Proteinase activated receptor 2: role of extracellular loop 2 for ligand-mediated activation

Abstract

1. Rat proteinase-activated receptor-2 (PAR₂) variants were stably expressed in rat KNRK cells: (a) wild-type (wt)–PAR₂; (b) PAR₂PRR, with the extracellular loop 2 (EL-2) sequence P₂₃₁E₂₃₂E₂₃₃mutated to PRR and (c) PAR₂NET, with the EL-2 sequence, PEEV changed to NETL. Cell lines were evaluated for their sensitivity (calcium signalling) towards trypsin and the receptor-activating peptides, SLIGRL-NH₂, SLIGEL-NH₂, trans-cinnamoyl(tc)-LIGRLO-NH₂, and SFLLR-NH₂.

2. SLIGEL-NH₂ exhibited low potency (1 : 200 relative to SLIGRL-NH₂) in wild-type PAR₂. Its activity was increased 5 fold in PAR₂PRR, but it was inactive in PAR₂NET.

3. In PAR₂PRR, the potencies of SLIGRL-NH₂, tc-LIGRLO-NH₂, and SFLLR-NH₂ were decreased by 80–100 fold. But, the potency of trypsin was decreased by only 7 fold.

4. In PAR₂NET, highly homologous in EL-2 with proteinase-activated receptor-1 (PAR₁), the potency of the PAR₁-derived peptide, SFLLR-NH₂, was reduced by 100 fold compared with wt-PAR₂, whereas the potency of the PAR₂-derived AP, SLIGRL-NH₂ was reduced 10 fold. In contrast, the potency of trypsin in PAR₂NET was almost the same as in wt-PAR₂.

5. We conclude that the acidic EL-2 tripeptide, PEE, in PAR₂ plays an important role in governing agonist activity.

6. The data obtained with the PEEV→NETL mutation suggested: (a) that SLIGRL-NH₂ and SFLLR-NH₂ interact in a distinct manner with PAR₂ and (b) that SFLLR-NH₂ may interact differently with PAR₂ than it does with PAR₁.

7. The differential reductions in the potencies of SLIGRL-NH₂, compared with trypsin in the PAR₂PRR and PAR₂NET cell lines point to differences between the interactions of the trypsin-revealed tethered ligand and the free receptor-activating peptide with PAR₂.

Introduction

Proteinases such as thrombin, trypsin and tryptase are now known to regulate target tissues via the activation of proteinase-specific cell surface G-protein-coupled-receptors (Vu et al., 1991; Rasmussen et al., 1991; Nystedt et al., 1994; Ishihara et al., 1997; Xu et al., 1998; Kahn et al., 1998; Dery et al., 1998). The novel mechanism whereby the proteinase-activated receptors (PAR₁, PAR₃ and PAR₄ activated by thrombin; PAR₂ activated by trypsin and tryptase) are triggered involves the proteolytic unmasking of a nascent N-terminal receptor-activating sequence that acts as a tethered ligand (Vu et al., 1991; Chen et al., 1994). For three of the four PARs cloned to date, synthetic receptor-activating peptides (PAR-APs) based on the proteolytically revealed tethered ligand are able to act as surrogate receptor agonists, so as to mimic the effects of protease-mediated receptor activation. In our own work, in keeping with the results of others, we have explored the structure-activity relationships (SARs) for short (four to seven amino acids) receptor-activating peptides based on the tethered ligand sequences present in PAR₁ and PAR₂ (Scarborough et al., 1992; Vassallo et al., 1992; Hollenberg et al., 1992; 1993; 1996; 1997; Blackhart et al., 1996; Chao et al., 1992; Hui et al., 1992; Natarajan et al., 1995; Kawabata et al., 1999). Such SAR studies of the PAR-APs have pointed out the importance of the arginine residue at position 5 of a PAR₁-AP such as SFLLR-NH₂ (Hollenberg et al., 1996; Natarajan et al., 1995). In contrast, the substitution of an amino acid with an acidic side chain (e.g. glutamic acid) at position 5 of a PAR₁-AP (e.g. SFLLEN-NH₂) has been found to reduce peptide potency for activating PAR₁ by at least two orders of magnitude (Natarajan et al., 1995; Nanevicz et al., 1995). Although derived from the tethered ligand of PAR₁, SFLLR-NH₂ is also a relatively potent activator of PAR₂ (Hollenberg et al., 1997; Blackhart et al., 1996; Kawabata et al., 1999). Nonetheless, the PAR₂ derived activating peptide, SLIGRL-NH₂, is unable to activate PAR₁. As discussed below, extracellular loop 2 is believed to play an important role in governing the selectivity of PAR₁ for PAR-APs (Lerner et al., 1996).

