Coordinate Transcriptional Regulation of Type I Procollagen Genes by Rous Sarcoma Virus*

Chicken embryo fibroblasts infected with a strain of Rous sarcoma virus containing a temperature-sensitive mutation in the gene coding for pp608'", a protein kinase, undergo changes in collagen synthesis within 4 h after a temperature shift. Cells shifted from the restrictive to the permissive temperature for transformation show decreasing levels of collagen synthesis and increasing levels of kinase activity; the reverse occurs when infected cells are shifted from the permissive to the restrictive temperature. Levels of type I procollagen mRNAs coding for proal and proa2 chains, meas- ured by hybridization to nick-translated cloned al and a2 cDNA, decreased simultaneously soon after a reduc- tion in temperature and reached a new steady state at about 50 h after the shift.

Chicken embryo fibroblasts infected with a strain of Rous sarcoma virus containing a temperature-sensitive mutation in the gene coding for pp608'", a protein kinase, undergo changes in collagen synthesis within 4 h after a temperature shift. Cells shifted from the restrictive to the permissive temperature for transformation show decreasing levels of collagen synthesis and increasing levels of kinase activity; the reverse occurs when infected cells are shifted from the permissive to the restrictive temperature. Levels of type I procollagen mRNAs coding for proal and proa2 chains, measured by hybridization to nick-translated cloned al and a2 cDNA, decreased simultaneously soon after a reduction in temperature and reached a new steady state at about 50 h after the shift.
In order to test for regulation at the transcriptional level, nuclei were isolated from normal and Rous sarcoma virus-transformed chicken embryo fibroblasts and allowed to transcribe in the presence of [a-32PIUTP.
Procollagen mRNA sequences in newly synthesized and in total RNA from transformed cell preparations were both about 5-fold lower than the levels in normal cell preparations. We conclude that the coordinate decrease in procollagen mRNAs observed in Rous sarcoma virus-transformed chicken embryo fibroblasts is caused primarily by a decrease in the transcription of the type I procollagen genes, a decrease which is directly or indirectly mediated by the pp6081' protein kinase.
If cells are infected with a mutant of the virus which has a src gene deletion, the virus integrates and replicates normally, but the biochemical changes associated with transformation do not occur (for a review, see Ref. 14).
It seems likely that a single viral gene which can produce such pleiotropic effects must code for a function which mimics a key element in cellular regulation. Investigators have found * This work was supported in part by National Institutes of Health Grants AM 11248 and DE 02600. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
that the src gene codes for a phosphorylated polypeptide of Mr = 60,000, denoted pp60.rc (15)(16)(17), which has an associated protein kinase activity (17)(18)(19). In addition, a structurally related protein kinase, termed pp60'rc, has been found in uninfected vertebrate cells (20)(21)(22)(23). pp60s rc activity is present in normal cells at about 1/5o the level of pp60`fC activity in RSVtransformed CEF (21,23,24). Because transformation specifically affects collagen synthesis, a major differentiated function of chicken embryo fibroblasts, it provides a system in which one can study the action of the pp60src kinase and perhaps the nature of normal regulation as well.
In this paper we describe a series of experiments undertaken to elucidate the mechanism of the decrease in collagen synthesis observed in transformed cells. We were particularly interested in a more detailed study of the events that take place in CEF undergoing transformation, the increase in pp60`rc kinase activity, the decrease in procollagen mRNA, and the decrease in collagen synthesis, and in determining whether the decrease in procollagen mRNA is correlated with a decrease in messenger RNA synthesis. Our results show that there is an inverse correlation between pp60src kinase activity and the levels of type I mRNA and collagen synthesis and that the reduced mRNA levels can be accounted for by a reduction in the transcription of al and a2 procollagen genes.

Vi ruses
The Prague A strain of RSV, PR-RSV-A, was a gift from Maxine Linial, Fred Hutchinson Cancer Research Center, Seattle, WA. RSV tsLA24 (25), a derivative of the Prague A strain which contains a temperature-sensitive mutation in the src gene, was also obtained from Dr. Linial. Infected cells are normal at the restrictive temperature of 41' C, but transformed at the permissive temperature, 35' C. This virus was subcloned and the four subclones producing the greatest temperature-dependent change in morphology were assayed for collagen synthesis at 351 C and 41' C. The clone which caused infected cells to have the lowest collagen synthesis (most transformed) at 35' C and the highest collagen synthesis (most normal) at 41°C was designated tsLA24-10 and was used in the studies described here.
