Proteolytic action of duodenase is required to induce DNA synthesis in pulmonary artery fibroblasts A role for phosphoinositide 3-kinase Alan D. Pemberton 1 , Tatyana S. Zamolodchikova 2 , Cheryl L. Scudamore 3 , Edwin R. Chilvers 4 , Hugh R. P. Miller 1 and Trevor R. Walker 5 1 Department of Veterinary Studies, University of Edinburgh, Easter Bush Veterinary Centre, Roslin, Edinburgh, UK; 2 Shemyakin- Ovchinnikov Institute of Bioorganic Chemistry, Russian Academy of Sciences, Moscow, Russia; 3 Department of Veterinary Pathology, University of Edinburgh, Easter Bush Veterinary Centre, Roslin, Edinburgh, UK; 4 Respiratory Medicine Unit, Department of Medicine, University of Cambridge School of Clinical Medicine, Addenbrooke’s and Papworth Hospitals, Cambridge, UK; 5 Rayne Laboratory, Respiratory Medicine Unit, University of Edinburgh Medical School, Edinburgh, UK Duodenase is a 29-kDa serine endopeptidase that displays selective trypsin- and c hymotrypsin-like s ubstrate specificity. This enzyme has been localized to epitheliocytes of B runner’s glands, a nd as described h ere, to mast cells within the intestinal mucosa and lungworm-infected lung, implying an important additional role in inflammation and tissue remodelling. In primary c ultures of pulmonary artery fibroblasts, duodenase induced a concentration-dependent increase in [ 3 H]thymidine incorporation with a maximal effect observed at 30 n M . Pretreating duodenase with soy- bean trypsin inhibitor abolished DNA synthesis, confirming that proteolytic a ctivity w as an essential requirement for this response. PAR1, P AR2 and PAR4 activating p eptides were unable to induce [ 3 H]thymidine incorporation in pulmonary artery fibroblasts. Likewise, pretreatment of fibroblasts with TNFa, known to up-regulate PAR2 expression in other systems, and IL-1b, did not enhance the potential of duodenase to induce DNA synthesis. Furthermore, duo- denase increased GTPcS binding to fibroblast membranes indicating that a G-protein-coupled receptor may mediate the effects of duodenase. Duodenase-induced DNA syn- thesis and GTPcS b inding were both f ound to be inhibited by pertussis toxin, implying a role for G i/o . Selective inhi- bitors of MEK1 (PD98059) and protein kinase C (GF109203X) only partially inhibited duodenase-induced DNA synthesis, but both wortmannin (100 n M )and LY294002 (10 l M ) inhibited this r esponse completely, indicating a key role for PtdIns 3-kinase. Furthermore, duodenase induced a 2.3 ± 0.1-fold increase in PtdIns 3-kinase activity in p85 immu noprecipitates, which was sensitive t o inhibition by wortmannin. These results suggest that duodenase can i nduce pulmonary artery fibroblast DNA synthesis in a PtdIns 3-kinase-dependent manner via a G-protein-coupled receptor which is activated by a proteo- lytic m echanism. Keywords: duodenase; fibroblasts; phosphoinositide 3-kinase; protease-activated receptor. Duodenase is a serine endopeptidase, originally isolated from bovine duodenum, with a dual trypsin-like and chymotrypsin-like primary substrate specificity, i.e. cleaving the C-terminal to both basic and hydrophobic amino-acid residues [1]. The closely related enzyme, sheep mast cell proteinase-1 (sMCP-1) is 85% identical at the amino-acid level [2] a nd, due to close similarity of the primary substrate binding region, has a strikingly similar cleavage specificity [3]. Duodenase was o riginally immunolocalized to epithelial cells of Brunner’s glands within the duodenum, and was an activator of enteropeptidase [4]. Oth er studies employing esterase staining have provided evidence for the expression of an enzyme with trypsin-like properties, distinct from tryptase, in intestinal mucosal mast cells, and in the lung around bronchioles and w ithin the alveolar septa [5]. This is consistent with data in sheep showing that sMCP-1 is located to mucosal mast cells of the gastrointestinal tract, and around small bronchi and a lveolar walls in the lung [6]. One of the many identified effects o f mast cell proteinases is their ability to induce cellular p roliferation. For example, both human mast cell tryptase and sMCP-1 h ave been shown t o b e m itogenic for fibroblasts [7,8]. In patients with chronic inflammatory lung disord ers, significant accumula- tion of mast c ells occurs within the lungs and is b elieved to underlie the generation of pulmonary fibrosis, involving proliferation of mesenchymal cells to form the basis of a fibrotic scar [9]. Recruitment of mast cells and release of their proteinases may therefore play a central role in the initiation of a p roliferative response following injury or inflammation within the lung. The field of proteinase-mediated cellular activation has expanded rapidly following the discovery that a-thrombin mediates its actions through a receptor which contains a Ôtethered-ligandÕ, with activation occurring consequent to Correspondence to T. R. Walker, Rayne Laboratory, Respiratory Medicine Unit, University of Edinburgh Medical School, Teviot Place, Edinburgh EH8 9AG, UK. Fax: + 131 6504384, Tel.: + 131 6511320, E-mail: tw@srv1.med. ed.ac.uk Abbreviations: PAR, protease-activated receptor; sMCP-1, sheep mast cell proteinase-1; DMEM, Dulbecco’s modified Eagle’s medium; TNFa, tumour necrosis factor a; PtdIns 3-kinase, phosphoinositide 3-kinase. (Received 1 4 September 2001, revised 7 December 2001, accepted 19 December 20 01) Eur. J. Biochem. 269, 1171–1180 (2002) Ó FEBS 2002 proteolytic cleavage of the N-terminal exodomain. This thrombin receptor has since been termed PAR1 (protease- activated receptor-1) and is known to mediate the actions of thrombin on platelets and other cell types [10]. Subse- quently, RT-PCR and Northern analysis have identified mRNA for three additional members of this receptors family termed PAR2, PAR3 and PAR4 [11]. Interestingly, thrombin has now been demonstrated to cleave and activate PAR1, PAR3 and PAR4 whereas trypsin a nd tryptase activate PAR2 [12]. Certain other proteases, including chymotrypsin and cathepsin G, appear to ÔdisarmÕ PAR1 by cleaving the exodomain of the receptor without inducing activation and t hus preventing activation by thrombin [13]. All four receptors have a classical heptahelical structure within the plasma membrane and are known to couple to both G q/11 and G i/o and stimulate phosphoinositide turn- over although their other potential downstream signalling targets h ave not been fully established [ 12]. In this study we have investigated the ability of duodenase to induce DNA synthesis in bovine pulmonary artery fibroblasts, attempting to elucidate which PAR subtype and signalling pathways may be involved in mediating this effect. We also provide evidence for an a dditional mast cell o rigin of duodenase, which has important implications with regard to the potential in vivo role of this enzyme. MATERIALS AND METHODS Purification of