The Stimulation of Collagen Secretion by Ascorbate Increased Proline Hydroxylation in Chick Embryo THOMAS Graduate
School of Arts and National Institutes
J. J. BLANCK’ Sciences, University of Health, National
of Pennsylvania, Cancer Institute,
as a Result of Fibroblasts
PETERKOFSKY’ Philadelphia, Bethesda,
Pennsylvania Maryland 20014
Collagen secretion by chick embryo fibroblasts was measured by incorporating [14C]proline into proteins and then analyzing the amount of collagen in the cell and medium separately by using purified bacterial collagenase. In order to produce varying levels of hydroxylation, cells were incubated with varying concentrations of ascorbate or with varying concentrations of cY,a’-dipyridyl in the presence of saturating ascorbate. Ascorbate stimulated both the hydroxylation of proline in collagen and the secretion of collagen; the concentration of ascorbate required for half-maximal stimulation of both processes was approximately 4.5 x lo-’ M. Since the cells could concentrate ascorbate lofold, this Khl for proline hydroxylation is IOO-fold lower than values reported for purified prolyl hydroxylase (Abbot, M. T., and Udenfriend, S. (1974) in Molecular Mechanisms of Oxygen Activation (Hayaishi, O., ed.), p. 173, Academic Press New York; Kivirikko K. I., et al. (1968) Biochim. Biophys. Acta 151, 5588567). Conversely, a,a’-dipyridyl inhibited both proline hydroxylation and collagen secretion; half-maximal inhibition of both processes was observed at 7 x 10m5 M. The results of the two types of experiments show that the secretion of collagen becomes directly proportional to proline hydroxylation when approximately 30% of the proline residues in collagen have been hydroxylated compared to maximal hydroxylation of 50%. Since the stability of triple-helical collagen at 37°C has been shown to be dependent on the hydroxyproline content of the molecule (Rosenbloom, J., et al. (1973) Arch. Biochem. Biophys. 158, 478-484), we suggest that the observed proportionality between secretion and hydroxylation is a reflection of the increased amount of stable triple helical collagen at 37°C. When the cells were incubated with a concentration of ascorbate that was saturating for secretion and hydroxylation, there was no significant activation of prolyl hydroxylase as measured in a cell-free extract. These experiments suggest that ascorbate effects collagen secretion by acting at the site of proline hydroxylation but not by increasing the activity of prolyl hydroxylase.
It has been known for 50 years that ascorbic acid is an essential vitamin in man for the maintenance of connective tissue (1). Recent investigations have shown that the action of ascorbate is related to the synthesis and metabolism of collagen. Robertson
demonstrated that ascorbic acid increased the amount of hydroxyproline in granuloma of scorbutic guinea pigs and interpreted this as an augmentation of collagen synthesis (2). At that time it was believed that hydroxyproline formation was indicative of collagen formation, while we know now that underhydroxylated collagen may be synthesized in the absence of ascorbate (3, 4). Stone and Meister (5) subsequently showed that ascorbic acid acted at the site of proline hydroxylation but assumed that the hydroxylation of proline occurred prior to incorporation of proline into peptide linkage (5). Based on their own observations and those of Robertson (21, they postu-
’ Presented to the Faculty of the Graduate School of Arts and Sciences of the University of Pennsylvania in partial fulfillment of the requirements for the degree of Doctor of Philosophy. Present address: Center for Muscle Biology and Physiopathology, Institute of General Pathology, University of Padua, Padua 35110, Italy. * To whom correspondence is to be addressed at: National Institutes of Health, National Cancer Institute, Laboratory of Biochemistry, Biosynthesis Section, Bethesda, Maryland 20014. 259 Copyright 0 1975 by Academic Press, Inc. All rights of reproduction in any form reserved.
