Degradation of the epidermal-dermal junction by proteolytic enzymes from human skin and human polymorphonuclear leukocytes.

The degradation of normal human skin by the human polymorphonuclear leukocyte proteinases cathepsin G and elastase, and by a human skin chymotrypsin-like proteinase that appears to be a mast cell constituent, was examined. Enzymes were incubated with fresh, split-thickness skin for up to 8 h; the tissue was examined ultrastructurally and immunohistochemically using antibodies to known basement membrane constituents. In all cases, the primary damage observed was at the epidermal-dermal junction. Elastase degraded the lamina densa leaving scattered and disorganized anchoring fibrils, dermal microfibril bundles, and normal-appearing collagen fibers. Immunohistochemically, type IV collagen, laminin, KF1 antigen, and EBA antigen were absent. The bullous pemphigoid antigen was present and localized on the basal cells. Epidermal-dermal separation produced by the chymotrypsin-like proteinases, cathepsin G, and the human skin proteinase, was confined to the lamina lucida. The lamina densa and sub-lamina densa fibrillar network remained intact. The human skin chymotrypsin-like proteinase produced extensive epidermal-dermal separation, while cathepsin G, at comparable concentrations, produced only focal separations. Immunohistochemically, all antigens were present after incubation with enzyme. The bullous pemphigoid antigen, however, was found on the epidermal side of the split, while laminin was found on the dermal side. These results show that the epidermal-dermal junction is highly susceptible to neutral serine proteinases located in mast cells and polymorphonuclear leukocytes. Although all the proteinases produce epidermal-dermal separation, the patterns and extent of degradation are different. The distinctive patterns of degradation may provide a clue to the involvement of these proteinases in skin diseases.

A human lung mast cell chymotrypsin-like enzyme. Identification and partial characterization.

We have used a high performance liquid chromatography assay, which detects chymotryptic cleavage of the phe8-his9 bond of angiotensin I to yield angiotensin II, in order to examine human lung mast cells for the presence of chymotryptic activity. Mast cells, purified from human lung by enzymatic dispersion, countercurrent elutriation, and Percoll gradient centrifugation, were lysed or challenged with goat anti-human IgE. In multiple experiments angiotensin II-converting activity was detected in lysates of 10-99% pure mast cell preparations. Regression analysis of net percent release values of histamine and the angiotensin I-converting activity from dose-response experiments demonstrated a correlation between the two parameters, indicating that the chymotrypsin-like enzyme is a constituent of the mast cell secretory granule. The chymotryptic activity was completely inhibited by 10(-3) M phenylmethylsulfonylfluoride but not by 10(-3) M Captopril, and the pH optimum of activity was 7.5-9.5. Gel filtration of released material separated the activity from tryptase and demonstrated an approximate molecular weight of 30-35,000. The mast cell enzyme, like a human skin chymotrypsin-like proteinase, can be distinguished from leukocyte cathepsin G by lack of susceptibility to inhibition by bovine pancreatic trypsin inhibitor. Thus, an enzyme with limited chymotryptic specificity is present in human lung mast cells. The Michaelis constant of the enzyme for angiotensin I of 6.0 X 10(-5) M is similar to that of endothelial cell angiotensin-converting enzyme and is consistent with a reaction of physiologic importance.

Human skin tryptase: purification, partial characterization and comparison with human lung tryptase.

Human skin tryptase was isolated using stepwise low- and high-salt extraction and further purified 448-fold with 33% yield using octyl-Sepharose CL-4B hydrophobic affinity chromatography, Sephacryl S-200 gel filtration and finally octyl-Sepharose CL-4B or cellulose phosphate ion exchange chromatography. The skin tryptase, which has an apparent Mr of 120,000 by gel filtration in high-salt buffer, consisted of polypeptide chains of Mr 34,000 and 38,000 when resolved on SDS gels. Both polypeptide chains, labelled with [3H]diisopropyl fluorophosphate, indicated that they were representative of subunits and that the native proteinase was an aggregate of subunits. However, in some preparations only one band with Mr 34,000 was seen. In low-salt buffer the enzyme was labile and at least 1.4 M KCl was needed to keep the enzyme stabile when incubated at 37 degrees C for 30 min. Heparin glycosaminoglycan partially stabilized the tryptase but addition of protein (e.g. albumin, 80 micrograms/ml) to the tryptase-heparin mixture was needed to keep the enzyme stabile. Tryptases purified by exactly the same method from human lung tissue and from human skin had identical molecular size in gel filtration and in SDS-polyacrylamide gel electrophoresis. They also revealed identical enzyme kinetic parameters with several synthetic peptide substrates. The inhibition profile was identical for both enzymes, and they also crossreacted completely in immunodiffusion plates. These studies strongly indicate that mast cells found in skin as well as lung contain closely related, possible identical trypsin-like proteinases.

