Histidines Are Critical for Heparin-Dependent Activation of Mast Cell Tryptase¹

Mast cells act in our innate defense against bacterial infections and parasites but are most known for their harmful effects during inflammatory conditions such as allergy and asthma (1, 2, 3, 4). When mast cells are activated, e.g., by cross-linking of the IgE molecules bound to their high-affinity receptor, FcεR1, on the cell surface, they release the contents of their secretory granules into the extracellular space (2). The secretory granules contain several preformed inflammatory mediators including histamine, heparin proteoglycan, cytokines, and proteases.

The mast cell proteases are divided into three categories: tryptases, chymases, and carboxypeptidase A, all of which are stored in their active form, i.e., with their activation peptides removed. In humans, the mast cell tryptases are further divided into α-, β-, and γ-tryptases. In mouse, three different tryptases, mouse mast cell protease (mMCP)³-6, mMCP-7, and a transmembrane tryptase, have been characterized (5). It was shown previously that β-tryptase, in its active form, is a tetramer (6), and subsequent crystallization studies provided evidence that all of the active sites were facing a narrow central pore of the tetramer (7, 8). This unique macromolecular arrangement results in an essentially complete resistance to macromolecular protease inhibitors (9) and a relatively narrow substrate cleavage profile. Tryptase has powerful proinflammatory properties and is gaining increased attention as a potential target for therapeutic intervention in a variety of diseases, e.g., allergic inflammation and fibrosis (10, 11).

In the mast cell granules, tryptase is stored in complex with negatively charged heparin proteoglycans. The critical importance of heparin proteoglycan for storage of tryptase (and other mast cell proteases) was highlighted in studies in which the gene for N-deacetylase/N-sulfotransferase 2 (NDST-2), an enzyme that holds a key position in the biosynthesis of mast cell heparin, was targeted. In the absence of functional NDST-2, peritoneal mast cells showed an essentially complete lack of stored mMCP-6, although the mRNA levels were not affected (12, 13). In contrast, the storage of mMCP-6 in bone marrow-derived mast cells was not affected by the NDST-2 knockout, probably due to the presence in this cell type of highly sulfated chondroitin sulfate, which can compensate for the lack of heparin (13). Although the tryptases are not basic proteins (pI values are close to neutral), they contain patches of positively charged amino acid residues on their surfaces, and it is thought that these regions mediate the interaction with glycosaminoglycans (8, 14, 15). The interaction of tryptase with heparin has been shown to have several important functional consequences. Apart from the critical role of heparin proteoglycan in storage of tryptase in the secretory granules, heparin has been implicated in the autocatalytic processing of protryptase into mature tryptase monomer (16). Furthermore, we have demonstrated that heparin is essential for the subsequent assembly of the mMCP-6 tetramer with concomitant enzymatic activation (17, 18), although other authors have reported that mMCP-6 activation/tetramerization may occur in the absence of heparin (19), and it has been reported that tetramerization/activation of human β-tryptase requires heparin (20). Moreover, it has been known for a long time that heparin is required for stabilization of the mature tryptase tetramer (21).

A previous study provided evidence for a striking pH dependence of human β-tryptase (22). It was found that β-tryptase was much more stable at slightly acidic (∼6) than at neutral pH. In a subsequent study, we showed that heparin-dependent activation and tetramerization of mMCP-6, the dominating tryptase in mouse, occurred only at acidic pH (below ∼6.5) (17). The reason behind the strong pH dependence of mast cell tryptase is not clear. However, a likely explanation is that His residues are involved in the interaction with heparin. At neutral pH, His residues (pKa, ∼6.5) are uncharged and thereby unable to engage in electrostatic interactions. In contrast, at pH 6.0, they are fully positively charged and can thus participate in electrostatic binding to heparin. Indeed, molecular modeling studies have previously suggested that His residues on the surface of human tryptase may engage in the binding to heparin (23). Furthermore, clustered His residues on the surface of mMCP-7 have been shown to mediate interaction with heparin, although mMCP-7 has been reported to be independent of heparin for activation and tetramerization (24). In the present study, we therefore investigated, by site-directed mutagenesis, whether surface-exposed His residues in mMCP-6 are involved in binding to heparin, tetramerization, and enzymatic activation.