As a counterpart to studies of the SARs for the activity of PAR-APs, other work has focused attention on the roles that the extracellular receptor domains of PAR₁ and PAR₂ play in determining agonist specificity (Nanevicz et al., 1995; Lerner et al., 1996; Gerszten et al., 1994). In particular, the extracellular loop 2 of PAR₁ and PAR₂ has been singled out as being a primary determinant of peptide agonist specificity (Nanevicz et al., 1995; Lerner et al., 1996). This receptor domain contains a short sequence (CHD) which is conserved amongst all PARs that have been cloned to date; and there is a strikingly high degree of homology between PAR₁ and PAR₂ for a stretch of about ten amino acid residues (Figure 1).

Figure 1.

Sequences of PAR₂ variants compared with PAR₂ and PAR₁ (upper panel) and their expression in KNRK cells (middle and lower panel). (Upper panel) The sequences of the extracellular loop 2 of wtPAR₂ (PAR₂) and the two receptor variants (PAR₂PRR, PAR₂NET) are compared with the extracellular loop 2 sequence of rat PAR₁. The bold print of the PAR₁ sequence denotes the region of high sequence homology between PAR₁ and PAR₂. (Middle panel) Cellular fluorescence intensity detected with the B5 anti-PAR₂ antiserum was monitored by fluorescence-activated cell sorting for untransfected KNRK cells (dashed lines; same signal was observed for vector-transfected cells) or for wt-PAR₂-expressing cells (solid lines). (Lower panel) The expression of receptor mRNAs for the host wild-type KNRK cells, (Wt-PAR₂), PAR₂PRR (PRR) and PAR₂NET (NET) were determined by RT–PCR. The RT–PCR signal observed for actin was the same in all cell lines (not shown). The positions of the oligonucleotide size markers (in base-pairs) are shown on the left (lane M). The position of the expected PCR product for PAR₂ (560 base-pairs) is shown on the right. Identical amounts of RNA obtained from each cell line were subjected to analysis by RT–PCR, as outlined in Methods.

In this extracellular loop 2 sequence, there is a conserved glutamic acid residue (E₂₃₂ in rat PAR₂; E₂₆₀ in human PAR₁) present in PAR₁ and PAR₂. In human PAR₁, E₂₆₀ (equivalent to E₂₆₇ in rat PAR₁) has been identified as potentially playing either a direct role in agonist binding or a regulatory function in modulating agonist access to a receptor docking site (Nanevicz et al., 1995). In the human PAR₁ thrombin receptor, low activity was observed for the PAR₁AP analogue, SFLLEN-NH₂, wherein an acidic side chain was substituted for the basic arginine side chain at position 5 in the parent agonist, SFLLRN-NH₂ (Nanevicz et al., 1995). However, functional complementation was achieved to enhance substantially the activity of SFLLEN-NH₂ by a single arginine substitution (E260R) in the extracellular loop 2 of human PAR₁. The Activity of the PAR₁-derived peptide, SFLLRN-NH₂, in the E260R human thrombin PAR₁ receptor mutant was only slightly lower than in the wild-type receptor (Nanevicz et al., 1995). Since E₂₆₀ in the human PAR₁ receptor is equivalent to E₂₃₂ in rat PAR₂, we wondered if this amino acid residue might play a comparable role in affecting the activity of PAR₂APs. We therefore synthesized and tested the activity of the PAR₂AP analogue, SLIGEL-NH₂, with a substitution of an acidic residue for a basic arginine side chain at position 5 of the parent PAR₂AP (SLIGRL-NH₂). Further, we prepared two rat PAR₂ mutants, one of which had a switch of acidic to basic residues (PAR₂PRR: E₂₃₂E₂₃₃→R₂₃₂R₂₃₃: Figure 1) in extracellular loop 2 and one of which was changed so that this amino acid portion of the PAR₂ extracellular loop 2 would have exactly the same sequence as human PAR₁ (PAR₂NET: P₂₃₁E₂₃₂E₂₃₃V₂₃₄→N₂₃₁E₂₃₂T₂₃₃L₂₃₄). The second substitution provides for an additional potential glycosylation site in the extracellular loop 2 of PAR₂. The activity of a number of PAR-APs as well as trypsin were assessed using a calcium signalling assay for the wild-type rat PAR₂ (wt-PAR₂) and the two receptor mutants (PEE→PRR; PEEV→NETL) expressed in KNRK cells.

In our experiments, we were interested in two main questions: (1) would the PAR₂PRR receptor be more sensitive to the PAR₂AP, SLIGEL-NH₂, compared with wt-PAR₂? and (2) would the PAR₂NET receptor, having an increased stretch of amino acid sequence homology with the extracellular loop 2 of PAR₁ have an increased sensitivity towards the PAR₁AP, SFLLR-NH₂?

