Control of membrane permeability by external and internal ATP in

Proc. Natl. Acad. Sci. USA Vol. 77, No. 4, pp. 2103-2107, April 1980 Cell Biology Control of membrane permeability by external and internal ATP in 3...
Author: Magnus Greene
2 downloads 0 Views 1MB Size
Proc. Natl. Acad. Sci. USA Vol. 77, No. 4, pp. 2103-2107, April 1980

Cell Biology

Control of membrane permeability by external and internal ATP in 3T6 cells grown in serum-free medium (insulin/epidermal growth factor/Rb+ efflux/membrane phosphorylation/nucleotide efflux)

PHILLIP DICKER, LEON A. HEPPEL*, AND ENRIQUE ROZENGURTt Imperial Cancer Research Fund, Lincoln's Inn Fields, London WC2A 3PX, United Kingdom

Contributed by Leon A. Heppel, January 18,1980

ABSTLACT Cultures of 3T6 cells were plated in serum-free medium and grown in the presence of insulin (1 gg/ml) and epidermal growth factor (0.5 ng/ml). External ATP (250 ,M) applied to such cultures caused a rapid efflux of acid-soluble pools labeled with [4Hjuridine, 2-deoxy(3Hjglucose, or 8"Rb+ and allowed the entry of p-nitrophenylphosphate. This increase in passive membrane permeability depended on ATP concentration, pH, and time of ATP contact with the cells, and it was not produced by GTP, UTP, or Pi. In the presence of compounds that decrease intracellular ATP, low concentrations of external ATP (40 jM) caused a massive synergistic stimulation of efflux. The efflux of acid-soluble pools was stopped (sealing) by bringing the cultures of 3T6 cells to neutral pH in the presence of Ca2+ and Mg2+. Exposure of 3T6 cells grown in serum-free medium to [y-32PJATP under the conditions of permeabilization led to the selective labeling of a membrane protein with a molecular weight of 44,000 as revealed by NaDodSO4 polyacrylamide gel electrophoresis and autoradiography. The results show that the control of membrane permeability by ATP is completely independent of serum-derived proteins. Furthermore, the protein band (Mr, 44 X 103) that shows selective labeling by [mPJATP during permeabilization is not an adsorbed serum component.

Addition of ATP to cultures of transformed mouse cells causes a striking increase in passive membrane permeability. Brief exposure of monolayers of 3T6, SV40-3T3, or Py-3T3 cells to ATP causes a massive efflux of nucleotide pools labeled with [3H]uridine or [3H]adenosine, of sugar phosphates, and of ions (1, 2). Treatment with ATP also allows the entry of normally impermeable molecules such as p-nitrophenyl phosphate, glucose 6-phosphate, or NAD (3-5). These effects are readily reversible and the treated cells grow normally afterward (1). The effect of exogenous ATP depends on the concentration of intracellular ATP; a decrease in intracellular ATP leads to a marked increase in the sensitivity of the cells to exogenous ATP (2). These findings have a number of implications. They provide a useful method by which to study internal metabolism in permeabilized cells (5, 6). Furthermore, these observations may have considerable relevance to the regulation of passive permeability, to the role of phosphorylation reactions in this control, and to the surface changes taking place after malignant

transformation. The complex nature of serum and its multiplicity of actions have frequently rendered difficult the interpretation of experiments with cultured animal cells grown in medium supplemented with serum. One aspect of this problem is that proteins foreign to the cell and derived from serum present in the nutrient medium become strongly associated with the cell surface (7, 8). Such adsorbed proteins may behave as membrane antigens in cell-mediated cytotoxicity (9), can be labeled by chemical probes (10) or enzymes (11, 12), and can endow the The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U. S. C. §1734 solely to indicate this fact.