The 2.2 A crystal structure of human chymase in complex with succinyl-Ala-Ala-Pro-Phe-chloromethylketone: structural explanation for its dipeptidyl carboxypeptidase specificity.

Human chymase (HC) is a chymotrypsin-like serine proteinase expressed by mast cells. The 2.2 A crystal structure of HC complexed to the peptidyl inhibitor, succinyl-Ala-Ala-Pro-Phe-chloromethylketone (CMK), was solved and refined to a crystallographic R-factor of 18.4 %. The HC structure exhibits the typical folding pattern of a chymotrypsin-like serine proteinase, and shows particularly similarity to rat chymase 2 (rat mast cell proteinase II) and human cathepsin G. The peptidyl-CMK inhibitor is covalently bound to the active-site residues Ser195 and His57; the peptidyl moiety juxtaposes the S1 entrance frame segment 214-217 by forming a short antiparallel beta-sheet. HC is a highly efficient angiotensin-converting enzyme. Modeling of the chymase-angiotensin I interaction guided by the geometry of the bound chloromethylketone inhibitor indicates that the extended substrate binding site contains features that may generate the dipeptidyl carboxypeptidase-like activity needed for efficient cleavage and activation of the hormone. The C-terminal carboxylate group of angiotensin I docked into the active-site cleft, with the last two residues extending beyond the active site, is perfectly localized to make a favorable hydrogen bond and salt bridge with the amide nitrogen of the Lys40-Phe41 peptide bond and with the epsilon-ammonium group of the Lys40 side-chain. This amide positioning is unique to the chymase-related proteinases, and only chymases from primates possess a Lys residue at position 40. Thus, the structure conveniently explains the preferred conversion of angiotensin I to angiotensin II by human chymase.

Limited proteolysis of C1-inhibitor by chymotrypsin-like proteinases.

Limited proteolysis of C1-inhibitor was observed with human skin chymase, human cathepsin G, and bovine chymotrypsin. In each case, the inhibitor was degraded to one major product migrating slightly faster than the native inhibitor in an SDS-polyacrylamide gel. The inhibitory activity of C1-inhibitor against human plasma kallikrein was not altered by the modification with chymase. Edman degradation of the proteolyzed inhibitor revealed two sequences in a 1:1 ratio: NPNATSSSQ, the N-terminus of native C1-inhibitor, and VEPILEVSSL. This second sequence showed that the Phe33-Val34 bond was hydrolyzed. Our results provide another example of the susceptibility of the N-terminal region of C1-inhibitor to proteolytic cleavage.

Detection of MCT and MCTC types of human mast cells by immunohistochemistry using new monoclonal anti-tryptase and anti-chymase antibodies.

We developed an improved immunohistochemical technique for distinguishing human mast cells of the MCT (tryptase-positive, chymase-negative) and MCTC (tryptase-positive, chymase-positive) types utilizing a biotinylated murine anti-chymase monoclonal antibody (MAb), termed B7, and an alkaline phosphatase-conjugated murine anti-tryptase MAb, termed G3. The B7 MAb also was used to show the selective presence of chymase in mast cells. The distribution of MCT and MCTC cells in Carnoy's fluid-fixed tissue sections of human lung, skin, small intestine, and tonsils was analyzed by the new technique and the results compared to those obtained with the older method using a rabbit polyclonal antichymase antibody and a mouse anti-tryptase MAb in indirect immunoperoxidase and indirect immunoalkaline phosphatase protocols, respectively. In tissues known to contain predominantly mature mast cells, there were no quantitative differences between the two techniques, although the staining intensity achieved with the anti-chymase MAb was greater and without development of high background, compared to results achieved with the polyclonal antibody. MCT cells were the predominant type seen in the alveoli of the lung (93%) and in the small intestinal mucosa (81%). MCTC cells predominanted in the skin (99%) and in the small intestinal submucosa (77%) and, to a lesser degree, in tonsils (60%). However, in newborn foreskin tissue which contains predominantly immature forms of mast cells, 75% of all mast cells were stained uniformly and intensely with B7, whereas only 43% were stained with the polyclonal anti-chymase antibody. Therefore, the use of MAb provides for better standardization of reagents and more accurate assessment of the distribution of human MCT and MCTC cells in tissues than previously available methods.