The sequences for various tryptases were aligned using the program GeneDoc (www.psc.edu/biomed/genedoc). A model of mMCP-6 was created with the program O (25) using the human β-tryptase (Brookhaven Protein Data Bank entry 1A0L) as a starting model. The coordinates of identical residues between mMCP-6 and human β-tryptase (77%) were not altered in the Brookhaven Protein Data Bank file, whereas the remaining residues were mutated manually in O. Each residue was then modeled in a conformation chosen from a rotamer library that made structural sense with the surrounding residues. Electrostatic potentials were calculated using the MEAD (26), and visualized with the program DINO (www.dino3d.org) on a surface calculated with MSMS (27). The electrostatic potential calculations were done both in the fully protonated state (pH 6.0) and in the deprotonated state (pH 7.0).

mMCP-6, cloned into pcDNA3 (Invitrogen Life Technologies, Carlsbad, CA), was used as a template. The mutants were prepared in two PCR steps using fragment primers and mutation-containing primers (Table I) purchased from DNA Technology (Aarhus, Denmark). PCR products were purified from 1% agarose gels using a gel extraction kit from Qiagen (Hilden, Germany). The obtained PCR products and the pCEP-Pu2 vector (17) were digested sequentially, first with NheI in 1× Y+/TANGO for 1.5 h, and then with EcoRI in 2× Y+/TANGO for 1.5 h at 37°C. Ligations were performed using Rapid DNA Ligation kit from Roche (Mannheim, Germany). Transformations into Escherichia coli (MC1061) were performed by heat shock. Preparations of plasmid DNA were made using either QIAprep Spin mini-prep or midi-prep kit from Qiagen. Positive clones were confirmed by sequencing using Big Dye Sequencing kit (Prism; Applied Biosystems, Foster City, CA) and sequenced in a 310 Genetic Analyzer (Applied Biosystems Prism).

The mMCP-6 mutants were expressed in a human kidney cell line, 293-EBNA, essentially as described (17). All of the constructs contained an N-terminal BM40 signal peptide followed by a 6× His tag (used for purification) and an enterokinase (EK) site substituting for the native activation peptide. In this study, a modified transfection strategy was used: 4 μg of pCEP Pu2 vector containing mutated mMCP-6 was diluted in 750 μl of serum-free DMEM (National Veterinary Institute, Uppsala, Sweden) supplemented with 2 mM l-glutamine (National Veterinary Institute, Uppsala, Sweden) and mixed thoroughly. Then, 20 μl of freshly mixed PLUS reagent from Invitrogen Life Technologies was added, and the DNA-containing solution was mixed and incubated at room temperature for 15 min. Lipofectamine reagent (20 μl; Invitrogen Life technologies) was diluted with serum-free DMEM containing l-glutamine (750 μl) and vortexed. The DNA containing solution was then added to the Lipofectamine solution and mixed, followed by incubation at room temperature for 15 min. During that time, serum-free DMEM containing l-glutamine (5 ml) was added to 70–90% confluent 293-EBNA cells, grown in 10-cm culture dishes. The DNA-Lipofectamine solution was added to the cells and incubated for 3 h at 37°C/5% CO₂. The volume of the medium was increased to 12 ml, and FCS was added to give a final concentration of 10%. After 15–20 h, the medium was changed, and after 36–48 h, the medium was changed again, but this time, puromycin (0.5 μg/ml) from Sigma-Aldrich (Steinheim, Germany) was added for selection, and gentamicin (50 μg/ml) from National Veterinary Institute (Uppsala, Sweden) was included as antibioticum. After ∼1 wk, the selection pressure was increased (5 μg/ml puromycin) to favor cells with high copy numbers of the vector. Medium was harvested, and His-tagged fusion protein was purified as previously described (17). Mature mMCP-6 monomers were obtained by EK digestion of the mMCP-6 fusion protein (17). After EK treatment, no signs of undigested fusion protein were observed, as judged by SDS-PAGE analysis.

