Significance of Prohormone Convertase 2, PC2, Mediated Initial Cleavage at the Proglucagon Interdomain Site, Lys70-Arg71, to Generate Glucag
http://www.100md.com
内分泌学杂志 2005年第2期
Department of Biochemistry and Molecular Biology (A.D., G.M.L., Y.R., C.N., J.S., R.C., D.F.S.) and the Howard Hughes Medical Institute (C.Z., D.F.S.), University of Chicago, Chicago, Illinois 60637
Address all correspondence and requests for reprints to: Donald F. Steiner, Department of Biochemistry and Molecular Biology, Uni-versity of Chicago, Chicago, Illinois 60637. E-mail: dfsteine@midway.uchicago.edu.
Abstract
To define the biological significance of the initial cleavage at the proglucagon (PG) interdomain site, K70-R71, we created two interdomain mutants, K70Q-R71Q and R71A. Cotransfection studies in GH4C1 cells show significant amounts of glucagon production by PC2 along with some glicentin, glicentin-related polypeptide-glucagon (GRPP-glucagon) and oxyntomodulin from wild-type PG. In contrast, a larger peptide, PG 33–158, and low amounts of GRPP-glucagon are predominantly generated from interdomain mutants. HPLC analysis shows a 5-fold increase in glucagon production by PC2 from wild-type PG and a corresponding 4-fold lower accumulation and secretion of unprocessed precursor relative to interdomain mutants. PC2 generates significant levels of glucagon from a glicentin (PG 1–69) expression plasmid, whereas PC1/3 produces only modest amounts of oxyntomodulin. Employing a major PG fragment (PG 72–158) expression plasmid, we show that PC1/3 predominantly generates glucagon-like peptide (GLP)-1, whereas PC2 produces only N-terminally extended GLP-1. Surprisingly, production of GLP-1 and GLP-2 by PC1/3 from interdomain mutants, compared with wild-type PG, is not significantly impaired. In addition to PC2 and PC1/3, PC5/6A and furin are also able to cleave the sites, K70-R71 and R107-X-R-R110 in PG. We show a much greater ability of furin to cleave the monobasic site, R77, than at the dibasic site, R124-R125, which is also weakly processed by PC5/6A, indicating overlapping specificities of these two convertases mainly with PC1/3. We propose here a trimer-like model of the spatial organization of the hormonal sequences within the PG molecule in which the accessibility to prohormone convertase action of most cleavage sites is restricted with the exception of the interdomain site, K70-R71, which is maximally accessible.
Introduction
PROGLUCAGON (PG, 18 kDa), an endocrine precursor protein, is synthesized mainly at two sites in mammals, the -cells in the pancreatic islets and the L cells in the intestinal mucosa (1, 2, 3, 4). It is also expressed in the brain and submandibular glands (5, 6, 7, 8). PG contains four major peptide segments and these are, from N to C terminus, glicentin-related polypeptide (GRPP) (PG 1–30), glucagon (PG 33–61), glucagon like peptide (GLP)-1 (PG 78–108/107-NH2), and GLP-2 (PG 126–158) (see Fig. 1). Glucagon and GLP-1 are separated by intervening peptide (IP)-1, whereas GLP-1 and GLP-2 are separated by IP-2 (see Fig. 1). PG is well conserved across many mammalian species, including man (9).
FIG. 1. Schematic representation of PG processing in GH4C1 cells and the role of prohormone convertases, PC2, PC1/3, PC5/6A, and furin. All the cleavage sites are underlined. The bold letters, K and R, represent Lys and Arg residues, respectively. The small arrowheads indicate the cleavage sites for convertases. The large and thick arrows represent the sequential steps of PG processing that generate smaller bioactive peptides.
In both islet -cells and TC1–6 cells, a mouse -cell-derived line, the major end products of PG processing are glucagon and major PG fragment (MPGF) resulting from the action of prohormone convertase (PC) 2, one of the two major neuroendocrine PCs (10), whereas in intestinal L cells, PG is cleaved predominantly by PC1/3, the other important neuroendocrine convertase (10), to generate GLP-1 and GLP-2 along with glicentin and other minor products (11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23). Glucagon and GLP-1 have opposing physiological functions (24, 25, 26, 27, 28, 29), i.e. glucagon causes hyperglycemia by stimulating two key metabolic processes, gluconeogenesis and glycogenolysis in the liver, thereby counteracting the hypoglycemic action of insulin, whereas GLP-1 enhances glucose-induced insulin secretion.
PG has many dibasic cleavage sites (K31-R32, K62-R63, K70-R71, and R124-R125), a single monobasic site, R77, and a single RXRR site, R107-X-R-R110 (see Fig. 1) that have been shown to be cleaved by PC2 and/or PC1/3 (13, 14, 15, 16, 20, 21, 22, 23). Although dibasic sites, KR and RR, and also KK and RK, present in prohormones and proneuropeptides, are generally processed by PC2 and PC1/3 (10, 30, 31), these cleavage sites located in several neuroendocrine precursors, including proenkephalin, proneurotensin/neuromedin, pro-TRH, promelanin-concentrating hormone, and chromogranin A, have been shown to be processed to some extent by either furin or PC5/6A in neuroendocrine cells (32, 33, 34, 35, 36, 37). Sites with upstream basic residues, RXX(K/R)R and R/KXRXX(K/R)R, found in many neuroendocrine precursors, including pro-GHRH, pro-SAAS, pro-7B2, and pro-PTH, have generally been shown to be cleavable by furin and PC5/6A in addition to PC1/3 or PC2 in neuroendocrine cells (38, 39, 40, 41, 42). Furin and PC5/6A have also been demonstrated to occasionally process at single basic residues, R or K, present in some prohormones/proneuropeptides such as prosomatostatin and procholecystokinin (43, 44, 45). Despite the presence of six of these typical cleavage motifs, previous studies from our laboratory using wild-type PG (expressed endogenously or exogenously) have shown that PG processing proceeds in a strict temporally ordered fashion with the initial cleavage starting at the interdomain site, K70-R71 (13, 14).
In our current studies, using the interdomain cleavage site mutants of PG, we confirm that the interdomain site, K70-R71, is first cleaved more efficiently by PC2 and PC1/3 relative to PC5/6A and furin. We also demonstrate that disruption of the interdomain site significantly impairs PC2-mediated glucagon production, whereas generation of GLP-1 and GLP-2 by PC1/3 remains unaltered. A three-dimensional model of PG is proposed that accounts well for the observed temporal order of cleavages. We have also defined the role of PC5/6A and furin in processing the cleavage sites located in the MPGF segment of PG and indicated functional redundancy of these two broadly distributed convertases with PC1/3 to generate GLP-1 and GLP-2.
Materials and Methods
Sources
Antisera to GLP-1 (2135–8, 89390–4, and 92165–3) and GLP-2 were the kind gifts of Dr. J. J. Holst (Panum Institute, University of Copenhagen, Denmark) and Dr. D. J. Drucker (Banting and Best Center, University of Toronto, Toronto, Canada), respectively. Glucagon antiserum (P7) was raised in K. Polonsky’s laboratory at the University of Chicago. The recombinant human furin expression plasmid was a kind gift from Dr. Gary Thomas (Vollum Institute, Oregon Health Sciences University, Portland, OR). Antisera to glucagon, GLP-1, and GLP-2 as well as synthetic glucagon (PG 33–61), GLP-1 (PG 78–107 amide and 72–107 amide), and GLP-2 (PG 126–159) peptides were purchased from Peninsula Laboratories Inc. (San Carlos, CA). 125I-glucagon, 125I-GLP-1, and dipeptidyl protease IV inhibitor were bought from Linco Research Inc. (St. Charles, MO).
Construction of recombinant plasmids
The wild-type hamster PG cDNA, 992 bp, was cut out from the vector, pshGlu1 (46), by DdeI restriction digestion and subcloned into the XhoI site of another vector, pCIneo Eco-, after Klenow treatment of both the insert and vector (XhoI digested). pCIneo Eco- was derived from pCIneo (Promega, Madison, WI) after EcoRI digestion, Klenow fill-in, and religation to disrupt the EcoRI site in the polylinker. Unlike glucagon and GLP-2, GLP-1 segment of PG does not have any Met residues. Thus, to enhance detection of GLP-1- and GLP-1-containing peptides derived from PG after transient transfection and metabolic pulse-chase labeling, two amino acids, Val87 and Leu91, in the wild-type PG were substituted by Met residues using site directed mutagenesis (47). We named the resulting construct, WtPG II, as also mentioned in the rest of the text. The primers used for V87M and V87M/L91M were: 5'-TTTACCAGCGATATGAGCTCTTACT-3' and 5'-ATGAGCTCTTACATGGAGGGCCAGG-3', respectively. Using WtPG II as a template, the PG interdomain cleavage site mutant, R71A, was created by site-directed mutagenesis employing a kit from Pharmacia (Uppsala, Sweden) and we call this interdomain mutant, MutPG II, as also mentioned in the subsequent text. The primer used for MutPG II was 5'-AACATTGCCAAAGCCCACGATGAGTTT-3'. The MPGF (PG 72–158) expression construct with the N-terminal signal peptide was made by a fusion PCR. For fusion, we used two PCR-generated fragments, one having the signal peptide coding sequence and a portion of the polylinker and the other containing part of the MPGF coding sequence including V87M and L91M and an EcoRI site. The fusion PCR product was gel purified, digested with NheI and EcoRI, and cloned into WtPG II expression vector already digested with the same two restriction enzymes. The primers used for the MPGF construct were 5'-TACGACTCACTATAGGCTAGCCTC-3', 5'-GTGCCTCTCAAACTCATCGTGCTGCCAGCTGCCTTGCACCAG-3', 5'-CTTTCACCAGCCAAGCAATGAATTCC-3', and 5'-CTGGTGCAAGGCAGCTGGCAGCACGATGAGTTTGAGAGGCAC-3'.
We subcloned the wild-type hamster PG cDNA into the vector, pcDNA3.1 (Invitrogen, Carlsbad, CA), creating the expression construct WtPG I, which was subsequently used as a template to generate the interdomain mutant, MutPG I, having K70, R71 changed to Q70, Q71 by multiple PCRs. The primers used were: I, 5'-AACAGGAACAACATTGCCCAACAGCACGATGAGTTTGAGAGG-3'; II, 5'-GGCGCTCTAGATGAATGCTGACACACTTA-3'; III, 5'-GCCGCGGATCCCAGCAGGCAGAAAAAAAAATG-3'; and IV, 5'-GGCAATGTTGTTCCTGTT-3'. The thermocycle used for the first PCR having the primers I+II in one set of reaction and III+IV in the other set was 94 C for 4 min followed by 30 cycles at 94 C for 30 sec, 55 C for 30 sec, 72 C for 45 sec, and finally one hold at 72 C for 7 min. The products from these two reactions were gel purified and used as templates for the second PCR using the primers II and III, and the thermocycle was similar to the above except elongation at 72 C for 2 min followed by final hold at 72 C for 10 min. The final PCR product after gel purification was digested with BamHI and XbaI and ligated into pcDNA3.1 vector digested with the same restriction enzymes. The expression plasmid for glicentin (PG 1–69) with the signal peptide at the N terminus was made by PCR using WtPG I as the template and either of the two following forward primers (without or with the Kozak sequence) and a common reverse primer. The primers were: forward primers, 5'-ACTTAAGCTTATGAAGAACATTTACATTGTG-3' and 5'-ACTTAAGCTTGCCACCATGAAGAACATTTACATTGTG-3' and reverse primer, 5'-TAGTGGATCCTCAGGCAATGTTGTTCCTGTTCCT-3'. The PCR thermocycle was 94 C for 5 min followed by 28 cycles at 94 C for 30 sec, 56 C for 30 sec, and 72 C for 1 min and finally 72 C for 7 min. The PCR products were gel purified, digested with HindIII and BamHI, and finally ligated into pcDNA3.1 digested with the same two restriction enzymes. All these expression constructs were verified by sequencing.
Cell culture, metabolic labeling, IP, and tricine-SDS-PAGE
GH4C1 cells were cultured as mentioned before (38). Cells were transiently transfected and cotransfected with either wild-type or different mutant PG expression plasmids in absence or presence of various PCs using the transfection reagent, effectene (Qiagen, Valencia, CA), following the company’s protocol. Cells were pulse labeled, in most cases with 35S-Met (Amersham Biosciences, Amersham, UK) and sometime with 3H-Leu and 3H-Val (Amersham Biosciences), and chased as previously described (48). In the chase medium, aprotinin (5 μg/ml) and dipeptidyl protease IV inhibitor (10 μg/ml) were added to inhibit degradation of PG-derived peptides. Preparation of whole-cell extracts as well as immunoprecipitation (IP) of metabolically labeled cell extracts and media using various rabbit antisera were performed as described before (48). PG and its processed peptides were resolved in Tricine-SDS-PAGE (49) under reducing conditions (Laemmli sample buffer + ?-mercaptoethanol) and fixed and fluorographed as described before (48) before being exposed to film.
Reversed phase-HPLC (RP-HPLC)
Rabbit antiglucagon antibody was used to IP metabolically labeled peptides from 2-h chase media of wild-type and interdomain mutant without or with PC2, and after IP bound peptides were extracted in 0.1 N HCl (150 μl) at room temperature for 10 min with resuspension of beads at a 2-min interval. For maximum elution, this procedure was repeated once. Each pooled sample was divided into two equal halves: half after drying in a vacuum centrifuge (Speed Vac) was resolved in Tricine-SDS-PAGE (49) under reducing conditions, and the corresponding half was injected onto a Vydac C4 RP-HPLC column. As internal standards, 2 μg each of synthetic glucagon, GLP-1, and GLP-2 as well as His-tag PG (made in our laboratory by inserting six His residues after the first two amino acids, Met and Lys, in the PG N terminus without the signal sequence) were injected. The peptides were resolved over a three-step gradient of acetonitrile as described before (13). One fourth volume of each collected fraction was measured for total incorporated radioactivity using a liquid scintillation counter (model 2200 CA, Packard Instruments, Meriden CT).
Molecular modeling
This was performed using the Insight and Discover Graphical Environments (Biosym Technologies Inc., San Diego, CA). The molecular mechanics energetic calculations employed the force field, cvff. For initial approximation of glucagon as well as GLP-1 and GLP-2, the molecular models for the active sites of PC1/3 and PC2 (50) and the recently solved x-ray structure of furin (51) as well as the x-ray structure of glucagon (52) were used. In determining and describing the angles of rotation around the C-N () and C-C' () bonds in the peptide chain, we followed the standard nomenclature (53).
Results
Mutation of the PG interdomain site significantly impairs glucagon production
To study the biological importance of the interdomain dibasic site, K70-R71, in PG processing, we created two interdomain mutants, MutPG I (K70Q-R71Q) and MutPG II (R71A), as described in detail in Materials and Methods. We used both the mutants, separately, in transient transfection/cotransfection studies using GH4C1 cells derived from rat anterior pituitary. These cells were reported to lack detectable PC2 and PC1/3 expression (41, 54, 55) as was also verified in our laboratory (data not shown). The cells were also cotransfected with either PC2 or PC1/3, or with both convertases, and subsequently metabolically pulse labeled and chased (see details in Materials and Methods). We selected only the 2-h chase period for these studies because in preliminary experiments, this time point provided an optimal assessment of the progress of PG processing in these cells in the presence of PCs. Here we describe only the results (representative of at least three experiments) employing WtPG I and MutPG I because similar findings were obtained with WtPG II and MutPG II (data not shown). Both the WtPG I and WtPG II expression constructs are described in detail in Materials and Methods.
It is noteworthy that whereas wild-type hamster PG contains two Met residues, glicentin, GRPP-glucagon, oxyntomodulin, and glucagon each contain only one Met residue. Because all the metabolic pulse-chase experiments reported here were done using 35S-Met, we considered these differences in Met content in converting band intensities of PG and peptides derived from it into quantitative estimates of cleavage products.
When WtPG I was transiently expressed without exogenous PCs in GH4C1 cells, we found cleavage only at the interdomain site, generating moderate amounts of glicentin, particularly in the 2-h chase medium (Fig. 2A, lane 4) after IP with antiglucagon antibody. We confirmed the endogenous expression of two PCs, PC5/6A and furin, in GH4C1 cells (data not shown) as also reported (41, 54). Our findings thus indicate that these two broadly distributed convertases, PC5/6A and/or furin, are most likely responsible for the observed cleavage at the interdomain site, K70-R71, in GH4C1 cells in the absence of both PC2 and PC1/3 (see below). Coexpression of PC2 more effectively processes the interdomain site, especially in 2-h chase medium, generating significantly greater amounts of glicentin as well as robust levels of GRPP-glucagon and glucagon and low amounts of oxyntomodulin (Fig. 2A, lane 7).
