Cleavage of Type VI1 Collagen by Interstitial Collagenase and Type IV ...

4 downloads 0 Views 2MB Size Report
Robert W. GlanvillellII , and Robert E. BurgesonT))**$$. From the $Division of Dermatology, Department of Medicine, Washington University School of Medicine, ...
THEJOURNALOF BIOLOGICAL CHEMISTRY 0 1989 by The American Society for Biochemistry and Molecular Biology, Inc

Vol. 264, No. 7, Issue of March 5, pp. 3822-3826,1989 Printed in U.S.A.

Cleavage of Type VI1 Collagen byInterstitial Collagenase andType IV Collagenase (Gelatinase) Derived from Human Skin* (Received for publication, September 21, 1987)

Jo Louise Seltzer$, ArthurZ. EisenS, Eugene A. Bauers, Nicholas P. Morris711 , Robert W.GlanvillellII, and Robert E.BurgesonT))**$$ From the $Division of Dermatology, Department of Medicine, Washington University School of Medicine, St. Louis, Missouri 63110, the §Department of Dermatology, Stanford University School of Medicine, Stanford, California 94305, and the Whriners Hospital for Crippled Children and the Departmentsof )IBiochemistry, **Cell Biology and Anatomy, and $$Dermatology, Oregon Health Sciences University, Portland, Oregon 97201

Type VI1 collagen is the major structural protein of anchoring fibrils, which are believed to be critical for epidermal-dermal adhesion in thebasement membrane zone of the skin. To elucidate possible mechanisms for the turnover of this protein, we examined the capacities of two proteases, human skin collagenase, which degrades interstitial collagens, and a protease with gelatinolytic andtype IV collagenase activities,to cleave type VI1 collagen. At temperatures below the denaturation temperature,pepsin cleaves type VI1 collagen into products of -95 and -75 kDa. Human skin collagenase cleaved type VI1 collagen into two stable fragments of -83 and -80 kDa, and the type IV collagenase (gelatinase) produced a broad band of -80 kDa as determined bysodium dodecyl sulfate-polyacrylamide gel electrophoresis. Cleavage of type VI1 collagen was linear with time and enzyme concentration for both enzymes. Although the K,,, values were similar for both enzymes, the catalyticrate of cleavage by type IV collagenase is much faster than by interstitial collagenase, and shows a greater rate of increase with increasingtemperature. Sequence analysis of the cleavage products from both enzymes showed typical collagenous sequences, indicating a relaxation in the helical part of the type VI1 collagen molecule at physiological temperature which makes it susceptible to gelatinolytic degradation. Interstitial collagenase from both normal skin cells and cells from patients with recessive dystrophic epidermolysis bullosa, a severe hereditary blisteringdisease in which both an anchoring fibril defect and excessive production of collagenase can be observed, produced identical cleavage products from type VI1 collagen. These data suggest a pathophysiological link between increased enzyme levels and theobserved decrease or absence of anchoring fibrils.

Type VI1 collagen was shown to be a unique gene product following purification from human amnion (1, 2). Each of the three identical a-chains/molecule consists of a 150-kDa globular domain at the carboxyl terminus attached to a helical * This work was supported by United States PublicHealth Service Grants AM 12129, AR 19537, AM 35532, and AM 35689, by the Shriners Hospitalfor Crippled Children, the Washington University/ DEBRA Centerfor Research and Therapyof Epidermolysis Bullosa, andtheWashingtonUniversity-Monsanto Biomedical Research Agreement. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement” in accordance with 18U.S.C. Section 1734 solely to indicate this fact.

