(Received for publication, September 12, 1996, and in revised form, November 6, 1996)
From the Departments of Pediatrics, ¶ Medicine,
and ** Surgery, Medical College of Virginia/Virginia Commonwealth
University, Richmond, Virginia 23298-0529 and the
Department of
Dermatology, University of Pennsylvania,
Philadelphia, Pennsylvania 19104
The ability of human mast cell chymase and
tryptase to process procollagen was examined. Purified human intestinal
smooth muscle cell procollagen was incubated with human mast cell
tryptase or human mast cell chymase. Purified chymase, but not
tryptase, exhibited procollagen proteinase activity in the presence of
EDTA. Addition of purified porcine heparin over a range of 0.1-100
µg/ml did not affect either the rate or the products of
procollagen chymase cleavage. The cleavage site of chymase on the
pro-1(I) collagen carboxyl terminus was found to be in the
propeptide region at Leu-1248-Ser-1249. Cleavage at this site suggested
that the collagen products would form fibrils and confirmed the
production of a unique carboxyl-terminal propeptide. Turbidometric
fibril formation assay demonstrated de novo formation of
chymase-generated collagen fibrils with characteristic lag, growth, and
plateau phases. When observed by dark field microscopy, these fibrils were similar to fibrils formed by the action of procollagen
proteinases. Thus, mast cell chymase, but not tryptase, exhibits
procollagen peptidase-like activity as evidenced by its ability to
process procollagen to fibril-forming collagen with concurrent
formation of a unique carboxyl-terminal propeptide. These data
demonstrate that mast cell chymase has a potential role in the
regulation of collagen biosynthesis and in the pathogenesis of
fibrosis.
Mast cells are abundant in connective tissues of skin, lung, and intestine (1, 2). Increased numbers (or increased activity) of mast cells are associated with fibrotic disorders such as scleroderma, pulmonary fibrosis, and Crohn's disease (3-5). The precise role for mast cells in connective tissue biology and fibrosis is not known. One potential mechanism for mast cell involvement in these processes is through the production of unique proteolytic enzymes that may act on matrix proteins. Mast cell enzymes include the neutral proteases tryptase, chymase, cathepsin G, and a zinc-dependent carboxypeptidase (6). Although chymase and tryptase cleave fibronectin (7-9) and chymase cleaves vitronectin (10), no direct effect of these enzymes on procollagen or collagen has been demonstrated.
Because procollagen (the precursor of collagen) requires proteolytic cleavage in the extracellular space prior to normal collagen fibril formation, we have speculated that newly secreted procollagen may be a substrate for mast cell proteolytic enzymes. The smooth muscle cells of the intestinal muscularis are a major contributor of collagen to the extracellular matrix in normal and fibrotic human intestine (11). Therefore, in vitro studies were performed to test the above hypothesis utilizing procollagen secreted by human intestinal smooth muscle cells.
Human intestinal smooth muscle cells (12) were grown to confluence in Dulbecco's modified Eagle's medium containing 200 units/ml penicillin, 0.2 mg/ml streptomycin, 50 units/ml Nystatin, and 10% fetal bovine serum. Cell monolayers were then rinsed with serum-free Dulbecco's modified Eagle's medium and incubated at 37 °C for 24 h in serum-free Dulbecco's modified Eagle's medium containing 0.1 µM/ml ascorbic acid and 10 µCi/ml tritiated 15-amino acid mixture (Sigma). The culture medium was harvested and filtered through a Nalgene 0.2-µM bottle-top filter (Nalge Co., Rochester, NY) and received 10% by volume of 10× stock filtration buffer containing 250 mM EDTA, 0.2% NaN3, 0.5 M Tris, pH 7.4. Protease inhibitors were added (25 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, and 10 mM N-ethylmaleimide), and the solution was stirred overnight at 4 °C. The pH of the medium solution was adjusted to 7.4, and fibronectin was removed by gelatin-Sepharose chromatography (13). The effluent was adjusted to 0.4 M NaCl and concentrated by pressure ultrafiltration using an Amicon S1Y100 spiral wound membrane cartridge (Amicon, Inc., Beverly, MA). Purification of type I procollagen was performed by DEAE-cellulose chromatography (14, 15). Fractions containing type I procollagen were pooled, dialyzed twice against 500 volumes of 0.4 M NaCl, 0.02% NaN3, 0.1 M Tris, pH 7.4, storage buffer, and concentrated by pressure ultrafiltration with Amicon Centriprep 100 filters. The final type I procollagen concentrations were determined by quantitative densitometric analysis using a Molecular Dynamics PDSI 486 laser densitometer and hydroxyproline assay, assuming 10.11% hydroxyproline by weight (15). Enriched culture medium harvested from 30 × 150-mm culture plates (~6 million cells/plate) yielded approximately 1 mg of purified type I procollagen.