lated that the inhibition of hydroxylation prevented the synthesis of collagen. In 1965 it was found that ascorbate was a cofactor for collagen prolyl hydroxylase (6) and subsequently Peterkofsky demonstrated that the addition of ascorbic acid to the incubation medium of 3T3 or chick embryo fibroblasts in culture resulted in a three-fold stimulation of collagen secretion into the medium (7). These studies showed that 3T3 cells produced underhydroxylated collagen in the absence of ascorbate and fully hydroxylated collagen in its presence, suggesting that the stimulation of collagen secretion by ascorbate was directly related to the function of ascorbate as a cofactor in proline hydroxylation. This conclusion, however, warranted further investigation in view of the fact that in logphase L929 cells, ascorbate stimulated proline hydroxylation while failing to increase collagen secretion (3). In addition to these functions of ascorbate, Stassen et al. (8) have shown that ascorbate activates the enzyme prolyl hydroxylase in logphase L929 fibroblasts in culture. The present paper describes experiments in which the role of ascorbate as a cofactor and activator of prolyl hydroxylase was further investigated in relation to collagen secretion by chick embryo fibroblasts in tissue culture. MATERIALS
Materials. L-[‘4C]proline (uniformly labeled), L [3,4-3Hlproline, and L-[i4Clascorbate were purchased from New England Nuclear Corp.; sodium ascorbate, colchicine, and cycloheximide from Calbiochem, and a&-dipyridyl from Sigma Chemical Co. Chromatographically purified bacterial collagenase, which was further purified (9), and crude bacterial collagenase were purchased from Worthington Biochemical Corp. Cell culture. Tissue culture medium was prepared by the NIH media section. All cells were cultivated in an atmosphere of CO#%)-air(95%) in Falcon tissue culture plates using Eagle’s minimal essential medium containing 10% fetal calf serum, modified by reducing the NaHC03 concentration to onehalf (0.013 M) and adding Tricine buffer, pH 7.4, 0.025 M. This medium is designated MEM-10; if serum is omitted it is designated MEM-0 (3). Isolation of fibroblasts. Chick embryo fibroblasts were obtained by treating 24 minced frontal bones from l&day embryos with crude bacterial collagenase (0.5 mg/ml) in 10 ml of MEM-0 for 30 min at 37°C
PETERKOFSKY while shaking. The cells were then centrifuged at 25% for 2 min, resuspended in MEM-10 and cultured as previously described (7). The cells were used in passages two through eight and were grown to approximately mid-log-phase for all experiments. Incorporation of [‘4Clproline into cell protein. Growth medium was removed from log-phase cells cultured for 2-3 days without ascorbate and the cell layer rinsed with 2 ml of warm (37°C) MEM-0. Three milliliters of MEM-0 were added to each plate followed by the addition of either ascorbate or a,&dipyridyl; control cultures received no further additions at this time. The plates were incubated at 37°C for 15 min at which time 1.5 &!i of [‘*C]proline (20 &i/~mol) in 75 ~1 of MEM-0 were added to each plate and incubation continued at 37°C for varying periods of time. Each sample consisted of pooled cell layer or medium from two or three plates, and duplicate samples were always used in analyses. After 2 h of incubation, medium was removed and the cell layer rinsed twice with 2 ml of cold MEM-0; the rinse was added to the original incubation medium for analysis. Two milliliters of cold 0.05 M Tris-HCl, pH 7.6, were then added to the cell layer and the cells detached from the plate by scraping with a rotating magnetic stirring bar. The plate was rinsed twice with 2 ml of the 0.05 M Tris-HCl, pH 7.6 buffer and the rinse added to the cell suspension. [‘YJascorbate uptake. Measurement of [‘4Clascorbate uptake as a function of concentration was carried out in a similar manner. Cells were rinsed as described above and 3 ml of MEM-0 added to each plate. [‘Qascorbate (3.4 @Zi/~rnoll was then added in increasing amounts and the culture incubated for 2 h at which time incubation medium was removed and the cell layer rinsed twice with cold MEM-0. The cells were harvested with 1 ml of Tris buffer per plate and the plates rinsed with 1 ml; the combined cell suspension was counted in a liquid scintillation counter. Collagen assay. Collagen and noncollagen protein in cell sonicates and dialyzed, lyophilized medium were analyzed by collagenase digestion of trichloroacetic acid-precipitated protein as previously described (3). Secretion is expressed as percent of the total collagen that is found in the medium: [disintegrations per minute (medium collagen)/(disintegrations per minute (medium collagen) + disintegrations per minute (cell collagen))] x 100%. Proline and hydroxyproline were determined in hydrolysates of collagenase digests by the method of Peterkofsky and Prockop (lo), with some modifications (9); hydroxylation is expressed as percent of total imino acids: [distintegrations per minute (hydroxyproline)/ (disintegrations per minute (hydroxyproline) + disintegrations per minute (proline))] x 100%. Cellular DNA was determined by the B.urton procedure (11) and protein by the Hartree moditication (12) of the Lowry method (13).
