Structure and expression of mouse furin, a yeast Kex2-related protease. Lack of processing of coexpressed prorenin in GH4C1 cells

Identification of chondromodulin I as a novel endothelial cell growth inhibitor. Purification and its localization in the avascular zone of epiphyseal cartilage

Cartilage is unique among tissues of mesenchymal origin in that it is resistant to vascular invasion due to an intrinsic angiogenic inhibitor. During endochondral bone formation, however, calcified cartilage formed in the center of the cartilaginous bone rudiment allows vascular invasion, which initiates the replacement of cartilage by bone. The transition of cartilage from the angioresistant to the angiogenic status thus plays a key role in bone formation. However, the molecular basis of this phenotypic transition of cartilage has been obscure. We report here purification of an endothelial cell growth inhibitor from a guanidine extract of bovine epiphyseal cartilage. The N-terminal amino acid sequence indicated that the inhibitor was identical to chondromodulin I (ChM-I), a cartilage-specific growth-modulating factor. Purified ChM-I inhibited DNA synthesis and proliferation of vascular endothelial cells as well as tube morphogenesis in vitro. Expression of ChM-I cDNA in COS7 cells indicated that mature ChM-I molecules were secreted from the cells after post-translational modifications and cleavage from the transmembrane precursor at the predicted processing signal. Recombinant ChM-I stimulated DNA synthesis and proteoglycan synthesis of cultured growth plate chondrocytes, but inhibited tube morphogenesis of endothelial cells. In situ hybridization and immunohistochemical studies indicated that ChM-I is specifically expressed in the avascular zone of cartilage in developing bone, but not present in calcifying cartilage. These results suggest a regulatory role of ChM-I in vascular invasion during endochondral bone formation.

Mechanism of the facilitation of PC2 maturation by 7B2: involvement in ProPC2 transport and activation but not folding

Among the members of the prohormone convertase (PC) family, PC2 has a unique maturation pattern: it is retained in the ER for a comparatively long time and its propeptide is cleaved in the TGN/ secretory granules rather than in the ER. It is also unique by its association with the neuroendocrine protein 7B2. This interaction results in the facilitation of proPC2 maturation and in the production of activatable proPC2 from CHO cells. In the present study, we have investigated the mechanism of this interaction. ProPC2 binds 7B2 in the ER, but exits this compartment much more slowly than 7B2. We found that proPC2 was also slow to acquire the capacity to bind 7B2, whereas 7B2 could bind proPC2 rapidly after synthesis. This indicated that proPC2 folding was the limiting step in the formation of the complex. Indeed, sensitivity of native proPC2 to N-glycanase F digestion and inhibition of proPC2 folding supported the notion that 7B2 is not involved in the early steps of proPC2 folding, and that proPC2 must fold before binding 7B2. Under experimental conditions that prevent propeptide cleavage, 7B2 expression increased proPC2 transport to the Golgi. This increase exhibited the same kinetics as the facilitation of the removal of the propeptide. Finally, proPC2 activation could be reconstituted in Golgi- enriched subcellular fractions. In vitro, 7B2 was required for proPC2 activation at an acidic pH. Taken together, our results demonstrate that rather than promoting proPC2 folding, 7B2 acts as a helper protein involved in proPC2 transport and is required in the proPC2 activation process. We propose, therefore, that 7B2 stabilizes proPC2 in a conformation already competent for these two events.

Stretch-induced hypertrophic growth of cardiocytes and processing of brain-type natriuretic peptide are controlled by proprotein-processing endoprotease furin

When hypertrophic growth is induced in neonatal rat cardiocytes by stretching, the cardiocytes express high levels of brain-type natriuretic peptide (BNP) and the proprotein-processing enzyme furin. A BNP precursor, gammaBNP, possesses a furin-cleavable Arg-X-X-Arg motif, which is cleaved when gammaBNP is processed to form BNP-45. The Arg-X-X-Arg motif is found in many precursors of growth factors and growth-related proteins. To determine if furin converts gammaBNP to BNP-45 as well as other unidentified growth-promoting protein precursors to their active form that may induce hypertrophic growth in cardiocytes, we used two protease inhibitor systems, synthetic peptidyl chloromethyl ketones (CMK) (dec-Arg-Val-Lys-Arg-CMK and dec-Phe-Ala-Lys-Arg-CMK; where dec is decanoyl) and vaccinia vector-integrated native and variant alpha1-antitrypsins. The furin-specific inhibitors, dec-Arg-Val-Lys-Arg-CMK and variant alpha1-antitrypsin with the inhibitory determinant Arg-X-X-Arg, suppressed the stretch-induced hypertrophic growth of cardiocytes as well as the processing of gammaBNP to BNP-45. The other serine protease inhibitors and variant alpha1-antitrypsin against elastase, or thrombin, however, neither suppressed the hypertrophic growth nor prevented the processing of gammaBNP to BNP-45. Thus, we suggest that furin catalyzes the conversion of gammaBNP to BNP-45 as well as growth-promoting proproteins to their active form, which might induce hypertrophic growth in cardiocytes.

Secretion of human furin into mouse milk

We have previously described the expression of the human proprotein convertase furin or paired basic amino acid-cleaving enzyme, in mice transgenic for paired basic amino acid-cleaving enzyme and human Protein C (HPC). Here we show 100-fold or higher expression of furin in the mammary gland, compared with endogenous furin. Furin and recombinant HPC were detected in the same regions of the mammary gland and regulated similar to the endogenous whey acidic protein. In addition to the expected intracellular localization, furin was secreted into the milk as an 80-kDa form lacking the transmembrane and cytoplasmic domains. Furin present at levels of up to 40,000 units/ml milk cleaved the t-butoxycarbonyl-RVRR-AMC substrate with a Km of 32 microM, and processed the recombinant HPC precursor at the appropriate sites. Surprisingly, the expression of an active protease was not toxic to the mammary gland. This is a rare example of an animal model secreting active truncated forms of a processing endoprotease into a bodily fluid.

Tissue targeting of angiotensin peptides

Angiotensin II (Ang II) is an octapeptide generated by the sequential proteolytic action of renin and angiotensin converting enzyme on the glycoprotein angiotensinogen. While numerous mammalian tissues have been shown to express some or all of the components of the renin-angiotensin system (RAS), the function of most of these tissue RAS remains a matter of conjecture. To test for tissue-specific functions of Ang II and as an alternative to co-expressing all the components of RAS, we have engineered a fusion protein that leads to direct Ang II release within specific tissues. The angiotensin peptide is cleaved from the fusion protein within the secretory pathway by the ubiquitous endoprotease furin and is released from the cell by constitutive secretion. Direct injection of an expression vector encoding such a fusion protein into rat cardiac ventricles results in a highly localized expression of atrial natriuretic peptide mRNA (an angiotensin responsive marker of cardiac hypertrophy), demonstrating the utility of this approach for local targeting of mature peptides to tissues in animal models.