duodenase from bovine jejunum The protocol used for the purification of duodenase from bovine jejunum was identical to that currently used for the isolation of sMCP-1 from ovine gastrointestinal tissue. In brief, fresh bovine jejunal tissue was finely chopped a nd then homogenized with 3 vol. of 20 m M Tris/HCl pH 7.5 (all procedures we re carried out on ic e). After centrifugation (30 0 00 g for 30 min) and repetition of the above low salt wash step, the pellet was homogenized with 3 vol. of 20 m M Tris/HCl (pH 7.5), 0.4 M NaCl, 0.1% (v/v) Brij 35. Following repeat centrifugation, the supernatant was diluted with 2 0 m M Tris/HCl (pH 7.5), 0.1% (v/v) Brij 35 to < 0.1 M NaCl, centrifuged again, and loaded onto a column containing CM–Sepharose FF (Pharmacia), equi- librated with t he buffer described above. After elution with a 0.1–0.5 M NaCl gradient, fractions containing both chymotrypsin-like and trypsin-like activity were pooled, then rechromatographed twice on a M ono-S c olumn (Pharmacia) using 0.05–0.35 M NaCl gradients in 20 m M Tris/HCl (pH 7.5), 0.1% (v/v) Brij 35, and then 20 m M sodium phosphate (pH 7.0), 0.1% (v/v) Brij 35. The final purification step involved gel fi ltration (Superdex 75, Phar- macia) in NaCl/P i (pH 7.4) containing 0.1% (v/v) Brij 35. The identity of the product was confirmed by N-terminal amino-acid sequence analysis (P. Barker, Babraham Insti- tute, Cambridge, UK), and by comparing its ability to hydrolyse specific peptide substrates (in 0.1 M Tris/HCl, pH 8.0) with duodenase. Immunohistochemical localization of duodenase in jejunum and lung Samples of fresh bovine jejunum were fixed in 10% (v/v) formalin and 4% (w/v) paraformaldehyde, and processed into paraffin blocks. Sections (4 lm thick) were stained using 0 .1% (w/v) toluidine blue (pH 0.5), followed by eosin counterstain, and duodenase detected using rabbit anti- duodenase serum (1 : 400), rabbit anti-(sMCP-1) IgG (1.2 lg ÆmL )1 ) or control rabbit serum (1 : 400) [14], using NaCl/P i (0.5 M NaCl) containing 0.5% (v/v) Tween 80 for blocking and antibody dilutions. The secondary antibody was biotinylated goat anti-(rabbit IgG) Ig (1 : 400; Vector Laboratories), followed b y treatment with avidin–horse radish peroxidase (Vectastain ABC kit, Vector Laborato- ries) and diaminobenzidine (DAB kit, Vector L aboratories). Following immunostaining, s ections were counterstained with 0.1% (w/v) Mayer’s hematoxylin (Sigma). Samples of lung parenchyma were obtained a t postmortem from a cow infected with the lungworm Dictyocaulus viviparus and fixed in 4% (v/v) paraformaldehyde in NaCl/P i . Sections (5 lm thick) w ere prepared a nd stained with toluidine blue, rabbit anti-duodenase serum and control rabbit serum, as described above. Isolation and culture of bovine pulmonary artery fibroblasts Sections of proximal bovine pulmonary artery were obtained from the local abattoir and pulmonary artery fibroblasts isolated using a primary explant procedure [ 15]. Cells were cultured in supplemented Dulbecco’s modified Eagle’s medium (DMEM) containing foetal bovine serum (10% v/v), penicillin/streptomycin (5 UÆmL )1 and 5 lgÆmL )1 , respectively) and amphotericin B ( 2.5 lgÆmL )1 ). Cells from passages 3–10 were used for all experiments. Cells were incubated i n serum-free DMEM for 48 h prior to experimentation. Assessment of [ 3 H]thymidine incorporation Pulmonary artery fi broblasts at % 80% confluence were quiesced for 48 h prior to addition of mitogens as indicated. The cells were then incubated for an additional 20 h, with [ 3 H]thymidine (0.1 lCiÆmL )1 ) added 4 h prior to harvest- ing. Cells were washed twice with ice-cold NaCl/P i , twice with trichloroacetic acid (5% w/v), twice with ethanol and finally were solubilized with NaOH (0.3 M ). [ 3 H]Thymidine incorporation was determined by liquid scintillation counting. [ 35 S]GTPcS binding to pulmonary artery fibroblast membranes Pulmonary artery fibroblasts were lysed in ice-cold buffer containing 10 m M Tris/HCl pH 7.4, 5 m M EDTA, homo- genized using a Polytron t issue homogeniser for 2 · 10 s on ice and centrifuged at 500 g for 10 min at 4 °C to remove intact cells. Supernatants containing cell membranes were centrifuged at 50 000 g for 1 0 min and pellets washed w ith the buffer described above; this washing procedure was repeated twice. The protein content of e ach pellet was determined after resuspension in 20 m M Hepes (pH 7.4) using a Pierce BCA protein assay reagent and the protein concentration adjusted to 1 mgÆmL )1 . Binding of [ 35 S]GTPcS was carried out by the addition of cell membra nes (10 lg) to binding buffer (100 lL) containing 20 m M Hepes pH 7.4, 100 m M NaCl, 3 m M MgCl 2 ,10l M 1172 A. D. Pemberton et al. (Eur. J. Biochem. 269) Ó FEBS 2002 GDP with 0.2 n M [ 35 S]GTPcS and incubating for 60 min at 4 °C. Bound radioactivity was determined by filtration of membranes onto Whatman GF-B filters using a Brandell Cell Harvester and counted by scintillation counting. Nonspecific binding was determined in the presence of 100 l M unlabelled GTPcS. Assay of immunoprecipitated PtdIns 3-kinase Bovine pulmonary artery fibroblasts were exposed to mitogens as detailed in the figure legends, and the reactions were terminated by rapid aspiration of the media followed by the addition of ice-cold lysis buffer (50 m M Hepes, pH 7.5, 150 m M NaCl, 10% v/v glycerol, 1% v/v Triton X-100, 1.5 m M MgCl 2 ,1m M EGTA, 10 lgÆmL )1 leupeptin, 10 lgÆmL )1 aprotinin, 1 m M phen- ylmethanesulfonyl fluoride, 200 l M Na 3 VO 4 ,10m M sodium pyroph osphate, 100 m M NaF). P tdIns 3-kinase was immunoprecipitated using antibodies specific to the p85a regulatory subunit of PtdIns 3-kinase complexed to Pansorbin (Calbiochem, Nottingham, UK). PtdIns 3-kinase activity in immunoprecipitates was a ssayed as described previously, u sing sonicated phosphtidylinositol/ phosphatidylserine (3 : 1, v/v, 0.2 mgÆmL )1 ) vesicles and [c- 32 P]ATP (10 lCiÆpoint )1 ) as substrates [16]. 