Assay of prolyl hydroxylase. Prolyl hydroxylase activity was measured by a micromodification (14) of the tritium release assay of Hutton et al. (151, in which tritiated water is collected after passage of trichloroacetic acid-soluble reaction products through a small Dowex-50 column. Unhydroxylated collagen substrate was prepared by incubating chick embryo frontal bones with L-[3,4-3Hlproline and 5 x 10m4 M a,a’-dipyridyl in MEM-0 for 2 h at 37°C. After the incubation, collagen was isolated from minced bones as previously described (16). Cells were incubated in MEM-0 for 2 h either in the presence or absence of 1 x lo-” M ascorbate. The medium was then removed, the cell layer rinsed with Earle’s basal salt solution lacking Ca2+ and Mg2+ and containing 0.013 M NaHCO,, 0.025 M Tritine, pH 7.4 and lo-’ M EDTA. The cells were removed in 1 ml of this solution by using a rubber policeman and were centrifuged at 1OOOg for 5 min. The cells were resuspended in 0.11 M NaCh0.05 M Tris-HCl, pH 7.4, and centrifuged again. The resulting pellet was suspended in 0.3 ml of 0.05 M TrisHCl, pH 7.6, containing 1 x 10m4 M dithiothreitol and was sonicated for 15 s at 0°C with a needle probe of a Biosonik sonicator at 20% maximum voltage on the low setting. Ten-microliter portions of the sonicates were assayed for prolyl hydroxylase in a total volume of 0.1 ml. This amount of enzyme was in the linear portion of a plot of tritium release per 15 min versus enzyme concentration. In all assays, less than 5% of tritium in the substrate added to the incubation was released as tritiated water, and substrate was present at saturating levels. RESULTS
The time course of collagen and noncollagen protein synthesis and secretion in the presence or absence of 1 x lop5 M ascorbate is depicted in Figs. 1 and 2. Figure 1A shows the incorporation of [‘*C]proline into total collagen as a function of time; Fig. 1B shows the incorporation into cell and medium collagen individually. In the absence of ascorbate, incorporation into collagen stopped at 2 h, after which the total collagen content decreased slightly, suggesting that collagen was degraded. In cells incubated in the presence of ascorbate, incorporation continued for 4 h, 2 h longer than in the untreated cells, with most of the additional radioactive collagen appearing in the medium (Fig. 1B). The reason for the cessation of incorporation into collagen at 2 h in the absence of ascorbate is not clear at this time, although it might be explained by increased degradation of underhydroxylated collagen. The
FIG. 1. Time course of incorporation of [Wlproline into collagen. Growth medium was removed from log-phase cells and replaced with 3 ml of MEM-0 and either with (circles) or without (triangles) 1 x lo-” M ascorbate. After 15-min preincubation at 37°C 1.5 PCi of 14C was added to each plate and the incubation continued for the time indicated. Duplicate samples of two plates (100 X 26 mm) of mid-log-phase cells were analyzed from each time point for collagen as described in Materials and Methods. (A), Total collagen; (B), cell layer (closed symbols, solid line); medium (open symbols, dashed line).