Purified Chymase and TryptaseHuman mast cell chymase was purified from dermal mast cells as described previously (16). Mast cell tryptase was purified from human lung by immunoaffinity chromatography as described previously (17).
Type I Procollagen N-proteinase1 and C-proteinaseType I procollagen amino-terminal propeptidase and carboxyl-terminal propeptidase (purified to functional purity from chick embryo leg tendons) were gifts from Dr. Yoshio Hojima (18, 19).
Procollagen Cleavage AssayChymase and mast cell lysate
were each assayed for procollagen proteinase activity in 0.01% (v/v)
Brij 35 or 2 M NaCl, 0.05 M Tris, pH 7.4, 0.02% (w/v) NaN3, and 25 mM EDTA at 35 °C,
a modified form of the assay published by Hojima et al. (18)
for the measurement of the specific activity of procollagen
carboxyl-terminal proteinase. Tryptase was preincubated on ice with
heparin:tryptase (5:1, w/w) before being added to assay solutions
containing 0.2 M NaCl. N-proteinase and C-proteinase were
incubated in procollagen cleavage assays containing 0.1 mM
CaCl2. Reactions were stopped by the addition of an equal
volume of non-reducing sample buffer containing 62.5 mM
Tris, pH 6.8, 10% (v/v) glycerol, 10% (w/v) SDS, and 0.001% (v/v)
bromphenol blue. In some cases 2% (v/v) 2-mercaptoethanol was added to
reduce samples. All samples were heated to 100 °C for 3 min and then
subjected to polyacrylamide gel electrophoresis using 4% stacking and
6 or 8% separating gels. All gels were fixed for 20 min with a mixture
of 10% (v/v) acetic acid and 25% (v/v) 2-propanol, incubated in 1 M sodium salicylate, dried, and exposed to Amersham
Hyperfilm-MP (Amersham International, Buckinghamshire, United Kingdom)
at 70 °C or stained with Coomassie Blue (20, 21). The protein
bands on the autoradiographs and Coomassie-stained gels were
quantitated using a Molecular Dynamics PDSI 486 laser densitometer.
Molecular weights were determined using FragmeNT analysis software
(Molecular Dynamics, Inc., Sunnyvale, CA).
By convention, the specific activity of
chymase is 2.7 µmol of product min1/nmol of chymase,
measured by cleavage of the synthetic substrate succinyl-Ala-Ala-Pro-Phe-p-nitroanilide (1 mM)
in 0.45 M Tris, pH 8.0, containing 1.8 mM NaCl,
and 0.9% Me2SO, 1.0 mM
succinyl-Ala-Ala-Pro-Phe-nitroanilide (
410
p-nitroaniline = 8800 M
1 cm
1) (16). Chymase used in
these studies had a specific activity of 2.7 µmol of product
min
1/nmol of chymase and was used at a concentration of
0.24 µg/ml, a concentration attainable in tissues with mast cell
degranulation (22).
Aliquots of human chymase were incubated for 20 min with serial concentrations of purified porcine heparin in siliconized microcentrifuge tubes on ice. Purified human type I procollagen was then added to initiate the reactions, which were carried out in 0.15 M NaCl assay solutions at 35 °C for 2 and 4 h. Reactions were quenched by addition of reducing sample buffer. Samples were analyzed by polyacrylamide gel electrophoresis and densitometry as above.
Amino Acid Sequencing of Reaction ProductsChymase (50 pmol) was incubated for 15 min with 50 µl of Affi-Gel heparin (Bio-Rad) in 0.12 M NaCl, 0.05 M Tris, 0.01% NaN3, pH 7.3, at 25 °C. Using these conditions, chymase bound tightly to Affi-Gel heparin (23). Bound chymase was then incubated for 24 h at 35 °C with 400 µg of purified type I procollagen in 0.12 M NaCl, 0.05 M Tris, 0.01% NaN3, pH 7.3. After 24 h, chymase bound to Affi-Gel was removed from the reaction mixture by centrifugation at 1,000 × g for 1 min. The reaction mixture was then prepared for gel electrophoresis under non-reducing conditions as described above. Cleavage products were purified by slab gel electrophoresis on a 9% polyacrylamide gel and blotted to a polyvinylidene difluoride membrane (Bio-Rad). The polyvinylidene difluoride membrane was stained with a solution of 40% methanol, 60% water, 0.025% Coomassie Blue and protein bands were visualized by destaining with a 50% methanol, 50% water solution. Product bands were labeled, and polyvinylidene difluoride membranes were shipped for commercial sequencing. Amino-terminal sequencing was performed by automated Edman degradation using an ABI 470A gas phase sequencer and an HP GS1000 HPLC (Protein Structure Laboratory, Davis, CA).