Demonstration of tryptase in bovine cutaneous and tumor mast cells.

We examined three tissue samples from each of four cows with non-lesional skin, tissue samples from a cow with multiple cutaneous mast cell tumors, and samples from another cow in which mast cells were infiltrating multiple lymphosarcomas of the skin, for the presence of tryptase and chymase by enzyme cytochemical and immunohistological methods. The enzyme activities of tryptase and chymase were tested using N-carbobenzoxy-glycilglycil-L-arginine-2-naphthylamide (Z-Gly-Gly-Arg-NA) and naphthol-AS-D-chloroacetate (N-AS-D-CA) as substrates, respectively. Tryptase reactivity could be demonstrated in frozen and Carnoy-fixed paraffin sections. Chymase reactivity was seen in neither frozen nor paraffin sections of formalin- or Carnoy-fixed skin tissues. Antibody linkage with a polyclonal rabbit anti-human skin tryptase antibody was highly specific in bovine normal cutaneous, infiltrating, and tumor mast cells. More than 90% of the tumor mast cells were distinctly tryptase-positive. With alcian blue, only slightly more than 10% of the mast cells stained clearly positive and with methylene blue hardly any staining of mast cell granules could be demonstrated. No antibody labeling of mast cell granules in any of the tissue sections was detected by the use of rabbit anti-dog chymase antiserum. These results indicate that there is a striking antigenic similarity of bovine tryptase to its canine and human equivalents. The demonstration of tryptase is an important tool in confirming the diagnosis of undifferentiated mast cell tumors. In contrast to other species, chymase appears to be completely absent in bovine skin mast cells.

Treatment of urticaria pigmentosa with corticosteroids.

Based on a previous observation that the long-term application of potent topical corticosteroids under occlusion to normal skin resulted in the loss of mast cells, we investigated the effects of intralesional and topical steroids in urticaria pigmentosa (UP). Three patients with UP had lesions that were injected with triamcinolone acetonide. Four weeks after injection, all patients showed a loss of Darier's sign, and, by eight weeks after injection, there was a dramatic clearing of the plaques and a decrease in brown hyperpigmentation. By 12 weeks, mast cells were undetectable by light microscopy and transmission electron microscopy (TEM). The injection sites remained dramatically improved for as long as one year after treatment, and histamine content was reduced 95% in one patient 48 weeks after injection. In six patients, the topical application of 0.05% betamethasone dipropionate under occlusion to limited areas induced almost complete clearing of UP lesions. Lesions treated with an emollient under occlusion, as a control, demonstrated no change. After treatment, no mast cells were seen by light microscopy or TEM, and this persisted for at least 24 weeks. There was also a significant decrease in tissue histamine levels in the treated areas. The treated areas remained clinically improved for at least nine to 12 months. These data indicate that steroid therapy dramatically decreases the excess number of normal-appearing mast cells in UP as well as induces a prolonged resolution of UP lesions. Local corticosteroids thus are a useful therapeutic modality for UP.

Serine proteinases in human cutaneous mastocytosis.

The main chymotryptic and tryptic proteinases of human skin were found in high-salt extracts of human dermis. The levels of these enzymes were markedly increased in salt extracts of human cutaneous mastocytosis as compared to the levels found in extracts of involved skin from the same patients, human cutaneous hemangiomas, and normal human skin. These data suggest that the chymotryptic and tryptic proteinases of human skin are primarily of mast-cell origin.

Structure of the dermal-epidermal junction and potential mechanisms for its degradation: the possible role of inflammatory cells.