intracellular level of collagen reached a maximum at 2 h in control cells while a maximum was reached at a lower level and earlier time (1 h) in cells treated with ascorbate, indicating that collagen accumulated in control cells (Fig. 1B). The total incorporation of [‘*Clproline into noncollagen protein is shown in Fig. 2A. The rate and amount of noncollagen protein synthesized was slightly higher at 2 and 3 h in the absence of ascorbate. There is no effect of ascorbate on the secretion of noncollagen protein (Fig. 2B). The percent of total collagen secreted into the medium in the presence and absence of ascorbate is depicted in Fig. 3. It is evident that both in the presence and absence of ascorbate there was a lag phase before any collagen appeared in the medium and that ascorbate markedly stimulated the secretion of collagen between 1 and 2 h. In order to determine whether collagen measured in the cell layer was true intracellular collagen as differentiated from collagen that had been secreted into the me-
FIG. 2. Time course of incorporation of [14C]proline into noncollagen protein. The experiment was performed as indicated in the legend to Fig. 1. Cells preincubated with ascorbate are indicated by circles and cells without ascorbate by triangles. (A), Total noncollagen protein; (B), cell layer (closed symbols, solid line); medium (open symbols, dotted line).
FIG. 3. Time course of secretion of collagen into the medium. The experiment was performed as described in the legend to Fig. 1. At each time indicated the percent collagen secreted was calculated from the following formula: Collagen secreted (o/o) = [dpmfmedium collagen)/(dpm(cell collagen) + dpm(medium collagen))] x 100%. Cells treated with 1 x 10m5 M ascorbate (0-O) and control cells (A-A).
dium and then precipitated on the cell surface, the following experiment was performed: [14C]proline-labeled collagen in the medium of cells incubated for 2 h was transferred to cells pretreated with 2 x
10e4 M cycloheximide for 15 min. The unlabeled cells with the medium containing radioactive collagen were allowed to incubate for an additional 2 h, and then cells and medium were analyzed as described in Materials and Methods. It was found that only 5-9% of the medium collagen became associated with the cell layer after 2 h of incubation (Table I). Therefore, in cells secreting 20% of the collagen synthesized in 2 h, only l-2% of the total collagen would adhere to the cell surface. In order to determine the relationship between proline hydroxylation and collagen secretion, experiments were performed in which proline hydroxylation and collagen secretion were studied as a function of ascorbate concentration. Using purified bacterial collagenase as a probe for collagen enabled us to measure the amount of collagen regardless of the degree of proline hydroxylation. Cells were incubated in the presence of varying concentrations of ascorbate for 2 h, and subsequently radioactivity in proline and hydroxyproline was measured in acid-hydrolyzed collagenase digests of the medium and cell fractions. Figure 4A shows the percent hydroxylation of proline in total collagen, and Fig. 4B, the percent hydroxylation of proline in the cells and the medium individually, while the percent secretion of collagen is shown in Fig. 5; each parameter is described as a function of ascorbate concentration. Due to the sharp transition from basal to maximal levels of hydroxylation, no experimental points could be obtained between 25 and 45% hydroxylation; however the estimated halfmaximal ascorbate concentration, K,,,, for stimulation of hydroxylation and secretion (Figs. 4A and 5) are similar for both processes, 4.5-8 X 10e7 M. In order to compare the effects of ascorbate on collagen hydroxylation and secretion with a system in which hydroxylation was varied by other means, experiments were carried out with varying concentrations of the iron chelator a,a’-dipyridyl. This compound has been shown to inhibit proline hydroxylation and collagen secretion (17, 18). These experiments were performed in the presence of 1 x lop5 M ascor-
(1 x 1O-5
I TO CELLSO
Collagen in cell layer (%)
(2 Medium from cells that had been treated with 1 x lo-’ M ascorbate or without ascorbate and containing radioactive collagen was added to log-phase cells that had been preincubated for 15 min with 2 x 10m4 M cycloheximide in MEM-0 and an additional 2-h incubation at 37°C was carried out. Medium was removed, the cell layer rinsed, and the rinse added to the incubation medium. Medium and cells were then analyzed for collagen as described in Materials and Methods. 60
FIG. 5. Collagen secretion as a function of ascorbate concentration. The abscissa is a logarithmic scale. Collagen secretion (%) = [medium collagen/(cell collagen + medium collagen)] x 100%.
