Studies on the Mechanism of Ca2+ Stimulation of Plasminogen Activator Synthesis/ Re lease by Swiss 3T3 Cells IIH-NAN (GEORGE) CHOU, ROBERT COX AND PAUL H. BLACK Infectious Disease Unit, Massachusetts General Hospital and Department of Medicine, Harvard Medical School, Boston, Massachusetts 02114

ABSTRACT Stimulation of postconfluent Swiss 3T3 cells in serum-free medium with 4.3 mM Ca2+results in marked increases in both released and cellassociated plasminogen activator (PA). Increased release of PA commenced approximately 10 to 12 hours post-stimulation and continued to increase steadily until 48 hours a t which time the stimulated cells (4.3 mM CaZ+)released approximately 14 times more PA than control cells (1.8mM Ca2+).Sr2+,like Ca2+, also stimulates PA synthesis/release either in the presence or in the absence of 1.8 mM Ca2+whereas an excess of Mg2+inhibits Ca2+stimulation. Supranormal [Pi1 in the medium stimulates PA synthesishelease in the presence of 1.8 mM Ca2+.Further, optimal stimulation by 4.3 mM Caz+requires a normal level of Pi (1.0 mM). Elevation of medium [Ca2+lor [Pi1 results in an enhanced uptake of Caz+.The facts that cycloheximide treatment completely abolishes the Ca2+ stimulatory effect and that an increase in cell associated PA precedes release indicate that PA release is coupled to synthesis of new PA. Ca2+stimulation of P A synthesishelease also requires continuous energy production and RNA as well as protein synthesis. A hypothesis is proposed to explain the relationship between stimulation of PA production and its enhanced release from cells stimulated by elevated [Caz+lor [Pi1 in the media. The possibility that PA release may be an example of the phenomenon of membrane shedding as opposed t o secretion is discussed. Elevated levels of plasminogen activator (PA) have been shown to occur in a variety of transformed or malignant cells (Unkeless et al., '73; Ossowski et al., '73; Rifkin et al., '74; reviewed by Roblin et al., '75a). PA, a cellular serine protease (Christman and Acs, '74; Unkeless et al., '74a), catalyzes the conversion of the serum zymogen, plasminogen, to plasmin which is then measured as fibrinolytic activity (Unkeless et al., '73). In addition to transformed cells, some untransformed cells including established mouse 3T3 cell lines and normal embryo cultures of various origins, are also known to produce PA (Chou et al., '74a; Unkeless et al., '74b; Goldberg, '74; Roblin et al., '75b; Rifkin and Pollack, '77; Rohrlich and Rifkin, '77). We have demonstrated that growing Swiss 3T3 cells release higher levels of PA than postconfluent cells and that Simian virus 40 (SV40) transformed 3T3 (SV3T3) cells release .

J. CELL.

PHYSIOL. (1979) 100: 457-466.

PA independent of cell density (Roblin et al., '75c; Chou et al., '77a). Recently, we showed that postconfluent Swiss 3T3 cells could be stimulated t o synthesize and release higher levels of PA by increasing [Ca2+lin the medium (Chou et al., '77b). In this paper, we report the results of our studies on the mechanism of Ca2+stimulation of PA synthesishelease and the factors influencing this phenomenon. The evidence accumulated indicates that Caz+stimulates synthesis of PA and that release is coupled to synthesis of PA. MATERIALS AND METHODS

Cells, media and culture conditions The origin of the Swiss 3T3 cells has been described previously (Chou et al., '74b). In this study, Swiss 3T3 cells were used between their seventh and fifteenth passages. Cells were Received Feb. 21, '79. Accepted May 1, '79.

