0014-2980/98/0303-1022$17.50+.50/0 © WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998

Characterization of mouse mast cell protease-8,the first member of a novel subfamily of mousemast cell serine proteases, distinct from both theclassical chymases and tryptases

Claudia Lützelschwab1, Mallen R. Huang1, Marika C. Kullberg2, Maria Aveskogh1 andLars Hellman1

1 Department of Medical Immunology and Microbiology, University of Uppsala, BiomedicalCenter, Uppsala, Sweden

2 Department of Immunology, Stockholm University, Stockholm, Sweden

Using a recently developed PCR-based strategy, a cDNA encoding a novel mouse mast cell(MC) serine protease (MMCP-8) was isolated and characterized. The MMCP-8 mRNA con-tains an open reading frame of 247 amino acids (aa), divided into a signal sequence of 18 aafollowed by a 2-aa activation peptide (Gly-Glu) and a mature protease of 227 aa. The matureprotease has an Mr of 25072, excluding post-translational modifications, a net positivecharge of +12 and six potential N-glycosylation sites. MMCP-8 showed a high degree ofhomology with mouse granzyme B in the critical regions for determining substrate cleavagespecificity, indicating that MMCP-8, similar to granzyme B, preferentially cleaves after Aspresidues. A comparative analysis of the aa sequence of MMCP-8 with other hematopoieticserine proteases shows that it is more closely related to cathepsin G and T cell granzymesthan to the MC chymases. We therefore conclude that MMCP-8 belongs to a novel sub-family of mouse MC proteases distinct from both the classical chymases and tryptases.Southern blot analysis of BALB/c genomic DNA indicated that only one MMCP-8 gene (orMMCP-8 like gene) is present in the mouse genome. Northern blot analysis of rodent hema-topoietic cell lines revealed high levels of MMCP-8 mRNA in a mouse connective tissue MC-like tumor line. However, MMCP-8 mRNA could not be detected in mouse liver, intestine,lung or ears, indicating very low expression in normal tissues. Analysis of the expression ofdifferent MMCP in the tissues of Schistosoma mansoni-infected BALB/c mice showed astrong increase in MMCP-8 levels in the lungs but not in the intestines of infected animals,suggesting the presence of a novel subpopulation of MC in the lungs that expressed MMCP-8, either alone or in combination with MMCP-5 and carboxypeptidase A. The dramaticincrease in MMCP-1 and MMCP-2 levels but not of MMCP-8 in the intestines of parasitizedanimals also shows that MMCP-8 is not expressed in mucosal MC in the mouse. This latteris in clear contrast to what has been observed in the rat where the MMCP-8 homologues,RMCP-8, -9 and -10, can be considered as true mucosal MC proteases.

Key words: Serine protease / Mast cell / MMCP-8

Received 8/9/97Revised 17/11/97Accepted 5/12/97

[I 17503]

Abbreviations: BMMC: Bone marrow-derived mast cellMC: Mast cell MTC: Mastocytoma tumor cell MMC:Mucosal mast cell CTMC: Connective tissue mast cellMMCP: Mouse mast cell protease RMCP: Rat mast cellprotease

1 Introduction

Mast cells (MC) are distributed throughout mammaliantissues where they serve as important effector cells dur-ing inflammatory responses, including allergic reactions.

MC are often found in close association with blood ves-sels and nerves, suggesting a role in the control of vas-cular permeability and in neurogenic induction ofinflammation. Based on phenotypical differences, histo-chemical staining properties and secretory granule com-ponents, it has become evident that in mammals, MCcan be discriminated into at least two separate sub-populations, mucosal (MMC) and connective tissue MC(CTMC) [2–3]. The most abundant secretory compo-nents are various neutral proteases. Due to their highdegree of tissue specificity, these proteases have beenused extensively as markers for the identification and

1022 C. Lützelschwab et al. Eur. J. Immunol. 1998. 28: 1022–1033

Figure 1. The figure shows the complete nucleotide sequence of the MMCP-8 cDNA clone and the deduced aa sequence(EMBL accession number X78545) in comparison with the rat homologues RMCP-8, RMCP-9, RNGL-II and RMCP-10. Thecleavage site for the signal sequence and the activation peptide are marked by arrows. The two aa of the activation peptide areunderlined. The positions of the three aa of the catalytic triad are shown within boxes. A schematic drawing of the cDNA isdepicted at the top of the figure. The coding region of the mature processed protein is depicted as a grey box. The signalsequence and the activation peptide are shown as black and white boxes, respectively.

characterization of MMC and CTMC. Among the MCneutral proteases we find several members of the serineprotease gene family as well as a MC-specific carboxy-peptidase A. MC serine proteases have been classifiedaccording to their cleavage specificities as chymases ortryptases. Several members of the mouse chymasefamily have been isolated as genes or cDNA. Theseinclude the chymases mouse mast cell protease(MMCP)-1 [4], MMCP-2 [5], MMCP-4A [6], MMCP-4B [4],

MMCP-L [6], and MMCP-5 [4, 7]. Two mouse MC trypta-ses have also been characterized: MMCP-6 [8–10] andMMCP-7 [11]. Cells of several other hematopoietic celllineages, i.e. cytotoxic T cells, NK cells, monocytes andneutrophils, express one or several serine proteases,which are all stored within the granules of the cells.Nearly all of these proteases are relatively closely relatedto the MC chymases [12].

Eur. J. Immunol. 1998. 28: 1022–1033 Mouse mast cell protease 8 1023

Figure 2. Amino acid sequence comparison of MMCP-8, RMCP-8, RMCP-9 and RNGL-II. The positions of the three aa of thecatalytic triad are marked with asterisks. The potential glycosylation sites are encircled. The first and the second arrows indicatethe cleavage sites for the removal of the signal sequence and the activation dipeptide, respectively. The aa of the correspondingposition in granzyme B (positions 195, 196, 198, 215, 218 and 226) of the six aa known to be of importance for the substrate siteselection of human proteinase 3 have been depicted below the sequence of RNGL-II (rat granzyme like protein II). The glycine inposition 218 of granzyme B, which differs significantly from the corresponding aa in MMCP-8, has been marked by a black dot.

