Basement Membrane Procollagen Is Not Converted to Collagen in Organ Cultures of’Parieta1 Yolk Sac Endoderm*

Basement membrane procollagen biosynthesis was studied in organ cultures of embryonic rat parietal yolk sac endoderm by following [“C]proline incorporation into nondialyzable proteins. After reduction with Smercaptoethanol the “C-proteins synthesized were characterized by agarose gel filtration and disc electrophoresis in the presence of sodium dodecyl sulfate. The labeled procollagen was identified by its content of hydroxy [,‘C Jproline, its sensitivity to digestion with bacterial collagenase, and its resistance to digestion with pepsin. In cultures which were continuously labeled for periods from 6 hours to 4 days, the pro-cu chains consistently eluted as a single peak with an apparent molecular weight of 160,000. After pepsin digestion the resultant OL chains had an apparent molecular weight between 125,000 and 140,000. This suggests that basement membrane procollagen either contains non-triple helical pepsin-resistant regions or a triple helical region which is larger than the corresponding region of interstitial procollagen. Two experiments were performed to determine whether the chains of newly synthesized basement membrane procollagen

Basement membrane procollagen biosynthesis was studied in organ cultures of embryonic rat parietal yolk sac endoderm by following ["C]proline incorporation into nondialyzable proteins. After reduction with Smercaptoethanol the "C-proteins synthesized were characterized by agarose gel filtration and disc electrophoresis in the presence of sodium dodecyl sulfate. The labeled procollagen was identified by its content of hydroxy [,'C Jproline, its sensitivity to digestion with bacterial collagenase, and its resistance to digestion with pepsin. In cultures which were continuously labeled for periods from 6 hours to 4 days, the pro-cu chains consistently eluted as a single peak with an apparent molecular weight of 160,000. After pepsin digestion the resultant OL chains had an apparent molecular weight between 125,000 and 140,000. This suggests that basement membrane procollagen either contains non-triple helical pepsin-resistant regions or a triple helical region which is larger than the corresponding region of interstitial procollagen. Two experiments were performed to determine whether the chains of newly synthesized basement membrane procollagen were cleaved to a smaller molecular species. In the first, the hydroxylation and secretion of procollagen were blocked with cr,a'-dipyridyl, and the resulting intracellular chains of basement membrane protocollagen were found to co-elute with fully hydroxylated and secreted pro-a chains. In the second, cultures were labeled for 1 day and chased for 3 days with unlabeled medium. Autoradiography had shown that most of the label was chased into new basement membrane. Agarose chromatography showed that after a 3%day chase the pro-n chains still elut.ed with an apparent molecular weight of 160,000. Thus, the data indicated that basement membrane procollagen was deposited in new basement membrane without undergoing a time-dependent extracellular conversion.
It is now well established that basement membranes consist cultures of PEM' were labeled with ["Clproline, 4-hydroxyof a collagen component in association with one or more ["C]proline constituted -20% of the total "C, and :Inon-collagen glycoproteins (:I, 4). In recent studies we have hydroxy [ "C lproline constituted w 10% of the total hydroxyfound that organ cultures of embryonic rat parietal yolk sac ["Clproline.