Tryptase activity was measured in 96-well microtiter plates. In standard conditions, EK-digested (17) mMCP-6 (wild type or mutants) was diluted with PBS (pH 6.0) to a final concentration of 1 μM. Protease was incubated either in the absence of heparin, or with 1 ng to 10 μg of pig mucosal heparin (commercial heparin; gift from U. Lindahl, Department of Medical Biochemistry and Microbiology, Uppsala University). After 30 min, 20 μl of the chromogenic peptide substrate S-2288 (2 mM in H₂O; Chromogenix, Mölndal, Sweden) was added, and tryptase activity was recorded using a Titertek Multiscan spectrophotometer (Flow Laboratories, McLean, VA). Control experiments showed that the tryptase activation process, both for wild-type mMCP-6 and for the H35A/H106A/H238A triple-mutant mMCP-6, was essentially completed within 5 min.

Fluorescence measurements were performed in an F-4000 fluorometer (Hitachi, Tokyo, Japan). Wild-type or mutant mMCP-6 protein was diluted in PBS (pH 6.0) containing 0.1% polyethylene glycol to a final concentration of 1 μM. Excitation was conducted at 280 nm, after which emission was recorded at 300–450 nm.

Size-exclusion gel chromatography was performed in a fast performance liquid chromatography system using a Superdex 200 column (10 × 300 mm). Both were from Amersham Pharmacia Biotech (Uppsala, Sweden). The column was equilibrated with PBS (pH 6.0), and run at a flow rate of 0.5 ml/min. Wild-type or mutants of mMCP-6 (20 μg in 200 μl) were analyzed either alone or after preincubation with 40 μg of heparin. Fractions (0.5 ml) were collected, and 100-μl samples of the fractions were analyzed for tryptase activity as above.

Affinity chromatography was performed on a 10-ml Poly-Prep column containing ∼0.5 ml of heparin-Sepharose (Amersham Pharmacia Biotech) equilibrated with binding buffer (PBS; pH 6.0 or 7.0). EK-digested wild-type or mutant mMCP-6 (40 μg in 200 μl PBS; pH 6.0 or 7.0) were applied to the column, followed by washing with 400 μl of binding buffer four times, and stepwise elution with PBS buffers containing increasing concentrations of NaCl. Protein concentration in eluted fractions was determined by the Quantigold assay (Diversified Biotech, Boston, MA), according to the procedure provided by the manufacturer.

RNA was prepared from 293 cells transfected with wild-type or H35A/H106A/H108A/H238A mMCP-6 using Nucleospin RNA II (Macherey-Nagel, Düren, Germany). First-strand synthesis and RT-PCR were performed using reagents and following the recommendations in the manual to SuperScript II RNase H⁻ Reverse Transcriptase (Invitrogen Life Technologies). Reverse mutant primer for H238A (used in first-strand synthesis) and forward mutant primer for H35A were used in the RT-PCR (see Table I).

Cells (0.9 × 10⁶) were grown to 50% confluence. Fresh medium (10 ml) was added, followed by addition of 100 μl of 0.5 M NH₄Cl or H₂O as a control. After 6-h incubation, the cells were resuspended, and the number of viable cells was examined by trypan blue exclusion. The NH₄Cl treatment did not lead to any reduction in cellular viability. Cells were centrifuged for 10 min (418 × g) and washed (one time), and cell extracts were prepared by adding 140 μl of 1× SDS-PAGE sample buffer. The supernatants and samples from cell medium were analyzed by SDS-PAGE (12% acryl amide) and subjected to Western blot analysis using a rabbit anti-mMCP-6 antiserum, as previously described (28).

