The Embryonic Rat Parietal Yolk Sac THE ROLE OF THE PARIETAL ENDODERM IN THE BIOSYNTHESIS OF BASEMENT MEMBRANE COLLAGEN AND GLYCOPROTEIN IN VITRO*

Basement membrane biosynthesis in vitro was studied in a rapidly growing embryonic tissue, the rat parietal yolk sac. This tissue consists of a thick, nonvascular basement membrane (Reichert’s membrane) separating two cellular layers (parietal endoderm and trophoblast). Morphologically, Reichert’s membrane appeared similar to other basement membranes. Previous analysis of the amino acid and carbohydrate composition of acellular Reichert’s membrane showed it to be typical of basement membranes isolated from other tissues and species. Analysis lens capsule and glomerular

glucosamine also served to localize the sy,nthesis of noncollagen basement membrane glycoprotein components to the parietal endoderm. As with the results reported for basement membrane collagen secretion in embryonic chick lens cells, there appeared to be approximately a 60-min delay between the incorporation of ['*C]proline into protein and the secretion of collagen as measured by the appearance of 4-hydroxy['Y2]proline in the culture medium. Experiments utilizing [3H]glucosamine to monitor glycoprotein synthesis did not show a delay between the incorporation of [3H]glucosamine and the secretion of nondialyzable 3H into the medium. The results obtained using the parietal yolk sac system to study basement membrane biosynthesis were compared to those previously obtained using the kidney glomerular and embryonic chick lens systems. It was concluded that the parietal yolk sac system is superior for a number of reasons: (a) the extracellular matrix appeared to contain only basement membrane components; there was no contamination by acid mucopolysaccharides or other types of collagen; (b) only a single cell type appeared to be responsible for the synthesis of basement membrane components; and (c) a relatively large percentage of the newly synthesized protein was basement membrane collagen.
Basement membranes are ubiquitous extracellular matrices cell layer, or both (3). Such cells are thought to be responsible which are usually contiguous with an epithelial or endothelial for the elaboration of the basement membrane upon which * This investigation was supported by United States Public Health they rest (4)(5)(6)(7). All of the basement membranes examined thus gestation since at earlier stages, uterine decidual tissue containing interstitial collagen contaminated preparations of PEMT.l As shown in Fig. 1, the ratio of labeled 4-hydroxyproline to total 14C varied markedly between the 12th and 16th days of gestation, but was constant from litter to litter with embryos of a particular age group. The maximum ratio appeared to be attained between Days 13.5 and 14.5. A pattern qualitatively similar to that presented in Fig. 1 also was obtained when the content of 4-hydroxy[14C]proline was expressed either on the basis of endodermal cell number or endodermal cell DNA (data not presented).
The inset to Fig. 1 shows further that the absolute values of total '"C (as a measure of protein synthesis) and 4 (Table I). Because purified basement membrane collagen contains an average of 70% of its imino acid as 3-plus 4-hydroxyproline (8), the results suggested that 20%/70%, or greater than onequarter (Table I) (Table II). Localization of Basement Membrane Biosynthesis in Vitro-In order to determine the contribution of PE and T to the biosynthesis of Reichert's membrane, PEMT, PEM, and T were isolated as described under "Experimental Procedure." Microscopic examination of these preparations (see Figs. 2 and 4) confirmed their homogeneity. The various tissues then were incubated in the presence of [*4C]proline for 4 hours and assayed for total "C incorporation and I-hydroxy ["Clproline synthesis. Analysis (Table III) showed that whereas T contributed approximately one-third of the total "C incorporated by PEMT, it contributed less than 1% (0.11/l&3) of the total 4-hydroxy['4C]proline in PEM + T.
A similar experiment was performed in the presence of [3H]glucosamine. Analysis to determine the distribution of label (Table IV) showed that PE and T incorporated approximately equal amounts of total 3H into glycoprotein.' However, a further localization of the synthesis of basement membrane glycoprotein was obtained autoradiographically (Fig. 2). It can be seen that grains are rather evenly distributed over T and are not concentrated along the trophoblastic surface of M. By contrast, there appears to be a concentration of grains over PE, and particularly along the endodermal surface of M.
The data in Tables III and IV further show that neither the hyaluronidase treatment nor the subsequent dissection had any deleterious effect on the system (compare PEMT with PEM + T). There does not appear to be a serum requirement 'It is presumed that [3H]glucosamine is predominantly incorporated into glycoprotein in this system. The presence of acid mucopolysaccharide-containing material is argued against by the finding of relatively large amounts of glucosamine, but only small amounts of hexuronic acid in preparations of M; and the observation that M does not stain metachromatically with toluidine blue (Ref. 14; Footnote 1). Delay. in the Secretion of Basement Membrane Collagen-The time course of incorporation of ["Clproline into protein and collagen by PEMT in vitro was examined. In four separate experiments, [ *K]proline incorporation and 4hydroxy['"C]proline synthesis by the tissue was linear for several hours (Fig. 3, A and B). Identical to the observations of Grant et al. (17), there appeared to be a delay of at least 60 min before significant nondialyzable [L4C]proline or 4-hydroxy[14C]proline was detected in the medium (Fig. 3, A and B). After this initial delay, the release of isotope into the medium increased linearly for the duration of the experiment.