FIG. 2. Processing of wild-type PG vs. interdomain mutant by PC2 and/or PC1/3 in GH4C1 cells. One microgram of hamster PG expression plasmid, WtPG I (A and B) or MutPG I (C and D), was transiently transfected in absence and presence of mouse PC2 (1 μg) or mouse PC1/3 (1 μg) or both the convertases (1 + 1 μg) (see Materials and Methods for details). After 44 h of transfection, cells were pulse labeled (30 min) with 35S-Met (150 μCi /ml) and chased (2 h) as described in the text. Equal amounts of cell extracts (CE) and equal volumes of chase medium (M) were immunoprecipitated with rabbit antiglucagon antibody, and immunoprecipitated peptides were resolved in tricine-SDS-PAGE (see Materials and Methods for details). After fixation and fluorography, gels were dried and exposed to Kodak X-Omat AR-5 films (Fisher Scientific, Pittsburgh, PA). Lane 1, A–D, Location of 125I-glucagon. The positions of the prestained protein molecular-weight markers are shown on the left of each figure. P, Pulse; Ch, chase.
Analysis of the levels of precursor and processed peptides derived from it by densitometric scanning (1D Image Analysis software, Kodak, Rochester, NY) revealed that in the absence of any exogenous PCs, secreted glicentin in medium constituted a moderate 10% of total levels of peptides, whereas unprocessed PG comprised 24% in chased cell extracts and 66% in corresponding medium samples (Fig. 2A, lanes 3 and 4), indicating efficient secretion of significant amounts of uncleaved precursor. In contrast, in the presence of PC2, the total levels of glicentin secreted in the medium constituted 28%, whereas the amounts of uncleaved precursor were only 7.5% in chased cell extracts and 22% in corresponding medium samples, confirming a significantly enhanced cleavage at the interdomain site by PC2 (Fig. 2A, lanes 6 and 7). Glicentin was subsequently processed by PC2 at the sites, K31-R32 and K62-R63, either simultaneously or sequentially, to generate significant amounts of both glucagon (18.5%) and GRPP-glucagon (19%) and low amounts of oxyntomodulin (5%) (Fig. 2A, lane 7). Coexpression of PC1/3 also liberated significant amounts of glicentin after the interdomain cleavage and subsequently glicentin was processed by PC1/3 only at the site, K31-R32, to generate high levels of oxyntomodulin, especially in the 2-h chase medium (Fig. 2B, lane 4). Densitometric scanning showed that, whereas secreted glicentin constituted 40% of total levels of peptides in the presence of PC1/3, the amounts of oxyntomodulin were 14% and uncleaved precursor constituted 16% in chased cell extracts and 30% in the corresponding medium (Fig. 2B, lanes 3 and 4). It is noteworthy that production of both GRPP-glucagon and glucagon was undetected in presence of PC1/3 in GH4C1 cells, suggesting that the cleavage site, K62-R63, is exclusively processed by PC2. Coexpression of PC1/3 with PC2 did not change the levels of GRPP-glucagon or glucagon that were generated by PC2 alone (Fig. 2B, lane7 vs. Fig. 2A, lane 7) confirming the absolute necessity of PC2 in cleaving the 62K-63R site in the glicentin segment.
When MutPG I was transiently expressed in GH4C1 cells without any exogenous PCs, we found no apparent processing, either in cells or medium, and correspondingly, significant amounts of unprocessed mutant precursor were secreted into the medium (Fig. 2C, lanes 2–4). Whereas PC2 alone was coexpressed, it generated a major intermediate fragment, PG 33–158, and low levels of GRPP-glucagon from the interdomain mutant both in cells and medium (Fig. 2C, lanes 6 and 7). Also robust amounts of unprocessed mutant precursor were secreted into the medium (Fig. 2C, lane 7). We detected very low levels of glucagon generated from the interdomain mutant in the presence of PC2, after a prolonged exposure of 16 d for the chase medium samples (Fig. 2C, lane 8) relative to an exposure of 4 d for the rest of Fig. 2. The findings confirm that the interdomain cleavage is important but not absolutely required to first generate glicentin, which is a far better substrate for subsequent cleavage by PC2 to produce glucagon. PC1/3 coexpression alone with MutPG I generated mainly three intermediate peptides, PG 33–158, PG 33–123, and PG 33–76, especially in chase medium (Fig. 2D, lane 4) and PC2 and PC1/3 together generated, along with these three intermediates, also low levels of GRPP-glucagon (Fig. 2D, lane 7) and very low amounts of glucagon seen only on longer exposure (data not shown). The findings altogether confirm that PG processing normally starts at the interdomain site despite the presence of several other similar cleavage motifs. It is noteworthy that the nonconventional peptides generated from mutant PG in presence of either PC2 or PC1/3 or both were identified on the basis of their individual mobility under reducing conditions in Tricine-SDS-PAGE relative to the mobility of purified iodinated markers, prestained low-molecular-weight markers and with the help of prior knowledge on PG processing, and by using specific antibodies against glucagon, GLP-1 or GLP-2. The conventional peptides including glucagon, GLP-1, GLP-2, PGs were identified in Tricine-SDS-PAGE corresponding to the locations of iodinated marker peptides and in RP-HPLC with reference to the internal standards as shown.
Next we quantitatively measured the extent of defects in glucagon production from the interdomain mutant relative to wild-type in the presence of PC2 by RP-HPLC. After cotransfection in GH4C1 cells without or with PC2, half of the chase media was IP with rabbit antiglucagon antibody and resolved in Tricine gradient gel (Fig. 3A, lanes 2–5). The corresponding half with PC2 after IP with the same antibody was extracted in 0.1 N HCl and subjected to RP-HPLC as described in detail in Materials and Methods. Analysis after scintillation counting showed a 5-fold reduction in glucagon production from MutPG I relative to WTPG I in presence of PC2 (Fig. 3, D vs. C). Correspondingly, we found almost 4-fold higher levels of unprocessed mutant PG secreted into the chase medium relative to the wild-type (Fig. 3, D vs. C). The elution peaks for PG in the experimental samples were slightly shifted to the right with reference to the location of the purified PG internal standard. This was possibly due to the presence of the His tag in the purified PG standard that was made as described in Materials and Methods. The results from RP-HPLC analysis were in agreement with the corresponding autoradiogram (Fig. 3A, lanes 5 vs. 3), confirming our hypothesis that the first cleavage at the PG interdomain site, K70-R71, is very important for efficient production of glucagon by PC2. Specific incorporation of radioactivity (counts per minute) after metabolic pulse-chase labeling with 35S-Met was much more robust relative to combination of two tritiated amino acids and thus the fold differences between the wild-type and mutant PG samples as shown in the representative RP-HPLC figures were calculated based on 35S-Met incorporation.
FIG. 3. Quantitative estimation of impaired glucagon production from interdomain mutant relative to wild type in presence of PC2. A, GH4C1 cells were transiently transfected with WtPG I (1 μg) or MutPG I (1 μg) without or with PC2 (1 μg) for 44 h and after 30 min pulse labeling, chased for 2 h. Equal amounts of media were immunoprecipitated with rabbit antiglucagon antibody, and half of the immunoprecipitated peptides after extraction in 0.1 N HCl were resolved in Tricine-SDS-PAGE. Lane 1 indicates the location of 125I-glucagon, and the positions of the prestained protein molecular weight markers are shown on left. C and D, Each corresponding half of the immunoprecipitated peptides from 2-h chase medium of wild-type and interdomain mutant, respectively, was analyzed by RP-HPLC. B, RP-HPLC peaks for the peptides used as internal standards (also see Ref.13 for further information on HPLC standardization). 35S and 3H are denoted by and , respectively.
In immunofluorescence assays employing AtT20 cells, we found significant and comparable levels of colocalization of wild-type and interdomain mutant PGs in secretory granules and Golgi using specific markers (data not shown). The findings suggest normal intracellular transport and sorting of the mutant PGs, similar to wild-type, after synthesis.
Next, we created an expression vector for glicentin with the PG signal sequence fused to the N terminus of glicentin (see Materials and Methods). This was transfected into GH4C1 cells in the absence and presence of PC2 or PC1/3. We showed that PC2 generated significant levels of glucagon along with high levels of GRPP-glucagon and moderate amounts of oxyntomodulin, especially in chase medium (Fig. 4A, lane 7), as we found earlier (Fig. 2A, lane 7). In contrast, PC1/3 produced only oxyntomodulin similar to what we found before (Fig. 4B, lane 4 vs. Fig. 2B, lane 4), and most of the unprocessed glicentin was secreted into the medium. The findings confirm the absolute requirement for PC2 in glucagon production due to its efficient action at both the dibasic sites, K31-R32 and K62-R63, relative to the restricted ability of PC1/3 to cleave only slowly at the site, K31-R32. The results indicate that amino acids present near the dibasic site, K62-R63, confer a unique specificity for PC2 cleavage.
FIG. 4. Processing of glicentin by PC2. An expression plasmid for glicentin (1 μg) (see Materials and Methods) was transiently transfected in the absence and presence of PC2 (1 μg) or PC1/3 (1 μg) in GH4C1 cells and after pulse-chase labeling, peptides from equal amounts of both cell extracts (CE) and medium (M) were immunoprecipitated with rabbit antiglucagon antibody. Immunoprecipitates were resolved in Tricine-SDS-PAGE. Lane 1, A and B, Position of 125I-glucagon; the locations of the prestained protein molecular markers are shown on the left. P, Pulse; Ch, chase.
We examined the effects of PC5/6A or furin on PG processing using cotransfection assays because we consistently found generation of moderate levels of glicentin in GH4C1 cells after transient expression of wild-type precursor in the absence of any exogenous PCs (see Fig. 2A, lanes 2–4). Coexpression of WtPG I with either PC5/6A or furin showed considerable enhancement in glicentin levels in both chased cells and medium and correspondingly much less accumulation and secretion of uncleaved precursor, compared with transfection of PG alone (Fig. 5, lanes 5–7 and 8–10 vs. 2–4), confirming considerable susceptibility of the interdomain site, K70-R71, to PC5/6A and/or furin. It is noteworthy that glicentin was not processed further by either of these two convertases, indicating the total inability of furin or PC5/6A to cleave the K31-R32 and K62-R63 sites located in the glicentin segment. However, when we examined the processing of MutPG I in similar cotransfection assays, we found small amounts of two major intermediates, PG 33–158 and PG 33–123, indicating some susceptibility to cleavage by PC5/6A or furin of the K31-R32 and R124-R125 sites in this mutant form (data not shown).
FIG. 5. Production of glicentin from PG by PC5/6A and furin. GH4C1 cells were transiently transfected with WtPG I (1 μg) in absence and presence of mouse PC5/6A (1 μg) or human furin (1 μg) expression constructs. After pulse-chase labeling, equal amounts of cell extracts (CE) and medium (M) were immunoprecipitated with rabbit antiglucagon antibody, and immunoprecipitates were analyzed in Tricine-SDS-PAGE. Lane 1 indicates the location of 125I-glucagon, whereas the prestained protein markers are shown on the left. P, Pulse; Ch, chase.
No impairment in GLP-1 and GLP-2 processing
To compare levels of GLP-1 generated from the interdomain mutant relative to wild type by various PCs and detect GLP-1 without ambiguity after pulse-chase labeling and IP (with anti GLP-1 antibody), we generated the expression vectors, WtPG II and MutPG II (see Materials and Methods for details). Transient expression of WtPG II alone in GH4C1 cells, in absence of any exogenous PCs, generated only moderate levels of MPGF in both cell extracts and medium (Fig. 6A, lanes 2–4), confirming the ability of both furin and PC5/6A to cleave the interdomain site, K70-R71, as we found for glicentin production using the antiglucagon antibody (see Fig. 2A, lanes 2–4). When PC2 was coexpressed, MPGF and N-terminally extended GLP-1 (PG 72–108/107-NH2) were generated and secreted prominently into the chase medium (Fig. 6B, lane 4). In contrast, coexpression of PC1/3 produced both MPGF and GLP-1 (PG 78–108/107-NH2), which were secreted into the chase medium (Fig. 6C, lane 4). Employing MutPG II, we found no significant impairment in GLP-1 production and secretion in the presence of PC1/3 (Fig. 6F, lanes 3 and 4) relative to wild type, suggesting that the first cleavage at the interdomain site of PG is critical only for glucagon production. Production of comparable amounts of GLP-1 from both interdomain mutant and wild-type precursors thus indicates unperturbed ability of PC1/3 to cleave the R77 and R107-X-R-R110 sites, irrespective of the interdomain mutation. The intermediate peptides generated from the mutant precursor by PC1/3 consisted mostly of PG 33–158 and partly of PG 33–123 and 33–108, whereas PC2 was found to generate two major intermediates, PG 33–158 and 64–158, and two minor intermediates, PG 33–108 and 64–108 (Fig. 6, F vs. E). Also similar to MutPG I (Fig. 2C, lanes 2–4), we found no apparent processing of MutPG II in the absence of any exogenous PCs (Fig. 6D, lanes 2–4). The nonconventional peptides were identified as discussed before.
FIG. 6. Processing of wild-type PG vs. interdomain mutant by PC2 or PC1/3 and assessment of N-terminally extended GLP-1 and GLP-1 (PG78–108/107-NH2) production. WtPG II (1 μg) or MutPG II (1 μg) was transiently transfected in absence and presence of PC2 (1 μg) or PC1/3 (1 μg) in GH4C1 cells. Cells were pulse labeled for 30 min and chased for 2 h. Equal amounts of cell extracts (CE) and medium (M) were immunoprecipitated with rabbit anti-GLP-1 antibody (2135–8), and immunoprecipitated peptides were resolved in Tricine-SDS-PAGE. Lane 1 in all the gels represents the location of the 125I-GLP-1, and the prestained low-molecular-weight protein markers are shown on the left of each gel. P, Pulse; Ch, chase.
Next we examined the levels of GLP-2. Again we found no significant defects in the production of GLP-2 (PG126–158) from the interdomain mutant, compared with the wild-type precursor in the presence of PC1/3, and the predominantly secreted peptide in both cases was GLP-2 (Fig. 7D, lanes 3 and 4 vs. 7B, lanes 3 and 4). Also, we found equivalent levels of N-terminally extended GLP-2 (PG 111–158) in chase media from both precursors (Fig. 7D, lane 4 vs. 7B, lane 4). The findings again suggest that regardless of interdomain cleavage, PC1/3 processes both the cleavage sites, R107-X-R-R110 and R124-R125, as effectively as we have previously seen for the production of GLP-1 after cleavages at the R77 and R107-X-R-R110 sites (see Fig. 6). Coexpression of PC2 produced equivalent levels of N-terminally extended GLP-2 from both WtPG I and MutPG I (Fig. 7A, lanes 6 and 7, and 7C, lanes 6 and 7), suggesting that the ability of PC2 to cleave the R107-X-R-R110 site was not perturbed by the interdomain mutation. The results taken altogether indicate that, whereas PC1/3 effectively cleaves the three sites, R77, R107-X-R-R110, and R124-R125, PC2 completely lacks the ability to process the R77 and R124-R125 sites.
FIG. 7. Assessment of GLP-2 production from wild-type PG vs. interdomain mutant in presence of PC2 or PC1/3. GH4C1 cells were transiently transfected by WtPG I (1 μg) or MutPG I (1 μg) in absence and presence of PC2 (1 μg) or PC1/3 (1 μg). After pulse-chase labeling, equal amounts of cell extracts (CE) and medium (M) were immunoprecipitated with rabbit anti-GLP-2 antibody, and immunoprecipitates were resolved in Tricine-SDS-PAGE. The lane 1 in A–D indicates the position of the 125I-glucagon and the locations of the prestained protein markers are shown on the left.
We also studied the processing of expressed MPGF (PG 72–158) (see Materials and Methods for details) in cotransfection assays in GH4C1 cells and showed again that PC2 only generated the N-terminally extended form of GLP-1, whereas PC1/3 produced GLP-1 (Fig. 8, A-C). The findings confirm that the two MPGF cleavage motifs, R77 and R107-X-R-R110, are considerably processed by PC1/3 as shown above (see Fig. 6). However, PC2 specifically cleaved the R107-X-R-R110 site as found previously (see Fig. 7) confirming its exclusive preference for a single cleavage site relative to a broader preference of PC1/3 for all three of the sites located in the MPGF segment. In examining GLP-2 immunoreactive peptides in cotransfection assays, PC2 was found to generate only N-terminally extended GLP-2, whereas PC1/3 produced GLP-2 from MPGF (data not shown).
FIG. 8. Processing of MPGF by PC1/3. MPGF (1–71 amino acids of PG) expression plasmid (1 μg) (see Materials and Methods) was transiently transfected in GH4C1 cells in absence and presence of PC2 (1 μg) or PC1/3 (1 μg) and pulsed and chased for 30 min and 2 h, respectively. Equal amounts of cell extracts (CE) and medium (M) were immunoprecipitated with rabbit anti-GLP-1, and immunoprecipitates were resolved in Tricine-SDS-PAGE. Lane 1 in A–C represents the location of the 125I-GLP-1. The prestained protein markers are located on the left on each gel. P, Pulse; Ch, chase.
