Growth Arrest Induced by Transforming Growth Factor 81 Is

THEJOURNAL OF BIOLOGICAL CHEMISTRY 0 1991 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 266, No. 6, Issue of February 25, pp. 3444-3448, 1991 Printed in (1. S.A.

Growth Arrest Induced by Transforming Growth Factor 81 Is Accompanied by Protein Phosphatase Activation in Human Keratinocytes” (Received for publication, October 17, 1990)

Philip A. GruppusoSQV,Ryoko Mikumoll, David L. Brautigans, and Lundy Braunll From the Departments of $Pediatrics, §Biochemistry, and 11 Pathology, Brown University, Providence, Rhode Island 02912

Protein phosphorylation and dephosphorylation are involved in regulation of cell growth. We tested the hypothesis that the growth inhibitory effect of transforming growth factor 81 (TGF-Dl)involves activation of protein phosphatases. Exposure of human keratinocytes in culture to 400 PM TGF-Dl for 48 h led to 80% inhibition of DNA synthesis as measured by nuclear labeling. Incubationof cultured keratinocytes with 400 PM TGF-Bl rapidly activated (within 30 min) protein serinelthreonine phosphatase, measured using phosphorylaseas a substrate. Based on several criteria, including neutralization of activity with specific antibodies and inhibitor-2, TGF-Dl-activated phosphorylase phosphatase was identified as protein phosphatase 1. TGF-01 did nothave rapid effects on protein serine/ threonine phosphatase activity(type 2A) measured with histone phosphorylated byprotein kinase C or on protein tyrosine phosphatase activity. However, protein tyrosine phosphatase was activated at 48 h, coincident with growth arrest. Differentiation, induced by the combination of TGF-81 plus calcium or by serum, was not accompanied by further serinelthreonine or tyrosine phosphatase activation. We conclude that induction of growth arrest in keratinocytes by TGF-Dl involves acute activation of protein phosphatase 1, while activation of protein tyrosine phosphatases may represent an additional mechanism for maintaining cells in a growth-arrested state.

The family of polypeptides designated by the term transforming growth factor beta (TGF-@)’ are ubiquitous regulators of cell growth, which havediverse, cell type-dependent biological effects (for reviews, see Refs. 9 and 10). In epithelial cells, including human keratinocytes, a member of the TGF-P family, TGF-@1, actsasapotent growth inhibitor (11). The mechanisms by which TGF-P1exertsthis effect are not known, although recent studies have demonstrated rapid effects on the expression of c-myc (12, 13) and on the phosphorylation state of the retinoblastoma gene product, RB (14). Given the important role of serine/threonine and tyrosine protein kinases in the stimulation of cell proliferation, we hypothesized that the inhibition of keratinocyte growth by TGF-@Iinvolves protein phosphatase activation. The present report describes studies on the effects of TGF-P1 on human keratinocyte protein phosphatase activities. EXPERIMENTAL PROCEDURES