section of 170 kDa. In tissue, type VI1 collagen is found arranged asantiparallel dimers, with the helical portions disulfide-bonded in a short overlap, and with the COOHterminal globular positions at either end (3). Immunologic methods have identified type VI1 collagen as the principal component of anchoring fibrils, structures which are believed to be important in attaching the epidermis to the dermis (4). Segment-long spacing crystallites of type VI1 collagen suggest that the triple helical domain can fully account for the banded pattern observed in the anchoring fibrils (2). Ultrastructurally, these fibrils are seen to insert directly into the lamina densa and either arch and reinsert intothat structure, or extend perpendicularly from the lamina densa and insert into anchoring plaques in thedermis which contain the 150-kDa globular domain. Additional anchoring fibrils bridge adjacent anchoring plaques to form an extended network of fibrils, entrapping large numbers of dermal fibrous components (5-8). Anchoring fibrils appear to be absent and/or altered in several disease states, most notably recessive dystrophic epidermolysis bullosa (RDEB),’ a hereditary blistering disorder (9-12). It is not known whether this is due to a genetic defect in the expression of the anchoring fibril complex or to its degradation by excessive proteolysis in the dermis (13, 14). Prolonged proteolysis of the helical portion of type VI1 collagen by pepsin results in two fragments, P1 and P2, of mass 95 and 75 kDa, respectively (1).Trypsin and chymotrypsin also cleave this molecule in the centralregion, giving products of slightly different sizes. Human skinfibroblast collagenase cleaves within the triple helix of interstitial collagens types I, 11, and 111, although it does not cleave two basement membrane collagens, types IV and V (15, 16).In thispaper we present evidence that human skin collagenase, as well as a protease derived from human skin which displays both gelatinolytic (17) and types IV and V collagenolytic activities, currently referred to as type IV collagenase (see Ref. 18), can at physiologic temperature degrade the type VI1 collagen helix into products similar in size to the pepsin cleavage products. Unless prolonged incubation at greater than physiologic temperature occurs, these products retain helicity and are stable to further proteolysis. MATERIALSANDMETHODS

Preparation of Collagens-Type VI1 collagen was purified from human chorioamniotic membranes as previously described (1) and upon solubilization was stored frozen in 0.05 M Tris-HCI, pH 7.5, 0.4 The abbreviations used are: RDEB, recessive dystrophic epidermolysis bullosa;NaDodSOr-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis.

3822

3823

Metalloprotease Degradation of Type VII Collagen M NaC1. "C-Labeled type I collagen was extracted from guinea pig skin by the method of Gross and Lapiere (15). Purification of Enzymes-Collagenase was purified to homogeneity as the proenzyme from serum-containing culture medium of normal and recessive dystrophic epidermolysis bullosa human skin fibroblasts, using a modification of the method of Stricklin et al. (19, 20). In this procedure, a concentration step on a Znz+/chelate affinity column was interposed between elution from CM-cellulose and gel filtration on Sephadex G-100. Type IV collagenase (gelatinase) was purified to homogeneity from medium of human skin inexplant cultures by the method of Seltzer et al. (21). Enzyme Assays-Procollagenase was fully activated by incubation with 1 mM phenyl mercuric chloride for 90 min at 37 "C (22). Type I collagenase activity was assayed by incubating the active enzyme with 25 p1 of a 2 mg/ml "C-labeled native reconstituted guinea pig collagen substrate gel (specific activity = 25,000 cpm/mg) at 37 "C as previously described (23). Type IV collagenase (gelatinase), which is fully activated when purified from skin explant culture medium, was assayed as described previously (17). Briefly, 25 pl of 14C-labeledguinea pig type I collagen was denatured by heating at 100 "C for 10 min. 75 pl of the purified enzyme was added, and after incubation at 37 "C, the reaction was terminated with 25 11of 100% trichloroacetic acid. Product peptides soluble in 20% trichloroacetic acid were counted in a liquid scintillation spectrometer. The cleavage of type VI1 collagen by interstitial collagenase or type IV collagenase was carried out at varying temperatures and under varying conditions of enzyme concentration and times of incubation. Cleavage products were analyzed by separation of total reaction mixtures on 8% NaDodS04-PAGE according to the method of King and Laemmli (24), assuming molecular weights of -170,000 for the intact oll(VII) chains and -75,000 and -95,000 for the two pepsin fragments (1).The digestion of the starting material and theappearance of product peptides were quantitated by scanning of the Coomassie Blue-stained gels on an LKB 2202 Ultroscan laser densitomer, using the Gelscan computer program for quantitation of the peaks. Electroblotting and Sequencing-Protein sequencing was performed on an Applied Biosystems 470A gas-phase Sequenator. Samples were electroblotted onto polyvinylidene difluoride membrane for NaDodS04-polyacrylamide gels and loaded directly onto the sequenator by the method of Aebersold et al. (25). Identification of the phenylthiohydantoin-derivativeswas performed by high pressure liquid chromatography as previously described (26). Other Assays-The protein concentration of enzyme preparations was determined spectrophotometrically by the method of Groves et 01. (27) using bovine serum albumin to establish a standard curve. Immunoreactive fibroblast collagenase was measured by enzymelinked immunosorbent assay using the method of Cooper et al. (28).