Cell culture Primary cells were obtained fron chicken embryos which were chicken helper factor negative and resistant to sarcomna virus subgroup E (C/E) (H and N Farms, Redmond, WA). CEF cultures were started by dispersing the body wall of 11-day embryos by shaking at 37' C in Temperature shift of ts124-1 -infected CEF Measurements of collagen synthesis, pp60t5 activity, and procollagen mRNA levels were made at various times after a temperature shift. Cells were infected with tsL24-10 and maintained at 35' C or 41' C. After four days, morphological transformation of cultures maintained at 35' C was always complete. Cells were replated at subconfluent density without temperature shift and the medium was changed on the second day. On the third day, sets of cells were shifted from 41' C to 35' C or from 35' C to 41' C or maintained at 41' C or 35' C.
Immunoprecipitation of pp605Q.C. kinase activity pp60AOkinase activity was measured as the transfer of 32P by immunoprecioitated 32sr pp6AAO.Ct from [Ey-P]ATP to the anti-pp6yl-rimmunoglobulin gamma chain. Antiserum against pp6Afrom tumor-bearing rabbits was a gift from Larry Rohrschneider. The preparation and properties of this serum have been described in detail (26). The antiserum we emoloyed precipitated be+' iral pp6Osrc and the endogenous pp601Acprotein from uninfected chicken -iZ2). The preparation of cell extracts for assay of op6SOnkinase activity, -. assay itself,were performed as described by Collett and Erikson (19). Cell lysates were handled at 4' C. Immunoprecipitation was performed with 4 p1 of antiserum and 100 .1 of 100 mg/ml fixed Staphylococcus aureus; this procedure has been described in detail (27). All kinase reactions were incubated for 10 min at 30' C with 0.1 PM [y-32P]ATP (3000 Ci/nmole, Anersham). Following the reaction, immunoorecipitates were washed three times with buffer (19) and once with 0.01 M Tris, pH 7.5; electrophoresis sample buffer (28) was then added and immunoprecipitates were boiled for 2 min. Samples were centrifuged and the supernatants containing the phosphorylated immunoglobulin were recovered for gel electrophoresis.
Using this assay we determined that incorporation of 32P increased linearly with increasing input of cellular extracts between 0 and 200 pa of cellular protein; all measurements were therefore performed within this range. Protein was determined by a dyebinding assay (29).
Polyacrylamide gel electrophoresis, autoradiography, and radioactivity in the phosphorylated heavy chain Phosphorylated protein from the immunoprecipitates described above was fractionated on a 12.5% polyacrylamide gel (28). The gel was stained, destained, and dried onto Whatman 3 M1 paDer. Radioactivity was detected by exposure to Kodak X-Omat R film. Each portion of the gel containing phosphorylated immunoglobulin heavy chain was cut out and dissolved by incubation in a scintillation vial in 1 ml of 30' H202 for 1.5 h at 80' C. The vials were cooled, toluene-based scintillant was added and the radioactivity was counted.
Col lagen synthesis Collagen synthesis in the CEF cells was determined by collagenase assay described by Peterkofsky and Diegelmann (30). Cells were preincubated in Dulbecco's modified Eagle's medium (DMDM) supplemented with 50 pg/ml ascorbate and antibiotics for 1 hr and labeled at various times during a time course by incubation for 30 min in 20-25 ICi/ml of [2,3-3H]proline in the above medium. Cell layers were washed three times with cold phosphate buffered saline, solubilized in 0.5 M ammonium hydroxide, precipitated by bringing the solution to 10 trichloroacetic acid, and further processed as described (30). Digests (34) to check for host DNA contamination. Each plasmid preparation was also identified by restriction digestion and gel electrophoresis (34).