Turbidometric Fibril Formation AssayStock procollagen (150 µg/ml) was twice dialyzed against 500 volumes of physiologic fibril formation buffer, pH 7.4, containing 20 mM NaHCO3, 110 mM NaCl, 5.4 mM KCl, 1.8 mM CaCl2, 0.81 mM MgSO4, 1.03 mM NaH2PO4, and 0.04% (w/v) NaN3. Procollagen solutions were then preheated in a temperature-controlled cuvette, and the reactions were initiated by the addition of chymase, charged with 10% CO2, 90% air, and sealed with a greased rubber stopper. The turbidity of the reactions was monitored at 313 nm using a Shimadzu UV160U spectrophotometer as described previously (24).
Fibril Formation for Dark Field MicroscopyCollagen fibril formation was determined by de novo fibril formation assay (25). 150 µg/ml procollagen in the fibril formation buffer described above was incubated with either 0.24 µg/ml chymase or 3 units each of N- and C-proteinases in siliconized microcentrifuge tubes. Tubes were charged with 10% CO2, 90% air, sealed, and incubated for 24 h at 34 °C. 300 µl of each solution was also incubated in the sealed well of a Lab-Tek chamber slide (Nunc, Inc., Naperville, IL) (25). Fibril formation was monitored with a Zeiss dark field microscope, and photographs were taken with an Olympus PM-A10 optical attachment.
Incubation of
purified human mast cell chymase with type I procollagen resulted in
the time-dependent cleavage of procollagen to collagen-size
chains. Four discrete intermediates of approximate molecular masses of
162, 143, 130, and 117 kDa were formed during the incubation (Fig.
1, 2 and 4 h). Proteins corresponding in size to
1(I)- and
2(I)-collagen chains were not further degraded under
the conditions employed (Fig. 1, 48 h)(9). The intermediates formed by chymase cleavage of procollagen were similar in size to
intermediates formed by cleavage of procollagen by the specific amino-terminal and carboxyl-terminal propeptidases (Fig. 1, lanes NP and CP). This result demonstrated that chymase
cleaved procollagen in both the N- and C-propeptide regions, producing
collagen products similar to those produced by the combined activity of
amino- and carboxyl-terminal propeptidases (Fig. 1, lane
N&C). Type I procollagen substrate was not processed in the
absence of enzyme (Fig. 1, lane C). Chymase was determined
to cleave 0.38 µg of procollagen/pmol of chymase/h at 35 °C in a
100-µl assay (assayed for quantitative cleavage using a modified form
of a carboxyl-terminal proteinase assay described previously (21)).
Although this number gives some indication of the activity of chymase
relative to type I procollagen proteinases, it cannot be used as a
direct comparison against N- or C-proteinase activity because chymase
appears to cleave in both the amino- and the carboxyl-terminal regions.
It is also important to note that the type I procollagen concentrations used in the time-course studies presented in Fig. 1 were an order of
magnitude greater than that used in the published assay (21). This
accounts for the nearly 5-fold increase in the rate of procollagen cleavage demonstrated in Fig. 1 compared to the cleavage by
C-proteinase assay.
Propeptide Products Produced by Chymase Cleavage of Procollagen
Cleavage of type I procollagen by C-proteinase
generated a heterotrimer of two 1 carboxy-propeptides
disulfide-linked to one
2 carboxy-propeptide (Fig. 2,
lane N&C-PICP). The molecular mass of the PICP
molecule was subsequently determined to be 96 kDa by FragmeNT analysis,
which corresponds to previous reports (18). In contrast, cleavage of
procollagen by mast cell chymase produced two very disparate propeptide
chains, one of 67-kDa and another of 25 kDa molecular mass (Fig. 2,
lane Chy).