Evidence was presented indicating that the DEJ as a basement membrane is highly susceptible to degradation by a variety of neutral proteolytic enzymes with different specificities. The effect of endoglycosidases which degrade heparan sulfate was also discussed. The latter enzymes are capable of removing heparan sulfate from the DEJ, but little gross alteration of structure, such as tissue detachment, appears to result from the loss of this component. Of the proteinases discussed, PMN elastase and probably type IV collagenase are the most destructive. This is likely related to their ability to degrade the type IV collagen network. Even though proteinases with chymotrypsinlike and trypsinlike specificity were not efficient at degrading the lamina densa or removing type IV collagen from intact basement membranes, these proteinases were capable of producing epidermal detachment from the lamina densa. Many inflammatory cells of the immune system contain proteinases and endoglycosidases with the potential to degrade the DEJ and other basement membrane zones, suggesting that these cells may have a significant pathologic role in basement membrane-related diseases. PMNs and mast cells may be of particular interest because they have stored within their secretory granules high concentrations of neutral serine proteinases which have been demonstrated to degrade the DEJ.

Cutaneous mast cell depletion results from topical corticosteroid usage.

The effect of long-term topical application of corticosteroids on human cutaneous mast cells was examined. Two potent corticosteroids, clobetasol-17-propionate and fluocinonide, produced a greater than 85% decrease in histamine content over a 6-wk treatment period, whereas betamethasone valerate, a less potent corticosteroid, produced a 66% decrease. Electron microscopic examination of the biopsies taken from sites after 6 wk of treatment indicate that the reduced levels of histamine were caused by the depletion of mast cells, as evidenced by: the inability to identify any cells representative of mast cells by detailed electron microscopy of the biopsies; and the marked acellularity around the vasculature where mast cells are certain to be detected. Histamine levels did not begin to decline until after 3 wk of corticosteroid treatment, indicating that corticosteroids are not immediately harmful to mast cells. Electron microscopic examination of biopsies taken at the beginning of treatment and 1 wk later showed normal-appearing mast cells, whereas at 3 wk, a small population of mast cells was detected with features usually associated with degenerating or dying cells. These observations suggest that protracted application of corticosteroids to skin is toxic to mast cells. After discontinuation of treatment, the drug-related atrophy associated with chronic application of potent corticosteroids to skin is rapidly reversed, and skin structure returns to near normal by 14 days. Over this time period, however, histamine levels did not increase and mature mast cells could not be observed by electron microscopy. At 14 days post-steroid treatment, the first signs of cells containing sparse amounts of granules having the characteristics of mast cell granules were seen. We interpret this to represent new mast cells beginning to mature in the skin. By 3 mo, histamine levels returned to normal, demonstrating the reversibility of the steroid-induced mast cell depletion. The studies presented here establish the deleterious effects of long-term topical corticosteroid treatment on cutaneous mast cells, and begin to establish a system in which the development of mast cells in tissue can be investigated.

Ultrastructural analysis of human T and TC mast cells identified by immunoelectron microscopy.

Tryptase and chymase were localized in human mast cells by immunoelectron microscopy, enabling the T (tryptase positive, chymase negative) and TC (tryptase positive, chymase positive) types of mast cells to be identified and ultrastructurally characterized. A double immunogold staining procedure was performed on samples of human skin, small intestine, and lung with rabbit polyclonal IgG anti-chymase and mouse monoclonal IgG anti-tryptase primary antibodies and gold-conjugated secondary antibodies. Approximately 225 mast cells were examined in this fashion; comparable sections from 170 of these mast cells along with approximately 200 additional mast cells also were examined using techniques optimized for ultrastructural detail. Each secretory granule of TC mast cells contained both tryptase and chymase; secretory granules of T mast cells stained strongly positive for tryptase alone. Extremely small amounts of chymase appeared to be present in an occasional T mast cell granule. Staining for the neutral proteases was more intense over electron-dense regions of the granules, particularly noticeable over the characteristic discrete scrolls of T mast cells. T and TC mast cells each had large numbers of cytoplasmic granules, nuclei with peripherally condensed chromatin and low nuclear/cytoplasmic ratios, indicating maturity of both cell types. TC mast cell granules generally were more uniformly electron dense, larger and more numerous than T mast cell granules, which were more variable in shape. Compact solid-core scrolls, peripheral parallel lamellae and amorphous electron-dense material were found in granules of both cell types. Only TC mast cells had granules with grating and lattice substructures; only T mast cells had granules containing discrete scrolls. Less commonly, T mast cells were detected containing granules with a characteristic beaded or particulate ultrastructure. The ultrastructural features noted above were observed in T and TC mast cells regardless of the tissue in which they were examined and thereby permit T and TC mast cells to be distinguished by ultrastructure alone.

Modification of integrin-mediated cell attachment to substrata by serine proteinases in the presence and absence of divalent cations.

