Extended cleavage specificity of mMCP-1, the major mucosal mast cell protease in mouse—High specificity indicates high substrate selectivity

Abstract

Mucosal mast cells are in the mouse predominantly found in the epithelium of the gastrointestinal tract. They express the β-chymases mMCP-1 and mMCP-2. During nematode infections these intraepithelial mast cells increase in numbers and high amounts of mMCP-1 appear in the jejunal lumen and in the circulation. A targeted deletion of this enzyme leads to decreased ability to expel the intraepithelial nematode Trichinella spiralis. A suggested role for mMCP-1 is alteration of epithelial permeability by direct or indirect degradation of epithelial and endothelial targets, however, no such substrates have yet been identified. To enable a screening for natural substrates we performed a detailed analysis of the extended cleavage specificity of mMCP-1, using substrate phage display technology. In positions P1 and P1′ distinct preferences for Phe and Ser, respectively, were observed. In position P2 a high selectivity for large hydrophobic amino acids Phe, Trp and Leu was detected, and in position P2′ aliphatic amino acids Leu, Val and Ala was preferred. In positions P3 and P4, N-terminal of the cleaved bond, mMCP-1 showed specificity for aliphatic amino acids. The high selectivity in the P2, P1, P1′ and P2′ positions indicate that mMCP-1 has a relatively narrow set of in vivo substrates. The consensus sequence was used to screen the mouse protein database for potential substrates. A number of mouse extracellular or membrane proteins were identified and cell adhesion and connective tissue components were a dominating subfamily. This information, including the exact position of potential cleavage sites, can now be used in a more focused screening to identify which of these target molecules is/are responsible for the increased intestinal permeability observed in parasite infected mice.

Introduction

Mast cells (MCs) are potent inflammatory cells that are known primarily for their involvement in allergic reactions. However, MCs have recently also been shown to be of vital importance for orchestrating counterattacks on invading bacteria and viruses. They have a prominent role in linking innate and adaptive immunity and have also been shown to neutralize toxins from snakebites and bees (Metz et al., 2006, Nakae et al., 2005, Nakae et al., 2006). MCs are found throughout the body, however primarily at body surfaces and in the mucosa. Based on their tissue location and the content stored in the cytoplasmic granules two major subpopulations have been identified, mucosal MCs (MMCs), and connective tissue MCs (CTMCs). The MMCs are in rodents primarily found in the mucosal layers of the gastrointestinal and respiratory tracts. These MCs store chondroitin sulfate proteoglycans and low amounts of histamine. Several serine proteases with chymotrypsin-like activity, i.e. chymases, are also found in these granules. Chymases cleave substrates C-terminal of aromatic amino acids (aa), i.e. Phe, Tyr and Trp. In rodents the second major MC population, the CTMCs are primarily found in the connective tissue and in the peritoneum. Their cytoplasmic granules contain heparin proteoglycans, high amounts of histamine and proteases. These proteases include the metalloprotease carboxypeptidase A (CPA), and serine proteases of the chymase and tryptase subfamilies. Tryptases cleave substrates after basic aa, i.e. Arg and Lys.

Chymases are the most abundant proteases in rodent MCs and in particular in the MMCs. Based on phylogenetic analyses the chymase family is divided in two subfamilies, the α-chymases and the β-chymases (Chandrasekharan et al., 1996). One single α-chymase has been identified in all species analyzed except for ruminants where two closely related α-chymase genes have been described (Gallwitz and Hellman, 2006). In contrast, the β-chymases has only been found in rodents, and all investigated rodent species seem to express several members of this protease subfamily. Rodent MMCs express only β-chymases, i.e. mouse mast cell protease (mMCP)-1 and mMCP-2 in the mouse and rat mast cell protease (rMCP)-2, rMCP-3 and rMCP-4 in the rat (Gibson and Miller, 1986, Huang et al., 1991, Lutzelschwab et al., 1997, Lutzelschwab et al., 1998a, Lutzelschwab et al., 1998b, Pemberton et al., 2003, Scudamore et al., 1997). The CTMCs of mouse and rat express one α-chymase (mMCP-5 and rMCP-5, respectively) and one β-chymase (mMCP-4 and rMCP-1, respectively). Interestingly, both mMCP-5 and rMCP-5 have secondarily changed specificity and instead become elastases (Karlson et al., 2003, Kunori et al., 2002). Based on the relatively large differences between the two major subtypes of rodent MCs, they are thought to have distinct biological functions.

In humans the distinction between MMCs and CTMCs is not as clear as in rodents. The two subtypes contain heparin and are distributed more equal in human than in rodent tissues. Both MC subtypes in humans contain tryptase but only the CTMCs express the α-chymase and CPA.