Procollagen-specific fragments were prepared by restricting the pCh54(al) plasmid with HindIII and Kpnl (32) according to conditions specified by the supplier (BRL), and pCg45(o2) with HindIII alone (31). Procollagen-specific fragments were separated from the remainder of the plasmids by electrophoresis in 0.71 agarose gels and were recovered from a slice of the gel by electroelution. Contaminants coelectroeluting from the agarose, which potentially interfere with subsequent enzymatic treatment of the fragment, were removed by chromatography over a small DEAE-cellulose column equilibrated in gel electrophoresis buffer. The column was washed further with this buffer and the DNA was eluted with buffer containing 1 M HaCl. The recovered fragment was preciDitated overnight with two volumes of ethanol at -20' C and centrifuged at 35K in an SW40 rotor (Beckman) at 4' C for 45 min.

Nick translation
Nick translation was carried out essentially as described by Thomashow et al. (35).
RNA isolation RNA was purified as described by Rowe et a]. (11). Total nucleic acid was obtained by proteinase K digestion in SDS followed by phenol/chloroform extraction and ethanol orecipitation. Nucleic acid was dissolved in water and precipitated with two volumnes of 3 M sodium acetate, 5 mM EDTA, pH 6.0, at 4' C for several hr. Precioitated RNA was washed once with the above salt buffer and twice with 60X ethanol, 0.1 M NaCl. The pellet was resuspended in water or 0.10 SDS.
Linkage of RNA and DNA to cellulose Aminobenzyloxymethyl cellulose (Miles) was converted to cellulose powder and diazotized to form diaminobenzyloxymethyl (DBM) cellulose as described by Noyes and Stark (36).
After diazotization, the DBM cellulose was washed once with cold water and twice with cold 0.2 M sodium phosphate, pH 5.5. The cellulose was then suspended at 10-50 mq/ml in 20 mM sodium phosphate, pH 5.5, for linkage to RNA. RNA, dissolved in 20 aM sodium phosphate, pH 5.5, was then added to the cellulose at a concentration of 10 ug Der mg. The linkage reaction proceeded with shaking at 4' C for 24-48 hr. RNA-cellulose was then incubated at 42' C for 4-12 hr in 50 deionized formamide, 0.75 M NaCl, 75 mM sodium citrate, 0.20 each of bovine serum albumin, ficoll, and polyvinylpyrrolidone, 100 pg denatured E. coli DNA/ml, and 10 glycine (37). This step was important in reducing backgrmwnd, probably by eliminating unreacted diazo groups. RNA linkage to cellulose as monltored by retention of [3H]RNA and A260 was greater than 951, up to concentrations of 16 og/ug cellulose.
DNA linkage was carried out in the same manner except that plasmid DNA was heated in a boiling water bath for 8 min to nick and denature the DNA, and the DOM cellulose was washed once with 20 mM sodium phosphate, pH 5.5, to reduce the salt concentration.
Hybridization assay of total cellular RNA Hybridizations were carried out between a nick-translated DNA fragment coding for a] and u2 sequences (see above) and RNA linked cowalently to cellulose. Hybridization reactions were performed in DNA excess in 80X deionized fornmaide, 0.4 M NaCl, 0.01 M PIPES, pH 6.4 (38), containing 100 og denatured E. coli DNAml. Hybridization assays contained 2-7 x 105 cpm of probe and 0.01-10 pg9 RPA in 200 plI. Reactions were covered with paraffin oil and shaken in round-bottom plastic tubes at SO' C for 24-72 hr. Under these conditions up to 900 of the cDNA could be driven into DNA-RNA hybrids. Following the hybridization reaction, the cellulose was washed three times in warm 0.60 M NaCl, 0.06 M sodium citrate containing 0.10 SDS, and 1 mM EDTA to remove most of the nonhybridized label. Background was further reduced by three additional centrifugations alternated with shaking in wash buffer for 30 min each at 60' C. Background, which was proportional to the quantity of cellulose, was determined by carrying a series of concentrations of blank cellulose through a mock hybridization and wash, and was subtracted from the raw hybridization cp.