Culture of Human Keratinocytes-Early passage cultures of human keratinocytes were prepared from foreskin tissue as previously described (15). Cultures were established initially on feeder layers of irradiated mouse fibroblast3T3 cells using Dulbecco’s modified Eagle’s medium supplemented with 5% fetal bovine serum and 0.4 pg/ ml hydrocortisone. After reaching confluence, the keratinocyte cultures were detached with trypsin and replated without a feeder layer in serum-free keratinocyte growth medium (KGM, Clonetics Corp., San Diego, CA) supplemented with 5 pg/ml insulin, 0.5 pg/ml hydrocortisone, 10 ng/ml epidermal growth factor, and 0.4% bovine pituitary extract (referred to ascomplete KGM). Cells were plated a t a density of 1 X IO5 cells/35-mm Petri dish (except where noted) in complete KGM. On the following day, me400 pM TGF-01 Based on advances of the last several years, it appears that dium was removed, and fresh KGM with or without stimulation of cell proliferation may involve multiple protein (R&D Systems,Inc., Minneapolis, MN)was added. To monitor DNA phosphorylation cascades. The role of receptor-associated and synthesis, cultures were labeled for the last 24 h before harvesting with 5 pCi of [3H]thymidine andprocessed for autoradiography (16). non-receptor tyrosine kinases in the initiation of cell growth The percentlabeled nuclei was determined by counting a t least 1000 is established (1).Tumor-promoting phorbol esters stimulate cells in four high power fields. cell growth via activation of the protein kinase C family (2). Histochemical staining for keratin was performed as described by Signal transduction pathways also involve protein kinases Ayoub and Shklar (17). In this procedure, differentiated cells stain such as ribosomal protein S6 kinase (3), raf (4), casein kinase brilliant red and undifferentiatedcells stain blue. Preparation of Keratinocyte Lysates-Cells were lysed with 0.5 ml/ I1 (5), and microtubule-associated protein kinase (6). In ad- 35-mm well of a buffer containing 50 mM HEPES, 5 mM EDTA, 2% dition, the cell cycle-dependent protein kinase, cdc2 (7), ap- Triton X-100, 250 pg/ml phenylmethylsulfonyl fluoride, and 10 pg/ pears to be directly involved in the control of DNA synthesis ml leupeptin, pH 7.4. Following scraping, cells were broken by several and mitosis. Finally, several lines of evidence point to involve- strokes ina ground glasshomogenizer. The resulting homogenatewas ment of protein phosphatases in cell cycle control (reviewed diluted 1:2 (v/v)inassay buffers appropriate for measurement of protein serine/threonine phosphatase(18)and protein tyrosine phosin Ref. 8). phatase (19) activities. An aliquot of the undiluted homogenate was * This work was supported by United States Public HealthService retained for measurement of protein phosphatase heat-stable inhibibicinchoninic Grants HD24455 (to P. G.), DK31374 (to D.B.), and CA46617 (to L. tor activity (seebelow) and protein determination using B.), and Juvenile Diabetes Foundation International Research Grant acid (Pierce Chemical Co.). Samples were frozen in liquid nitrogen 188541 (to P. G . ) .The costs of publication of this articlewere defrayed The abbreviations used are: TGF-P, transforming growth factor in part by the payment of page charges. This article must therefore be hereby marked“aduertisement”inaccordancewith 18 U.S.C. p; KGM, keratinocyte growth medium; HEPES, N-2-hydroxyethylpiperazine-N’-2-ethanesulfonic acid; 12, inhibitor-2; RCML, reduced Section 1734 solely to indicate this fact. protein; 7 To whom correspondence should be addressed Dept. of Pediat- carhoxyamidomethylated lysozyme; MBP, myelinbasic MOPS, 3-(N-morpholino)propanesulfonicacid. rics, Rhode Island Hospital, 593 Eddy St., Providence, RI 02903.

3444

Protein Phosphatase Activation and TGF-B and stored at -70 "C until analysis 1-3 days later. For all experiments parallel incubations were carried out SO that cell counts could bedone on duplicate wells. Cells were removed from the substratum using trypsin digestion and counted on a hemocytometer. Determination of Protein Phosphatase Activities-Phosphorylase phosphatase was measured asthe inactivation of rabbit skeletal muscle phosphorylase a (18). Protein serinelthreonine phosphatase activity was also measured as therelease of 32Pfrom histone type IIIS (Sigma), which was phosphorylated by partially purified rat liver protein kinase C. This represents a modification of our previously described method which used casein as substrate (20). Protein tyrosine phosphatase activity was measured as therelease of 32P from Tyr(P)-reduced carboxyamidomethylated lysozyme (Tyr(P)-RCML) orTyr(P)-myelin basic protein (Tyr(P)-MBP). These substrates were phosphorylated on tyrosine using partially purified human placental insulin receptors (19).Protein tyrosine phosphatase activity was measured as 32Preleased from the Tyr(P) substrates as described previously (19). Assignment of the activities measured with Tyr(P)-RCML lysozyme and Tyr(P)-MBP to protein tyrosine phosphatases, as opposed to acid or alkaline phosphatases, was supported by sensitivity to vanadate, zinc, and poly(g1utamic acid. tyrosine) (21). In addition, the activity of protein tyrosine phosphatases is insensitive to EDTA, an inhibitor of acid and alkaline phosphatases (21) and a constituent of our protein tyrosine phosphatase assay buffer. For all protein phosphatase activity determinations, samples were diluted so that substrate consumption was less than 50%. In this range, the relationship between substrate consumed and time is linear. The inhibition of phosphorylase phosphatase by heat-stable proteins, a measure of inhibitor-2 (12) activity, was determined using rabbit skeletal muscle protein phosphatase 1 catalytic fragment as described previously (18).For these analyses, undiluted homogenates were placed in a boiling water bath for 5 min and then centrifuged for 5 min at 12,000 X g to remove denatured proteins. One unit of I2 activity is defined as that amount which produces 50% inhibition of the standard amount of phosphorylase phosphatase activity (15 milliunits) included in the assay. Purified I2 used to characterize protein phosphatase activity was prepared from rabbit skeletal muscle (18). Antibodies against rabbit skeletal muscle protein phosphatase 1 (22, 23), also used to characterize lysate phosphorylase phosphatase activity, were prepared as described previously by ammonium sulfate precipitation followed by dialysis with 20 mM MOPS, 150 mM NaCl, pH 7.4 (22). Statistical Analyses-The significance of differences from control data was determined using the Rank Sum Test with the Bonferroni correction for multiple comparisons (24). Significance was defined as p < 0.05.