Methods

Preparation of transfected KNRK cell lines

Rat PAR₂ (Saifeddine et al., 1996) was cloned into the pcDNA3 mammalian expression vector (In Vitrogen, San Diego, CA, U.S.A.) and the ‘wild-type' and mutated receptors were expressed in Kirsten virus-transformed rat kidney cells (KNRK, American Tissue Type Culture Collection, Bethesda, MD, U.S.A.). The receptor mutants, PAR₂PRR and PAR₂NET were prepared by the overlapping PCR approach wherein P₂₃₁E₂₃₂E₂₃₃ in rat PAR₂ were changed to P₂₃₁R₂₃₂R₂₃₃ and N₂₃₁E₂₃₂T₂₃₃L₂₃₄ respectively. In the background nontransfected KNRK cell line, PAR₂ can be detected by a polymerase chain reaction (PCR) approach, but insufficient receptor is expressed in the nontransfected cells to yield an appreciable intracellular calcium signal in response to high concentrations of either trypsin or PAR₂APs (Böhm et al., 1996; Al-Ani et al., 1999). Nor are cell surface PAR₂ receptors detectable by immunofluorescence, using a PAR₂-targeted receptor antibody (Al-Ani et al., 1999). Cells were transfected using the Lipofectamine® method according to the manufacturer's instructions (Gibco BRL, Gaithersburg, MD, U.S.A.) with 5 μg of each construct used per KNRK cell monolayer (60-mm² flask, 60% confluent). Transfected cells were subcloned in geneticin (0.6 mg ml⁻¹)-containing medium and receptor-bearing cells were isolated with the use of the anti-receptor B5 antibody (Kong et al., 1997) and with fluorescence-activated cell sorting to yield three permanent cell lines: wild-type KNRK-PAR₂ (wt-PAR₂), PAR₂PRR and PAR₂NET, as illustrated in Figure 1. The cell lines were routinely propagated in geneticin (0.6 mg ml⁻¹) containing Dulbecco's modified Eagle's medium (DMEM) supplemented with 5% (v  v⁻¹) foetal calf serum, using 80 cm² plastic T-flasks. Cells were subcultured by resuspension in calcium-free isotonic saline/EDTA solution, without the use of trypsin. Background KNRK cells were similarly grown in geneticin-free medium. A KNRK cell line transfected with an ‘empty' pcDNA3 vector was also subcloned and grown in geneticin-containing medium, as for the other transfected cell lines. The B5 anti-receptor antibody, used in previous work (Kong et al., 1997) was raised in rabbits using a peptide corresponding to rat PAR₂ (G₃₀PNSKGRSLIGRLDTP₄₆-YGGC) coupled to keyhole limpet hemocyanin (YCGGC added for conjugation). Cyanine 3-coupled goat anti-rabbit IgG (Cedarlane Laboratory Hornby, ON, Canada) was used to detect cell-surface bound B5 anti-receptor antibody. To assess the presence of increased PAR₂ mRNA in the transfected cell lines, total RNA was prepared from confluent T-flasks using the TRI® reagent (Molecular Research Center, Cincinnati, OH, U.S.A.). The RNA was reverse-transcribed (RT) with a first strand cDNA synthesis kit using pd(N)6 primer (Pharmacia LKB Biotechnology, Uppsala, Sweden) according to manufacturer's recommendations at 37°C for 60 min; 3 μl of this solution was used with primer pairs targeted to (1) Rat PAR₂ forward primer, PAR₂F: 5′- CACCACCTGTCACGATGTGCT-3′ and reverse primer, 5′- CCCGGGCTCAGTAGGAGGTTTTAACAC-3′, and (2) to actin: forward primer, 5′- CGTGGGCCGCCCTAGGCACCA-3′ and reverse primer, 5′- TTGGCCTTAGGGTTCAGGGGG-3′. The detection of an intron-free 243 base pair product using this actin primer pair which spans an intron in genomic DNA, can confirm the absence of DNA-derived intron sequences in the RT product obtained from the cell-derived DNA preparation (Watson et al., 1992). Routinely, amplification was done using 2.5 units of Taq DNA polymerase (Promega, Madison, WI, U.S.A.) in a 10 mM Tris-HCl buffer, pH 9.0 (50 μl, final vol.) Containing 1.5 mM MgCl₂, 50 mM KCl, 0.1% v v⁻¹ triton X-100 and 0.2 mM each of deoxynucleotide triphosphates. Amplification was allowed to proceed for 25 cycles beginning with a 1 min denaturing period at 94°C, followd by a 1 min reannealing time at 55°C and a primer extension period of 1 min at 72°C. The PCR products were separated by 1.5% agarose gel electrophoresis and visualized with ethidium bromide. A 560 base-pair product yielded by the PAR₂ amplimers has been previously documented by sequence analysis to represent rat PAR₂ (Saifeddine et al., 1996). For each transfected cell line, a population isolated by fluorescence-activated cell sorting for subculturing was obtained from a fraction in which >95% of the population was found to exhibit cell surface fluorescence using the B5 anti-receptor antibody. For each cell line so isolated, the approximate number of anti-receptor antibody binding sites per cell was estimated fluorometrically with the use of Quantum Simply Cellular® microbeads (Flow Cytometry Standards Corp, San Juan, PR, U.S.A.) according to the manufacturer's instructions, and in keeping with previous work (Lopez et al., 1992; Zagursky et al., 1995). However, our polyclonal rabbit B5 anti-receptor antibody was used instead of a monoclonal mouse anti-receptor antibody, as described in previously published work (Lopez et al., 1992; Zagursky et al., 1995).