All serine proteases share a common catalytic triad, His-Asp-Ser, which is responsible for the proteolytic activityof the enzyme. Based on sequence conservation in theregion surrounding the histidine and serine residues, aPCR-based strategy for the isolation of novel membersof this gene family was recently developed [13]. The useof this technique led to the isolation and characterizationof several novel rat MC proteases [14], and mousehomologues to the human neutrophil proteases 3 andcathepsin G [13]. Using this technique we here report thecloning and structural analysis of MMCP-8, a novel MCprotease. A comparative analysis showed that MMCP-8is more closely related to cathepsin G and the T cellgranzymes than to the MC chymases, thereby forming anew subfamily of MC proteases.

2 Results

2.1 Isolation and nucleotide sequence analysisof cDNA clones for MMCP-8

To study the complexity of serine proteases expressedby mammalian hematopoietic cells, and to clone newmembers of this multigene family, we recently developeda strategy based on the use of PCR primers directedagainst highly conserved regions around the histidineand serine residues of the catalytic triad of these en-zymes [13, 14]. Using this technique, PCR amplificationwith cDNA from the mouse MC line MTC resulted in astrong PCR product in the region of 450–520 bp. DNAisolated from this PCR band was cloned into a plasmidvector and the resulting clones were plated on several

identical filters. To identify novel MC proteases, these fil-ters were probed with either the PCR primers or with amix of probes corresponding to the previously identifiedMC serine proteases MMCP-1 to MMCP-7. Clones thatwere found positive for the PCR primers but were nega-tive for previously isolated serine proteases wereselected for further analysis. One of the clones wasidentified as mouse cathepsin G and another cloneshown to encode a novel mouse protease which wasgiven the name MMCP-8. To obtain the sequence of theentire coding region for this protease, the approximately450-bp-long PCR product for MMCP-8 was used asprobe to screen a mouse MC cDNA library [4, 15]. Sev-eral clones were obtained, of which one contained theentire coding region of this novel serine protease. Thenucleotide sequence for the full-length cDNA clone ofthe protease is shown in comparison with the rat homo-logues in Fig. 1. The open reading frame of 741 bp en-codes a signal sequence of 18 amino acids (aa), a 2-aaactivation peptide (Gly-Glu) and the 227-aa mature pro-tein (Fig. 1). The coding region is followed by a 3'untranslated region of 85 bp with the canonical AATAAApolyadenylation signal located at position 815. The Mr ofthe mature protease is 25072 and it has a net positivecharge of +12. MMCP-8 contains an exceptionally highnumber of potential glycosylation sites (Fig. 2). The sixpotential N-linked glycosylation sites are at positions 21,51, 81, 131, 159 and 168 of the mature protein.

A comparative analysis of the evolutionary relationship ofMMCP-8 with other mouse, rat and human hemato-poietic serine proteases is presented as a phylogenetictree in Fig. 3. This analysis shows that MMCP-8 is a

1024 C. Lützelschwab et al. Eur. J. Immunol. 1998. 28: 1022–1033

Figure 3. A phylogenetic tree, based on the aa sequencesimilarities between mature MMCP-8 (i.e. devoid of signaland activation peptides) and all previously characterizedhematopoietic serine proteases in rodents and man, mouseand human adipsins (complement factor D), and threehuman pancreatic serine proteases (trypsin I, trypsin II andchymotrypsin). MC serine proteases are indicated in lightgrey and granzymes and cathepsin G are indicated in darkgrey. The position of MMCP-8 and its rat homologues aremarked by a black square to facilitate their identification.The sequence divergence in percent is shown at the bottomof the figure. The bootstrap tree was constructed from 1000independent comparisons.

member of a new subfamily of serine proteases moreclosely related to mouse cathepsin G and the T cell gran-zymes than to the MC chymases or tryptases.

2.2 Genomic Southern blot analysis of the genesfor MMCP-8

To determine the number of MMCP-8-related genes inthe mouse genome, a Southern blot analysis with liverDNA from BALB/c mice was performed (Fig. 4). Theresults show that there is most likely only one gene forMMCP-8 in the mouse genome (Fig. 4). This conclusionis also supported by the findings that identical sequen-ces originating from complete or partial nucleotidesequences were obtained from five independent MMCP-8 cDNA clones isolated from the same cDNA library, andthat identical partial DNA sequences were found in sev-eral genomic clones (data not shown).

2.3 Analysis of the expression of MMCP-8 andother mast cell-related differentiationmarkers in a panel of rodent hematopoieticcell lines

To study the tissue distribution of the newly clonedMMCP-8 and to relate its expression to other hemato-poietic differentiation markers, total RNA from a panel ofmouse and rat hematopoietic cell lines was studied byNorthern blot analysis. Six almost identical blots con-taining the RNA from the different cell lines were hybrid-ized in successive rounds with a panel of differentprobes (Fig. 5). Among the lines tested were cell linesexhibiting characteristics of mouse immature myeloidprecursors (M1D), mouse immature monocytes (WEHI-274.1), mouse monocytes or macrophages (IC-21 andJ774), mouse plasma cells (J558), mouse T helper cells(EL4), mouse cytotoxic T cells (CTLL), mouse connectivetissue mast cells (MTC), and rat mucosal mast cells(RBL-1). In addition to the cell lines, total RNA from bonemarrow-derived MC [M-BMMC(3w)] from a 3-week-oldculture maintained in the presence of conditioned mediafrom WEHI-3B and from purified mouse peritoneal mac-rophages (M-M ¤ ) was included in this analysis.

A strong hybridization signal was detected for MMCP-8in MTC cells and of the rat counterparts in RBL-1 cells,and a low level of MMCP-8 expression was detectedboth in BMMC cells and in the mouse CTLL line (Fig. 5).These results indicated that MMCP-8 and its rat homo-logues are expressed in both MMC and CTMC in rat andmouse and possibly at low levels in some T cell popula-tions.

Eur. J. Immunol. 1998. 28: 1022–1033 Mouse mast cell protease 8 1025

Figure 5. Northern blot analysis of total RNA from variousmouse and rat cell lines, mouse peritoneal macrophages (M-M ¤ ), and cells from a 3-week culture of mouse BMMC. Fora more detailed description of the different cell lines seeSect. 4.4. Only the relevant region of the Northern blots aredepicted (in general no other bands were seen on the blots).In the lane “MMCP-1” all mouse samples were hybridizedwith the mouse MMCP-1 probe and the rat RBL-1 cell linewas hybridized with a probe for RMCP-2, the rat homologueto MMCP-1. A human GAPDH probe was used as a refer-ence gene to equalize the amounts of RNA loaded in eachlane. nd (not determined).