3445

0.6 r

*

I

Time (hour61

FIG. 1. Effectof TGF-Bl on phosphorylase phosphatase. Phosphorylase phosphatase activity was measured in lysates of keratinocytes harvested at various times after the addition of 400 pM TGF-P1. Data for control and 48-h time points were compiled from six separate experiments utilizing three primary cultures. Data for time points up to 24 h represent measurements made during the course of three separate experiments using three primary cultures. Error burs represent the standard errorof the mean. Asterisks denote a significant difference from the control ( t = 0).

less. These findings are consistentwith those of Wikner etal. (25), who observed a %fold increase in protein synthesis in human keratinocytes exposed to TGF-01 for 22 h. However, these authorsdescribed different ratesof increased production forvarious proteins. Since increased proteinsynthesisin response to TGF-01 involves specific proteins and protein content did not vary during early time points in ourexperiments, we have expressed phosphatase activities per unit of cell number. DuetoTGF-01-induced growth arrest, cell densitiesin treated and untreated cultures differed significantly by the end of each experiment. To determineif the TGF-01-associated increase in phosphorylase phosphatase activity was independent of cell density, we examined phosphatase activity in keratinocytes plated atvarying densities (0.5, 1.0, and 3.0 X IO5 cells/35-mm well) and exposed to TGF-/31 for 48 h. Comparison of activities in wells with comparable cell numbers showed TGF-01-associated activation of phosphorylase phosphatase regardless of cell density (data not shown). Characterization of Protein SerinelThreonine Phosphatase Activity-In cell extracts, phosphorylase phosphatase activity is predominantly, if not exclusively, due to protein phosphaRESULTS tases 1 and 2A (26). We studied the sensitivity of phosphoEffect of TGF-01-induced Growth Arrest on Phosphorylase rylase phosphatase activity to the protein phosphatase reguPhosphatase Actiuity-We found that TGF-01 induced a re- latory protein, 12, which selectively blocks protein phosphaversible state of growth arrest in rapidly proliferating, early tase 1 (26). Using keratinocyte lysates obtained before and passage human keratinocytes. This effect was maximal within after exposure to TGF-01, phosphorylase phosphatase activity 48 h after the additionof 400 PM TGF-01, as indicatedby an was inhibited by approximately 50% in the presence of 2 ng 80% decrease in labeled nuclei detected by autoradiography of I2 and was maximally inhibited at concentrations of 10(data notshown). To determine if growth inhibition of human 100 ng/assay tube (approximately 17-170 nM I2 (21)). The keratinocytes was accompanied by changes in the activityof 12-resistant activity (measured in the presence of 100 ng of protein serinelthreonine phosphatases,we measured activity 12) did not change with2- or 48-h exposure to 400 PM TGFusing phosphorylase a as a substrate in keratinocyte lysates 01 (Fig. 2). Therefore, the increase in total phosphorylase at various time points after the addition of 400 PM TGF-01 phosphatase activity in response to TGF-01 represented an inserum-free complete KGM.Phosphorylasephosphatase increase in protein phosphatase 1 activity that was sensitive activity increased significantly within 30 min after treatment t o 12. Phosphorylasephosphataseactivity measured after of subconfluent keratinocyte monolayers with TGF-01 (Fig. trypsin activation (Fig. 2)also increased after exposure to 1). This activation increasedprogressively for up to 48 h, by TGF-81. Trypsin treatmentmay expose protein phosphatase which time, a 3-4-fold increase in phosphorylase phosphatase 1 catalytic activity by removing associated inhibitors or regactivity was observed. ulatory subunits, including I2 (see "Discussion"). Expression of phosphorylase phosphatase activity per unit Phosphorylase phosphatase activity that was insensitive to of protein yielded similar results for incubation periods up to I2may representproteinphosphatase 2A. However, itis 6 h. After24 and 48 h of exposure to TGF-01, protein contentpossible that this 12-resistant activity was due to a form of per unit of cell number increased 1.5-2.5-fold, such that the protein phosphatase 1 catalytic subunit that was associated apparent increase in phosphatase activity was proportionately with regulatory subunits and therefore resistant to inhibition