TO determine if the delay in the appearance of 14C-labeled basement membrane collagen in the medium was due to an extracellular accumulation of protein on M, PEMT were incubated with ['4C]proline for either 30 or 180 min and the isolated tissue was treated with bacterial collagenase and trypsin (17). The results (Table V) showed that proteolytic digestion of the tissue during the delay period (e.g. 30 min) did not release 4-hydroxy[14C]proline in excess of the control, whereas prior lysis of the cells permitted the release of nearly all the 4-hydroxy [14C Jproline synthesized. On the other hand, proteolytic digestion of the tissue after longer incubation periods (e.g. 180 min) showed that a significant portion of the 4-hydroxy[14C]proline was extracellular. This suggested that labeled basement membrane collagen was being deposited on M during the incubation period.  by PEM in vitro was also examined. PEM was used rather than PEMT since it was previously shown that while T contributed approximately 50% of the total 3H incorporated by PEMT (Table IV), it probably does not contribute significantly to the synthesis of M (Fig. 2) (26). In two separate experiments, [3H]glucosamine incorporation into the tissue fraction and subsequent release into the medium appeared to be linear for at least 120 min (Fig. 5). In contrast to incubations with ['*C]proline (Fig. 3), no delay period was observed in the appearance of 3H in the medium. DISCUSSION Basement membrane biosynthesis in vitro has recently been studied using either isolated kidney glomeruli (18-21) or whole embryonic chick lenses and isolated lens cells (17,27,28). While the basement membranes of both of these systems have been extensively characterized (for review, see Ref. 8), these systems appear to suffer from several disadvantages when it comes to studying basement membrane biosynthesis.
Therefore, we have examined the efficacy of the rat embryo parietal yolk sac for this purpose.
In the discussion which follows, we shall note the similarities in results found in all these in vitro systems, as well as emphasize the apparent advantages of the latter system.
Previous studies' (4, 6, 11, 13, 14, 29) have shown that the rodent parietal yolk sac is a relatively homogeneous system consisting of a single morphologically identifiable extracellular  Fig. 2). On the basis of histochemical studies (14,15), it was demonstrated that rat Reichert's membrane was composed of glycoprotein, but not acid mucopolysaccharides, elastin or lipid.
Our chemical studies of acellular Reichert's membrane' confirm the presence of glycoprotein and absence of mucopolysaccharides and also show the presence of basement membrane collagen. Initial biosynthetic experiments showed that intact PEMT (or PEM) incorporate ['*C]proline and synthesize hydroxy['4C]proline in a reproducible fashion at all stages of gestation examined (Fig. 1). Maximum protein synthesis appeared to occur around Day 14.5 so all subsequent biosynthetic studies were perf0rme.d using tissues from that stage of gestation. These observations contrast those reported for the whole chick embryo lens system where there was a considerable variation among results obtained with different batches of embryos on different weeks (17).
The cellular origin of rodent Reichert's membrane has long been in dispute. On a morphological basis, some investigators (30, 31) suggested an ectodermal (trophoblastic) derivation, while others (13, 32-34) favored an endodermal origin. On the basis of electron microscopic and immunohistochemical studies on murine parietal yolk sac, Pierce et al. (5,6) concluded that Reichert's membrane was a secretion of normal PE. Pierce and his colleagues (4-7, 11) also concluded on the basis of experiments with a cultured murine testicular teratocarcinema, that a hyaline material similar to Reichert's membrane (M) was secreted solely by the tumor cells which were likened to PE. This tumor system has been referred to as a "parietal yolk sac carcinoma" (35,36). Biochemical evidence in a nonpathological tissue, however, has not been directed toward resolution of this controversy. Therefore, we examined the question of the cellular origin of M in normal rat embryos by means of chemical and autoradiographical studies. Of considerable importance in our studies is the fact that the parietal endoderm and trophoblast were completely separated and studied independently.
We believe that this is the first report that these tissue elements have been isolated in a viable condition for in vitro biosynthetic studies. The data presented in Table III Fig. 2 suggest that labeled material is being secreted onto M by PE, but not by T. Similarly, newly synthesized basement membrane components appear to be deposited on M even in the absence of T (Fig. 4). Since this region on the endodermal surface of M has been shown to contain newly deposited basement membrane3 (26), we also must conclude that all of the components of M are elaborated solely by PE.
A major disadvantage in using either the glomerular or embryonic lens systems for studying basement membrane collagen biosynthesis appears to be the fact that basement membrane collagen constitutes a very small percentage of the total protein synthesized by these tissues.