Measurements of calcium signalling using fluorescence emission

Cells to be used of measurements of peptide-stimulated fluorescence emission (reflecting an increase in intracellular calcium) were grown to about 85% confluency in 80 cm² T-flasks and were disaggregated with calcium-free isotonic phosphate-buffered saline containing 0.2 mM EDTA. Disaggregated cells were pelleted by centrifugation and were resuspended in 1 ml DMEM/10% FCS for loading with the intracellular calcium indicator, Fluo-3 (Molecular Probes Inc., Eugene, OR, U.S.A.) at a final concentration of 22 μM (25 μg ml⁻¹) of fluo-3AM ester. Indicator uptake was established over 20–25 min at room temperature in the presence of 0.25 mM sulphinpyrazone, after which time cells were washed two times by centrifugation and resuspension with the buffer described below, to remove excess dye. Fluo-3-loaded cells were then resuspended to yield a stock solution (about 6×10⁶ cells ml⁻¹) in a buffer of the following composition (mM): NaCl 150, KCl 3, CaCl₂ 1.5, HEPES 20, glucose 10 and sulphinpyrazone 0.25. Fluorescence measurements, reflecting elevations of intracellular calcium, were conducted at 24°C using a Perkin-Elmer fluorescence spectrometer, with an excitation wavelength of 480 nm and an emission recorded at 530 nm. Cell suspensions (about 2 ml of approx. 3×10⁵ cells ml⁻¹) were maintained in suspension with a stirred (magnetic flea bar) thermostatted cuvette (total volume, 4 ml) and peptide stock solutions were added to monitor peptide-induced changes in fluorescence. To construct concentration-response curves for fluorescence yield, the signals caused by the addition of test peptides were expressed as a percentage (%A23187) of the fluorescence peak height yielded by replicate cell suspensions when treated with 2 μM of the ionophore A23187 (Sigma, St. Louis, MD, U.S.A.). This concentration of A23187 was at the plateau of its concentration-response curve for a fluorescence response. Under the assay conditions, the addition of proteinase inhibitors (e.g. Amastatin) did not potentiate or diminish the fluorescence response caused by the PAR-APs. Thus, routinely, proteinase inhibitors were not added to the assay cuvettes. The relative potencies of the PAR-APs studied, SLIGRL-NH₂, trans-cinnamoyl-LIGRLO-NH₂, SLIGEL-NH₂, SFLLR-NH₂, were expressed in two ways: (1) relative to the activity of the parent tethered ligand, SLIGRL-NH₂, in each of the three receptor constructs examined (wt-PAR₂, PAR₂PRR; PAR₂NET) and (2) for each peptide agonist acting in the mutant receptors, relative to their activity in the wild-type KNRK-PAR₂ cell line. Activity ratios for the PAR-AP peptides R_EC,SL-NH2 in each of the receptor systems were determined as outlined previously (Hollenberg et al., 1997) by using several points along the linear portions of the concentration-effect curves (e.g. Figures 3, 4 and 5) to estimate for each agonist, an average concentration ratio (R_EC,SL-NH2=EC_peptide÷EC_SL-NH2) related to a concentration of SLIGRL-NH₂ that caused the same calcium signal response as did the peptide studied. R_EC,SLNH2 values greater than unity denote peptides that were lower in potency than SLIGRL-NH₂. Similarly an activity ratio (R_EC,PAR2) for an individual PAR-AP as well as trypsin in each of the three expressed receptor systems was calculated using several points along the calcium signalling concentration-effect curves to estimate for each of the two receptor mutants, an average peptide concentration ratio (R_EC,PAR2=EC_mutant÷EC_PAR2) related to a concentration of agonist that in the PAR₂ mutant receptor caused a calcium signal equivalent to that observed in the wild-type receptor. Values of R_EC,PAR2 greater than unity denote a receptor that is less sensitive to the tested agonist.