32 P-Labelled phosphoinositide 3-phosphate was then separated a nd quantified by thin layer chromatography using a solvent system containing chloroform/methanol/ ammonia/water (20 : 15 : 3 : 5, v/v/v/v) and autoradiog- raphy; 32 P incorporation w as determined by liquid scintillation counting. Ca 2+ measurements using Fura-2 Bovine pulmonary artery fibroblasts (P4-10) w ere g rown to confluence in supplemented DMEM as d escribed above, washed with NaCl/P i , and gently harvested into a solution containing BSA (0.2% w /v), glucose ( 0.1% w/v) and CaCl 2 (1 m M )inNaCl/P i (NaCl/P i + ). Following centrifugation, the cells were washed twice in NaCl/P i + and resuspended in the same buffer at a concentration of 1.5 · 10 6 cellsÆmL )1 . The cells were then incubated for 1 h at 37 °C with an equal volume of 4 l M Fura-2 AM (Sigma) in NaCl/P i + , washed three times with NaCl/P i + and resuspended at 1–2 · 10 6 cellsÆmL )1 . The cell suspension was allowed to equilibrate to room temperature for % 30 min and 2 mL aliquots of cells then used for Ca 2+ measurements over the following 2–3 h. M easurements were made in 1 · 1 cm quartz cuvettes, equipped with a magnetic stirrer, using a PerkinElmer LS 50B fluorimeter with fast-filter accessory. This allowed measurement of emission at 510 nm for quasi-simultaneous excitation at 340 and 380 nm, for Fura2 bound and unbound to Ca 2+ , respectively. Additions of agonists (trypsin, thrombin, duodenase, chymotrypsin and bradykinin) were made in small volumes (5–20 lL). At the end of each experiment, the maximum fluorescence was obtained b y disrupting the cells by addition of 10% (v/v) Triton X-100 (40 lL), and minimum fluorescence then determined using 20 lLof 0.4 M EGTA in 3 M Tris base. Results were analysed, and conversions to intracellular Ca 2+ concentration p erformed, using FL WINLAB software (PerkinElmer). In vitro comparison of PAR2 peptide cleavage by duodenase, tryptase and trypsin The peptide Gly-Pro-Asn-Ser-Lys-Gly-Arg-Ser-Leu-Ile- Gly-Arg-Leu-Asp-Thr-Pro corresponding to residues 5–20 of rat PAR2 [PAR2(5–20)] was synthesized (G. Bloom- berg, University of Bristol, UK). The activities of bovine trypsin, human s kin tryptase (stabilized wi th heparin, a gift from Axis Ph armaceuticals, San Francisco, USA) and bovine duodenase were first standardized against the substrate CBZ-Lys-thiobenzyl ester. This was undertaken using suitably diluted enzyme (10 lL) added to a cuvette containing 170 lLof0.1 M Hepes (pH 7.5), 10 lL 5,5¢-dithiobis-(2-nitrobenzoic acid) (10 m M in dimethylsulf- oxide) and 10 lLofN-carbobenzyloxy-Lys-thiobenzyl ester (10 m M in dimethylsulfoxide ). Initial cleavage rates at 405 nm were measured over 90 s at 23 °C, and specific activities calculated, with 1 U of activity defined as the amount of enzyme required to produce an absorbance increase of 1.0 UÆmin )1 . For each enzyme, i ncubations were in 0.05 M Hepes (pH 7.5), 0.15 M NaCl, containing rat PAR2(5–20) (0.475 mgÆmL )1 ), alanyl-tryptophan (in- ternal standard, 0.05 mgÆmL )1 ) and 0.13 U of enzyme (total assay volume 200 lL). Samples (30 lL) were removed at varying time-points, and reactions terminated by the addition of 30 lL 10% acetic acid. These samples were then chilled on ice , and frozen ( )20 °C) prior to analysis. Intact PAR2(5–20) and internal standard peak heights were quantified in samples following RP-HPLC (Jupiter C5 column, Phenomenex) using a water/acetonit- rile gradient containing 0.1% trifluoroacetic acid. The ratio of intact PAR2(5–20) to internal standard peak heights was plotted against time. Fractions collected from some runs were subjected to mass spectrometry (I. Davidson, University of Aberdeen, Scotland, UK). Materials Anti-duodenase serum and affinity-purified anti-(sMCP-1) IgG were prepared as described previously [4,8]. Anti-(p85 PtdIns 3-kinase) Ig was obtained from TCS Biologicals (Botolph Claydon, UK) and [c- 32 P]ATP from Amersham (Amersham). PAR activating peptides, Ser-Phe-Leu-Leu- Arg-Asn for PAR1 and Gly-Tyr-Pro-Gly-Lys-Phe for PAR4 were obtained from Bachem Ltd (Saffron Walden, Essex, UK) and Ser-Leu-Ile-Gly-Arg-Leu and Ser-Leu-Ile- Gly-Arg-Leu-NH 2 for PAR2 were supplied by G. Bloom- berg (University of Bristol, UK). All other chemicals were of the highest commercial quality. RESULTS Identification of duodenase The N-terminal amino-acid sequence of the product isolated from jejunum (Ile-Ile-Gly-Gly-His-Gl u-Ala-Lys-Pro- His-Ser-Arg-Pro-Tyr-Met-Ala-Phe-Leu-Leu-Phe) was iden- tical t o t hat originally described for duodenase [1]. The first of two p eptide substrates analysed, bee ve nom melittin, was cleaved preferentially at Lys7, with secondary cleavage at Lys23, as previously described for duodenase [17]. Porcine angiotensinogen (1–14) was rapidly and s pecifically cleaved Ó FEBS 2002 Duodenase induces DNA synthesis via PtdIns 3-kinase (Eur. J. Biochem. 269) 1173 Fig. 1. Histochemical detection o f j ejunal an d lung mast c ells and immunoperoxidase l ocalization o f d uodenase i n bov ine intestine and l ung. Repre- sentative positively s tained mast cells ar e indicated by large arrows. Panels ( a–e) show localization o f duodenase in bovine jejenum. In panel (a) ma st cells surrounding crypts in the jejunal mucosa are toluidine blu e (pH 0.5)-p ositive (counterstained with eosin). Anti- duodenase Ig stainin g is shown at low magnificatio n in panel (b), with ab undant stainin g of c ells with morphology and distribution similar to th at shown for to luidin e blu e p anel ( a). Hi gher m agnification in panels (c–e), show s bovine j ejenum stain ed with control rabbit seru m, rabb it an ti- duodenase Ig and rabbit anti-(sMCP-1) Ig, r espectively. A s imilar p attern o f s taining i s s een w ith a nti-duo denase Ig and anti-(sMCP-1) Ig, and no staining is observed in the control. Sections of a bronchiole from bovine l ung infected with the lungworm Dic tyocaulus vi viparus are shown in panels (f–h). I n panel (f), duodenase-positive cells are abund ant in the granulomatous reaction around the bronchiole. Panel (g) shows a n adjacent section incubated with control serum. Panel (h) sh ows toluidine blue and eosin staining of a s ection adjacent to (f). Note the s imilar distribution of mast cells in (f) a nd (h) and the presence of numerous eosinophils (small arrows) in the parasitized lung. In association with the accumulation of mas t cells there i s increased fib rosis (*) an d smooth m uscle h ypertrophy (arrowhead). All of the tissues w ere fixed in 4% ( v/v) paraformaldehyde. 1174 A. D. Pemberton et al. (Eur. J. Biochem. 269) Ó FEBS 2002 at Phe8, a s has b een shown for duodenase [4]. Therefore, the jejunal enzyme we purified was identified as duodenase, or a highly similar variant of the enzyme. Immunolocalization of duodenase Toluidine blue staining identified abundant spindle or stellate-shaped m ast cells in bovine jejunum samples. The se cells were l ocate d principally in the lamina propria (Fig. 1a) and submucosa (not shown). Immunostaining o f paraform- aldehyde-fixed sections with rabbit anti-duodenase serum and affin ity-purified rabbit anti-(sMCP-1) IgG detected cells only in the lamina propria. These strongly staining cells showed a similar distribution and morphology to those seen with toluidine blue within the lamina propria (compare Fig. 1a with Fig. 1b,d,e). The distribution of positive cells after labelling with anti-duodenase Ig or anti-(sMCP-1) IgG was very similar, and in neither instance was there any labelling of submucosal tissues. Occasional intraepithelial cells were weakly labelled (Fig. 1d), and the identity of these toluidine blue negative cells was not confirmed. Tissues fixed in neutral buffered formalin showed negligible mast cell staining by comparison, and control rabbit serum was