3446

Protein PhosphataseActivation and TGF-p phohistone (data not shown). Activities ranged from 1.5-2.1 pmol/min/105 cells in samples obtained before and after 30min to 48-h exposure t o TGF-Dl. I2 activity, measured in heat-treated lysates, was unchanged after either acute (30 min to 6 h) or 48-h exposure to TGF-Dl.Activities in all samples ranged from 15 to36 units/105 cells. Effect of TGB-Dl-induced Growth Arrest on Protein Tyrosine Phosphatase Actiuity-Growth arrest following 48 h of exposure t o TGF-Dl was associated with a 3-4-fold increase inproteintyrosinephosphataseactivity measuredwith Tyr(P)-RCML (Fig. 4A). However, the kinetics of protein

0.51

0

2

48

hour8

FIG. 2. Phosphorylase phosphatase activity measured in the presence of I2 and following activation by trypsin. Keratinocyte lysates obtained after 0-, 2-, and 48-h exposure to 400 pM TGF01 were used to measure phosphorylase phosphatase activity in the presence of 100ng of I2 (open bars), without pretreatment or additions (hatched bars), and following digestion by trypsin (20 pg/ml for 10 min at 30 "C, solid bars). Measurements were made in triplicateusing lysates from separate wells. Error bars represent 1standard deviation.

t

'.4A I

I

Bo

Cells per 35 mm well (x 10-5)

0

1:8

1: 14 : 12 :l

PP-1 Antibody Diiutlon

FIG. 3. Neutralization of phosphorylase activity by protein phosphatase (PP-1)antibodies. Keratinocyte lysates (20 pl) from cells exposed to 400 PM TGF-@1for 0 h (circles), 2 h (squares), or 48 h (triangles)were incubated with protein phosphatase 1antibody (10 pl) at varying dilutions for 30 min a t 30 "C. The mixture was then assayed for phosphorylase phosphatase activity.

FIG. 4. Effect of TGF-Dl and cell density on protein tyrosine phosphatase (PTPase).Panel A shows keratinocyte lysate protein tyrosine phosphatase activity measured with Tyr(P)-RCML before (hatched bar, n = 10) and after (solid bar, n = 9) exposure for 48 h to TGF-@l. Error bars represent 1 standard deviation. The difference between the means was significant ( p < 0.001). Panel B shows protein tyrosine phosphatase activities in lysates of cells harvested before (solid circles) and 48 h after (open circles) exposure to 400 p~ TGF-01. Results are expressed as a function of cell density determined at theend of the experiment.

=

'SO0 O0L

T

PP-1

by 12. To test this possibility,phosphorylase phosphatase activity was measured without prior activation and in the presence of antibodies against rabbit skeletalmuscle protein phosphatase 1. We have demonstrated previously that these antibodies neutralize the activityof purified protein phospha'$ 6 0 0 t T of purified protein tase 1while having no effect on the activity phosphatase 2A (23). Lysates of cells harvested a t 0, 2, and 48 h after the addition of TGF-Dl t o culture medium were assayed in the presence of varying concentrations of antibody (Fig. 3). Neutralization was nearly complete in the 0- and 2h lysates. The maximum concentration of antibody was insufficient to inhibit phosphatase activity completely in the TGF-P - + + - 48-h lysate. However, activities in all three lysateswere neuC a* + + + tralized in parallel.These results are most consistent with the Serum - - - - + conclusion that the keratinocyte lysate phosphorylase phosFIG. 5. Effect ofkeratinocyte differentiation on protein phatase activity is nearly completely accounted for by protein phosphatase 1 ( P P - I ) and protein tyrosine phosphatase phosphatase 1. A portion of this activity is sensitive to 12. (PTPase).Protein phosphatase 1 and protein tyrosine phosphatase After exposure to TGF-Dl, 12-resistant protein phosphatase activities were measured in cells treated with various combinations of 400 PM TGF-@l,1.8 mM calcium, and 10%fetal bovine serum (see 1 activity remains constant, while 12-sensitive activity in"Experimental Procedures"). Results are expressed as a percentage creases. of control activity. Error bars represent 1 standard deviation. AsterWe employed a second substrate, histone phosphorylated isks denote a significant difference from the corresponding control probe forprotein (no additions) activity. Control and TGF-Pl results are compiled by rat liver protein kinaseC, as an additional serine/threonine (type 2A) phosphatase activity. TGF-Dl had fromsix experiments, while other data represent three combined no effect on protein phosphatase activity measured with phos- experiments. "

