Structural Components and Characteristics of Reichert's Membrane

THE JOURNAL OF BIOLOGICAL CHEMISTRY Val. 256, No. 9, Issue of May 10, pp. 46544661, 1981 Printed in U.S.A.

Structural Components and Characteristics of Reichert’s Membrane, an Extra-embryonic Basement Membrane*

(Received for publication, October 23, 1980)

Karen Kaye Smith and Sidney Strick€and From The Rockefeller University, New York, New York 10021

This paper presents an analysis of mouse and rat Reichert’s membrane, a thick basement membrane formed between the trophoblast and parietal endoderm cells of early mammalian embryos. When analyzed by polyacrylamide gel electrophoresis, Reichert’s membranes from rat and mouse conceptuses appear simple and consist primarily of collagen and 4 noncollagenous glycoproteins of approximate M, = 415,000, 245,000, 170,000, and 50,000. The proteins at 415,000 and 245,000 are similar in molecular weight to laminin and are immunoprecipitated by anti-laminin antiserum. The protein at 170,000 is co-precipitated with laminin and represents a novel form of a laminin-like protein, whereas the protein at 50,000 is unrelated to laminin. Using metabolic labeling experiments, it is shown that parietal endoderm cells synthesize these components and incorporate them into the matrix in organ cultures but do not degrade Reichert’s membrane in the conditions used over a 6-day culture period. Evidence is presented that the high molecular weight collagenase- sensitive proteins (greater than 420,000) observed on polyacrylamide gel electrophoresis are due to lysyl oxidase-derived cross-links between Reichert’s membrane components.

The parietal yolk sac is the outermost extra-embryonic membrane of midgestation rodent embryos. It is a three- layered structure composed of trophoblast cells on the uterine side and parietal endoderm cells on the embryonic side, separated by a basement membrane known as Reichert’s membrane (see Ref. 1). This extracellular matrix is a proteinaceous structure containing collagen (2,3) and lacking proteoglycans (4, 5).

Situated as it is between the yolk sac cavity and the maternal blood sinuses of the trophoblast layer, RM’ is in an important position to act as a molecular sieve for the passage of macromolecules between the embryo and the mother. Many authors have observed that macromolecules pass from the maternal into the embryonic circulation through the inverted yolk sac placenta (parietal and visceral yolk sacs) rather than

* This work was supported by grants from the American Cancer Society (CD-48), the National Institute of Child Health and Human Development (HD-IIW), and the Rockefeller Foundation. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “adver- tisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

The abbreviations used are: RM, Reichert’s membrane; DME, Dulbecco’s modified Eagle’s medium; Hepes, 4-(2-hydroxyethyl)-l- piperazineethanesulfonic acid; PAGE, polyacrylamide gel electrophoresis; PE + RM, Reichert’s membrane with attached parietal endoderm cells after removal of trophoblast cells, SDS, sodium dodecyl sulfate.

the chorioallantoic placenta (6-10). Furthermore, Jollie has proposed that the only barrier to filtration of macromolecules across the parietal yolk sac exists at RM (11). Permeability studies with tracer molecules suggest that RM has a rather large pore size, since y-globulins (6), ferritin (ll), and horse- radish peroxidase (8) can all pass across RM. Structural characteristics which affect fitration properties are therefore likely to be important in the nutrition of the embryo.

Apart from its importance in embryogenesis, RM has many properties which make it useful for the study of basement membranes in general. First, it is bounded by cell monolayers on both sides, which makes the membrane easy to isolate by a simple dissection and to prepare in an acellular form. Sec- ond, since the parietal endoderm cells which synthesize RM (2, 12) can be maintained in culture for several days and remain biosynthetically active in basement membrane pro- duction (13, 14), this system can provide information about both the molecules present and the regulation of their biosyn- thesis. Third, RM expands as the embryo grows and can therefore be used for studying growth and restructuring of basement membranes during tissue remodeling. Clark et al. (5) have shown that the surface area of the capsular portion of rat RM increases from 100 to 1000 mm2 in 6 days of gestation. The thickness also increases until day 15 of gestation when it starts to decrease markedly until the capsular parietal yolk sac ruptures on day 18 (15). The process of rupture itself is another example of tissue remodeling in this system.

RM is also of interest in developmental biology, since the parietal endoderm cells that synthesize this basement membrane differentiate from the inner cell mass very early after the development of the blastocyst (1). Likewise, teratocarcinoma stem cells can differentiate in culture into parietal endoderm-like cells spontaneously or after induction, and RM proteins have been used as a marker for this process (16-21).

In this work, various aspects of RM structure have been studied and characterized. First, the major glycoprotein components of mouse and rat RM are described. These fall into two classes of proteins, collagen and noncollagen glycoproteins. By metabolic labeling of RM, it is shown that the parietal endoderm cells synthesize the basement membrane components but do not degrade them once the basement membrane has been deposited. Finally, the existence of covalently cross-linked aggregates of basement membrane proteins derived from the action of lysyl oxidase is described.

EXPERIMENTAL PROCEDURES

Materials Mice used were ICR females from Flow Labs, McLean, VA and

either SJL males from Jackson Labs, Bar Harbor, MA or NCR males from Rockefeller University. Females for mating were superovulated (22) using 2 units of pregnant mare serum gonadotropin and 2 units

4654

Structural Components and Characteristics of Reichert's Membrane 4655

of human chorionic gonadotropin. Pregnant rats were purchased from zivic Miller, Allison Park, PA. Reagents were purchased as follows: Triton x-100, phenylmethylsulfonyl fluoride, ethylenediaminetetra- acetic acid, N-ethylmaleimide, a,d-dipyridyl, iodoacetamide, NaBH4, glycerol, periodic acid, bovine serum albumin (crystalline and lyophilized), basic Fuchsin, 4-(2-hydroxyethyl)-l-piperazineethanesulfonic acid, L-proline, trans-4-hydroxy-~-proline, glycine, dimethylamino- naphthalene-1-sulfonyl chloride, putrescine, P-aminoproprionitrile fu- marate, and tannic acid from Sigma; fluorescamine from Roche Diagnostics, Nutley, NJ; sodium dodecyl sulfate from GS-BDH, Carle Place, NY; dithiothreitol from Aldrich hydrochloric acid, trichloroacetic acid, glacial acetic acid, and methanol from Fisher; benzene and acetone from Mallinckrodt; sodium arsenite from Baker; acryl- amide, bisacrylamide, and Coomassie brilliant blue from Eastman; N,N,N',N'-tetramethylethylenediamine and ammonium persulfate from Bio-Rad; protein A-Sepharose and protein molecular weight standards from Pharmacia; [U-'4C]proline (250-260 mCi/mmol) from Schwarz/Mann; [35S]methionine (600-1400 Ci/mmol), I4C-methyl-