negative regardless of the fixation procedure (Fig. 1c). Lungworm-infected lung parenchyma showed the presence of large numbers of eosinophils and toluidine blue-positive mast c ells. An example of their distribution around a bronchiole is shown in Fig. 1h, in which fibrosis and smooth muscle hyperplasia was also evident. Numerous cells were also lab elled with duodenase antiserum around bronchioles (Fig. 1f) and within the alveolar septa (not shown). Their size and distribution as observed in adjacent sections was similar to that of toluidine blue-positive cells (compare Fig. 1f,h). Control r abbit serum gave no labelling (Fig. 1g). Duodenase induces DNA synthesis in pulmonary artery fibroblasts The effect of duodenase on DNA synthesis w as assessed using [ 3 H]thymidine incorporation in bovine primary pul- monary artery fibroblasts. Treatment of cells for 24 h with duodenase induced a concentration-dependent increase in [ 3 H]thymidine incorporation w hich was m aximal at 3 0 n M , achieving a 5.5 ± 0.8-fold increase above control values (Fig. 2 A). Pretreatment o f duodenase with soybean t rypsin inhibitor ( 3 mg ÆmL )1 , 1 5 min), an effective inhibitor of t his enzyme [1] was found to inhibit completely the ability of this enzyme to induce [ 3 H]thymidine incorporation i n pulmo- nary artery fibroblasts (Fig. 2B), confirming that the proteolytic activity of duodenase is essential for induction of DNA synthesis. Importantly, treatment of cells with duodenase (30 n M ) for 10 min followed by the addition of soybean t rypsin inhibitor (3 mgÆmL )1 , 1 5 min) induced [ 3 H]thymidine incorporation to a similar extent as addition of duodenase alone (Fig. 2B), suggesting a rapid signalling mechanism. Furthermore, conditioned media generated by this method was used to assess whether duodenase could cleave and release a cell surface molecule that could interact Fig. 2. Duodenase induces DNA synthesis in pulmonary artery fibro- blasts. (A) quiescent cells were treated with duodenase (3–100 n M )as indicated f or 20 h prior to ad dition of [ 3 H]thymidine (0.1 lCiÆwell )1 ): incorporation was assessed after 4 h as detailed in Materials and methods. (B) [ 3 H]Thymidine incorporation tested in c ells treated with duodenase ( duod, 3 0 n M )whichhadbeenpretreatedwithorwithout soybean trypsin inhibitor (+ STI, 0.2 mgÆmL )1 ) for 15 min. To examine a role for duo denase-induced release of a mitogenic f actor and generation of con ditioned media, duodenase was a dded to cells for 10 min prior to addition of soybean trypsin inhibitor for 15 min, media removed and r eplaced w ith f resh quiescent media (duod + STI removed). This conditioned media was transferred to untreated cells (cond. media) and [ 3 H]thymidine incorporation assessed as before. (C) [ 3 H]Thymidine incorporation tested in cells treated with duodenase (30 n M ), PAR1 activating peptide (Ser-Phe-Leu-Leu-Arg-Asn, 100 l M ), PAR2 activating peptide ( a, Ser-Leu-Ile-Gly-Arg-Leu; b, Ser- Leu-Ile-Gly-Arg-Leu-NH 2 ;both100l M ) or P AR4 activating peptide (Gly-Tyr-Pro-Gly-Lys-Phe, 100 l M ). [ 3 H]Thymidine incorporation was assessed as detailed in Materials and methods. R esults are expressed as m ean ± SEM-fold increase ov er control cells fro m four separate exper iments, each performed i n triplicate. Ó FEBS 2002 Duodenase induces DNA synthesis via PtdIns 3-kinase (Eur. J. Biochem. 269) 1175 with cell surface receptors or induce secretion o f a bioac tive molecule to induce DNA synthesis. In these experiments, addition of conditioned media to pulmonary artery fibro- blasts had no significant effect on [ 3 H]thymidine in corpor- ation above control levels (Fig. 2B). Hence duodenase, purified from bovine jejenum is mitogenic for bovine pulmonary artery fibroblasts and this effect is dependent on the direct proteolytic activity of th is enzyme. As the PARs d escribed to date are activated by cleavage of trypsin-like primary specificity, and as duodenase, (like sMCP-1, which is also mitogenic in this system [8]), has a trypsin-like component, a ctivating peptides selective for PAR1, PAR2 and PAR4 were used to investigate whether the mitogenic effect of duodenase was mediated v ia a known PAR mechanism. Surprisingly, all PAR peptides were unable to induce [ 3 H]thymidine incorporation in pulmonary artery fibroblasts (Fig. 2C). It should be noted that two forms of the PAR2 activating peptide were assessed, the f ree form a nd the a mido form, neither of which showed ability to induce DNA synthesis ( Fig. 2C). Lack of activation by these peptides is unlikely to be a c onsequence of species differences in receptor sequences as Ser-Leu-Ile- Gly-Arg-Leu (PAR2 activating peptide, mouse-derived sequence, 100 l M ) was reported to mobilize Ca 2+ in bovine coronary artery smooth muscle cells [18]. The PAR1 activating peptide Ser-Phe-Leu-Leu-Arg-Asn (human- derived sequence, 100 l M ) activated phospholipase C in bovine tracheal smooth muscle cells (T. R. W alker & E. R. Chilvers, unpublished observations). Furthermore, this PA R1 activating p eptide (100 l M ) w as found to induce aggregation o f isolated bovine platelets s imilar to that induced by thrombin (T. R. Walker, unpublished observa- tions). Degradation of PAR2 model peptide To further assess t he po tential interaction between duoden- ase a nd PAR2, the ability of this enzyme to cleave a PAR2 substrate was investigated. Under the experimental condi- tions used, the known P AR2 activators bovine t rypsin and human mast cell tryptase rapidly cleaved the model PAR2 substrate PAR2(5–20) (t ½ ¼ 3.5 and 3.4 min, respective - ly). One cleavage product was resolved by HPLC and identified by mass spectrometry a s Ser-Leu-Ile-Gly-Arg- Leu-Asp-Thr-Pro (m/z ¼ 971) (the other product Gly- Pro-Asn-Ser-Lys-Gly-Arg was not resolved under the chromatographic conditions used). This confirms the capacity of t rypsin and t ryptase to cleave at the appropriate activation site. However, PAR2(5–20) was cleaved much more slowly by duodenase (t ½ % 1200 min). Moreover, the cleavage mixture exhibited HPLC peaks corresponding both to the activation product (Ser-Leu-Ile-Gly-Arg-Leu- Asp-Thr-Pro) a nd to other unidentified products, suggesting multiple sites of cleavage of this substrate. Together, these results suppo rt the hypothesis that duodenase acts independently of the known trypsin/ tryptase-sensitive PAR2 receptor. Duodenase induces GTPcS binding in pulmonary artery fibroblast membranes To establish the mechanism of action of duodenase, [ 35 S]GTPcS binding to fibroblast membranes was used as an index of G protein activation. Duodenase (30 n M ) induced a 57.0 ± 2.3% increase in guanine nucleotide binding to pulmonary artery fibroblast cell membranes compared to controls, suggesting that the effects of duodenase are indeed mediated through a G-protein- coupled rec eptor. Pre-treatment of cells with