We analyzed the role of PC5/6A and furin in the generation of GLP-1 and GLP-2 using only the wild-type precursor because we failed to detect any significant defects in production of these two peptides from the interdomain mutant by PC1/3 (see Figs. 6 and 7). We showed in cotransfection assays that, whereas PC5/6A generated only considerable levels of MPGF and N-terminally extended GLP-1 in both cells and medium (Fig. 9B, lanes 2–4), furin produced, along with MPGF, both N-terminally extended GLP-1 and GLP-1 that were efficiently secreted into the medium at a ratio of 1:1 (Fig. 9C, lanes 2–4). The results indicate a relatively lower ability of furin than PC1/3 to cleave the single basic site, R77, and also suggest a complete inability of PC5/6A to process the same site. However, both furin and PC5/6A are able to cleave the interdomain site, K70-R71, as well as the site R107-X-R-R110. When we studied the effects of PC5/6A or furin on GLP-2 production, we found considerable levels of MPGF and N-terminally extended GLP-2 but only very low levels of GLP-2 (Fig. 9D, lanes 5–7, and E, lanes 2–4), suggesting much stronger activities of PC5/6A or furin at the site R107-X-R-R110, relative to the more distal cleavage site R124-R125.
FIG. 9. Processing of PG by PC5/6A or furin and generation of GLP-1 (PG78–108/107-NH2) and GLP-2 (PG126–158). GH4C1 cells were transiently transfected with wild-type PG, either WtPG II (1 μg) or WtPG I (1 μg), in absence and presence of PC5/6A (1 μg) or furin (1 μg). Cells were pulse labeled (30 min) and chased (2 h) and equal amounts of cell extracts (CE) and medium (M) were immunoprecipitated with either rabbit anti-GLP-1 (A–C) or GLP-2 (D and E). Immunoprecipitates were analyzed in Tricine-SDS-PAGE. Lane 1 in A–C and D and E represents the location of the 125I-GLP-1 and 125I-glucagon, respectively, and the positions of the prestained low-molecular-weight protein markers are shown on the left of each figure. P, Pulse; Ch, chase.
Molecular modeling of PG
Although crystallization of PG has not yet been successful, an x-ray structure of crystalline glucagon was reported almost three decades ago wherein the glucagon moiety was mainly -helical (residues 6–28) (52). However, in solution at high dilutions, glucagon does not have a highly ordered conformation (56), whereas in more concentrated solutions (at 3 mg/ml), glucagon monomers associate into trimers (57). The association of monomers to trimers stabilizes the -helical conformation of glucagon. We propose that PG in solution forms a structure analogous to the trimer in crystalline glucagon in which the three homologous segments, residues 38–59 of glucagon, 83–104 of GLP-1, and 131–152 of GLP-2, in the PG sequence form -helices. The presence of peptide connections between these homologous regions is equivalent to the concentration effect in glucagon solutions, in which the increase in concentration of glucagon leads to the trimeric arrangement of -helical glucagon molecules (56, 57). The idea that the glucagon region of PG is mainly -helical is confirmed by the observation that, despite the presence of a dibasic site in the middle of the glucagon moiety (R49-R50), proteolytic cleavage at this position by prohormone convertases occurs only at very low levels in pancreas (58). If organized in this fashion, the inner core of PG would exist in the form of a triple complex among glucagon, GLP-1, and GLP-2, analogous to the trimer in crystalline glucagon (52), whereas the intervening peptides, dividing glucagon and GLP-1, and also GLP-1 and GLP-2, form nonregular external loops (residues 62–77 and 109–125). The N-terminal peptide segment, GRPP (PG1–30), is predicted to be highly disordered [prediction based on PONDR Protein Disorder Predictor (www.pondr.com)] and lies outside this predicted PG core structure.
To model PG, we simulated the ordered docking of the -helical segments of glucagon, GLP-1, and GLP-2, respectively, in such a triple complex taking into consideration the optimal packing of hydrophobic residues on the N and C ends of the corresponding predicted -helices. A possible arrangement of the individual glucagon, GLP-1 and GLP-2 units, found by optimization of the energy of nonbonded interactions in this complex and the corresponding three hydrophobic knots maintaining their trimer-like assembly with 3-fold symmetry, is shown in Fig. 10A. In this model, the hydrophobic patch near the C terminus of glucagon interacts with a similar N-terminal hydrophobic patch of GLP-1. Similarly the C-terminal end of GLP-1 interacts with the N-terminal end of GLP-2 and so on, as expected because the hydrophobic patches near the ends of each of the predicted -helical segments in GLP-1 and GLP-2 are composed of the same or very similar bulky hydrophobic and aromatic amino acid residues as are present in glucagon. The hydrophobic knot between glucagon and GLP-1 is formed by contacts between amino acids Phe54, Trp57, and Leu58 of glucagon and Phe83, Val87, and Tyr90 of GLP-1. Similarly, the knot between GLP-1 and GLP-2 is formed by contacts of Phe99, Trp102, and Leu103 of GLP-1 with Phe131, Met135, and Ile138 of GLP-2. Because the dibasic site within glucagon moiety, R49-R50, is substituted by nonpolar amino acids in GLP-1 and GLP-2, Q94-A95 and L142-A143, respectively, electrostatic interactions within the trimer are ruled out (see Fig. 10A). In turn, these hydrophobic patches could stabilize the formation of even longer -helices within the PG structure. Given the fact that the dibasic site R49-R50, located within the -helical segment of glucagon, is cleaved very little, if at all, during biosynthesis either in intestinal L cells or islet -cells (13, 14, 16, 58), we propose that the other dibasic sites, which are not cleaved initially in PG processing, are also located on the N and C ends of -helices, which were extended to include residues 31–63, 77–107,and 124–155, at which the C end of the glucagon segment has the highest -helical potency. On this basis, the -helical content of PG is estimated to be about 50%. This figure is consistent with results of CD measurements made on recombinant human PG (Zhu, X., C. Zhang, and D. F. Steiner, unpublished data).
FIG. 10. A, Trimeric structure of PG. PG contains three homologous peptide sequences, glucagon, GLP-1, and GLP-2, which are modeled as -helices including the dibasic sites on their N and C ends. The hydrophobic knot between glucagon and GLP-1, shown in the structure of PG, is formed by the hydrophobic patches on the C end of glucagon (Phe-54, Trp-57, Leu-58) and the N end of GLP-1 (Phe-83, Val-87, and Tyr-90). B, IP-1 forms an external loop that acts as a docking site for PCs. The IP-1, 64–76 amino acids of PG, is modeled as an external loop that corresponds to a conformation required for the formation of a complex with the active site of PCs. Here the sequence, IAKRHD, residues 68–73 of IP-1, corresponding to the docking site of furin, is shown by space-filled images. C, A modeled spatial structure for PG.
After definition of the -helical segments in PG, we also considered the specific stereochemical requirements for a polypeptide chain to enter an -helix at the N end and to exit from it at the C end (59). Accordingly, amino acids 63–65 (Arg-Asn-Arg) at the C end of the glucagon moiety have been modeled in a specific conformation corresponding to the parallel exit of the polypeptide chain from the C ends of -helices, ? (59). Likewise, amino acids 76–78 (Glu-Arg-His) preceding -helix of GLP-1 segment have been modeled as a parallel entry at the N end of this -helix, ? (59).
Recently the crystal structures of two proprotein convertases, furin and kexin, have been solved (51, 60). The catalytic domain of furin has been shown to be very similar to subtilisin and related convertases in overall structural organization. Like the subtilisins (61), the active site groove of each convertase can bind at least six residues, P4-P2' of a substrate in an extended conformation. In reconstructing a possible conformational variant of the external loop of IP-1 (PG 66–75) that contains the initial cleavage site K70-R71, we considered only those variants that would facilitate formation of an enzyme-substrate complex with the substrate binding groove of furin as a model PC. Because a productive convertase interaction typically requires an extended peptide conformation for residues P4-P2' of a substrate, the IP-1 sequence, IAKRHD (68–73 amino acids), corresponding to P4-P2' positions has been arranged in such an energetically optimized conformation to fit the active site of furin. Its location in the groove corresponds to that of the peptide inhibitor of furin (51). The peptide bond between R71 and H72 also is arranged in an orientation appropriate for catalytic attack by the side chain of S368 of furin. The angles of rotation (C-N) and (C-C’) around the main chain bonds for residues IAKRHD are found to be –77,114; –88,124; –78,128; –57,75; –114,170; –105,–114. At the same time, the segment IAKRHD forms the top of the external loop of IP-1. In this situation, in which residues P4-P2', IAKRHD, are arranged in the active site, two dipeptide units, Asn66-Asn67 and Glu74-Phe75 of IP-1, which connect the top hexapeptide with glucagon and GLP-1, act as spacers to prevent any van der Waals overlap between the volume of the glucagon-GLP-1 moiety and the external surface of the convertase. The proposed docking interaction is shown in Fig. 10B, in which an unobtrusive conformation of IP-1 relative to the convertase is achieved by formation of inner half-turns by the above-mentioned residues, Asn-Asn and Glu-Phe on either side of the bound peptide segment 68–73 (P4-P2'). It directs the remaining PG structure away from the convertase surface. The volume of the catalytic domain of the prohormone convertase is modeled according to the catalytic domain of furin (51).
We predict that IP-2 (PG 111–123), connecting GLP-1 and GLP-2, also contains a short -helical segment, PEEVAIV (PG 113–119), because Glu residues have a high potential for initiation of -helices, and Pro residues are often located at their N ends (53). The flexible amino acids, Asp111 and Gly123, at the N and C ends of this external loop of IP2 help to form the exit from the -helix formed by the GLP-1 domain and the entry into the -helix formed by the GLP-2 domain, respectively. Then side chains of the -helical residues, VAIV at the top of the IP-2 loop, may contribute to the hydrophobic core of PG.
Employing the proposed initial approximations for the glucagon, GLP-1 and GLP-2, moieties as well as for the two connecting peptides, IP-1 and IP-2, we carried out an optimization of the inner potential energy of the PG structure, excluding the extraneous highly disordered N-terminal peptide segment, PG 1–30, and derived the model as projected in Fig. 10C. In this proposed structural motif, all dibasic sites are screened by intramolecular van der Waals contacts with neighboring amino acids except for the interdomain site, K70-R71, which is located on the top of the external loop formed by IP-1 and, due to its high accessibility, easily adopts a conformation appropriate for binding with the active sites of prohormone convertases. These considerations thus predict cleavage of the interdomain site as the first step in the temporally ordered processing of intact PG.
After the cleavage of the dibasic site K70-R71 of PG, the newly formed glicentin moiety is processed to glucagon in -cells, which express PC2 at high levels (13). In the model, both sites K31-R32 and K62-R63, which are located at the ends of the single -helix in the glucagon segment, now are open and likely have similar stereochemical possibilities for interactions with PC2. Unexpectedly, in AtT-20 cells, expressing a high level of PC1/3, the newly formed MPGF was cleaved first at the monobasic site R77 with formation of truncated MPGF and only after that, at the R109-R110 site, leading to the formation of GLP-1 (16). Therefore, the N-terminal segment of MPGF up to and including R77, is most likely in an unfolded conformation, which is more readily accessible for interactions with the active site of PC1/3, than the inner dibasic site, R109-R110, in IP2. This region, separating two -helical segments of MPGF (GLP-1 and GLP-2) is still screened by nonbonded interactions inside an antiparallel helix-turn-helix hairpin, which in turn, can be stronger, than in the PG structure itself. This residual structure in free MPGF may therefore account for the lack of any apparent effect of inhibition of interdomain cleavage on the processing of this half of the PG molecule.
Discussion
We conclude from our current studies that despite the presence of several cleavage sites (see Fig. 1), PG processing mediated by PCs including PC2, PC1/3, PC5/6A, and furin begins at the interdomain site, K70-R71, generating two major peptides, glicentin and MPGF. Glicentin is then processed by PC2 to generate glucagon. The exclusive role of PC2 in generating glucagon was illustrated in our previous studies using PC2 null animals and an -cell line (17, 19, 62) and also in mice null for 7B2 (63), a protein required for activation of PC2. For glucagon production, glicentin must be processed at its two cleavage sites, K31-R32 and K62-R63. In the current studies, we have shown that the site K62-R63 is exclusively processed by PC2, whereas the K31-R32 site is cleaved by both PC2 and PC1/3. Employing interdomain cleavage site mutants, we find production of only very low amounts of glucagon (5-fold less) by PC2 relative to wild type. However, PC2 efficiently cleaves the K31-R32 site generating considerable levels of PG 33–158 from mutant precursor. These results indicate that some kind of steric hindrance strongly masks the K62-R63 site in intact PG but not in glicentin. Contrary to PC2, PC1/3 is unable to cleave the K62-R63 site located in either PG or glicentin, and this is in agreement with our studies done in TC-PC2 cells that contain only low levels of PC1/3 (19). The findings are also consistent with studies on PC1/3 null mice in which glucagon production is unchanged from that in wild-type animals (18). The results taken altogether confirm the absolute requirement of PC2 for glucagon production but are in contrast with the findings of a few other investigators (21, 22, 23) who were either unable to adequately substantiate the presence of enzymatically active PC2 in vaccinia virus preparations or failed to demonstrate loss of PC2 enzymatic activities in antisense studies.
When we examined the levels of GLP-1 and GLP-2 produced from interdomain mutant relative to wild type in presence of PC1/3, we found comparable amounts of these peptides. These results may indicate a lack of any significant steric hindrance for PC1/3 in accessing the three C-terminal cleavage sites after the interdomain site, R77, R107-X-R-R110, and R124-R125, despite the interdomain mutation. On the other hand, they could also be consistent with the persistence of a comparable level of steric hindrance in the free MPGF domain after its release from PG as suggested by the modeling experiments. Likewise, the generation of equivalent amounts of N-terminally extended GLP-2 by PC2 from both interdomain mutant and wild-type may be on a similar basis. Expressing either intact PG or MPGF, we confirmed that GLP-1 is generated by PC1/3, whereas PC2 produces only the N-terminally extended form of GLP-1. The results are in agreement with previous in vitro and in vivo studies implying an important role for PC1/3 in generating GLP-1 and GLP-2 (14, 16, 18, 22, 23, 64) and PC2 in producing the N-terminally extended forms of GLP-1 and GLP-2 (13, 19). The ability of PC1/3 to process a single basic site without a P2 or P4 basic residue has been shown for a few other neuroendocrine precursors (48, 65). However, in vitro cell-free studies employing PC1/3 showed either lack of conversion or very low levels of processing of N-terminally extended GLP-1 (PG 72–108) into GLP-1 (PG 78–108) (14, 16), indicating weaker ability of PC1/3 to cleave the monobasic site, R77. In a living human subject lacking active PC1/3, both GLP-1 and GLP-2 are found in plasma (66). Also low levels of GLP-2 and GLP-1 were found in intestinal tissue extracts of PC1/3 null mice (18). Whereas expression of PC2 has not been detected immunohistochemically in any regions of the small intestine (18), PC5/6A is widely expressed in the gut (67, 68), and furin, the ubiquitously expressed convertase (69), is coexpressed with PC1/3 and PC5/6A in small intestine (70). The results altogether suggest the existence of overlapping specificities among PC5/6A, furin, and PC1/3 to generate GLP-1 and GLP-2.
We have consistently seen production of some glicentin and MPGF from transiently expressed PG without any exogenous PCs in GH4C1 cells, indicating that either PC5/6A or furin or both are responsible for the observed moderate levels of cleavage at the interdomain site, as also suggested by others (14, 15, 16, 18, 22). Coexpression of PC5/6A or furin generated greater amounts of glicentin and MPGF from PG. Remarkably, neither PC5/6A nor furin appeared to process glicentin further (see Fig. 5), indicating the inability of these two broadly distributed convertases to cleave the sites K31-R32 and K62-R63. The most likely explanation might be lack of upstream basic residues or other features of local amino acid sequence near the two cleavage sites. The findings also explain the lack of any detectable ability of PC5/6A to liberate any glucagon despite a report of its colocalization with glucagon in secretory granules of pancreatic islets (71). In contrast, MPGF was efficiently cleaved by PC5/6A or furin under similar experimental conditions. We found in cotransfection assays that PC5/6A generated considerable levels of both N-terminally extended GLP-1 and GLP-2, whereas furin produced, along with N-terminally extended GLP-1 and GLP-2, also considerable amounts of GLP-1 (PG 78–108/107-NH2). However, the ability to produce GLP-2 (PG 126–158) by either PC5/6A or furin was rather weak, compared with PC1/3. In the current studies, we confirmed the ability of furin to cleave the single basic site R77 and the R107-X-R-R110 site, which is also processed by PC5/6A to generate GLP-1. Furin was previously shown to cleave a single Arg residue in prosomatostatin to produce SS-28 (45). Also in studies of PC2 knockout mouse pancreatic extracts, SS-28 was found as the major end product of prosomatostatin processing (62), although islet -cells were shown to lack PC1/3 expression (72). We suggest, based on these results, that efficient normal production of both GLP-1 and GLP-2 from PG requires redundant functions of PC1/3, furin, and PC5/6A. In this context, conditional knockouts for furin and/or PC5/6A in intestinal L cells would be of great biological interest. However, expression of both PC2 and PC1/3 in canine L cells was reported (73). In our laboratory, we were unable to detect any PC2 expression in mouse intestine by RT-PCR (Zhu, X., and C. Zhang, unpublished data), and these findings were confirmed in a recent independent study (18).