Figure 3.

Concentration-effect curves for PAR-APs and trypsin in wt-PAR₂ cells. The fluorescence responses relative to the signal caused in identical cell suspensions by 2 μM A23187 (E₅₃₀: %A23187) were measured in replicate cell suspensions, as outlined in Methods, for increasing concentrations of the indicated PAR-APs and trypsin. Values represent the averages (±s.e.mean, bars) for measurements done with four or more replicate cell suspensions coming from two or more independently grown crops of cells. Error bars smaller than the symbols are not shown.

Figure 4.

Concentration effect curves for PAR-APs and trypsin in PAR₂PRR cells. The fluorescence responses relative to the signal caused by 2 μM A23187 (E₅₃₀: %A23187) were measured for increasing concentrations of the indicated PAR-APs and trypsin, exactly as described in the legend to Figure 3.

Figure 5.

Concentration-effect curves for PAR-APs and trypsin in PAR₂-NET cells. Concentration-effect curves for the fluorescence response relative to 2 μM A23187 (E₅₃₀: %A23187) were obtained for the same PAR-APs and trypsin, as outlined in the legends to Figures 3 and 4.

Peptides and other reagents

All peptides were synthesized by solid phase methods at the peptide synthesis Facility, University of Calgary, Faculty of Medicine (Calgary AB, Canada) (Director, Dr Denis McMaster), or were provided through the courtesy of Dr L. Leblond, via the peptide synthesis Facility at BioChem Therapeutic, Laval PQ, Canada. The composition and purity of all peptides were confirmed by HPLC analysis, mass spectral analysis and quantitative amino acid analysis. Stock solutions, prepared in 25 mM HEPES buffer, pH 7.4 were standardized by quantitative amino acid analysis to verify peptide concentration and purity. Porcine trypsin (14,900 U mg⁻¹, Cat. No. T7418) was from Sigma (St. Louis, MO, U.S.A.). A maximum specific activity of 20,000 U mg⁻¹ was used to calculate the approximate molar concentration of trypsin in the incubation medium.

Results

Receptor expression and comparison of agonist potencies

Using the fluorescence-activated cell sorting approach to obtain receptor-expressing KNRK cell lines and estimate receptor density, we observed, with the B5 anti-receptor antibody, that there was a comparable abundance of cell surface receptors (about 75,000 sites cell⁻¹) in the three cell lines, designated wt-PAR₂ (wild-type receptor), PAR₂PRR (PEE→PRR mutation), and PAR₂NET (PEEV→NETL mutation). In the vector-transfected KNRK cells, and in the background KNRK cells, no fluorescence above that observed with non-immune serum was detected with the use of the B5 anti-receptor antibody (Figure 1, middle panel; and see Al-Ani et al., 1999). In all receptor transfected cell lines, there was a marked shift to the right in the fluorescence intensity curve, using the B5 anti-receptor antibody probe (solid tracing, Figure 1, middle panel). The sequences confirmed for the extracellular loop 2 domains of the three receptor clones are shown in Figure 1 (upper panel), along with the sequence of rat PAR₁. A comparable abundance of receptor RNA was found for each receptor cell line using an RT–PCR approach (Figure 1, lower panel). Under the same conditions, the background KNRK cells yielded a very low PCR signal for PAR₂ (Figure 1, lower panel). The PCR signal obtained for actin (not shown) was comparable for all cell lines, for which the PAR₂ signals are shown in Figure 1 (lower panel). All three receptor-bearing cell lines yielded a calcium signal in response either to trypsin (20–100 nM) or to the PAR₂AP, SLIGRL-NH₂ (see below). The non-transfected KNRK cells and the vector-transfected KNRK cells did not yield a calcium signal at concentrations of trypsin (20 nM) and SLIGRL-NH₂ (50 μM) that were at the plateau for the concentration-effect curves for wt-PAR₂ (not shown and see Al-Ani et al., 1999). Thrombin (100 nM) was similarly unable to cause a calcium signal in the non-transfected or vector-transfected cells (not shown).

Since the three receptor-bearing cell lines were observed to have comparable receptor densities according to the fluorescence signal yielded by the B5 anti-receptor antibody, it was possible, as outlined in Methods, to compare the relative potencies of the agonists we tested in two ways: (1) In an individual cell line, the activity of each agonist could be measured relative to the action of the tethered ligand-derived PAR₂AP, SLIGRL-NH₂ (R_EC,SL-NH2: Table 1); and (2) for an individual agonist, an activity could be determined in the two mutant receptor cell lines, relative to its activity in the wild-type PAR₂KNRK cell line (R_EC,PAR2: Table 2). The relative activities were determined according to the relative concentrations of agonists that in each cell suspension caused an equivalent calcium signal, normalized to the signal generated by the addition of the ionophore, A23187, as summarized in Tables 1 and 2. Typical comparative responses of the three cell lines (calcium signals) to the PAR₂APs, SLIGRL-NH₂ and SLIGEL-NH₂ are shown in Figure 2. In Figure 2, the concentrations of peptides were selected so as to cause a comparable calcium signal, relative to the ionophore A23187, in each of the cell lines. The PAR₂NET cell line responded poorly to SLIGEL-NH₂, even at high concentrations (Figures 2f and 5).