hydro~y[2-'~C]proline (18.7 mCi/mmol), and NCS Tissue Solubilizer ated protein molecular weight markers for gel electrophoresis, L-4-

from Amersham/Searle; [1,4-14C]putrescine (89.5 mCi/mmol), ['HI- NaBH4 (250 mCi/mmol), En'Hance, Aquasol, and Liquifluor from New England Nuclear; Dulbecco's modified Eagles' medium (4.5 g of glucose/liter), penicillin and streptomycin (lyophilized), and ascorbic acid from GIBCO; fetal bovine serum from Reheis Chemical Co. (Armour Pharmaceuticals), Kankakee, IL; trypsin from Nutritional Biochemicals; tissue culture dishes from Falcon; chromatographically purified bacterial collagenase from Worthington (390-500 units/mg); and polyamide sheets for thin layer-amino acid analysis from the Cheng Chin Trading Co. Ltd., Taipei, Taiwan. Sheep anti-laminin antiserum was generously provided by Dr. George Martin of the National Institutes of Health.

Methods Preparation of Reichert's Membranes with and without Parietal

Endoderm Cells-Conceptuses of mice I2 days in gestation (the day of the seminal plug is designated day 1) or of rats 14 days in gestation (positive vaginal smear is day 1) were dissected in sterile phosphate- buffered saline (23) to obtain the parietal yolk sacs. Trophoblast was separated from the RM by dissection without the use of enzymes. RM with parietal endoderm attached is referred to as PE + RM. To obtain acellular RM, the PE + RM were washed in 2 ml of 0.1% Triton X-100 in water for 15 min in a 35-mm Petri dish, and then transferred with forceps to 1.5 mM EDTA for 15 min. The following were added to all washes as protease inhibitors: 2 mM phenylmethylsulfonyl fluoride, 2.5 mM N-ethylmaleimide, 3.0 mM EDTA, and 0.3 mM a,a'-dipyridyl. The samples were shaken on a rotator for 15 min and triturated several times during the procedure using a Pasteur pipette. This method removes most of the endoderm cells as judged by visual inspection. Although molecules extracted in the wash solu- tions were not analyzed, RM components exhibited a limited extract- ability in Triton, since RM metabolically labeled with ["'S]methionine and incubated for 12 h in 0.1% Triton at 37°C did not release significant amounts of :jSS-labeled protein. Samples were then frozen and stored in 1.5-ml Eppendorf tubes for several hours to several days. After thawing, samples were resuspended in approximately 0.5 nd of 0.05 M NaPO,, pH 7.4, and to rinse the pellet, were centrifuged and resuspended.

TO solubilize the samples for polyacrylamide gel electrophoresis, a volume of 0.05 M NaP04 buffer, pH 7.4, 0.5% SDS (w/v), 50 m~ dithiothreitol, and 2 m~ phenylmethylsulfonyl fluoride was added to give a final protein concentration in the range of 0.25-1.0 mg/ml (approximately 40 pl/mouse RM and 80 pl/rat RM). The samples were placed in a boiling water bath for 15-20 min, during which time most of the solid material had dissolved. Insoluble residue was removed by centrifugation in a Brinkman Eppendorf centrihge Model 5412 for 2 min. Aliquots of the supernatant were removed for protein determination (see below). This method of solubilization dissolves over 98% of the RM proteins as determined by the amount of insoluble 35S-labeled residue from RM which had been metabolically labeled with [35S]methionine. Samples for PAGE analysis were then adjusted to 2.5% SDS and 0.1 M dithiothreitol and boiled for 3 min. Protein determinations were performed by the method of Lowry et al. (24) on aliquots which were precipitated in 10% trichloroacetic acid for 10 min at 0 "C, pelleted by centrifugation, washed in 10% trichloroacetic acid, re-pelleted, and finally redissolved in 0.1 N NaOH prior to addition to the Lowry assay. If protein determinations were not

performed, as for example with radioactively labeled RM, samples were dissolved directly in 2.5% SDS, 0.1 M dithiothreitol, and 2 mM phenylmethylsulfonyl fluoride (sample buffer).

For collagenase digestions of RM shown in Figs. 1 and 2 (see following for a description of the collagenase), acellular RM was incubated with 8.5 pg/ml of purified bacterial collagenase, 2.5 m~ N- ethylmaleimide, 0.5 mM CaCI2, and 2 mM phenylmethylsulfonyl fluoride in 0.05 M NaP04 buffer, pH 7.4, before the addition of solubilizing agents. Digestion was at 37 "C for 7 h.

Biosynthetically Labeled RM and Culture Methods-PE + RM were cultured in Dulbecco's modified Eagle's Medium supplemented with 5% dialyzed fetal bovine serum, 50 pg/ml of glycine, and 50 p g / ml of ascorbic acjd. In all cases, DME contained penicillin (100 units/ ml) and streptomycin (100 pg/ml). Routinely, several (5-10 rat and 10-20 mouse) PE + RM were cultured in a 35-mm dish in 2 ml of medium. Samples were preincubated for 1 h at which time the medium was changed and 5 pCi/ml of [I4C]proline or 10 pCi/ml of ["S]methionine were added. In the latter case DME minus methionine containing 40 pg/ml of proline was used. The samples were incubated for 24 h at 37 "C in a humid atmosphere of 5% COz-95W air. They were prepared for analysis by rinsing in DME, and then remov- ing parietal endoderm cells and dissolving the membrane as described. After solubilization, aliquots were precipitated with trichloroacetic acid to determine the incorporation of radioactive precursors into RM proteins. Precipitation with trichloroacetic acid was performed by spotting an aliquot on fdter paper, drying, washing twice in approximately 10 ml of 10% trichloroacetic acid/fdter disc, rinsing in ethanol, drying, and counting in Liquifluor in a Packard scintillation counter Model 3385.

For [I4C]putrescine incorporation, 8 mouse PE + RM were incubated in 1 ml of DME containing 0.5 mg/ml of bovine serum albumin, 50 pg/ml of glycine, 50 p g / d of ascorbic acid, and 10 pCi of ["CI- putrescine for 12 h at 37 "C. RM were then washed for 20 min in phosphate-buffered saline (23) containing 100 pg/ml of unlabeled putrescine to remove nonspecifically bound material. Membranes were frozen, thawed, and washed to remove parietal endoderm as described except that 100 pg/ml of putrescine were present in all washes. The sample was suspended in 200 p1 of 0.05 M NaP04, pH 7.4, and treated with 4 pmol of NaBH4 at room temperature for 1 h. RM were washed by centrifugation and resuspension, and then dissolved in sample buffer. Aliquots were removed for determination of protein content and of "C-protein precipitable in trichloroacetic acid.