2103

outer cell surface with enzymatic activity. For this reason, it is particularly important to define whether the effects of extracellular agents, which act on the outer cell surface, are mediated by molecules that are derived from serum or are synthesized by the cell. This question has an obvious connection

with the permeabilizing effect of external ATP. We approached this problem by studying the effect of exogenous ATP on the passive permeability of 3T6 cells plated and grown in serum-free synthetic medium to which the polypeptides insulin and epidermal growth factor (EGF) were added. We found that addition of ATP to such cultures induced a dramatic increase in membrane permeability as revealed by exit of nucleotide pools and by entry of p-nitrophenyl phosphate and have confirmed all of the basic features of the phenomenon of ATP permeability. Furthermore, in transformed mouse cells grown in serum-free medium, the effect of ATP was elicited at low concentrations, could be clearly seen at physiological pH, and was associated with the labeling of a 44,000-dalton protein upon exposure of these cultures to [,y-32P]ATP. The results provide unambiguous evidence that the effect of external ATP on the cell surface is completely independent of serum-derived proteins. MATERIALS AND METHODS Cell Culture. Stock cultures of Swiss mouse 3T6 cells (13) were propagated in Dulbecco's modified Eagle's (DME) medium containing 10% fetal calf serum as described (1, 2). Plating and Growth of 3T6 Cells in Serum-Free Medium. 3T6 cells were removed from stock dishes by incubation with 0.25% solution of bovine pancreas crystalline trypsin (Sigma) in phosphate-buffered saline. As soon as the cells detached from the dish, a 10-fold excess of soybean trypsin inhibitor (Sigma) was added and the cells were centrifuged 1000 X g for 10 min. The cell pellet was resuspended at 105 cells per ml in a 1:1 mixture of DME medium and Waymouth's medium supplemented with 1.6 ,M ferrous sulfate, insulin at 1 Mg/mi, and EGF at 0.5 ng/ml. This suspension was added to Nunc dishes (2 ml for 30-mm dishes, 10 ml for 90-mm dishes) and cultures were used for experiments after a confluent layer had formed (7-12 days). Such cultures contained an average of 2 X 105 cells and 30,Mg of protein per cm2. Measurement of Efflux of Acid-Soluble Pools. The acidsoluble pool was labeled with [3H]uridine (1 ,M; 0.5 ,uCi/ml; 1 Ci = 3.7 X 1010 becquerels) or 86Rb+ (1 MCi/ml) in DME medium/Waymouth's medium, 1:1 (vol/vol), containing 1 ,g of insulin per ml. For 2-deoxy[3H]glucose labeling of the Abbreviations: EGF, epidermal growth factor; DME medium, Dulbecco's modified Eagle's medium; CCCP, carbonyl cyanide-m-chlorophenylhydrazone. * Present address: Section of Biochemistry and Molecular and Cell Biology, Cornell University, Ithaca, NY 14853. t To whom reprint requests should be addressed.

2104

Cell Biology: Dicker et al.