457

458

IIH-NAN (GEORGE) CHOU, ROBERT COX AND PAUL H. BLACK

routinely grown and maintained in Dulbecco's modified Eagle's medium (Gibco) containing 10% fetal calf serum, penicillin and streptomycin. Special Ca2+-freeand phosphate-free Dulbecco's media, as well as human sera used as plasminogen source, were obtained from Gibco. Cells used in these studies were free from mycoplasma contamination as determined by tritiated thymidine incorporation followed by autoradiography (Culp and Black, '72). Plasminogen activator sample preparation and assay Swiss 3T3 cells were plated in 60-mm Petri dishes (Falcon) a t 8-10 X l o 4 cellddish and grown to confluence prior to use (Chou et al., '77b). Procedures for preparing serum-free harvest fluids (HF) and Triton X-100 cell lysates containing released and cell-associated PA, respectively, have been described (Chou et al., '77b) except that 1%Triton X-100 in 0.1 M Tris-HC1, pH 8.1, was used in the experiments reported herein. The divalent cations or drugs utilized were added to the cultures a t the time of switching cells to serumfree medium (0 hours) and unless indicated otherwise, the cells were incubated for 40 to 44 hours to maximize the Ca2+stimulatory effect before collecting HF's and preparing cell lysates. All HF and lysate samples were frozen in aliquots at - 70°C until assayed. The assays for released and cell-associated PA in HFs and Triton X-100 lysates were performed essentially as described (Chou et al., '77b). The assay mixture contained: 0.1% Triton X-100 in 0.1 MTris-HCl, pH 8.1,5-10 pg of purified human plasminogen (Deutsch and Mertz, '70), and HF or lysate sample in 1.0 ml final volume. One unit of PA is defined as described (Chou e t al., '77b) and the total PA units were normalized to lo5 cells. Plasminogen blanks were included in all assays and subtracted before calculating the net PA activity. In addition, plasminogen independent fibrinolytic activity by HF or lysate alone was routinely found to be negligible. The data presented were the average of two assays of samples prepared from duplicate plates.

and various concentrations of Caz+(1.8 or 4.3 mM) and phosphate (Pi, 1.0 or 7.0 mM). At the end of a 20-minute pulse, the medium was removed from duplicate plates and the cells were washed on the plates six times with 5 ml of a solution containing 200 mM choline chloride, 1 mM ethylene glycol-bis (P-aminoethyl ether) tetra-acetic acid (EGTA, disodium salt) and 1mM Tris-HC1, pH 7.2 (Tupper and Zorgniotti, '77). It should be noted that the washing procedure employed did not remove an appreciable number of cells from the plates. The cells were then lysed with 0.2 N NaOH in 2% Na,CO, and aliquots of cell lysates were neutralized and counted in Ready-Solv E P scintillation fluid (Beckman). Chemicals a n d drugs CaCl,, SrC12, MnC1,.4H20, BaC1,. 2H20, MgSO,. 7H20, ZnS0,.7Hz0, Na4P20,.10H20 and NaF were of reagent grade (Mallinckrodt). Stock solutions of actinomycin D and 2,4-dinitrophenol (Sigma) were prepared in ethanol, frozen and diluted with serum-free medium before use. Stock solutions of cycloheximide, cytosine arabinoside and NaN (Sigma) were prepared in phosphate buffered saline, frozen and diluted in medium. 45CaC12 was obtained from New England Nuclear Corp. RESULTS

Time course of Ca 2+-stimulatedPA synthesishelease Previous dose response studies showed that optimal stimulation of PA synthesishelease by postconfluent Swiss 3T3 cells occurred when the [Ca2+lof the medium was increased to 4.3 mM (Chou et al., '77b). The time course of Ca2+stimulated PA synthesislrelease was determined and the results from a typical experiment are shown in figure 1.Very little PA was released during the first eight to ten hours incubation in serum-free medium containing 1.8 mM (control) or 4.3 mM (stimulated) Ca2+.However, a marked increase in PA release commenced approximately 12 hours poststimulation and by 24 hours the amount of released PA was approximately 8 times greater than that released by 12 hours. The 45ca uptake Caz+-stimulatedrelease of PA continued to 45Ca uptake by the cell was measured by a increase steadily and at 48 hours, when t h e procedure modified from Tupper and Zorg- experiment was terminated, t h e amount of niotti ('77). Postconfluent cells were washed released PA was approximately 24-fold and incubated in serum-free Dulbecco's me- greater than that released at 12 hours. In condium containing 45Ca(4 pCi per mM cold Ca2+) trast, in control cells, released PA remained at