RESULTS

We initially incubated the 510-kDa triple helical portion of type VI1 collagen withhumanskin collagenase at 28 "C. Product peptidesof approximately half the size of the a-chain were observed. Increasing the incubation temperature to as high as 37 "C revealed more extensive cleavage without the formation of additional products(see below). Fig. 1shows the human skin collagenase cleavage products of M, -83,000 and -80,000 as determined by gel electrophoresis. It is of interest that thecollagenase produced by recessive dystrophic epidermolysis bullosa fibroblasts makes a cleavage that appears to be qualitatively identicalto thatof the normal collagenase. The existence of two primary collagenase cleavage products of a size similar to thoseof the two pepsin productssuggested a possible relaxation of the triple helix in the mid-portionof the typeVI1 monomer. If this were the case,we reasoned that type VI1 collagen might also be susceptible to human skin gelatinase, a protease which has recently been shown to have type IV collagenolytic activity (18). Indeedthis enzyme cleaved type VI1 collagen, yielding a broad band of -80 kDa (Fig. 1).A minor cleavage product of slightly smaller size was also produced (Fig. 2). Circular dichroism (at 221 nm) of disulfide-bonded type VI1 collagen indicates that a biphasic denaturation occurs with

I

Normal Collagenase

I

0

I

I

1

I

I

L

C

a-

RDEB Collagenase

1

2

C IViG

a-

I

64-

Type IV Collagenase Gelatinase

20.

I

I

10

I

I

20

I

I

30

I

I

L

40

P o s i t i o n Number

FIG. 1. Densitometric scan of type VI1 collagen cleavage by various human skin proteases. Type VI1 collagen (3 pg)was incubated at 37 "C for 18 h with purified normal human skin collagenase (0.1 pg), RDEB collagenase (0.05 pg), or human skin type IV collagenase (gelatinase) (0.12 pg). The procollagenase was fully activated by preincubation with 1 mM phenyl mercuric chloride at 37 "C for 90 min as described (21). Densitometric scan of type VI1 collagen cleavage products: Control, intact type VI1 collagen substrate; C, principal collagenase cleavage products; CEB, RDEB collagenase; C W I G , principal type IV collagenase (gelatinase) cleavage products.

values of T,,,a t 41 and55 "C. The amplitudesof the transitions suggest that themajor portion of the triple-helix meltsat the lower temperature while the disulfide bond-stabilized 60-nm overlap exhibits a higher thermal stability.' To examine the possible effects of temperature on proteolytic susceptibility, we incubated type VI1 collagen with interstitial collagenase or type IV collagenase a t varying temperatures. Breakdown of type VI1 collagen by both enzymes was extremely temperature-dependent. Byusing a high ratio of substrate to enzyme, the same cleavage products as shown in Fig. 1 were obtained at 28 and 32 "C (data not shown). Fig. 2 demonstrates the dramatic increase in cleavage rates with small temperature increments. At 34 "C, interstitial collagenase, at anenzymembstrate ratio of1:5, cleaved 4 0 % of the type VI1 collagen in a 20-h incubation (Fig. 2 A ) . With an increase in incubation temperature of only three degrees (Fig. 2B), the substrate was completely cleaved, with very littlefurther breakdown of the products. Breakdown of the reaction products from interstitial collagenase was somewhat more pronounced at 39 "C (Fig. 2C), a temperature approaching the H. P. Bachinger, N. P. Morris, and R. E. Burgeson, unpublished observations.