To study the effects of @-aminoproprionitrile, 5 mouse PE + RM were incubated in each of two 16-mm cell culture wells containing 500 p1 of DME minus methionine supplemented with 50 pg/ml of glycine, 50 pg/ml of ascorbic acid, 40 pg/ml of proline, and 10 pCi/ml of ["S]methionine; 5 pg of @-aminoproprionitrile was added to one well. Cells were incubated for 5 h at 37 "C before preparation of RM for gel analysis.

For studies on RM stability after deposition, PE + RM were dissected in sterile Simms balanced salt solution (25) and cultured in 16-mm wells (four 11-day mouse or three 13-day rat RM/well) in 150 pl of DME:Simms balanced salt so1ution:fetal bovine serum (2:2:1) with 5 pCi/ml of [I4C]proline and 50 p g / d of ascorbic acid. Cultures were labeled for 24 h, then washed for 1 h in DME:Simms balanced salt solution as above but without ,['4C]proline and with 0.28 mg/ml of proline, and incubated in this medium over the next 5 days with a medium change every 24 h. After the 1-h preincubation period and every 24 h thereafter, samples from 2 or 3 wells were removed, washed free of parietal endoderm cells as described above, and the acellular RM processed for the amount of collagen (see following).

PAGE Analysis-Gradient slab gels were made according to the methods described by Laemmli (26) (3.5-11%), and for molecular weight determinations gels prepared by the method of Neville (27) (4-11%) were also used. Electrophoresis was performed at 8 mA constant current overnight or 25 mA for 4-5 h. The gels were stained for protein with 0.25% Coomassie brilliant blue in an aqueous solution of 40% methanol/8% acetic acid and for glycoproteins with periodic acid-Schiff stain (28). Gels containing radioactively labeled samples were analyzed by autoradiography using flashed Kodak SB-5 x-ray film (29). The gel in Fig. 4 was analyzed by fluorography at -80 "C after impregnating the gel with En'Hance. The lengths of exposures of films are given in the figure legends. The gels in Figs. 3 and 5 were analyzed using a position-sensitive @-detector (30) equipped with an Ortec Model 6220 pulse height analyzer with a 4-chamber memory.

Immunoprecipitation-Immunoprecipitation was performed by a modification of the method of Kessler (31, 32). Antigen was prepared from 15 mouse RM which had been biosynthetically labeled with 10

pCi/ml of ["S]methionine for 5 h and then stirred for 11 h in 0.05 M Tris, pH 7.4, with 0.5 M NaCI, at 4 "C, to extract RM components. The soluble extract was treated with Collagenase a t 37 "C for 6 h. Antiserum (5 pl) and antigen (50 pl containing 10,000 trichloroacetic acid-precipitable cpm) were added to a final volume of 250 p1 of 2% Triton X-100 and 0.5% SDS in 0.05 M NaP04, pH 7.4. Carrier bovine serum albumin was added to a final concentration of 1 mg/ml. After approximately 16 h at 4 "C, 2 p1 of protein A-Sepharose was added per pl of mtiserum. The suspension was gently shaken for 30 min at room temperature. The protein-A-Sephirose was then pelleted by centrifugation in a Brinkman Eppendorf centrifuge for 2 min and washed 4 times in the buffer and once in 0.05 M NaP04, pH 7.4. The immunoprecipitate was finally resuspended in sample buffer with 15% glycerol (v/v) and placed in a boiling water bath for 3 min. The protein A was removed by centrifugation and the supernatant analyzed by PAGE. An aliquot was removed for the quantitation of 3sS- protein precipitable in trichloroacetic acid.

Determination of the Fraction of Collagen in RM-Bacterial collagenase was prepared as described by Peterkofsky and Diegel- mann (33) using Sephadex G-150 instead of G-200. It is important that the enzyme be free of contaminating proteases which contribute to protein degradation; this was measured by the release of Iz5I-fibrin- degradation products from Iz5I-fibrin on 16-mm plates (34, 35). Col- lagenase (8 pg) was added to 250 p1 of 0.12 M Hepes, pH 7.2,0.032 M NaCI, 2.5 mM N-ethylmaleimide, and 0.5 m~ CaCIz in '2sII-fibrin- coated 16-mm wells and incubated at 37 "C. After 24 h, aliquots were removed and the radioactivity measured in a Packard 5160 gamma counter; in the absence of collagenase 3.4% of total Iz5I cpm solubilized by 0.25% trypsin in 0.12 M Hepes buffer was released and in the presence of collagenase, 3.9%, indicating a very low level of protease contamination. The enzyme is further characterized in Table I. The time course of the collagenase reaction on V - R M was determined from the release of I4C-peptides into trichloroacetic acid-soluble material; the release of '"C-peptides in the absence of collagenase is low and constant over 48 h, and the reaction in the presence of enzyme reached a plateau by 12 h and was not further stimulated by addition of fresh enzyme at 36 h. The insoluble material left after 24 h of collagenase digestion was analyzed for 4-hydroxyproline to measure the remaining undigested collagen (Table I). The results show that after digestion no significant amounts of 4-hydroxyproline remain trichloroacetic acid-insoluble, suggesting that almost all of the collagen has been degraded.

The fraction of collagen in RM (Tables I and 11) was determined by a collagenase assay based on that described by Diegelmann and Peterkofsky (36). Acellular RM metabolically labeled with ['"Clpro- line was transferred to Eppendorf tubes containing 250 p1 of 0.12 M Hepes, pH 7.2,0.032 M NaC1,2.5 m~ N-ethylmaleimide, 0.5 m~ CaClz, and 8 pg of collagenase, and incubated at 37 "C with vigorous shaking. As shown by the time course of the reaction in Table I, 24 h was sufficient for completion of the collagenase reaction, and this is the length of time used for the experiments in Table 11. After digestion, the reaction was stopped by adjusting the solution to 10% trichloroacetic acid-0.5% tannic acid. The fraction of 14C-proteins digested by the collagenase was determined as follows: trichloroacetic acid-insoluble material (TCA,) was collected on a Millipore filter (HA, 0.45 pm), and the flow-through from the filtration was collected directly into scintillation vials for determination of trichloroacetic acid-soluble material (TCA,). Filters were washed with 1 ml of 10% trichloroacetic acid/0.5% tannic acid, then 1 ml of 5% trichloroacetic acid/0.25% tannic acid the washes were monitored for I4C-peptides by scintillation counting, and these values were added to TC&. Samples were counted in Aquasol. Insoluble RM (A) remaining on the filter papers was transferred with forceps to 100 pl of NCS tissue solubilizer in counting vials and incubated at 37 "C until dissolved, and fdter papers (B) were dried under a heat lamp before being counted; TCA, is the sum of A + B. The percentage of collagen was determined using the following formula (36):

% collagen = 100 X TCA,

(TCA, x 5.4) i TCA,

TCAi is corrected by a factor of 5.4 due to the enrichment of collagen for proline by that amount (36).