The remarkable feature of PG processing, irrespective of its sites of synthesis, is that the interdomain site, K70-R71, located in the middle of the IP-1 sequence, PG 62–77, is cleaved first despite the presence of three other dibasic sites, KR, a single monobasic site, R, and one R-X-R-R site (see Fig. 1) as previously predicted in a study from our laboratory (13) and also shown in the current studies. The findings suggest the possibility that in native PG, all the cleavage sites except the interdomain site are screened by intramolecular contacts, whereas the interdomain site, located at the top of the external loop connecting glucagon and GLP-1, is more available to interact with the active site of PCs including PC2, PC1/3, PC5/6A, and furin. In our model of the spatial structure of PG, IP-1, PG 62–77, is an external loop in an optimal conformation for interacting with the active sites of PCs in general (50, 51) (see Fig. 10, B and C). Cleavage at this site allows the molecule to open, giving rise to glicentin and MPGF and thus setting the stage for the next processing steps involving either PC2 mediated cleavage of glicentin to generate primarily glucagon in the pancreatic -cells or PC1/3, PC5/6A, and furin assisted processing of MPGF to produce predominantly GLP-1 and GLP-2 in intestinal L cells.
Acknowledgments
We thank Paul Gardner for making oligonucleotide primers, Xiaorong Zhu for providing the His-tag PG, and Rama Boddipalli for his technical assistance in running the Kodak 1D Image Analysis software. We also thank Rosie Ricks for her expert secretarial assistance.
References
Patzelt C, Tager HS, Carroll RJ, Steiner DF 1979 Identification and processing of proglucagon in pancreatic islets. Nature 282:260–266
Mojsov S, Heinrich G, Wilson IB, Ravazzola M, Orci L, Habener JF 1986 Proglucagon gene expression in pancreas and intestine diversifies at the level of post-translational processing. J Biol Chem 261:11880–11889
Novak U, Wilks A, Buell G, McEwen S 1987 Identical mRNA for preproglucagon in pancreas and gut. Eur J Biochem 164:553–558
Orskov C, Holst JJ, Poulsen SS, Kirkegaard P 1987 Pancreatic and intestinal processing of proglucagon in man. Diabetologia 30:874–881
Larsen PJ, Tang-Christensen M, Holst JJ, Orskov C 1997 Distribution of glucagon-like peptide-1 and other preproglucagon-derived peptides in the rat hypothalamus and brainstem. Neuroscience 77:257–270
Lui EY, Asa SL, Drucker DJ, Lee YC, Brubaker PL 1990 Glucagon and related peptides in fetal rat hypothalamus in vivo and in vitro. Endocrinology 126:110–117
Tang-Christensen M, Larsen PJ, Thulesen J, Romer J, Vrang N 2000 The proglucagon-derived peptide, glucagon-like peptide-2, is a neurotransmitter involved in the regulation of food intake. Nat Med 6:802–807
Egea JC, Hirtz C, Deville de Periere D 2003 Preproglucagon mRNA expression in adult rat submandibular glands. Diabetes Nutr Metab 16:130–133
Seino S, Welsh M, Bell GI, Chan SJ, Steiner DF 1986 Mutations in the guinea pig preproglucagon gene are restricted to a specific portion of the prohormone sequence. FEBS Lett 203:25–30
Zhou A, Webb G, Zhu X, Steiner DF 1999 Proteolytic processing in the secretory pathway. J Biol Chem 274:20745–20748[Free Full Text]
Bromer WW 1972 In: Steiner DF, Freinkel N, eds. Chemistry of glucagon and gastrin. Handbook of physiology, Section 7: Endocrinology, Vol 1, Endocrine Pancreas. Washington, DC: American Physiological Society Chap 7:133–138
Orskov C 1992 Glucagon-like peptide-1, a new hormone of the entero-insular axis. Diabetologia 35:701–711
Rouille Y, Westermark G, Martin SK, Steiner DF 1994 Proglucagon is processed to glucagon by prohormone convertase 2 in TC1–6 cells. Proc Natl Acad Sci USA 91:3242–3246
Rouille Y, Martin S, Steiner DF 1995 Differential processing of proglucagon by the subtilisin-like prohormone convertases PC2 and PC3 to generate either glucagon or glucagon-like peptide. J Biol Chem 270:26488–26496
Rouille Y, Bianchi M, Irminger JC, Halban PA 1997 Role of the prohormone convertase PC2 in the processing of proglucagon to glucagon. FEBS Lett 413:119–123
Rouille Y, Kantengwa S, Irminger JC, Halban PA 1997 Role of the prohormone convertase PC3 in the processing of proglucagon to glucagon-like peptide-1. J Biol Chem 272:32810–32816
Furuta M, Zhou A, Webb G, Carroll R, Ravazzola M, Orci L, Steiner DF 2001 Severe defect in proglucagon processing in islet A-cells of prohormone convertase 2 null mice. J Biol Chem 276:27197–27202
Ugleholdt R, Zhu X, Deacon CF, Orskov C, Steiner DF, Holst JJ 2004 Impaired intestinal proglucagon processing in mice lacking prohormone convertase 1. Endocrinology 145:1349–1355
Webb GC, Dey A, Wang J, Stein J, Milewski M, Steiner DF 2004 Altered proglucagon processing in an -cell line derived from prohormone convertase 2 null mouse islets. J Biol Chem 279:31068–31075
Mineo I, Matsumura T, Shingu R, Mamba M, Matsuzawa Y 1995 The role of prohormone convertases PC1(PC3) and PC2 in the cell-specific processing of glucagon. Biochem Biophys Res Commun 207:646–651
Dhanvantari S, Brubaker PL 1998 Proglucagon processing in an islet cell line: effects of PC1 overexpression and PC2 depletion. Endocrinology 139:1630–1637
Dhanvantari S, Seidah NG, Brubaker PL 1996 Role of prohormone convertases in the tissue-specific processing of proglucagon. Mol Endocrinol 10: 342–355
Rothenberg ME, Eilertson CD, Klein K, Mackin RB, Noe BD 1996 Evidence for redundancy in propeptide/prohormone convertase activities in processing proglucagon: an antisense study. Mol Endocrinol 10:331–341
Mojsov S, Weir GC, Habener JF 1987 Insulinotropin: glucagon-like peptide-1 (7–37) co-encoded in the glucagon gene is a potent stimulator of insulin release in the perfused rat pancreas. J Clin Invest 79:616–619
Drucker DJ, Philippe J, Mojsov S, Chick WL, Habener JF 1987 Glucagon-like peptide-1 stimulates insulin gene expression and increases cyclic AMP levels in a rat islet cell line. Proc Natl Acad Sci USA 84:3434–3438
Kreymann B, Ghatei MA, Williams G, Bloom SR 1987 Glucagon-like peptide-1 7–36: a physiological incretin in man. Lancet 2:1300–1304
Orskov C, Holst JJ, Nielsen OV 1988 Effect of truncated glucagon-like peptide-1 [proglucagon-(78–107)amide] on endocrine secretion from pig pancreas, antrum, and nonantral stomach. Endocrinology 123:2009–2013
Orskov C, Bersani M, Johnsen AH, Hojrup P, Holst JJ 1989 Complete sequences of glucagon-like peptide-1 from human and pig small intestine. J Biol Chem 264:12826–12829
Drucker DJ 2001 Glucagon-like peptide-2. J Clin Endocrinol Metab 86:1759–1764
Cameron A, Apletalina EV, Lindberg I 2001 The enzymology of PC1 and PC2. In: Dalbey RE, Sigman, DS, eds. The enzymes. 3rd ed. San Diego: Academic Press; 291–328
Seidah NG, Chretien M 1999 Proprotein and prohormone convertases: a family of subtilases generating diverse bioactive polypeptides. Brain Res 848:45–62
Breslin MB, Lindberg I, Benjannet S, Mathis JP, Lazure C, Seidah NG 1993 Differential processing of proenkephalin by prohormone convertases 1(3) and 2 and furin. J Biol Chem 268:27084–27093
Barbero P, Rovere C, De Bie I, Seidah NG, Beaudet A, Kitabgi, P 1998 PC5-A-mediated processing of pro-neurotensin in early compartments of the regulated secretory pathway of PC5-transfected PC12 cells. J Biol Chem 273:25339–25346
Villeneuve P, Lafortune L, Seidah NG, Kitabgi P, Beaudet A 2000 Immunohistochemical evidence for the involvement of protein convertases 5A and 2 in the processing of pro-neurotensin in rat brain. J Comp Neurol 424:461–475
Schaner P, Todd RB, Siedah NG, Nillni EA 1997 Processing of prothyrotropin-releasing hormone by the family of prohormone convertases. J Biol Chem 272:19958–19968
Viale A, Ortola C, Hervieu G, Furuta M, Barbero P, Steiner DF, Seidah NG, Nahon JL 1999 Cellular localization and role of prohormone convertases in the processing of pro-melanin concentrating hormone in mammals. J Biol Chem 274:6536–6545
Eskeland NL, Zhou A, Dinh TQ, Wu H, Parmer RJ, Mains RE, O’Connor DT 1996 Chromogranin A processing and secretion: specific role of endogenous and exogenous prohormone convertases in the regulated secretory pathway. J Clin Invest 98:148–156
Dey A, Norrbom C, Zhu X, Stein J, Zhang C, Ueda K, Steiner DF 2004 Furin and prohormone convertase 1/3 are major convertases in the processing of mouse pro-growth hormone-releasing hormone. Endocrinology 145:1961–1971
Sayah M, Fortenberry Y, Cameron A, Lindberg I 2001 Tissue distribution and processing of proSAAS by proprotein convertases. J Neurochem 76:1833–1841
Mzhavia N, Qian Y, Feng Y, Che FY, Devi LA, Fricker LD 2002 Processing of proSAAS in neuroendocrine cell lines. Biochem J 361:67–76
Paquet L, Bergeron F, Boudreault A, Seidah NG, Chretien M, Mbikay M, Lazure C 1994 The neuroendocrine precursor 7B2 is a sulfated protein proteolytically processed by a ubiquitous furin-like convertase. J Biol Chem 269:19279–19285
Hendy GN, Bennett HPJ, Gibbs BF, Lazure C, Day R, Seidah NG 1995 Proparathyroid hormone is preferentially cleaved to parathyroid hormone by the prohormone convertase furin. A mass spectrometric study. J Biol Chem 270:9517–9525
Cain BM, Vishnuvardhan D, Beinfeld MC 2001 Neoronal cell lines expressing PC5, but not PC1 or PC2, process pro-CCK into glycine-extended CCK 12 and 22. Peptides 22:1271–1277
Cain BM, Vishnuvardhan D, Wang W, Foulon T, Cadel S, Cohen P, Beinfeld MC 2002 Production, purification and characterization of recombinant prohormone convertase 5 from bacuolovirus-infected insect cells. Protein Expr Purif 24:227–233
Galanopoulou AS, Seidah NG, Patel YC 1995 Direct role of furin in mammalian prosomatostatin processing. Biochem J 309:33–40
Bell GI, Santerre RF, Mullenbach GT 1983 Hamster proglucagon contains the sequence of glucagon and two related peptides. Nature 302:716–718
Deng WP, Nickoloff JA 1992 Site-directed mutagenesis of virtually any plasmid by eliminating a unique site. Anal Biochem 200:81–88
Dey A, Zhu X, Carroll R, Turck CW, Stein J, Steiner DF 2003 Biological processing of the cocaine and amphetamine-regulated transcript precursors by prohormone convertases, PC2 and PC1/3. J Biol Chem 278:15007–15014
Schagger H, Jagow GV 1987 Tricine-sodium dodecyl sulfate-polyacrylamide gel electrophoresis for the separation of proteins in the range from 1 to 100 kDa. Anal Biochem 166:368–379
Lipkind G, Gong Q, Steiner DF 1995 Molecular modeling of the substrate specificity of prohormone convertases SPC2 and SPC3. J Biol Chem 270:13277–13284
Henrich S, Cameron A, Bourenkov GP, Kiefersauer R, Huber R, Lindberg I, Bode W, Than ME 2003 The crystal structure of the proprotein processing proteinase furin explains its stringent specificity. Nat Struct Biol 10:520–526
Sasaki K, Dockerill S, Adamiak DA, Tickle IJ, Blundell T 1975 X-ray analysis of glucagon and its relationship to receptor biding. Nature 257:751–757
Creighton TE 1993 Proteins: structures and molecular properties. 2nd ed. New York: WH Freeman and Co.
Seidah NG, Chretien M, Day R 1994 The family of subtilisin/kexin like pro-protein and pro-hormone convertases: divergent or shared functions. Biochimie (Paris) 76:197–209
Posner SF, Vaslet CA, Jurofcik M, Lee A, Seidah NG, Nillni EA 2004 Stepwise posttranslational processing of progrowth hormone-releasing hormone (pro GHRH) polypeptide by furin and PC1. Endocrine 23:199–213
Pohl SL, Birnbaumer L, Rodbell M 1969 Glucagon-sensitive adenyl cyclase in plasma membrane of hepatic parenchymal cells. Science 164:566–569
Panijpan B, Gratzer WB 1974 Conformational nature of monomeric glucagon. Eur J Biochem 45:547–553
Dalle S, Fontes G, Lajoix AD, LeBrigand L, Gross R, Ribes G, Dufour M, Barry L, LeNguyen D, Bataille D 2002 Miniglucagon (glucagon 19–29), a novel regulator of the pancreatic islet physiology. Diabetes 51:406–412
Efimov AV 1993 Standard structures in proteins. Prog Biophys Mol Biol 60:201–239
Holyoak T, Wilson MA, Fenn TD, Kettner CA, Petsko GA, Fuller RS, Ringe D 2003 A resolution crystal structure of the prototypical hormone-processing protease Kex2 in complex with an Ala-Lys-Arg boronic acid inhibitor. Biochemistry 42:6709–6718
McPhalen CA, James MN 1988 Structural comparison of two serine proteinase-protein inhibitor complexes: eglin-c-subtilisin Carlsberg and CI-2-subtilisin Novo. Biochemistry 27:6582–6598
Furuta M, Yano H, Zhou A, Rouille Y, Holst JJ, Carroll R, Ravazzola M, Orci L, Furuta H, Steiner DF 1997 Defective prohormone processing and altered pancreatic islet morphology in mice lacking active SPC2. Proc Natl Acad Sci USA 94:6646–6651
Westphal CH, Muller L, Zhou A, Zhu X, Bonner-Weir S, Schambelan M, Steiner DF, Lindberg I, Leder P 1999 The neuroendocrine protein 7B2 is required for peptide hormone processing in vivo and provides a novel mechanism for pituitary Cushing’s disease. Cell 96:689–700
Zhu X, Zhou A, Dey A, Norrbom C, Carroll R, Zhang C, Laurent V, Lindberg I, Ugleholdt R, Holst JJ, Steiner DF 2002 Disruption of PC1/3 expression in mice causes dwarfism and multiple neuroendocrine peptide processing defects. Proc Natl Acad Sci USA 99:10293–10298
Dupuy A, Lindberg I, Zhou A, Akil H, Lazure C, Chretien M, Seidah NG, Day R 1994 Processing of prodynorphin by the prohormone convertase PC1 results in high molecular weight intermediate forms. FEBS Lett 337:60–65
Jackson RS, Creemers JWM, Farooqi S, Raffin-Sanson ML, Varro A, Dockray GJ, Holst JJ, Brubaker PL, Corvol P, Polonsky KS, Ostrega D, Becker KL, Bertagna X, Hutton JC, White A, Dattani MT, Hussain K, Middleton SJ, Nicole TM, Milla PJ, Lindley KJ, O’Rahilly S 2003 Small-intestinal dysfunction accompanies the complex endocrinopathy of human proprotein convertase 1 deficiency. J Clin Invest 112:1550–1560
Lusson J, Vieau D, Hamelin J, Day R, Chretien M, Seidah NG 1993 cDNA structure of the mouse and subtilisin/kexin-like PC5: a candidate proprotein convertase expressed in endocrine and neuroendocrine cells. Proc Natl Acad Sci USA 90:6691–6695
Nakagawa T, Hosaka M, Torii S, Watanabe T, Murakami K, Nakayama K 1993 Identification and functional expression of a new member of the mammalian kex2-like processing endoprotease family: its striking structural similarity to PACE4. J Biochem 113:132–135
Thomas G 2002 Furin at the cutting edge: from protein traffic to embryogenesis and disease. Nat Rev Mol Cell Biol 3:753–766
Keller P, Zecca L, Boukamel R, Zwicker E, Gloor S, Semenza G 1995 Furin, PC1/3, and/or PC6A process rabbit, but not human, pro-lactase-phlorizin hydrolase to the 180-kDa intermediate. J Biol Chem 270:25722–25728
De Bie I, Marcinkiewicz M, Malide D, Lazure C, Nakayama K, Bendayan M, Seidah NG 1996 The isoforms of proprotein convertase PC5 are sorted to different subcellular compartments. J Cell Biol 135:1261–1275
Tanaka S, Kurabuchi S, Mochida H, Kato T, Takahashi S, Watanabe T, Nakayama K 1996 Immunocytochemical localization of prohormone convertases PC1/PC3 and PC2 in rat pancreatic islets. Arch Histol Cytol 59:261–271
Damholt AB, Buchan AM, Holst JJ, Kofod H 1999 Proglucagon processing profile in canine L cells expressing endogenous prohormone convertase 1/3 and prohormone convertase 2. Endocrinology 140:4800–4888(Arunangsu Dey, Gregory M.)