Figure 4. Genomic Southern blot analysis of the gene forMMCP-8. Ten micrograms of BALB/c mouse chromosomalDNA were digested with the restriction enzymes Eco RI, Hin-d III, Bam HI and Xba I. The blot was hybridized with the full-length cDNA probe for MMCP-8, and was washed atmedium stringency.

The serine proteases MMCP-5 and MMCP-6 showed aclear tissue specificity for connective tissue like MC(Fig. 5). Expression of the MMCP-4 gene was detected inthe MTC cell tumor, consistent with the finding that alarge percentage of the cells within this tumor appear tohave a phenotype of relatively mature CTMC. The gen-eral view is that MMCP-4 is specifically expressed inhighly differentiated or terminally differentiated CTMC[16]. Accordingly, the BMMC cells grown in WEHI-3B-conditioned media, which are considered to representrelatively immature MC, did not express this serine pro-tease. The expression of MMCP-2 in the MTC tumor is inagreement with previously published results on theexpression of the various mast cell proteases in otherCTMC-like cell lines [5, 16].

One clear discrepancy with previously published resultswas seen in the expression pattern for MMCP-7, one ofthe MC tryptases. In this study the expression of MMCP-7 was not, as reported [11, 16] limited to CTMC-like celllines and to BMMC cultures, but was also detected inmouse peritoneal macrophages (Fig. 5). The MMCP-7probe was actually isolated from peritoneal macro-phages by PCR amplification, allowing us to concludethat MMCP-7 and not a closely related protease isexpressed by these cells. As a control, the same macro-phage RNA preparation was negative for both MMCP-5

and MMCP-6. In addition, lack of MMCP-7 expression inmature serosal MC has been reported previously [11].These results rule out the possibility that a small amountof contaminating serosal MC in the macrophage prepa-ration were responsible for the MMCP-7 hybridizationsignal.

The mast cell carboxypeptidase A was detected in theimmature MC of the BMMC culture and in the CTMC-likecell line MTC but not in the rat MMC line RBL-1, resultswhich are in agreement with previously publishedobservations [17]. In addition, carboxypeptidase A wasdetected in the T helper cell line EL4 (Fig. 5), an unex-pected finding which possibly can be explained as an

1026 C. Lützelschwab et al. Eur. J. Immunol. 1998. 28: 1022–1033

Figure 6. Northern blot analysis of normal BALB/c mousetissues (lung, intestine, spleen, ear, thymus and liver). Amouse § -actin probe was used as a reference gene toequalize the amounts of RNA loaded in each lane. Depend-ing on strong cross-hybridization between the § -actin probeand the g -actin mRNA, the signal detected on the blot is thecytoplasmic g -actin.

Table 1. A summary of our present view of the expressionof various proteases in different mouse MC populations

CTMC MMC BMMCIL-3

Lung MC(Schistosoma-

infected)

MMCP-2a) MMCP-1 MMCP-2a) MMCP-8MMCP-4 MMCP-2 MMCP-5MMCP-5 MMCP-6 (MMCP-5?)MMCP-6 MMCP-7b)

MMCP-7Carb.Ac)

Carb.A(Carb.A?)

a) MMCP-2 is not expressed in CTMC or BMMC of BALB/c mice but it seems to be expressed in most otherstrains of mice.

b) MMCP-7 is expressed transiently in BMMC culturespropagated in the presence of IL-3 (or IL-3-containingWEHI-3B-conditioned medium).

c) Carb.A: carboxypeptidase A.

aberrant expression most likely limited to this particularcell line.

Low levels of cathepsin G mRNA were detected in theearly myeloid M1 (subline D), in WEHI-274.1 cells and inBMMC cells, whereas high mRNA levels were detectedin the CTMC-like MC tumor, MTC (Fig. 5).

Mouse proteinase 3 (myeloblastin) was expressed inseveral cell lines of both myelo-monocytic origin and inthe B cell line J558 but not in any of the MC lines.

A high level of expression was detected for T cell gran-zyme B in both the mouse CTLL line and in the mucosalMC line RBL-1 (Fig. 5). However, only a few passages ofthe RBL-1 cell line expressed granzyme B, indicatingthat granzyme is only rarely expressed in this cell line(data not shown). A summary of our present view of theexpression of various proteases in different mouse MCpopulations is shown in Table 1.

The expression of a few additional markers was investi-gated to confirm the lineage commitment of the differentcell lines. The IgE receptor § and g chains genes werefound to be expressed only in the MC lines and in cells ofthe BMMC culture which confirms the lineage commit-ment of these cell lines to the MC lineage (Fig. 5). Abroader distribution was detected for the IgE receptor+ chain. Both of these results are in good agreement withpreviously published investigations [18]. In addition, theexpression of the myeloid cell surface marker Mac-2 wasanalyzed (the IgE-binding protein). This marker was

found to be restricted in its expression to the myeloidcells WEHI-274.1, IC-21 and to MC lines including themucosal MC line RBL-1, and no mRNA for this proteinwas detected in any of the lymphoid lines tested.

2.4 Expression of MMCP-8 and other mast cellproteases in normal mouse tissues

To determine the expression patterns for the variousmast cell proteases in normal mouse tissues, variousorgans from BALB/c and NMRI mice were excised andtotal RNA was prepared for Northern blot analysis. Toincrease the sensitivity of the probes, longer cDNA frag-ments from the different protease clones were used.These probes were tested on dot blots for their specific-ity and all were found to be monospecific (data notshown). Total RNA was prepared from lungs, intestines,ears, spleen, thymus and liver and probed with frag-ments of MMCP-1, MMCP-5, MMCP-6, MMCP-7,MMCP-8 and carboxypeptidase A.