acid-soluble pool, cells were incubated with 2-deoxy[3H]glucose (0.1 nM, 2 piCi/ml) in DME medium lacking glucose but containing 1 jug of insulin per ml. In 8 hr, about 20% of the added radioactivity was in the acid-soluble pool, and in 24 hr, all of it was intracellular and nearly all was acid-soluble. After labeling, the cells were washed four times with 0.15 M NaCl/0.05 mM CaC12 and incubated with 1.5 ml of medium A (0.1 M Tris-HCl/0.05 M NaCl/50 ,gM CaCl2/dextran 500 at 5 mg/ml, pH 8.2 at 230C). Incubation was at 37.50C for times specified in each experiment. At the end of incubation, 1-ml portions of the medium were removed for measurement of the radioactivity as described (1). When ethanolic solutions of uncouplers [carbonyl cyanide-m-chlorophenylhydrazone (CCCP)], inhibitors of electron transport (rotenone, antimycin A), and inhibitors of energy transfer (rutamycin) were used, the control culture received an equivalent amount of ethanol. Usually, the final concentration of ethanol was 0.2%, which does not affect the efflux of acid-soluble pools under the conditions used in our experiments (2). Measurements of Entry of p-Nitrophenyl Phosphate. Entry of p-nitrophenyl phosphate into permeabilized cells was measured by following hydrolysis by internal alkaline phosphatase (3). [y-32PJATP Labeling of Membrane Proteins. Cultures of 3T6 cells were produced in 90-mm dishes in the absence of serum as described above. Cultures were washed four times in 0.15 M NaCl/50 ,gM CaC12 and then incubated at 37.50C as described in figure legends. At the end of the incubation period, cultures were washed twice in ice-cold 0.15 M NaCl/50,uM CaC12 and left for 15 min in 1 ml of ice-cold homogenization buffer [10 mM sodium phosphate, pH 7.4/2 mM EDTA/10 mM sodium fluoride/0.01 mM ammonium molybdate/2 mM phenylmethylsulfonyl fluoride (a protease inhibitor)]. Cell membranes were then isolated by a method similar to that of Brunette and Till (14) as modified by Stone et al. (15). Cells were scraped off the dish with a rubber policeman and homogenized by 150 complete cycles in a Dounce homogenizer. The homogenate was poured into 15-ml Corex tubes and centrifuged at 8000 rpm for 15 min at 4'C in an HB-4 swinging bucket rotor with a Sorvall RC2-B centrifuge (this rotor and centrifuge were used in all centrifugations described in this procedure). The pellet was resuspended in 10 ml of ice-cold two-phase polymer solution [2.3 ml of 20% (wt/wt) dextran T 500, 1.3 ml of 30% (wt/wt) polyethylene glycol 6000, 4.2 ml of 0.22 M sodium phosphate at pH 6.6, 0.01 ml of 1 M NaCI, and 2.0 ml of H20] and centrifuged at 10,000 rpm for 10 min. The entire supernatant was then decanted into another 15-ml Corex tube and recentrifuged at 10,000 rpm for 10 min. The white interface band between the two phases (cell membranes) was removed by using a pasteur pipette, diluted 1:10 with ice-cold water, and centrifuged at 8000 rpm for 15 min. The supernatant was discarded and the pellet was air dried. NaDodSO4/Polyacrylamide Gel Electrophoresis. The cell membranes were dissolved in 50 Al of electrophoresis sample buffer [2% (wt/vol) NaDodSO4/10% (vol/vol) glycerol/0.08 M Tris-HCl, pH 6.8/0.1 M bromophenol blue/0.1 M dithiothreitol). An aliquot was taken to determine its total radioactivity and protein content. The rest of the solution was analyzed by NaDodSO4/polyacrylamide gel electrophoresis on a 5-15% polyacrylamide gradient gel, with a 3% polyacrylamide stacking gel. Electrophoresis was carried out at 50 V until the sample had reached the running gel and then at 100 V. The gel was stained with Coomassie blue in 45% methanol/9% acetic acid, destained in 45% methanol/9% acetic acid, dried, and autoradiographed on Kodak X-Omat H film.

Proc. Natl. Acad. Sci. USA 77 (1980)

Materials. All radioisotopes were obtained from Amersham. ATP was obtained-from Boehringer unless indicated as from Sigma or P-L Biochemicals. CCCP and rotenone were obtained from Sigma. Rutamycin was kindly provided by R. J. Holsley (Eli Lilly).