In cultures labeled with ["Cllysine, -85'i; of the tissues offer many advantages for studies of the biosynthesis of hydroxy["C]lysine residues were glycosylated, and -95% of basement membrane (5)(6)(7)(8)(9). For example, the initial studies these residues were glucosylgalactosylhydroxy [ "C ]lysine (5,7, showed that the parietal endodermal cells on the surface of 9). Autoradiography showed that in cultures of PEM labeled existing basement membrane could be cleanly isolated without with ['H]proline or [SH]glucosamine, more than 10 pm of exposure to enzymes, that it was only the endodermal cells heavily labeled new basement membrane were deposited on which synthesized basement membrane, that basement mem-the surface of existing membrane during the first 4 days of brane was the only structured extracellular matrix synthe-culture (5,8). Because of the relatively high rate of basement sized, and that new basement membrane was deposited on the membrane synthesis, cultures of PEM therefore appeared to be surface of existing basement membrane (5)(6)(7)(8)(9). When organ appropriate for further studies of both basement membrane collagen biosynthesis, and the steps involved in the deposition

Basement
Membrane Procollagen in Parietal Yolk Sac Cultures lens and glomeruli, Grant et al. (10,11) suggested that this molecule, like interstitial collagens, was synthesized as a precursor or procollagen molecule which appeared to undergo a time-dependent extracellular conversion to basement membrane collagen. The pro-a chains of basement membrane procollagen secreted by the lens had an apparent molecular weight of -140,000 and these chains appeared to undergo an extracellular conversion to chains with a molecular weight of -115,000 (12). In contrast, our initial studies of the collagen component of basement membrane deposited in cultures of PEM showed a' higher molecular weight for the procollagen molecule, and failed to show any evidence for conversion to a lower molecular weight species (1).
The present study deals with investigations of the molecular weight and lack of conversion of basement membrane procollagen. The pro-cu chains extracted from new basement membrane had an apparent molecular weight of 160,000, and the (Y chains resulting from pepsin digestion had an apparent molecular weight between 125,000 and 140,000. However, there was no evidence of an extracellular conversion of this basement membrane procollagen in the cultures. Unhydroxylated intracellular pro-a chains, hydroxylated pro-a chains in g-hour to 4-day continuously labeled cultures, and pro-a chains extracted from new basement membrane after a l-day pulse and s-day chase all had the same apparent molecular size.  Lash et al. (13).
Organ Culture Preparations-All cultures consisted of the single layer of cells of the parietal yolk sac endoderm on the surface of the existing parietal yolk sac basement membrane, which were isolated by dissection without the aid of enzymes (8). Six halves of PEM were grown on the surface of 1 ml of a nutrient agar substrate and fed 80 ~1 of liquid nutrient medium at I%-hour intervals as described previously (5, 8. 9). Sodium ascorbate was added to a concentration of 200 &ml in the liquid nutrient medium, giving a concentration of 16 eg of ascorbate/ml of nutrient substrate. P-Aminopropionitrile was added to a concentration of 50 &ml in both the nutrient agar and liquid nutrient medium. For radioisotopic labeling, L-proline was omitted from the F12X and 5 rCi/ml of ["Clproline or 10 NCi/ml of ]SH]proline .~ was added to both t,he nutrient agar and liquid nutrient medium.

Cultures
were continuously labeled for 6 hours to 4 days, or were labeled for 1 day and then chased for 3 days on unlabeled agar. For the preparation of unhydroxylated basement membrane procollagen (protocollagen), the tissues were preincubated 20 min at 37" in 0.5 mM n,a'-dipyridyl in Simms balanced salt solution before they were pulsed on nutrient agar containing labeled proline and 0.5 mM a,cu'-dipyridyl (14).
The procedures described by Grant et al. (10)  The remainder of the sample was chromatographed on a calibrated column of agarose A-5m (200 to 400 mesh) as described previously (10). Fractions of 2 ml were collected, and aliquots of each fraction were taken for the determination of total "C; the remainder of appropriate fractions was hydrolyzed and aliquots were taken for the determination of labeled and unlabeled 4-hydroxyproline.
The positions of fi and a components from the added rat tail tendon collagen were used to calibrate the column for molecular weight determinations.