Ontogeny of prohormone convertases in rat prenatal development

It has been well established that peptide precursors usually undergo limited proteolysis at pairs or single basic amino acids during their biosynthetic process. This posttranslational modification paradigm is common for numerous membrane-spanning and secreted proteins, neuropeptides, and peptide hormones of physiological significance, in which endoproteolytic cleavage is invariably essential for the accurate biosynthesis and full activity of the mature products. Establishment of an effective peptide profile is dependent on not only the presence of peptide precursor, but also the presence and the enzymatic specificities of cleavage enzymes. We have, therefore, characterized the spatial and temporal patterns of six subtilisin-like serine endoproteases known to be involved in proprotein processing, including furin, PC1, PC2, PC4, PC5, and PACE4, in rat prenatal development and related the results to the expression patterns of several peptide precursors. We have observed largely distinct and sometimes complementary expression patterns of individual PCs in various embryonic structures, suggesting PCs may be functionally distinct in processing different sets of proprotein substrates in development. From these studies, numerous tentative enzyme-substrate relationships in various embryonic structures have been proposed and should encourage more studies to test the in vitro cleavage potentialities of individual PCs toward these precursors. In the future, knowledge gained from these studies, when combined with insights gained from in vivo perturbation and genetic ablation studies, should lead to final comprehensive understanding of specific precursors cleaved by specific enzymes at specific cleavage sites in known spatial and temporal expression patterns during development.

Dexamethasone induces neuropeptide Y (NPY) expression and impairs insulin release in the insulin-producing cell line RINm5F. Release of NPY and insulin through different pathways

Neuropeptide Y (NPY) occurs in adrenergic as well as in non-adrenergic nerves innervating the islets of Langerhans and inhibits glucose-stimulated insulin secretion. Recently we demonstrated that NPY is expressed within islet beta cells of the rat pancreas following treatment with dexamethasone in vivo. In this study we examined the cellular expression of NPY following dexamethasone treatment of the insulin-producing cell line RINm5F, which under control conditions does not express or release NPY. The cells were cultured with or without dexamethasone (100 nM) for 5 days. Over the 5-day culture period, dexamethasone time dependently induced an increased release of NPY with a concomitant decrease in the release of insulin. Northern blot and in situ hybridization revealed a corresponding time-dependent increase in the amount of NPY transcripts and in the number of cells labeled for NPY mRNA, whereas immunocytochemistry for NPY revealed only a few immunoreactive cells, indicating a rapid release of the formed peptide. Following 5 days of culture with dexamethasone, acute stimulation with D-glyceraldehyde (10 mM) or KCl (20 mM) Ca2+ dependently stimulated the release of insulin. In contrast neither stimulation with D-glyceraldehyde or KCl nor removal of extracellular Ca2+ affected the release of NPY. Furthermore the D-glyceraldehyde- and KCl-induced increase in cytosolic Ca2+, evident in control RINm5F cells, was impaired after dexamethasone treatment. We conclude that RINm5F cells show steroid-sensitive plasticity and express NPY after dexamethasone treatment concomitantly with a decreased insulin secretion and impaired increase in cytosolic Ca2+ upon depolarization with KCl or stimulation with D-glyceraldehyde. We also conclude that NPY and insulin secretion are regulated differently and suggest that the inability of the removal of extracellular Ca2+ to inhibit NPY secretion and the failure of D-glyceraldehyde and KCl to stimulate NPY secretion reflect a constitutive release of this peptide from the cells in contrast to the regulated release of insulin.

Processing of proendothelin-1 at the C-terminus of big endothelin-1 is essential for proteolysis by endothelin-converting enzyme-1 in vivo

Production of endothelin-1 is thought to be a three-step process consisting of an initial proteolytic cleavage of the proendothelin-1 precursor to big endothelin-1-Lys-Arg, C-terminal trimming by a carboxypeptidase and further processing of the big endothelin-1 peptide to endothelin-1 by endothelin-converting enzyme (ECE). To further clarify the mechanism of processing in the biosynthesis of endothelin-1, we introduced a point mutation into endothelin-1 cDNA to replace the Arg in the -4 position of the recognition motifs of furin-like convertase in human preproendothelin-1 (Arg49 or Arg89) by Gly. When mutant cDNAs were expressed in Chinese hamster ovary (CHO)-K1 cells, they failed to be processed at the mutated processing signal, suggesting that the Arg-Ser-Lys-Arg motifs of preproendothelin-1 are recognized by CHO-K1 furin-like convertase. Co-transfection with ECE-1 cDNA revealed that cleavage at Arg52 is not essential for cleavage by ECE-1, but that cleavage at Arg92 is critical. Although a high-molecular-mass form of endothelin-1 is produced by processing by ECE-1 without cleavage at Arg52, it did not evoke Ca2+ transient in endothelinA-receptor-expressing cells. In conclusion, prior cleavage at Arg92 by furin-like convertase is absolutely necessary for cleavage by ECE-1 at Trp73 to produce mature endothelin-1.

Production of bioactive salmon calcitonin from the nonendocrine cell lines COS-7 and CHO

To produce bioactive salmon calcitonin from the conventional nonendocrine cell lines, COS-7 and CHO, we devised a salmon calcitonin expression vector by combining the amino-terminus of human calcitonin precursor with a salmon calcitonin sequence, inserting the efficient furin-cleavable processing sequence Arg-X-Arg-X-Lys-Arg before salmon calcitonin, and deleting the carboxyl-terminal extension peptide. This chimeric calcitonin precursor terminates at glycine to easily receive an amidation reaction. COS-7 and CHO produced a high level of bioactive calcitonin by the resorption pit formation assay. Although amidating activity is highly expressed in CHO, but only a little in COS-7 cells, both cells produced a similar level of bioactive calcitonin. Thus, the engineered salmon calcitonin expression vector enables nonendocrine cells even with low amidation activity to produce bioactive calcitonin.