MECHANISM OF CAZ+STIMULATION OF PLASMINOGEN ACTIVATOR

0

8

16

24

40

32

459

48

/NCUBAT/ON T/M€ fHours 1 Fig. 1 Time course of Ca2' stimulation of PA synthesishelease. At time 0, postconfluent Swiss 3T3 cells were washed and incubated in serum-free Dulbecco's medium containing 1.8 mM (0- - -0 and 0-0) or 4.3 (A - - - A and A-A) Caz+.At each time point, HF and Triton X-100lysates containing released (dashed lines) and cell-associated (solid lines) PA, respectively, were prepared and assayed as described under MATERIALS AND METHODS.

low levels and relatively unchanged between 12 and 48 hours post-incubation. Furthermore, as shown in table 1, the ratio of released PA from Ca2+-stimulatedcells to PA released from control cells increased from 1.5 to 13.5 over a 48-hour incubation. Thus, the longer the cells were stimulated, the greater the difference in the amount of released PA between stimulated and control cells. As also shown in figure 1,in the presence of 1.8 mM Ca2+,the amount of cell-associated P A (intracellular pool) remained low and increased only slightly throughout the entire incubation period. In contrast, upon stimulation with 4.3 mM Ca2+,the cell-associated PA began to increase approximately six to eight hours post-stimulation and continued to increase for approximately 36 hours; intracellular levels of PA remained essentially unchanged thereafter. Thus, an increase in cellassociated PA appeared to precede an increase in released PA (fig. 1).

While the cell-associated PA remained essentially unchanged between 36 and 48 hours post-stimulation, the cells continued to release large amounts of PA during this same period (fig. 1). These results suggest that the intracellular pool size of PA may be relatively limited, i.e., it can be increased only by 2- to 3fold (table 1). Furthermore, the total PA (cell released) in Ca2+-stimulated associated cells also increased considerably between 12 and 48 hours post-stimulation (fig. 1, table 1) indicating synthesis of new PA. Since the cellassociated PA did not increase by more than 2to 3-fold, these results suggest that enhanced PA release is coupled to synthesis of new PA.

+

Divalent cation specificity of stimulation In order to determine the specificity of stimulation of PA release, various divalent cations including Sr2', Ba2+,Mg2+,Mn2+,Hg2+ and Zn2+were tested in the presence of 1.8 mM Ca2+.As shown in table 2, Sr2+,like Ca2+,stim-

460

IIH-NAN (GEORGE) CHOU, ROBERT COX AND PAUL H. BLACK TABLE 1

TABLE 3

Ca 2 + stirnulation of plasminogen activator synthesishelease

Substitution of S r 2 +for Caz+in stimulation of plasminogen activator synthesishelease

Ratio of PA activity at Ca2+ 4.3 mMICA1+ 1.8 mM '

Incubation hours Total PA

0 8 12 16 24 36 48

1.0 1.1 1.8 2.3 4.7 3.7 6.3

Released

Cell-associated

-

1.0 1.2 1.9 1.9 3.1 2.0 2.7

1.0 1.5 3.9 8.1 8.4 13.5

Data were calculated from figure 1. Total PA = released + cell-associated PA in units/105 cells calculated from figure l I

TABLE 2

Cu2+stimulation o f PA synthesishelease: Divalent cation specificity and inhibition by high LVgz+l % ! of control PA activity

Ca" (mM) Additions

1.8 4.3 1.8 1.8 1.8 1.8 1.8 1.8 1.8 1.8 1.8 4.3 4.3 4.3

'

0 (control)

0

SrZ+1.2 Sr2+2.4 Sr2' 3.6 BaZ' 0.8 Ba2+1.2 Ba2+1.6 BaZ+2.0 Mn2+0.025 Mn2+ 0.1 0 (control) Mg2+3.8 M g Z +7.8

Released

Cell-associated

100 604 150 374 572 207 282 236 71 172 208 100 71 21

100 211 105 147 183 101 143 118 80 109 106 100 62 39

' 0.8 mM Mg'+ was present in all experiments. 2 Precipitation of very fine particles occurred at [Ba"l 3 2.0 mM. 3Toxic at 1Mn"l 3 0.2 mM.

