated by PCR-amplifying the region between bp 1 and 2043 of the rat sequence. The sense and antisense primers used were 5′-GACCGCGAATTCGCACAGACAATAGTG (bp15) and 5′-ATGTCGACACCATGGCCTG (bp2029), respectively. The antisense primer had a mutated stop codon. The amplified region was subcloned in frame into the HindIII and SalI sites of the pEGFP-N1 vector (Clontech).
The pGL-01 and pGL-75 vectors were generated by PCR-amplifying the cytomegalovirus (CMV) promoter along with either 1 or 75 bp from the 5′UTR of HDC, using plasmid pEP-HDC2.4 as the template. PCR conditions were as described above. The sense primer 5′-CGGGGTACCGTATTACCGCCATGCAT, which contains aKpnI restriction site, was used for both amplifications. For the pGL-UTR01 insert, the antisense primer used was 5′-CCACGCGTCGAATTCGAAGCTTGAGCTCG, and for the pGL-UTR75 insert, the antisense primer used was 5′-CCACGCGTCTTTCTTGACTTGGCTTGC. Antisense primers containedMluI restriction sites. Amplified regions were initially cloned into the pCRscript vector and subsequently subcloned into theKpnI and MluI sites of the pGL basic vector (Promega).
The pGF-PEST/ODC and pGF-PEST/DDC constructs were generated by reverse transcription-PCR using total rat liver RNA (Ambion) as the template. Oligo(dT)-primed reverse transcription was performed on 1 μg of RNA using Superscript reverse transcriptase (Gibco) as previously described (16). The PEST domain of ODC was PCR amplified withpfu polymerase using specific sense (5′-GCCGGGGTACCACCGGCTCGGACGATGAAG, bp1077) and antisense (5′-CGGCGGGATCCTACATTGATACTAGCAGAAGC, bp1521) primers. The PEST domain of dopa decarboxylase (DDC) was PCR amplified using specific sense (5′-GCCGGGGTACCATGGATTCCCGTGAATTCC, bp95) and antisense (5′-CGGCGGGATCCCCCAGCTCTTCCAGCC, bp477) primers. In both cases the PCR product was cloned in frame into theKpnI and BamHI sites of the pEGFP-C1 vector, and the integrity of the reverse transcription-PCR protocol was confirmed by sequencing in both directions.
Plasmid DNA for cell transfections was prepared using the Qiafilter method of preparation (Qiagen).
Cell culture.Cos-7 cells were maintained in complete medium, which consisted of Dulbecco's modified Eagle's medium (DMEM; BioWhittaker) containing 10% fetal bovine serum and 1% penicillin-streptomycin solution (Life Technologies). Cells were cultured in a humidified incubator with 5% CO2 at 37°C.
For transient-transfection experiments, cells were seeded at a density of 106 per 100-mm dish. After 24 h, cells were cotransfected with 5 μg of pEP-GR and either 20 μg of the pEP-HDC constructs, 10 μg of 5′UTR-luciferase constructs (pGL-01 or pGL-75), 10 μg of pBF-PEST constructs (pBF-PEST1 or pBF-PEST2), or 20 μg of pGF-PEST constructs (pGF-PEST/ODC or pGF-PEST/DDC), using the calcium phosphate method (5′-3′, Inc.). After 16 h the transfection mix was removed, and complete medium was added to the cells. For stimulation experiments, the medium was replaced 24 h later with serum-free Ultraculture medium (BioWhittaker) supplemented withl-glutamine (2 mM) and 1% penicillin-streptomycin solution (Life Technologies). Fresh actinomycin D (20 μg/ml, final concentration) was added to the Ultraculture medium just before addition to the cells. Cells were treated with actinomycin D for a minimum of 2 h before the addition of 10−7 M gastrin (Peninsula Laboratories), 10−8 M phorbol myristate acetate (PMA; Sigma), 10−4 M forskolin (Sigma), 10−6M thapsigargin (Sigma), or 10−5 M lactacystin (BioMol). When appropriate, 10−6 M staurosporin (Calbiochem), 10−4 M PD98085 (New England Biolabs), or 10−5M cycloheximide (Sigma) was added at the same time as actinomycin D for 2 h before gastrin stimulation.
Isolation of stomach tissue from rats.Male Sprague-Dawley rats weighing 200 to 250 g were fasted with free access to water; 48 h later, standard dietary nuts were added to cages containing test animals, and the rats were allowed to feed ad libitum. Control and test rats were sacrificed 3 h later by lethal injection (100 ng of ketamine-HCl per 100 g) and cervical dislocation. Whole stomachs were isolated and cleaned in phosphate-buffered saline (PBS), (136 mM NaCl, 2.7 mM KCl, 8.1 mM Na2HPO4, 1.5 mM KH2PO4). The tissue was homogenized in ice-cold 0.1 M sodium phosphate buffer (pH 7.4) containing 0.2 mM dithiothreitol. The tissue homogenate was centrifuged at 10,000 × g for 15 min, and the supernatant was transferred to a fresh tube. Protein content was estimated by the method of Bradford (Bio-Rad) for subsequent fractionation. Experiments were performed in accordance with local animal welfare regulations.
Assay of HDC activity.Cells were harvested in 2 ml of PBS, with a 500-μl aliquot removed for RNA analysis. The remaining cells were assayed for HDC activity using an adaptation of previously described methods (8). Briefly, cells were pelleted and resuspended in ice-cold 0.1 M sodium phosphate buffer (pH 7.4) containing 0.2 mM dithiothreitol and sonicated for 10 s. Whole-cell extracts were assayed for HDC activity by incubating 40-μl aliquots with 40 μl of 2× reaction buffer (2 μCi ofl-[14C]histidine [Amersham], 0.5 mMl-histidine, 0.01 mM pyridoxal phosphate, 0.1 M sodium phosphate [pH 6.8]). Reactions were performed in an open microcentrifuge tube placed in a closed scintillation vial.14CO2 generated by the enzymatic reaction was trapped in 50 μl of 80% (vol/vol) Soluene 350 (Packard). The enzyme reaction was stopped by the addition of 50 μl of 3 M perchloric acid, and the reaction was incubated at room temperature for 30 min. Levels of radiolabeled CO2 were determined by liquid scintillation counting. All samples were normalized to total protein content, and the protein concentration was determined by the method of Bradford.
Assay of luciferase activity.Cells were harvested in 2 ml of PBS, with a 500-μl aliquot removed for RNA analysis. The remaining cells were lysed in 1× cell lysis buffer and assayed for luciferase activity as advised by the manufacturer (Promega). All samples were normalized for protein content, and protein concentration was determined using a detergent-compatible protein estimation kit (Bio-Rad).