RESULTS Growth of 3T6 Cells in the Absence of Serum. 3T6 cells were transferred from stock dishes and plated into experimental dishes with crystalline trypsin which, after cell detachment, was neutralized with excess soybean trypsin inhibitor. No serum was used. In the presence of EGF and insulin, the number of cells increased about 20-fold in 6 days. The kinetics of proliferation are shown in Fig. 1. In the exponential phase, the cell number doubled every 24 hr. The saturation density achieved was 2 X 105 cells per cm2. These cultures, plated and grown in the complete absence of serum, were then used for the experiments described below. ATP Increases Permeability in 3T6 Cells Grown Under Chemically Defined Conditions. When cultures of 3T6 cells whose acid-soluble pools were labeled with [3H]uridine, deoxy[3H]glucose, or 86Rb+ were exposed to 0.25 mM ATP, there was a sharp increase in the rate of exit of the radioactively labeled pools (Fig. 2). Efflux of 86Rb+ was much more rapid than efflux of nucleotides or of 2-deoxyglucose 6-phosphate. Entry of p-nitrophenyl phosphate was also greatly stimulated by external ATP, as had previously been observed for serum-fed cells (3). Because the cells were grown in serum-free medium, the effect of ATP was not due to enzymes or other proteins contributed by serum nor could it be due to specific stimulation by serum components. Influence of ATP Concentration, pH, and Time of Contact in 3T6 Cells. A detailed kinetic analysis of the effect of ATP was carried out in 3T6 cells grown under chemically defined conditions. Decreased ATP concentration in the medium led to two clear kinetic modifications: (i) the lag period before permeabilization was longer, and (ii) the linear rate of efflux after the lag period was decreased (Fig. 3A). A detectable effect could be seen with as little as 50 ,uM ATP, and this threshold was reduced in experiments involving synergistic effects (see below). The effect of pH reported here is somewhat different from what we have reported previously (1). We still observed an alkaline pH optimum, but we found that the reaction is as rapid with Tris-buffered medium at pH 7.8 (measured at 230C) as with medium at pH 8.2 (Fig. 3B). At 37.5°C, the "pH 7.8" I_

100 100

°xf 10

E 1E 10 _

v

3

6

9

12

Time, days FIG. 1. Growth of 3T6 cells in serum-free medium. 3T6 cells were plated in the absence of serum at 104 cells per CM2 in serum-free medium supplemented with EGF and insulin. Cell number was determinebyusin a outrw Couner o theolwn dlay (daoy 0) and on subsequent days as indicated.

Proc. Natl. Acad. Sci. USA 77 (1980)

Cell Biology: Dicker et al.

2105

10_

!00_ c C

E E E

I-

0 0 x

r-

*5TP X

00

wJ

0

0

D

0 0 a:

Time, min

FIG. 2. Effect of ATP on the efflux of acid-soluble pools labeled with [3H]uridine (A), 2-deoxy[3Hlglucose (B), or 86Rb (C) or on the rate of p-nitrophenyl phosphate entry and hydrolysis (D). Cultures of 3T6 cells grown in serum-free medium containing 0.5 ng of EGF and 1 Mg of insulin per ml, were incubated in media containing the isotopes for 6-12 hr. After four washes with 0.15 M NaCI/50 AM CaCl2 and one wash with medium A (pH 8.2), the dishes received 1.5 ml of medium A (pH 8.2) and were incubated at 37.50C. At intervals as indicated, a control dish was removed and the supernatant medium analyzed for radioactivity (0). The experimental dishes received 250 MM ATP at the time indicated by arrows and incubation was continued. At various time intervals, a dish was removed from the bath and again the supernatant was assayed for radioactivity (-). Other dishes were washed as described and immediately extracted with 5% trichoroacetic acid to determine the size of the acid-soluble pool at the time of the experiment with the following results expressed as cpm per dish: [3H]uridine-labeled pool, 200,000; 2-deoxy[3H]glucoselabeled pool, 70,000; 86Rb pool, 7000. For (D), cells were incubated with 3.6 mM p-nitrophenyl phosphate in the presence (-) or absence (0) of 250 MM ATP. p-Nitrophenol production was measured at various time points.