MMCs are involved in several pathological conditions, i.e. atopic asthma, stress-induced enteropathies and reperfusion injuries (Boros et al., 1995, Castagliuolo et al., 1998, Schwartz, 1990). However, the primary role of rodent MMCs is believed to be in providing a first line of defence against intestinal nematodes. These infections are namely associated with intraepithelial MMC hyperplasia in the host (Miller, 1996, Woodbury et al., 1984). The MMC-specific β-chymases mMCP-1 and rMCP-2 are released during nematode expulsion and are found systemically and in the jejunal lumen at the time of worm expulsion (Huntley et al., 1990, King and Miller, 1984, Miller et al., 1983, Tuohy et al., 1990). The exact role of mMCP-1 and rMCP-2 in nematode expulsion is not yet known. Anaphylactic release of MC-derived mediators in Nippostrongylus brasiliensis-sensitized rats and also intravascular perfusion of purified rMCP-2 results, however, in macromolecular leakage over the mucosal epithelium and the vascular endothelium (Scudamore et al., 1995). An mMCP-1 knockout model has also been used to study the in vivo relevance of this enzyme during nematode infections. The expulsion of the parasitic nematode Trichinella spiralis was found to be delayed in mMCP-1^−/− mice compared to the control and an increased deposition of muscle larvae was also detected (Knight et al., 2000). However, the targeted deletion of mMCP-1 did not affect the expulsion of N. brasiliensis (Knight et al., 2000).

The MMC-derived chymases mMCP-1 and rMCP-2 seem to provide a mechanism important in the clearance of nematode infections. By increasing the permeability of mucosal epithelium and vascular endothelium, an efflux of inflammatory cells and plasma components are allowed into the intestinal lumen. The proposed proteolytic targets have been tight junction proteins or activation of matrix metalloproteases (MMPs). rMCP-2 has been shown to activate MMP-1 and -3 and to alter the tissue distribution of tight junction proteins ZO-1 and occludin, but the in vivo relevance of these effects remain to be investigated (Scudamore et al., 1998, Suzuki et al., 1995). Decreased degradation of occludin following T. spiralis infection in mice lacking mMCP-1 (McDermott et al., 2003) suggests that occludin may be a target, but apart from this indirect evidence, the substrate(s) for mMCP-1 that cause the increase in permeability have not yet been identified. The only known substrate found for this enzyme is angiotensin (Ang) I, which is converted to the potent vasoconstrictor Ang II, with only marginal degradation of this product (Saito et al., 2003).

While mMCP-1 has been manifested as a mediator in nematode expulsion in mice, the role of mMCP-2 in this process is still uncertain. Compared to mMCP-1, mMCP-2 is expressed at similar levels in jejunal MMCs, and at higher levels in the gastric mucosa (Pemberton et al., 2003). mMCP-2 is also upregulated during primary infection by N brasiliensis, and found in serum of infected mice, although at lower levels than mMCP-1, due to an inability to form complexes with serpins (Pemberton et al., 2003, Pemberton et al., 2006).

When mMCP-1 was identified it was characterized as an enzyme with chymotryptic activity (Newlands et al., 1987, Newlands et al., 1993). Like mMCP-1, mMCP-2 was also characterized as an active serine hydrolase, with a DFP-binding active site (Serafin et al., 1990). However, this enzyme seems to have very weak proteolytic activity, based on a screening of chromogenic substrates for chymotrypsin, trypsin, elastase, metase or aspartase activity (Pemberton et al., 2003). Only very low activity towards a chymotrypsin-sensitive substrate was detected, and no activity against these other substrates was observed (Pemberton et al., 2003). This can be the result of a proteolytically inactive enzyme, or possibly one with an extremely strict substrate specificity that prevents proteolytic cleavage of the substrates tested.

The primary specificity of mMCP-1 and mMCP-2 has thus been investigated, but the interactions with substrate residues distant from the P1 position have never been addressed. Information concerning positions P1–P4, N-terminal of the cleaved bond, for the corresponding rat enzyme rMCP-2 has however been obtained using chromogenic substrates (Powers et al., 1985). In the present study we have determined the extended substrate specificity of mMCP-1 using a substrate phage display approach. This methodology provides information of up to eight amino acids surrounding the cleavage site and thereby a more complete view of the substrate specificity. This is also the only technique that unbiased can detect cooperative effects of amino acid position N and C terminal of the cleavage site (manuscript in preparation Gallwitz et al.).

Our results show that mMCP-1 displays a high selectivity in the P2, P1, P1′ and P2′ positions, which indicates that this enzyme has a relatively narrow set of in vivo substrates. This information was used to screen the entire mouse proteome for potential substrates, a screening that led to the identification of a relatively narrow set of potential substrates of which a large fraction is cell adhesion molecules, connective tissue and immunologic components. Information concerning the location within the protein of the potential cleavage sites was also obtained. This information can now be used for a more detailed analysis of the potential targets and their role in the observed phenotypic changes observed in animals after injection or natural release of this important MC enzyme.