Address all correspondence and requests for reprints to: Donald F. Steiner, Department of Biochemistry and Molecular Biology, Uni-versity of Chicago, Chicago, Illinois 60637. E-mail: dfsteine@midway.uchicago.edu.
Abstract
To define the biological significance of the initial cleavage at the proglucagon (PG) interdomain site, K70-R71, we created two interdomain mutants, K70Q-R71Q and R71A. Cotransfection studies in GH4C1 cells show significant amounts of glucagon production by PC2 along with some glicentin, glicentin-related polypeptide-glucagon (GRPP-glucagon) and oxyntomodulin from wild-type PG. In contrast, a larger peptide, PG 33–158, and low amounts of GRPP-glucagon are predominantly generated from interdomain mutants. HPLC analysis shows a 5-fold increase in glucagon production by PC2 from wild-type PG and a corresponding 4-fold lower accumulation and secretion of unprocessed precursor relative to interdomain mutants. PC2 generates significant levels of glucagon from a glicentin (PG 1–69) expression plasmid, whereas PC1/3 produces only modest amounts of oxyntomodulin. Employing a major PG fragment (PG 72–158) expression plasmid, we show that PC1/3 predominantly generates glucagon-like peptide (GLP)-1, whereas PC2 produces only N-terminally extended GLP-1. Surprisingly, production of GLP-1 and GLP-2 by PC1/3 from interdomain mutants, compared with wild-type PG, is not significantly impaired. In addition to PC2 and PC1/3, PC5/6A and furin are also able to cleave the sites, K70-R71 and R107-X-R-R110 in PG. We show a much greater ability of furin to cleave the monobasic site, R77, than at the dibasic site, R124-R125, which is also weakly processed by PC5/6A, indicating overlapping specificities of these two convertases mainly with PC1/3. We propose here a trimer-like model of the spatial organization of the hormonal sequences within the PG molecule in which the accessibility to prohormone convertase action of most cleavage sites is restricted with the exception of the interdomain site, K70-R71, which is maximally accessible.
Introduction
PROGLUCAGON (PG, 18 kDa), an endocrine precursor protein, is synthesized mainly at two sites in mammals, the -cells in the pancreatic islets and the L cells in the intestinal mucosa (1, 2, 3, 4). It is also expressed in the brain and submandibular glands (5, 6, 7, 8). PG contains four major peptide segments and these are, from N to C terminus, glicentin-related polypeptide (GRPP) (PG 1–30), glucagon (PG 33–61), glucagon like peptide (GLP)-1 (PG 78–108/107-NH2), and GLP-2 (PG 126–158) (see Fig. 1). Glucagon and GLP-1 are separated by intervening peptide (IP)-1, whereas GLP-1 and GLP-2 are separated by IP-2 (see Fig. 1). PG is well conserved across many mammalian species, including man (9).
FIG. 1. Schematic representation of PG processing in GH4C1 cells and the role of prohormone convertases, PC2, PC1/3, PC5/6A, and furin. All the cleavage sites are underlined. The bold letters, K and R, represent Lys and Arg residues, respectively. The small arrowheads indicate the cleavage sites for convertases. The large and thick arrows represent the sequential steps of PG processing that generate smaller bioactive peptides.
In both islet -cells and TC1–6 cells, a mouse -cell-derived line, the major end products of PG processing are glucagon and major PG fragment (MPGF) resulting from the action of prohormone convertase (PC) 2, one of the two major neuroendocrine PCs (10), whereas in intestinal L cells, PG is cleaved predominantly by PC1/3, the other important neuroendocrine convertase (10), to generate GLP-1 and GLP-2 along with glicentin and other minor products (11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23). Glucagon and GLP-1 have opposing physiological functions (24, 25, 26, 27, 28, 29), i.e. glucagon causes hyperglycemia by stimulating two key metabolic processes, gluconeogenesis and glycogenolysis in the liver, thereby counteracting the hypoglycemic action of insulin, whereas GLP-1 enhances glucose-induced insulin secretion.
PG has many dibasic cleavage sites (K31-R32, K62-R63, K70-R71, and R124-R125), a single monobasic site, R77, and a single RXRR site, R107-X-R-R110 (see Fig. 1) that have been shown to be cleaved by PC2 and/or PC1/3 (13, 14, 15, 16, 20, 21, 22, 23). Although dibasic sites, KR and RR, and also KK and RK, present in prohormones and proneuropeptides, are generally processed by PC2 and PC1/3 (10, 30, 31), these cleavage sites located in several neuroendocrine precursors, including proenkephalin, proneurotensin/neuromedin, pro-TRH, promelanin-concentrating hormone, and chromogranin A, have been shown to be processed to some extent by either furin or PC5/6A in neuroendocrine cells (32, 33, 34, 35, 36, 37). Sites with upstream basic residues, RXX(K/R)R and R/KXRXX(K/R)R, found in many neuroendocrine precursors, including pro-GHRH, pro-SAAS, pro-7B2, and pro-PTH, have generally been shown to be cleavable by furin and PC5/6A in addition to PC1/3 or PC2 in neuroendocrine cells (38, 39, 40, 41, 42). Furin and PC5/6A have also been demonstrated to occasionally process at single basic residues, R or K, present in some prohormones/proneuropeptides such as prosomatostatin and procholecystokinin (43, 44, 45). Despite the presence of six of these typical cleavage motifs, previous studies from our laboratory using wild-type PG (expressed endogenously or exogenously) have shown that PG processing proceeds in a strict temporally ordered fashion with the initial cleavage starting at the interdomain site, K70-R71 (13, 14).
In our current studies, using the interdomain cleavage site mutants of PG, we confirm that the interdomain site, K70-R71, is first cleaved more efficiently by PC2 and PC1/3 relative to PC5/6A and furin. We also demonstrate that disruption of the interdomain site significantly impairs PC2-mediated glucagon production, whereas generation of GLP-1 and GLP-2 by PC1/3 remains unaltered. A three-dimensional model of PG is proposed that accounts well for the observed temporal order of cleavages. We have also defined the role of PC5/6A and furin in processing the cleavage sites located in the MPGF segment of PG and indicated functional redundancy of these two broadly distributed convertases with PC1/3 to generate GLP-1 and GLP-2.
Materials and Methods
Sources
Antisera to GLP-1 (2135–8, 89390–4, and 92165–3) and GLP-2 were the kind gifts of Dr. J. J. Holst (Panum Institute, University of Copenhagen, Denmark) and Dr. D. J. Drucker (Banting and Best Center, University of Toronto, Toronto, Canada), respectively. Glucagon antiserum (P7) was raised in K. Polonsky’s laboratory at the University of Chicago. The recombinant human furin expression plasmid was a kind gift from Dr. Gary Thomas (Vollum Institute, Oregon Health Sciences University, Portland, OR). Antisera to glucagon, GLP-1, and GLP-2 as well as synthetic glucagon (PG 33–61), GLP-1 (PG 78–107 amide and 72–107 amide), and GLP-2 (PG 126–159) peptides were purchased from Peninsula Laboratories Inc. (San Carlos, CA). 125I-glucagon, 125I-GLP-1, and dipeptidyl protease IV inhibitor were bought from Linco Research Inc. (St. Charles, MO).
Construction of recombinant plasmids
The wild-type hamster PG cDNA, 992 bp, was cut out from the vector, pshGlu1 (46), by DdeI restriction digestion and subcloned into the XhoI site of another vector, pCIneo Eco-, after Klenow treatment of both the insert and vector (XhoI digested). pCIneo Eco- was derived from pCIneo (Promega, Madison, WI) after EcoRI digestion, Klenow fill-in, and religation to disrupt the EcoRI site in the polylinker. Unlike glucagon and GLP-2, GLP-1 segment of PG does not have any Met residues. Thus, to enhance detection of GLP-1- and GLP-1-containing peptides derived from PG after transient transfection and metabolic pulse-chase labeling, two amino acids, Val87 and Leu91, in the wild-type PG were substituted by Met residues using site directed mutagenesis (47). We named the resulting construct, WtPG II, as also mentioned in the rest of the text. The primers used for V87M and V87M/L91M were: 5'-TTTACCAGCGATATGAGCTCTTACT-3' and 5'-ATGAGCTCTTACATGGAGGGCCAGG-3', respectively. Using WtPG II as a template, the PG interdomain cleavage site mutant, R71A, was created by site-directed mutagenesis employing a kit from Pharmacia (Uppsala, Sweden) and we call this interdomain mutant, MutPG II, as also mentioned in the subsequent text. The primer used for MutPG II was 5'-AACATTGCCAAAGCCCACGATGAGTTT-3'. The MPGF (PG 72–158) expression construct with the N-terminal signal peptide was made by a fusion PCR. For fusion, we used two PCR-generated fragments, one having the signal peptide coding sequence and a portion of the polylinker and the other containing part of the MPGF coding sequence including V87M and L91M and an EcoRI site. The fusion PCR product was gel purified, digested with NheI and EcoRI, and cloned into WtPG II expression vector already digested with the same two restriction enzymes. The primers used for the MPGF construct were 5'-TACGACTCACTATAGGCTAGCCTC-3', 5'-GTGCCTCTCAAACTCATCGTGCTGCCAGCTGCCTTGCACCAG-3', 5'-CTTTCACCAGCCAAGCAATGAATTCC-3', and 5'-CTGGTGCAAGGCAGCTGGCAGCACGATGAGTTTGAGAGGCAC-3'.
We subcloned the wild-type hamster PG cDNA into the vector, pcDNA3.1 (Invitrogen, Carlsbad, CA), creating the expression construct WtPG I, which was subsequently used as a template to generate the interdomain mutant, MutPG I, having K70, R71 changed to Q70, Q71 by multiple PCRs. The primers used were: I, 5'-AACAGGAACAACATTGCCCAACAGCACGATGAGTTTGAGAGG-3'; II, 5'-GGCGCTCTAGATGAATGCTGACACACTTA-3'; III, 5'-GCCGCGGATCCCAGCAGGCAGAAAAAAAAATG-3'; and IV, 5'-GGCAATGTTGTTCCTGTT-3'. The thermocycle used for the first PCR having the primers I+II in one set of reaction and III+IV in the other set was 94 C for 4 min followed by 30 cycles at 94 C for 30 sec, 55 C for 30 sec, 72 C for 45 sec, and finally one hold at 72 C for 7 min. The products from these two reactions were gel purified and used as templates for the second PCR using the primers II and III, and the thermocycle was similar to the above except elongation at 72 C for 2 min followed by final hold at 72 C for 10 min. The final PCR product after gel purification was digested with BamHI and XbaI and ligated into pcDNA3.1 vector digested with the same restriction enzymes. The expression plasmid for glicentin (PG 1–69) with the signal peptide at the N terminus was made by PCR using WtPG I as the template and either of the two following forward primers (without or with the Kozak sequence) and a common reverse primer. The primers were: forward primers, 5'-ACTTAAGCTTATGAAGAACATTTACATTGTG-3' and 5'-ACTTAAGCTTGCCACCATGAAGAACATTTACATTGTG-3' and reverse primer, 5'-TAGTGGATCCTCAGGCAATGTTGTTCCTGTTCCT-3'. The PCR thermocycle was 94 C for 5 min followed by 28 cycles at 94 C for 30 sec, 56 C for 30 sec, and 72 C for 1 min and finally 72 C for 7 min. The PCR products were gel purified, digested with HindIII and BamHI, and finally ligated into pcDNA3.1 digested with the same two restriction enzymes. All these expression constructs were verified by sequencing.
Cell culture, metabolic labeling, IP, and tricine-SDS-PAGE
GH4C1 cells were cultured as mentioned before (38). Cells were transiently transfected and cotransfected with either wild-type or different mutant PG expression plasmids in absence or presence of various PCs using the transfection reagent, effectene (Qiagen, Valencia, CA), following the company’s protocol. Cells were pulse labeled, in most cases with 35S-Met (Amersham Biosciences, Amersham, UK) and sometime with 3H-Leu and 3H-Val (Amersham Biosciences), and chased as previously described (48). In the chase medium, aprotinin (5 μg/ml) and dipeptidyl protease IV inhibitor (10 μg/ml) were added to inhibit degradation of PG-derived peptides. Preparation of whole-cell extracts as well as immunoprecipitation (IP) of metabolically labeled cell extracts and media using various rabbit antisera were performed as described before (48). PG and its processed peptides were resolved in Tricine-SDS-PAGE (49) under reducing conditions (Laemmli sample buffer + ?-mercaptoethanol) and fixed and fluorographed as described before (48) before being exposed to film.
Reversed phase-HPLC (RP-HPLC)
Rabbit antiglucagon antibody was used to IP metabolically labeled peptides from 2-h chase media of wild-type and interdomain mutant without or with PC2, and after IP bound peptides were extracted in 0.1 N HCl (150 μl) at room temperature for 10 min with resuspension of beads at a 2-min interval. For maximum elution, this procedure was repeated once. Each pooled sample was divided into two equal halves: half after drying in a vacuum centrifuge (Speed Vac) was resolved in Tricine-SDS-PAGE (49) under reducing conditions, and the corresponding half was injected onto a Vydac C4 RP-HPLC column. As internal standards, 2 μg each of synthetic glucagon, GLP-1, and GLP-2 as well as His-tag PG (made in our laboratory by inserting six His residues after the first two amino acids, Met and Lys, in the PG N terminus without the signal sequence) were injected. The peptides were resolved over a three-step gradient of acetonitrile as described before (13). One fourth volume of each collected fraction was measured for total incorporated radioactivity using a liquid scintillation counter (model 2200 CA, Packard Instruments, Meriden CT).
Molecular modeling
This was performed using the Insight and Discover Graphical Environments (Biosym Technologies Inc., San Diego, CA). The molecular mechanics energetic calculations employed the force field, cvff. For initial approximation of glucagon as well as GLP-1 and GLP-2, the molecular models for the active sites of PC1/3 and PC2 (50) and the recently solved x-ray structure of furin (51) as well as the x-ray structure of glucagon (52) were used. In determining and describing the angles of rotation around the C-N () and C-C' () bonds in the peptide chain, we followed the standard nomenclature (53).
Results
Mutation of the PG interdomain site significantly impairs glucagon production
To study the biological importance of the interdomain dibasic site, K70-R71, in PG processing, we created two interdomain mutants, MutPG I (K70Q-R71Q) and MutPG II (R71A), as described in detail in Materials and Methods. We used both the mutants, separately, in transient transfection/cotransfection studies using GH4C1 cells derived from rat anterior pituitary. These cells were reported to lack detectable PC2 and PC1/3 expression (41, 54, 55) as was also verified in our laboratory (data not shown). The cells were also cotransfected with either PC2 or PC1/3, or with both convertases, and subsequently metabolically pulse labeled and chased (see details in Materials and Methods). We selected only the 2-h chase period for these studies because in preliminary experiments, this time point provided an optimal assessment of the progress of PG processing in these cells in the presence of PCs. Here we describe only the results (representative of at least three experiments) employing WtPG I and MutPG I because similar findings were obtained with WtPG II and MutPG II (data not shown). Both the WtPG I and WtPG II expression constructs are described in detail in Materials and Methods.
It is noteworthy that whereas wild-type hamster PG contains two Met residues, glicentin, GRPP-glucagon, oxyntomodulin, and glucagon each contain only one Met residue. Because all the metabolic pulse-chase experiments reported here were done using 35S-Met, we considered these differences in Met content in converting band intensities of PG and peptides derived from it into quantitative estimates of cleavage products.
When WtPG I was transiently expressed without exogenous PCs in GH4C1 cells, we found cleavage only at the interdomain site, generating moderate amounts of glicentin, particularly in the 2-h chase medium (Fig. 2A, lane 4) after IP with antiglucagon antibody. We confirmed the endogenous expression of two PCs, PC5/6A and furin, in GH4C1 cells (data not shown) as also reported (41, 54). Our findings thus indicate that these two broadly distributed convertases, PC5/6A and/or furin, are most likely responsible for the observed cleavage at the interdomain site, K70-R71, in GH4C1 cells in the absence of both PC2 and PC1/3 (see below). Coexpression of PC2 more effectively processes the interdomain site, especially in 2-h chase medium, generating significantly greater amounts of glicentin as well as robust levels of GRPP-glucagon and glucagon and low amounts of oxyntomodulin (Fig. 2A, lane 7).
FIG. 2. Processing of wild-type PG vs. interdomain mutant by PC2 and/or PC1/3 in GH4C1 cells. One microgram of hamster PG expression plasmid, WtPG I (A and B) or MutPG I (C and D), was transiently transfected in absence and presence of mouse PC2 (1 μg) or mouse PC1/3 (1 μg) or both the convertases (1 + 1 μg) (see Materials and Methods for details). After 44 h of transfection, cells were pulse labeled (30 min) with 35S-Met (150 μCi /ml) and chased (2 h) as described in the text. Equal amounts of cell extracts (CE) and equal volumes of chase medium (M) were immunoprecipitated with rabbit antiglucagon antibody, and immunoprecipitated peptides were resolved in tricine-SDS-PAGE (see Materials and Methods for details). After fixation and fluorography, gels were dried and exposed to Kodak X-Omat AR-5 films (Fisher Scientific, Pittsburgh, PA). Lane 1, A–D, Location of 125I-glucagon. The positions of the prestained protein molecular-weight markers are shown on the left of each figure. P, Pulse; Ch, chase.