Sodium Dodecyl Sulfate-Polyacrylamide Gel Electrophoresis-Aliquots of the reduced homogenate described above were dialyzed against several changes of water containing Dowex 1 resin prior to lyophiliza-Lion. Membrane-Sodium dodecyl sulfate-agarose gel filtration and sodium dodecyl sulfate-acrylamide gel electrophoresis were used to determine the size of the polypeptide chains of the collagen component of new basement membrane which was deposited in continuously labeled 1, 2, and 4-day cultures 01 PEM. The results showed that all of the cultures yielded a single major peak of both "C and 4-hydroxy[1*C]proline which eluted between the a chains and /3 components of the rat tail tendon collagen (Fig. 1). The apparent molecular weight of this peak was -160,000 with both gel filtration (Fig. IA) and disc gel electrophoresis (Fig. 1B). This peak on sodium dodecyl sulfate-agarose contained -45'1 of the total "C and -75%, of the 4-hydroxy["C]proline. The highest ratio of 4-hydroxy["C]proline to total "C in the peak fractions was -0.65, and in the entire peak (between fractions 30 and 40) this ratio was -0.4. Since this ratio has been shown to be -0.6 in other basement membrane collagens (3,4), these results suggested that approximately two-thirds (0.4/0.6) of the total "C in this peak was in basement membrane collagen chains with an apparent molecular weight of 160,000.
In addition to the one major peak of 4-hydroxy["C]proline, smaller amounts were also present in two larger molecular aggregates. With sodium dodecyl sulfate-agarose, -3% of the 4-hydroxy ["C lproline eluted in the void volume of the column and -10% eluted in a peak with an apparent molecular weight of -220,060 (Fig. 1A). Corresponding peaks of "C were present in the sodium dodecyl sulfate-acrylamide gels (Fig. 1B). The ratio of 4-hydroxy['E]proline to total "C in these peaks was between 0.2 and 0.3, which was approximately the same as that in the whole tissue.
To show that the 4.hydroxy["C]proline in the chromatograms was indicative of the distribution of collagen chains, cultures of PEM were digested with bacterial collagenase prior to denaturation and reduction with sodium dodecyl sulfatemercaptoethanol.
As shown in Fig. IC, all of the 4-hydroxy-[*'C]proline-containing proteins in these cultures were digested. Approximately 90% of the 4-hydroxy["C]proline became dialyzable, and the only remaining nondialyzable 4-hydroxy["C]proline in the chromatograms was in peptides which were too small to be resolved by the sodium dodecyl sulfateagarose (Fig. 1C) compared to that of chains which were secreted and deposited in new basement membrane. In order to block secretion, unhydroxylated basement membrane procollagen (protocollagen) was prepared by labeling cultures of PEM with either [aH]-or ["Clproline in the presence of 0.5 mM a,&-dipyridyl. The ratio of 4-hydroxy["C] proline to total "C in these cultures was less than 0.005. Autoradiography showed that at the end of the 6 hour-pulse in the presence of ol,rr'-dipyridyl nearly all of the label remained in the endodermal cells (data not shown), whereas in control cultures much of the label was deposited in new membrane (see Ref. 8). Cultures labeled with l"C]proline in the presence of a,a'-dipyridyl were therefore reduced with sodium dodecyl sulfate-mercaptoethanol and co-chromatographed on sodium dodecyl sulfate-agarose with control cultures labeled with ['Hlproline in the absence of a,a'-dipyridyl. The results of these experiments showed that the ['*C]protocollagen chains eluted in the same position as the [9H]procollagen chains, with an apparent molecular weight of 160,000 (Fig. 2).