ulated both release and intracellular accumulation of PA in a dose-dependent manner. In contrast, optimal concentrations of Ba2+(1.2 mM) and Mn2+(0.1 mM) enhanced PA release by approximately 3- and %fold, respectively, with little effect on intracellular PA accumulation. When [Ba2+lwas raised to 2 2 mM, it resulted in precipitation of very fine particles as well as total loss of the Ba2+stimulatory effect. Neither BaZ+nor Mn2+showed a dose-dependent stimulation comparable to that of CaZ+or Sr2+.Non-toxic concentrations of Hg2+ (0.0005-0.002 mM) and Zn2+(0.005-0.03 mM) were without effect. It should be pointed out that 0.8 mM Mg2+was normally present in the

Sr"

CaZ'

% of control PA activity

(mM)

1.8 1.8 1.8 4.2 0' 0 0

0 2.4 3.6 0 1.8 4.2 5.0

Released

Cell-associated

100 374 572 564 218 520 620

100 147 183 239 101 254 232

I The actual 1Ca"I in Ca2+-freemedium was < 0.005 mM as determined by atomic absorption spectrophotometry.

Dulbecco's medium used in these experiments. However, an excess of Mgz+(up to 4.8 mM) in the presence of 1.8 mM CaZ+had no stimulatory effect on PA synthesishelease (data not shown). On the other hand, a n excess of Mg2+ was found to depress the stimulation caused by 4.3 mM Ca2+(table 2).

Sr 2 + substitution for Ca 2+ Since Sr2+could stimulate PA synthesishelease in the presence of 1.8mM Ca2+,we determined whether Sr2+was stimulatory in the absence of Ca2+.Ca2+-freeDulbecco's medium was supplemented with various concentrations of Ca2+andlor SrZ+as shown in table 3. As can be seen, 1.8 mM Sr2+in Caz+-deficient medium (see footnote, table 31, caused an approximately %fold increase in released PA as compared to 1.8 mM Ca2+alone. In addition, 4.2 mM Sr2+appeared to be as effective as 4.2 mM Ca2+in stimulation of PA synthesisirelease. A combination of 1.8 mM Ca2+with 2.4 mM Sr2+(total 4.2 mM) was slightly less stimulatory than 4.2 mM of either Ca2+or Sr2+ alone. These results indicate that Sr2+is essentially as effective as CaZ+in stimulating PA synthesishelease. Role o f phosphate (Pi)in Ca 2+-stimulated PA synthesishelease Supranormal [Ca2+l (Dulbecco and Elkington, '75; Boyton and Whitfield, '76; Rubin and Sanui, '77; Barnes and Colowick, '77) or [Pi] (Rubin and Sanui, '77; Barnes and Colowick, '77) have been shown to stimulate DNA synthesis and sugar uptake in confluent 3T3 cells. In addition, Rubin and Sanui ('77) and Barnes and Colowick ('77) have demonstrated that the stimulatory effect of supranormal [Ca2+lis

46 1

MECHANISM OF CA2' STIMULATION OF PLASMINOGEN ACTIVATOR TABLE 4

TABLE 5

Effect of varying Pi concentrations on plasninogen activator synthesishelease I

Effect of drugs on Ca 2+ stimulation of plasrninogen activator synthesishelease

Released

1.8

1.0 (control)

1.8 1.8 1.8 1.8 1.8 1.8 4.3 4.3 4.3 4.3

0 0.25 4.0 5.0 7.5 10.0 1.0 2.0 3.0 0

% of control PA activity

% of control PA activity

CaZ+ (mM) Pi

100

97 85 437 1,132 1,179 476 1,162(100) 395(34) 42( 4) 148(13)

Treatment

Cell-associated

Released

Cellassociated

100 27 4 8

100 36 6 7

9

22

100 80 82 220 279 361 240 353(100) 49(14) 13(4) 80(23)

' Phosphate free Dulbecco's medium containing 1.8 mM Ca" was supplemented with various amounts of a sterile stock solution of 100 mM Na2HP0, andlor 250 mM CaCI, to give the desired concentrations of Ca2+and Pi. ZValues in ( f are % PA activity relative to PA a t 4.3 mM Ca" and 1.0 mM Pi (= 100).