Analysis of the levels of precursor and processed peptides derived from it by densitometric scanning (1D Image Analysis software, Kodak, Rochester, NY) revealed that in the absence of any exogenous PCs, secreted glicentin in medium constituted a moderate 10% of total levels of peptides, whereas unprocessed PG comprised 24% in chased cell extracts and 66% in corresponding medium samples (Fig. 2A, lanes 3 and 4), indicating efficient secretion of significant amounts of uncleaved precursor. In contrast, in the presence of PC2, the total levels of glicentin secreted in the medium constituted 28%, whereas the amounts of uncleaved precursor were only 7.5% in chased cell extracts and 22% in corresponding medium samples, confirming a significantly enhanced cleavage at the interdomain site by PC2 (Fig. 2A, lanes 6 and 7). Glicentin was subsequently processed by PC2 at the sites, K31-R32 and K62-R63, either simultaneously or sequentially, to generate significant amounts of both glucagon (18.5%) and GRPP-glucagon (19%) and low amounts of oxyntomodulin (5%) (Fig. 2A, lane 7). Coexpression of PC1/3 also liberated significant amounts of glicentin after the interdomain cleavage and subsequently glicentin was processed by PC1/3 only at the site, K31-R32, to generate high levels of oxyntomodulin, especially in the 2-h chase medium (Fig. 2B, lane 4). Densitometric scanning showed that, whereas secreted glicentin constituted 40% of total levels of peptides in the presence of PC1/3, the amounts of oxyntomodulin were 14% and uncleaved precursor constituted 16% in chased cell extracts and 30% in the corresponding medium (Fig. 2B, lanes 3 and 4). It is noteworthy that production of both GRPP-glucagon and glucagon was undetected in presence of PC1/3 in GH4C1 cells, suggesting that the cleavage site, K62-R63, is exclusively processed by PC2. Coexpression of PC1/3 with PC2 did not change the levels of GRPP-glucagon or glucagon that were generated by PC2 alone (Fig. 2B, lane7 vs. Fig. 2A, lane 7) confirming the absolute necessity of PC2 in cleaving the 62K-63R site in the glicentin segment.
When MutPG I was transiently expressed in GH4C1 cells without any exogenous PCs, we found no apparent processing, either in cells or medium, and correspondingly, significant amounts of unprocessed mutant precursor were secreted into the medium (Fig. 2C, lanes 2–4). Whereas PC2 alone was coexpressed, it generated a major intermediate fragment, PG 33–158, and low levels of GRPP-glucagon from the interdomain mutant both in cells and medium (Fig. 2C, lanes 6 and 7). Also robust amounts of unprocessed mutant precursor were secreted into the medium (Fig. 2C, lane 7). We detected very low levels of glucagon generated from the interdomain mutant in the presence of PC2, after a prolonged exposure of 16 d for the chase medium samples (Fig. 2C, lane 8) relative to an exposure of 4 d for the rest of Fig. 2. The findings confirm that the interdomain cleavage is important but not absolutely required to first generate glicentin, which is a far better substrate for subsequent cleavage by PC2 to produce glucagon. PC1/3 coexpression alone with MutPG I generated mainly three intermediate peptides, PG 33–158, PG 33–123, and PG 33–76, especially in chase medium (Fig. 2D, lane 4) and PC2 and PC1/3 together generated, along with these three intermediates, also low levels of GRPP-glucagon (Fig. 2D, lane 7) and very low amounts of glucagon seen only on longer exposure (data not shown). The findings altogether confirm that PG processing normally starts at the interdomain site despite the presence of several other similar cleavage motifs. It is noteworthy that the nonconventional peptides generated from mutant PG in presence of either PC2 or PC1/3 or both were identified on the basis of their individual mobility under reducing conditions in Tricine-SDS-PAGE relative to the mobility of purified iodinated markers, prestained low-molecular-weight markers and with the help of prior knowledge on PG processing, and by using specific antibodies against glucagon, GLP-1 or GLP-2. The conventional peptides including glucagon, GLP-1, GLP-2, PGs were identified in Tricine-SDS-PAGE corresponding to the locations of iodinated marker peptides and in RP-HPLC with reference to the internal standards as shown.
Next we quantitatively measured the extent of defects in glucagon production from the interdomain mutant relative to wild-type in the presence of PC2 by RP-HPLC. After cotransfection in GH4C1 cells without or with PC2, half of the chase media was IP with rabbit antiglucagon antibody and resolved in Tricine gradient gel (Fig. 3A, lanes 2–5). The corresponding half with PC2 after IP with the same antibody was extracted in 0.1 N HCl and subjected to RP-HPLC as described in detail in Materials and Methods. Analysis after scintillation counting showed a 5-fold reduction in glucagon production from MutPG I relative to WTPG I in presence of PC2 (Fig. 3, D vs. C). Correspondingly, we found almost 4-fold higher levels of unprocessed mutant PG secreted into the chase medium relative to the wild-type (Fig. 3, D vs. C). The elution peaks for PG in the experimental samples were slightly shifted to the right with reference to the location of the purified PG internal standard. This was possibly due to the presence of the His tag in the purified PG standard that was made as described in Materials and Methods. The results from RP-HPLC analysis were in agreement with the corresponding autoradiogram (Fig. 3A, lanes 5 vs. 3), confirming our hypothesis that the first cleavage at the PG interdomain site, K70-R71, is very important for efficient production of glucagon by PC2. Specific incorporation of radioactivity (counts per minute) after metabolic pulse-chase labeling with 35S-Met was much more robust relative to combination of two tritiated amino acids and thus the fold differences between the wild-type and mutant PG samples as shown in the representative RP-HPLC figures were calculated based on 35S-Met incorporation.
FIG. 3. Quantitative estimation of impaired glucagon production from interdomain mutant relative to wild type in presence of PC2. A, GH4C1 cells were transiently transfected with WtPG I (1 μg) or MutPG I (1 μg) without or with PC2 (1 μg) for 44 h and after 30 min pulse labeling, chased for 2 h. Equal amounts of media were immunoprecipitated with rabbit antiglucagon antibody, and half of the immunoprecipitated peptides after extraction in 0.1 N HCl were resolved in Tricine-SDS-PAGE. Lane 1 indicates the location of 125I-glucagon, and the positions of the prestained protein molecular weight markers are shown on left. C and D, Each corresponding half of the immunoprecipitated peptides from 2-h chase medium of wild-type and interdomain mutant, respectively, was analyzed by RP-HPLC. B, RP-HPLC peaks for the peptides used as internal standards (also see Ref.13 for further information on HPLC standardization). 35S and 3H are denoted by and , respectively.
In immunofluorescence assays employing AtT20 cells, we found significant and comparable levels of colocalization of wild-type and interdomain mutant PGs in secretory granules and Golgi using specific markers (data not shown). The findings suggest normal intracellular transport and sorting of the mutant PGs, similar to wild-type, after synthesis.
Next, we created an expression vector for glicentin with the PG signal sequence fused to the N terminus of glicentin (see Materials and Methods). This was transfected into GH4C1 cells in the absence and presence of PC2 or PC1/3. We showed that PC2 generated significant levels of glucagon along with high levels of GRPP-glucagon and moderate amounts of oxyntomodulin, especially in chase medium (Fig. 4A, lane 7), as we found earlier (Fig. 2A, lane 7). In contrast, PC1/3 produced only oxyntomodulin similar to what we found before (Fig. 4B, lane 4 vs. Fig. 2B, lane 4), and most of the unprocessed glicentin was secreted into the medium. The findings confirm the absolute requirement for PC2 in glucagon production due to its efficient action at both the dibasic sites, K31-R32 and K62-R63, relative to the restricted ability of PC1/3 to cleave only slowly at the site, K31-R32. The results indicate that amino acids present near the dibasic site, K62-R63, confer a unique specificity for PC2 cleavage.
FIG. 4. Processing of glicentin by PC2. An expression plasmid for glicentin (1 μg) (see Materials and Methods) was transiently transfected in the absence and presence of PC2 (1 μg) or PC1/3 (1 μg) in GH4C1 cells and after pulse-chase labeling, peptides from equal amounts of both cell extracts (CE) and medium (M) were immunoprecipitated with rabbit antiglucagon antibody. Immunoprecipitates were resolved in Tricine-SDS-PAGE. Lane 1, A and B, Position of 125I-glucagon; the locations of the prestained protein molecular markers are shown on the left. P, Pulse; Ch, chase.
We examined the effects of PC5/6A or furin on PG processing using cotransfection assays because we consistently found generation of moderate levels of glicentin in GH4C1 cells after transient expression of wild-type precursor in the absence of any exogenous PCs (see Fig. 2A, lanes 2–4). Coexpression of WtPG I with either PC5/6A or furin showed considerable enhancement in glicentin levels in both chased cells and medium and correspondingly much less accumulation and secretion of uncleaved precursor, compared with transfection of PG alone (Fig. 5, lanes 5–7 and 8–10 vs. 2–4), confirming considerable susceptibility of the interdomain site, K70-R71, to PC5/6A and/or furin. It is noteworthy that glicentin was not processed further by either of these two convertases, indicating the total inability of furin or PC5/6A to cleave the K31-R32 and K62-R63 sites located in the glicentin segment. However, when we examined the processing of MutPG I in similar cotransfection assays, we found small amounts of two major intermediates, PG 33–158 and PG 33–123, indicating some susceptibility to cleavage by PC5/6A or furin of the K31-R32 and R124-R125 sites in this mutant form (data not shown).
FIG. 5. Production of glicentin from PG by PC5/6A and furin. GH4C1 cells were transiently transfected with WtPG I (1 μg) in absence and presence of mouse PC5/6A (1 μg) or human furin (1 μg) expression constructs. After pulse-chase labeling, equal amounts of cell extracts (CE) and medium (M) were immunoprecipitated with rabbit antiglucagon antibody, and immunoprecipitates were analyzed in Tricine-SDS-PAGE. Lane 1 indicates the location of 125I-glucagon, whereas the prestained protein markers are shown on the left. P, Pulse; Ch, chase.
No impairment in GLP-1 and GLP-2 processing
To compare levels of GLP-1 generated from the interdomain mutant relative to wild type by various PCs and detect GLP-1 without ambiguity after pulse-chase labeling and IP (with anti GLP-1 antibody), we generated the expression vectors, WtPG II and MutPG II (see Materials and Methods for details). Transient expression of WtPG II alone in GH4C1 cells, in absence of any exogenous PCs, generated only moderate levels of MPGF in both cell extracts and medium (Fig. 6A, lanes 2–4), confirming the ability of both furin and PC5/6A to cleave the interdomain site, K70-R71, as we found for glicentin production using the antiglucagon antibody (see Fig. 2A, lanes 2–4). When PC2 was coexpressed, MPGF and N-terminally extended GLP-1 (PG 72–108/107-NH2) were generated and secreted prominently into the chase medium (Fig. 6B, lane 4). In contrast, coexpression of PC1/3 produced both MPGF and GLP-1 (PG 78–108/107-NH2), which were secreted into the chase medium (Fig. 6C, lane 4). Employing MutPG II, we found no significant impairment in GLP-1 production and secretion in the presence of PC1/3 (Fig. 6F, lanes 3 and 4) relative to wild type, suggesting that the first cleavage at the interdomain site of PG is critical only for glucagon production. Production of comparable amounts of GLP-1 from both interdomain mutant and wild-type precursors thus indicates unperturbed ability of PC1/3 to cleave the R77 and R107-X-R-R110 sites, irrespective of the interdomain mutation. The intermediate peptides generated from the mutant precursor by PC1/3 consisted mostly of PG 33–158 and partly of PG 33–123 and 33–108, whereas PC2 was found to generate two major intermediates, PG 33–158 and 64–158, and two minor intermediates, PG 33–108 and 64–108 (Fig. 6, F vs. E). Also similar to MutPG I (Fig. 2C, lanes 2–4), we found no apparent processing of MutPG II in the absence of any exogenous PCs (Fig. 6D, lanes 2–4). The nonconventional peptides were identified as discussed before.
FIG. 6. Processing of wild-type PG vs. interdomain mutant by PC2 or PC1/3 and assessment of N-terminally extended GLP-1 and GLP-1 (PG78–108/107-NH2) production. WtPG II (1 μg) or MutPG II (1 μg) was transiently transfected in absence and presence of PC2 (1 μg) or PC1/3 (1 μg) in GH4C1 cells. Cells were pulse labeled for 30 min and chased for 2 h. Equal amounts of cell extracts (CE) and medium (M) were immunoprecipitated with rabbit anti-GLP-1 antibody (2135–8), and immunoprecipitated peptides were resolved in Tricine-SDS-PAGE. Lane 1 in all the gels represents the location of the 125I-GLP-1, and the prestained low-molecular-weight protein markers are shown on the left of each gel. P, Pulse; Ch, chase.
Next we examined the levels of GLP-2. Again we found no significant defects in the production of GLP-2 (PG126–158) from the interdomain mutant, compared with the wild-type precursor in the presence of PC1/3, and the predominantly secreted peptide in both cases was GLP-2 (Fig. 7D, lanes 3 and 4 vs. 7B, lanes 3 and 4). Also, we found equivalent levels of N-terminally extended GLP-2 (PG 111–158) in chase media from both precursors (Fig. 7D, lane 4 vs. 7B, lane 4). The findings again suggest that regardless of interdomain cleavage, PC1/3 processes both the cleavage sites, R107-X-R-R110 and R124-R125, as effectively as we have previously seen for the production of GLP-1 after cleavages at the R77 and R107-X-R-R110 sites (see Fig. 6). Coexpression of PC2 produced equivalent levels of N-terminally extended GLP-2 from both WtPG I and MutPG I (Fig. 7A, lanes 6 and 7, and 7C, lanes 6 and 7), suggesting that the ability of PC2 to cleave the R107-X-R-R110 site was not perturbed by the interdomain mutation. The results taken altogether indicate that, whereas PC1/3 effectively cleaves the three sites, R77, R107-X-R-R110, and R124-R125, PC2 completely lacks the ability to process the R77 and R124-R125 sites.
FIG. 7. Assessment of GLP-2 production from wild-type PG vs. interdomain mutant in presence of PC2 or PC1/3. GH4C1 cells were transiently transfected by WtPG I (1 μg) or MutPG I (1 μg) in absence and presence of PC2 (1 μg) or PC1/3 (1 μg). After pulse-chase labeling, equal amounts of cell extracts (CE) and medium (M) were immunoprecipitated with rabbit anti-GLP-2 antibody, and immunoprecipitates were resolved in Tricine-SDS-PAGE. The lane 1 in A–D indicates the position of the 125I-glucagon and the locations of the prestained protein markers are shown on the left.
We also studied the processing of expressed MPGF (PG 72–158) (see Materials and Methods for details) in cotransfection assays in GH4C1 cells and showed again that PC2 only generated the N-terminally extended form of GLP-1, whereas PC1/3 produced GLP-1 (Fig. 8, A-C). The findings confirm that the two MPGF cleavage motifs, R77 and R107-X-R-R110, are considerably processed by PC1/3 as shown above (see Fig. 6). However, PC2 specifically cleaved the R107-X-R-R110 site as found previously (see Fig. 7) confirming its exclusive preference for a single cleavage site relative to a broader preference of PC1/3 for all three of the sites located in the MPGF segment. In examining GLP-2 immunoreactive peptides in cotransfection assays, PC2 was found to generate only N-terminally extended GLP-2, whereas PC1/3 produced GLP-2 from MPGF (data not shown).
FIG. 8. Processing of MPGF by PC1/3. MPGF (1–71 amino acids of PG) expression plasmid (1 μg) (see Materials and Methods) was transiently transfected in GH4C1 cells in absence and presence of PC2 (1 μg) or PC1/3 (1 μg) and pulsed and chased for 30 min and 2 h, respectively. Equal amounts of cell extracts (CE) and medium (M) were immunoprecipitated with rabbit anti-GLP-1, and immunoprecipitates were resolved in Tricine-SDS-PAGE. Lane 1 in A–C represents the location of the 125I-GLP-1. The prestained protein markers are located on the left on each gel. P, Pulse; Ch, chase.
We analyzed the role of PC5/6A and furin in the generation of GLP-1 and GLP-2 using only the wild-type precursor because we failed to detect any significant defects in production of these two peptides from the interdomain mutant by PC1/3 (see Figs. 6 and 7). We showed in cotransfection assays that, whereas PC5/6A generated only considerable levels of MPGF and N-terminally extended GLP-1 in both cells and medium (Fig. 9B, lanes 2–4), furin produced, along with MPGF, both N-terminally extended GLP-1 and GLP-1 that were efficiently secreted into the medium at a ratio of 1:1 (Fig. 9C, lanes 2–4). The results indicate a relatively lower ability of furin than PC1/3 to cleave the single basic site, R77, and also suggest a complete inability of PC5/6A to process the same site. However, both furin and PC5/6A are able to cleave the interdomain site, K70-R71, as well as the site R107-X-R-R110. When we studied the effects of PC5/6A or furin on GLP-2 production, we found considerable levels of MPGF and N-terminally extended GLP-2 but only very low levels of GLP-2 (Fig. 9D, lanes 5–7, and E, lanes 2–4), suggesting much stronger activities of PC5/6A or furin at the site R107-X-R-R110, relative to the more distal cleavage site R124-R125.