95 kDa + 7 5 kDa,

of Type VII Collagen

Metalloprotease Degradation

3824

-

II)

67 kDa

-

4

95 kDa 75 kDa4 _I

-94kDa

67 kDa

m ! "67

-

94 kDa

kDa

43 kDa

-43 kDa

a43 kDa

30 kDa

-30kDa

e30 kDa

FIG. 2. Effect of temperature on the cleavageof type VI1 collagen by various proteases. In each case type VI1 collagen (9 pg) was incubated in buffer alone ( V U ) , with previously activated normalhuman skin collagenase (CN = 2 pg), or with normal human skin type IV collagenase (gelatinase) (C IV/C = 0.2 pg), at the indicated temperature prior to examination using NaDodS0,-PAGE. The M,scales from collagen helical (left)and globular (right) peptides are indicated. The collagen helical scale was determined using type VI1 collagen monomer and the pepsin cleavage fragments P1 and P2. A, 34; B, 37; C, 39 'C.

denaturation temperatureof the helix. collagenase. The K, for type VI1 trimer for each enzyme is The reaction rate of type VI1 collagen breakdown with type about 0.3 p ~ but , kat(moles of substrate degraded per mol of IV collagenase, alsoshown in Fig.2, A, B , and C, clearly enzyme/h) for type IV collagenase is 160 uersus 0.05 for demonstrates a relaxation in the helicity of the central portion collagenase. Sincethe relativeeffectiveness of interstitial of the typeVI1 monomer with increasing temperature. It has collagenase to type IV collagenase was greater a t lower tembeen shown that type IV collagenase is unable to cleave the peratures, the increased katfor the gelatinolytic enzyme a t helical type I collagen molecule (17). Although this enzyme 37 "C is presumably due to partialmelting of the helix. can apparently cleave within the helix of types IV and V DISCUSSION collagen (18),the reaction rateis more than 2 orders of magnitude slower than with denatured type I collagen (or From the data presented, it clear is that human skin interindeed with any denaturedcollagen). Since even a t 39 "C the stitial collagenase and type IV collagenase (gelatinase) can product peptide producedby type IV collagenase is fairly cleave type VI1 collagen monomer. Both extracellular neutral resistant to further breakdown, and since the NH2-terminal proteases are specific for collagenous substrates. Even when sequence of theproductischaracteristic of collagen (see enzyme andsubstratearepresentin nearlyequimolar below), it is highly probable that susceptibility to cleavage is amounts, the cleavage fragments maintain the helicity prea t least partially determined by some looseningof the collagen dicted from the aminoacid composition (Fig. 2). Any comprohelix. mise in thehelical stability through heat denaturationwould T o determine the nature of the cleavage specificity the have resulted in degradation into small peptides by the gelaNH2-terminal sequence for both the-83-kDa interstitial col- tinolytic activity of the type IV collagenase. The significant lagenase cleavage peptide and the type IV collagenase cleavage temperature dependence of proteolysis supports the hypothproduct was obtained.Undertheexperimentalconditions esis that the middle region of the intact type VI1 monomer, used, the NH2 terminus of the type VI1 collagen monomer which is susceptible to pepsin cleavage at 4 "C withprolonged was blocked. The data suggest very little loss of Coomassie incubation, has a relaxed helical character. The >3000-fold staining material in the conversion to the product peptides difference in rate between type IV collagenase and interstitial (Fig. 3); we assume that the single sequenceobtained for both collagenase, reflecting the difference between the capacities product peptide bands represents both the peptide from which of the two enzymes to cleave type I gelatin (18, 291, also sequence was obtained and another peptide of approximately supports this hypothesis. It should be emphasized that while equal length beginning at theoriginal blocked NH2 terminus. the size of the product peptidesfrom both enzymes indicates Fig. 4 demonstrates schematically the approximate sites of proteolytic cleavage within the type VI1 collagen monomer. cleavage near the middle of the 170-kDa monomer, our data As shown in Table I the product sequences, while different, do notpreclude the possibility of additional cleavages. The slight size differences between interstitial collagenase were both indicative of cleavage within the collagenous portion of the molecule. About 10% of the amino acid sequence and type IV collagenase cleavage products reflected bygel differof type VI1 collagen has been determined, including the pepsin electrophoresis (Fig. 1) are almost certainly related to cleavage siteandtheNH2terminus ofP2." Thesites of ences in the respective enzymespecificities. It hasbeen shown interstitial collagenase and type IV collagenase cleavage are that collagenase cleaves only at Gly-Ile and Gly-Leu bonds different from portions of the molecule heretofore sequenced. (30), while type IV collagenase (gelatinase) also cleaves GlyBreakdown of type VI1 collagen catalyzed by human skin Phe, Gly-Val, and Gly-Ala bonds at similar rates (31). Thus, collagenase is a linear functionof both enzyme concentration the slightly smaller size of type IV collagenase cleavage prod(Fig. 3A) andtime (Fig. 3B),asisthe casefor type IV ucts may reflect more cleavages in the middle region of the collagenase (Fig. 3C). At 37 "C cleavage of the typeVI1 trimer type VI1 molecule, producing small peptides which are not by type IV collagenase is much faster than by interstitial readily visible under the conditions of NaDodS04-PAGE. Nevertheless, the slightly smaller sizes of the apparent initial R. Burgeson and N. P. Morris, unpublished data. type IV collagenase products could alternatively be due to