E y rs g
0 IO Y 01 0
I I ASCORSATE
I I 5 IO lx IO’M)
FIG. 4. The hydroxylation of proline in collagen as a function of ascorbate concentration; the abscissa is a logarithmic scale. (A), Total collagen; (B), cell (0) and medium (0) collagen. Broline hydroxylation (%) = [hydroxyproline/(hydroxyproline + proline)] x 100%. Radioactivity in collagen proline and hydroxyproline was determined in collagenase digests of cell and medium protein.
bate in order to observe a broader range of inhibition of hydroxylation. The concentration of a,a’-dipyridyl resulting in half-maximal inhibition of proline hydroxylation
(Fig. 6) and secretion (Fig. 7), K,,,, were identical: 5.6 x 10m5M. Dehm and Prockop have found previously that hydroxylation and secretion of collagen in isolated tendon cells were simultaneously inhibited by a&-dipyridyl at approximately this same concentration (17). A composite of the data resulting from experiments varying ascorbate and CY,(Y’dipyridyl and plotted in the form of collagen secretion as a function of proline hydroxylation is presented in Fig. 8. These data were obtained from four individual experiments with chick embryo fibroblasts at different passages and derived from three separate groups of embryos over a 6month period. There is no response of secretion to hydroxylation until approximately 30% of the proline residues are hydroxylated. Collagen secretion is directly proportional to proline hydroxylation at any point over 30% hydroxylation. The constant level of collagen secretion at all levels of proline hydroxylation below 30% suggested that underhydroxylated col-
FIG. 6. The hydroxylation of proline in collagen as a function of qa’-dipyridyl concentration in the presence of 1 x 1W5 M ascorbate. The abscissa is a logarithmic scale. (A), Total collagen; (B), cell (0) and medium (0) collagen. Proline and hydroxyproline were determined as described in the legend to Figure 4. I 90
a c 5 ;
hydroxylated collagen are secreted via a pathway involving microtubules. In order to determine whether ascorbate stimulated proline hydroxylation and secretion by causing the formation of more active enzyme in chick embryo fibroblasts, as has been found in other cell lines (8,22), prolyl hydroxylase activity was measured in sonicates of chick embryo cells which had been incubated in the presence or absence of 1 x 10m5 M ascorbate for l-3 h. Table III demonstrates that the specific activity of prolyl hydroxylase did not increase over control levels during the period of ascorbate treatment. Log-phase L929 cells treated similarly showed a three-fold stimulation of activity (not shown). The specific activity of prolyl hydroxylase in chick cells was five times higher than maximal levels in L929 cells indicating that prolyl hydroxylase in chick embryo fibroblasts might already be fully activated. The variation of ascorbate transport as a function of ascorbate concentration was studied in order to determine whether the sharp transition in proline hydroxylation observed between 4 and 5 x 10m7 M (Fig. 4A) might be due to a sharp increase in ascorbate transport in that concentration range. In addition, the fact that the K,,, for hydroxylation which we observed was 700-fold lower than the K, for purified enzyme (23, 24) suggested that the cells might be concentrating ascorbate. We measured the [14C]ascorbate uptake as a
I I a.&D,pyr,dyl
FIG. 7. Collagen secretion as a function of a,o’dipyridyl concentration in the presence of 1 x 10m5 M ascorbate. The abscissa is a logarithmic scale.
lagen might be secreted into the medium by a pathway different from that of fully hydroxylated collagen. Colchicine, which has been shown to inhibit collagen secretion without inhibiting proline hydroxylation (19-21) was used to examine this question. Colchicine at 1 x 10m5 M inhibited the secretion of collagen from both control cells and those cells treated with ascorbate to an equal extent (Table II) suggesting that both hydroxylated and under-
FIG. 8. Collagen secretion as a function of collagen proline hydroxylation. Composite of data from experiments shown in Figs. 4-7.