Protein Determination-The amount of cell protein in the mouse

the Triton extract of washed RM by the fluorescamine method (37). organ cultures in Table I1 was determined by measuring protein in

Amino Acid Analysis-For the determination of lysyl oxidase- derived cross-links, 82 rat RM were dissected from 14-day conceptuses

TABLE I Characteristics of the collagenase reaction used to determine the

fraction of collagen in RM

1 4 ~ cpm re- "C cpm remaining in RM

genase" PROh OH-PRO' Time Collagenase leased by colla- present as

h % total "C cpm % % 2 1

+ 8 12 - 2

+ 60 24 - 2

i 60 51.5 48.9 95.6 4.4

48 2 + 62

-

-

a The time courSe of release of [14C]proline and ['4C]hydroxyproline from mouse RM was measured using RM labeled in organ culture for 4 days and fed every 24 h with [I4C]proline as described under "Experimental Procedures" for the studies on RM stability. Acellular RM were then prepared and incubated in the absence or presence of collagenase for the times indicated, and then assayed for trichloroacetic acid-insoluble I4C-protein, with reaction conditions exactly as described under "Experimental Procedures." The addition of fresh collagenase after 36 h did not affect the percentage of '"C-protein solubilized.

Eighteen mouse RM were biosynthetically labeled with ['"I- proline for 24 h, incubated + or - collagenase, and precipitated in trichloroacetic acid according to the procedures described for deter- mining the fraction of collagen in RM (Experimental Procedures). The trichloroacetic acid-insoluble material was washed three times with ether and hydrolyzed in 6 N HC1 in uacuo for 23 h at 105OC. Amino acids were dansylated and analyzed by two-dimensional thin layer chromatography on polyamide sheets (4 X 4 cm) exactly as described by Woods and Wang (55). Positions of migration of proline and 4-hydroxyproline were determined using dansylated ['4C]proline (5 X lo5 cpm/pg) and 4-hydroxy-[2-I4C]proline (lo5 cpm/pg) standards. Dansylated amino acids (500-3000 cpm) from hydrolyzed '"C- RM were run on the polyamide sheets with 0.02 pg of dansylated proline and 0.04 pg of dansylated 4-hydroxyproline standards. Dan- sylated amino acids were detected by ultraviolet irradiation, and spots corresponding to the standards were cut out of the polyamide sheets, dried, and counted in a toluene-based scintillation mixture in a Packard 3383 scintillation counter; results are expressed as the percentage of total radioactivity co-migrating with proline and hydroxyproline standards. The rest of the sheets were also counted but contained no cpm above background (10-25 cpm).

TABLE I1 Stability of labeled RM in organ culture

PE + RM were labeled with [I4C]proline for 24 h, cultured in unlabeled medium for the times shown in the left column, and analyzed for the fraction of RM proteins degraded by collagenase (expressed as a percentage) as described under "Experimental Pro- cedures.'' The background of the assay (release of I4C-peptides in the absence of collagenase) was negligible.

Days of chase Collagen k S.E. Cell protein

% cpm/mg Mouse

1 2 3 4 5

Rat

1 3 5 6

0 23.8 f 0.8 (2)" 1250 & 230 (3)" 1050 f 140 (3) 1350 f 140 (3) 1015 f 315 (2) 960 f 435 (3)

23.5 f 0.4 (2) 26.8 f 0.0 (2) 26.0 f 3.0 (2) 25.2 f 2.6 (2) 25.1 rfr 0.9 (2) 1250 f 15 (2)

0 26.5 f 2.0 (2)" 27.7 f 0.7 (3) 26.0 f 3.5 (2) 27.6 f 1.0 (2) 28.2 & 0.6 (2)

Numbers in parentheses represent the number of samples used in the determination.

Structural Components and Characteristics of Reichert’s Membrane 4657

and washed free of cells as described. RM were suspended in 1 ml of 0.2 M NaP04, pH 7.4, and treated with 630 pg of [“HlNaBH, dissolved in dimethylformamide (previously dried over sodium sulfate) at 3 mg/ ml. The [:’H]NaBH4 was added in 3 equal aliquots 3 min apart, with constant stirring. The solution was then adjusted to pH 3 with 10% acetic acid to terminate the reaction. The sample was rinsed exhaus- tively in 0.2 M NaI’04, pH 7.4, and then twice in Hz0 before being lyophilized. Dried RM were hydrolyzed in 3 N toluene sulfonic acid for 24 h a t 100 “C. The sample was then filtered through a sintered glass filter and applied to the amino acid analyzer developed by the method of Mechanic (38) for the separation of lysyl oxidase-derived cross-links. The effluent was monitored for radioactivity and the elution times of radioactive products were compared to those of known standards (38).

RESULTS

Protein Components of RM-The protein components of RM have been analyzed by a variety of methods. Initially, RM were dissected from embryos and studied directly by PAGE; Fig. 1 shows the proteins from rat and mouse RM soluble in denaturing and reducing conditions. Each sample was incubated in the absence (-) or presence (+) of collagenase to distinguish collagen from other proteins. The collagenase-insensitive molecules have M , = approximately 410,000, 240,000, 165,000, and 50,000 in mouse and 420,000, 250,000, 170,000, and 50,000 in rat RM. There are two major collagenase-sensitive proteins, a t 220,000 and 370,000.’ The collagen a t 220,000 is resolved in various preparations as either a doublet or a triplet.

PE + RM were also cultured in the presence of radioactive amino acids, and the metabolically labeled proteins of the acellular RM were subsequently analyzed by PAGE and autoradiography (Fig. 2). The pattern of major proteins generated (lanes I and 3) is seemingly more complex than that shown in Fig. 1; however, after collagenase treatment, the complex array of high molecular weight proteins (greater than 420,000) disappears, resulting in four principal collagenase- insensitive molecules designated A, B, C, and D (lanes 2 and 4), which correspond in molecular weight to those in Fig. 1. Collagenase-sensitive proteins are marked by arrows in Fig. 2 and also correspond in molecular weight in some instances to the collagen in unlabeled RM. Evidence is presented below that the high molecular weight, collagenase-sensitive proteins result from lysyl oxidase-derived cross-links. Hence, the major glycoproteins present in the basement membrane are all synthesized and deposited in the RM by cultured parietal endoderm cells, consistent with the reports that these cells synthesize RM (2, 12). In some cases there is also a diffuse protein migrating above glycoprotein D (see Fig. 2B). I t is not consid- ered as a major component of RM because it is present in low amounts and is not detected in all preparations; it may represent a degradation product or cellular contamination. The intensities of bands below 50,000 are also variable.