RNA analysis.Cells to be analyzed for RNA were pelleted and stored at −70°C until required. RNeasy kits (Qiagen, La Jolla, Calif.) were used to extract total RNA from in vitro-cultured cells. Total RNA was fractionated on 1% agarose denaturing formaldehyde gels, and the RNA was blotted to nylon membranes (Amersham Pharmacia Biotech) using established capillary blotting methods. DNA probes for Northern blot analysis were labeled with [α-32P]dCTP (3,000 mCi/mmol; NEN) using a random primer labeling kit as advised (Megaprime; Amersham). A full length (2.36 kb) HDC cDNA fragment generated by EcoRI and SalI digestion of the pEP-HDC2.4 vector was used as the template for an HDC-specific probe. For luciferase-specific hybridizations, a 500-bp fragment generated byHindIII and EcoRV digestion of the pGL75 vector was used as the template for labeling reactions. For glyceraldehyde-3-phosphate dehydrogenase (G3PDH), a commercially available human G3PDH probe was used (Clontech). Hybridizations were performed at 65°C using Quickhyb solution (Stratagene) following the manufacturer's instructions, and membranes were washed to high stringency in 0.1× SSC (1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate)–0.1% sodium dodecyl sulfate (SDS) at 65°C. After exposure to X-Omat LS autoradiographic film (Kodak), blots were stripped with 0.1% SDS at 95°C for 5 min before rehybridization with other probes. Quantitation was performed using appropriate computer software (NIH Image).
Generation of an anti-HDC antibody.A custom-prepared anti-HDC antibody was raised in a rabbit and affinity purified (Immunodynamics). Peptides corresponding to residues 30 to 44, 133 to 147, and 321 to 335 of the rat HDC protein sequence were used for immunizations. To confirm the specificity of the antibody, lysates from Cos-7 cells transfected with the pEP-empty vector were compared with lysates from cells transfected with either pEP-HDC1.5 or pEP-HDC2.0 as described in the text (see Fig. 6C). Comparisons were performed using both Western blotting and immunoprecipitation methods.
Western blotting analysis.Cell and tissue lysates were diluted in 2× sample buffer (130 mM Tris-HCl containing 4% SDS, 20% glycerol, 10% 2-mercaptoethanol, and 0.1% bromophenol blue), boiled for 5 min, and electrophoresed on SDS-polyacrylamide (10%) gels. Fractionated proteins were transferred electrophoretically (60 mA) to a polyvinylidene difluoride membrane (NEN Dupont) in 24 mM Tris buffer containing 40 mM glycine and 20% methanol. The transfer was performed overnight at 4°C. Membranes were blocked in PBS containing 1% Tween 20 (PBS-T) and 5% nonfat milk overnight at 4°C. The membranes were washed briefly with room temperature PBS-T before the addition of primary antibodies. Anti-HDC antibody was added at a dilution of 1:20,000 in PBS-T containing 5% nonfat milk, while the anti-blue fluorescent protein (anti-BFP) antibody (Clontech Living Colors) was added at a dilution of 1:1,000 in PBS-T containing 1% bovine serum albumin. After 2 h, the membranes being probed for HDC were washed three times for 20 min at 39°C in PBS-T, while membranes being probed for BFP were washed at room temperature. After washing, the blots were incubated for 1 h with horseradish peroxidase-conjugated anti-rabbit immunoglobulin G (Amersham) at a dilution of 1:1,000 in PBS-T containing 1.0% bovine serum albumin. The membrane was washed three times in PBS-T at room temperature, and immunoreactive proteins were detected using the Renaissance kit (NEN). Quantitation was performed using appropriate computer software (NIH Image).
Immunoprecipitation of HDC isoforms and GFP-PEST chimeras.Cos-7 cells were seeded at a density of 106 cells per 100-mm dish and cotransfected with 5 μg of pEP-GR and 20 μg of pEP-HDC2.4, pGF-PEST-ODC, or pGF-PEST-DDC as described above. Twenty-four hours after removal of the transfection mix, the cells were washed twice with PBS and incubated in cysteine- and methionine-free DMEM (Cys−/Met− medium; BioWhittaker) supplemented with 10% dialyzed fetal bovine serum (Gibco), 2 mMl-glutamine, and 1% penicillin-streptomycin solution (Gibco). After 1 h the medium was replaced with Cys−/Met− medium supplemented with actinomycin D (20 μg/ml, final concentration) and 200 μCi of Easytag Express 35S-labeled methionine-cysteine mix (1,175 mCi of [35S]methionine per mmol; New England Nuclear). After a 4-h pulse, cells were washed twice with PBS, and complete medium supplemented with actinomycin D (20 μg/ml, final concentration) was added. For test cells, gastrin (10−7 M) or lactacystin (10−5 M) was added. Cells were harvested at appropriate time points in 1 ml of radioimmunoprecipitation (RIPA) buffer (150 mM NaCl, 20 mM Tris-HCl [pH 7.5], 2 mM EDTA, 0.1% SDS, 0.25% deoxycholate, 1% Triton X-100) supplemented with protease inhibitors (Boehringer Mannheim). After a minimum incubation of 1 h on ice, cell lysates were centrifuged at 12,000 × g for 10 min at 4°C, and a 950-μl aliquot of the supernatant was added to 30 μl of protein A-Sepharose CL-4B (Pharmacia; 30 mg/ml in PBS). Samples were incubated overnight at 4°C with constant agitation. Precleared samples were centrifuged for 1 min at 12,000 × g, and a 900-μl aliquot of supernatant was transferred to a fresh tube containing 2 μl of affinity-purified anti-HDC antibody or 4 μl of polyclonal anti-GFP antibody (Molecular Probes). After a 1-h incubation at 4°C, 30 μl of protein A-Sepharose CL-4B was added, and samples were incubated for a further hour. Samples were centrifuged at 12,000 × g for 1 min, and the resulting precipitate was washed four times with 1 ml of RIPA buffer. The pellet was diluted in an equal volume of 2× sample buffer, and radiolabeled HDC isoforms were fractionated by electrophoresis on SDS-polyacrylamide (10%) gels. Gels were dried under vacuum and exposed to Biomax MR film using appropriate LS intensifying screens (Kodak). Quantitation was performed using appropriate computer software (NIH Image).
Intracellular localization of GFP-HDC chimeras.Cos-7 cells were seeded at a low density on poly-d-lysine-coated glass microscope wells. The next day, the cells were transfected for 2 h with 20 μg of the pGF constructs (pGF-empty, pGF-ER1, pGF-ER2, pGF-antiER2, and pHDC2.0-GF) and Superfect reagent, as advised by the manufacturer (Qiagen). After 24 h, ER-Tracker Blue-White DPX was added to the cells at a concentration of 100 nM in complete medium as advised by the manufacturer (Molecular Probes). The cells were incubated for a further 30 min at 37°C before being washed twice with PBS and visualized at ×100 and ×60 magnification by fluorescent microscopy.