pertussis toxin (100 ngÆmL )1 , 18 h) prior to cell fractionation and membrane isolation inhibited [ 35 S]GTPcS binding by 80.8 ± 10.3%, suggesting that the predominant G-protein mediating this signal is a member o f the G i/o family (Fig. 3). Intracellular signalling pathway underlying duodenase-stimulated fibroblast proliferation In order to identify a role for a downstream signalling pathway that may mediate the effect of duodenase on pulmonary artery fibroblasts, we examined a number of diverse signalling pathways that have been implicated in agonist-stimulated DNA synthesis in other cell systems. Pulmonary artery fibroblasts preloaded with the Ca 2+ -binding dye fura-2 were stimulated with duodenase and fluorescence analy sed as an index of Ca 2+ mobilization. As demonstrated in Fig. 4, duodenase at concentrations up to 90 n M was unable to induce Ca 2+ mobilization. In addition, thrombin, trypsin and chymotrypsin were also unable to induce C a 2+ mobilization. However, addition of bradykinin (5 l M ) to these cells induced a rapid Ca 2+ transient indicating that these cells were responsive t o activation through other G-protein-coupled receptors (Fig. 4). As anticipated, this response to bradykinin could be desensitized by prior exposure to the agonist (Fig. 4). These results suggest that this group of proteases do not appear to cause acute Ca 2+ mobilization or influx in these cells. Of note, addition of a PAR2-activating peptide or addition of thrombin, which will act t hrough PAR1, PAR3 and P AR4, all h ad no effect on Ca 2+ mobilization (Fig. 4). These results demonstrate that Ca 2+ mobilization is unlikely to be involved in mediating cell growth in pulmonary artery fibroblasts. Wortmannin (100 n M ) and LY294002 (10 l M ), two structurally distinct and selective inhibitors of PtdIns 3-kinase, completely blocked duodenase-induced [ 3 H]thymidine incorporation, suggesting a key role for PtdIns 3-kinase in this response (Fig. 5). In contrast, PD98059, a MEK1 inhibitor, caused only a partial Fig. 3. Effect of duodenase on [ 35 S]GTPcS b inding. Pul monary artery fibroblasts were untreated (open bars) or pretreated with pertussis toxin (PTX, 100 ngÆmL )1 , 18 h , hatch ed bars) p rior to ce ll lysis and membrane isolation. [ 35 S]GTPcS bin ding was carried out a s detailed i n the Method s s ection, results are expressed as mean fold increase a bove control ± SEM from three ex periments performed in triplicate. 1176 A. D. Pemberton et al. (Eur. J. Biochem. 269) Ó FEBS 2002 inhibition of DNA synthesis, r educing the response t o duodenase by 54% ± 13% (Fig. 5), implying that acti- vation of MEK1 and its downstream e ffectors may have a modulatory role in duodenase-stimulated responses. Pre- incubation of cells with the protein kinase C inhibitor GF109203X, at a concentration previously shown to f ully inhibit protein kinase C activity [19] again resulted in only a modest reduction in duodenase-stimulated [ 3 H]thymi- dine incorporation (29% ± 1 0% inhibition from stimu- lated control values), indicating that, although required for a full mitogenic response, protein kinase C activation does not ap pear to be critical for the initiation of this response. Pretreatment of pulmonary artery fibroblasts for 18 h w ith pertussis toxin, which ADP-ribosylates the a subunit of G i and G o resulting in blockade of G protein activation, inh ibited duodenase-induced [ 3 H]thymidine incorporation by 52 ± 2.5%, suggesting involvement of G i /G o in mediating this component of cell growth (Fig. 5). In subsequent experiments, duodenase (30 n M ) was found to activate p85a-associated PtdIns 3-kinase in pulmonary artery fibroblasts by 2.28 ± 0.14- fold above control values, and pretreatment of these cells with wortmannin (100 n M , 20 min) inhibited this activity to below basal levels (Fig. 6). In combination with the major inhibitory effects of wortmannin and LY294002 on duodenase-stimulated [ 3 H]thymidine incorporation, these results indicate a key role for a G-protein-coupled receptor/PtdIns 3-kinase p athway in mediating duoden- ase-stimulated DNA synthesis. Effect of inflammatory cytokines on duodenase-induced DNA synthesis Pretreatment of pulmonary artery fibroblasts with the cytokines IL-1b and TNFa was undertaken to elucidate whether th e effect of duodenase on DNA synthesis in our model system could be augmented by factors which are released at a site of inflammation. Furthermore, TNFa has been reported to increase PAR2 expression and hence would allow further insight into a potential role of PAR2 in mediating the effects of duodenase [20]. Exposure of pulmonary artery fibroblasts to duodenase resulted in a Fig. 5. Effect of signalling inhibitors on duodenase-induced DNA syn- thesis. Pulmonary artery fibroblasts were pretreated with wortmannin (100 n M , 20 m in) or LY294002 (10 l M , 20 m in), PD98059 (10 l M , 30 min), GF109203X (1 l M , 5 min ) or pertussis toxin ( PTX, 100 ngÆmL )1 , 18 h) prior to addition of duodenase (30 n M ). [ 3 H]Thymidine incorporation was assessed as indicated in Methods, results are expressed as percentage mean ± S EM relative to untreated cells st imulated with duoden ase. Results are from four independent experiments e ach performed in triplicate. Fig. 6. Duodenase a ctivates PtdIn s 3-kinase i n pulmonary artery fibro- blasts. Pulmonary artery fibroblasts were incubated in the presence (hatched bars) or absence (open bars) of wortmannin (100 n M )for 20 min prior to additio n of du odenase (30 n M , duod). Reactions were terminated and PtdIns 3-kinase activity was assayed in p85a immunoprecipitates as detailed in Materials and methods. Results are expressed as mean c .p.m. ± SEM from a single experiment performed in quadruplicate, representative of two others with s imilar results. Fig. 4. Effect of duodenase on Ca 2+ mobilization. Pulmonary artery fibroblasts preloaded with Fura-2 were stimulated with agonists as indicated. Intracellular Ca 2+ was analysed and plotted over time as indicated. Traces are representative of three separate experiments which a ll gave ve ry similar results. Ó FEBS 2002 Duodenase induces DNA synthesis via PtdIns 3-kinase (Eur. J. Biochem. 