TABLE THE INHIBITION Ascorbate
237 203 117 163
Cell collagen (dpmlpg of DNA)
collagen of DNA)
Total collagen (dpm/pg of DNA)
Collagen secreted (%I
371 226 335 193
36.1 10.3 67.3 15.6
134 23.3 238 30.1
Inhibition of aecretion (%I
a Mid-log-phase cells were preincubated for 15 min with no additions, 1 x 1O-5 M ascorbate, 1 X 10m5 M colchicine or a combination of 1 x 10e5 M ascorbate and 1 x 10e5 M colchicine. The cells were labeled for 2 h with [Wlproline, and the medium and cells analyzed as described in Materials and Methods. TABLE THE
ASCORBATE Ascorbate + + +
ON PROLYL Incubation time (h) 1 1 2 2 3 3
Specific activity ((dpm/mg x 10-5) c SD) 1.64 1.78 1.65 1.67 1.76 1.81
2 ” + + k 2
0.21(4jb 0.01(2) 0.12(4) O.lO(4) 0.0X2) 0.05(4)
a Duplicate samples consisting of two 100 x 20mm plates per sample of mid-log-phase chick embryo flbroblasts were incubated for varying periods of time in the presence or absence of 1 x 10m5 M ascorbate in MEM-0. Cells were harvested as described in Materials and Methods. The cells were sonicated for 15 s and lO+l portions of the sonicate were assayed for prolyl hydroxylase activity as described in Materials and Methods. Incubations contained 0.025 mg protein and 1.5 x lo5 dpm of substrate of which 80% was unhydroxylated collagen. * The numbers in parentheses indicate the number of experimental determinations.
function of extracellular ascorbate concentration (Fig. 9) and noted that log-phase cells could concentrate [‘4C]ascorbate approximately lo-fold. The [14C]ascorbate uptake was exponentially dependent on the concentration of ascorbate in the medium, but there was no sharp transition in uptake with increasing concentration. In calculating the intracellular volume, as indicated in the legend of Fig. 9, a variation of a factor of 1.5 was found between the packed volume at 2000g and the extrapolated volume at lg. When one considers the ability of the cells to concentrate V4Clascorbate lo-fold and the uncertainty in our volume determination, the K,,,
2 ASCORBATE l107x Ml
FIG. 9. The cellular [Wlascorbate uptake as a function of the medium ascorbate concentration; the abscissa is a logarithmic scale. [Wlascorbate at increasing concentrations in 3 ml of MEM-0 was added to single 100 x 20-mm tissue culture plates. Cell volume was determined in the following manner: 10’ cells were suspended in 1 ml of MEM-0 and placed in a l-ml graduated macrohematocrit tube. The suspension was centrifuged at various speeds resulting in centrifugal forces varying from 3Og to 2OOOg. The volume of the cell pellet was determined from the graduations on the hematocrit tube. A plot of cell volume as a function of centrifugal force was drawn and extrapolated to lg which was assumed to be the best approximation of cell volume. Cellular ascorbate concentration could then be calculated from the data (amount of ascorbate in cell layerl/[(volume per cell) x (number of cells)].