To identify candidate His residues that may take part in the interaction with heparin, the mMCP-6 amino acid sequence was aligned with sequences of other mast cell tryptases. In particular, because human β-tryptase has been shown to display a similar pH dependence as mMCP-6 (22), it was important to identify His residues that were conserved between these two tryptases. From Fig. 1,A, it is apparent that altogether 13 His residues are present in mMCP-6. Of these, H27, H35, H44, H80, H106, H108, H163, and H238 show conservation between mMCP-6 and human β-tryptase, as well as considerable conservation among tryptases from other species. H5, H49, H82, H176, and H244 were not conserved between mMCP-6 and human β-tryptase, and were therefore not considered for mutational analysis, and the active site His (H44) was also deselected. To determine whether any of the conserved His residues are located in a way that is compatible with an ability to interact with heparin, mMCP-6 was subjected to molecular modeling on the basis of the known three-dimensional structure of human β-tryptase (7) (Fig. 1,B). It is evident from the three-dimensional model of mMCP-6 that H35, H106, H108, and H238 are all exposed on the surface of the mMCP-6 tetramer, and they were therefore considered as likely candidates for involvement in the interaction with heparin. H27, H80, and H163 are located at the A-D/B-C subunit interfaces and are not exposed to the exterior. Therefore, they were not considered as being likely binding partners for heparin. It is also apparent that H35, H106, H108, and H238 form a relatively contiguous region of positive charge connecting the A-B and the equivalent C-D subunits, respectively. Mapping of the overall electrostatic potential of the mMCP-6 tetramer showed that, at neutral pH (deprotonated His), the region connecting the A-B subunits (and C-D) displays a relatively low positive charge (Fig. 1,C, upper panel). However, at lower pH, i.e., with His residues in their protonated state, there is a clear contiguous region of positive charge in the corresponding region, due to the protonation of H35, H106, H108, and H238 (Fig. 1,C). Lowering of the pH does not result in any obvious major change in electrostatic potential in other parts of the mMCP-6 tetramer (Fig. 1, C and D; lower panels).

To investigate whether the conserved and surface-exposed His residues affect the activation and tetramer formation of mMCP-6, we constructed nine mutants using PCR-based site-directed mutagenesis (Table I). Four single mutants (H35A, H106A, H108A, and H238A), two double mutants (H35A/H106A and H106A/H108A), two triple mutants (H35A/H106A/H238A and H106A/H108A/H238A), and one quadruple mutant (H35A/H106A/H108A/H238A) were prepared. All mutants except the quadruple mutant (see below) were expressed at high levels in 293-EBNA cells and were purified in high yields.

Tetramerization of mMCP-6 was studied by size-exclusion gel chromatography in a fast performance liquid chromatography system. In the absence of heparin, neither of the single or double mutants nor wild-type mMCP-6 showed any signs of tetramerization (Fig. 2,A) or enzymatic activity (B; shown for wild-type mMCP-6). Note, however, that high-molecular-mass material was present in preparations of the triple mutants (H35A/H106A/H238A and H106A/H108A/H238A; Fig. 2,A). However, these compounds were not enzymatically active (Fig. 2,B), indicating aggregation into nonactive complexes. In the presence of heparin, tetramerization was observed both for wild-type mMCP-6 as well as for the mutants (Fig. 2,A), with active enzyme found almost exclusively in the tetramer position (B). However, the degree of tetramerization varied markedly between the different mMCP-6 variants. A major fraction of wild-type mMCP-6 was transferred into tetramer upon heparin addition, and a similar degree of tetramerization was seen for the single mutant H35A, indicating that the H35A mutation alone does not affect the interaction with heparin. In contrast, addition of heparin to the H106A or H108A mutants resulted in somewhat lower degrees of tetramer formation compared with wild-type mMCP-6. A small but detectable defect in tetramerization was also observed for the H238A mutant. A more pronounced defect in the tetramerization was seen for the H106A/H108A double mutant, whereas the H35A/H106A double mutant showed a similar degree of tetramerization as did the H106A single mutant, again indicating that H35A does not provide a major contribution to the heparin-induced tetramerization. An even more pronounced defect in tetramerization was seen for the H35A/H106A/H238A triple mutant, and the H106A/H108A/H238A triple mutant was almost completely unable to form active tetramers. Importantly, the specific activities of the formed tetramers were similar for wild-type mMCP-6 and H35A/H106A/H238A mutant (63 and 47 mOD/min⁻¹·μM⁻¹, respectively), indicating that histidine residues primarily influence the efficiency of the tetramerization process, but do not affect the actual enzymatic activity of the generated tetramers. Tryptophane fluorescence measurements indicated that the mutations did not cause any detectable changes in overall conformation of the mMCP-6 monomer (Fig. 2 C).