Localization of endogenous furin in cultured cell lines

Furin is a dibasic endopeptidase responsible for the proteolytic maturation of many precursor proteins in the secretory and endocytic pathways of mammalian cells. The levels of furin expression in most cells are very low, and this has hampered attempts to identify the intracellular compartments in which endogenous furin is localized. We have used a specific antibody reagent to a sequence in the carboxy terminus of furin to perform immunofluorescent staining of mammalian cell lines. This antibody was sensitive enough to detect staining for furin in various cell lines. For the most part, furin staining was confined to a juxtanuclear structure characteristic of the Golgi complex. Analyses by video microscopy and confocal microscopy showed that the distribution of furin was distinct from that of mannosidase II, a marker of the Golgi stack, and most closely resembled that of TGN38, a marker of the trans-Golgi network. Therefore, our results suggest that endogenous furin is predominantly localized to the area of the Golgi complex, most likely within the trans-Golgi network.

Structural and physiologic characterization of the mid-region secretory species of parathyroid hormone-related protein

Parathyroid hormone-related protein (PTHrP) is initially translated as a preprohormone which is posttranslationally processed to yield a family of mature secretory forms. Most attention has focused on the amino-terminal portion of the molecule which is homologous to parathyroid hormone. It is clear, however, that a mid-region species of PTHrP is posttranslationally cleaved from the highly conserved mid-region of PTHrP, and that the amino terminus of this peptide is Ala38. The purposes of the current study were three: 1) to confirm that Arg37 immediately preceding Ala38 serves as a posttranslational processing site in the PTHrP precursor, 2) to determine the carboxyl terminus of the mid-region secretory species of PTHrP, and 3) to synthesize this authentic mid-region secretory form of PTHrP and determine whether it is biologically active. The results indicate that: 1) Arg37 is indeed a processing site in the PTHrP precursor; 2) three distinct mid-region PTHrP species are generated by posttranslational processing, PTHrP(38-94)amide, PTHrP(38-95), and most likely, PTHrP(38-101); and 3) synthetic mid-region PTHrP(38-94)amide is active in four different biological systems. These studies confirm the finding that PTHrP is a prohormone. More importantly, they define a novel, biologically active highly conserved mid-region secretory form of PTHrP.

Activation of a recombinant membrane type 1-matrix metalloproteinase (MT1-MMP) by furin and its interaction with tissue inhibitor of metalloproteinases (TIMP)-2

Membrane type 1-matrix metalloproteinase (MT1-MMP) initiates the activation of the zymogen progelatinase A/ 72-kDa type IV collagenase by cleavage of the Asn66-Leu peptide bond. We previously pointed out that MT1-MMP possesses a unique amino acid sequence Arg-Arg-Lys-Arg111 which is a potential recognition sequence for furin-like proteases (Nature, 370 (1994) 61-65). Here, using a recombinant MT1-MMP expressed in Escherichia coli we demonstrated that furin specifically cleaves MT1-MMP between Arg111-Tyr in vitro, which resulted in a stimulation of progelatinase A-activation function. Tissue inhibitor of metalloproteinases (TIMP)-2 inhibited activation of progelatinase A by forming a stable complex with activated MT1-MMP.

A protease processing site is essential for prorenin sorting to the regulated secretory pathway

Transfected mouse pituitary AtT-20 cells were used to examine the sorting of human prorenin to dense core secretory granules and the regulated secretory pathway. These cells secrete prorenin constitutively and sort a portion of the prorenin to secretory granules, where it is converted to active renin by proteolytic processing. Pulse-chase labeling of transfected AtT-20 cells demonstrated that regulated secretion of prorenin was prevented by: 1) the mutagenic deletion of the prosegment, 2) the premature proteolytic removal of the prosegment by a Golgi-resident processing protease, or 3) the mutation of the native cleavage site so as to prevent removal of the prosegment. In addition, expression of fusion proteins containing portions of the prorenin prosegment demonstrated that exposure of potential proteolytic cleavage sites was sufficient to confer cleavage-dependent regulated secretion of the corresponding protein. These data implicate the protease cleavage event in the regulated secretion of prorenin and are consistent with the involvement of a subclass of processing proteases in the sorting of certain proteins to secretory granules in AtT-20 cells.

Human lactase-phlorizin hydrolase is not processed by furin, PC1/PC3, PC2, PACE4 and PC5/PC6A of the family of subtilisin-like proprotein processing proteases

Human lactase-phlorizin hydrolase (LPH, EC 3.2.1.23/62) is synthesized as a single-chain precursor glycoprotein (pro-LPH) with a relative molecular mass of just over 200 kDa. Maturation to the mature enzyme (m-LPH, 160 kDa) occurs after passage of pro-LPH through the Golgi complex and involves the proteolytic removal of a 849 amino acid propeptide. The role of this propeptide as well as its removal is not fully understood and the proteolytic enzyme or enzymes involved are unknown. We studied the potential role of five different members of the family of subtilisin-like proprotein processing proteases in the maturation process of human LPH using a vaccinia virus based coexpression system in pig kidney PK(15) cells. Infected/transfected PK(15) cells expressed full-length pro-LPH but no maturation to m-LPH was observed. Coexpression of human pro-LPH with human furin, human PC1/PC3, human PC2, human PACE4 and mouse PC6A in PK(15) cells did not result in maturation of the enzyme. Cleavage and secretion of von Willebrand factor precursor (pro-vWF) was used as a positive control. None of the five proprotein processing proteases tested were capable of cleaving human pro-LPH, strongly suggesting that they are not involved in the maturation of this enzyme.

Limulus kexin: a new type of Kex2-like endoprotease specifically expressed in hemocytes of the horseshoe crab

A Kex2-like protease was identified in hemocytes of the horseshoe crab (Tachypleus tridentatus), named limulus kexin, and a full-length cDNA was obtained from a hemocyte cDNA library. The deduced amino acid sequence contains 752 residues, composed of five domains with a signal sequence, a propeptide, a catalytic domain, a Ser/Thr-rich domain, and a transmembrane domain. The domain organization is very similar to that of the yeast Kex2 except that limulus kexin does not have a cytoplasmic tail. The catalytic domain exhibits striking sequence identities with those of furins, especially Drosophila furin1 (79%). Northern blotting showed specific expression of limulus kexin in hemocytes, suggesting the involvement in proteolytic processing of the granule components of hemocytes.