Protein Phosphatase Activation and TGF-P

3447

observed following the addition of TGF-P1 to human keratinocyte cultures. All the activity that increased following exposure to TGF-/31 was sensitive to inhibition by 12. This has been usedas a primary criterion for classification of phosphorylase phosphatase activity as type 1 (28). To further confirm that increases in phosphorylase phosphatase activity following TGF-Pl were due to protein phosphatase 1, the activity was neutralized with protein phosphatase 1-specificantibodies. The conclusion that TGF-(31 led to rapid activation of protein phosphatase 1 and not protein phosphatase 2A was also supported by the finding that protein phosphatase activity measured with histone phosphorylated by protein kinase C did not change with exposure to TGF-01. The latter substrate has been used for the selective measurement of protein phosphatase 2A activity (29). A direct relationship between TGF-Pl and regulation of protein tyrosine phosphatase is less clear. Although exposure to TGF-Dl did lead to increased protein tyrosine phosphatase activity, this effect was only observed at 24 and 48 h. TGF61 arrests keratinocyte growth during progression from GI to S phase (11).It is possible that theincreased protein tyrosine phosphatase activity associated with TGF-Dl-induced growth arrest is a function of the point in the cell cycle at which the PM) . cells are arrested rather than direct involvement in TGF-P1 Effects of Keratinocyte Differentiation on Protein Phospha- signal transmission. We originally considered a role for protase Actiuities-Growth arrest in human keratinocytes by tein tyrosine phosphatase activation during keratinocyte difTGF-Dl is not accompanied by terminal differentiation, even ferentiation. Such an effect has been seen with monocytic after prolonged (5-12 days) exposure (27).2 Differentiation, differentiation of HL-60 leukemia cells (30). However, calhowever, can be induced in cells exposed to TGF-Dl within 2 cium-induced keratinocyte differentiation was not accomdays by the addition of 1.8 mMCa". In the present studies, panied by protein tyrosine phosphatase activation. Indeed, we confirmed that growth arrest of keratinocytes induced by treatment with 1.8 mM calcium following growth arrest with TGF-Dl was not accompanied by differentiation, as indicated TGF-01 was associated with areturn of protein tyrosine by morphologic evidence of stratification aswell as histochem- phosphatase activity to levels seen in proliferating, undifferical staining for keratin. The addition of 1.8 mM calcium to entiated cells. This result is consistent with the findings of nonproliferating cells cultured in the presence of 400 pM TGF- Filvaroff et al. (31), who observed that differentiation of p l for 48 h produced differentiation without DNA synthesis mouse keratinocytes in response to calcium wasaccompanied (as determined by autoradiography). In contrast, the addition by protein tyrosine phosphorylation and that tyrosine kinase of 1.8 mM calcium to proliferating cultures produced minimal inhibitors blocked differentiation. signs of differentiation and no growth arrest. Treatment of The above studies on protein tyrosine phosphatase regulacells with 10% fetal bovine serum led to full differentiation tion utilized two substrates, Tyr(P)-RCMLandTyr(P)-MBP. while proliferation continued. Multiple classes of protein tyrosine phosphatase have now The greatest increase above control levels in protein phos- been described based on molecular cloning. We utilized the phatase 1 activity (Fig. 5, top panel) was seen in growth- protein tyrosine phosphatase substrate, Tyr(P)-MBP, as a arrested, undifferentiated cells (400 PM TGF-Pl for 48 h). probe for the transmembrane class of protein tyrosine phosInduction of differentiation in growth-arrested cells (i.e.TGF- phatases ( M , -200,000), of which the leukocyte common D l plus 1.8 mM calcium for 48 h after exposure to TGF-Dl antigen, CD45 (32), and LAR (leukocyte antigen-related (33)) alone for 48 h) did not lead to a further increase in protein are members. Our own findings using rat liver membranes phosphatase 1activity. Neither the addition of 1.8 mM calcium (34) indicate that Tyr(P)-MBP is a better substrate than to proliferating cells nor the differentiation of proliferating Tyr(P)-RCML for a proteintyrosine phosphatase resembling cells with fetal bovine serum plus 1.8 mM calcium resulted in LAR, while it is a poor substrate relative to Tyr(P)-RCML changes in protein phosphatase 1 activity from control levels. for a rat liver homologue of the M , -50,000 enzyme which is Protein tyrosine phosphatase activity (Fig. 5, bottom panel) termed protein tyrosine phosphatase 1B (35). Our protein increased only in growth-arrested, undifferentiated keratinotyrosine phosphatase measurements would include the activicytes (exposure to TGF-Pl alone). Neither exposure to 1.8 mM calcium nor differentiation in proliferating or nonproli- ties of either LAR-related or protein tyrosine phosphatase ferating cells led to a change in protein tyrosine phosphatase 1B-related enzymes, or both. It is of interest that thestructure activity from control levels. Similar results were obtained of the extracellular domain of LAR has characteristics of cell adhesion molecules (33). The effect of cell density on protein using Tyr(P)-RCML and Tyr(P)-MBP as substrate. tyrosine phosphatase activity may reflect changes mediated via the cell-surface domain of a transmembrane protein tyDISCUSSION rosine phosphatase. Our datasupport the hypothesis that the induction of In summary, our results show the rapid activation of protein growth arrest in keratinocytes by TGF-Pl involves activation phosphatase type1in human keratinocytes treated with TGFof protein phosphatases. More specifically, a rapid increase (31.This finding is consistent with a role for protein phosphain type 1 protein serine/threonine phosphatase activity was tases inthe mechanism of action of TGF-Dl. It seems reasonable to speculate that protein phosphatase activation might * L. Braun and R. Mikumo, unpublished observations. lead to growth arrest via dephosphorylation of multiple sub-