Hybridization conditions were determined by a number of factors. A temperature of 50' C and addition of 100 wg/nl E. coli DNA favored specific hybridization. The formamide, salt and buffer conditions favored RNA-DNA hybridization and minimized degradation (38). Nick translation results in a significant percentage of snapback sequences (39). We observed 7-10 snapback sequences with circular plasmid substrate and as much as 25-30X snapback sequences with linear substrates such as the procollagen-specific restriction fragments used here. Since this would create an unacceptable background in solution hybridization, and greatly reduce the sensitivity of measurements which could be mode, we linked the RNA to cellulose so that only the DNA probe hybridized to RNA would be recovered. Cellulose powder was used in order to optimize contact between the DNA probe and matrix-bound R0A, and because on a weight basis the powder has a far higher capacity to bind RNA than does filter paper. Customarily, RNA was linked to cellulose in a batch reaction and then assayed at concentrations from 0.01-10 vg of RNA per reaction. Although hybridization is readily observed below 1 pg of R02A per hybridization reaction, we found that higher amounts of RNA gave greater sensitivity.
Nuclear transcription, isolation of R0A, and hybridization Nuclei were isolated essentially as described by Nulvihill and Palmiter (40)  RNA was isolated as described (41). The RNase-free DNase used in the isolation was a gift from Richard Palmiter, University of Washington, Seattle, WA. The procollagen CDNA plasmids, pCg54 and pCg45, were baked onto 7 me dianeter circles of nitrocellulose paper (Beckman) at and 1.2 pg/filter, respectively, according to the method of Gillespie (42). Hybridizations were in DNA excess. The conditions of hybridization were as described [0.5 M NaCl, 50  McKnight and Palmiter (41)], but volumes varied from 40-60 p1/assay. Three filters were included in each hybridization: a pCg54(wl), pCg45(u2), and p8R322 or blank filter.
Reactions were incubated at 45' C for three days. Nonspecific label was removed from the filters by washing and Roase A and T1 digestion and filters were counted as described (41). Backgrounds were between S and 10 ppm and were subtracted from al and a2 filter hybridizations.

RESULTS
Immunoprecipitation of Protein Kinase Activity from Uninfected and Infected Cells-A pp60Arc-associated protein kinase activity has been detected in both RSV-infected and uninfected vertebrate cells (20)(21)(22). This kinase activity has been measured by precipitation of the enzyme with antiserum against pp60`rc from tumor-bearing rabbits. The activity is about 50-to 100-fold greater in transformed than in normal cells (21,23,24). It is also temperature-sensitive in extracts from cells infected with a virus possessing a temperaturesensitive mutation in the src gene (17,19,(43)(44)(45). The phosphorylation of the IgG by this kinase may reflect the in vivo activity of the pp600rC protein, although the natural substrate(s) in the cell is unknown.
Nonimmune serum, reacted with infected and uninfected CEF extracts, did not result in phosphorylation of the IgG heavy chain. However, the heavy chain of IgG from anti-pp6Osrf serum was phosphorylated by extracts from uninfected cells and to a much larger extent by extracts from infected cells (data not shown). Quantitation of the cpm incorporated into the heavy chain revealed a 60-fold greater phosphoryla-tion of the IgG by infected than uninfected cell extracts. As a control, antiserum incubated with Staphylococcus aureus cells, but without cell extract, showed no ability to transfer 32p from [-y_32P]ATP to IgG.
Measurements of Kinase Activity and Collagen Synthesis in Normal and Transformed Cells after Shifts to Temperatures Permissive and Restrictive for Transformation-We have measured levels of procollagen mRNAs early after infection of CEF with RSV and have shown that these mRNA levels begin to decline about 24 h after infection (13). In order to eliminate the time required for provirus synthesis, integration, and gene expression (see Ref. 46 for a review) and to observe directly the effects of transformation on collagen synthesis and mRNA levels, we have infected CEF with a mutant temperature-sensitive (ts) for transformation, tsLA24 (25). We then shifted the cells to temperatures both permissive and restrictive for transformation and measured levels of collagen protein and mRNA and pp60Src kinase activity over a period of 2-3 days.