Ca2+4.3 mM (control) Actinomycin D (0.025 pg/ml) Cycloheximide (0.5 pg/ml) 2,4-dinitrophenol (0.5 mM) NaF (2.5 mM) NaN, (1.0 mM)

+

' The

solvent control (0.005-0.5% ethanol) had no stimulatory

effect. At the end of the experiment, the drug-treated cells remained attached to the plate and, after trypsinization, 95%were found to be viable as judged by the ability to exclude Trypan Blue.

required a normal level of Pi (approximately 1.0 mM) since Pi deprivation in the presence of 4.3 mM Ca2+resulted in marked diminution in both released and cell-associated P A (87 and 77%,respectively). dependent on [Pil, and vice versa, and atPyrophosphate (PPi) has also been shown to tributed this to the action of Ca-Pi precipi- stimulate DNA synthesis and sugar uptake in tates formed in the medium. We, therefore, de- confluent 3T3 cells (Rubin and Sanui, '77; termined whether increased [Pi] in the pres- Barnes and Colowick, '77). In this connection, ence of 1.8mM Ca2+stimulated PA synthesis/ it is of interest to note that 0.1-0.5 mM PPi in release and whether Pi was required for Ca2+ the presence of 1.8 mM Ca2+and 1.0 mM Pi stimulation. As shown in table 4, when [Pil also stimulated PA release by as much as 6-t o was raised from 1.0 mM (normal level) to 4.0 %fold although i t increased the level of cellmM in the presence of 1.8 mM Ca2+,an in- associated PA only 2-fold (data not shown). crease of approximately 4- and 2-fold in the reRequirements for energy, protein and R N A leased and cell-associated PA, respectively, synthesis for Ca 2+stirnuZation was observed. Maximal stimulation occurred and, in fact, plateaued when [Pil was raised to To assess further the metabolic require5.0-7.0 mM which caused approximately 11- ments for Ca2+stimulation, we tested the efto 12- and 3- to 4-fold increases in PA release fect of a variety of drugs including inhibitors and intracellular accumulation, respectively. of oxidative phosphorylation, protein, DNA Some granular precipitates were visible when and RNA synthesis, on PA synthesidrelease. [Pi] was 2 8 mM. Higher [Pi1 ( > l o mM) The results are shown in table 5. As can be caused heavy precipitation and a marked de- seen, 0.5 mM 2,4-dinitrophenol inhibited Ca2+crease in the amount of released PA compared stimulated release and intracellular accumuwith optimal stimulation. lation of PA by greater than 90% whereas a As also shown in table 4, in the presence of combination of NaF (2.5 mM) with NaN, (1.0 4.3 mM Ca2+,optimal stimulation of PA syn- mM) diminished the released and cell-associthesishelease occurred a t 1.0 mM Pi. How- ated PA to 9% and 22%, respectively, of the ever, when [Pi] was raised to 2.0 mM and 3.0 control values. These results indicate that mM, respectively, moderate and heavy pre- continual energy generation is required for cipitation occurred. Under these conditions, sustained synthesishelease of PA stimulated the levels of both released and cell-associated by 4.3 mM Ca2+.In addition, the CaZ+stimuPA were severely and progressively dimin- lated PA synthesis/release was also blocked ished compared with optimal stimulation. In by low concentrations of cycloheximide (0.5 addition, the results in table 4 also show that pg/ml) and actinomycin D (0.025 pg/ml); the maximal stimulation of PA by 4.3 mM Ca2+ former caused 96% and 94% while the latter

462

IIH-NAN (GEORGE)CHOU, ROBERT COX AND PAUL H. BLACK

caused 73%and 64%inhibition of released and intracellular accumulation of PA, respectively. These results indicate that continued protein as well as RNA synthesis is required for sustained PA synthesishelease stimulated by Ca2+.Thus, in the presence of cylcoheximide or actinomycin D, the cells released less P A presumably due to diminished synthesis of new PA. In fact, actinomycin D or cycloheximide treatment in the presence of 4.3 mM Ca2+ caused the cells t o behave like unstimulated cells in terms of their ability t o synthesize and release PA. The effect of cytosine arabinoside was also studied and the results showed that at concentrations up to 10 Fglml, cytosine arabinoside had little to no effect on PA synthesislrelease although DNA synthesis was inhibited by 90-95%. These results indicate that DNA synthesis induced by CaZ+treatment is not necessary for Ca2+stimulation of PA synthesishelease. Thus, our results are in good agreement with those obtained by other investigators using different cell systems (Rohrlich and Riain, '77).