Proprotein-processing endoprotease furin decreases regulated secretory pathway-specific proteins in the pancreatic beta cell line MIN6

Prohormone convertases PC2 and PC3, yeast Kex2-family endoproteases specific to the regulated secretory pathway, cleave proinsulin to insulin in the secretory granules of pancreatic beta cells. The well-differentiated beta cell line MIN6 expresses PC2 and PC3 and another regulated secretory pathway-specific protein chromogranin A. Furin, another yeast Kex2 endoprotease, exists in the trans-Golgi networks of many cell types. The beta cell line RINm5F (a cell line that is less differentiated than the MIN6 cell line) does not express the regulated pathway-specific proteins, but strongly expresses furin. We suspected that furin expression may cause the decrement of regulated secretory pathway-specific proteins. To test this hypothesis, we expressed a furin cDNA with a metallothionein promoter in MIN6 cells. With Zn2+ stimulation of furin expression, the messages of PC2, PC3, and chromogranin A decreased, and the processing of proinsulin to mature insulin became less efficient. The furin-expressing MIN6 cells exhibited less insulin content and weakened insulin secretion in response to a high glucose concentration. The conditioned medium from furin-expressing MIN6 cells also exerted a decrease of PC2 and PC3 expression in unaltered MIN6 cells. Thus, proteins cleaved by furin inside the cells or by truncated furin shed into the culture medium appear to cause decreased PC2 and PC3 expression, insulin content, and glucose-responsive insulin secretion in MIN6 cells.

Cathepsin B is a prorenin processing enzyme

Conversion of prorenin to renin results from proteolytic cleavage of a 43-amino-acid prorenin prosegment in renal juxtaglomerular cells. The enzyme that performs this processing is not known. Of several enzymes proposed, cathepsin B is a candidate because it colocalizes with renin in juxtaglomerular cell secretory granules and accurately cleaves the prosegment of human prorenin in vitro. It is not known whether cathepsin B can perform this function in the cell. We examined this using secretory granule-containing rat GH4C1 cells transfected with a human preprorenin expression vector. When treated with secretagogue (KCl 50 mmol/L + forskolin 10 micromol/L), these cells secrete 95% prorenin and 5% active renin into the medium, indicating little prorenin processing activity. In contrast, when the cells are cotransfected with a vector that expresses human preprocathepsin B or mouse prohormone convertase 1, secretagogue-induced secretion of active renin increased to 12% and 16.5%, respectively. With antisera that recognize the prosegment and renin, prorenin and renin were identified as proteins of 47 and 43 kD, respectively, and an antibody specific to the prosegment precipitated only the 47-kD species. These results do not address whether cathepsin B is the authentic renal prorenin processing enzyme. However, the results do demonstrate that cathepsin B can localize to the appropriate subcellular compartment and process prorenin to renin in GH4C1 cells and are consistent with a role for this enzyme in prorenin processing.

Furin/PACE/SPC1: a convertase involved in exocytic and endocytic processing of precursor proteins

One of the most exciting breakthroughs of the 90's in the fields of biochemistry, cell biology and neuroendocrinology is the identification of a novel family of proteolytic enzymes called mammalian subtilisin-like convertases. This family is comprised so far of seven distinct endoproteases responsible for the proteolytic excision of biologically active polypeptides from inactive precursor proteins. Six years after the initial observation of a structural conservation between a characterized yeast enzyme (kexin) and a human gene product (furin), it is now well accepted that one of these convertases, furin, has the enzymatic capabilities to efficiently and correctly process a great variety of precursors. Furin's ability to cleave precursors within both the exocytic and endocytic pathways will require sustained efforts in order to delineate all of its physiological roles.

Processing of mutated human proinsulin to mature insulin in the non-endocrine cell line, CHO

Heterologous genes encoding proproteins, including proinsulin, generally produce mature protein when expressed in endocrine cells while unprocessed or partially processed protein is produced in non-endocrine cells. Proproteins, which are normally processed in the regulated pathway restricted to endocrine cells, do not always contain the recognition sequence for cleavage by furin, the endoprotease specific to the constitutive pathway, the principal protein processing pathway in non-endocrine cells. Human proinsulin consists of B-Chain-C-peptide-A-Chain and cleavage at the B/C and C/A junctions is required for processing. The B/C, but not the C/A junction, is recognised and cleaved in the constitute pathway. We expressed a human proinsulin and a mutated proinsulin gene with an engineered furin recognition sequence at the C/A junction and compared the processing efficiency of the mutant and native proinsulin in Chinese Hamster Ovary cells. The processing efficiency of the mutant proinsulin was 56% relative to 0.7% for native proinsulin. However, despite similar levels of mRNA being expressed in both cell lines, the absolute levels of immunoreactive insulin, normalized against mRNA levels, were 18-fold lower in the mutant proinsulin-expressing cells. As a result, there was only a marginal increase in absolute levels of insulin produced by these cells. This unexpected finding may result from preferential degradation of insulin in non-endocrine cells which lack the protection offered by the secretory granules found in endocrine cells.

Proinsulin cleaved by furin is processed to chromatographically mature insulin by carboxypeptidases in nonneuroendocrine cells

Proinsulin is converted to mature insulin by two reactions, cleavage by the prohormone convertases PC2 and PC3, and removal of basic residues by carboxypeptidase H. These reactions are performed in the secretory granules of pancreatic beta cells. When we replaced the processing sites of proinsulin with furin-cleavable sites, the three nonneuroendocrine cell lines Hep G2, CHO, and NIH/3T3 produced insulin with the same size as synthetic human insulin. Although the three cell lines expressed different quantities of carboxypeptidase H mRNA, the cytosol fractions of the cells exhibited similar levels of carboxypeptidase activity, suggesting that additional carboxypeptidases were active. The insulins resulting from the three cell lines were eluted as a single peak on a cation-exchange chromatography column, indicating that proinsulin can be maturated to insulin even in nonneuroendocrine cells.

Localization of furin to the trans-Golgi network and recycling from the cell surface involves Ser and Tyr residues within the cytoplasmic domain

Furin is a membrane-associated endoprotease that catalyzes cleavage of precursor proteins at Arg-X-Lys/Arg-Arg sites. Although, at steady state, furin is predominantly found in the trans-Golgi network (TGN), it also cycles between the TGN and the cell surface. Recently, the cytoplasmic tail of furin has been shown to be sufficient for its localization to the TGN. Within the cytoplasmic domain, there are Ser residues, which we now show are sites for phosphorylation by casein kinase II in vitro, and a Tyr-containing sequence, both of which have been shown to be important for other TGN proteins to localize to this compartment. In the present study, we show by site-directed mutagenesis that these residues are important for TGN localization and recycling of furin. Mutation of the Ser residues abrogated the TGN localization. By contrast, mutation of the Tyr residue did not affect the TGN localization but impaired the internalization from the plasma membrane. These observations suggest that distinct cytoplasmic determinants are responsible for retention in the TGN and retrieval from the cell surface of furin.