RESULTS
HDC activity is stimulated by gastrin in a transcription-independent manner.In order to examine the posttranscriptional regulation of HDC, we have generated an HDC expression construct (pEP-HDC2.4) that codes for the rat HDC protein. The construct contains the complete HDC cDNA sequence cloned downstream from the CMV promoter and upstream from the simian virus 40 polyadenylation signal. The inserted fragment includes 75 bp of 5′UTR, 317 bp of 3′UTR, and 1,968 bp of coding sequence (total, 2.36 kb).
To determine whether gastrin can increase HDC activity in Cos-7 cells in a transcription-independent manner, the pEP-HDC2.4 construct and a second construct expressing the CCK-B/gastrin receptor (pEP-GR) were transiently cotransfected, and the cells were stimulated with 10−7 M gastrin. Stimulation was performed in the presence of actinomycin D to prevent increases in pEP-HDC2.4 transcription. Gastrin induced an immediate increase in HDC activity that was detectable within half an hour of stimulation (Fig.1A). The activity continued to increase gradually through the 3-h time point, at which stage activity levels appeared to plateau and remained constant at 2.6 ± 0.3 nmol/mg/h through the 5-h time point (Fig. 1A). This 2.6- ± 0.2-fold increase in activity occurred without any increase in the levels of HDC mRNA (Fig.1B).
Time course showing the effect of gastrin stimulation on HDC enzyme activity and HDC mRNA levels. Cos-7 cells transfected with pEP-HDC2.4 and pEP-GR were treated for 2 h with actinomycin D and stimulated in a reverse time course with gastrin (10−7 M) for the time periods shown. (A) Effect of stimulation on HDC enzyme activity. Data show the combined results from four independent experiments ± SEM. (B) Northern blot showing the expression of HDC mRNA after gastrin stimulation. -ve, expression in cells transfected with the pEP-empty vector and stimulated with gastrin for 5 h. G3PDH shows relative loading. The 18S and 28S arrows indicate the migration of rRNA subunits. The graph shows HDC expression corrected for loading. Similar Northern blotting results were obtained in three other experiments.
Gastrin stimulation of HDC involves activation of PKC and MEK1 pathways.We next wanted to identify which intracellular signalling pathways are involved in this gastrin effect. Cells transfected with pEP-HDC2.4 and pEP-GR were pretreated for 2 h with actinomycin D and stimulated with forskolin, PMA, and thapsigargin in order to activate protein kinase A (PKA), protein kinase C (PKC), and intracellular calcium release, respectively. The results from these experiments, which are shown in Fig. 2A, indicated that gastrin-induced increases in HDC activity were most closely mimicked by stimulation with 10−8 M PMA, with levels increased from 1.04 ± 0.18 nmol/mg/h to 3.5 ± 0.65 nmol/mg/h. Stimulation with either thapsigargin or forskolin had no significant effect on HDC activity. Northern blotting confirmed that HDC mRNA levels were unaltered by treatment with these compounds (Fig.2B).
Effect of signaling pathway activators and inhibitors on HDC enzyme activity. (A and B) Cos-7 cells transfected with pEP-HDC2.4 and pEP-GR were treated for 2 h with actinomycin D and stimulated for 4 h with 10−7 M gastrin, 10−4 M forskolin (Forsk.), 10−8 M PMA, or 10−6 M thapsigargin (TG). -ve, cells transfected with the pEP-empty vector. Control, unstimulated cells transfected with pEP-HDC2.4. (A) Effect of stimulation on HDC enzyme activity. Data are the means of three independent experiments ± SEM. (B) Northern blot showing the expression of HDC mRNA after stimulation with the signaling pathway activators shown in A. G3PDH shows relative loading. The graph shows HDC expression corrected for loading. Similar Northern blotting results were obtained in three other experiments. (C) Cos-7 cells transfected with pEP-HDC2.4 and pEP-GR were treated for 2 h with actinonmycin D and specific inhibitors; PKC inhibitor staurosporine (Stauro., 50 μM), MEK1 inhibitor PD98059 (20 μM), or the translation inhibitor cycloheximide (CHX, 20 μg/ml). After 2 h of treatment, the cells were stimulated for 4 h with gastrin (10−7 M). Data, which show the fold increase in HDC enzyme activity after gastrin stimulation, are the combined results from three independent experiments ± SEM.
To confirm the involvement of PKC and downstream pathways in gastrin activation of HDC, transfected and actinomycin D-treated cells were incubated with one of a range of inhibitors and stimulated for 4 h with 10−7 M gastrin (Fig. 2C). Compared with control cultures, which showed a 2.3- ± 0.4-fold increase in activity, cells treated with the PKC inhibitor staurosporin exhibited only a 0.8- ± 0.2-fold increase (Fig. 2C, lane 2). Cells treated with the MAP or ERK kinase 1 (MEK1) inhibitor PD98059 showed partial inhibition, with a 1.6- ± 0.2-fold increase (Fig. 2C, lane 3). The gastrin effect was also significantly inhibited by treatment of transfected cells with the translation inhibitor cycloheximide, demonstrating a requirement for novel protein synthesis (Fig. 2C, lane 4).
Gastrin stimulation leads to increased levels of HDC isoforms.A total of five HDC isoforms have so far been identified in rat tissue extracts. This includes the full-length 74-kDa primary translation isoform, a 63-kDa isoform, two isoforms having molecular masses close to 54 kDa, and a smaller isoform of about 36 kDa (8). We wished to investigate the molecular basis for transcription-independent increases in HDC activity and consequently developed an affinity-purified polyclonal antibody to study HDC expression at the protein level. The antibody was raised against three peptide regions in the amino terminus of rat HDC, and its specificity was confirmed by immunoprecipitations and Western blots (see Materials and Methods section).
In order to examine the effect of gastrin treatment on the expression of different HDC isoforms, pEP-HDC2.4- and pEP-GR-transfected Cos-7 cells were treated with actinomycin D and stimulated for 4 h with gastrin. Western blotting was performed on total-cell lysates and showed a 1.9- ± 0.2-fold increase (n = 3, mean ± standard error of the mean [SEM]) in the levels of the 74-kDa HDC isoform (Fig. 3A, 10-min exposure). Similar increases in expression were observed for a 63-kDa isoform (Fig. 3A, 10-min exposure), and also for a number of smaller HDC isoforms that could only be detected after longer exposure of the blots to film (Fig. 3A, 30-min exposure). These smaller isoforms were estimated to be 54, 48, 40, and 36 kDa in size. Northern blotting confirmed that HDC mRNA levels were unchanged by gastrin stimulation (data not shown).