which we observed for proline hydroxylation, 4.5 x lop7 M, in the medium, in fact, may be 45-68 x lop7 M in the cell. This approximated K,,, for hydroxylation in the cell is still 50-70-fold lower than values reported for purified enzyme (23, 24). DISCUSSION
Our present studies with chick embryo fibroblasts clarify the mechanism of ascorbate stimulation of collagen secretion and the role that proline hydroxylation plays in the secretion of collagen. The results lead to the conclusion that ascorbate stimulates secretion through its effect on the hydroxylation of proline and that there is no intermediary effect on prolyl hydroxylase. Even in the absence of ascorbate, control cells showed a low level of hydroxylation (Fig. 4A). This can probably be attributed to the presence of an endogenous cofactor rather than to the presence of ascorbate in the medium since incubations were carried out in the absence of serum, the only possible source of exogenous ascorbic acid. This low level of proline hydroxylation resulted in secretion of collagen at a slow rate. The secretion of collagen occurred more rapidly in cells treated with ascorbate compared to untreated control cells. The time course of collagen synthesis and secretion showed that the biggest difference between the relative amount of collagen secreted by ascorbate treated and control cells occurred l-2 h after [14C]proline addition (Fig. 3). All subsequent experiments were performed 2 h after P4Clproline incorporation. In several experiments, we observed that the amount of collagen secreted after 2 h in control cells varied from 20 to 40%. This variation was related to a variation in the hydroxylation of the collagen suggesting that different cultures might contain varying amounts of endogenous cofactor. Experiments in which ascorbate and cy,a’-dipyridyl concentrations in the culture medium were varied over a wide range clearly showed the direct relationship between proline hydroxylation and collagen secretion. The control by ascorbate of proline hydroxylation and collagen secretion occurred over exactly the same concentra-
tion range. The estimated K,,, for secretion was approximately the same as it was for hydroxylation, 4.5-8 x lo-’ M. The stimulation of hydroxylation by ascorbate which we observed does not appear to be due to the cellular activation of prolyl hydroxylase that has been described in L929 and 3T6 fibroblasts (8, 22) but rather by ascorbate directly participating in the hydroxylation reaction. When proline hydroxylation was inhibited by a,a’-dipyridyl, secretion was inhibited in a parallel fashion. The concentration range over which a,a’dipyridyl in the presence of saturating ascorbate inhibited both hydroxylation and secretion was exactly the same, and the K,,, was the same for both processes. Even in the presence of saturating ascorbate, cY,a’-dipyridyl inhibited secretion and hydroxylation, indicating that ascorbate could not stimulate the secretion of underhydroxylated collagen. It is obvious that ascorbate does not function to stimulate secretion directly but only through its effect on proline hydroxylation. The functional relationship between hydroxylation and secretion is emphasized in Fig. 8 which shows that collagen secretion is directly dependent on proline hydroxylation after approximately 30% of the proline residues have been hydroxylated; this is about 60% of the proline residues available for hydroxylation in the collagen molecule. Several laboratories have shown recently that the hydroxylation of proline residues results in an increase in the thermal stability of the collagen triple-helical structure (25-27), and that the melting temperature, T,, of the collagen molecule is directly proportional to the extent of proline hydroxylation (28). It has been suggested that the reason why the inhibition of proline hydroxylation prevents optimal secretion is because underhydroxylated chains cannot form stable helical molecules at 37°C (26) and that the triple-helical conformation is necessary for secretion (4, 29). Estimates made from the data of Rosenbloom et al. (28) indicate that at 37°C essentially no triple-helical molecules exist if the proline hydroxylation of the collagen molecules is not above 35%. Their melting curves show that at 37°C the amount of
triple-helical molecules would increase with increasing hydroxylation. Based on these experimental facts, our data in Fig. 8 might reflect a linear relationship between secretion and the amount of helical collagen formed above hydroxylation levels of 30% at 37°C. The dependence of secretion on the level of proline hydroxylation is also evident in Fig. 4B, in which it may be seen that at ascorbate concentrations below 4 x 10m7M the proline in secreted collagen is at least 30% hydroxylated and is approximately 1.5 times more hydroxylated than the proline in intracellular collagen. These data further support the suggestion that the molecules which are secreted are those which can preferentially form stable triple-helical molecules. The half-maximal concentration at which ascorbate stimulated hydroxylation, 4.5 x 10e7 M, is 700-fold lower than values obtained for purified enzymes (23, 24). Although this discrepancy could in part be due to the ability of the chick embryo fibroblasts to concentrate ascorbate, we have shown that there is at most a lo15-fold greater concentration of ascorbate in the cells than in medium. There is evidence showing that prolyl hydroxylase is localized in the endoplasmic reticulum (30-32) and it is possible that ascorbate could be further concentrated in this organelle, which would explain the apparent lower K,. A second possibility is that the K, determined in in vitro enzyme assays is inaccurate because of oxidation of a large fraction of the added ascorbate. ACKNOWLEDGEMENT lent
We thank Miss assistance.
for her excel-
REFERENCES G. C. (1967) in The Vitamins (Sebrell, W. H., Jr., and Harris, R. S., eds.), Vol. 1, p. 415, Academic Press, New York. ROBERTSON, W. VAN B. AND HEWITT, J. (1961) Biochim. Biophys. Acta 49, 404-406. PETERKOFSKY, B. (1972)Arch. Biochem. Biophys. 152,318-328. RAMALEY, P. B., JIMENEZ, S. A., AND ROSENBLOOM, J. (1973) FEBS Lett. 33, 187-191. STONE, N., AND MEISTER, A. (1962) Nature (London) 194, 555-557.