High expression of MMCP-1 and MMCP-2 was detectedin the intestine of BALB/c mice, while no signal wasobserved in intestinal samples from NMRI mice (Fig. 6and data not shown). Low and high levels of the CTMC

Eur. J. Immunol. 1998. 28: 1022–1033 Mouse mast cell protease 8 1027

Figure 7. Northern blot analysis of total RNA from intestinesand lungs of uninfected (Nml) and 4, 6, 8 and 10 weeks S.mansoni-infected BALB/c mice. In “Intestine” we show theresults from the analysis of intestines from two individualanimals. One additional experiment showing the up-regulation of MMCP-8 in lung of 8-week-infected mice isalso depicted in the right panel of the figure. A mouse § -actin probe was used as a reference gene to equalize theamounts of RNA loaded in each lane. Depending on strongcross-hybridization between the § -actin probe and the g -actin mRNA, the signal detected on the blot is the cyto-plasmic g -actin.

protease MMCP-5 and carboxypeptidase A were foundin the lungs and ears, respectively, of both strains ofmice. Expression of the MC tryptases, MMCP-6 and -7,was detected in the ears of both BALB/c and NMRI mice(Fig. 6 and data not shown). No detectable MMCP-8mRNA was found in the organs of normal BALB/c (Fig. 6)or NMRI mice (data not shown), suggesting very lowlevels of this protease in mature MC of normal tissues.

2.5 Expression of MMCP-8 and other MCproteases in intestines and lungs ofSchistosoma mansoni-infected animals

To study the expression of the mouse MC proteases dur-ing a parasitic infection, BALB/c mice were infected withthe trematode S. mansoni. Total RNA from intestines andlungs from 4, 6, 8, and 10 week-infected as well as fromuninfected controls were investigated for the expressionof MMCP-1, MMCP-2, MMCP-5, MMCP-6, MMCP-7,MMCP-8 and carboxypeptidase A (Fig. 7).

Our results show a marked up-regulation of both MMCP-1 and MMCP-2 expression in intestines 4 weeks afterinfection (Fig. 7). The levels of transcripts of these twoproteases declined by 6 weeks post infection andreached maximal levels by 8 weeks post infection (Fig. 7and data not shown). MMCP-5 and carboxypeptidase Afollowed a similar pattern of expression as the two MMC

proteases (Fig. 7). No hybridization signal for the twomajor MMC proteases, MMCP-1 and MMCP-2, wasdetected in the lungs at any point after infection (Fig. 7).Furthermore, transcripts for the MC tryptases MMCP-6and -7 were absent in both intestine and lungs ofschistosome-infected animals (data not shown). MMCP-8 was not detected in the intestines during the entireinfection period. However, its expression increasedmarkedly in the lungs of the same animals from 4 to 8weeks post infection. MMCP-5 and carboxypeptidase Awere detected simultaneously with MMCP-8 in lungs,although at lower levels (Fig. 7). These results were con-firmed by duplicate animals from two independentexperiments.

To rule out the possibility that cytotoxic T cells and NKcells were the source of MMCP-8, RNA form parasitizedintestines and lungs were hybridized with a granzyme Aprobe. Hybridization signals for this serine protease weredetected in intestines with similar levels at the differenttime points tested, and at low levels in lungs (Fig. 7). Theminor increase in granzyme A expression in both intes-tine and lung following parasite infection and the differ-ence in kinetics in this increase compared to the increaseof MMCP-1 and MMCP-2 in intestines makes it unlikelythat a T cell or NK cell population is the source of theMMCP-8 expressed in the lung.

3 Discussion

A new subfamily of murine MC proteases distinct fromboth the chymases and the tryptases has been identifiedthrough the present cloning of MMCP-8. In a recentpaper we reported the presence of three homologousproteases, denoted RMCP-8, -9 and -10 in the rat [14].Northern blot analysis of a panel of hematopoietic celllines showed that the MMCP-8 expression was confinedto the MC lineage, although low levels of expressionwere detected in the CTLL line. This latter findingtogether with the expression of granzyme B in certain(rare) passages of the RBL-1 cell line and the expressionof MC carboxypeptidase A in the mouse T helper cell lineEL4 can most likely be explained as an effect of theimmortalization of these cell lines (Fig. 5). However, wecannot at present rule out the possibility that certain sub-populations of normal T cells or NK cells express lowlevels of MMCP-8.

The substrate specificity of this novel mast cell proteaseis as yet unknown. However, based on sequence com-parison with other hematopoietic serine proteases, con-sidering the amino acid positions of importance forcleavage site selection in human proteinase 3 [19], it isevident that the binding pocket of MMCP-8 shows clear

1028 C. Lützelschwab et al. Eur. J. Immunol. 1998. 28: 1022–1033

similarities to the T cell granzyme B. By comparingMMCP-8 with mouse granzyme B, three out of the sixpositions of importance for cleavage selection in protei-nase 3 are identical in sequence (positions 195, 215 and226 in Fig. 2), two positions have conservative changes(positions 196 and 198 in Fig. 2), and one shows a mar-ked difference in the type of aa (position 218, arginine inMMCP-8 and glycine in granzyme B). This latter differ-ence indicates that MMCP-8 has an even higher positivecharge in the binding pocket compared to granzyme B,and thereby may have an even higher preference fornegatively charged aa in the binding pocket. It is there-fore likely that MMCP-8 also shows aspartate cleavagespecificity [19]. The members of the MMCP-8 subfamilyof serine proteases have an exceptionally high number ofpotential glycosylation sites, several of which are con-served between mouse and rat [14]. As can be seen fromFig. 2, the three glycosylation sites found in the C-terminal half of the protein are all conserved between thefour different members of this subfamily of serine prote-ases. However, no information is yet available on theactual amount of carbohydrate present in these prote-ases.

In several independent screenings of a human MC cDNAlibrary and of a human genomic library at low stringency,we have not yet been able to identify a human homo-logue of MMCP-8. Although not a proof, this suggeststhat such a homologue does not exist. A similar situationhas been observed for the mouse MC chymases MMCP-1 to MMCP-4, for which no human homologues havebeen identified. The only human chymase so far isolatedis the homologue of MMCP-5 [20].

A rapid increase in MC numbers is often observed duringparasitic worm infections [21–24]. To study the expres-sion of MMCP-8 during a parasitic infection and to relateits expression to other MC differentiation markers, weanalyzed the expression of the various MC proteases inintestine and lung during the course of a S. mansoniinfection in BALB/c mice. S. mansoni has been fre-quently used as a model to study the phenotype andfunction of MC and other inflammatory cells in the hostreaction against the invading parasite [25–29].