HDC isoform expression in transfected Cos-7 cells and in rat stomach cell lysates. (A) Cos-7 cells transfected with pEP-HDC2.4 and pEP-GR were treated for 2 h with actinomycin D and then stimulated with 10−7 M gastrin for 4 h. Western blots were probed for HDC isoform expression using an anti-HDC antibody raised against antigens in the amino terminus of HDC. Immunoreactive HDC isoform expression was detected after 10 and 30 min of exposure of blots to autoradiographic film. The results shown are representative of four independent experiments. Asterisks indicate secondary bands as described in the text. (B) Rats that had been fasted for 2 days were given free access to food for 3 h. HDC isoform expression in the stomach cell lysates of fasted and refed animals was analyzed by Western blotting. The results shown are representative of three independent experiments performed on paired fasted and refed rats.
A total of six HDC isoforms were also observed in lysates derived from the stomachs of both fasted and refed rats (Fig. 3B). These isoforms were similar in size to those described for transfected Cos-7 cells.
While we report here the identification of six HDC isoforms, our experiments with Cos-7 cells suggested that a secondary band exists for some isoforms. In particular we noticed 68-kDa, 58-kDa, and 51-kDa bands located just above the 63-kDa, 54-kDa, and 48-kDa bands, respectively (Fig. 3A, 30 min). These differences of ∼3 to 5 kDa could result from a secondary cleavage step or significant posttranslational modifications, such as glycosylation. These secondary bands were not as obvious in rat tissue extracts.
5′UTR of HDC mRNA does not mediate a gastrin-stimulated increase in mRNA translation.Gastrin stimulation of HDC activity in Cos-7 cells is sensitive to cycloheximide treatment (Fig. 2C) and leads to an increase in the levels of the primary 74-kDa HDC translation product (Fig. 3A). Therefore it was hypothesized that gastrin could act to increase the translation of HDC mRNA. Previous studies have shown that gastrin can act via 5′UTR sequences and in a cycloheximide-sensitive manner to increase the translation of ODC mRNA, strengthening the rationale for this hypothesis (33).
To test whether this type of translational control could be relevant to HDC, we generated the expression constructs pGL-01 and pGL-75, which are shown diagrammatically in Fig. 4A. These vectors contain either 1 or 75 bp of the HDC 5′UTR cloned downstream from the CMV promoter and upstream from the luciferase reporter cassette. Following transient transfection into Cos-7 cells and actinomycin D treatment, the effect of gastrin stimulation on luciferase mRNA translation was examined by assaying for luciferase enzyme activity.
Effect of gastrin stimulation on 5′UTR-mediated translation. (A) Diagrammatic representation of 5′UTR-luciferase (Luc.) constructs containing either 1 bp (pGL-01) or 75 bp (pGL-75) of HDC 5′UTR. (B) Cos-7 cells were transfected with gastrin receptor (pEP-GR) and 5′UTR-luciferase (pGL-01 and pGL-75) constructs. Cells were treated for 2 h with actinomycin D and cultured in the presence or absence (control) of 10−7 M gastrin for 4 h. The graph shows the effect of gastrin stimulation on absolute luciferase activities in a typical experiment (mean ± range for two different transfections), and data are representative of six independent experiments. RLU, relative light units.
Figure 4B shows the absolute luciferase values obtained in a representative experiment. Luciferase expression tended to increase slightly after gastrin stimulation. This occurred for both the pGL-01 (1.2- ± 0.2-fold mean ± SEM, n = 6) and pGL-75 (1.1- ± 0.2-fold mean ± SEM, n = 6) vectors. The fact that this trend was observed in both the presence and absence of 5′UTR suggested that it was more related to the luciferase protein than to the 5′UTR of HDC.
Interestingly, the values obtained for cells transfected with the pGL-75 construct were consistently greater (2- to 3-fold) than for cells transfected with the pGL-01 vector (compare lanes 1 and 2 with lanes 3 and 4 in Fig. 4B). These differences were not due to differences in luciferase mRNA expression (data not shown), suggesting that the 5′UTR is indeed critical for regulation of basal HDC translation (23) but is unlikely to be sensitive to gastrin.
Gastrin stimulation stabilizes HDC isoforms.Having demonstrated that gastrin stimulation of the 5′UTR is unlikely to increase HDC translation, we wanted to investigate whether increased isoform expression arises as a result of increased protein stability. In order to test this, we pulse-labeled transfected Cos-7 cells for 4 h, and the pattern of HDC isoform degradation was chased in the presence and absence of gastrin. Protein lysates from pulse-labeled cells were subjected to immunoprecipitation using our anti-HDC antibody, which confirmed the expression of multiple HDC isoforms (Fig.5A). However, some of the isoforms were more weakly detected than by Western blotting, presumably reflecting differences in the affinity of the antibody for the native isoforms in solution.
Pulse-chase showing the effect of gastrin stimulation on HDC isoform stability. (A) Cos-7 cells transfected with pEP-HDC2.4 and pEP-GR were pulse labeled with 35S-labeled methionine and cysteine in the presence of actinomycin D. After a 4-h pulse, cells were cultured in control medium in the absence (control) and presence of 10−7 M gastrin for the time intervals shown. HDC isoforms were immunoprecipitated and electrophoretically fractionated. (B) Graphs showing the effect of gastrin stimulation on the degradation of the 48-kDa and 63-kDa HDC isoforms. Each point represents the mean ± SEM for three independent experiments. Values for isoform half-lives (t 1/2) represents the mean ± SEM of values determined for each of the three individual experiments. Statistically significant differences relative to the corresponding control values are indicated by a superscript a (P < 0.05, Student's t test).
Densitometric analysis of these pulse-chase studies allowed us to conclude that HDC isoforms degrade at different rates. In control, unstimulated cells, the degradation half-life of the 48-kDa isoform was 4.2 ± 0.1 h (Fig. 5B, top graph). In contrast, the half-life of the 63-kDa HDC isoforms was only 3.0 ± 0.1 h (Fig. 5B, bottom graph). Gastrin stimulation increased the half-lives of both isoforms, maintaining the differential degradation rates of the two isoforms. For the 48-kDa isoform, the half-life was increased to 5.8 ± 0.5 h (Fig. 5B, top graph), while the half-life of the 63-kDa isoform was increased to 3.7 ± 0.1 h (Fig. 5B, bottom graph). The half-lives of the other HDC isoforms also appeared to increase with gastrin stimulation, although low levels of expression prevented exact quantification (Fig. 5A).