Table 1.

Agonist activities^a relative to SLIGRL-NH₂

Table 2.

Relative receptor activities^a for each agonist

Figure 2.

Calcium signalling by wt-PAR₂ (a, b), PAR₂PRR (c, d) and PAR₂NET (e, f). Comparative responses to SLIGRL-NH₂ (SL-NH₂, ○) and SLIGEL-NH₂ (SE-NH₂, ▪) [fluorescence (E₅₃₀), reflecting increases in intracellular calcium] were monitored in fluo-3-loaded cell lines stimulated by either SLIGRL-NH₂ (a, c, e) or SLIGEL-NH₂ (b, d, f). The concentrations of peptide agonists were adjusted in an attempt to show comparable increases in fluorescence, relative to the signal yielded in each cell sample by 2 μM of the ionophore A23187 (▴). (a, b) wt-PAR₂; (c, d) PAR₂PRR; (e, f) PAR₂NET. The PAR₂NET cell line responded poorly to relatively high concentrations of SLIGEL-NH₂ (f and Figure 5).

Reduced activity of SLIGEL-NH₂, partially compensated by substituting PRR for PEE in PAR₂

In wt-PAR₂, the PAR₂AP analogue, SLIGEL-NH₂ was over two orders of magnitude less potent than the parent PAR₂AP, SLIGRL-NH₂, that has a basic instead of an acidic side chain at the fifth position corresponding to the revealed PAR₂ tethered ligand (Figure 3 and Table 1). When the two acidic residues in the P₂₃₁EE sequence of the extracellular loop of PAR₂ were changed to arginines (PAR₂PRR), the potency of the PAR₂AP analogue, SLIGEL-NH₂, in PAR₂PRR, was increased about 5 fold, relative to its potency in wild-type-PAR₂ (compare Figures 3 and 4 and see Table 2). In contrast, the potency of all other peptide agonists was reduced 80 to 100 fold, relative to their activities in wild-type PAR₂ (compare Figures 3 and 4 and see Table 2). Nonetheless, in the mutant PAR₂PRR receptor, the potencies relative to SLIGRL-NH₂ of the PAR₁-derived agonist, SFLLR-NH₂, and the PAR₂-derived agonist, trans-cinnamoyl-LIGRLO-NH₂ were equivalent to their relative potencies in wild-type PAR₂ (Table 1). In contrast, relative to the activity of SLIGRL-NH₂, SLIGEL-NH₂ was over twice as potent in the PAR₂PRR receptor, whereas it displayed a 180 fold lower potency than SLIGRL-NH₂ in the wild-type receptor (Figures 3 and 4 and see Table 1).

Differential changes in the potencies of trypsin and SLIGRL-NH₂ in PAR₂PRR

Surprisingly, on a molar basis, the activity of trypsin, which liberates the tethered PAR₂ ligand, SLIGRLDTP., was reduced by only about 8 fold in PAR₂PRR compared with the wild-type receptor (Table 2). This small reduction in potency contrasted with the 110 fold reduction in the potency of the PAR₂AP, SLIGRL-NH₂, derived from the tethered ligand sequence itself (Table 2). Thus, expressed relative to the potency of the free peptide, SLIGRL-NH₂, trypsin appeared to be almost an order of magnitude more effective in PAR₂PRR than it was relative to SLIGRL-NH₂ in the wild-type receptor (For trypsin, 10³×R_EC,SL-NH2≈percnt;0.04 in PAR₂PRR and ≈percnt;0.6 in wt-PAR₂: Table 1). Thus, the PEE→PRR mutation in extracellular loop 2 markedly reduced the potency of all PAR-APs tested that had a basic side chain at position 5, but did not, in the same proportion, appear to affect the activity of the tethered ligand revealed by trypsin. It is to be noted that the tethered ligand also has a basic side chain at position 5.