Recently, the isolation and characterization of two high molecular weight proteins from basement membranes have been reported (42,43). These proteins, called GP1 and 2 (42) and laminin (43), migrate with the same approximate molecular weights as glycoproteins A and B described previously. We performed immunoprecipitation of mouse RM with anti- laminin antiserum to determine the immunological cross-reactivity of RM proteins with laminin. RM was first extracted with 0.5 M NaCl to solubilize the glycoproteins without de-

‘The apparent molecular weight for collagen reported here is higher than that reported by others (3,39,40), which may be due to the fact that we have used globular standards for molecular weight determination in order to characterize the noncollagenous components. Furthmayr and Timpl (41) have reported that collagen migrates a t an anomalously high molecular weight in PAGE when compared to globular protein standards.

A. RAT E. MOUSE C. MOUSE I

- TOP -

- ’67kA 60k

- + - + - c

FIG. 1. Polyacrylamide gel electrophoretic analysis of protein components of Reichert’s membrane. Acellular rat ( A ) and mouse ( B and C ) HM were prepared as described under “Experimen- tal Procedures.” Thirty-one pg (rat) and 40 pg (mouse) of protein were loaded/lane. Symbols under the lanes indicate incubation with (+) or without (-) collagenase for 7 h at 37 “C. Gels A and B are Coomassie blue-stained, C is periodic acid-Schiff-stained to indicate glycoproteins. The minor Coomassie blue-stained band in A (+) and B (+) a t M, = 100,OOO is due to the added collagenase. The molecular weight markers are, from high to low molecular weight, thyroglobulin, ferritin half-subunit, phosphorylase b, albumin, catalase, ovalbumin, and lactate dehydrogenase.

naturation or reduction, since the antiserum does not precip- itate laminin which is reduced (43). The extract was treated with collagenase and immunoprecipitated with sheep anti- laminin antiserum as described under “Experimental Proce- dures.” The results shown in Fig. 3 indicate that anti-laminin antiserum recognized and precipitated glycoproteins A, B, and C but not D. Due to their similar molecular weights and immunological cross-reactivity, glycoproteins A and B appear to be the same as laminin and will henceforth be referred to as such.

Studies on RM Stability in the Presence of Endoderm Cells-In light of the fact that parietal endoderm cells synthesize RM, and in order to investigate the processes by which RM expands and eventually ruptures, we were interested in studying the fate of the newly synthesized RM components. Of additional interest in this regard is the fact that in primary culture (44) and in organ culture,” parietal endoderm cells secrete a protease, plasminogen activator, which is implicated in extracellular proteolysis in other systems (45). We have therefore performed pulse-chase experiments to determine whether RM is degraded following its deposition by parietal endoderm cells. PE + RM organ cultures were incubated for 24 h in the presence of [I4C]proline to label RM and then cultured for 1 to 5 days in a medium without [I4C]proline, with a medium change every 24 h (see “Experimental Proce- dures”). PE + RM were removed from culture throughout the chase and acellular RM were prepared and analyzed for the fraction of collagen present. In this study, the amount of collagen was determined using a collagenase assay which measures the amount of I4C-peptides released by bacterial collagenase from [14C]proline-labeled RM (36). The data in Table I and “Experimental Procedures” show that there is no contaminating proteolytic activity in the collagenase reaction

’’ K. H. Marotti, unpublished observations.

4658 Structural Components and Characteristics of Reichert’s Membrane

A ‘“C-proline

” -

330K-

--

t + ~ + FIG. 2. Polyacrylamide gel electrophoretic analysis of bio-

synthetically labeled Reichert’s membrane. Rat and mouse RM were biosynthetically labeled in culture with [‘4C]proline (A) or [““Slmethionine (B) for 24 h. Acellular KM were then prepared for PAGE analysis. The radioactivity (cpm) of the trichloroacetic acid- precipitable proteins applied was in A: 1, 19,ooO. 2, 7,500; 3, 16,ooO, and 4, 7,ooO; and in B: 1, and 3, 20,000; 2, and 4, 12,000. The autoradiogram of A was exposed for 11 days and of B for 6 days. The positions of the major glycoproteins are indicated to the sides of A and B: collagenase-sensitive proteins are marked by arrows and collagenase-insensitive proteins are labeled A through D in order of decreasing molecular weight. The molecular weight markers are the same as in Fig. 1.

and that the reaction goes to completion after 24 h of incubation with ‘%-RM.

Our studies into the stability of RM components are similar to turnover experiments by Minor et al. (14) with the following modifications. One, measurement of the percentage of collagen in RM was performed on acellular RM rather than PE + RM to ensure that the measurements reflected events in the RM rather than the cells. Two, the culture method differed from that of Minor et al. in that the PE + RM was floated in medium as a monolayer rather than grown on agar as an organ culture.

The results in Table II indicate that the percentage of collagen remained constant in rat and mouse RM and, furthermore, there was no significant turnover of protein in mouse RM since the cpm/mg of celI protein remained constant. Thus RM degradation did not occur in the presence of metabolically active parietal endoderm cells over a 6-day culture period. PE + RM were labeled and chased in the presence of serum which contained plasminogen. However, under these conditions neither the endoderm cells nor the plasminogen activator catalyzed the breakdown of collagen or RM glycoproteins.

Minor et al. (14) have reported a significant turnover of proteins in organ cultures of rat PE + RM in the presence or absence of trophoblast cells over a 6-day culture period. Both collagen and noncollagen proteins were degraded in their experiments, and the rate of noncollagen protein degradation was much greater such that the percentage of collagen in- creased with time. The rate of collagen degradation was accelerated in the presence of trophoblast. These differences from the data reported here may be explained by the fact that Minor et al. were studying protein turnover in RM with parietal endoderm cells present; therefore a significant portion of the degradation they observed could have been intracellu-

lar. The experiments reported in Table II were performed on acelhdar RM prepared from PE + RM at the end of each chase period and therefore should reflect the fate of proteins laid down in RM following the 24-h labeling period. Our results indicate that this RM was stable in the presence of metabolically active parietal endoderm cells in the conditions of culture used. If degradation does exist as a mechanism for RM growth and restructuring, it may depend on the presence of other components of the embryo such as the trophoblast.