Distinct domains in the primary protein sequence regulate HDC expression levels.Gastrin stimulation increased the half-lives of both the 48- and 63-kDa isoforms (Fig. 5); however, in view of the fact that these increases occurred in the setting of different basal rates of protein turnover, it appeared likely that expression of the different HDC isoforms can be differentially regulated. There was some evidence for this in gastrin-stimulated Cos-7 cells, where the increase in expression of the 36-kDa isoform appeared greater than increases observed in expression of some of the other isoforms (Fig. 3A). Refeeding of fasted rats also appeared to preferentially increase the expression of the smaller HDC isoforms in rat stomach cell lysates. For example, we noted a 3.5- ± 1.0-fold increase in the 48-kDa isoform, compared to a 1.7- ± 0.5-fold increase in the 74-kDa isoform (mean ± SEM, n = 3, Fig. 3B). This led us to believe that there are domains present in some but not all isoforms which are capable of differentially regulating enzyme stability and protein expression levels.
To help identify domains involved in the regulation of HDC protein expression, we have generated a series of deletion constructs that express carboxy-truncated HDC isoforms. The primary translation products predicted for these constructs are shown diagrammatically in Fig. 6A. The vector pEP-HDC2.0 is similar to pEP-HDC2.4 used earlier in this study but lacks any 3′UTR sequence. It was therefore predicted to generate a full-length primary translation product of 74 kDa. The second construct, pEP-HDC1.7, lacks the sequence coding for the carboxy-terminal 87 amino acids. By removing this region, which contains a putative intracellular targeting domain (ER2), it was predicted that the resulting protein would be 64 kDa in size (Fig. 6A). Further 3′ deletion of the cDNA led to generation of the pEP-HDC1.6 vector. The protein expressed from this plasmid was predicted to be 58 kDa in size and devoid of a C-terminal PEST domain (PEST2) that has been hypothesized to influence HDC stability (Fig. 6A). Finally, the primary translation product generated by the pEP-HDC1.5 vector lacks a total of 170 amino acids from the carboxy terminus of the primary HDC sequence. This construct is similar in design and size to constructs which have been used in previous studies (8, 47, 48), where it was assumed that HDC cleavage involves only carboxy-terminal processing. It was predicted to generate a primary translation product of 54 kDa (Fig. 6A).
Effect of carboxy-terminal truncations on HDC expression. (A) Diagrammatic representation of the carboxy-truncated constructs used in panels B, C, and D. Predicted sizes of the primary translation products are shown on the right-hand side. Amino acid numbers in the primary rat sequence are shown underneath. (B, C, and D) Expression constructs pEP-HDC2.0, pEP-HDC1.7, pEP-HDC1.6, and pEP-HDC1.5 were transiently transfected into Cos-7 cells. After 48 h, the cells were harvested for (B) HDC enzyme activity analysis, (C) HDC protein expression analysis by Western blotting, and (D) HDC mRNA expression by Northern blotting. G3PDH shows relative loading. -ve, cells transfected with the pEP-empty vector. The enzyme activities (B) and Western blot (C) are derived from the same experiment and are representative of six independent experiments. Enzyme activities show mean ± range for two readings from the same sample. The Northern blot is representative of four independent experiments.
The HDC deletion constructs were transiently transfected into Cos-7 cells, and enzyme activities were compared to protein expression levels for each of the constructs (Fig. 6B and C). These studies showed relatively low levels of the 74-kDa HDC isoform, which is the primary translation product generated by the pEP-HDC2.0 construct (Fig. 6C, lane 5). This corresponded closely to low but detectable levels of enzyme activity (Fig. 6B, lane 5). Removing the putative carboxy-terminal targeting domain (ER2) led to an increase in the levels of the HDC protein expressed from the pEP-HDC1.7 vector (Fig.6C, lane 4) and increased enzyme activity (Fig. 6B, lane 4). The protein expressed from the pEP1.7 vector was 60 kDa in size, not 64 kDa as predicted (Fig. 6C, lane 4). The greatest protein expression and enzyme activity increases occurred for the pEP-HDC1.6 (lane 3) and pEP-HDC1.5 (lane 2) vectors (Fig. 6C and B, respectively). In these instances, the primary sequence had been truncated to remove the putative targeting domain (ER2) and additional amino acids 517 to 568, which roughly correspond to PEST2 (Fig. 6A). The proteins expressed from the pEP-HDC1.5 and pEP-HDC1.6 vectors were 49 and 54 kDa in size, not 54 and 58 kDa as predicted.
While these results suggest that the specific activity of the 74-kDa isoform is similar to that of other, smaller isoforms, our inability to block cleavage to smaller, potentially more active isoforms makes it impossible to draw definitive conclusions. However, it was noted that increasing HDC expression by carboxy-terminal truncation did not lead to corresponding increases in the levels of the processed 48-kDa, 40-kDa, or 36-kDa HDC isoforms. Northern blotting confirmed expression of HDC mRNAs of the correct sizes (Fig. 6D).
Amino acids 568 to 656 of the rat HDC protein encode an endoplasmic reticulum signaling sequence.Cell fractionation experiments generally suggest that the 74-kDa and 53- or 54-kDa rat HDC isoforms are differentially localized within the cell, although contradictory results have been reported (44, 47). Because these earlier studies assumed that the 54-kDa isoform is generated as a consequence of carboxy-terminal cleavage only, it has been proposed that amino acids 485 to 656 of the rat protein sequence are involved in intracellular targeting (8, 44).
Our results in this study showed that removal of ER2 (amino acids 555 to 656) led to an increase in HDC protein expression levels (Fig. 6A and C). To test whether this region that regulates expression is also responsible for intracellular localization, we have generated a series of GFP chimeras which contain in-frame HDC sequences. The GFP chimeras were transiently transfected into Cos-7 cells, and the pattern of intracellular localization was observed by fluorescent microscopy. In the absence of any additional sequence, the GFP protein was localized throughout the cell but predominantly to the nucleus (Fig. 7A). Fusion of the entire HDC protein sequence to the amino terminus of GFP changed this pattern, with expression clearly localized to a vesicular network outside the nucleus (Fig. 7B). An identical pattern of localization was observed for a GFP chimera containing the ER2 region alone (Fig. 7C, left-hand panel), suggesting that this domain contributes to the pattern of localization observed for the full-length protein. A specific probe (ER-Tracker) confirmed endoplasmic reticulum localization of GFP-ER2 (Fig. 7C, right-hand panel). It is noteworthy that fusion of ER2 by itself to GFP greatly reduced protein expression levels, as indicated by low levels of fluorescence of GFP-ER2 and the requirement for extended exposure times for photography (Fig. 7C, left-hand panel). The GFP-ER2 chimera was fluorescently tagged at the amino terminus (Fig. 7C, left-hand panel), whereas the HDC2.0-GFP chimera was fluorescently tagged at the carboxy end (Fig. 7B). This alternate pattern of tagging confirmed that localization does not occur as a consequence of cleavage within the ER2 domain of the fluorescent chimeras.