Relative activities of agonists in PAR₂NET: reduced activity of SFLLR-NH₂ and lack of effect for trypsin-mediated activation

In the PEEV→NETL mutant receptor (PAR₂NET), which has a 15 amino acid sequence stretch identical to PAR₁ (Figure 1), both the PAR₂AP, SLIGRL-NH₂, and the PAR₁AP, SFLLR-NH₂ displayed lower potencies than in the wild-type receptor (compare Figures 3 and 5, and see Tables 1 and 2). Compared with wt-PAR₂, the reduction in the potency of the PAR₁AP, SFLLR-NH₂ in the PAR₂NET receptor was more pronounced (about 40 fold lower, Table 2) than the reduction observed for the PAR₂APs, SLIGRL-NH₂ and trans-cinnamoyl-LIGRLO-NH₂ (about a 10 fold reduction: Table 2). The result of this differential shift in the potencies of SFLLR-NH₂ and SLIGRL-NH₂ was that the PAR₁AP was about 10 fold less potent than the PAR₂AP, in PAR₂NET, whereas these peptides had comparable potencies in either wild-type PAR₂ or the mutant PAR₂PRR receptor (Table 1). The PAR₂AP analogue with the acidic side chain, SLIGEL-NH₂ was essentially inactive in PAR₂NET (Figure 5). Significantly, although the potency of the PAR₂AP, SLIGRL-NH₂ was reduced by about 10 fold in PAR₂NET, the potency of trypsin in PAR₂NET was close to that in wt-PAR₂ itself (Table 2). Thus, as was the case for the PRR receptor mutant, the potency of the free peptide in PAR₂NET appeared to be much more affected than the potency of the trypsin-revealed tethered ligand (for trypsin, relative to SLIGRL-NH₂ in PAR₂NET: 10³× R_EC,SL-NH2=0.10, compared with 0.60 in wt-PAR₂, Table 1).

In summary, in PAR₂NET, there was a differential shift in the potencies of the PAR-APs, relative to the activity of trypsin. As indicated in Table 1, the relative orders of potencies for all agonists in the three cell lines were: (a) for wt-PAR₂: trypsin>> tc-LIGRLO-NH₂⩾SLIGRL- NH₂>SFLLR-NH₂ >>SLIGEL-NH₂, (b) for PAR₂PRR: trypsin>>>SLIGEL-NH₂>tc-LIGRLO-NH₂⩾SLIGRL-NH₂⩾SFLLR-NH₂, and (c) for PAR₂NET: trypsin>>>tc-LIGRLO-NH₂⩾SLIGRL-NH₂>>SFLLR-NH₂>>>>SLIGEL-NH₂.

Discussion

Complementation of SLIGEL-NH₂ activity in the PAR₂PRR mutant

One main finding of our study was that the PAR₂PRR receptor did exhibit an increased sensitivity towards the PAR₂AP, SLIGEL-NH₂ compared with wt-PAR₂. Previous work (Nanevicz et al., 1995; Lerner et al., 1996) had highlighted the importance in PAR₁ and PAR₂ of extracellular loop 2 (EL-2) as a key determinant of peptide agonist specificity. Further, the negatively-charged glutamic acid residue at position 260 in the EL-2 of human PAR₁ was found to play a governing role for the activity of the PAR₁APs, SFLLRN-NH₂ and SFLLEN-NH₂, in that the E₂₆₀R mutant of PAR₁ displayed a reduced (4 fold) sensitivity towards SFLLRN-NH₂, that has a basic side chain, and a substantially increased (about 100 fold) sensitivity towards SFLLEN-NH₂, that has an acidic side chain (Nanevicz et al., 1995). Qualitatively, our results with PAR₂PRR mirrored the data obtained previously with the human PAR₁E₂₆₀R mutant, in that compared with the wild-type PAR₂ receptor, the rat PAR₂PRR mutant showed about a 5 fold increased sensitivity for SLIGEL-NH₂. Like the PAR₁AP, SFLLEN-NH₂, the PAR₂AP, SLIGEL-NH₂ has an acidic side chain (Table 2). Our results were thus consistent with our working hypothesis that the P₂₃₁EE sequence in the EL-2 of PAR₂ plays an agonist docking or access role, akin to that proposed for the N₂₅₉ET sequence in extracellular loop 2 of human PAR₁ (Nanevicz et al., 1995). That said, the complementation of the low activity of SLIGEL-NH₂ in the wild-type receptor caused by the substitution of basic for acidic side chains in the PAR₂PRR receptor mutant was quite modest (about a 5 fold increase in potency) and was much less impressive than the large increase in the potency (about 100 fold) of SFLLEN-NH₂ caused by the E₂₆₀R mutation in human PAR₁ (Nanevicz et al., 1995). Further, the PAR₂PRR mutant, relative to wild-type PAR₂, exhibited a much more marked reduction (about 100 fold) in sensitivity towards SLIGRL-NH₂ than would have been expected from the data obtained with the E₂₆₀R mutant of PAR₁, where there was only a small (3 fold) reduction in sensitivity towards the PAR₁AP, SFLLRN-NH₂ (Nanevicz et al., 1995). Moreover, the PAR₂PRR mutant was slightly less able to discriminate between SFLLR-NH₂ and SLIGRL-NH₂ than was the wild-type receptor (R_EC,SL-NH2=1.9 in wt-PAR₂, compared with 1.4 in PAR₂PRR: Table 1). These quantitative differences between the results obtained with the ‘acidic' PAR-APs (containing glutamic acid) acting on the PAR₂ and PAR₁ receptor mutants containing E→R substitutions suggest that the mechanism of docking of the ligands with extracellular loop 2 of the PAR₁ and PAR₂ receptors may differ considerably. Notwithstanding, in general accord with the data obtained with mutant PAR₁ receptors, our data support the working hypothesis that in rat PAR₂ as in PAR₁, the charge properties of extracellular loop 2 (P₂₃₁E₂₃₂E₂₃₃ of PAR₂) play an important role in ligand recognition.