The differences in collagen content are reflected in the ratio of labeled 4-hydroxyproline to total radioactivity which ranged around 0.015 in both the glomerular and lens systems (17, 18) compared to 0.192 in the yolk sac system (Table I). The latter value contrasts markedly with the value of 0.015 reported by Priest (37) using cells derived from a mouse "parietal yolk sac carcinoma." From these ratios, it is possible to estimate the percentage of the newly synthesized protein represented by basement membrane collagen using a modification of the equation presented by Diegelmann and Peterkofsky5 (38). From such calculations, it appears that only 0.5% of the newly synthesized protein in the glomerular, lens, and "parietal yolk sac carcinoma" systems is basement membrane collagen compared to 6% in PEMT cultures. Moreover, since T contributes about one-third of the total '"C incorporated but essentially none of the 4-hydroxy[14C]proline synthesized by PEMT (Table III), the basement membrane collagen represents at least 9% (6%/0.67) of the total newly synthesized protein in PEM cultures.
The collagen synthesized by PEMT was shown to be typical of basement membrane collagen in that it contained relatively large amounts of 3-hydroxyproline (Table I)  secretion into the medium appear to be due to an extracellular adsorption of the collagen became linear (Fig. 3). This observation is identical to that on either PE or M since digestion of the tissue with trypsin and reported by Grant et al. (17) and suggests that a 60-min lag collagenase indicated that extracellular matrix was not labeled period is characteristic of the secretion of basement membrane in excess of the control ( analysis also showed that the extracellular matrix was not significantly labeled during the lag period (Fig. 4, A and C). On the other hand, at times subsequent to the lag period (e.g. 180 min) when protein secretion is occurring (Figs. 3 and 5), the extracellular matrix (M) appeared to be labeled as indicated both by the release of labeled hydroxyproline in excess of the control (Table V) and by autoradiographic analysis (Figs. 2 and 4, B and D). After 120 min of incubation, approximately 17% of the total 4-hydroxy ["C Jproline was present in the medium. This value is similar to that reported by Grant  It has been suggested (17) that one possible explanation for the delay in the secretion of basement membrane collagen may be the time required to synthesize the relatively large amount of hydroxyproline and hydroxylysine in this collagen. Indeed, Peterkofsky (41) has shown that when cultures of 3T3 fibroblasts are supplemented with ascorbate there was a 45min lag in the secretion of collagen, while in the absence of the vitamin the lag period increased to 90 to 120 min. In addition, the collagen synthesized and secreted in the absence of ascorbate was extensively underhydroxylated (41). In PEMT cultures, however, supplementation of the incubation medium with 0.25 mM ascorbate had no effect on either the percent of the [14C]proline hydroxylated or the delay in the secretion of 14C-labeled basement membrane collagen (data not presented). The delay also does not appear to be due to time required for the extensive glycosylation of hydroxylysine since, after only a 5-min incubation with ["Cllysine, the newly synthesized basement membrane collagen appeared to be completely glycosylated (Table II). It is concluded, therefore, that at least for short term incubations this system does not require exogenous ascorbate for basement membrane collagen biosynthesis, and that the lag period is probably not due to a delay in the hydroxylation of either proline or lysine, or to a delay in the glycosylation of hydroxylysine 6 C. C. Clark, E. A. Tomichek, T. R. Koszalka, R. R. Minor, and N. A. Kefalides, unpublished observations.

Continuous
labeling of PEMT for up to 480 min showed that the maximum ratio of 4-hydroxy['4C]proline to total 14C in the medium (21 to 28%) appeared to be attained between 120 to 240 min. This indicated that nearly 40% (24%/60%) of the [*4C]proline in the medium at this time is present in newly synthesized basement membrane collagen. The ratio in the tissue fraction remained constant around 12 to 15% up to 240 min. The ratios in both fractions began to fall after longer incubation periods. A time course experiment performed in the presence of [3H]glucosamine (Fig. 5) did not show a lag period for the release of nondialyzable 3H into the medium. While the precise nature of the glucosamine-containing material was not investigated, previous studies have shown that a relatively large amount of glucosamine was associated with protein components of M; mucopolysaccharide contamination was not detected. 4 Experiments designed to determine either the presence or absence of glucosamine in newly synthesized basement membrane collagen are presently in progress.
From these considerations, it may be inferred that there exists both a temporal and a cellular compartmentalization of the collagen and glycoprotein components of M. That is, basement membrane glycoprotein may be present extracellularly before basement membrane collagen is secreted. Interaction between these components, perhaps through disulfide bonds, then would occur extracellularly to form a functional basement membrane. This interpretation is consistent with the observation in embryonic chick lens cells that in the absence of disulfide reducing agents, the extracellular hydroxy ['"C ]proline is present in large aggregates while the intracellular hydroxy [14C lproline is not aggregated (28).
Based on the results and comparisons presented, the parietal yolk sac system appears well suited for studying the synthesis and assembly of basement membrane components.