269) 1177 5.23 ± 0.47-fold increase in [ 3 H]thymidine incorporation above control levels (Table 1). While pretreatment of cells with IL-1b (10 ngÆmL )1 ) alone for 24 h induced an increase in DNA synthesis by 4 9 ± 12% (n ¼ 8, p < 0.05), it also resu lted in a significant inhibition of duodenase-induced [ 3 H]thymidine incorporation relative to IL-1b-treated control cells (Table 1). In contrast, pretreatment with TNFa (10 ngÆmL )1 ) alone reduced the level of [ 3 H]thymidine incorporation by 69 ± 3% (n ¼ 12, p < 0.05) (Table 1) but had no significant effect on the relative magnitude of DNA synthesis induced by duodenase: 6.19 ± 1.03-fold increase above control values, respectively (Table 1). DISCUSSION Mast cells present within the intestinal mucosa of rodents express subset-specific chymases which are thought to act as part of the innate immune response against intestinal nematodes by increasing epithelial permeability [21]. Similar mucosal-specific mast cell subsets also e xist in the sheep and goat intestine [21,22] and are typified by the expression of sMCP-1 and goat mast cell proteinase-1, respectively. Expulsion o f intestinal nematodes i n the sheep is associated with simultaneous release of sMCP-1 into the gut lumen and circulation [23]. The immunolocalization of duodenase to bovine intesti- nal mucosal mast cells described h ere would suggest that it too belongs to the ruminant mucosal mast cell proteinase family, w hich are notable for their dual c hymase and tryptase-like activities. It was possible to isolate duodenase from bovine jejunum using methodology identical to that employed for the purification of sMCP-1 from gastrointes- tinal tissues. However, duodenase has previously been localized only to the epithelial cells of Brunner’s glands located in the duodenal wall [4]. This suggests either that duodenase is present in both cell types, or that each site produces distinct enzymes that are nonetheless highly similar structurally, functionally and immunologically. Lungworm infection in sheep is known to involve a pronounced mastocytosis [24], and sMCP-1 is upregulated in mast c ells recruited t o s ites of allergic lung inflammation [25]. The current observation of abundant duodenase- positive mast cells in lungworm-infec ted bovine lung shows the potential for local duodenase release by mast cells recruited to i nflammatory sites i n the bovine lung and i s consistent with a putative role in tissue modelling. In this study, we have shown t hat the similarity between duodenase and sMCP-1 e xtends to the stimulation of pulmonary artery fibroblasts, with both enzymes able to induce DNA synthesis over a similar concentration range. As soybean t rypsin inhibitor was able to completely inhibit the duodenase effect, this demonstrates that the catalytic activity is essential for its action. However, only a short exposure to duodenase is required to induce maximal DNA synthesis suggesting a rapid activation p rofile. Conditioned media from duodenase and soybean trypsin inhibitor- treated cells had no mitogenic effect, implying that duodenase acts directly on fibroblasts and does not release a mitogenic mediator from the cell or medium to act in an autocrine or paracrine manner. Furthermore, duodenase induces [ 3 H]thymidine incorporation preferentially in sub- confluent cell cultures suggesting t hat close cell–cell contac t and intercellular activation is not a requirement for DNA synthesis. It is now recognized that the mitogenic effect of other proteases such as thrombin and trypsin are mediated by protease-activated receptors [12]. These are a family of seven-transmembrane o r heptahelical receptors, which cou- ple to heterotrimeric G-proteins to transduce their signal, and are activated by cleavage of an extracellular portion of the receptor close to the N-terminus, thus exposing a new N-terminus that interacts with, and activates the receptor. Receptors identified to date are PAR1, PAR2, PAR3 and PAR4; each has a similar mechanism of action has a d istinct sequence at its cleavage site. As a consequence, synthetic peptides have been developed t hat mimic the newly exposed N-terminus and a ct as specific activators [12]. Howe ver, no selective ligand for PAR3 exists implying that this rece ptor requires other structural interactions to achieve activation [26]. Indeed, r ecent data suggest that PAR3 does not mediate signal transduction directly but instead acts as a cofactor for the clea vage and activation of PAR4 [27]. Thrombin has been shown t o cleave and a ctivate PAR1, PAR3 and PAR4, whereas trypsin cleaves and activates PAR2. As duodenase is capable of cleaving certain substrates with trypsin-like primary specificity, we initially hypothesized that induction of DNA synthesis by duoden- ase is mediated through a PAR2 mechanism. Surprisingly, we could find no evidence to support the involvement of a classic PAR2 in mediating the mitogenic effects of duodenase, specifically: (a) the synthetic peptide Ser-Leu-Ile-Gly-Arg-Leu, which is specific for PAR2, was unable to i nduce [ 3 H]thymidine incorporation in fibroblasts, and a similar lack of mimickery was evident for peptides specific for PAR1 and PAR4; and (b) duodenase cleave d t he model PAR2 substrate more slowly than either trypsin or tryptase, and generated a very different array of peptides, suggesting that duodenase may cleave PAR2 at different sites. Activation of PAR3 by duodenase seems unlikely, as this receptor has limited intrinsic signalling capacity [27] and so far has only b een found to be activated by thrombin [12]. Schechter et al. [28] have described the action of mast cell tryptase on keratinocytes, as acting through a subpopula- tion of PAR2 receptors, suggesting the existence o f subtypes Table 1. Effect of cytokines on duodenase-induced DNA synthesis. Bovine p ulmonary artery fibroblasts were assessed for [ 3 H]thymidine incorporation induced by d u odenase (30 n M ), following pretreatment for 24 h with TNFa (10 n gÆmL )1 )orIL-1b(10 ngÆmL )1 )asindicated. The values qu oted represent the ratio of [ 3 H]thymidine incorporation to the mean [ 3 H]thymidine incorporation for the corresponding un - treated control w ells in the s ame experiment. Results a re expressed as mean ± SEM. Results in parentheses are corrected for the effects of cytokines o n baseline c ell growth, and are expressed as the ratio o f [ 3 H]thymidine incorporation in proteinase-treated wells to that in control wells for each cytokine treatment. Untreated a + TNF-a a + IL-1b b Control 1.00 ± 0.03 0.31 ± 0.03 1.49 ± 0.12 Duodenase 5.23 ± 0.47 1.90 ± 0.32 (6.19 ± 1.03) 3.96 ± 0.57 (2.65 ± 0.38)* a n ¼ 12, over three separate experiments. b n ¼ 8, over two sepa- rate experiments. * p > 0.001. 1178 A. D. Pemberton et al. (Eur. J. Biochem. 