2. 3. 4. 5.
7. PETERKOFSKY, Commun. 8.
9. 10. 11.
16. 17. 18.
B. (1972) Biochem. Biophys. Res. 49, 1343-1350. STASSEN, F. L. H., CARDINALE, G. J., AND UDENFRIEND, S. (1973) Proc. Nut. Acad. Sci. USA 70, 1090-1093. PETERKOFSKY, B., AND DIEGELMANN, R. F. (1971) Biochemistry 10, 988-994. PETERKOFSKY, B., AND PROCKOP, D. J. (1962) Anal. Biochem. 4, 400-406. BURTON, K. (1956) Biochem. J. 62, 315-323. HARTREE, E. F. (1972) Anal. Biochem. 48, 422427. LOWRY, 0. H., ROSEBROUGH, N. J., FARR, A. J. AND RANDALL, R. J. (1951) J. Biol. Chem. 193, 265-275. PETERKOFSKY, B., AND DIBLASIO, R. (1975)AnaZ. Biochem. 66, 279-286. HUTTON, J., TAPPEL, A. L., AND UDENFRIEND, S. (1966) Anal. Biochem. 16, 384-394. BERG, R. A., AND PROCKOP, D. J. (1973) Biochemistry 12, 3395-3401. DEHM, P., AND PROCKOP, D. J. (1971) Biochim. Biophys. Actu 240, 358-369. MARGOLIS, R. L., AND LUKENS, L. N. (1971) Arch. Biochem. Biophys. 147, 612-618. DEHM, P., AND PROCKOP, D. J. (1972) Biochim. Biophys. Actu 264, 375-382.
20. DIEGELMANN, R. F., AND PETERKOFSKY, B. (1972) Proc. Nut. Acud. Sci. USA 69, 892-896. 21. EHRLICH, H. P., AND BORNSTEIN, P. (1972) Nature New Biol. 238, 257-260. 22. LEVENE, C. I., ALEO, J. J., PRYNNE, C. J., AND BATES, C. J. (1974) Biochim. Biophys. Actu 338, 29-36. 23. ABBOT, M. T., AND UDENFRIEND, S. (1974) in Molecular Mechanisms of Oxygen Activation (Hayaishi, O., ed.), p. 173, Academic Press New York. 24. KIVIRIKKO, K. I., BRIGHT, H. J., AND PROCKOP, D. J. (1968) Biochim. Biophys. Actu 151, 558567. 25. BERG, R. A., AND PROCKOP, D. J. (1973) Biochem. Biophys. Res. Commun. 49, 1343-1350. 26. JIMENEZ, S. A., HARSCH, M., AND ROSENBLOOM, J. (1973) B&hem. Biophys. Res. Commun. 52, 106-114. 27. WARD, A. R., AND MASON, P. (1973) J. Mol. Biol. 79, 431-433. 28. ROSENBLOOM, J., HARSCH, M., AND JIMENEZ, S. (1973) Arch. Biochem. Biophys. 158, 478-484. 29. UITTO, J., DEHM, P., AND PROCKOP, D. J. (1972) Biochim. Biophys. Actu 278, 601-605. 30. OLSEN, B. R., BERG, R. A., KISHIDA, Y., AND PROCKOP, D. J. (1973) Science 182,825-827. 31. DIEGELMANN, R. F., BERNSTEIN, L., AND PETERKOFSKY, B. (1973) J. Biol. Chem. 248, 65146521. 32. CUTRONEO, K. R., GUZMAN, N. A., AND SHARAWY, M. M. (1974) J. Biol. Chem. 249, 59895994.