Carboxyl-terminal Chymase Cleavage Site in Procollagen
The
precise cleavage site of chymase in the carboxyl terminus of the type I
procollagen molecule was determined by amino acid sequence analysis of
the 67- and 25-kDa peptide products (Fig. 2). The analyses yielded an
identical sequence that was found in the 1(I) C-propeptide region
(Fig. 3). The analysis demonstrated the
carboxyl-terminal chymase cleavage site on the
1(I) chain to be
between Leu and Ser at positions 1248-1249 (Fig. 3). Leu in the P1
position and Ser in the P2 position are compatible with the known
preferences of chymase for its substrates (22). The chymase cleavage
site, therefore, is 20 amino acids carboxyl to the C-proteinase
cleavage site that is between residues 1228 and 1229. The only other
potential chymase-compatible cleavage sites more proximal to this one
would lie within less than 10 residues of the triple helix region,
where steric problems would be expected to hinder chymase
approximation. These data strongly suggest that 1248-1249 is the
carboxyl-terminal chymase cleavage site in the
1 procollagen chain
and that this cleavage would effect the removal of the globular
C-propeptide, which is the critical step for collagen assembly into
fibrils (17, 28). No secondary sequences corresponding to the
2
C-propeptide were identified, suggesting that the
2 C-propeptide is
more extensively degraded by chymase or that the amino-terminal residue
is blocked, inhibiting the Edman reaction. The identification of
identical sequences for both the 67- and 25-kDa products suggests that
the 25-kDa fragment is the product of further cleavage of the 67-kDa
fragment. It is likely that the 67-kDa fragment is a dimer of two
1
chains linked by disulfide bonds through a small fragment of the
2
chain. Cleavage of the
1 chain (amino to the disulfide link) frees a
25-kDa fragment and may generate the transient 46-kDa fragment seen in
Fig. 2, lane Chy.
Effects of Heparin on Chymase Activity
Chymase has a high affinity for charged proteoglycans and is secreted either bound to heparin or bound extracellularly and immobilized shortly after secretion (23). Such binding may enhance or inhibit proteolysis of a large molecule such as procollagen. Therefore, it was important to test the proteolytic action of chymase-heparin complexes as well as chymase alone. To determine whether the action of purified chymase on procollagen was augmented by heparin (26), chymase was incubated with increasing concentrations of heparin prior to the addition of type I procollagen substrate in the procollagen cleavage assay (0.15 M NaCl). Both the rate of cleavage and the products formed remained unchanged in the presence of heparin over a 1000-fold concentration range (data not shown) suggesting that the binding of chymase to heparin in vivo would not affect the affinity or specificity of chymase for type I procollagen substrate. Thus, procollagen and C-propeptides may be suitable substrates in the extracellular space for free chymase or chymase bound to heparin.
Turbidometric Fibril Formation AssayA turbidity assay was
used to determine whether chymase-generated collagen was capable of
de novo fibril formation. Following addition of purified
chymase to procollagen in fibril formation buffer, a lag phase (0-8
h), a growth phase (8-16 h), and a plateau phase (16-24 h) were
observed (Fig. 4B). The lag phase represents the time necessary for chymase to generate a critical concentration of
collagen molecules before fibril assembly can begin, the growth phase
represents the assembly of collagen molecules into collagen fibrils
causing increased turbidity of the reaction solution, and the plateau
phase represents an equilibrium between collagen fibrils and collagen
molecules in solution. The turbidity profile obtained conformed to what
has been described previously for collagen fibril formation (14, 24).
Aliquots taken from a parallel reaction solution were analyzed by
SDS-PAGE in 8% separating gels under non-reducing conditions. The
turbidity profile shown in Fig. 4B corresponded to the
cleavage of procollagen to collagen-size chains shown in Fig.
4A.
Fibril Morphology by Dark Field and Electron Microscopy
Collagen fibrils were generated by the de
novo cleavage of type I procollagen by chymase, or N- and
C-proteinases. Fibrils were visible to the eye in both enzyme solutions
following a 24-h incubation period at 34 °C. When observed by dark
field microscopy, fibrils formed by the activity of N- and
C-proteinases ranged in diameter from approximately 0.25 to 1 µm
(Fig. 5, A and B). Fibrils formed
by chymase cleavage of type I procollagen were similar in shape, but
tended to be smaller in diameter, ranging from approximately 0.1 to 0.5 µm (Fig. 5, C and D). Type I procollagen incubated without chymase or N- and C-proteinases did not form fibrils.
When negatively stained (27) and observed by transmission electron
microscopy, fibrils of both types displayed 67 nm D-periods as
previously reported for native fibrils (28).
We have hypothesized that mast cell proteases play a role in connective tissue biology and contribute to the pathogenesis of fibrosis by cleaving secreted type I procollagen to collagen. To begin testing this hypothesis, we explored the effect of isolated mast cell enzymes on purified type I procollagen in vitro. Purified human mast cell chymase was found to cleave type I procollagen to collagen-size products that can then spontaneously form fibrils. These results suggest a novel role for chymase and confirm that mast cell chymase does not degrade the helical region of the collagen molecule (9).