The sensitivity to serine proteinases of cellular proteins involved in cell-matrix adhesion was investigated using C32 melanoma cells. Cells dissociated from monolayers by the metal chelator ethylenediaminetetraacetic acid were incubated with proteolytic enzymes, and then attachment was quantified by standard cell adhesion assays. The effect of proteinases was found to depend on the presence of Ca2+ in the incubations. Incubation with 100 nM trypsin or chymotrypsin for 1-2 h (37 degrees C) in the absence of Ca2+ reduced cell attachment to vitronectin (Vn), fibrinogen (Fb), laminin, and fibronectin by approximately 80, 80, 40, and 30%, respectively. Viability studies indicated that such treatment with proteinases was not cytotoxic. Inclusion of 0.1 nM CaCl2 in the incubations prevented the loss in attachment to all substrata. In the case of Fb, proteinase treatment in the presence of Ca2+ had an additional effect; it improved cell attachment to this substratum by about 50%. C32 cells have been shown to express the integrin alpha v beta 3 (Vn receptor) which mediates attachment to Vn and Fb in a GRGDS-sensitive manner. Attachment of C32 cells to Vn and Fb prior to proteinase treatment and after proteinase treatment in the presence of Ca2+ was 90% inhibited by the addition of GRGDS peptide to the attachment assays. These results suggest that the adhesion observed both before and after proteinase treatment was mediated by this integrin. Analysis of the Vn receptor from proteinase-treated cells by immunoblotting of cell extracts and by SDS gel electrophoresis of immunoprecipitated receptor revealed no detectable change in either the alpha v or beta 3 subunit that correlated with loss in attachment. Similarly proteinase treatment in the presence of Ca2+ did not produce detectable alterations in the subunits which might correlate with the improved attachment to Fb. Consistent with these results, an enzyme-linked immunoassay to quantify cell surface receptors revealed little difference in the amount of Vn receptor on cells treated with proteinase in the presence or absence of Ca2+. Degradation of the alpha v subunit was demonstrated, however, at proteinase concentrations higher than those required to affect cell attachment. Thus, treatment of cells with serine proteinases can affect integrin-mediated attachment to matrix proteins in a manner moderated by Ca2+, but the alterations in attachment do not appear to be accompanied by detectable proteolytic modification of the integrin.

Ultrastructural analysis of maturing human T and TC mast cells in situ.

Mast cells at immature stages of development were identified in human tissues by electron microscopic techniques. General morphologic criteria of immaturity (such as a high apparent nuclear:cytoplasmic ratio and small cell size), the presence of few granules (those present being smaller than those in mature mast cells) and a lack of features of mast cell activation were used together to determine the level of maturity. Mast cells were identified as being of the T or TC type by immunogold staining with polyclonal rabbit IgG anti-chymase and murine monoclonal anti-tryptase primary antibodies and the appropriate gold-labeled secondary antibodies. Only those cells with tryptase-positive granules were recognized as mast cells. Immature T mast cell granules contained the same characteristic discrete scrolls found in their mature counterparts and all stained positive for tryptase. The presence of trace amounts of chymase in a minority of these granules, as in mature T mast cells, could not be ruled out. The majority of granules in immature TC mast cells had one or more amorphous electron-dense cores rather than the grating and lattice substructures characteristic of granules in mature TC mast cells. Secretory granules in immature TC mast cells stained positively for tryptase and chymase. Occasional immature TC mast cells contained a complete granule or a portion of a granule with the substructure characteristic of mature TC mast cells, favoring the concept that these TC mast cell forms are developmentally related. Essentially all mast cells in foreskin of newborns appeared immature, whereas 10, 5, 10, and 15% of the mast cells in adult lung, foreskin, bowel mucosa and bowel submucosa, respectively, appeared immature. The distribution of T and TC types of immature mast cells seemed to parallel that of the mature mast cell types. These compositional and ultrastructural differences between immature T and TC types of mast cells suggest that from the time granule formation begins, and possibly before this time, each type of human mast cell follows a distinct developmental pathway.

Trypsin-induced, neurokinin-mediated contraction of guinea pig bronchus.