Based on our published data on the expression of the rathomologues of MMCP-8 in normal rat tissues [14], weexpected to see a dramatic increase in the expression ofMMCP-8 together with MMCP-1 and MMCP-2 in theintestine of the S. mansoni-infected mice, late duringinfection when eggs start to accumulate. However, nosuch increase was detected. In contrast, a dramaticincrease of MMCP-8 was observed in the lung ofinfected animals, and this at week 8 of infection. Nosimultaneous increase in the expression of MMCP-1 and

MMCP-2 was observed in this tissue and only a minorincrease in the CTMC proteases, MMCP-5 and carboxy-peptidase A was seen. MMCP-5 and carboxypeptidaseA are the first proteases to appear in immature MC,which may indicate that the MMCP-8-positive cells origi-nate from an expansion of immature MC in the lung andthat the cytokines of this particular microenvironmentpromote development of a novel subtype of MC, expre-ssing MMCP-5, carboxypeptidase A and MMCP-8.However, the late appearance of MMCP-8 in this tissueis still difficult to explain.

Eosinophils do not seem to be involved since no signs ofan increase in MMCP-8 levels was detected in the intes-tines of parasite-infected animals, where a massiveinfiltration of eosinophils often are observed.

An important question in MC biology today is how muco-sal and connective tissue MC are related and how theydevelop from different or common precursors during dif-ferentiation. Immature mast cells grown in vitro in thepresence of IL-3 (BMMC) were initially, due to theirsafranin-negative phenotype, considered to representmucosal MC or MMC-like cells. However, after a moredetailed analysis, based on their protease expression,they have been shown to have a phenotype which moreresembles immature CTMC [6, 11, 30] (Table 1). Additionof recombinant IL-10 to the BMMC cultures induces,within 24–48 h, the expression of high levels of bothMMCP-1 and MMCP-2 [30]. This shows that the imma-ture, MMCP-5 and carboxypeptidase A-positive MC inthe culture probably develop into relatively maturemucosal MC. In addition, a recent report has describedthe appearance of various MC subtypes during Trichi-nella spiralis infections of BALB/c mice, indicating thatimmature MC expressing MMCP-5 and carboxypepti-dase A also can develop into mature mucosal MC in vivo[31]. The increase of MMCP-5 and carboxypeptidase A,together with MMCP-1 and MMCP-2 in the intestines ofschistosome-infected animals, and the lack of a simulta-neous increase in the CTMC-specific MC tryptasesMMCP-6 and MMCP-7 adds further evidence to this pic-ture of a relatively close interrelation between MMC andCTMC during differentiation (Fig. 7). A MMCP-5- andcarboxypeptidase A-positive population of immature MCmay thereby be the precursor of both mature CTMCexpressing MMCP-4, -5, -6, -7 and carboxypeptidase Aand of mature MMC expressing only MMCP-1 andMMCP-2 (Table 1). How the MMCP-8-positive MCpopulation is related to these two classical MC popula-tions in the mouse is still an open question. However, it isevident that MMCP-8 cannot be considered a regularmucosal MC protease in the mouse. It is here interestingto note that there is a clear species difference betweenmouse and rat in expression pattern of MMCP-8 and its

Eur. J. Immunol. 1998. 28: 1022–1033 Mouse mast cell protease 8 1029

rat homologues, RMCP-8, -9 and -10. In the rat, RMCP-8, -9 and -10 are preferentially expressed in mucosal MCand only very low levels have been detected in normalserosal MC isolated from the peritoneal cavity ofSprague Dawley rats [14]. We have recently also shownthat the RMCP-8 family increases dramatically, togetherwith RMCP-2, in intestinal tissues of Nippostrongylusbrasiliensis-infected rats, which shows that the membersof the RMCP-8 family are true mucosal MC proteases inthe rat (Lützelschwab et al., unpublished observations).

Most of the present data concerning the expression pat-terns for the various MC proteases and other hemato-poietic differentiation markers in various rodent cell linesand tissues, presented in Fig. 5, are in agreement withresults reported in previous studies [4, 5, 7, 9, 11, 13, 16,18, 32–36]. However, there are several minor discrepan-cies. MMCP-7 was not only detected in mouse MC butalso in normal purified peritoneal macrophages. Theexpression levels and the distribution in various macro-phage subpopulations of MMCP-7 is, however, not yetknown. The detection of MMCP-7 in mouse macro-phages is interesting in the light of the detection andcharacterization of the human counterpart to MMCP-7 intwo human monocytic cell lines [10].

In summary, the cloning of MMCP-8 has led to the identi-fication of a third subfamily of mouse MC proteases, dis-tinct from both the classical chymases and tryptases. Inaddition, the massive increase in expression of MMCP-8in lung but not in intestines of S. mansoni-infected ani-mals indicates the existence of either a novel MMCP-8single-positive MC subpopulation or alternatively a novelMMCP-5-, carboxypeptidase A- and MMCP-8-positivesubpopulation of MC. This needs, however, to be con-firmed with immunohistochemical techniques to detectthe potential novel MC population at the single-cell level.A clear species difference in the expression of MMCP-8and its rat homologous has also been identified. In addi-tion, the detection of MMCP-7 mRNA in normal perito-neal macrophages indicates that MMCP-7 is alsoexpressed in certain macrophage populations andthereby cannot be considered an entirely MC-specificserine protease.

4 Materials and methods

4.1 Cloning and nucleotide sequence analysis of acDNA for MMCP-8

The cloning of MMCP-8 was performed according to previ-ously described protocols [13, 14]. The nucleotide sequencepresented in this study was established by sequencing bothstrands of the DNA insert.

4.2 Amino acid sequence alignments

Amino acid sequence alignment was performed as previ-ously described [13]. All references for the protease sequen-ces included in this study have been listed in two recentlypublished articles [13, 14].

4.3 Genomic Southern blot analysis

Genomic DNA was prepared from BALB/c mice as previ-ously described [37]. Ten micrograms of genomic DNA wasdigested to completion with Eco RI, Hind III, Bam HI andXba I restriction enzymes and was fractionated through0.7 % agarose gels. The DNA was transferred to N+ Nylonmembranes (Amersham, Amersham, GB) by blotting in0.4 M NaOH. Blotting with a “VacuBlot” unit (Pharmacia,Uppsala, Sweden) was completed in 3 h and the DNA wascross-linked to the membrane by UV irradiation (GS GeneLinker, Bio-Rad Laboratories, Hercules, CA). Hybridizationswere carried out overnight in a solution containing 0.25 MNa2HPO4 (pH 7.2), 7 % SDS, at 65 °C, after which the filterswere washed first at 65 °C, in 20 mM Na2HPO4 (pH 7.2), 5 %SDS, and then in 20 mM Na2HPO4 (pH 7.2), 1 % SDS, for1 h. The filters were hybridized with a 32P random-labeledprobe consisting of a purified cDNA fragment from themouse MMCP-8 clone.