The amount of immunoprecipitable kinase activity in uninfected and infected cells during the course of a temperature shift was determined. Since the antiserum reacts with endogenous pp60sarc, as well as with viral pp60src, the activity of uninfected cells was subtracted from that of infected cells. Normal and transformed cells were at approximately the same density throughout the time course, as determined by protein content/plate, so no corrections were necessary for changes in activity of pp60sarc with different cell densities. The normalized results are shown in relation to changes in collagen synthesis over the time course (Fig. 1, A and B). Pp60src kinase activity increased abruptly during the shift to the permissive temperature (Fig.1A) and decreased abruptly during the shift to the restrictive temperature (Fig.1B). The 3-to 6-fold range of kinase activity between restrictive and permissive conditions has been observed by other investigators (44,47). The difference is less than that observed between wild type RSVtransformed CEF and normal CEF, probably because the temperature-sensitive transforming protein is less active than the wild type protein at the permissive temperature and has residual activity at the restrictive temperature. transcriptional control we measured al and a2 procollagen mRNA levels in normal and transformed cells by hybridization with nick-translated procollagen-specific fragments of pCg54 and pCg45, respectively. Fig. 2 shows that the extent of hybridization was directly proportional to RNA up to 20,tg of total RNA/assay. The extent of hybridization was also dependent on DNA input and time. RNA from transformed cells contained 7.6-fold less al procollagen mRNA and 4.3-fold less a2 procollagen mRNA than RNA from normal cells.
Since the al and a2 collagen chains show considerable homology (for a review, see Ref. 48), it was necessary to show that the two mRNAs could be measured independently in our assay. We prepared nick-translated fragments from pCg54 and pCg45, representing the al and a2 sequences, respectively, and hybridized 500,000 cpm of each to whole pCg54 and pCg45 plasmids attached to cellulose (10 yg of pCg54/mg and 12. gag of pCg45/mg). Low cross-reactivity between the probes was, in fact, observed (Table I). The al probe hybridized to pCg45 at 1.3% of the level to which it hybridized to its complementary plasmid, pCg54. The ca2 probe hybridized to pCg54 at 8.1% of the level at which it hybridized to its complementary plasmid, pCg45. Since some contamination of the nick-translated fragments with pBR322 sequences is likely, these estimates provide upper limits of the extent of cross-hybridization in the actual RNA assays.
Temperature Dependence ofProcollagen mRNA Levels in tsLA24-10-infected Cells oal and a2 procollagen mRNA levels in tsLA24-10-infected CEF shifted from  'C were measured by hybridization assay. Fig. 3 shows the kinetics of the change in relative concentration of procollagen mRNA in RNA isolated from cells before the temperature change and up to 50 h after the shift to 35 2C. Even at early times after the shift to the permissive temperature, al and a2 procollagen mRNA levels had dropped from values observed at 41 nC. The decrease is most dramatic between 5 and 10 h after temperature reduction but there was some variability in timing. By 50 h after the temperature shift, proul and proC52 mRNA levels had decreased to 26 and 32% of the original levels, respectively. In contrast, uninfected cels subjected to a similar shift in temperature showed no reduction in at and a2 procogtagen mRNA levels as measured by hybridization (data not shown).
Regulation of Transcription of the Type I Procollagen Genes by RSVrTo ascertain whether the decrease in coliagen synthesis and procooagen mRNA levels is due to a decrease in the rate of transcription of the type I procollagen genes, we isolated nuclei from normal and transformed CEF and allowed  3. Effect of shift to the permissive temperature on al and a2 procollagen mRNA levels in tsLA24-10-infected CEF. tsLA24-10-infected CEF were infected and maintained at 41 'C as described under "Experimental Procedures." Control infected cells were also kept at 35 'C to monitor the course of the infection. After 4 days cells were replated at 2 X 106 cells/135-mm dish. Cells were temperature-shifted on the second day after replating. RNA was isolated from tsLA24-10-infected CEF at 0, 0.5, 1, 2, 5, 10, 27, and 50 h after shifting from the restrictive temperature (41 'C) to the permissive temperature (35 'C) and attached to cellulose at a concentration of 10 gg/mg. al (-4) and a2 (O---0) mRNA levels were assayed by hybridization to al and a2 nick-translated probes for 2 days; 5 x 105 cpm were used in each assay. them to transcribe in the presence of [ac-32P]UTP; transcripts were isolated and hybridized to al and a2 plasmids bound to filters. Fig. 4 shows that hybridization of al and a2 RNA sequences synthesized in vitro by nuclei from normal CEF increased with input up to 3 x 106 cpm, indicating that plasmid DNA is in excess.