Furin, PC1/3, and/or PC6A process rabbit, but not human, pro-lactase-phlorizin hydrolase to the 180-kDa intermediate

Small intestinal lactase-phlorizin hydrolase (LPH) is synthesized as a large precursor (prepro-LPH) of 1926 amino acids. In the endoplasmic reticulum, prepro-LPH is split by signal protease. The resulting pro-LPH is cut to mature LPH directly (human) or via a 180-kDa intermediate (rabbit), most likely in the trans-Golgi network or in a later compartment. Antibodies directed against different regions of rabbit pro-LPH locate the cleavage site resulting in the 180-kDa intermediate between amino acid residues 79 and 286. This stretch contains the two sequences -Arg-Cys-Tyr-Arg114 approximately -Arg-Ala-Ser-Arg191 approximately, which are potential cleavage sites for subtilisin-like proprotein convertases. These sites are not conserved in human pro-LPH. By coexpression in COS 7 cells of rabbit prepro-LPH and proprotein convertases (PC 1/3, PC2, PC6A, PC6B, furin), we show that furin, PC 1/3, and PC6A generate a processing intermediate that is immunologically indistinguishable from the one observed in vivo. Furin, PC 1/3, and PC6A are all expressed in the small intestine as shown by a polymerase chain reaction-based approach and, more specifically, in enterocytes, as shown by in situ hybridization. These results suggest that furin, PC 1/3, and/or PC6A are responsible for the in vivo processing of rabbit pro-LPH to the 180-kDa intermediate.

Characterization of multiple prohormone convertase PC1/3 transcripts in porcine ovary

Overlapping cDNAs encoding porcine prohormone convertase, PC1/3, have been isolated from a pregnant sow ovary cDNA library using a mouse PC1/3 cDNA as a probe. Nucleotide sequence analysis of these cDNAs predicts a PC1/3 precursor protein of 753 amino acid residues, which shares an overall sequence homology of 96, 92, and 92% with the human, rat, and mouse counterparts, respectively. Furthermore, five different polyadenylation sites have been observed. The utilization of these polyadenylation sites results in a length difference of 40-440 bp in the 3' untranslated regions of the transcripts.

Strain-specific presence of two TGN38 isoforms and absence of TGN41 in mouse

TGN38 and TGN41 are isoforms of an integral membrane protein that is predominantly localized to the trans-Golgi network (TGN) in rat cells. They have been proposed to form a heterodimer and to be involved in the budding of exocytic transport vesicles from the TGN. By cDNA cloning and analysis using polymerase chain reaction, we found that there were two TGN38 isoforms in a strain of mouse (ICR), whereas other strains examined (BALB/c, DBA/2, and C57BL/6) had only one TGN38. The major difference between the two isoforms was in the number of characteristic octapeptide repeats. Apart from this, there were several nucleotide substitutions between them. The two isoforms appeared to be derived from two distinct genes but not from one gene via alternative splicing. Furthermore, we failed to show the presence of TGN41 in all the strains examined. This result suggests that TGN38 may function as a monomer or a homodimer in mouse cells.

Furin is important but not essential for the proteolytic maturation of gp160 of HIV-1

The envelope glycoproteins of HIV are required for viral infectivity. Proteolysis of the precursor envelope glycoprotein gp160 results in the formation of gp120 and gp41. Cleavage occurs after the sequence Arg-Glu-Lys-Arg. This sequence is expected to be a substrate for the cellular protease furin. We examined whether furin is responsible for cleavage of gp160 by using a furin-deficient CHO cell line and the same cell line transfected with furin cDNA. Data obtained from viral transmission assays suggested that furin increased viral infectivity but was not essential for the maturation of gp160, implying that other proprotein processing enzymes also recognize this putative furin cleavage site.

Molecular cloning and cellular localization of a furin-like prohormone convertase from the atrial gland of Aplysia

Prohormone convertases (PCs) are Ca(2+)-dependent subtilisin-related endoproteases that have been implicated in the post-translational processing of prohormones and other proproteins. Furin is an ubiquitously expressed PC that has been shown to hydrolyze a wide variety of precursor proteins in secretory pathways. We have screened an Aplysia atrial gland cDNA library using a furin probe prepared by polymerase chain reaction (PCR) and have isolated an Aplysia furin-related 6.7-kb cDNA partial clone that was truncated on the 5' end. The remaining 5' atrial gland furin nucleotide sequence was obtained by two stages of reverse transcription PCR. The final composite nucleotide sequence of the atrial gland furin cDNA was 7,837 bp in length. This sequence encoded a putative preproendoprotease (Afurin2) of 824 amino acid residues that was related to other eukaryotic furins, and showed a high sequence identity with a recently reported Aplysia nervous system furin-like sequence. In situ hybridization demonstrated extensive expression of Afurin2 in atrial gland secretory cells. The cDNA clone contained a relatively long 3' untranslated region of 5,230 nucleotides that included a microsatellite repeat region (TG)n. The characterized Aplysia Afurin2 is a candidate PC that may play an important role in the processing of egg-laying hormone (ELH)-related precursors in the secretory cells of the atrial gland. In addition, comparative structural studies of Afurin2, together with previously reported localization studies, argue for the occurrence of a furin-like convertase within secretory granules.

Processing of mouse proglucagon by recombinant prohormone convertase 1 and immunopurified prohormone convertase 2 in vitro

The mouse tumor cell line alpha TC1-6 was used as a model system to examine the post-translational processing of proglucagon. Determination of the mouse preproglucagon cDNA sequence and comparison with the published sequences of rat and human preproglucagons revealed nucleic acid homologies of 89.1 and 84%, respectively, and amino acid homologies of 94 and 89.4%, respectively. Immunohistochemical analyses with antibodies directed against PC2 and glucagon colocalized both the enzyme and substrate within the same secretory granules. PC1 was also immunolocalized in secretory granules. Cells were metabolically labeled with [3H]tryptophan, and extracts were analyzed by reverse-phase high pressure liquid chromatography. Radioactive peptides with retention times identical to those of synthetic peptide standards were recovered and subjected to peptide mapping to verify their identities. To determine the potential role of PC1 and PC2 in proglucagon processing, 3H-labeled proglucagon was incubated in vitro with recombinant PC1 and/or immunopurified PC2. Both enzymes cleaved proglucagon to yield the major proglucagon fragment, glicentin, and oxyntomodulin, whereas only PC1 released glucagon-like peptide-I from the major proglucagon fragment. Neither PC1 nor PC2 processed glucagon from proglucagon in vitro. These results suggest a potential role for PC1 and/or PC2 in cleaving several of the normal products, excluding glucagon, from the mouse proglucagon precursor.