4.4 Cell lines, tissues and in vitro cultures of BMMC

The following mouse and rat cell lines were used in thisstudy: M1D, a mouse early myeloid cell line [38]; WEHI-274.1, a mouse monocytic cell line [39]; IC-21, a mousemacrophage cell line [40]; J558, a mouse myeloma cell line[41]; EL4, a mouse T cell line [42]; RBL-1, a rat mucosal MCline [43]; MTC, a mouse mastocytoma (CTMC-like MCtumor) maintained as a solid tumor in (Leaden × A1) F1 miceby intramuscular transplantation in the hind legs every10–12 days [44]; J774, a mouse monocyte/macrophage cellline [45] and M-CTLL, a mouse cytotoxic T cell line [46].BMMC were obtained according to a previously describedprotocol [13].

Macrophages were isolated as previously described [13].The original cell fractions were analyzed for the presence ofMC. No MC were detected among the interface cells, whilethe original cells from the peritoneal lavage containedapproximately 2 % MC.

Lungs, intestine, ears, spleen, thymus, and liver fromnormal 5–8-week-old female BALB/c and NMRI mice,and lungs and small intestine from 4, 6, 8 and 10 week S.mansoni-infected BALB/c mice were used to study thedistribution and kinetics of appearance of a panel ofmouse MC proteases. S. mansoni (Puerto Rican strain)cercariae were obtained from the Laboratory of Parasito-logy, SMI (Stockholm, Sweden). Seven to eight-week-

1030 C. Lützelschwab et al. Eur. J. Immunol. 1998. 28: 1022–1033

old female BALB/c mice (Alab, B&K Universal AB, Sol-lentuna, Sweden) were anesthetized with Pentobarbital(0.1 ml i.p./10 g body weight of a 1:7 dilution in NaCl ofa 60 mg/ml Mebumal vet stock) and then percutaneouslyinfected via exposure of the abdominal skin to 50 cer-cariae. At different time points after schistosome infec-tion, organs (lungs, intestines) from individual mice wereremoved, immediately frozen in liquid nitrogen, andstored at −70 °C until further processed. Intestines wererinsed/flushed with sterile PBS before being frozen.Uninfected littermates were included as controls.

4.5 Northern blot analysis

Northern blot analysis was performed as previouslydescribed [13, 14]. The different probes used in the Northernblot analysis are listed below. The rat high-affinity IgE recep-tor § chain probe was isolated from the RBL-1 cell line byPCR amplification. The probe for the high-affinity IgE recep-tor g chain was a 60-nucleotide oligonucleotide directedagainst the 5' end of mouse and rat IgE receptor g chainmRNA. The high-affinity IgE receptor + chain probe was anearly complete cDNA for the mouse IgE receptor + chainisolated from a mouse MC cDNA library. The MMCP-1,MMCP-4 and MMCP-5 probes were purified fragmentsderived from the 3' end of the corresponding cDNA, andthese probes have been shown not to cross-hybridize withother MC chymase mRNA [4]. The MMCP-2 probe was acDNA fragment from the 3' end of a MMCP-2 cDNA whichwe isolated by PCR amplification from the MTC cell line.This probe has also been tested for cross-hybridization withother chymase cDNA and been found to be monospecificfor MMCP-2. The probe for MMCP-6 was a fragment ofapproximately 300 bp from the 3' non-coding region of thefull-length cDNA [9]. The MMCP-7 probe is an approxi-mately 300-bp fragment from the 3' end of a full-lengthclone for MMCP-7, isolated from a mouse MC cDNA library.The MC carboxypeptidase A probe was a partial cDNAclone isolated by PCR amplification from mRNA from theCTMC-like tumor MTC. The probe for MMCP-8 used wasthe full-length cDNA. The probes for mouse cathepsin G,mouse proteinase 3 (myeloblastine) and RNK-A (granzymeB) were the His-Ser region probes isolated by PCR ampli-fication with the degenerate primers described in Aveskoghet al. [13]. The probe for the heparin core protein was anearly full-length cDNA isolated from a mouse mastocytoma[15, 34]. The probe for Mac-2 was a nearly full-length cDNAfor rat Mac-2 isolated from a RBL-1 cDNA library. Theglyceraldehyde-3-phosphate-dehydrogenase (GAPDH)probe was a fragment from the plasmid pHcGAP3 contain-ing cDNA for GAPDH (a gift from Dr. R. Wu, Cornell Univer-sity, NY); the actin probe was a 1.6-kb Pst 1 fragment iso-lated from the clone 91 [47] (a gift from Dr. M. Buckingham).To increase the sensitivity for the total RNA Northern blotanalysis of normal and parasitized animals, the probes for

MMCP-1, MMCP-2, MMCP-6 and MMCP-7 were increasedin size (460 bp, 250 bp, 300 bp, 260 bp, respectively).

Acknowledgements: We would like to thank Dr. GunnarPejler and Dr. Allen Cheever for valuable discussions and forlinguistic revision of the manuscript. This investigation wassupported by grants from the Swedish Natural SciencesResearch Council (grant no B-AA/BU 09400-309). M. C. K.was supported by grants from the Swedish Agency forResearch Cooperation with Developing Countries (SAREC)and the Swedish Medical Research Council.

5 References

1 Enerbäck, L., Mast cells in rat gastrointestinal mucosa1. Effects of fixation. Acta Pathol. Microbiol. Scand.1966. 66: 289–302.

2 Enerbäck, L., Mast cells in rat gastrointestinal mucosa2. Dye-binding and metachromatic properties. ActaPathol. Microbiol. Scand. 1966. 66: 303–312.

3 Stevens, R. L., Katz, H. R., Seldin, D. C. and Austen,K. F. (Eds.), Biochemical characteristics distinguish sub-classes of mammalian mast cells. Raven Press, NewYork 1986, pp. 183–203.

4 Huang, R. Y., Blom, T. and Hellman, L., Cloning andstructural analysis of MMCP-1, MMCP-4 and MMCP-5,three mouse mast cell-specific serine proteases. Eur. J.Immunol. 1991. 21: 1611–1621.

5 Serafin, W. E., Reynolds, D. S., Rogelj, S., Lane, W. S.,Conder, G. A., Johnson, S. S., Austen, K. F. and Ste-vens, R. L., Identification and molecular cloning of anovel mouse mucosal mast cell serine protease. J. Biol.Chem. 1990. 265: 423–429.