Cross-links-The presence of a complex array of very high molecular weight (greater than 420,000) collagenase-sensitive proteins (see Fig. 2) suggested that the RM proteins were being covalently cross-linked after synthesis. There are at least two mechanisms by which this could occur. One is that active transglutaminase (Factor XIIIa) was present and was diffusing into the RM. However, transglutaminase from the fetal bovine serum in the culture medium was not responsible for the presence of the high molecular weight proteins since culture in the absence of serum resulted in the same protein profile (data not shown). The other possibility is that lysyl oxidase was present. Lysyl oxidase is found in connective tissue and is involved in collagen cross-linking (for a review see Ref. 46). This enzyme catalyzes the oxidative deamination of the e-amino group of lysine yielding an aldehyde which is then available for reaction with amines (Schiff base formation) or with other aldehydes (aldol condensation).

Since both transglutaminase and lysyl oxidase could cata- lyze amine incorporation into RM, we incubated [%]putres-

a

C‘---

d Y-L Frc. 3. Immunoprecipitation of Reichert’s membrane pro-

teins with anti-laminin antiserum. The soluble salt extract from biosynthetically labeled mouse RM was immunoprecipitated with a l/50 dilution of either normal rabbit serum or sheep anti-laminin antiserum as described under “Experimental Procedures.” The gel was analyzed using a position-sensitive P-detector (Ref. 30 and “Ex- perimental Procedures”). The horizontal axis represents the relative positions of radioactivity along the gel and the uerlical axis the number of events (counts) in each channel. The figure is a photograph of the 4-chamber pulse height analyzer screen. For display purposes, a line has been drawn through the digital values generated by the pulse height analyzer. The top (left) and bottom (right) of the gel are marked by slashes above lane a. Samples in lane a are “‘C-methvl- ated molecular weight standards; they are from the left: myosin (200,000), phosphorylase (100,090 and 92,500). albumin (67,009). ovalbumin (46,000), carbonic anhydrase (30,000), and lysozyme (14,300). Lane b is [““Slmethionine-labeled RM (2,506 cpm loaded) prepared as described under “Experimental Procedures.” The major peak at approximately 200,000 corresponds to collagen, and the positions of glycoproteins A through D are marked. Lane c is the immunoprecipitate of 10,009 cpm of RM extract treated with an irrelevant serum (rabbit nonimmune). Lane d is a parallel sample immunoprecipitated with anti-laminin antiserum. The peaks from left to right, identified both by mobility and by their co-migration with radioactive and Coomassie blue-stained protein in b, correspond to glycoproteins A. B and C. In c, 3.2% of the total trichloroacetic acid-precipitable cpm were immunoprecipitated and in d, 24%. Data were collected from each lane for the following times tin minutes): a. 15: b. 60: c. 70; and d, 70.

Structural Components and Characteristics of Reichert’s Membrane 4659

A B A B - A B

I .oo- 0.46-

0.96 -

+ - P-APN IOpg /m l

FIG. 4 (lefl). Incorporation of [“C]putrescine into Reichert’s membrane. Mouse PE + RM were incubated with [l4C]putrescine as described under “Experimental Procedures.” Lane A is 1000 cpm of [‘4C]putrescine-labeled RM. Lane B is 3000 cpm of RM biosynthetically labeled with [14C]proline as in Fig. 2. The heavily labeled doublet is the 220,000-collagen doublet. The labeling pattern in lane B reflects the preferential uptake of proline into collagen and does not reflect the relative amounts of RM proteins present (compare the patterns of Fig. 1 with Fig. 2). Lane A was exposed for 1 month and lane B for 10 days.

FIG. 5 (right). The effects of &aminoproprionitrile on the gel profile of [S6S]methionine-labeled Reichert’s membrane. HM + PE were incubated in serum-free conditions for 5 h in the presence or absence of P-aminoproprionitde (B-APN) (lOpg/ml) and prepared for PAGE analysis as described under “Experimental Pro- cedures.” A total of 10,000 trichloroacetic acid-precipitable cpm of protein was applied to each lane; the autoradiogram was exposed for 5 days. To quantitate the amounts of radioactivity in various regions of the gel, lanes A and B were analyzed on a p-sensitive detector (30) until approximately 39,000 /3 particles had been recorded and localized from each lane. The total number of events (p particles) in the peaks in regions of interest were determined by summing the events in the pulse height analyzer channels spanning the peaks. The values for each region of the two lanes are expressed as a ratio (A/B) to the left of the autoradiogram. The peak area was corrected for the amounts of radioactive background in nearby regions of the gel by subtracting the background from the number of events in each channel. There were two protein bands which were sensitive to the presence of /3- aminoproprionitrile (marked by arrows to the right of the autoradiogram), whereas glycoproteins A and C had ratios of 1.0 and 0.96, respectively, and served as internal standards.

Derived from

norleucine condensation aldol

Schiff base oroducl cross-links 1

Froction number

FIG. 6. Radioactive elution profile after cation exchange chromatography of ptoluenesulfonic acid hydrolysate of [SH]NaBH,-treated rat RM. The “H-hydrolysate was chromato- graphed on column 1 as described by Mechanic (38). The identity of hydroxynorleucine was established by i ts co-elution with dentin hydroxynorleucine on column 1. On rechromatography, the radioactive peak in fractions 150-160 eluted with the acid-hydrolysis breakdown product of the aldol.“ The peaks in fractions 250-300 elute in the region of reduced Schiff base cross-links. The presence of hydroxylysinonorleucine was established by rechromatography of a pool of peaks in this region on the basic column designed for resolution of Schiff base cross-links (38); a radioactive peak from rat RM eluted at the same position as hydroxylysinonorleucine from sponge collagen. These analyses were performed by Lila Graham and Dr. G . Mechanic.

cine (1,4-diaminobutane) with PE + RM in the absence of serum for 12 h, treated PE + RM with N a B a to reduce Schiff bases, and analyzed the RM for [‘4C]putrescine incorporation. The acellular RM contained a total of 3,900 cpm in 60 pg of protein, or 65 cpm/pg. The results in Fig. 4 of PAGE analysis of the labeled RM indicate that radioactivity was incorporated specifically into collagen. Since the RM was washed a t each step in 100 p g / d of unlabeled putrescine, and since the [‘4C]putrescine co-migrates with only collagen in the presence of SDS, the co-migration probably represents covalent incorporation rather than nonspecific adsorption. If the covalent bond is a Schiff base, it could not have formed with the aldehydes of galactose or glucose, the carbohydrates of collagen, because it has been shown that the C-1 of each sugar in basement membrane collagen is involved in a glycosidic linkage (47). The fact that all of the detectable radioactivity was incorporated into collagen supports the possibility that the enzyme responsible was lysyl oxidase since that enzyme is specific for collagen.