In another experiment, we compared the relative level of procollagen mRNA transcription and steady state mRNA levels of normal and transformed cells (Fig. 5). The results of hybridization of nick-translated al and a2 cDNA fragments to RNA bound to cellulose are shown in Panel A (Fig. 5). The al and a2 mRNA in transformed cells are present at 13 and 24% of the level observed in normal cells, respectively. These levels correlate well with the level of collagen synthesis observed in transformed cells, about one-seventh of the level in normal cells. Nuclei from duplicate plates were isolated and transcribed in vitro. Total incorporation was similar for normal and transformed nuclei. On the basis of ppm of input hybridized, the relative rate of al and a2 transcription in transformed nuclei was about 15 and 21%, respectively, of that C) W obtained with control nuclei (Fig. 5, Panel B). These data suggest that reduced transcription of procollagen genes is a major factor in determining the levels of al and a2 messenger RNAs in RSV-transformed cells.
To explore the possibility that nuclear RNA degradation was a factor contributing to the reduction in the levels of newly synthesized procollagen RNA sequences in nuclei from transformed cells, we performed two types of experiments. Since a time course of total incorporation indicated that most rapid incorporation occurred during the first 10 min of transcription (Fig. 6A), we reasoned that procollagen mRNAspecific nuclease activity would be apparent as a decrease in parts per million hybridized to the procollagen plasmid filters with increasing time of transcription incubation. Nuclei from normal and transformed cells were incubated for 10, 20, and 30 min and their respective RNAs were isolated. No significant difference was observed in the parts per million of procollagen RNA hybridized over this time course (Fig. 6B). Hybridization input wa bNuclei were mixed 10 r c Nuclei were mixed 10 r In the second experim normal and transformec fore or after the transcr transformed cell or con cleases responsible for seemed likely that we w tion of RNA in preparat (B) the transcription r pp60src protein somehom nuclei, this would be apr not after (B), the trans determine the expected transformed cell nuclei RNAs were isolated fro transformed cell nucleair filters. Some reduction hybridization to the axl, RNA from nuclei mixed cant difference was obse: the transcription reacti the possibility that pro( could account for the d scription observed in tr2 Although it was previ( necessary for maintena review, see Ref. 14), this be correlated, in the cou effect of viral transforn protein, collagen. Sefton pp60src activity, as meas assay, correlated genera They reported kinase a that in tsNY68-infected 30-50% of the activity of pp60srC kinase activity v formants than in norn infected cells also showe even at restrictive temI and a 3to 6-fold differer tive and permissive ter with a 2to 4-fold differ between the tsLA24-10 permissive temperature though a time course of change in pp60src activity has not 32Pphybridizedto procollagencDNA been correlated with a specific transformation phenotype, Radke and Martin (49) have reported a phosphoprotein of al plasmid % con-2 plasmid % con-36,000 daltons which increases as soon as 20 min after shifting trol cpm tro tsLA29-infected cells to permissive temperatures. At Table II, nuclei from h, when the first measurements were made. Activity continued 1 cells were mixed and incubated beto decrease or increase until between 24 and 36 h when it iption labeling period. If nuclei from appeared to level off. This activity correlated inversely with Ltaminating cytoplasm contained nu-the changes we observed in collagen synthesis. degrading procollagen transcripts, it The low levels of collagen synthesis in CEF after RSV ould observe a decrease in hybridiza-transformation, about 14% of normal levels, closely parallel tions mixed both before (A) and after the decrease in procollagen mRNA levels. Hybridization aseaction. If, on the other hand, the says revealed that transformed cell mRNA contained about v reduced transcription in the normal 13% of the al sequences and 23% of the a2 sequences present parent in nuclei mixed before (A), but in normal CEF (Fig. 5). The difference in mRNA levels ;cription labeling period. In order to between transformed and normal CEF is less than the 10-fold levels of hybridization, normal and decrease in translatable procollagen mRNAs (9,11) and the were allowed to transcribe separately; 20-fold decrease in procollagen mRNA sequences observed by m each, and a mixture