Wild-type mMCP-6 and the various mMCP-6 mutants were incubated with increasing amounts of heparin, followed by determination of enzymatic activity toward the chromogenic substrate, S-2288. The results displayed in Fig. 3 demonstrate that the triple-mutant H106A/H108A/H238A shows a drastically reduced degree of activation compared with wild-type mMCP-6, well in line with its defects in tetramerization (see above). A marked reduction in heparin-dependent activation was seen also for the H35A/H106A/H238A mutant. Even at higher levels of heparin (100 μg to 1 mg), the activity of the H35A/H106A/H238A mutant was 2- to 4-fold lower than that of wild-type mMCP-6 (not shown). Moreover, the double mutant, H106A/H108A, i.e., the double mutant with the largest defect in tetramerization (see Fig. 2 A), showed a low degree of heparin-induced activation. Some defects in heparin-induced activation were also seen for the H35A/H106A double mutant and for the H106A single mutant. Note that the differences in heparin-induced activities between wild-type and mMCP-6 mutants were most pronounced at low heparin concentrations. This, together with the finding that both mutant and wild-type mMCP-6 tetramers had comparable specific activities (see above), indicates that the targeted His residues affect the affinity for heparin, but that formed wild-type and mutant tetramers have similar characteristics in terms of enzymatic properties.

Experiments were performed to investigate whether the His mutations affected the affinity for heparin. At pH 6.0, wild-type mMCP-6 bound strongly to heparin-Sepharose, eluting at ∼1 M NaCl (Fig. 4,A). In contrast, at pH 7.0, wild-type mMCP-6 eluted between 0.14 and 0.3 M NaCl (Fig. 4,B), indicating low affinity for heparin. Affinity chromatography (at pH 6.0) of the H106A mutant showed that the material was divided into two distinct fractions (Fig. 4 C): approximately two-thirds of the material eluted at 1 M NaCl, whereas one-third was recovered in a lower affinity position, eluting between 0.3 and 0.5 M NaCl. The H106A/H108A double mutant showed a similar division into high- and low-affinity material as the H106A single mutant, but with a higher proportion (∼60%) eluting in the low-affinity position. Rechromatography of the high-affinity portion of the H106A/H108A mutant on heparin-Sepharose showed that it again bound with high affinity to the column (not shown). In contrast, when the low-affinity portion of the H106A/H108A mutant was rechromatographed on the affinity matrix, a division into approximately equal amounts of high- and low-affinity peaks was obtained (results not shown). When the H35A/H106A/H238A triple mutant was chromatographed on the heparin-Sepharose column, all of the material was found at the 0.3–0.5 M NaCl position, and the H106A/H108A/H238A mutant displayed an even more pronounced reduction in affinity for heparin, with most of the material appearing in the flow-through fraction.

Despite repeated transfections, it was not possible to purify even a small amount of the quadruple mutant (H35A/H106A/H108A/H238A) from the transfected 293-EBNA cells. Although the cells appeared normal after transfection, Western blot analysis of the collected medium failed to detect any immunoreactive mMCP-6 fusion protein. To investigate whether the cells expressed the construct, we performed RT-PCR on extracts from cells transfected with either wild-type mMCP-6 or the H35A/H106A/H108A/H238A mutant. In both cases, PCR products of the correct sizes were obtained, and control sequencing of the PCR products revealed that the quadruple mutant contained all desired mutations and that the sequence did not contain any errors (data not shown). To investigate whether the mutant mMCP-6 is translated but shows defects in secretion, Western blot analysis was performed on cellular extracts. Although it was evident that mMCP-6 protein of correct size was present in extracts from cells expressing wild-type mMCP-6, cells expressing the quadruple mutant contained a lower-molecular-mass immunoreactive protein, indicating that mMCP-6 protein is present but in a partially degraded form (data not shown). In addition, small amounts of an immunoreactive protein that is slightly larger than the wild-type protein was observed. Possibly, the latter band corresponds to fusion protein still containing the BM40 signal peptide (see Materials and Methods). When cells were exposed to 5 mM NH₄Cl for 6 h, small amounts of correctly sized quadruple mutant fusion protein was observed in the cell fraction but not in the culture medium (data not shown). Taken together, these results suggest that the quadruple mutant is produced but is, at least partly, degraded by the lysosomal pathway.