Processing of proendothelin-1 by human furin convertase

Endothelin-1 (ET-1) is the most potent vasoactive peptide known to date. The peptide is initially synthesized as an inactive precursor (proET-1) which undergoes proteolysis at specific pairs of basic amino acids to yield bigET-1. Production of ET-1 then proceeds by cleavage of bigET-1 by the endothelin converting enzyme (ECE). Here, we demonstrate that the in vitro cleavage of proET-1 by furin, a mammalian convertase involved in precursor processing, produced bigET-1. Upon further processing, bigET-1 was converted to biologically active ET-1. Furthermore, we demonstrate that the furin inhibitor, decanoyl-Arg-Val-Lys-Arg chloromethylketone, abolished production of ET-1 in endothelial cells.

Purification and characteristics of the candidate prohormone processing proteases PC2 and PC1/3 from bovine adrenal medulla chromaffin granules

The prohormone-processing proteases PC1/3 and PC2 belong to the family of mammalian subtilisin-related proprotein convertases (PC) possessing homology to the yeast Kex2 protease. The presence of PC1/3 and PC2 in secretory vesicles of bovine adrenal medulla (chromaffin granules) implicates their role in the processing the precursors of enkephalin, neuropeptide Y, somatostatin, and other neuropeptides that are present in chromaffin granules. In this study, PC1/3 and PC2 were purified to apparent homogeneity from the soluble fraction of chromaffin granules by chromatography on concanavalin A-Sepharose, Sephacryl S-200, pepstatin A-agarose, and anti-PC1/3 or anti-PC2 immunoaffinity resins. PC1/3 and PC2 were monitored during purification by measuring proteolytic activities with 35S-enkephalin precursor and Boc-Arg-Val-Arg-Arg-methylcoumarin amide (MCA) substrates and by following PC1/3 and PC2 immunoreactivity with specific anti-PC1/3 and anti-PC2 sera generated in this study. Purified PC1/3 and PC2 on SDS-polyacrylamide gels each show a molecular mass of 66 kDa. PC2 in the soluble fraction of chromaffin granules was present at 5- and 10-fold higher enzyme protein and activity, respectively, compared with that of PC1/3. PC1/3 and PC2 cleaved paired basic and monobasic sites within peptide-MCA substrates, with Boc-Arg-Val-Arg-Arg-MCA and pGlu-Arg-Thr-Lys-Arg-MCA as the most effectively cleaved peptides tested. PC1/3 and PC2 showed pH optima of 6.5 and 7.0, respectively. Kinetic studies indicated apparent Km values for hydrolysis of Boc-Arg-Val-Arg-Arg-MCA as 66 and 40 microM, with Vmax values of 255 and 353 nmol/h/mg for PC1/3 and PC2, respectively. Specificity of the PC enzymes for dibasic sites was confirmed by potent inhibition by the active site-directed peptide inhibitors (D-Tyr)-Glu-Phe-Lys-Arg-CH2Cl and Ac-Arg-Arg-CH2Cl. Inhibition by EGTA and activation by Ca2+ indicated PC1/3 and PC2 as Ca(2+)-dependent proteases. In addition, PC enzymes were activated by dithiothreitol and inhibited by thiol-blocking reagents, p-hydroxymercuribenzoate and mercuric chloride. These results illustrate the properties of endogenous PC1/3 and PC2 as prohormone-processing enzymes.

The distinct gene expression of the pro-hormone convertases in the rat heart suggests potential substrates

The present study examined the distribution of the pro-hormone convertases PC1, PC2, furin, PACE4 and PC5 in the rat heart. Northern blot analysis of RNA extracted from cardiac tissues showed high levels of furin and PACE4 mRNA in the atria and ventricles, while PC5 mRNA was found to be expressed at high levels in the dorsal aorta. Although undetectable by Northern blot analysis, both PC1 and PC2 mRNA were detected by in situ hybridization and immunohistochemistry in discrete regions of the intracardiac para-aortic ganglia. In situ hybridization studies also showed that furin mRNA was observed in all cardiac tissues and cells, consistent with the previously reported ubiquitous expression of this gene. PACE4 mRNA was highly abundant in both the atria and ventricular cardiomyocytes, with low to undetectable levels observed in blood vessels. Finally, PC5 transcripts were expressed in the endothelial cells lining coronary vessels and the valve leaflets of the heart. The present localization studies in the heart and cardiac blood vessels suggests potential roles for each convertase in the processing of various neuropeptides, hormones and growth factors.

Recent evolution of genes encoding the prohormone-like protein SMR1 in the rat submandibular gland

The Variable Coding Sequence (VCS) multigene family of Rattus norvegicus, is composed of at least 10 members, and shows extensive evolutionary divergence in the protein-coding region. Three members of the VCSA subclass, have been characterized: one of them, the VCSA1 gene mainly expressed in the submandibular gland (SMG) encodes the prohormone-like protein, SMR1-VA1. As VCSA-related genes have not been detected in Mus musculus, the VCSA genes subclass is presumed to have recently emerged. To study the evolution of this subclass, we have looked for VCSA genes in a closely related species, Rattus rattus. By Northern analysis, we demonstrate that VCS-related mRNAs are present in the SMG, and that the level of VCSA mRNA accumulation is approximately equal in both sexes. By contrast, in R. norvegicus, males accumulate about 3,000 times more VCSA1 mRNA than females. Using total SMG mRNA, an almost full-length cDNA, homologous to the cDNA of the R. norvegicus VCSA1 gene, was cloned by reverse transcriptase polymerase chain reaction (RT-PCR). The putative corresponding SMR1-VA1 protein is 146 amino acids long and presents the features characteristic of a secreted protein, with a potential signal peptide of 22 amino acids in the amino-terminal portion. The presence of potential processing multibasic sites suggests that small peptides could be generated (particularly a hexapeptide: Arg-Gln-His-Asn-Leu-Arg), as in the case of the SMR1-VA1 protein of R. norvegicus. From Southern blot analysis there appears that species-species modifications of VCSA gene copy number have occurred; R. rattus contains a greater VCSA1 copy number than R. norvegicus (two or three and one, respectively).