Metalloprotease Degradation of Type VIZ Collagen A 100

100

NORMAL COLLAGENAYL

0

. .

B -

0.25 050 1.00

0'

110' 220' 330'

HUMANSKIN IPE ECOLLAGENASElGELATlNASE ST d 120' 240' 360'

-

--

" "

" "

-

NORMAL COLLAGENASE

XU

3825

80 -

"

--

-

2 0

60

-

LL

"

n

I

240

Collagenase Added

Time (min)

(119)

I

120

360

Time (min)

FIG. 3. Effect of protease concentration and time of incubation onthe cleavage of type VI1 collagen. A, effect of interstitial collagenase concentration. Type VI1 collagen (8.5 pg) was incubated with varying concentrations of activated normal human skin collagenase for 330 min a t 37 "C. Reaction products were analyzed by NaDodS0,-PAGE and product peakswere quantitated using an LKB 2202 Ultroscan laser densitometer and the Gelscan computer program. B, effect of time of incubation with interstitial collagenase. Type VI1 collagen (8.5 pg) was incubated with activated normal human skin collagenase (1 pg) a t 37 "C.At the indicated times, one-fourthof the total reaction mixture was removed and stopped by the addition of 30 pg/ml EDTA. Products were analyzed as indicated in A above. C, effect of time of incubation with type IV collagenase (gelatinase). Type VI1 collagen (5 pg) was incubated with human skin gelatinase (0.4 pg) for the indicated times. Portions of the mixture were removed and analyzed as described in I? above.

II I

D7

t y p e IV c o l l a g e n a s e

J

interstitial

1L

PI)

pepsln

collagenase

FIG.4. This schematic representation of the type VI1 collagen triple-helical domain illustrates the most likely relationships between the cleavage sites of type IV collagenase, interstitial collagenase, and pepsin. Both type IV and interstitial collagenase cleave within helical sequences while the pepsin site is within a non-helical disruption that presumably contributes increased flexibility to thisregion. Additional sites may occur a t both the NH, or COOH termini.