The second set of experiments in this series was designed to determine whether there was a delayed conversion of basement membrane procollagen to basement membrane collagen. Cultures were labeled for 1 day with ["Clproline and chased for 3 days on unlabeled agar. Autoradiography had previously shown that in such cultures the majority of the label was chased out of the endodermal cells into heavily labeled basement membrane which subsequently became buried under poorly labeled new basement membrane (8). As shown in Fig.  3, the pro-cu chains in these cultures eluted in the same position as those in both continuously labeled cultures (Fig. lA), and cultures labeled in the presence of Lu,a'-dipyridyl (Fig. 2). These results indicated that the procollagen in these cultures was secreted and deposited in new basement membrane without undergoing an extracellular cleavage. Since Grant et al. (10) Fig. 4, the major peak of 4-hydroxy["C]proline synthesized by the lenses eluted in the same position as the 'H-labeled pro-n chains synthesized by the endodermal cells. There were some smaller molecular weight (<85,000) peptides containing 4-hydroxy["C]proline in these chromatograms, but there was no evidence of an extracellular conversion of the pro-a chains synthesized by either lenses or PEM (Fig. 4). Apparent

Region of Basement
Membrane Procollagen-Sodium dodecyl sulfate-agarose gel filtration was also used to determine the apparent size of the (Y chains resulting from pepsin digestion of basement membrane procollagen. In l-day cultures of PEM, -30% of the total "C and -70% of the 4.hydroxy["C]proline remained in the retentate after pepsin digestion, reduction with sodium dodecyl sulfate-mercaptnethanol, and dialysis against sodium dodecyl sulfate-phosphate buffer. When this retentate was chromatographed on sodium dndecyl sulfate-agarose, between 70 and 75'2 of both the total "C and 4.hydroxy["C]proline eluted in a peak with an apparent molecular weight between 125,OOOand 140,OW (Fig. 5). The ratio of 4-hydrnxy["C]proline tn total "C in this entire peak ranged from 0.65 to 0.68. This suggested that the pepsin-resistant material eluting in this region was free of non-collagen proteins.

DISCUSSION
Numerous studies have shown that the interstitial cnllagens are synthesized as precursor or procollagen molecules which undergo an extracellular cleavage to form collagen which is then deposited in fibrils and fibers (21-23). It is also well established that the oeotides which are cleaved from intersti- The chromatographic conditions were the same as in Fig. 1A. 04, total "C; O---O, Chydroxy["C lproline.
tial procollagens are not triple helical and hence are removed by proteolytic enzymes such as pepsin, trypsin, or chymotrypsin. On the other hand, the triple helical region of native interstitial procollagen is digested by bacterial collagenase, but is resistant to the aforementioned proteolytic enzymes. The studies presented here have shown that the basement membrane collagen molecules synthesized by the parietal endoder-ma1 cells consist of both a collagenase-sensitive, pepsin-resistant, triple helical region, and pepsin-sensitive non-triple helical extensions. Because of this similarity to the interstitial procollagens, we refer to the collagenous molecules synthesized by the endodermal cells as "basement membrane procollagen" and the polypeptide chains which comprise these molecules as "pro-a chains." With both sodium dodecyl sulfate-agarose gel filtration and sodium dodetiyl sulfate-polyacrylamide disc gel electrophoresis, the pro-a chains synthesized by-the endodermal cells had an apparent molecular weight of 160.000 ( Fig. 1A and B). This is larger than the apparent size of 140,000 obtained by Grant et al. (IO) for the chains of procollagen synthesized by embryonic chick lens. and it is smaller than the apparent size of 180,000 obtained in our initial studies of PEM (1). In neither of these previous studies, however, were the positions ofelution of a chains and /3 components of tendon collagen used to calibrate the sodium dodecyl sulfate-agarose chromatograms. This method of internal calibration has minimized the variation in the estimates of molecular weight noted in the previous studies (1. IO). Our results showed that the pro-a chains synthesized by the lens co-elmed wit.h those synthesized by the endodermal cells with an apparent molecular weight of 160,000 (Fig. 4). A comparable value was reported by Kefalides et al. (24) for the chains of basement membrane procollagen synthesized by cornea1 endothelium, and by Byers et al. (25) for the chains of interstitial procollagen synthesized by embryonic chick calvaria.