6 Serafin, W. E., Sullivan, T. P., Conder, G. A., Ebhrahimi,A., Marcham, P., Johnson, S. S., Austen, K. F. andReynolds, D. S., Cloning of the cDNA and Gene forMouse Cell Protease 4. J. Biol. Chem. 1991. 266:1934–1941.

7 McNeil, H. P., Austen, K. F., Somerville, L. L., Gurish,M. F. and Stevens, R. L., Molecular cloning of themouse mast cell protease-5 gene. A novel secretorygranule protease expressed early in the differentiation ofserosal mast cells. J. Biol. Chem. 1991. 266:20316–20322.

8 Reynolds, D. S., Gurley, D. S., Austen, K. F. and Sera-fin, W. E., Cloning of the cDNA and Gene of the MouseMast Cell Protease-6. J. Biol. Chem. 1991. 266:3847–3853.

9 Huang, R. and Hellman, L., Cloning and structural anal-ysis of a novel mouse mast cell specific protease, theMMCP-6: a connective tissue mast cell specific trypt-

Eur. J. Immunol. 1998. 28: 1022–1033 Mouse mast cell protease 8 1031

ase. Abstracts of the XIVth International Congress ofAllergology and Clinical Immunology 1991, Kyoto,Japan, 1991. Abstract no. 125.

10 Huang, R., Abrink, M., Gobl, A. E., Nilsson, G., Aves-kogh, M., Larsson, L. G., Nilsson, K. and Hellman, L.,Expression of a mast cell tryptase in the human mono-cytic cell lines U-937 and Mono Mac 6. Scand. J. Immu-nol. 1993. 38: 359–367.

11 McNeil, H. P., Reynolds, D. S., Schiller, V., Ghildyal, N.,Gurley, D. S., Austen, K. F. and Stevens, R. L., Isola-tion, characterization, and transcription of the geneencoding mouse mast cell protease 7. Proc. Natl. Acad.Sci. USA 1992. 89: 11174–11178.

12 Huang, R. and Hellman, L., Genes for mast-cell serineproteases and their molecular evolution. Immunogenet-ics 1994. 40: 397–414.

13 Aveskogh, M., Lützelschwab, C., Huang, M. R. andHellman, L., Characterization of cDNA clones encodingmouse proteinase 3 (myeloblastine) and cathepsin G.Immunogenetics 1997. 46: 181–191.

14 Lützelschwab, C., Pejler, G., Aveskogh, M. and Hell-man, L., Secretory granule proteases in rat mast cells.Cloning of 10 different serine proteases and a carboxy-peptidase A from various rat mast cell populations. J.Exp. Med. 1997. 185: 13–29.

15 Kjellen, L., Pettersson, I., Lillhager, P., Steen, M.-L.,Pettersson, U., Lehtonen, P., Karlsson, T., Rouslahti,E. and Hellman, L., Primary structure of a mouse mas-tocytoma proteoglycan core protein. Biochem. J. 1989.263: 105–113.

16 Stevens, R. L., Friend, D. S., McNeil, H. P., Schiller, V.,Ghildyal, N. and Austen, K. F., Strain-specific andtissue-specific expression of mouse mast cell secretorygranule proteases. Proc. Natl. Acad. Sci. USA 1994. 91:128–132.

17 Reynolds, D. S., Stevens, R. L., Gurley, D. S., Lane, W.S., Austen, K. F. and Serafin, W. E., Isolation and mole-cular cloning of mast cell carboxypeptidase A. J. Biol.Chem. 1989. 264: 20094–20099.

18 Ra, C., Jouvin, M.-H. E., Blank, U. and Kinet, J.-P., Amacrophage Fc + receptor and the mast cell receptor forIgE share an identical subunit. Nature 1989. 341:752–754.

19 Brubaker, M. J., Groutas, W. C., Hoidal, J. R. and Rao,N. V., Human neutrophil proteinase 3: mapping of the sub-strate binding site using peptidyl thiobenzyl esters. Bio-chem. Biophys. Res. Commun. 1992. 188: 1318–1324.

20 Caughey, G. H., Zerweck, E. H. and Vanderslice, P.,Structure, chromosomal assignment, and deducedamino acid sequence of a human gene for mast cellchymase. J. Biol. Chem. 1991. 266: 12956–12963.

21 Gerken, S. E., Mota, S. T. and Vaz, N. M., Evidence forthe participation of mast cells in the innate resistance ofmice to Schistosoma mansoni: effects on in vivo treat-ment with the ionophore 48–80. Braz. J. Med. Biol. Res.1990. 23: 559–565.

22 Amerik, A., Yarovoi, S. V., Grigorenko, V. G. and Anto-nov, V. K., Identification, sequence analysis and charac-terization of cDNA clones encoding two granzyme-likeserine proteinases from rat duodenum. FEBS Lett. 1993.324: 226–230.

23 Mitchell, L. A., Wescott, R. B. and Perryman, L. E.,Kinetics of expulsion of the nematode, Nippostrongylusbrasiliensis, in mast-cell deficient W/WV mice. ParasiteImmunol. 1983. 5: 1–12.

24 Finkelman, F. D., Shea-Donohue, T., Goldhill, J., Sulli-van, C. A., Morriz, S. C., Madden, K. B., Gause, W. C.and Urban, J. F. J., Cytokine regulation of host defenceagainst parasitic gastrointestinal nematodes. Lessonsfrom studies with rodent models. Annu. Rev. Immunol.1997. 15: 505–533.

25 Bentley, A. G., Carlisle, A. S. and Phillips, S. M., Ultra-structural analysis of the cellular response to Schisto-soma mansoni: initial and challenge infections in the rat.Am. J. Trop. Med. Hyg. 1981. 30: 102–112.

26 Gerken, S. E., Mota-Santos, T. A., Vaz, N. M., Correa-Oliveira, R., Dias-da-Silva, W. and Gazzinelli, G.,Recovery of schistosomula of Schistosoma mansonifrom mouse skin: involvement of mast cells and vasoac-tive amines. Braz. J. Med. Biol. Res. 1984. 17: 301–307.

27 Henderson, W. R., Chi, E. Y., Jong, E. C. and Kleba-noff, S. J., Mast cell-mediated toxicity to schistosomulaof Schistosoma mansoni: potentiation by exogenousperoxidase. J. Immunol. 1986. 137: 2695–2699.