269) Ó FEBS 2002 of this receptor. In addition, it has been demonstrated that regulation of intestinal ion transport in rat jejenum is mediated by a P AR that, a lthough similar in many r espects to PAR2, showed distinct and atypical orders of potency when a range of peptide agonists were assessed [29]. These reports and the data from this study, in particular the pertussis toxin-sensitivity of DNA synthesis induction and the ability o f duodenase to stimulate [ 35 S]GTPcS binding to pulmonary artery fibroblast membranes, would suggest that the m itogenic action of duodenase is mediated via direct interaction with a proteolytically activated G i/o -coupled receptor. While the precise PAR subtype remains to be fully identified, it may be an atypical P AR2 that is not activated by existing classic PAR2 peptides. To date, no bo vine PAR sequences have been published and analysis o f cleavage sites on these receptors may reveal species-specific activation motifs that are distinct from those in mouse, rat and humans an d explain the lack of efficacy of current PAR2- activating peptides in our model system. A number of s ignalling pathways and intermediates such as Ca 2+ mobilization, the E RK pathway, PtdIns 3-kinase and protein kinase C h ave all been identified a s mediators of proliferative signals in a variety of cell t ypes. I n pulmonary artery fibroblasts, duodenase, trypsin, chymotrypsin, thrombin and PAR2 peptides were unable to mobilize Ca 2+ from intracellular stores. As duodenase induces DNA synthesis for these cells, these data would app ear to dissociate mobilization o f intrac ellular Ca 2+ from induction of DNA synthesis, a situation very similar to that p reviously demonstrated in bovine a irway s mooth muscle [19,30]. These results also parallel those described for the effects of human tryptase on fibroblasts, where tryptase is mitogenic for these cells but does not act via PAR2 or Ca 2+ - dependent pathways [28]. Employing a range of selective inhibitors, we investigated the r oles of PtdIns 3-kinase, MEK1/ERK, p rotein kinase C a nd pertussis t oxin-sensitive G-proteins. Only partial inhibition of DNA synthesis was achieved with maximally effective concentrations of PD98059 (MEK1 inhibitor) and G F109203X (protein kinase C inhibitor) indicating that each of these pathways has a modulatory r ather than a mandatory role to play in mediating the proliferative response. In contrast, a recent report has shown that tryptase induces DNA synthesis in canine tracheal smooth muscle through an ERK1/2-depen- dent mechanism, proliferation being inhibited completely by PD98059 [31]. Moreover, in pulmonary artery fibroblasts, inhibition of PtdIns 3-kinase by wortmannin or LY294002 inhibited completely duodenase-induced [ 3 H]thymidine in- corporation. This would suggest that activation of P tdIns 3- kinase is the key regulatory step in the proliferative p athway and that each of the other pathways interacts with this pathway with the magnitude of the cellular response determined by the integrated sum of each of these components. Our data is supported by previous reports demonstrating t hat thrombin a cts i n a PtdIns 3-kinase- a nd p70 s6k -dependent manner to induce DNA synthesis in pulmonary artery fibroblasts [32]. In addition, this report noted that downregulation of protein kinase C partially attenuated thrombin-induced p70 s6k activation, which would concur with our findings that inhibition of protein kinase C results in partial inhibition of DNA synthesis. To date, identification of downstream signalling path- ways for PARs have principally concentrated on PAR1 and PAR2. P AR1 c ouples to members of the G 12/13 ,G q and G i families, interacting with various signalling pathways including phospholipase Cb, adenylyl cyclase, PtdIns 3-kinase and nonreceptor tyrosine kinases such as Src [33]. PAR2 activation is associated with MAPK activation, phospholipase C activation and Ca 2+ mobilization. How- ever, trypsin-induced MAPK activation was reported to occur independently of PAR2 in bovine pulmonary artery fibroblasts [34]. Inflammatory cytokines have previously been shown to induce selective upregulation of P AR2 receptors without affecting the thrombin rece ptor in human umbilical vein endothelial cells [20]. To determine whether the prolifera- tive response of pulmonary artery fi broblasts could be modulated under inflammatory conditions, cells were treated for 24 h with TNFa; this resulted in a reduction in the level of [ 3 H]thymidine incorporation under control conditions, but had no significant influence on the relative magnitude of this response to duoden ase. In contrast, pretreatment of cells with IL-1b resulted in significant inhibition of duodenase-induced DNA synthesis and an enhanced level of baseline [ 3 H]thymidine incorporation under control conditions. These results suggest that these cytokines cause the fibroblasts either to become refractory to mitogens or to enter into S-phase more slowly over the time period examined. It remains to be established whether chronic exposure to TNFa and IL-1b would result in a sensitization of these cells to mitogenic stimuli. These results support further our hypothesis that duodenase is not acting via a classical PAR2. In summary, this study has demonstrated that duoden- ase induces DNA synthesis in pulmonary artery fibro- blasts and that this response may be mediated by an atypical PAR, either an i soform of PAR2 or an uniden- tified receptor. It is important to recognize that the current study was undertaken in a fully homologous system, using a bovine serine protease a nd bovine pulmonary fibroblasts. This would indicate that the proteolytic event and subsequent downstream signalling and functional responses we have described m ay be an important consequence o f duodenase release from Brun- ner’s glands, or of mast cell activation in vivo. Indeed, mast cell hyperplasia is known to be a prominent event in many forms of chronic inflammation in the lung such as cryptogenic fibrosing alveolitis, and fibroblast proliferation is the most significant feature in the pathology of these clinical conditions [9]. The precise nature and character- ization of the receptor that mediates the effects of duodenase requires further investigation. ACKNOWLEDGEMENTS This work was funded by the Norman Salvesen Emphysema Research Trust, the Wellcome Trust, and the National Asthma Campaign (UK). We thank Dr Joh n Huntley and Ms Anne Mackellar for providing the bovine lung se ctions and Dr Jeremy B rown for h elp in p reparing Fig. 1. REFERENCES 1. Zamolodchikova, T.S., Vorotyntseva, T.L. & Antonov, V.K. (1995) Duodenase, a new serine protease of unusual specificity from bovine duodenal mucosa. Purification and properties. Eur. J. Biochem. 227, 866– 872. Ó FEBS 2002 Duodenase induces DNA synthesis via PtdIns 3-kinase (Eur. J. Biochem. 269) 1179 2. McAleese, S .M., Pemberton, A.D., McGrath, M.E., Huntley, J.F. & Miller, H.R.P. (1998) Sheep mast-cell proteinases-1 and -3: cDNA cloning, prim ary structure and m olecular modellin g of the enzymes and further s tudies on su bstrate sp ecificity. Biochem. J. 333, 801–809. 3. Pemberton, A.D., Huntley, J.F. & Miller, H.R.P. (1997) Sheep mast cell proteinase-1: characterization a s a m ember of a new class of dual-specific ruminant chymases. Biochem. J. 321, 665–670. 4. Zamolodchikova, T.S., Sokolova, E.A., Alexandrov, S.L., Mikhaleva, I.I., Prudchenko, I.A., Morozov, I.A., Kononenko, N.V., Da Mirgorodskaya, O.A.U., Larionova, N.I., Pozdnev, V.F. et al. (1997) Subcellular localization, substrate s pecificity and crystallization o f duodenase, a potential a ctivator of enteropepti- dase. Eur. J. Biochem. 