Previous autoradiographic and biochemical studies have shown that the procollagen synthesized in cultures of PEM is deposited with non-collagen glycoproteins in new basement membrane (5,7,8). The results obtained here show that the procollagen synthesized by the endodermal cells is deposited in the new basement membrane without undergoing an extracellular cleavage. Even aft.er a &day chase (Fig. 3). the chains of procollagen extracted from new basement membrane still had the same apparent molecular weight as the intracellular chains of basement membrane protocollagen (Fig. 2). This absence of cleavage of basement membrane procollagen may account for the absence of banded collagen fibrils in most basement membranes (4).
The absence of cleavage of procollagen in this study differs from the observations of Grant et al. (10) which suggested that during a 4-hour incubation of embryonic chick lenses there was a cleavage of the newly synthesized procollagen to a collagen consisting of cy chains with an apparent molecular weight ot 115,000. In contrast, we failed to see any evidence of a conversion of either endodermal procollagen, or the procollagen synthesized by embryonic chick lenses which were incubated for 6 hours under the conditions described by Grant et al. (10) (Fig. 4). Some nonspecific degradation undoubtedly occurs in all culture systems, and this may explain the apparent conversion of basement membrane procollagen which was reported by Grant et al. (10,11).
It was interesting to find that although the size of basement membrane procollagen appears to be comparable to that of interstitial procollagen (25), there is a notable difference in the apparent size of the pepsin-resistant region of these procollagens. Byers et al. (25) and Fessler et al. (26) have recently shown that the procollagen which is synthesized by calvaria consists of (a) a pepsin-resistant triple helical region of OL chains with an apparent molecular weight of 95,000, (b) a disulfide-bonded, pepsin-sensitive, non-triple helical, COOHterminal extension in which the chains each have an apparent molecular weight of -35,000, and (c) a pepsin-sensitive, non-triple helical, NH,-terminal extension in which the chains each have an apparent molecular weight of -20,000. In contrast, we find that the pepsin-resistant region of basement membrane procollagen consists of chains with an apparent molecular weight between 125,000 and 140,000 (Fig. !I), and this suggests that the pepsin-sensitive extensions of the pro-o chains have a molecular weight between 20,000 and 35,000. Consequently, the pepsin-sensitive propeptides may constitute 15 to 20% of the basement membrane procollagen molecule, whereas they may constitute as much as 40% of the total mass of an interstitial procollagen molecule (25,26). Presumably this is indicative of differences in the proportion 01' the molecule which is triple helical in basement membrane and interstitial procollagen. The possibility must be considered, however, that basement membrane procollagen molecules may also contain non-triple helical, pepsin-resistant regions. This possibility is currently under investigation.
Since 80 to 854-of the hydroxylysine residues in basement membrane procollagen are glycosylated and 95% of these residues are the disaccharide glucosylgalactosylhydroxylysine (4,7,9), one would expect that the apparent size of the pro-0 chains extracted from basement membrane should be larger than that of the unhydroxylated, unglycosylat.ed intracellular chains of protocollagen.
Nevertheless, the experiments with ti,a'-dipyridyl failed to show a reduction in size of the pro-tr chains in the absence of the sugars. This may be due to a failure of the sugars to affect the apparent size of the procollagen chains in the presence of sodium dodecyi sulfate. It is possible, however, that there is an additional peptide ext,ension on the intracellular chains of basement membrane protocollagen which has the same effect on their apparent size as the sugar residues have on the extracellular pro-n chains. If such an extension does exist it may be cleaved intracellularly. This possibility is also currently under investigation.
The studies described in this report have shown that the collagenous component of the parietal volk sac basement membrane is synthesized as a procollagen molecule which is deposited in new basement membrane without undergoing an extracellular cleavage to a smaller molecular species. In addition, these studies have shown that a major difference between basement membrane and interstitial procollagen consists of the fact that a larger proportion of the basement membrane procollagen molecule is resistant to pepsin. Finally, these studies have shown that organ cultures of PEM on a nutrient agar substrate should permit further examination of the steps involved in the synthesis, secretion, deposition, and stabilization of basement membrane procollagen.