Maintained PC1 and PC2 expression in the AtT-20 variant cell line 6T3 lacking regulated secretion and POMC: restored POMC expression and regulated secretion after cAMP treatment

Two variant cell lines were recently established from parent AtT-20 cells. Whereas HYA.15.10.T.2 have a reduced level of secretory granules, HYA.15.6.T.3 are completely devoid of both the regulated pathway of secretion and of dense-core secretory granules. AtT-20 cells normally express the processing enzymes PC1, PC2, furin, carboxypeptidase E, and peptidylglycine alpha-amidating monooxygenase, as well as proopiomelanocortin, chromogranin B, and 7B2. We measured the expression of these mRNAs in both variant cell lines. Although some differences in mRNA level were noted, HYA.15.10.T.2 and HYA.15.6.T.3 cell lines maintained their expression of the processing enzymes and of 7B2. Furthermore, PC1 and PC2 were shown to be functionally active in the HYA.15.6.T.3 cells. In contrast, proopiomelanocortin and chromogranin B mRNA levels were no longer detectable in HYA.15.6.T.3 cells. Interestingly, stimulation of the HYA.15.6.T.3 cells with cAMP restored proopiomelanocortin mRNA, beta-endorphin immunoreactivity, and dense-core granules. Furthermore, at the ultrastructural level, beta-lipotropin immunoreactivity was detected in granules of cAMP-induced HYA.15.6.T.3 cells. Finally, depolarization of cAMP-induced HYA.15.6.T.3 cells with 56 mM potassium chloride resulted in a marked increase in the release of beta-endorphin immunoreactivity. These observations demonstrate that cAMP restores the regulated pathway of secretion in HYA.15.6.T.3 cells, which under untreated conditions do not demonstrate regulated release. These variant cell lines are unique models to understand better the relationship of the regulated pathway and the expression of the processing enzymes.

Role of amino acid sequences flanking dibasic cleavage sites in precursor proteolytic processing. The importance of the first residue C-terminal of the cleavage site

The amino acid sequences flanking 352 dibasic moieties contained in 83 prohormones and pro-proteins listed in a database were examined. Frequency calculations on the occurrence of given residues at positions P6 to P'4 allowed us to delineate a number of features which might be in part responsible for the in vivo discrimination between cleaved and uncleaved dibasic sites. These include the following: amino acids at these positions were characterized by a large variability in composition and properties; no major contribution of a given precursor subsite to endoprotease specificity was observed; some amino acid residues appeared to occupy preferentially certain precursor subsites (for instance, Met in P6 and P3, Asp and Ala in P'1, Pro in P6, Gly in P3 and P'2 etc.) whereas some others appeared to be excluded. Most amino acid residues occupying the P'1 position in these precursor cleavage sites were tolerated. But the beta-carbon branched side chain residues (Thr, Val, Leu, Ile) and Pro, Cys, Met and Trp were either totally excluded or poorly represented, suggesting that they might be unfavourable to cleavage. The biological relevance of these observations to the efficacy of dibasic cleavage by model propeptide convertases was in vitro tested using both pro-ocytocin convertase and Kex2 protease action on a series of pro-ocytocin related synthetic substrates reproducing the Pro7-->Leu15 sequence of the precursor in which the Ala13 residue (P'1 in the LysArg-Ala motif) was replaced by various amino acid residues. A good correlation was obtained on this model system indicating that P'1 residue of precursor dibasic processing sites is an important feature and may play the role of anchoring motif to S'1 convertase subsite. We tentatively propose that the present database, and the corresponding model, may be used for further investigation of dibasic endoproteolytic processing of propeptides and pro-proteins.

Immortalized hypothalamic luteinizing hormone-releasing hormone (LHRH) neurons: a new tool for dissecting the molecular and cellular basis of LHRH physiology

1. Two LHRH neuronal cell lines were developed by targeted tumorigenesis of LHRH neurons in vivo. These cell lines (GN and GT-1 cells) represent a homogeneous population of neurons. GT-1 cells have been further subcloned to produce the GT1-1, GT1-3, and GT1-7 cell lines. While considerable information is accumulating about GT-1 cells, very little is currently known about the characteristics and responses of GN cells. 2. By both morphological and biochemical criteria, GT-1 cells are clearly neurons. All GT-1 cells immunostain for LHRH and the levels of prohormone, peptide intermediates, and LHRH in the cells and medium are relatively high. 3. GT-1 cells biosynthesize, process, and secrete LHRH. Processing of pro-LHRH appears to be very similar to that reported for LHRH neurons in vivo. At least four enzymes may be involved in processing the prohormone to LHRH. 4. LHRH neurons are unique among the neurons of the central nervous system because they arise from the olfactory placode and grow back into the preoptic-anterior hypothalamic region of the brain. Once these neurons reach this location, they send their axons to the median eminence. With respect to the immortalized neurons, GN cells were arrested during their transit to the brain. In contrast, GT-1 cells were able to migrate to the preoptic-anterior hypothalamic region but were unable correctly to target their axons to the median eminence. These problems in migration and targeting appear to be due to expression of the simian virus T-antigen. 5. While GT-1 cells are a homogeneous population of neurons, they are amenable to coculture with other types of cells. Coculture experiments currently under way should help not only to reveal some of the molecular and cellular cues that are important for neuronal migration and axonal targeting, but they should also highlight the nature of the cellular interactions which normally occur in situ. 6. GT-1 cells spontaneously secrete LHRH in a pulsatile manner. The interpulse interval for LHRH from these cells is almost identical to that reported for release of LH and LHRH in vivo. GT-1 cells are interconnected by both gap junctions and synapses. The coordination and synchronization of secretion from these cells could occur through these interconnections, by feedback from LHRH itself, and/or by several different compounds that are secreted by these cells. One such compound is nitric oxide. 7. GT-1 cells have Na+, K+, Ca2+, and Cl- channels.(ABSTRACT TRUNCATED AT 400 WORDS)

Analysis of mutations in alleles of the fur gene from an endoprotease-deficient Chinese hamster ovary cell strain

RPE.40 mutant cells differ from wild-type Chinese hamster ovary (CHO-K1) cells in their increased resistance to Pseudomonas exotoxin A and their inability to process the insulin proreceptor and certain viral envelope proproteins. Northern analysis revealed that RPE.40 cells maintained a substantially lower steady-state level of 4.0 kb fur mRNA than did CHO-K1 cells. Analysis of fur cDNAs showed that RPE.40 cells were diploid at the fur locus, and RPE.40 cells had a Cys (TGC) to Tyr (TAC) mutation in codon 196 of one allele (allele I). Approximately 25-30% of the CHO-K1 cells were also heterozygous (Tyr/Cys) at codon 196, and pre-mRNAs transcribed from the second allele (allele II) in RPE.40 cells were defectively spliced. All other pre-mRNAs were correctly spliced. Rapid turnover of defectively spliced transcripts may account for the reduced steady-state level of fur mRNA observed in RPE.40 cells. Our results provide a mechanistic basis for the endoprotease-deficient phenotype of RPE.40 cells.