tyrosine phosphatase activation differed markedly from activation of protein phosphatase 1, with no change in protein tyrosine phosphatase activity occurring until 24 h of exposure (data not shown). Unlike protein phosphatase 1, the increase in protein tyrosine phosphatase activity seen at 48 h was affected by cell density (Fig. 4B). The greatest increase induced byTGF-/31was seen at the lowest cell densities. In contrast, cells cultured without TGF-81 appeared to have increased activity at higher densities. Longer culture times in the absence of TGF-P1 ( 9 7 2 h), which resulted in the highest cell densities we studied (7-14 X lo5 cells/35-mm well), were associated with no further increase in protein tyrosine phosphatase activity. Protein tyrosine phosphatase activities were also measured with a second substrate, Tyr(P)-MBP. Proteintyrosine phosphatase activities measured with 1 PM Tyr(P)-RCML and Tyr(P)-MBP were approximately equal in multiple keratinocyte lysates prepared from cells in various stages of growth arrest anddifferentiation (see below). Activity measured with both substrates was inhibited 50% by 105 PM Zn2+and was fully inhibited at 250 FM. Protein tyrosine phosphatase activity was also sensitive to vanadate (50% inhibition at 250 p M ) and poly(g1utamic acid.tyrosine (4:l)) (50% inhibition at 1.3