We have tested the effect of P-aminoproprionitrile, a well described inhibitor of lysyl oxidase (see Ref. 46), on the gel profile of RM proteins. When PE + RM were incubated with 10 p g / d of P-aminoproprionitrile for 5 h in the presence of [?3]methionine, the amount of radioactivity in the high molecular weight collagenase-sensitive proteins was decreased (Fig. 5). Thus the presence of an inhibitor of lysyl oxidase affected the fate of newly synthesized collagen to the extent that less was incorporated into the high molecular weight

Rat RM was analyzed chemically to determine the presence of lysyl oxidase-derived aldehydes and aldehyde cross-links. Acellular rat RM treated with [:‘H]NaBH4 was prepared as described under “Experimental Procedures” and analyzed for amino acids and derivatives using a column system described by Mechanic (38). The elution profile of radioactive peaks is shown in Fig. 6 and supports the conclusion that RM proteins

A similar conclusion has been reported by others in abstract form (48).

4660 Structural Components and Characteristics of Reichert’s Membrane

were cross-linked by a mechanism involving lysyl oxidase. For example, hydroxynorleucine and Schiff base cross-links were present, and when a pool of peaks in the region of Schiff base cross-links was rechromatographed on a basic column (38), a radioactive peak eluted in the position of hydroxylysinonorleucine. In addition there was radioactive material which co- migrated with the acid-hydrolysis breakdown product of the aldol5 formed between lysine-derived aldehydes, the identity of which was also confirmed by rechromatography. These derivatives all result from the action of lysyl oxidase on lysine. It cannot be determined from these experiments whether the lysyl oxidase was synthesized by parietal endoderm or was derived from another source.

DISCUSSION

The method described here is useful for preparing and solubilizing acellular RM in a form suitable for biochemical characterization. Since the conditions for dissolving the RM samples solubilizes almost all of the visible structure, and since the remaining insoluble pellet contains less than 2% of the RM proteins, it is likely that the proteins analyzed by this method include all the major species of basement membrane proteins. This technique provides the important advantage that no proteolysis is needed for preparation of the components.

The composition of rodent RM is very simple, consisting after reduction of two major and several minor collagen molecules and four noncollagen glycoproteins (Figs. l and 2). Parietal endoderm cells synthesize (Refs. 2 and 12; Fig. 2) but do not degrade (Table 11) RM components in the culture conditions used. The composition of rat and mouse RM is remarkably similar. The collagen molecules migrate in SDS gels as a doublet or triplet at M , = 200,000-220,000, a singlet at 370,000, and several high molecular weight proteins extend- ing up to the top of the gel (Fig. 2). The collagen has not been characterized as type IV in this study, but this has been done by other investigators who have examined both rat (2, 3, 5) and mouse (39, 40) parietal yolk sac.

The high molecular weight RM glycoproteins A, B, and C of M , = approximately 415,000, 245,000, and 165,000 on reduced PAGE are immunoprecipitated by anti-laminin anti- bodies (Fig. 3). Laminin has been shown by immunofluores- cence to exist in skin, glomerulus, capillary, and embryonic basement membranes including RM (42, 43, 49, 50). On the basis of this accumulated evidence, laminin is a general basement membrane protein which is synthesized and deposited into RM by parietal endoderm cells. Our results support the suggestion (51) that RM proteins which co-migrate with laminin are related to it chemically. The synthesis of laminin or a molecule with very similar mobility characteristics has been observed in the culture medium of mouse parietal endoderm primary cultures (19-21,51), parietal yolk sac carcinoma cells (19), and F9 cells induced to differentiate (20,21). The results of this study suggest that these soluble extracellular products are destined for incorporation into basement membranes.

The properties of glycoprotein C and D are unknown. Although glycoprotein C is precipitated by anti-laminin antiserum (Fig. 3), it is not the same molecular weight as that reported for laminin in the literature (43). Glycoprotein D has a molecular weight very close to those of two of the major structural proteins of the cell, actin and tubulin; however, glycoprotein D does not co-migrate with actin or tubulin on polyacrylamide gels (data not shown). We have also observed that there is a protein in RM which migrates between glyco- ’ Aldol is the aldol condensation product of two residues of pepti-

dyl-a-amino-adipoyl-6’-semialdehyde to form the interchain, intra- molecular cross-link.

proteins A and B at the same molecular weight as fibronectin synthesized by F9 teratocarcinoma stem cells6 Fibronectin has been detected in mouse RM by immunofluorescent tech- niques (52, 53). However, the amounts of this protein in biosynthetically labeled RM are variable and usually minor, which is consistent with reports that parietal endoderm cells do not secrete significant amounts of fibronectin in culture (19,53). Fibronectin in RM may derive from maternal plasma in the sinuses of the trophoblast; it is also possible that it is synthesized by trophoblast cells and its minor presence in biosynthetically labeled RM is due to variable contamination of PE + RM with trophoblast cells. Based on the observations presented here, however, fibronectin is not a major component of RM.

The presence of the very high molecular weight collagenase- sensitive proteins in RM (see arrows in Fig. 2) has been investigated.’ On the basis of the evidence presented, there exists a mechanism for the generation of covalent bonds between RM proteins which is due to lysyl oxidase. Lysyl oxidase-derived aldehyde cross-links have been reported in other basement membranes (54) and have been postulated to exist in RM (48, 51). The distribution of aldehyde-derived cross-links in RM is unusual in the relative abundance of aldehyde precursors and aldols compared to Schiff base derivatives (compare for example the profile in Fig. 6 with that of insoluble bovine collagen fibrils in Ref. 38). The significance of this difference is unknown, although it may reflect differences in the conditions governing cross-link formation in basement membranes and in interstitial collagen. In any case, it is likely that lysyl oxidase-derived aldehyde cross-links are important in stabilizing the insoluble matrix structure; they may also affect the pore size characteristics of RM.

Acknowledgments-We would like to thank Dr. E. Reich for his suggestions and interest, Dr. George Martin for the gift of antiserum and for many helpful discussions, Dr. James Manning for much help and advice in the preliminary investigations into lysyl oxidase, and Dr. Gerald Mechanic and Lila Graham for performing the column analyses for lysyl oxidase-derived cross-links and for their generous cooperation.