Localization of fluorescent chimeras. Cos-7 cells were transfected with the expression constructs pGF-empty (A), pHDC2.0-GF (B), pGF-ER2 (C), pGF-antiER2 (D), and pGF-ER1 (E), as diagrammatically represented with numbered amino acids from the HDC protein sequence. Panels A and B and the left-hand panels of C, D, and E show patterns of intracellular localization of GFP chimeras 24 h after transfection. The right-hand panels of C, D, and E show endoplasmic reticulum localization of the ER-Tracker probe in the cells shown in the corresponding left-hand panels. The magnifications and exposure times are shown underneath the photographs.
As additional confirmation that ER2 contains an endoplasmic reticulum targeting domain that promotes degradation, transfections were performed with a GFP chimera that contained the ER2 sequence in the antisense direction. For this chimera, GFP-antiER2, the localization of fluorescence to the nucleus and the intensity of fluorescence (Fig. 7D, left-hand panel) were similar to that observed for the GFP alone (Fig.7A). Staining with the ER-Tracker probe confirmed that localization of GFP-antiER2 was not specific to the endoplasmic reticulum (Fig. 7D, right-hand panel). Interestingly, the region described here as ER2 contains a 20-amino-acid sequence (N580 to G600) which is 74% homologous to an endoplasmic reticulum targeting sequence very recently identified in the mouse HDC protein sequence (41).
Our results indicated that endoplasmic reticulum localization of the GFP-ER2 chimera actually promotes degradation. Taken together with the Western blots shown in Fig. 6C, where removal of ER2 increased protein expression levels, these results suggested that degradation is a function of localization. We therefore wanted to test whether gastrin stabilization of HDC isoforms occurs as a consequence of altered ER2-mediated localization within the cell. Cos-7 cells were transfected with pEP-GR and pGF-ER2, and after treatment with actinomycin D, the cells were stimulated for 4 h with gastrin. Endoplasmic reticulum localization of the GFP-ER2 chimera, as shown in the left-hand panel of Fig. 7C, was unaltered by gastrin stimulation (data not shown).
PEST domains of HDC regulate protein expression levels in a gastrin-responsive manner.Our results in Fig. 6 showed that the HDC sequence between amino acids 517 and 568 also regulates protein expression levels. This region corresponds in part to a PEST domain previously identified by computer analysis (15) and is one of two such regions identified in the primary rat HDC protein sequence (Fig. 6A). To study the role of this PEST domain, PEST2, in regulating HDC enzyme expression and to test whether this region might mediate gastrin regulation of isoform expression, we fused the PEST2 peptide in frame to the carboxy terminus of the BFP (Fig.8A, BFP-PEST2). This pBF-PEST2 construct was transiently transfected into Cos-7 cells, and after actinomycin D treatment, the cells were stimulated with 10−7 M gastrin. Western blotting with an antibody raised against the BFP was used to examine protein expression levels.
Effect of gastrin stimulation and proteasome inhibition on BFP-PEST expression levels. (A) Diagrammatic representation of BFP-PEST2 and BFP-PEST1, with numbered amino acids from the primary rat HDC protein sequences shown underneath. (B, C, D, and E) Cos-7 cells were cotransfected with pEP-GR and either pBF (B and C), pBF-PEST2 (B and D), or pBF-PEST1 (C and E). Cells were treated for 2 h with actinomycin D before incubation for 4 h with 10−7 M gastrin (B and C) or for 6 h with 10−5 M lactacystin (D and E), a proteasome inhibitor. Control cells were left untreated (B, C, D, and E). Western blotting using an anti-BFP antibody was used to examine the patterns of chimeric protein expression. The results shown are representative of not less than three independent experiments, as detailed in the text.
The fusion of PEST2 sequences to the BFP resulted in a 5- ± 2-fold reduction (mean ± SEM, n = 6) in protein expression levels (Fig. 8B, compare lanes 1 and 3). This effect was partially reversed by 4 h of gastrin stimulation, which led to a 3.9- ± 0.9-fold (mean ± SEM, n = 4) increase in fusion protein expression levels (Fig. 8B, compare lanes 3 and 4). These results indicated not only that this PEST2 region is capable of regulating protein expression levels, but also that it is highly gastrin responsive.
A similar series of experiments were conducted with the BFP fused in frame to amino acids 1 through 119 of the HDC protein sequence (Fig.8A, BFP-PEST1). This region also contains a previously defined PEST domain (PEST1) between amino acids 43 and 74 (15). However, rather than fusing this region alone in frame to the BFP, the construct pBF-PEST1 was designed to include the entire amino terminus. In this way it was also possible to test for the presence of an amino-terminal cleavage site.
The effect of this PEST1 domain on BFP expression levels was examined by transient transfection of pBF-PEST1 and pEP-GR into Cos-7 cells and stimulation of actinomycin D-treated cells with gastrin for 4 h. Western blots using the anti-BFP antibody confirmed that levels of this PEST1 chimera could also be regulated by gastrin, with stimulation leading to a 1.4- ± 0.1-fold (mean ± SEM, n = 3) increase in expression (Fig. 8C). In addition to detecting a 45-kDa BFP-PEST1 protein of the predicted size, Western blotting also detected a smaller protein of about 35 kDa (Fig. 8C). This is larger than the BFP, indicating the presence of a chimera containing additional HDC protein sequence (∼4 kDa).
The results shown in Fig. 8B and C identify the PEST domains as independent and portable elements, both of which promote protein degradation and both of which respond to gastrin stimulation. They are unrelated sequentially other than that they define highly hydrophilic regions within the protein. While both elements respond to gastrin stimulation, it was noted that the response of the BFP-PEST2 chimera was greater than that of the BFP-PEST1 chimera. Disproportionate increases in expression levels were also observed when transfected and actinomycin D-treated cells were cultured for 6 h with the proteasome inhibitor lactacystin. Once again the BFP-PEST2 fusion protein appeared to be the most sensitive (Fig. 8D), with proteasome inhibition leading to a 1.8- ± 0.2-fold (mean ± SEM, n = 4) increase in expression, compared to a 1.4- ± 0.2-fold (mean ± SEM, n = 5) increase in expression for the BFP-PEST1 fusion protein (Fig. 8E).