Differential effects on peptide and trypsin-mediated activation of PAR₂NET and PAR₂PRR

Since the PAR₁AP, SFLLR-NH₂ displays a potency lower than, but close to that of SLIGRL-NH₂ for activating PAR₂, (Blackhart et al., 1996; Kawabata et al., 1999 and Figure 3) we expected that the PAR₂NET mutant, wherein a 15 residue sequence of extracellular loop 2 is identical to PAR₁, would be even more sensitive to the PAR₁AP, SFLLR-NH₂, than was the wild-type PAR₂. To our surprise, compared with wild-type PAR₂, PAR₂NET was about 40 fold less sensitive to SFLLR-NH₂ whereas there was only about a 9 fold reduction in sensitivity towards SLIGRL-NH₂ and the PAR₂-selective peptide analogue, tc-LIGRLO-NH₂. It should be noted that SLIGRL-NH₂ cannot activate/interact with PAR₁, unless the extracellular loop 2 of PAR₂ is substituted into PAR₁ (Lerner et al., 1996). We were thus able to conclude that although the PAR₁ sequence LNITTCHDVLNETLL, is involved in the interaction of PAR₁ with either the PAR₁AP, SFLLR-NH₂ or the PAR₂AP, SLIGRL-NH₂, it would appear that in general, SFLLR-NH₂ seems to dock differently with the extracellular loop 2 of PAR₂ than it does with the extracellular loop 2 of PAR₁.

It is remarkable that the PEEV→NETL mutation in PAR₂NET did not noticeably affect the activity of trypsin (and presumably of the tethered ligand revealed by trypsin), whereas the activity of the synthetic receptor-activating peptide, SLIGRL-NH₂, was reduced by about 9 fold (Table 2). This result echoed the much lower reduction in the activity of trypsin (8 fold) compared with SLIGRL-NH₂ (110 fold) in PAR₂PRR (Table 2). One possibility, although unlikely, was that the accessibility of trypsin to the receptor cleavage/activation site might have been enhanced in PAR₂NET and PAR₂PRR, so as to compensate for a lowered intrinsic activity of the revealed tethered ligand. Notwithstanding, we believe that, taken together, the data indicate that the rat PAR₂ tethered ligand sequence revealed by trypsin cleavage (SLIGRLDTPP), remaining attached to the receptor, interacts differently with the receptor than does the free peptide, SLIGRL-NH₂ in solution. These possible differences between the interaction of the free and tethered ligand with the receptor may prove of importance in the context of developing PAR₂ receptor antagonists.

In summary, our data obtained from the assay of the activities of PAR-APs and trypsin in PAR₂PRR and PAR₂NET point to an important role for the acidic extracellular loop 2 tripeptide, PEE, in governing agonist activity in PAR₂. The results also suggest that there may be differences in the receptor activation interactions, between those of the tethered ligand and those of the synthetic activating peptides derived from the tethered ligand sequence. Further work with peptide based receptor crosslinking reagents may allow for a more precise analysis of the docking sites at which the activating peptides and the tethered ligand can interact.

Abbreviations

(Amino acids are abbreviated by their one-letter codes) AP

receptor-activating peptide

EL-2

extracellular loop 2

KNRK

Kirsten virus-transformed normal rat kidney cells

PAR

proteinase-activated receptor

PAR₁

proteinase-activated receptor-1 (activated by thrombin)

PAR₂

proteinase-activated receptor-2 (activated by trypsin & tryptase)

PAR₂NET

PEEV→NETL mutant of PAR₂ expressed in KNRK cells

PAR₂PRR

PEE→PRR mutant of PAR₂ expressed in KNRK cells

PAR-AP

PAR-activating peptide

SAR

Structure-activity relationship

tc

trans-cinnamoyl

Wt-PAR₂

wild-type rat PAR₂ expressed in KNRK cells