medium actually is at pH 7.55. The effect of ATP was rapid but not instantaneous (Fig. 3C); a contact period of several minutes was necessary to produce a maximal rate of efflux. Effect of ATP on Cells Grown in the Absence of Serum is Specific. The permeabilization by ATP was specific; GTP, UTP, and inorganic pyrophosphate were ineffective (data not shown). In earlier work with serum-fed cells, many compounds were tested. A positive effect was observed only with ATP (1-5) and with y-thio-ATP (4). A considerable literature has accumulated on the effect of vanadate on a number of enzyme systems (16, 17), and vanadate has been reported to be a powerful inhibitor present in ATP isolated from muscle (18, 19). Accordingly, we tested ATP supplied by Sigma and P-L Biochemicals as well as the Boehringer product; they were equivalent in their action on transformed mouse cells. Furthermore, sodium orthovanadate in the concentration range 0.2 to 10 MM caused no stimulation of efflux (results not shown). Reciprocal Effects of External and Internal ATP on Rate of Efflux of Labeled Pools in 3T6 Cells Grown in the Absence of Serum. Previously, we reported (2) on the reciprocal effects of external and internal ATP on rate of efflux of labeled pools. When the concentration of intracellular ATP was decreased by treatment of 3T6 cells with uncouplers, respiratory inhibitors, or energy-transfer inhibitors, the cells became much more sensitive to external ATP. We have confirmed and extended these observations by using cells plated and grown in serum-free

10 20 30 Time, min

01 7.0

4 0-z 8.2 0 7.6 pH

8 Time, min

16

FIG. 3. (A) Effect of time of contact with ATP on stimulation of efflux of the [3H]uridine pool from 3T6 cells. Cultures were incubated for 12 hr with [3H]uridine, washed, and then incubated for various times in medium A with 250 MiM ATP (-), 125MM ATP (A), 62 uM ATP (v), 31 MM ATP (o), or no ATP (-). At intervals, the supernatant was removed and assayed for radioactivity. (B) Effect of pH on efflux of [3Hjuridine from 3T6 cells in the presence or absence (0) of 250 MM ATP. Experimental conditions were as in A except that Tris-HCl buffer was adjusted to different pH values and the time of incubation was 20 min. (C) Effect of time of ATP contact on the rate of [3H]uridine efflux. 3T6 cells were exposed to 250MM ATP in medium A for the times indicated (phase 1) after which solution was removed and assayed for radioactivity; the old medium was replaced with medium A without ATP. Incubation at 37.50C was continued (phase 2) so that each dish remained in the bath for a total of 20 min, after which the second supernatant solution was removed and assayed for radioactivity. Ordinate shows the rate of uridine efflux in phase 2. Acid-soluble pools were 200,000,168,000, and 180,000 cpm in A, B, and C, respectively.

medium. The acid-soluble pools were labeled with [3H]uridine 2-deoxy[3H]glucose. We consider the results with 2deoxy[3H]glucose to be somewhat more reliable because we are dealing here with a single labeled phosphorylated compound in the acid-soluble pool. In the case of [3Hjuridine labeling, some breakdown of UTP to UDP and UMP is possible in the presence of inhibitors, and these products may exit from the cells more rapidly than UTP. In any case, the results with [3H]uridine were similar. Excellent synergism was observed between 40 ,gM ATP and rotenone, rutamycin, or CCCP (Table 1). Direct measurement with the luciferase reaction established that the inhibitors decreased intracellular ATP levels under the conditions used (results not shown). In the presence of 2 1AM rotenone, a massive stimulation of efflux of 2-deoxy[3H]glucose 6-phosphate occurred with only 30 tM external ATP, but when the conor

Table 1. Synergism between inhibitors and a low concentration of ATP in causing efflux of deoxy[3H]glucose-labeled pools

Efflux, Inhibitor

ATP

cpm X 103/dish

+

27.2 34.2

+

24.8 114.5

-

39.8 101.0

None

Rutamycin (2 Mg/ml) Rotenone (4MM)

+

CCCP (2 MM)

-

26.7 86.5 The cultures were preincubated with or without the various inhibitors for 5 min and then some of the cultures received 40MM ATP. The incubation was continued for an additional 15 min. +

2106

Cell Biology: Dicker et al.