additional cleavages near the NH2- or COOH-terminal portion of the helix. Anchoring fibrils, thought tobe important for the integrity of the epidermal-dermal junction (2, 4), are conspicuously

absent in certain diseases. Thus, the ability of interstitial collagenase and type IV collagenase to cleave the major protein component of anchoring fibrils may have clinical significance. For example, areas invaded by the basalcell carcinomas often show disruption of basement membrane, and concomitant loss of anchoring fibrils can be demonstrated immunologically (32). Likewise, although reportsdiffer in the extent to which a defect in anchoring fibrils is present, all RDEB patients show a t least some alteration or decrease in the number of anchoring fibrils (9-12). Although it has been suggested that the primary defect causing absenceof anchoring fibrils might be failure to express the type VI1 collagen gene, in a recent reportfour of four RDEB patientsexamined were shown to have the COOH-terminal globular domain of type VI1 collagen in their basement membrane,even though no anchoring fibrils were present (33). Since it has alsobeen suggested that the blistering and skin fragility exhibited by patients with recessive dystrophic epidermolysis bullosa may

TABLE I Comparison of interstitial collagenase and type IV collagenuse (gelntinase) cleavage sites with known type VII collagen sequence 40 pg of type VI1 collagen was incubated for 20 h a t 36 "Cwith 9 pg of interstitial collagenase and 0.4 pg of type IV collagenase (gelatinase). The reactions were terminated with EDTA, concentrated, subjected to NaDodS0,PAGE, and sequenced as described under "Materials andMethods." X, cysteine or tryptophan; Z, hydroxyproline; -, unidentified residue; ( ), ambiguity in sequencing (letter within parentheses is most likely residue). NH2-terminalsequence

Interstitial collagenase cleavage peptide Type IV collagenease cleavage peptide P1-pepsin cleavage site, type VI1 monomer P2-pepsin cleavage site CB5

GX-G-"-Z-G-K--ZcG--I G7-G-P-Z-G-P-Z-G-P-K A-(T)-Q-P-R-P-E-P-G-P-V "Y---G(K)"G-Q-Q-G-E ---E-A-G-D-K-K-GG-P G7A-P-V-G-F~Z-G-~A

3826

of Type VII Collagen

Metalloprotease Degradation

be related to excessive production of collagenase (13, 14, 20) our findings may provide a pathophysiologic link between the increased levels of the enzyme and thedecrease in anchoring fibrils. Furthermore, patients with RDEB frequently have a marked increase in blistering following a febrile episode. The significant temperature dependence of the cleavage of type VI1 collagen by these enzymes (Fig. 2) could also provide a plausible explanation for this phenomenon. The fact that extracellular neutral metalloproteases known to be present in connective tissue are capable of degrading anchoring fibril components seems significant, especially in light of the known elevations of collagenase in some disease situations. However, our current in vitro findings do not exclude the possibility that otherproteases may play a role in type VI1 collagen turnover. Because the globular portions of type VI1 collagen also contribute tothe structure of anchoring fibrils, anchoring plaques, and basement membranes (5,7), it is probable that proteases secreted by inflammatory infiltrates could also be destructive to anchoring fibrils, especially through attack on the globular domains of the molecule.

9. Pearson, R. W. (1962) J. Inuest. Dermatol. 39,551-575 and Wheeler, C.E., Jr. (1975) J. Inuest. 10. Briggaman, R.A., Derrnatol. 65, 203-211 11. Hashimoto, I., Schnyder, U. W., Anton-Lamprecht, I., GeddeDahl, T., and Ward, S. (1976) Arch. Dermatol. Res. 256, 137150 12. Tidman, M. J., and Eady, R. A. J. (1985) J . Invest. Dermatol. 84, 374-377 13. Eisen, A. Z. (1969) J . Inuest. Dermatol. 52, 449-453 14. Bauer, E. A., Gedde-Dahl, T., Jr., and Eisen, A. Z. (1977) J. Inuest. Dermatol. 68, 119-124 15. Gross. J.. and LaDiere.C.M. (1962) Proc.Natl.Acad.Sei. U. A: 48, 1014-1022 16. Weleus. H. G.. Jeffrev. J. J.. and Eisen. A. Z. 11981) . , J. Biol. C L m . 256,9511-9gl5 ' 17. Seltzer, J. L., Adams, S. A., Grant, G. A., and Eisen, A. Z. (1981) J. B i d . Chem. 256, 4662-4668 18. Collier, I. E., Wilhelm, S. M., Eisen, A. Z., Marmer, B. L., Grant,

s.