FIG. 9. Processing of PG by PC5/6A or furin and generation of GLP-1 (PG78–108/107-NH2) and GLP-2 (PG126–158). GH4C1 cells were transiently transfected with wild-type PG, either WtPG II (1 μg) or WtPG I (1 μg), in absence and presence of PC5/6A (1 μg) or furin (1 μg). Cells were pulse labeled (30 min) and chased (2 h) and equal amounts of cell extracts (CE) and medium (M) were immunoprecipitated with either rabbit anti-GLP-1 (A–C) or GLP-2 (D and E). Immunoprecipitates were analyzed in Tricine-SDS-PAGE. Lane 1 in A–C and D and E represents the location of the 125I-GLP-1 and 125I-glucagon, respectively, and the positions of the prestained low-molecular-weight protein markers are shown on the left of each figure. P, Pulse; Ch, chase.
Molecular modeling of PG
Although crystallization of PG has not yet been successful, an x-ray structure of crystalline glucagon was reported almost three decades ago wherein the glucagon moiety was mainly -helical (residues 6–28) (52). However, in solution at high dilutions, glucagon does not have a highly ordered conformation (56), whereas in more concentrated solutions (at 3 mg/ml), glucagon monomers associate into trimers (57). The association of monomers to trimers stabilizes the -helical conformation of glucagon. We propose that PG in solution forms a structure analogous to the trimer in crystalline glucagon in which the three homologous segments, residues 38–59 of glucagon, 83–104 of GLP-1, and 131–152 of GLP-2, in the PG sequence form -helices. The presence of peptide connections between these homologous regions is equivalent to the concentration effect in glucagon solutions, in which the increase in concentration of glucagon leads to the trimeric arrangement of -helical glucagon molecules (56, 57). The idea that the glucagon region of PG is mainly -helical is confirmed by the observation that, despite the presence of a dibasic site in the middle of the glucagon moiety (R49-R50), proteolytic cleavage at this position by prohormone convertases occurs only at very low levels in pancreas (58). If organized in this fashion, the inner core of PG would exist in the form of a triple complex among glucagon, GLP-1, and GLP-2, analogous to the trimer in crystalline glucagon (52), whereas the intervening peptides, dividing glucagon and GLP-1, and also GLP-1 and GLP-2, form nonregular external loops (residues 62–77 and 109–125). The N-terminal peptide segment, GRPP (PG1–30), is predicted to be highly disordered [prediction based on PONDR Protein Disorder Predictor (www.pondr.com)] and lies outside this predicted PG core structure.
To model PG, we simulated the ordered docking of the -helical segments of glucagon, GLP-1, and GLP-2, respectively, in such a triple complex taking into consideration the optimal packing of hydrophobic residues on the N and C ends of the corresponding predicted -helices. A possible arrangement of the individual glucagon, GLP-1 and GLP-2 units, found by optimization of the energy of nonbonded interactions in this complex and the corresponding three hydrophobic knots maintaining their trimer-like assembly with 3-fold symmetry, is shown in Fig. 10A. In this model, the hydrophobic patch near the C terminus of glucagon interacts with a similar N-terminal hydrophobic patch of GLP-1. Similarly the C-terminal end of GLP-1 interacts with the N-terminal end of GLP-2 and so on, as expected because the hydrophobic patches near the ends of each of the predicted -helical segments in GLP-1 and GLP-2 are composed of the same or very similar bulky hydrophobic and aromatic amino acid residues as are present in glucagon. The hydrophobic knot between glucagon and GLP-1 is formed by contacts between amino acids Phe54, Trp57, and Leu58 of glucagon and Phe83, Val87, and Tyr90 of GLP-1. Similarly, the knot between GLP-1 and GLP-2 is formed by contacts of Phe99, Trp102, and Leu103 of GLP-1 with Phe131, Met135, and Ile138 of GLP-2. Because the dibasic site within glucagon moiety, R49-R50, is substituted by nonpolar amino acids in GLP-1 and GLP-2, Q94-A95 and L142-A143, respectively, electrostatic interactions within the trimer are ruled out (see Fig. 10A). In turn, these hydrophobic patches could stabilize the formation of even longer -helices within the PG structure. Given the fact that the dibasic site R49-R50, located within the -helical segment of glucagon, is cleaved very little, if at all, during biosynthesis either in intestinal L cells or islet -cells (13, 14, 16, 58), we propose that the other dibasic sites, which are not cleaved initially in PG processing, are also located on the N and C ends of -helices, which were extended to include residues 31–63, 77–107,and 124–155, at which the C end of the glucagon segment has the highest -helical potency. On this basis, the -helical content of PG is estimated to be about 50%. This figure is consistent with results of CD measurements made on recombinant human PG (Zhu, X., C. Zhang, and D. F. Steiner, unpublished data).
FIG. 10. A, Trimeric structure of PG. PG contains three homologous peptide sequences, glucagon, GLP-1, and GLP-2, which are modeled as -helices including the dibasic sites on their N and C ends. The hydrophobic knot between glucagon and GLP-1, shown in the structure of PG, is formed by the hydrophobic patches on the C end of glucagon (Phe-54, Trp-57, Leu-58) and the N end of GLP-1 (Phe-83, Val-87, and Tyr-90). B, IP-1 forms an external loop that acts as a docking site for PCs. The IP-1, 64–76 amino acids of PG, is modeled as an external loop that corresponds to a conformation required for the formation of a complex with the active site of PCs. Here the sequence, IAKRHD, residues 68–73 of IP-1, corresponding to the docking site of furin, is shown by space-filled images. C, A modeled spatial structure for PG.
After definition of the -helical segments in PG, we also considered the specific stereochemical requirements for a polypeptide chain to enter an -helix at the N end and to exit from it at the C end (59). Accordingly, amino acids 63–65 (Arg-Asn-Arg) at the C end of the glucagon moiety have been modeled in a specific conformation corresponding to the parallel exit of the polypeptide chain from the C ends of -helices, ? (59). Likewise, amino acids 76–78 (Glu-Arg-His) preceding -helix of GLP-1 segment have been modeled as a parallel entry at the N end of this -helix, ? (59).
Recently the crystal structures of two proprotein convertases, furin and kexin, have been solved (51, 60). The catalytic domain of furin has been shown to be very similar to subtilisin and related convertases in overall structural organization. Like the subtilisins (61), the active site groove of each convertase can bind at least six residues, P4-P2' of a substrate in an extended conformation. In reconstructing a possible conformational variant of the external loop of IP-1 (PG 66–75) that contains the initial cleavage site K70-R71, we considered only those variants that would facilitate formation of an enzyme-substrate complex with the substrate binding groove of furin as a model PC. Because a productive convertase interaction typically requires an extended peptide conformation for residues P4-P2' of a substrate, the IP-1 sequence, IAKRHD (68–73 amino acids), corresponding to P4-P2' positions has been arranged in such an energetically optimized conformation to fit the active site of furin. Its location in the groove corresponds to that of the peptide inhibitor of furin (51). The peptide bond between R71 and H72 also is arranged in an orientation appropriate for catalytic attack by the side chain of S368 of furin. The angles of rotation (C-N) and (C-C’) around the main chain bonds for residues IAKRHD are found to be –77,114; –88,124; –78,128; –57,75; –114,170; –105,–114. At the same time, the segment IAKRHD forms the top of the external loop of IP-1. In this situation, in which residues P4-P2', IAKRHD, are arranged in the active site, two dipeptide units, Asn66-Asn67 and Glu74-Phe75 of IP-1, which connect the top hexapeptide with glucagon and GLP-1, act as spacers to prevent any van der Waals overlap between the volume of the glucagon-GLP-1 moiety and the external surface of the convertase. The proposed docking interaction is shown in Fig. 10B, in which an unobtrusive conformation of IP-1 relative to the convertase is achieved by formation of inner half-turns by the above-mentioned residues, Asn-Asn and Glu-Phe on either side of the bound peptide segment 68–73 (P4-P2'). It directs the remaining PG structure away from the convertase surface. The volume of the catalytic domain of the prohormone convertase is modeled according to the catalytic domain of furin (51).
We predict that IP-2 (PG 111–123), connecting GLP-1 and GLP-2, also contains a short -helical segment, PEEVAIV (PG 113–119), because Glu residues have a high potential for initiation of -helices, and Pro residues are often located at their N ends (53). The flexible amino acids, Asp111 and Gly123, at the N and C ends of this external loop of IP2 help to form the exit from the -helix formed by the GLP-1 domain and the entry into the -helix formed by the GLP-2 domain, respectively. Then side chains of the -helical residues, VAIV at the top of the IP-2 loop, may contribute to the hydrophobic core of PG.
Employing the proposed initial approximations for the glucagon, GLP-1 and GLP-2, moieties as well as for the two connecting peptides, IP-1 and IP-2, we carried out an optimization of the inner potential energy of the PG structure, excluding the extraneous highly disordered N-terminal peptide segment, PG 1–30, and derived the model as projected in Fig. 10C. In this proposed structural motif, all dibasic sites are screened by intramolecular van der Waals contacts with neighboring amino acids except for the interdomain site, K70-R71, which is located on the top of the external loop formed by IP-1 and, due to its high accessibility, easily adopts a conformation appropriate for binding with the active sites of prohormone convertases. These considerations thus predict cleavage of the interdomain site as the first step in the temporally ordered processing of intact PG.
After the cleavage of the dibasic site K70-R71 of PG, the newly formed glicentin moiety is processed to glucagon in -cells, which express PC2 at high levels (13). In the model, both sites K31-R32 and K62-R63, which are located at the ends of the single -helix in the glucagon segment, now are open and likely have similar stereochemical possibilities for interactions with PC2. Unexpectedly, in AtT-20 cells, expressing a high level of PC1/3, the newly formed MPGF was cleaved first at the monobasic site R77 with formation of truncated MPGF and only after that, at the R109-R110 site, leading to the formation of GLP-1 (16). Therefore, the N-terminal segment of MPGF up to and including R77, is most likely in an unfolded conformation, which is more readily accessible for interactions with the active site of PC1/3, than the inner dibasic site, R109-R110, in IP2. This region, separating two -helical segments of MPGF (GLP-1 and GLP-2) is still screened by nonbonded interactions inside an antiparallel helix-turn-helix hairpin, which in turn, can be stronger, than in the PG structure itself. This residual structure in free MPGF may therefore account for the lack of any apparent effect of inhibition of interdomain cleavage on the processing of this half of the PG molecule.
Discussion
We conclude from our current studies that despite the presence of several cleavage sites (see Fig. 1), PG processing mediated by PCs including PC2, PC1/3, PC5/6A, and furin begins at the interdomain site, K70-R71, generating two major peptides, glicentin and MPGF. Glicentin is then processed by PC2 to generate glucagon. The exclusive role of PC2 in generating glucagon was illustrated in our previous studies using PC2 null animals and an -cell line (17, 19, 62) and also in mice null for 7B2 (63), a protein required for activation of PC2. For glucagon production, glicentin must be processed at its two cleavage sites, K31-R32 and K62-R63. In the current studies, we have shown that the site K62-R63 is exclusively processed by PC2, whereas the K31-R32 site is cleaved by both PC2 and PC1/3. Employing interdomain cleavage site mutants, we find production of only very low amounts of glucagon (5-fold less) by PC2 relative to wild type. However, PC2 efficiently cleaves the K31-R32 site generating considerable levels of PG 33–158 from mutant precursor. These results indicate that some kind of steric hindrance strongly masks the K62-R63 site in intact PG but not in glicentin. Contrary to PC2, PC1/3 is unable to cleave the K62-R63 site located in either PG or glicentin, and this is in agreement with our studies done in TC-PC2 cells that contain only low levels of PC1/3 (19). The findings are also consistent with studies on PC1/3 null mice in which glucagon production is unchanged from that in wild-type animals (18). The results taken altogether confirm the absolute requirement of PC2 for glucagon production but are in contrast with the findings of a few other investigators (21, 22, 23) who were either unable to adequately substantiate the presence of enzymatically active PC2 in vaccinia virus preparations or failed to demonstrate loss of PC2 enzymatic activities in antisense studies.
When we examined the levels of GLP-1 and GLP-2 produced from interdomain mutant relative to wild type in presence of PC1/3, we found comparable amounts of these peptides. These results may indicate a lack of any significant steric hindrance for PC1/3 in accessing the three C-terminal cleavage sites after the interdomain site, R77, R107-X-R-R110, and R124-R125, despite the interdomain mutation. On the other hand, they could also be consistent with the persistence of a comparable level of steric hindrance in the free MPGF domain after its release from PG as suggested by the modeling experiments. Likewise, the generation of equivalent amounts of N-terminally extended GLP-2 by PC2 from both interdomain mutant and wild-type may be on a similar basis. Expressing either intact PG or MPGF, we confirmed that GLP-1 is generated by PC1/3, whereas PC2 produces only the N-terminally extended form of GLP-1. The results are in agreement with previous in vitro and in vivo studies implying an important role for PC1/3 in generating GLP-1 and GLP-2 (14, 16, 18, 22, 23, 64) and PC2 in producing the N-terminally extended forms of GLP-1 and GLP-2 (13, 19). The ability of PC1/3 to process a single basic site without a P2 or P4 basic residue has been shown for a few other neuroendocrine precursors (48, 65). However, in vitro cell-free studies employing PC1/3 showed either lack of conversion or very low levels of processing of N-terminally extended GLP-1 (PG 72–108) into GLP-1 (PG 78–108) (14, 16), indicating weaker ability of PC1/3 to cleave the monobasic site, R77. In a living human subject lacking active PC1/3, both GLP-1 and GLP-2 are found in plasma (66). Also low levels of GLP-2 and GLP-1 were found in intestinal tissue extracts of PC1/3 null mice (18). Whereas expression of PC2 has not been detected immunohistochemically in any regions of the small intestine (18), PC5/6A is widely expressed in the gut (67, 68), and furin, the ubiquitously expressed convertase (69), is coexpressed with PC1/3 and PC5/6A in small intestine (70). The results altogether suggest the existence of overlapping specificities among PC5/6A, furin, and PC1/3 to generate GLP-1 and GLP-2.
We have consistently seen production of some glicentin and MPGF from transiently expressed PG without any exogenous PCs in GH4C1 cells, indicating that either PC5/6A or furin or both are responsible for the observed moderate levels of cleavage at the interdomain site, as also suggested by others (14, 15, 16, 18, 22). Coexpression of PC5/6A or furin generated greater amounts of glicentin and MPGF from PG. Remarkably, neither PC5/6A nor furin appeared to process glicentin further (see Fig. 5), indicating the inability of these two broadly distributed convertases to cleave the sites K31-R32 and K62-R63. The most likely explanation might be lack of upstream basic residues or other features of local amino acid sequence near the two cleavage sites. The findings also explain the lack of any detectable ability of PC5/6A to liberate any glucagon despite a report of its colocalization with glucagon in secretory granules of pancreatic islets (71). In contrast, MPGF was efficiently cleaved by PC5/6A or furin under similar experimental conditions. We found in cotransfection assays that PC5/6A generated considerable levels of both N-terminally extended GLP-1 and GLP-2, whereas furin produced, along with N-terminally extended GLP-1 and GLP-2, also considerable amounts of GLP-1 (PG 78–108/107-NH2). However, the ability to produce GLP-2 (PG 126–158) by either PC5/6A or furin was rather weak, compared with PC1/3. In the current studies, we confirmed the ability of furin to cleave the single basic site R77 and the R107-X-R-R110 site, which is also processed by PC5/6A to generate GLP-1. Furin was previously shown to cleave a single Arg residue in prosomatostatin to produce SS-28 (45). Also in studies of PC2 knockout mouse pancreatic extracts, SS-28 was found as the major end product of prosomatostatin processing (62), although islet -cells were shown to lack PC1/3 expression (72). We suggest, based on these results, that efficient normal production of both GLP-1 and GLP-2 from PG requires redundant functions of PC1/3, furin, and PC5/6A. In this context, conditional knockouts for furin and/or PC5/6A in intestinal L cells would be of great biological interest. However, expression of both PC2 and PC1/3 in canine L cells was reported (73). In our laboratory, we were unable to detect any PC2 expression in mouse intestine by RT-PCR (Zhu, X., and C. Zhang, unpublished data), and these findings were confirmed in a recent independent study (18).