centration of ATP was further decreased to 15 AM, no stimulation was observed. Sealing of Permeabilized 3T6 Cells. It was particularly important to establish that sealing of the permeability barrier did not depend on components contributed by serum. In earlier studies (1), it was shown that the effect of external ATP could be reversed by replacing the supernatant fluid with medium of neutral pH and containing divalent cations. The possibility existed that serum proteins attached to the cell surface might contribute to the sealing reaction in the presence of divalent cations at neutral pH. In the present investigation, we observed sealing in the complete absence of serum components. The permeabilization process was completely reversed by replacing the supernatant medium with DME medium (Fig. 4). Less complete and slower sealing was obtained with medium A adjusted to pH 6.8 and containing 20 mM glucose, 1 mM of MgCl2, and 1 mM CaCl2. Sealing of cells grown in the absence of serum could also be demonstrated by measuring permeability for the entry of p-nitrophenyl phosphate. The ATP-treated cells were permeable, and hydrolysis of this ester by monolayer cultures could be demonstrated (Fig. 2). After treatment with DME medium or medium A at pH 6.8 supplemented with glucose, Mg2+, and Ca2+, the permeability barrier was restored (results not shown). Membrane Phosphorylation in 3T6 Cells Grown in Serum-Free Medium. It has been suggested that phosphorylation of a membrane protein by extracellular ATP brings about the formation of an aqueous channel in the plasma membrane through which the acid-soluble pools leak out (2). Therefore, it was especially appropriate to determine whether 3T6 cells grown in the absence of serum catalyze the transfer of the terminal phosphate of [Ty-32P]ATP to acceptor membrane proteins. A 10-min exposure of 3T6 cells to radioactive ATP in the presence of 2 MM rotenone led to the labeling of a major plasma membrane protein band that migrated with an apparent molecular weight of about 44,000 as revealed by NaDodSO4/ 200

x100 _

0

y

20 30 Time, min FIG. 4. Effect of transfer to DME medium or medium A (pH 6.8) containing 20 mM glucose, 1 mM CaCl2, and 1 mM MgCl2 on the stimulation of efflux produced by ATP. Cells were labeled with [3H]uridine, washed, and treated with 1 ml of medium A (pH 8.2) plus 250 ,uM ATP for 7.5 min at 37.5°C (phase 1). The medium A was replaced by fresh medium A (pH 8.2) (0), medium A (pH 6.8) supplemented with 20 mM glucose, 1 mM MgCl2, and 1 mM CaCl2 (A), or DME medium (pH 7.2) (v), and incubation at 37.5°C was continued (phase 2). At intervals, the supernatant medium was removed and assayed for radioactivity. DME medium was kept at pH 7.2 by passing a stream of CO2 over it. The ordinate shows the efflux of uridine in the second phase (i.e., time 0 is the end of the first phase). The total acid-soluble pool was 280,000 cpm in this experiment, of which 180,000 cpm remained in the cells after phase 1. 0

10

Proc. Natl. Acad. Sci. USA 77 (1980)

e

FIG. 5. Radioautographs of NaDodSO4/polyacrylamide gel electrophoretograms of 3T6 cells exposed to [y-32P]ATP. Serum-free 3T6 cultures in 90-mm Nunc dishes were produced, were washed four times in 0.15 M NaCl/0.05 mM CaCl2, and then treated as follows. (i) Incubated with 3 ml of medium A (pH 8.2) plus 2 MM rotenone, to which ['y-32P]ATP (50 ,M; 30 ,uCi/ml) was added after 7 min. After a total of 17 min, the culture was washed twice with cold 0.15 M NaCl/0.05 mM CaCl2, and 1 ml of homogenization buffer was added. (ii) Incubated with 3 ml of medium A (pH 8.2) plus 2 ,M rotenone, to which [932P]-ATP (50 ,uM, 30 ,Ci/ml) was added after 7 min. After a total of 17 min, the culture was washed twice with 0.15 M NaCl/0.05 mM CaCl2 at 37°C and then incubated with DME medium (pH 7.2) for 10 min. The culture was then washed twice in cold 0.15 M NaCl/ 0.05 mM CaCl2, and 1 ml of homogenization buffer was added. Cell membranes from these cultures were isolated and electrophoresed on NaDodSO4/polyacrylamide gels which then were autoradiographed. The arrows indicate the position, after electrophoresis, of the following proteins: a, phosphorylase A (Mr, 100,000); b, bovine serum albumin (Mr, 60,000); c, ovalbumin (Mr, 45,000); d, carbonic anhydrase (Mr, 29,000); and e, myoglobin (Mr, 17,000).