19. 20. 21.

Acknowledgements-We thank M. L. Eschbach, K. Akers, and S. White for excellent technical help. We thank Dr. Gregory Grant, Director of the Washington University Protein Chemistry Laboratory, for performing amino acid sequence analyses. REFERENCES 1. Bentz, H., Morris, N. P., Murray, L. W., Sakai, L. Y., Hollister, D. W., and Burgeson, R. E. (1983) Proc.Natl.Acud.Sci. U. 5'. A . 80,3168-3172 2. Burgeson, R. E., Morris, N. P., Murray, L.W., Duncan, K.G., Keene, D. R., and Sakai, L. Y. (1986) Ann. N . Y. Acad. Sci. 460,47-57 3. Morris, N. P., Keene, D. R., Glanville, R. W., and Burgeson, R. E. (1986) J. Biol. Chem. 261, 5638-5644 4. Briggaman, R. A,, Dalldorf, F. G., and Wheeler, C. E., Jr. (1971) J. Cell Biol. 51, 384-395 5. Lunstrum, G. P., Sakai, L. Y., Keene, D. R., Morris, N. P., and Burgeson, R. E. (1986) J. Biol. Chem. 261, 9042-9048 6. Sakai, L. Y., Keene, D. R., Morris, N. P., and Burgeson, R. E. (1986) J. Cell Biol. 103, 1577-1586 7. Keene, D. R., Sakai, L. Y., Lunstrum, G. P., Morris, N. P., and Burgeson, R. E. (1987) J.Cell Biol. 104,611-621 8. Gipson, I. K., Spurrmichaud, S. J., and Tisdale, A. S. (1987) Inuest. Ophthal. Visual Sci. 28,212-220

22. 23.

G.A., Seltzer, J. L., Kronberger, A., Chengshi, H., Bauer, E. A., and Goldberg, G . I. (1988) J . Biol. Chem. 263,6579-6587 Stricklin, G. P., Bauer, E. A., Jeffrey, J. J., and Eisen, A. Z. (1977) Biochemistry 16, T607-1615 Stricklin, G. P., Welgus, H. G., and Bauer, E. A. (1983) J. Clin. Inuest. 69, 1373-1383 Seltzer, J. L., Eschbach, M. L., and Eisen, A. Z. (1985) J. Chromatogr. 326, 147-155 Stricklin, G. P., Jeffrey, J. J., Roswit, W. T., and Eisen, A. 2. (1983) Biochemistry 22,61-68 Nagai, Y., Lapiere, C.M., and Gross, J. (1966) Biochemistry 5,

3123-3130 24. King, J., and Laemmli, U. K. (1971) J. Mol. Biol. 62,465-477 25. Aebersold, R. H., Teplow, D. B., Hood, L. E., and Kent, S. B. H. (1986) J. Biol. Chem. 261,4229-4238 26. Grant, G. A., Sacchettini, J. C., and Welgus, H. G. (1983) Biochemistry 22,354-358 27. Groves, W. E., Davis, F. C., and Sells, B. (1968) Anal. Biochem. 22,195-210 28. Cooper, T. W., Bauer, E. A., and Eisen, A. Z. (1983) Collagen Relat. Res. 3, 205-216 29. Welgus,H.G., Jeffrey, J. J., Stricklin, G. P., and Eisen, A. Z. (1982) J. Biol. Chem. 257, 11534-11539 30. Gross, J., Harper, E., Harris, E. D., McCroskery, P. A., Highberger, J. H., Corbett, C. A., and Kang, A. H. (1974) Biochern. Biophys. Res. Commun. 61,605-612 31. Weingarten, H., and Feder, J. (1986) Biochem.Biophys.Res. Cornrnun. 139,1184-1187 32. Lane, A. T., Goldsmith, L. A., McCoon, P. E., and Muhlbauer, J . E. (1985) Arch. Dermatol. Res. 277,499-501 33. Rusenko, K. W., Gammon, W. R., Fine, J. D., and Briggaman, R. A. (1988) Clin. Res. 36,691A