Gastrin stimulation stabilizes chimeric proteins that contain the PEST domains from ODC and DDC.The experiments shown in Fig. 8demonstrated that gastrin disproportionally increases the expression of sequentially unrelated HDC-PEST domain chimeras, with the most rapidly turned-over BFP-PEST2 protein also being the most stimulated by gastrin. Consequently it was proposed that gastrin might regulate the stability of other rapidly turned-over PEST domain-containing proteins. To test this, we cloned the PEST domains of two other decarboxylase enzymes, ODC and DDC, and fused them in frame to the carboxy terminus of GFP (pGF-PEST/ODC and pGF-PEST/DDC) (Fig.9A). The effect of gastrin stimulation on GFP-PEST fusion protein stability was examined in pulse-chase experiments. These experiments confirmed that gastrin was able to regulate the stability of other PEST domain-containing proteins, although to differing degrees (Fig. 9B). For the GFP-PEST/ODC protein, this gastrin regulation led to a 2.3- ± 0.5-fold increase in expression after a 6-h chase (mean ± SEM, n = 3; top panel, lane 2). Similar increases in GFP-PEST/ODC expression were observed when the proteasome was inhibited by lactacystin (2.6- ± 0.2-fold, mean ± SEM, n = 3; top panel, lane 3). Gastrin stabilization of the GFP-PEST/DDC fusion protein was less marked after the 6-h gastrin chase (1.4- ± 0.1-fold increase in expression; mean ± SEM, n = 3; middle panel, lane 2), and lactacystin treatment had no detectable effect on expression levels (1.1- ± 0.3-fold, mean ± SEM, n = 3; middle panel, lane 3).
Effect of gastrin stimulation and proteasome inhibition on the stability of GFP chimeras containing the PEST domains of ODC and DDC. (A) Diagrammatic representation of GFP-PEST/ODC and GFP-PEST/DDC, with numbered amino acids from the primary rat ODC and rat DDC protein sequences shown underneath. (B) Cos-7 cells transfected with pEP-GR and either pGF-PEST/ODC, pGF-PEST/DDC, or pGF-empty vector were pulse labeled with 35S-labeled methionine and cysteine in the presence of actinomycin D. After a 4-h pulse, cells were chased in control medium supplemented with 10−7 M gastrin or 10−5 M lactacystin for 6 h. Control cells were left untreated. GFP chimeras were immunoprecipitated and electrophoretically fractionated on a single gel. The results shown are derived from a common autoradiograph (i.e., common exposure time) from a representative experiment (n = 3).
DISCUSSION
Here we establish that the peptide hormone gastrin acts through G protein-stimulated pathways to regulate the stability of a group of rapidly turned-over PEST domain-containing proteins. The PEST domains used to prove this are sequentially unrelated other than that they define hydrophilic regions in their native proteins. This argues against a direct effect at the level of substrate recognition but instead suggests that gastrin inhibits some common component of the general protein degradation machinery. We further demonstrate that the PEST domain proteins most susceptible to this gastrin stabilization are also the proteins most affected by proteasome inhibition. This is again supportive of a gastrin effect on protein degradation, with rapidly turned-over PEST-containing proteins the most affected. The HDC enzyme, which contains PEST domains at both the amino- and carboxy-terminal ends, provides a unique model with which to study this type of hormonal regulation. Our results indicated that gastrin stimulation leads to increased HDC enzyme activity and that this transcription-independent increase requires the activation of PKC and MEK1 pathways. Although regulation is dependent upon novel protein synthesis, our experiments showed that it is unlikely to relate to increased translation of HDC. Instead, our results point to a dependence on the translation of one or more protein factors capable of inhibiting the degradation of HDC and other PEST domain-containing proteins. To our knowledge, this is the first report of a hormone regulating the degradation of multiple PEST domain-containing proteins and thus delineates a novel mechanism of hormone regulation of protein function.
In this study we characterized two discrete and transferable HDC PEST domains which promote degradation of heterologous proteins and regulate stabilization by gastrin. However, our studies identified additional features of HDC regulation that are likely to determine whether these domains are buried within the primary structure or exposed flush against the end of the enzyme. In particular, we identified a key role for the carboxy-terminal ER2 domain in intracellular localization and patterns of protein processing. First, we found that carboxy-terminal truncation to sequentially remove the ER2 and PEST2 domains successively increased expression of HDC protein, with corresponding increases in enzyme activity. While PEST2 regulation of protein expression might have been expected based on studies with other rapidly degraded proteins (35), it was not anticipated that removing the C-terminal 87 amino acids would, by itself, also lead to increases in protein expression levels. Fluorescent protein chimeras containing this 87-amino-acid region colocalize with endoplasmic reticulum-specific probes, suggesting that ER2-mediated targeting may be associated with degradation of the 74-kDa primary HDC translation product. This degradation is known to occur via the ubiquitin-proteasome pathway (43), in a process that is likely to involve ubiquitin-conjugating enzymes. Many of these conjugation enzymes have been localized to the endoplasmic reticulum membrane (1, 40), and thus an endoplasmic reticulum-localizing sequence could hypothetically target the 74-kDa isoform for conjugation and degradation. It is interesting that NF-κB activation involves partial degradation of the p105 isoform by the ubiquitin-proteasome pathway (7, 30, 31), raising the possibility of partial degradation and subsequent stabilization and activation of HDC by this mechanism.
The carboxy-terminal ER2 targeting domain also appears to influence protein processing, and it was noted that truncation of HDC to exclude this region led to substantially altered cleavage patterns (see Fig. 6C for processing of pEP-HDC1.5, -1.6, and -1.7 vector products). In the first instance, it was anticipated that carboxy-terminal truncations to generate stable HDC proteins would also lead to increased levels of the smaller processed HDC isoforms. We found no evidence for this, suggesting that generation of processed isoforms either directly or indirectly involves the ER2 region.
Our experiments with truncated HDC proteins also suggested that the carboxy-terminal ER2 domain somehow inhibits an amino-terminal cleavage, which is a novel form of processing not previously considered for HDC. Specifically, it was noted that expression of the full-length protein led to the detection on Western blots of a protein of the predicted 74-kDa size. In contrast, when we expressed truncated proteins that lacked the carboxy-terminal ER2 domain, the major protein band detected was about 4 to 5 kDa shorter than anticipated. Northern blots confirmed HDC transcripts of the correct sizes, and Western blots confirmed that the predicted pattern of carboxy-terminal truncations had been maintained. This suggested that discrepancies between predicted and actual sizes arise as a consequence of amino-terminal cleavage, which occurs shortly after translation. These results therefore provided preliminary evidence for an amino-terminal processing step which is promoted when the ER2 domain is either engineered to be absent (as was done here) or removed by cleavage (as is likely to occur in vivo).
Experiments with the PEST1 fluorescent chimera provided further evidence for this type of regulation. In addition to detecting the primary ∼45-kDa primary translation product for BFP-PEST1, low levels of a smaller specific band of ∼35 kDa were also detected. This band is slightly larger than the BFP protein, indicating the presence of additional HDC protein sequence, again consistent with an amino-terminal processing step.