Production of bioactive enkephalin from the nonendocrine cell lines COS-7, NIH3T3, Ltk-, and C2C12

Enkephalin is synthesized from proenkephalin in neuroendocrine cells. For the attempt to induce nonneuroendocrine origin cells to produce enkephalin, we used a mammalian expression vector for fusion peptides, pMEproCT beta, in which a fused peptide is designed to be cleaved by a yeast Kex2-like endoprotease furin. Met-Enkephalin was expressed in four nonneuroendocrine cell lines: COS-7, C2C12, Ltk-, and NIH3T3. The four cell lines produced a marked amount of Met-enkephalin, which appeared as a single peak on reverse-phase HPLC. Because transplantation of adrenal medullary cells to the subarachnoid space has been used to alleviate terminal cancer pain, and enkephalin appears to play a central role in relieving pain, this enkephalin expression vector may be useful for direct enkephalin expression in pericancerous tissues.

A survey of furin substrate specificity using substrate phage display

The substrate specificity of furin, a mammalian enzyme involved in the cleavage of many constitutively expressed protein precursors, was studied using substrate phage display. In this method, a multitude of substrate sequences are displayed as fusion proteins on filamentous phage particles and ones that are cleaved can be purified by affinity chromatography. The cleaved phage are propagated and submitted to additional rounds of protease selection to further enrich for good substrates. DNA sequencing of the cleaved phage is used to identify the substrate sequence. After 6 rounds of sorting a substrate phage library comprising 5 randomized amino acids (xxxxx), virtually all clones had an RxxR motif and many had Lys, Arg, or Pro before the second Arg. Nine of the selected sequences were assayed using a substrate-alkaline phosphatase fusion protein system. All were cleaved after the RxxR, and some substrates with Pro or Thr in P2 were also found to be cleaved as efficiently as RxKR or RxRR. To further elaborate surrounding determinants, we constructed 2 secondary libraries (xxRx(K/R)Rx and xxRxPRx). Although no consensus developed for the latter library, many of the sequences in the the former library had the 7-residue motif (L/P)RRF(K/R)RP, suggesting that the furin recognition sequence may extend over more than 4 residues. These studies further clarify the substrate specificity of furin and suggest the substrate phage method may be useful for identifying consensus substrate motifs in other protein processing enzymes.

Isolation and in situ localization of a cDNA encoding a Kex2-like prohormone convertase in the nematode Caenorhabditis elegans

1. A cDNA that encodes a Kex2-like prohormone convertase (PC) containing an active site similar to that of mammalian PC2 has been isolated from C. elegans. Total RNA was isolated from a mixed population of strain BA713 worms. After poly-(A)-selection and reverse transcription, degenerate/nested polymerase chain reactions (PCR) were performed using primers based on conserved regions within the active sites of the known vertebrate and invertebrate endoproteases. 2. Two distinct 300-bp PCR products that shared homologies with the active sites of known Kex2-like endoproteases were isolated. These two PCR products were used to screen a C. elegans cDNA library. 3. The complete cDNA for a Kex2-like endoprotease, designated CELPC2, was isolated and determined to be 2527 bp in length. This size was confirmed by northern analysis. The deduced amino acid sequence for the CELPC2 cDNA is very similar to the known Kex2-like endoproteases, especially at conserved regions within the active sites, but not identical to any one of them. The strongest structural homology was to vertebrate and invertebrate PC2 sequences. 4. In situ hybridization suggests that CELPC2 is synthesized primarily in cells associated with the circumpharyngeal nerve ring and the dorsorectal ganglion.

Furin has the proalbumin substrate specificity and serpin inhibitory properties of an in situ hepatic convertase

Furin, a KEX2 protease homolog with high RNA expression in the liver is an excellent candidate as a hepatic proprotein convertase. Here we show that purified recombinant furin has the same proalbumin specificity and serpin inhibitory properties as the in situ hepatic convertase. There was rapid cleavage at the -RRD- site of normal human proalbumin but there no significant cleavage of natural unprocessed variants with cleavage site sequences of -RRV-, -HRD-, -RQD-, or -CRD-. Cleavage of the latter was not increased by S-aminoethylation. Furin was specifically inhibited by alpha 1-antitrypsin Pittsburgh (358 Met-->Arg), (K1/2 = 3 microM) but not by 50 microM normal antitrypsin M or by antithrombin, however, antithrombin/heparin was a good inhibitor (K1/2 = 9 microM). The pH optimum for proalbumin cleavage was between pH 5.5 and 6.0, indicating that furin is potentially fully active within secretory vesicles, the site of proalbumin cleavage.

Production of bioactive gastrin from the non-endocrine cell lines CHO and COS-7

We made a mutated progastrin cDNA construct that contains a cleavage site (-Arg(-4)-Arg(-3)-Lys(-2)-Arg-1) specific for the Kex2-like endoprotease furin, located ahead of the bioactive gastrin. For expressing the mutated progastrin cDNA, we used two non-endocrine cell lines, CHO and COS-7. CHO cells exhibit amidating enzyme activity and levels of amidation enzyme mRNA as high as those in the pituitary-derived endocrine cell line GH3, whereas COS-7 cells have far less amidating activity and lower amounts of mRNA. Mutant progastrin-expressing CHO cells produced mostly amidated gastrin. Gel filtration showed the size of this gastrin corresponded to that of the synthetic human gastrin-17. In contrast, COS-7 cells produced glycine-extended gastrin and only a small amount of amidated gastrin. The difference in the amount of amidated gastrin products produced by the two non-endocrine cell lines is due to differing amounts of the amidation enzyme contained in each cell line.