The remarkable feature of PG processing, irrespective of its sites of synthesis, is that the interdomain site, K70-R71, located in the middle of the IP-1 sequence, PG 62–77, is cleaved first despite the presence of three other dibasic sites, KR, a single monobasic site, R, and one R-X-R-R site (see Fig. 1) as previously predicted in a study from our laboratory (13) and also shown in the current studies. The findings suggest the possibility that in native PG, all the cleavage sites except the interdomain site are screened by intramolecular contacts, whereas the interdomain site, located at the top of the external loop connecting glucagon and GLP-1, is more available to interact with the active site of PCs including PC2, PC1/3, PC5/6A, and furin. In our model of the spatial structure of PG, IP-1, PG 62–77, is an external loop in an optimal conformation for interacting with the active sites of PCs in general (50, 51) (see Fig. 10, B and C). Cleavage at this site allows the molecule to open, giving rise to glicentin and MPGF and thus setting the stage for the next processing steps involving either PC2 mediated cleavage of glicentin to generate primarily glucagon in the pancreatic -cells or PC1/3, PC5/6A, and furin assisted processing of MPGF to produce predominantly GLP-1 and GLP-2 in intestinal L cells.
Acknowledgments
We thank Paul Gardner for making oligonucleotide primers, Xiaorong Zhu for providing the His-tag PG, and Rama Boddipalli for his technical assistance in running the Kodak 1D Image Analysis software. We also thank Rosie Ricks for her expert secretarial assistance.
References
Patzelt C, Tager HS, Carroll RJ, Steiner DF 1979 Identification and processing of proglucagon in pancreatic islets. Nature 282:260–266
Mojsov S, Heinrich G, Wilson IB, Ravazzola M, Orci L, Habener JF 1986 Proglucagon gene expression in pancreas and intestine diversifies at the level of post-translational processing. J Biol Chem 261:11880–11889
Novak U, Wilks A, Buell G, McEwen S 1987 Identical mRNA for preproglucagon in pancreas and gut. Eur J Biochem 164:553–558
Orskov C, Holst JJ, Poulsen SS, Kirkegaard P 1987 Pancreatic and intestinal processing of proglucagon in man. Diabetologia 30:874–881
Larsen PJ, Tang-Christensen M, Holst JJ, Orskov C 1997 Distribution of glucagon-like peptide-1 and other preproglucagon-derived peptides in the rat hypothalamus and brainstem. Neuroscience 77:257–270
Lui EY, Asa SL, Drucker DJ, Lee YC, Brubaker PL 1990 Glucagon and related peptides in fetal rat hypothalamus in vivo and in vitro. Endocrinology 126:110–117
Tang-Christensen M, Larsen PJ, Thulesen J, Romer J, Vrang N 2000 The proglucagon-derived peptide, glucagon-like peptide-2, is a neurotransmitter involved in the regulation of food intake. Nat Med 6:802–807
Egea JC, Hirtz C, Deville de Periere D 2003 Preproglucagon mRNA expression in adult rat submandibular glands. Diabetes Nutr Metab 16:130–133
Seino S, Welsh M, Bell GI, Chan SJ, Steiner DF 1986 Mutations in the guinea pig preproglucagon gene are restricted to a specific portion of the prohormone sequence. FEBS Lett 203:25–30
Zhou A, Webb G, Zhu X, Steiner DF 1999 Proteolytic processing in the secretory pathway. J Biol Chem 274:20745–20748[Free Full Text]
Bromer WW 1972 In: Steiner DF, Freinkel N, eds. Chemistry of glucagon and gastrin. Handbook of physiology, Section 7: Endocrinology, Vol 1, Endocrine Pancreas. Washington, DC: American Physiological Society Chap 7:133–138
Orskov C 1992 Glucagon-like peptide-1, a new hormone of the entero-insular axis. Diabetologia 35:701–711
Rouille Y, Westermark G, Martin SK, Steiner DF 1994 Proglucagon is processed to glucagon by prohormone convertase 2 in TC1–6 cells. Proc Natl Acad Sci USA 91:3242–3246
Rouille Y, Martin S, Steiner DF 1995 Differential processing of proglucagon by the subtilisin-like prohormone convertases PC2 and PC3 to generate either glucagon or glucagon-like peptide. J Biol Chem 270:26488–26496
Rouille Y, Bianchi M, Irminger JC, Halban PA 1997 Role of the prohormone convertase PC2 in the processing of proglucagon to glucagon. FEBS Lett 413:119–123
Rouille Y, Kantengwa S, Irminger JC, Halban PA 1997 Role of the prohormone convertase PC3 in the processing of proglucagon to glucagon-like peptide-1. J Biol Chem 272:32810–32816
Furuta M, Zhou A, Webb G, Carroll R, Ravazzola M, Orci L, Steiner DF 2001 Severe defect in proglucagon processing in islet A-cells of prohormone convertase 2 null mice. J Biol Chem 276:27197–27202
Ugleholdt R, Zhu X, Deacon CF, Orskov C, Steiner DF, Holst JJ 2004 Impaired intestinal proglucagon processing in mice lacking prohormone convertase 1. Endocrinology 145:1349–1355
Webb GC, Dey A, Wang J, Stein J, Milewski M, Steiner DF 2004 Altered proglucagon processing in an -cell line derived from prohormone convertase 2 null mouse islets. J Biol Chem 279:31068–31075
Mineo I, Matsumura T, Shingu R, Mamba M, Matsuzawa Y 1995 The role of prohormone convertases PC1(PC3) and PC2 in the cell-specific processing of glucagon. Biochem Biophys Res Commun 207:646–651
Dhanvantari S, Brubaker PL 1998 Proglucagon processing in an islet cell line: effects of PC1 overexpression and PC2 depletion. Endocrinology 139:1630–1637
Dhanvantari S, Seidah NG, Brubaker PL 1996 Role of prohormone convertases in the tissue-specific processing of proglucagon. Mol Endocrinol 10: 342–355
Rothenberg ME, Eilertson CD, Klein K, Mackin RB, Noe BD 1996 Evidence for redundancy in propeptide/prohormone convertase activities in processing proglucagon: an antisense study. Mol Endocrinol 10:331–341
Mojsov S, Weir GC, Habener JF 1987 Insulinotropin: glucagon-like peptide-1 (7–37) co-encoded in the glucagon gene is a potent stimulator of insulin release in the perfused rat pancreas. J Clin Invest 79:616–619
Drucker DJ, Philippe J, Mojsov S, Chick WL, Habener JF 1987 Glucagon-like peptide-1 stimulates insulin gene expression and increases cyclic AMP levels in a rat islet cell line. Proc Natl Acad Sci USA 84:3434–3438
Kreymann B, Ghatei MA, Williams G, Bloom SR 1987 Glucagon-like peptide-1 7–36: a physiological incretin in man. Lancet 2:1300–1304
Orskov C, Holst JJ, Nielsen OV 1988 Effect of truncated glucagon-like peptide-1 [proglucagon-(78–107)amide] on endocrine secretion from pig pancreas, antrum, and nonantral stomach. Endocrinology 123:2009–2013
Orskov C, Bersani M, Johnsen AH, Hojrup P, Holst JJ 1989 Complete sequences of glucagon-like peptide-1 from human and pig small intestine. J Biol Chem 264:12826–12829
Drucker DJ 2001 Glucagon-like peptide-2. J Clin Endocrinol Metab 86:1759–1764
Cameron A, Apletalina EV, Lindberg I 2001 The enzymology of PC1 and PC2. In: Dalbey RE, Sigman, DS, eds. The enzymes. 3rd ed. San Diego: Academic Press; 291–328
Seidah NG, Chretien M 1999 Proprotein and prohormone convertases: a family of subtilases generating diverse bioactive polypeptides. Brain Res 848:45–62
Breslin MB, Lindberg I, Benjannet S, Mathis JP, Lazure C, Seidah NG 1993 Differential processing of proenkephalin by prohormone convertases 1(3) and 2 and furin. J Biol Chem 268:27084–27093
Barbero P, Rovere C, De Bie I, Seidah NG, Beaudet A, Kitabgi, P 1998 PC5-A-mediated processing of pro-neurotensin in early compartments of the regulated secretory pathway of PC5-transfected PC12 cells. J Biol Chem 273:25339–25346
Villeneuve P, Lafortune L, Seidah NG, Kitabgi P, Beaudet A 2000 Immunohistochemical evidence for the involvement of protein convertases 5A and 2 in the processing of pro-neurotensin in rat brain. J Comp Neurol 424:461–475
Schaner P, Todd RB, Siedah NG, Nillni EA 1997 Processing of prothyrotropin-releasing hormone by the family of prohormone convertases. J Biol Chem 272:19958–19968
Viale A, Ortola C, Hervieu G, Furuta M, Barbero P, Steiner DF, Seidah NG, Nahon JL 1999 Cellular localization and role of prohormone convertases in the processing of pro-melanin concentrating hormone in mammals. J Biol Chem 274:6536–6545
Eskeland NL, Zhou A, Dinh TQ, Wu H, Parmer RJ, Mains RE, O’Connor DT 1996 Chromogranin A processing and secretion: specific role of endogenous and exogenous prohormone convertases in the regulated secretory pathway. J Clin Invest 98:148–156
Dey A, Norrbom C, Zhu X, Stein J, Zhang C, Ueda K, Steiner DF 2004 Furin and prohormone convertase 1/3 are major convertases in the processing of mouse pro-growth hormone-releasing hormone. Endocrinology 145:1961–1971
Sayah M, Fortenberry Y, Cameron A, Lindberg I 2001 Tissue distribution and processing of proSAAS by proprotein convertases. J Neurochem 76:1833–1841
Mzhavia N, Qian Y, Feng Y, Che FY, Devi LA, Fricker LD 2002 Processing of proSAAS in neuroendocrine cell lines. Biochem J 361:67–76
Paquet L, Bergeron F, Boudreault A, Seidah NG, Chretien M, Mbikay M, Lazure C 1994 The neuroendocrine precursor 7B2 is a sulfated protein proteolytically processed by a ubiquitous furin-like convertase. J Biol Chem 269:19279–19285
Hendy GN, Bennett HPJ, Gibbs BF, Lazure C, Day R, Seidah NG 1995 Proparathyroid hormone is preferentially cleaved to parathyroid hormone by the prohormone convertase furin. A mass spectrometric study. J Biol Chem 270:9517–9525
Cain BM, Vishnuvardhan D, Beinfeld MC 2001 Neoronal cell lines expressing PC5, but not PC1 or PC2, process pro-CCK into glycine-extended CCK 12 and 22. Peptides 22:1271–1277
Cain BM, Vishnuvardhan D, Wang W, Foulon T, Cadel S, Cohen P, Beinfeld MC 2002 Production, purification and characterization of recombinant prohormone convertase 5 from bacuolovirus-infected insect cells. Protein Expr Purif 24:227–233
Galanopoulou AS, Seidah NG, Patel YC 1995 Direct role of furin in mammalian prosomatostatin processing. Biochem J 309:33–40
Bell GI, Santerre RF, Mullenbach GT 1983 Hamster proglucagon contains the sequence of glucagon and two related peptides. Nature 302:716–718
Deng WP, Nickoloff JA 1992 Site-directed mutagenesis of virtually any plasmid by eliminating a unique site. Anal Biochem 200:81–88
Dey A, Zhu X, Carroll R, Turck CW, Stein J, Steiner DF 2003 Biological processing of the cocaine and amphetamine-regulated transcript precursors by prohormone convertases, PC2 and PC1/3. J Biol Chem 278:15007–15014
Schagger H, Jagow GV 1987 Tricine-sodium dodecyl sulfate-polyacrylamide gel electrophoresis for the separation of proteins in the range from 1 to 100 kDa. Anal Biochem 166:368–379
Lipkind G, Gong Q, Steiner DF 1995 Molecular modeling of the substrate specificity of prohormone convertases SPC2 and SPC3. J Biol Chem 270:13277–13284
Henrich S, Cameron A, Bourenkov GP, Kiefersauer R, Huber R, Lindberg I, Bode W, Than ME 2003 The crystal structure of the proprotein processing proteinase furin explains its stringent specificity. Nat Struct Biol 10:520–526
Sasaki K, Dockerill S, Adamiak DA, Tickle IJ, Blundell T 1975 X-ray analysis of glucagon and its relationship to receptor biding. Nature 257:751–757
Creighton TE 1993 Proteins: structures and molecular properties. 2nd ed. New York: WH Freeman and Co.
Seidah NG, Chretien M, Day R 1994 The family of subtilisin/kexin like pro-protein and pro-hormone convertases: divergent or shared functions. Biochimie (Paris) 76:197–209
Posner SF, Vaslet CA, Jurofcik M, Lee A, Seidah NG, Nillni EA 2004 Stepwise posttranslational processing of progrowth hormone-releasing hormone (pro GHRH) polypeptide by furin and PC1. Endocrine 23:199–213
Pohl SL, Birnbaumer L, Rodbell M 1969 Glucagon-sensitive adenyl cyclase in plasma membrane of hepatic parenchymal cells. Science 164:566–569
Panijpan B, Gratzer WB 1974 Conformational nature of monomeric glucagon. Eur J Biochem 45:547–553
Dalle S, Fontes G, Lajoix AD, LeBrigand L, Gross R, Ribes G, Dufour M, Barry L, LeNguyen D, Bataille D 2002 Miniglucagon (glucagon 19–29), a novel regulator of the pancreatic islet physiology. Diabetes 51:406–412
Efimov AV 1993 Standard structures in proteins. Prog Biophys Mol Biol 60:201–239
Holyoak T, Wilson MA, Fenn TD, Kettner CA, Petsko GA, Fuller RS, Ringe D 2003 A resolution crystal structure of the prototypical hormone-processing protease Kex2 in complex with an Ala-Lys-Arg boronic acid inhibitor. Biochemistry 42:6709–6718
McPhalen CA, James MN 1988 Structural comparison of two serine proteinase-protein inhibitor complexes: eglin-c-subtilisin Carlsberg and CI-2-subtilisin Novo. Biochemistry 27:6582–6598
Furuta M, Yano H, Zhou A, Rouille Y, Holst JJ, Carroll R, Ravazzola M, Orci L, Furuta H, Steiner DF 1997 Defective prohormone processing and altered pancreatic islet morphology in mice lacking active SPC2. Proc Natl Acad Sci USA 94:6646–6651
Westphal CH, Muller L, Zhou A, Zhu X, Bonner-Weir S, Schambelan M, Steiner DF, Lindberg I, Leder P 1999 The neuroendocrine protein 7B2 is required for peptide hormone processing in vivo and provides a novel mechanism for pituitary Cushing’s disease. Cell 96:689–700
Zhu X, Zhou A, Dey A, Norrbom C, Carroll R, Zhang C, Laurent V, Lindberg I, Ugleholdt R, Holst JJ, Steiner DF 2002 Disruption of PC1/3 expression in mice causes dwarfism and multiple neuroendocrine peptide processing defects. Proc Natl Acad Sci USA 99:10293–10298
Dupuy A, Lindberg I, Zhou A, Akil H, Lazure C, Chretien M, Seidah NG, Day R 1994 Processing of prodynorphin by the prohormone convertase PC1 results in high molecular weight intermediate forms. FEBS Lett 337:60–65
Jackson RS, Creemers JWM, Farooqi S, Raffin-Sanson ML, Varro A, Dockray GJ, Holst JJ, Brubaker PL, Corvol P, Polonsky KS, Ostrega D, Becker KL, Bertagna X, Hutton JC, White A, Dattani MT, Hussain K, Middleton SJ, Nicole TM, Milla PJ, Lindley KJ, O’Rahilly S 2003 Small-intestinal dysfunction accompanies the complex endocrinopathy of human proprotein convertase 1 deficiency. J Clin Invest 112:1550–1560
Lusson J, Vieau D, Hamelin J, Day R, Chretien M, Seidah NG 1993 cDNA structure of the mouse and subtilisin/kexin-like PC5: a candidate proprotein convertase expressed in endocrine and neuroendocrine cells. Proc Natl Acad Sci USA 90:6691–6695
Nakagawa T, Hosaka M, Torii S, Watanabe T, Murakami K, Nakayama K 1993 Identification and functional expression of a new member of the mammalian kex2-like processing endoprotease family: its striking structural similarity to PACE4. J Biochem 113:132–135
Thomas G 2002 Furin at the cutting edge: from protein traffic to embryogenesis and disease. Nat Rev Mol Cell Biol 3:753–766
Keller P, Zecca L, Boukamel R, Zwicker E, Gloor S, Semenza G 1995 Furin, PC1/3, and/or PC6A process rabbit, but not human, pro-lactase-phlorizin hydrolase to the 180-kDa intermediate. J Biol Chem 270:25722–25728
De Bie I, Marcinkiewicz M, Malide D, Lazure C, Nakayama K, Bendayan M, Seidah NG 1996 The isoforms of proprotein convertase PC5 are sorted to different subcellular compartments. J Cell Biol 135:1261–1275
Tanaka S, Kurabuchi S, Mochida H, Kato T, Takahashi S, Watanabe T, Nakayama K 1996 Immunocytochemical localization of prohormone convertases PC1/PC3 and PC2 in rat pancreatic islets. Arch Histol Cytol 59:261–271
Damholt AB, Buchan AM, Holst JJ, Kofod H 1999 Proglucagon processing profile in canine L cells expressing endogenous prohormone convertase 1/3 and prohormone convertase 2. Endocrinology 140:4800–4888(Arunangsu Dey, Gregory M.)