Many proteins contain short amino-terminal sequences that are cleaved in response to specific stimuli (37) or following translocation across the endoplasmic reticulum (6, 41), and we confirm here (Fig. 7E) that amino acids 1 through 40 (hereafter referred to as ER1) are at least capable of targeting a chimeric fluorescent protein to the endoplasmic reticulum. The presence of the carboxy-terminal ER2 domain therefore appears to inhibit an amino-terminal cleavage that removes the ER1 domain. We conclude that more than one processing pathway is available for the primary translation product depending on whether or not the carboxy-terminal ER2 is present. It is conceivable that tissue-specific factors will regulate whether amino- or carboxy-terminal processing patterns dominate, which might explain why some of the isoforms identified in transfected Cos-7 cells were not as clearly detected in rat stomach cell lysates. It might also explain why only 74-kDa and 54-kDa HDC isoforms have been reported in rat basophilic cells (43).
Interestingly, an amino-terminal cleavage such as that described here would place the amino-terminal PEST domain flush against the end of the enzyme, potentially altering its susceptibility to degradation or, for that matter, regulation by gastrin. We have considered cleavage steps that would initially expose and then remove the PEST domains at both the amino- and carboxy-terminal ends. Possible cleavage permutations are shown diagrammatically in Fig. 10. The isoforms predicted by this model are almost identical in size to some of the isoforms identified in transfected Cos-7 cells (Fig. 3A), which strongly supports this model of protein processing. Previous studies have highlighted the importance of PEST domain exposure (ODC) (20, 28) and removal (NF-κB) (30, 31) on degradation function. It is interesting therefore that we propose a pattern of processing for HDC in which both steps feature in regulation. This regulation could occur as part of either the degradation or activation pathway. While future studies will need to address the relative enzyme activities of isoforms generated by this hypothesized cleavage scheme, it is noteworthy that previous in vitro translation studies suggest that a 48-kDa core protein located between PEST1 and PEST2 is highly enzymatically active (15); however, this is the first study to show that generation of such an isoform might have any physiological relevance. While previous studies suggest that the production of smaller HDC isoforms requires the initial translation of the 74-kDa isoform (44), it is noteworthy that our results do not completely rule out the use of alternative translational start sites in the generation of multiple HDC isoforms.
Diagrammatic representation of hypothesized cleavage pattern of HDC, based on inclusion or exclusion of endoplasmic reticulum targeting and PEST domains. Numbered amino acids from the primary rat HDC protein sequence are shown underneath. The approximate locations of the antigenic determinants used in the generation of the anti-HDC antibody used in this study are shown by solid triangles above the sequence.
The presence of targeting and degradation elements (ER1/PEST1 and ER2/PEST2) (Fig. 10) at both ends of the enzyme raises some interesting questions regarding the relevance of successive processing steps on HDC function. One possibility is that the 74-kDa HDC is initially targeted to the endoplasmic reticulum for degradation. Under certain conditions, partial degradation of the 74-kDa isoform to remove ER2 could allow retrograde translocation and amino-terminal cleavage of a more stable (and consequently more active) isoform back to the cytosol, where it is believed that histamine is produced (3, 4). Such retrograde translocation to the cytosol is an accepted model of intracellular protein trafficking (50) and could result in degradation regulated by PEST1, which is shown here to be less susceptible to proteasome degradation than PEST2.
It is in this context, therefore, where one or more functional domains are likely to have been exposed or removed, that we need to consider gastrin regulation of HDC. Pulse-chase experiments demonstrated that gastrin disproportionally increases the turnover rates of the different HDC isoforms. We propose that this form of regulation depends on the presence or absence of amino- and carboxy-terminal PEST domains, which differentially stabilize heterologous proteins in response to gastrin stimulation. In this model, the expression of isoforms that lack PEST domains should also increase due both to stabilization of precursor isoforms and to slower rates of degradation. This model assumes a constant rate of conversion between isoforms and the potential for independent degradation of each of the isoforms. At the present time, however, we cannot determine categorically whether selective conversion of isoforms occurs or whether parameters known to regulate stability (such as covalent modification or dimerization [2, 36, 17]) might represent an additional level of regulation.
In our experiments we observed that the gastrin effect was cycloheximide sensitive. While this could imply the translation of factors that interact with specific HDC domains to regulate stability, our experiments showed that gastrin can affect the expression of heterologous proteins containing PEST domains from a variety of different sources, including ODC and, to a lesser extent, DDC. This suggests a more general effect on the turnover of PEST-containing proteins. In preliminary competition cotransfection experiments performed with pEP-HDC2.4 and increasing amounts of pBF-PEST2, we found no effect on basal HDC enzyme activity (data not shown), which further argues against specific factors interacting with the PEST2 domain to specifically regulate HDC degradation. Our results therefore suggest that gastrin stabilizes HDC isoforms by regulating factors common to the degradation of a number of rapidly turned-over proteins. This interpretation is consistent with experiments showing that the heterologous proteins most susceptible to gastrin stimulation were also the proteins most affected by lactacystin inhibition of the proteasome (BFP-PEST2 and GFP-PEST/ODC). This type of regulation could involve the gastrin-stimulated translation of one or more inhibitors of degradation, although such a pattern of regulation has not, to our knowledge, been described previously. Such regulation could have important implications for other proteins as well. For example, transcription factors (such as Fos and NF-κB [35, 31]) and cell cycle regulators (such as cyclin D1 and cyclin G [35]) are also known to contain PEST domains. It is possible that gastrin could affect the steady-state levels of these proteins; indeed, such stabilization could contribute to the documented trophic effect of gastrin on ECL cells (38).
Interestingly, our results have identified an additional level at which ODC stability can be regulated. The GFP-PEST/ODC chimera used in these experiments lacks the antizyme-binding domain (27). Therefore, even though antizyme is required for the exposure of the carboxy-terminal PEST domain during normal degradation of the native ODC protein (14, 20, 28), our results indicate that there are additional factors, such as gastrin, which can regulate stability at points thereafter. Taken together with studies that demonstrate a role for gastrin response elements in transcriptional regulation (34, 49) and gastrin stimulation in the regulation of mRNA translation (33), our studies here confirm and extend evidence for multilayered control of gene expression by stimulation of G protein-coupled receptors. They also revealed a complex pattern of posttranslational regulation, with activity and stability determined by the presence or absence (through cleavage) of functional domains located within the tertiary protein sequence. Gastrin acts with an unidentified factor(s) to increase the half-lives of specific HDC isoforms. Extrapolation of these results suggests that increased histamine production, observed immediately after gastrin stimulation of ECL cells, occurs as a consequence of stabilization of HDC isoforms.
ACKNOWLEDGMENTS
We thank Ted Koh and Rocchina Colucci for help with animal experiments and useful discussion. We also thank Bill Rees at the Rowett Institute for critically reviewing the manuscript and Jeffrey Sussman (T.M.W.A.C.) for technical assistance.
T.C.W. is supported by NIH RO1 grant DK48077.
FOOTNOTES
- Received 12 November 1999.
- Returned for modification 17 January 2000.
- Accepted 17 March 2000.
- Copyright © 2000 American Society for Microbiology