Proteases may act as cell signaling molecules via protease-activated receptors (PARs). PAR1, PAR3, and PAR4, but not PAR2, are activated by thrombin, whereas trypsin can activate PAR2 and PAR4. In this study, trypsin (3-100 nM) evoked concentration-dependent contractions of guinea pig isolated bronchus, however, thrombin (3-300 nM) was a weak spasmogen. Neither the PAR2-activating peptide SLIGRL (100 microM) nor mast cell tryptase (100 nM), a trypsin-like protease known to activate PAR2, evoked contraction. A role for neurokinins in trypsin-induced contraction is suggested by our observation that contractions to trypsin were markedly attenuated in the presence of neurokinin receptor antagonists. Depletion of neurokinins in sensory nerves with capsaicin also markedly reduced the ability of trypsin to evoke contraction. In electrophysiological studies, trypsin did not evoke action potentials in C-fiber afferents whose receptive fields were located in the trachea or main bronchi. The results from this study support the hypothesis that trypsin activates a mechanism allowing for local release of sensory neurokinins from afferent C-fibers and that this release occurs independently of the sensory function of these nerves.

Inhibition of human chymase by 2-amino-3,1-benzoxazin-4-ones.

A series of 2-sec.amino-4H-3,1-benzoxazin-4-ones was evaluated as acyl-enzyme inhibitors of human recombinant chymase. The compounds were also assayed for inhibition of human cathepsin G, bovine chymotrypsin, and human leukocyte elastase. Introduction of an aromatic moiety into the 2-substituent resulted in strong inhibition of chymase, cathepsin G, and chymotrypsin. Extension of the N(Me)CH2Ph substituent by one methylene unit was unfavourable to inhibit these proteases. Towards chymase, 2-(N-benzyl-N-methylamino)-4H-3,1-benzoxazin-4-one (32) and 2-(N-benzyl-N-methylamino)-6-methyl-4H-3,1-benzoxazin-4-one (33) were found to exhibit Ki values of 11 and 17 nM, respectively, and form stable acyl-enzymes with half-lives of 53 and 25 min, respectively. Benzoxazinone 33 also inhibited the human chymase-catalyzed formation of angiotensin 11 from angiotensin I.

Production of active recombinant human chymase from a construct containing the enterokinase cleavage site of trypsinogen in place of the native propeptide sequence.

Human chymase, a chymotrypsin-like proteinase found in mast cells, was produced in an enzymatically active recombinant form. The protein was expressed in Escherichia coli as part of an insoluble fusion protein which was solubilized and renatured. The structure of the fusion protein was NH2-ubiquitin-enterokinase cleavage site-chymase-COOH. The enterokinase cleavage site of trypsinogen replaced the native propeptide sequence of chymase, allowing for activation by a readily available proteinase (enterokinase) of known specificity. Characterization of refolded-activated recombinant chymase with substrates and inhibitors demonstrated properties identical to that of the native proteinase isolated from skin.

Serine proteinases are regionally segregated within mast cell granules.

BACKGROUND: By ultrastructure, human skin mast cell granules exhibit amorphous regions situated next to crystalline regions with scroll, lamellar, and grating/lattice-like configurations. The composition and function of these structural domains is unknown. These granules also contain large amounts of three serine proteinases termed tryptase, chymase, and cathepsin G. In this study, subgranular proteinase distribution is correlated with granule ultrastructure. EXPERIMENTAL DESIGN: Human skin sections were immunolabeled for each proteinase, visualized with gold-conjugated antibodies, and then stained with heavy metals to enhance granule features. Both single and double labeling experiments were analyzed. RESULTS: Comparison of the immunogold labeling patterns revealed nonrandom distributions for each proteinase within most granules. At low magnification, chymase and cathepsin G immunoreactivity was found preferentially over more electron-dense granule subregions, whereas tryptase immunoreactivity was found preferentially over less electron-dense subregions. The complimentary staining pattern of the proteinases was also observed in double labeling experiments where tryptase immunoreactivity was compared directly with that of chymase or cathepsin G. Higher magnification of sections revealed that electron-dense granule areas corresponded to amorphous regions, whereas less dense regions most often demonstrated crystalline structures. CONCLUSIONS: These observations correlate the subgranular distribution of serine proteinases with specific granule ultrastructure and suggest that the different morphologic features within granules may be related, in part, to the packaging of serine proteinases. Human mast cells differ from human basophils and rat mast cells by virtue of the presence of high amounts of tryptase and the presence of crystalline substructures. This raises the possibility that association of tryptase with crystalline structures reflects a specialized form of packaging permitting efficient storage of high levels of this proteinase within mast cell granules.