A large number of proteins have been ascribed heparin-binding properties. Noteworthy, in most cases where heparin-binding properties have been ascribed to a certain protein, the true physiological ligand is not heparin (located exclusively within mast cells) but rather heparan sulfate, a closely related but lower sulfated glycosaminoglycan that has a much broader tissue distribution than heparin (29). Because the major defect observed in animals lacking heparin is the drastic reduction in stored mast cell proteases (12, 13), it is evident that the mast cell proteases constitute major physiological binding partners for (mast cell) heparin. Previous studies based on molecular modeling and crystal structures have suggested potential heparin-binding sites in various mast cell proteases (8, 23, 30, 31). However, the actual binding site for heparin in the mast cell proteases has only been subject to limited biochemical investigations previously (14). The present study describes the heparin-binding site in mMCP-6, and thus provides the identification of a heparin-binding site in a physiological ligand to mast cell heparin.

In a previous report, we provided evidence that the tetramerization and accompanying activation of mMCP-6 was dependent on heparin (17) or similar anionic polysaccharides (18). In a subsequent study, human β-tryptase was also shown to require heparin for activation/tetramerization (20), indicating that the human and mouse tryptases are similar as regards dependence on heparin. In a more recent study, we examined the structural requirements, in terms of heparin, for the activation/tetramerization of tryptase. We found that the interaction of heparin with mMCP-6 was strongly dependent on anionic charge of the polysaccharide and that the minimal size of heparin that was required for binding to mMCP-6 was approximately a decasaccharide (18). Interestingly, although the decasaccharides bound strongly to mMCP-6, they were unable to induce tetramerization. Instead, the minimal size required for tetramerization was approximately twice the size of the minimal size needed for binding, i.e., a 20-saccharide. This indicates that heparin induces tetramerization by binding simultaneously to at least two tryptase monomers, thereby bridging them and facilitating the tetramerization process. In this study, we used intact pig mucosal heparin, i.e., heparin of a size that preferentially yields active tetramers in favor of active monomers.

Although the interaction between mast cell tryptase and heparin has been characterized to some extent in terms of the polysaccharide structures required (18, 32), the structures in tryptase that are involved have not been the subject of extensive investigation previously. When the crystal structure of human β-tryptase was solved, patches of positively charged residues (Lys/Arg) were found that, via the interaction with heparin, could bridge A-B and C-D regions (see Fig. 1 C, upper part of subunit A and lower part of subunit B) (8). Interestingly, the A-B and C-D interfaces display relatively few protein-protein interactions, indicating a need for other stabilizing factors such as bridging by heparin. In contrast, the A-C and B-D interfaces display a multitude of stabilizing contacts, suggesting that these interfaces are sufficiently stable in the absence of heparin.