Mast cell chymase is selectively present in human mast cells of the MCTC type (tryptase-positive, chymase-positive) and comprises about 4.5 pg of the total protein/mast cell of this type (29). Localization of chymase-containing MCTC cells to intestinal submucosa and dermis of skin is suggestive of a specific role for chymase in the connective tissue of these organs, but the precise function of the chymase is not known. The studies presented here suggest that mast cell chymase may function to augment normal procollagen processing in order to facilitate tissue repair.
The types and concentrations of enzyme inhibitors around the mast cells
after degranulation are not known, making it difficult to predict the
functional lifetime of chymase. Nevertheless, chymase is inhibited by
serine protease inhibitors in the plasma (serpins). The serpins
(1-antichymotrypsin and
1-protease
inhibitor) responsible for 80% of the inhibitory effect of plasma on
chymase, inhibit chymase at a rate 3000-fold less than rates calculated
for their inhibition of either neutrophil proteinase cathepsin G or
elastase (30). Because serpins prove to be better substrates than
inhibitors for chymase (30), chronic low level mast cell degranulation may deplete local inhibitor concentrations and provide a constant level
of active chymase to the affected tissue. In this scenario, microenvironments of chymase would be in contact with secreted procollagen. The results of these studies suggest that such a scenario
may lead to accelerated processing of procollagen to fibril-forming
collagens and to the formation of as yet unidentified procollagen-derived peptides with potential bioactivity (31, 32).
Telopeptide degradation studies have shown that N- and C-telopeptides function in determining the morphology of the collagen fibril. Specifically, the amino-telopeptide was shown to direct the orientation of collagen monomers and the carboxy-telopeptide was shown to have a role in the determination of fibril diameter (33). Fibrils of smaller diameter form when pN-collagen is allowed to assemble, due to steric hindrance from the N-propeptide folding back on the triple helix region (34). The fibril formation studies reported here demonstrate that de novo cleavage of procollagen by chymase leads to the formation of collagen fibrils smaller in diameter than procollagen proteinase-generated fibrils. This suggests that the 20-amino acid C-telopeptide extension, indicated by sequence data (Fig. 3) may limit fibril diameter by a similar mechanism. Investigation by electron microscopy into the assembly pattern and orientation of chymase-generated collagen monomers has demonstrated parallel assembly and 67-nm D-periodic symmetry of these fibrils as described previously for native collagen fibrils (28). These data demonstrate that chymase-generated collagen monomers may contribute to aberrant fibril architecture. Similarly decreased collagen fibril diameter has been reported in vivo in remodeled anterior cruciate ligament grafts (35, 36).
The data presented in this paper demonstrate clearly that human mast
cell chymase directly cleaves type I procollagen and generates novel
propeptide products. Classical propeptidases generate amino- and
carboxyl-terminal propeptides that are thought to function as feedback
inhibitors of procollagen biosynthesis (37-40). Synthetic peptide
subfragments derived from the 1 carboxyl-terminal propeptide have
been shown to be potent positive and negative effectors of matrix
protein biosynthesis in fibroblasts (32). In those studies, the
pentapeptide sequence Lys-Thr-Thr-Lys-Ser was the minimum sequence
necessary to stimulate procollagen and fibronectin production. Interestingly, the potent stimulatory effect of this pentapeptide sequence was largely dependent on the NH2- and
COOH-flanking peptide sequences. The propeptides generated by chymase
may have similar potent feedback properties and therefore could
influence not only the processing of procollagen propeptides, but also
the feedback regulation of procollagen biosynthesis. Whether
chymase-generated propeptides will have functions similar to those
generated by classical propeptidases is currently being examined.
The current study demonstrates for the first time that mast cell
chymase cleaves type I procollagen to a fibril-forming collagen molecule. This may have important implications with regard to the
increased numbers, or activation, of mast cells reported in association
with fibrotic diseases such as scleroderma, pulmonary fibrosis, and
Crohn's disease. Initial recruitment of mast cells by chemotactic
factors such as transforming growth factor- (41) and soluble stem
cell factor (42) generated at such sites may be followed by the
secretion of proteases that contribute to the production and
accumulation of collagen. The effect of chymase on procollagen
processing, collagen fibril formation, and C-propeptide degradation may
prove important for understanding the role of mast cells in connective
tissue metabolism and in these fibrotic disorders.
We thank Dr. Karl Kadler and Dr. Alexander Sieron for helpful advice and Dr. Rod Watson for the kind gift of chick procollagen.