polyacrylamide gel electrophoresis and autoradiography (Fig. 5). A most significant observation is that, when cells that were exposed to [y-32P]ATP were washed and transferred to DME medium for 10 min, which reverses the permeability change (see Fig. 4), a loss of the label in the 44,000-dalton band occurred (Fig. 5, lane ii). DISCUSSION It is generally accepted that membranes are composed of a lipid bilayer in which are inserted proteins (integral proteins) and attached to which are other proteins (peripheral proteins) (20). Proteins present in the outer cell membranes of tissue culture cells can be either produced by the cells or adsorbed from serum in the surrounding medium. A considerable adsorption of serum proteins to cell surfaces has been demonstrated in various cell types (7-12). These proteins might be of particular importance in processes taking place at the outer cell surface, specifically in reactions involving ectoenzymes like protein kinase (21) or protein phosphatase (22) and in the change of permeability induced by exogenous ATP. A direct approach to ascertain the role of serum-derived proteins in the extensive permeability change produced by ATP or in the reactions catalyzed by ectoenzymes is to grow the cells under chemically defined conditions. In the present experi-

Cell Biology: Dicker et al. ments, 3T6 cells were grown in a mixture of DME medium and Waymouth's medium supplemented with only two polypeptides, EGF and insulin, which were found to stimulate DNA synthesis and cell division in these cells (23). The cultures were passaged without serum and therefore could not contain any serum left over after plating. Thus, all the results in this paper were obtained in the complete absence of serum proteins. The observation that mouse 3T6 fibroblasts can grow in serum-free medium supplemented with only EGF and insulin agrees with findings in hamster fibroblasts (24, 25). When cultures of 3T6 cells grown under chemically defined conditions are exposed to ATP, there is a dramatic increase in permeability as revealed by exit of [3H]uridine, 2-deoxy[3H]glucose, and 86Rb+ labeled pools or entry of p-nitrophenyl phosphate. The results clearly demonstrate that ATP'interacts with an effector system that is derived from the cells. The kinetics and extent of the change in permeability induced by ATP in these cultures are similar to those previously observed in cells grown in the presence of serum, but some differences should be pointed out. The effect of ATP is elicited by concentrations lower than those used in previous experiments with serumgrown cells, and the optimum pH is around 7.55 (measured at the incubation temperature). Cells rendered permeable by external ATP can be sealed by replacing the medium with DME medium or with a medium at lower pH (6.8) containing cations (Ca2+ and Mg2+) and glucose. All these observations and the previous studies (1-5) indicate that external ATP, when present, can reversibly regulate the permeability of the cell surface and that external and internal ATP exert a reciprocal effect on the permeability barrier of the cells. All these properties are independent of serum-derived proteins. The molecular mechanism of the regulation of membrane permeability by ATP is unknown. Some recent evidence (2, 4, 21, 22, 26, 27) is consistent with the involvement of membrane phosphorylation in the control of membrane permeability. We speculated that a critical protein regulates the passive permeability of the cell membrane and that such a protein can be phosphorylated at both the inner and outer surfaces of the plasma membrane (2). The conformations achieved in this way have an opposing function in controlling cell permeability (2). In this context, a most interesting observation is that permeabilization of 3T6 cells grown in serum-free medium appears to be related to the phosphorylation of a 44,000-dalton membrane protein by the y phosphate of ATP. Such a reaction had been seen earlier when serum-fed cells were used. This observation was made by N. Makan (personal communication) and confirmed and extended by T. Kitagawa (personal communication). In view of the fact that membrane-bound kinases of 3T6 cells can phosphorylate exogenous proteins such as histones (21), it was especially important to establish that the 44,000-