28 Miller, H. R., Newlands, G. F., McKellar, A., Inglis, L.,Coulson, P. S. and Wilson, R. A., Hepatic recruitment ofmast cells occurs in rats but not mice in infected withSchistosoma mansoni. Parasite Immunol. 1994. 16:145–155.

29 Cohn, R. G., Williams, M., Sher, A. and Caulfield, J. P.,Schistosoma mansoni: characterization of an Fc epsilonR+ population of granule-containing splenocytes iso-lated from infected mice. Exp. Parasitol. 1995. 80:339–341.

30 Ghildyal, N., McNeil, H. P., Stechschulte, S., Austen,K. F., Silberstein, D., Gurish, M. F., Somerville, L. L.and Stevens, R. L., IL-10 induces transcription of thegene for mouse mast cell protease-1, a serine proteasepreferentially expressed in mucosal mast cells of Trichi-nella spiralis-infected mice. J. Immunol. 1992. 149:2123–2129.

1032 C. Lützelschwab et al. Eur. J. Immunol. 1998. 28: 1022–1033

31 Friend, D. S., Ghildyal, N., Austen, K. F., Gurish, M. F.,Matsumoto, R. and Stevens, R. L., Mast cells thatreside at different locations in the jejunum of miceinfected with Trichinella spiralis exhibit sequentialchanges in their granule ultrastructure and chymasephenotype. J. Cell. Biol. 1996. 135: 279–290.

32 Eklund, K. K., Ghildyal, N., Austen, K. F., Friend, D. S.,Schiller, V. and Stevens, R. L., Mouse bone marrow-derived mast cells (mBMMC) obtained in vitro from micethat are mast cell-deficient in vivo express the samepanel of granule proteases as mBMMC and serosal mastcells from their normal littermates. J. Exp. Med. 1994.180: 67–73.

33 Frigeri, L. G. and Liu, F. T., Surface expression of func-tional IgE binding protein, an endogenous lectin, onmast cells and macrophages. J. Immunol. 1992. 148:861–867.

34 Angerth, T., Huang, R., Aveskogh, M., Pettersson, I.,Kjellen, L. and Hellman, L., Cloning and structural anal-ysis of a gene encoding a mouse mastocytoma proteo-glycan core protein; analysis of its evolutionary relationto three cross hybridizing regions in the mouse genome.Gene 1990. 93: 235–240.

35 Heusel, J. W., Hanson, R. D., Silverman, G. A. and Ley,T. J., Structure and expression of a cluster of humanhematopoietic serine protease genes found on chromo-some 14q11.2. J. Biol. Chem. 1991. 266: 6152–6158.

36 Heusel, J. W., Scarpati, E. M., Jenkins, N. A., Gilbert,D. J., Copeland, N. G., Shapiro, S. D. and Ley, T. J.,Molecular cloning, chromosomal location, and tissue-specific expression of the murine cathepsin G gene.Blood 1993. 81: 1614–1623.

37 Sambrock, J., Fritsch, E. F. and Maniatis, T. (Eds.),Molecular cloning, a laboratory manual. Cold Spring Har-bor Laboratory Press, Cold Spring Harbor, New York1989, pp. 9.14–9.21.

38 Ichikawa, Y., Differentiation of a cell line of myeloid leu-kemia. J. Cell. Physiol. 1969. 74: 223–234.

39 Warner, L. L., Burchiel, S. W., Walker, S. B., Richey, J.B., Leary, J. F. and McGlaughlin, S., Characterization ofmurine B cell lymphomas and macrophage tumors byflow microfluorometry and functional assays. In W. Terryand Y. Yamamura (eds.), Immunobiology and Immuno-lotherapy of Cancer. Elsevier North Holland PublishingCo., New York 1979, pp 243–261.

40 Mauel, J. and Defendi, V., Infection and transformationof mouse peritoneal macrophages by simian virus 40. J.Exp. Med. 1971. 134: 335–350.

41 Lundblad, A., Steller, R., Kabat, E. A., Hirst, J. W., Wei-gert, M. G. and Cohn, M., Immunochemical studies onmouse myeloma proteins with specificity for dextran orfor levan. Immunochemistry 1972. 9: 535–546.

42 Farrar, J. J., Fuller-Farrar, J., Simon, P. L., Hilifiker, M.L., Stadler, B. M. and Farrar, W. L., Thymoma produc-tion of T cell growth factor (interleukin 2). J. Immunol.1980. 125: 2555–2559.

43 Eccleston, E., Leonard, B. J., Lowe, J. S. and Welford,H. J., Basophilic leukaemia in the albino rat and ademonstration of the basopoietin. Nature 1973. 244:73–76.

44 Furth, J., Hagen, P. and Hirch, E. I., Transplantablemastocytoma in the mouse containing histamine, hepa-rin, 5-hydroxytryptamine. Proc. Soc. Exp. Biol. Med.1957. 95: 824–828.

45 Ralph, P., Prichard, J. and Cohn, M., Reticulum cellsarcoma: an effector cell in antibody-dependent cell-mediated immunity. J. Immunol. 1975. 114: 898–905.

46 Gillis, S. and Smith, K. A., Long-term culture of tumor-specific cytotoxic T cells. Nature 1977. 268: 154–156.

47 Minty, A. J., Caravatti, M., Robert, B., Cohen, A., Dau-bas, P., Weydert, A., Gros, F. and Buckingham, M. E.,Mouse actin messenger RNAs. Construction and char-acterization of a recombinant plasmid molecule contain-ing a complementary DNA transcript of mouse alpha-actin mRNA. J. Biol. Chem. 1981. 256: 1008–1014.

Present addresses: M. R. Huang, Department of ClinicalImmunology, University Hospital, S-75185 Uppsala, Swe-den; M. C. Kullberg, Immunobiology Section, Laboratory ofParasitic Diseases, National Institute of Allergy and Infec-tions Diseases, National Institute of Health, Bethesda, MD20892, USA.

Correspondence: Lars Hellman, Department of MedicalImmunology and Microbiology, University of Uppsala,Biomedical Center, Box 582, S-751 23 Uppsala, SwedenFax: +46-18-55 75 79; e-mail: Lars.Hellman@immun.uu.se

Eur. J. Immunol. 1998. 28: 1022–1033 Mouse mast cell protease 8 1033