REFERENCES

1. Snell, G. D., and Stevens, L. C. (1968) in Biology ofthe Labora- tory Mouse (Green, E. L., ed) pp. 205-245, Dover Publications, New York

2. Clark, C. C., Tomichek, E. A. Koszalka, T. R., Minor, R. R., and Kefalides, N. A. (1975) J. Biol. Chem. 250,5259-5267

3. Minor, R. R., Clark, C. C., Strause, E. L., Koszalka, T. R., Brent, R. L., and Kefalides, N. A. (1976) J . Biol. Chem. 251, 1789- 1794

4. Wislocki, G. B., and Padykula, H. A. (1953) Am. J. Anat. 92,117- 151

5. Clark, C. C., Minor, R. R., Koszalka, T. R., Brent, R. L., and Kefalides, N. A. (1975) Deu. Biol. 46, 243-261

6. Anderson, J. W. (1959) Am. J. Anat. 104,403-429 7. Robertson, T. A., Archer, J . M., Papadimitrou, J . M., and Walters,

M. N-I. (1971) J. Pathol. 103, 141-147 8. Brunschwig, A. E. (1927) Anat. Rec. 34,237-244 9. Everett, J . W. (1935) J. Exp. 2001. 70,243-285

10. Brambell, F. W. R., and Halliday, R. (1956) Proc. Roy. Soc. LO^.

11. Jollie, W. P. (1968) Am. J. Anat. 122, 513-532 12. Pierce, G. B., Midgley, A. R., Sri Ram, J., and Feldman, J. D.

B . Biol. Sci. 145,170-178

(1962) Am. J. PathoE. 41,549-566

‘’ Unpublished observations. ’ The large number and the molecular weight range of very high

molecular weight (> 420,000) collagenase-sensitive proteins suggests that cross-links may exist between collagen and other RM components. Although this has not been studied in detail, the amount of glycoproteins A and B appears to increase after collagenase digestion, suggesting their release from high molecular weight forms complexed with collagen.

Structural Components and Chal

13. Minor, R. R., Hoch, P. S., Koszalka, T. R., Brent, R. L., and Kefalides, N. A. (1976) Deu. Biol. 48,344-364

14. Minor, R. R., Strause, E. L., Koszalka, T. R., Brent, R. L., and Kefalides, N. A. (1976) Deu. Biol. 48, 365-376

15. Jensh, R. P., Koszalka, T. R., Jensen, M., Biddle, L., and Brent, R. L. (1977) J. Embryol. Exp. Morphol. 39,9-21

16. Strickland, S., and Mahdavi, V. (1978) Cell 15,393-403 17. Adamson, E. D., Gaunt, S. J., and Graham, C. F. (1979) Cell 17,

18. Solter, D., Shevinsky, L., Knowles, B. B., and Strickland, S. (1979)

19. Hogan, B. L. M. (1980) Deu. Biol. 76,275-285 20. Howe, C. C., and Solter, D. (1980) Deu. Biol. 77, 480-487 21. Strickland, S., Smith, K. K., and Marotti, K. R. (1980) Cell 21,

22. Runner, M. N., and Palm, J . (1953) J. Exp. 2001. 124, 303-316 23. Dulbecco, R., and Vogt, M. (1954) J. Exp. Med. 99, 167-182 24. Lowry, 0. H., Rosebrough, N. J., Farr, A. L., and Randall, R. J .

(1951) J. Biol. Chem. 193, 265-275 25. Simms, H. S., and Sanders, M. (1942) Arch. Pathol. 33,619-635 26. Laemmli, U. K. (1970) Nature 227,680-685 27. Neville, D. M., Jr. (1971) J. Biol. Chem. 246,6328-6334 28. Fairbanks, G., Steck, T. L., and Wallach, D. F. H. (1971) Bio-

29. Laskey, R., and Mills, A. (1975) Eur. J. Biochem. 56, 335-341 30. Goulianos, K., Smith, K. K., and White, S. N. (1980) Anal.

31. Kessler, S. W. (1975) J. Zmmunol. 115, 1617-1624 32. Kessler, W. S. (1976) J. Immunol. 117, 1482-1490 33. Peterkofsky, B., and Diegelmann, R., (1971) Biochemistry 10,

34. Unkeless, J . C., Tobia, A., Ossowski, L., Quigley, J. P., Rifkin, D.

35. Strickland, S., and Beers, W. H. (1976) J. Biol. Chem. 251, 5694-

469-476

Deu. Biol. 70,515-521

347-355

chemistry 10,2606-2617

Biochem. 103,64-69

988-994

B., and Reich, E. (1973) J. Exp. Med. 137, 85-111

5702

eacteristics of Reichert’s Membrane 4661

36. Diegelmann, R. F., and Peterkofsky, B. (1972) Deu. Biol. 28,443- 453

37. Udenfriend, S., Stein, S., Bohlen, P., Dairman, W., Leimgruber, W., and Weigele, M. (1972) Science 178, 871-872

38. Mechanic, G. (1974) Anal. Biochem. 61,355-361 39. Adamson, E. D., and Ayers, S. E. (1979) Cell 16,953-965 40. Tryggvason, K., Gehron Robey, P., and Martin, G. R. (1980)

41. Furthmayr, H., and Timpl, R. (1971) Anal. Biochem. 41,510-516 42. Chung, A. E., Jaffe, R., Freeman, I. L., Vergnes, J. -P., Braginski,

J. E., and Carlin, B. (1979) Cell 16,277-287 43. Timpl, R., Rohde, H., Gehron Robey, P., Rennard, S. I., Foidart,

J.-M., and Martin, G. R. (1979) J. Biol. Chem. 254,9933-9937 44. Strickland, S., Reich, E., and Sherman, M. I. (1976) Cell 9, 231-

240 45. Reich, E. (1978) in Biological Markers of Neoplasia: Basic and

Applied Agents (Ruddon, R. W., ed) pp. 491-500, Elsevier North Holland, New York

Biochemistry 19, 1284-1289

46. Siegal, R. C. (1979) Znt. Rev. Connect. Tissue Res. 8, 73-118 47. Spiro, R. G. (1967) J. Biol. Chem. 242,4813-4823 48. Minor, M. M. (1976) J. Cell Biol. 70, 109a 49. Ekblom, P., Alitalo, K., Vaheri, A., Timpl, R., and Saxen, L. (1980)

50. Leivo, I., Vaheri, A., Timpl, R., and Wartiovaara, J . (1980) Deu.

51. Hogan, B. L. M., Cooper, A. R., and Kurkinen, M. (1980) Deu.

52. Wartiovaara, J., Leivo, I., and Vaheri, A. (1979) Deu. Biol. 69,

53. Jetten, A. M., Jetten, M. E. R., and Sherman, M. I. (1979) Deu.

54. Tanzer, M. L., and Kefalides, N. A. (1973) Biochem. Biophys.

55. Woods, K. R., and Wang, K.-T. (1967) Biochim. Biophys. Acta

Proc. Natl. Acad. Sei. U. S. A . 77, 485-489

Biol. 76, 100-114

Biol. 80,289-300

247-257

Bi01. 70.89-104

Res. Commun. 51, 775-780

133,369-370

Documents

Structural Components and Characteristics of Reichert's Membrane