249, 612– 621. 5. Jolly, S ., Coignoul, F., Gabriel, A. & Desmecht, D. (1999) Detection of t ryptase i n b ovine mast c ells: c omparison o f e nzyme- and immuno-histochemistry. J. Comp. Path. 120 , 269–279. 6. Pemberton, A.D., McAleese, S.M., Huntley, J.F., Collie, D.D.S., Scudamore, C.L., McEuen, A.R., Walls, A.F. & Miller, H.R.P. (2000) cDNA sequence of two s heep mast cell tryptases a nd the differential expression of tryptase and s heep mast cell proteinase-1 in lung, dermis and gastrointestinal tract. Clin. E xp Allergy. 30, 818–832. 7. Ruoss, S.J., Hartmann, T. & Caughey, G.H. (1991) Mast cell tryptase is a mitogen for cultured fi broblasts. J. Clin. Invest. 88, 493–499. 8. Pemberton, A.D., Belham, C.M., Huntley, J.F., Plevin, R. & Miller, H.R.P. ( 1997) Sheep mast cell proteinase-1, a serine pro- teinase with both tryptase- and chymase-like p roperties, is inhib- ited by plasma proteinase inhibitors an d i s m itogen ic fo r b ovine pulmonary artery fib roblasts. Biochem. J. 323, 7 19–725. 9. Pesci., A., Bertorelli, G., Gabrielli, M. & Olivieri, D. (1993) Mast cells i n fibrotic lung disorders. Chest 10 3, 989–996. 10. Vu, T K.H., Hung, D.T., Wheaton, V.I. & Coughlin, S .R. ( 1991) Molecular cloning of a functional thrombin receptor reveals a novel proteolytic mechanism of receptor activation. Cell 64, 1057–1068. 11. Hollenberg, M.D. (1999) Protease-activated receptors: PAR4 and counting: how long is the course? Trends Pharmacol. Sci. 20, 271–273. 12. Coughlin, S.R. (1999) How t he protease thrombin tal ks t o cells. Proc. N atl Acad. Sci. USA 96 , 11023–11027. 13. Parry, M.A., Myles, T., Tschopp, J . & Stone , S.R. (1996) Cleavage of th e thrombin receptor: identification of potential activators and inactivators. Bio c hem. J. 320, 335– 341. 14. Sture, G.H., Huntley, J.F., Mackellar, A. & Miller, H.R.P. (1995) Ovine mast cell heterogeneity is defined by the distribution of sheep mast cell proteinase . Vet Immunol. Immunopathol. 48 , 275–285. 15. Freshney, R.I. (1983) Cul ture of A nimal Cells. A Manual of B asic Techniques, pp. 99–118. Alan R . Liss I nc., New York. 16. Carter, A.N. & Downes, C.P. (1993) Signaling by neurotrophic factors: activation o f pho spho inositide 3 -kinase b y n erve growth factor. Neuroprotocols 3 , 107–118. 17. Mirgorodskaya, O., Kazanina, G., Mirgorodskaya, E., Vorotyntseva, T., Zamolodchikova, T. & Alexandrov, S. (1996) A comp arative study of the specificity of m elittin hydrolysis by duodenase, trypsin and plasmin. Protein Peptide Lett. 3 , 315–320. 18. Bretschneider, E., Kaufmann, R., Braun, M., Wittpoth, M., Glusa,E.,Nowak,G.&Schror,K.(1999)Evidenceforprotein- ase-activated receptor-2 (PAR-2)-mediated mitogenesis in coro- nary artery s mooth muscle cells. Br. J. Pharmacol. 126, 1735–1740. 19. Walker, T.R., Moore, S.M., Lawson, M.F., Panettieri J r, R.A. & Chilvers, E.R. (1998) Platelet-derived growth factor-BB and thrombin ac tivate phosphoinositide 3 -kinase and protein kinase B: role in mediating airway smooth m uscle proliferation. Mol. Pharm. 54, 1 007–1015. 20. Nystedt, S., Ramakrishnan, V. & Sundelin, J. (1996) The pro- teinase-activated receptor 2 is induced by i nflammatory mediators in human endothelial cells. Comparison with the t hrombin receptor. J. Biol . Chem. 271, 1 4910–14915. 21. Miller, H .R.P., Huntley , J.F. & N ewlands, G.F.J. (1995) Mast cell chymases in helminthosis and hypersensitivity. In Mast Cell Proteases in Immunology and Biology (Caughey, G.H., ed.), pp. 203 –235. Marcel Dekke r, New York. 22. Macaldowie, C.N., Mackellar, A. & Huntley, J .F. (1998) The isolation and purification of a dual specific mast cell-derived protease from parasitised caprine jejunal t issue. Res. Vet. Sci. 64, 17–24. 23. Huntley, J.F., G ibson, S ., Brown, D., Sm ith, W.D ., Jackson, F. & Miller, H.R. (1987) Systemic release of a mast cell proteinase following nematode infections in sheep. Parasite Immunol. 9, 603–614. 24. Mansfield,L.S.&Gamble,H.R.(1995)Alveolarmastocytosisand eosinophilia in l ambs with naturally acquired nematode infections of Protostrongylus rufescens and H aemonchu s contortus. Vet. Immunol. I mmunopathol. 49 , 251–262. 25. Collie, D.D., McAldowie, C.N., Pemberton, A. D., Woodall, C.J., McLean, N., Hodgson, C. & Kennedy, M. (2001) Lo cal lung responses following local l ung challenge with r ecombinant lu ng- worm antigen i n systemically s ensitised s heep. Clin. Exp. Allergy 31, 1636–1647. 26. Ishihara, H ., Connolly, A.J., Zeng, D., Kahn, M .L., Zheng, Y.W., Timmons, C., Tram, T . & Coughlin, S.R. (1997) Protease-acti- vated recep tor 3 is a second thromb in re ceptor i n hu mans. Nature 386, 502–506. 27. Nakanishi-Matsui, M ., Zheng, Y W., Sulciner, D.J., Weiss, E.J., Ludeman, M.J. & Coughlin, S.R. (2000) PAR3 is a cofactor f or PAR4 activation by thrombin. Nature 404, 609–613. 28. Schechter, N.M., B rass, L.F., Lavker, R.M. & Jensen, P.J. (1998) Reaction of mast cell proteases tryptase and chymase with pro- tease activated receptors (PARs) on keratinocytes and fibroblasts. J. Cell. P hysiol. 176, 365– 373. 29. Vergnolle, N., MacNaughton, W.K., Al-Ani, B., Saifeddine, M., Wallace, J.L. & Hollenberg, M.D. (1998) Proteinase-activated receptor 2 (PAR2) -activating peptides: identification o f a receptor distinct from PAR2 th at regulates intestinal transport. Proc. Natl Acad. Sci. USA 95, 7766–7771. 30. Panettieri Jr, R .A., Hall, I.P., Maki, C .S. & Murray, R.K. (1995) Alpha-thrombin increases cytosolic calcium and induces human airway smooth muscle cell proliferation. Am. J. Respir. Cell Mol. Biol. 13, 205 –216. 31. Brown, J .K., Jones, C.A., Rooney, L.A. & Caughey, G.H. (2001) Mast cell tryptase activates extracellular-regulated kinases (p44/ p42) in airway smoo th-muscle c ells. Am. J . Respir. Cell Mol. Biol. 24, 146–154. 32. Belham, C.M., Scott, P.H., Twomey, D.P., Gould, G.W., Wads- worth, R.M. & Ple vin, R. (1997 ) Evidence t hat thrombin-stimu- lated DNA synthesis in pulmonary arterial fibroblasts involves phosphatidylinositol 3-kinase-dep endent p70 ribosomal S6 kinase activation. Cell Signal. 9, 109–116. 33. Coughlin, S .R. ( 2000) T hrombin s ignalling and protease-activated receptors. Nature 407, 258–264. 34. Belham, C.M., Tate, R .J., Scott, P.H., P emberton, A.D., Miller, H.R., Wadsworth, R.M., Gould, G.W. & Plevin, R. (1996) Trypsin stimulates proteinase-activated receptor-2-dependent and -independent activation of mitogen-act ivated protein kinases. Biochem. J . 320, 9 39–946. 1180 A. D. Pemberton et al. (Eur. J. Biochem. 269) Ó FEBS 2002 . Proteolytic action of duodenase is required to induce DNA synthesis in pulmonary artery fibroblasts A role for phosphoinositide 3-kinase Alan D. Pemberton 1 ,. G-protein-coupled receptor/PtdIns 3-kinase p athway in mediating duoden- ase-stimulated DNA synthesis. Effect of in ammatory cytokines on duodenase- induced DNA synthesis Pretreatment