Proc. Natl. Acad. Sci. USA 77 (1980)

2107

dalton band on the gels did not represent an adsorbed serum protein. These findings raise the possibility that the 44,000-dalton protein is directly involved in the regulation of membrane permeability. The potential significance of this possibility for control of membrane permeability warrants further experimental work. We thank Ann Legg and Patricia Pettican for excellent technical assistance. L.A.H. acknowledges with thanks a Wellcome Research Travel Grant. 1. Rozengurt, E., Heppel, L. A. & Friedberg, I. (1977) J. Biol. Chem.

252,4584-4590. 2. Rozengurt, E. & Heppel, L. A. (1979) J. Biol. Chem. 254, 708-714. 3. Rozengurt, E. & Heppel, L. A. (1975) Biochem. Biophys. Res. Commun. 67, 1581-1588. 4. Makan, N. (1978) Exp. Cell Res. 114, 417-427. 5. Makan, N. & Heppel, L. A. (1978) J. Cell. Physiol. 96,87-94. 6. Kitagawa, T. (1980) J. Cell. Physiol., in press. 7. Patterson, M. K., Jr. (1974) J. Natl. Cancer Inst. 53, 14931498. 8. Graham, J. M. (1979) Surface Membrane Enzymes in Neoplasm in Surfaces of Normal and Malignant Cells, ed. Hynes, R. 0. (Wiley, New York), 199-242. 9. Forni, G. & Green, I. (1976) J. Immunol. 116, 1561-1565. 10. Tarone, G. & Comoglio, P. M. (1976) FEBS Lett. 67,364-367. 11. Hynes, R. 0. (1973) Proc. Natl. Acad. Sci. USA 70, 31703174. 12. Hogg, N. M. (1974) Proc. Natl. Acad. Sci. USA 71,489-492. 13. Todaro, G. J. & Green, H. (1963) J. Cell Biol. 17, 299-313. 14. Brunette, D. M. & Till, J. E. (1971) J. Membr. Biol. 5, 215221. 15. Stone, K. R., Smith, R. E. & Joklik, W. K. (1974) Virology 58, 86-100. 16. Fagan, J. B. & Racker, E. (1977) Biochemistry 16, 152-158. 17. Cantley, L. C., Jr., Cantley, L. G. & Josephson, L. (1978) J. Biol. Chem. 253, 7361-7368. 18. Cantley, L. C., Jr., Josephson, L., Warner, R., Yanagisawa, M., Lechene, C. & Guidotti, G. (1977) J. Biol. Chem. 252, 74217423. 19. Beauge, L. A. & Glynn, I. M. (1978) Nature (London) 272, 551-552. 20. Singer, S. J. (1974) Annu. Rev. Biochem. 43, 805-833. 21. Mastro, A. M. & Rozengurt, E. (1976) J. Biol. Chem., 251, 7899-7906. 22. Makan, N. (1979) Biochim. Biophys. Acta 485,379-390. 23. Mierzejewski, K. & Rozengurt, E. (1976) Biochem. Biophys. Res. Commun. 73, 271-277. 24. Hayashi, I. & Sato, G. (1976) Nature (London) 259, 132-134. 25. Cherington, P. V., Smith, B. I. & Pardee, A. B. (1979) Proc. Natl. Acad. Sci. USA 76,3937-3941. 26. Greengard, P. (1978) Science 199, 146-152. 27. Makan, N. (1979) J. Cell. Physiol. 101, 481-492.

Suggest Documents