of normal and Northern blotting with cloned cDNAs (10). That we observed r RNA was hybridized to the plasmid a smaller decline in collagen specific mRNA sequences is in hybridization is apparent in the probably due to the fact that our assay measures total (intact but not to the A2, plasmid filters, by and partially degraded) mRNA sequences while the transla-1 before the transcription. No signifi-tion and blot assays measure unnicked, complete mRNA rved when the nuclei were mixed after molecules. on. These experiments argue against We were interested in determining the mechanism of this collagen-specific mRNA degradation decrease in procollagen mRNA levels. We began by examining lecrease in procollagen mRNA tran-the change in mRNA levels in tsLA24-10-infected cells shifted ansformed nuclei. from a temperature restrictive to one permissive for transformation. Both al and a2 mRNA levels decreased most rapidly DISCUSSION between 5 and 10 h after the temperature shift (Fig. 3). These ously known that src gene function is results are in general agreement with the decrease in collagen Lnce of the transformed state (for a synthesis and probably also reflect the increase in pp60`rC study shows that pp60srC activity can activity. The change in procollagen RNA levels which accomirse of a kinetic experiment, with the panies transformation may be attributable to changes in pronation on the synthesis of a specific collagen mRNA synthesis since RNA isolated from transet al. (47) reported that the extent of formed cell nuclei contained 15 and 21% of the level of newly 3ured by the y chain phosphorylation synthesized al and a2 sequences, respectively, found in nuclei tlly with the state of transformation. from normal cells. These reductions paralleled those in xl and Lctivity in normal CEF at 2-3% and a2 mRNA in total RNA in transformed cells, suggesting that 1 CEF at restrictive temperatures at the decreased transcription of the type I procollagen genes is fully transformed cells. In our studies, directly responsible for the observed reduction in al and a2 was also much higher in RSV trans-mRNAs.
ial CEF. Our data with tsLA24-10-Several experiments were performed to examine the possid significantly higher kinase activity, bility that pulse-labeled procollagen mRNA sequences were peratures, than with uninfected cells degraded in transformed nuclei. First, we demonstrated that nce between pp60src activity at restricthe parts per million of labeled nuclear RNA which hybridized nperatures (Fig. 1). This correlated to al1 and a2 plasmids were constant over a 30-mmn period for ence in the level of collagen synthesis preparations from both normal and transformed cells, despite -infected cells at the restrictive and the fact that synthesis is greatest early in the time course (Fig.   !s. We observed no evidence for the 6). In a second control experiment, nuclei from normal and transformed cells were mixed before or after transcription. When nuclear RNA, isolated separately from normal and transformed cells and from cells mixed before transcription, was compared, there was no significant difference in the extent to which the RNA hybridized to cDNA plasmids containing al or a2 sequences (Table II). We therefore concluded that procollagen-specific nucleases were not responsible for the decrease observed in procollagen RNA sequences synthesized in transformed nuclei. However, we currently cannot exclude additional regulatory elements which could function at the level of the processing of mRNA precursors or alter the stability of mRNA thereby affecting translational efficiency.
The RSV-transformed chicken cell offers a unique system for studying collagen regulation. We have observed an inverse correlation between collagen synthesis and pp60src kinase activity in these cells and have shown that transformation alters expression of the collagen genes at the level of transcription. We would like to know whether this change occurs at the level of chromosome structure and are attempting to determine whether the type I collagen genes have decreased DNase I sensitivity (51) after cellular transformation by RSV. We are also examining whether RNA polymerase II or high mobility group proteins 14 and 17, which confer DNase I sensitivity to actively transcribed genes (52,53), are differentially phosphorylated upon transformation. Since there is a highly conserved homologous kinase in most higher eukaryotic cells, it would also be interesting to determine the correlation between this kinase and collagen synthesis at different cell densities in tissue culture or during development.