Lys and Arg residues are obvious candidates for engagement in tryptase:heparin interaction. However, the strong pH dependence for the activation/tetramerization of mMCP-6 (17) appears to be in some disagreement with Lys and Arg as the principal players in the binding to heparin, because the electrostatic charges of the Lys/Arg residues are not affected by lowering the pH from neutral to ∼6.0. This led us to hypothesize that His residues may be involved in the interaction with heparin and subsequent tetramerization/activation. A dependence of human β-tryptase on His residues for tetramerization has also been suggested, based on a similar strong pH dependence of the human protease for stability and reactivation (22). Furthermore, based on molecular modeling studies, surface-exposed His residues in human β-tryptase have been proposed to participate in the binding to heparin, at least intracellularly (23). Our results indeed demonstrate that His residues are of critical importance for binding to heparin, tetramerization, and activation of mMCP-6. Importantly, the mutants that were the most defective in either of the studied parameters were equally affected in the other parameters studied, indicating that binding to heparin, tetramerization, and activation are strongly associated events. It is also important to note that the His residues that were shown to be involved in the interaction with heparin are located in a way that, if interacting with heparin, bridging of the A-B and C-D interfaces will be accomplished. These results thus support the previous hypothesis (8) that heparin is required to connect the A-B and C-D interfaces, whereas the A-C and B-D interfaces are sufficiently stable without bridging by heparin. Of the single mutants, the H106A mMCP-6 was the mutant that displayed the most pronounced defects in the interaction with heparin. Interestingly, H106 is the His residue that is closest to the subunit interface (see Fig. 1 B), suggesting that this region may be of particular importance for binding to heparin. Furthermore, the combined mutation of H106 and its close neighbor, H108, resulted in a marked reduction in tetramerization, activation, and heparin binding. Importantly, the defects displayed by the H106A/H108A mutant were more pronounced than the defects exhibited by the other double mutant tested, further supporting that the His residues that are closest to the A-B and C-D interfaces are of particular importance for the interaction with heparin. Even more pronounced defects were obtained by the triple mutation of H106, H108, and H238, clearly indicating a central role for surface-exposed His residues in the interaction with heparin. It is important to stress that mutation of the three His residues in each single monomer will cause a reduction in total surface-exposed His residues from 16 to 4 in each tryptase tetramer.

At acidic pH, wild-type mMCP-6 binds very tightly to heparin-Sepharose (Fig. 4), with ∼1 M NaCl being required for elution from the column. Possibly, mMCP-6 tetramers are being formed during the affinity chromatography and the high-affinity binding of wild-type mMCP-6 to heparin-Sepharose could thus be explained by cooperative His-dependent binding of mMCP-6 tetramers to the affinity matrix. The central role for His residues in the interaction of mMCP-6 with heparin suggests that mMCP-6 interacts strongly with heparin whenever the pH is low, e.g., in the secretory granule. In contrast, when the protease is exposed to neutral pH, e.g., after mast cell degranulation, the protein should be dissociated from the proteoglycan. In line with this notion, it has been shown that mMCP-7 rapidly diffuses away following mast cell degranulation (15), in agreement with His residues being responsible for the interaction with heparin (14). In contrast, mMCP-6 was retained in the extracellular matrix after degranulation (15). The reason why mMCP-6 is retained despite being exposed to neutral pH is not clear. Possibly, the much higher affinity of mMCP-6 for heparin than that of mMCP-7 (17, 19, 24) could be related to the differing tendencies to be retained by the matrix. We may also note that mMCP-6 shows partial and low-affinity binding to heparin-Sepharose also at neutral pH (Fig. 4). The basis for this low-affinity binding is not known, but we may speculate that the high-affinity binding to heparin is mainly due to interaction with His residues, but that Lys/Arg residues may give some contribution to the binding energy. Furthermore, retention of mMCP-6 could be explained by interaction with other extracellular matrix components than proteoglycans.

The critical importance of surface-exposed His residues in mMCP-6 is further accentuated by the experiments performed on the mMCP-6 mutant where all of the identified surface-exposed His residues were mutated. It was found that this mutant protein was degraded intracellularly by the 293-EBNA cells, indicating that the absence of all of the surface-exposed His residues causes major defects in the protein that prevents its secretion, and that the defective protein is routed to intracellular degradation. The exact mechanism behind this observation is not clear. However, we note that both the H35A/H106A/H238A and, in particular, the H106A/H108A/H238A mutants show tendencies to aggregate into inactive compounds (Fig. 2 A). It is thus possible that the introduction of one extra His mutation in the quadruple mutant causes an even increased tendency for nonproductive aggregation or misfolding, and that the severely defective protein is sorted into degradative pathways. On a different angle, we may speculate that mMCP-6 needs to interact with heparin (or similar polysaccharides) during proper transport along the secretory pathway, and that the absence of the surface-exposed His residues, and an accompanying reduced ability to interact with heparin, leads to missorting of the protein.