3448

Phosphatase and Activation Protein

strates, possibly including theanti-oncogeneproduct RB, which is dephosphorylated in response to TGF-(31 (14). Acknowledgments-We thank Joan Boylan for protein tyrosine phosphatase substrate preparation and Beth Smiley for assistance in these studies. REFERENCES 1. Hunter, T., and Cooper, J. A. (1985) Annu. Reu. Biochem. 5 4 , 897-930 2. Nishizuka, Y. (1989) Cancer 6 3 , 1892-1903 3. Susa, M., Olivier, A. R., Fabbro, D., and Thomas, G. (1989) Cell 57,817-824 4. Rap, U. R., Heidecker, G., Huleihel, M., Cleveland, J. L., Choi, W. C., Pawson, T., Ihle, J. N., and Anderson, W. B. (1988) Cold Spring Harbor Symp. Quunt.Biol. 53, 173-184 5. Krebs, E. G., Eisenman, R. N., Kuenzel, E. A., Litchfield, D. W., Lozeman, F. J., Luscher, B., and Sommercorn, J. (1988) Cold Spring Harbor Symp. Quant.Biol. 5 3 , 77-84 6. Ray, L. B., and Sturgill, T. W. (1988) J. Bwl. Chem. 263,1272112727 7. Draetta, G., Brizuela, L., Moran, B., and Beach, D. (1988) Cold Spring Harbor Symp. Quunt. Biol. 53,195-201 8. Cyert, M. S., and Thorner, J. (1989) Cell 57,891-893 9. Roberts, A. B., and Sporn, M. B. (1988) Biofactors 1,89-93 10. Rizzino, A. (1988) Deu. Biol. 130,411-422 11. Shipley, G.D., Pittelkow, M. R., Wille, J. J., Jr., Scott, R. E., and Moses, H. L. (1986) Cancer Res. 4 6 , 2068-2071 12. Pietenpol, J. A., Holt, J. T., Stein, R. W., and Moses, H. L. (1990) Proc. Natl. Acud. Sci. U. S. A. 8 7 , 3758-3762 13. Braun, L., Durst, M., Mikumo, R., and Gruppuso, P. A. (1990) Cancer Res. 50, 7324-7332 14. Laiho, M., DeCaprio, J. A., Ludlow, J. W., Livingston, D. M., and Massagu6, J. (1990) Cell 6 2 , 175-185 15. Hawley-Nelson, P., Sullivan, J. E., Kuny, M., Hennings, H., and Yuspa, S. H. (1980) J. Inuest. Dermatol. 75, 176-182 16. Freshney, R. I. (1987) Culture of Animal Cells: A Manual of Basic

TGF-p Technique, Alan R. Liss, Inc., New York 17. Ayoub, P., and Shklar, G. (1963) Oral Surg. 16, 580-581 18. Gruppuso, P. A., Johnson, G. L., Constantinides, M., and Brautigan, D. L. (1984) J. Biol. Chem. 260, 4288-4294 19. Gruppuso, P. A., Boylan, J. M., Posner, B. I., Faure, R., and Brautigan, D. L. (1990) J . Clin. Inuest. 8 5 , 1754-1760 20. Gruppuso, P. A. (1990) Pediatr. Res. 27,599-603 21. Brautigan, D. L., and Shriner, C.L. (1988) Methods Enzymol. 159,339-346 22. Brautigan, D. L., Shriner, C. L., and Gruppuso, P. A. (1985) J. Biol. Chem. 260,4295-4302 23. Brautigan, D. L., Gruppuso, P. A., and Mumby, M. (1986) J. Bwl. Chem. 261,14924-14928 24. Wallenstein, S., Zucker, C. L., and Fleiss, J. L. (1980) Circ. Res. 4 7 , 1-9 25. Wikner, N. E., Persichitte, K. A., Baskin, J. B., Nielsen, L. D., and Clark, R. A. F. (1988) J. Inuest. Dermatol. 9 1 , 207-212 26. Ingebritsen, T. S., and Cohen, P. (1983) Science 2 2 1 , 331-338 27. Moses, H. L., Tucker, R. F., Leof, E. B., Coffey, R. J., Halper, J., and Shipley, G.D. (1985) in CancerCelk 3 (Feramisco, J., Ozanne, B., and Stiles, C., eds) pp. 65-71, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY 28. Ingebritsen, T. S., Stewart, A. A,, and Cohen, P. (1983) Eur. J . Biochem. 132,297-307 29. Jakes, S., and Schlender, K. K. (1988) Biochim. Biophys. Acta 967,ll-16 30. Frank, D. A., and Sartorelli, A. C. (1986) Biochem. Biophys. Res. Commun. 140,440-447 31. Filvaroff, E., Stern, D. F., and Dotto, G. P. (1990) Mol. Cell. Biol. 10, 1164-1173 32. Tonks, N. K., Diltz, C. D., and Fischer, E. H. (1990) J. B i d . Chem. 265,10674-10680 33. Streuli, M., Kreuger, N. X., Tsai, A. Y. M., and Saito, H. (1989) Proc. Natl. Acud. Sci. U. S. A. 8 6 , 8698-8702 34. Gruppuso, P. A., Boylan, J. M., Smiley, B. L., Fallon, R. J., and Brautigan, D. L. (1991) Biochem. J., in press 35. Cool, D. E., Tonks, N. K., Charbonneau, H., Walsh, K. A., Fischer, E. H., and Krebs, E. G. (1989) Proc. Natl. Acad. Sci. U. S. A. 86,5257-5261