TABLE 6

Enhanced EGTA insensitiue V a uptake by eleuated [Cal'l or Pi1 in the medium Ca"

Pi

'%a uptake in 20-min pulse

(mM1

1.8 4.3 1.8

1.0 (control) 1.0 7.0

cpmlplate

%

1,984 11,732 26,716

100 591 1,347

uptake and presumably an increased intracellular Ca content. DISCUSSION

We have shown that Ca2+stimulation of postconfluent Swiss 3T3 cells causes marked increases in both extracellular release and intracellular accumulation of PA. Thus, the total amount of PA is markedly increased following Ca2+stimulation. However, the amount of intracellular (cell-associated) PA did not increase by more than 2- to 3-fold suggesting that the intracellular pool size of PA Enhanced EGTA-insensitive 45cauptake at may be relatively limited. Thus, prolonged elevated medium [Ca2+lor[Pi1 stimulation resulted in more synthesis and Elevation of the medium [Ca2+lmay result subsequent release of P A indicating that PA in an increased CaZ+influx and intracellular release is linked to PA synthesis. Further[Ca2+las suggested by Dulbecco and Elkington more, CaZ+ stimulated PA release was ('75). Indeed, Moore and Pastan ('77) have markedly diminished when synthesis of new shown that microsomal Ca2+uptake activity P A was prevented by actinomycin D or cycloof 3T3 cells is enhanced when the medium heximide. In addition, optimal stimulation of [Ca2+lis raised. In addition, Borle ('72) has sustained PA synthesishelease is dependent shown that the total intracellular Ca content on continual energy generation. Taken tois markedly increased by raising the medium gether, we conclude that the CaZ+stimula[Ca2+lor [Pi1 and that the Ca2+efflux is inhib- tory effect occurs by increasing the syntheited by elevating the medium [Pi]; conversely, sis of new P A followed by increased release of Ca2+efflux is stimulated by depleting Pi from PA. Sr2+,like Ca2+,can stimulate PA synthethe medium. Therefore, we measured the sislrelease in a dose-dependent manner either uptake of Ca" which is located intracellularly in the presence of 1.8 mM CaZ+or in the abas defined by Tupper and Zorgniotti ('771, sence of Ca2+.In contrast, in control cells (1.8 when the medium [CaZ+lor [Pi] was elevated. mM CaZ+), the amount of total PA produced by The results from a typical 20-minute pulse ex- the cell remained low and relatively unperiment are shown in table 6. It should be changed. The majority remained as intracelnoted that this experiment was not meant to lular PA and only a small and relatively conanalyze the kinetics of Ca2+uptake nor its stant fraction of the total was released by the compartmentalization. As can be seen, 20 cell. minutes after raising the medium [CaZ+lfrom In the presence of 1.8 mM Caz+,elevated Pi 1.8 mM to 4.3 mM, the EGTA-resistant 45Ca concentrations stimulate PA synthesisheuptake was increased by approximately 6-fold. lease. In addition, optimal stimulation of PA Similarly, when the medium [Pi]was elevated synthesishelease by 4.3 mM CaZ+requires a from 1.0 mM to 7.0 mM for 20 minutes, the normal level of Pi (1.0 mM). Furthermore, the EGTA-insensitive 45Ca uptake was enhanced stimulation of PA synthesis/release by eleby 13.5-fold.Thus, an increase in the medium vated [Ca2+land [Pi] appears to involve an en[Ca2+lor [Pi1 results in an enhanced CaZ+ hanced Ca2+uptake by the cells which occurs

MECHANISM OF CA2' STIMULATION OF PLASMINOGEN ACTIVATOR

when the medium [Ca2+l or [Pi] is raised. These results are similar to those obtained by others (Borle, '72; Moore and Pastan, '771, and agree with our earlier finding that the ionophore A23187 in the presence of 1.8 mM Ca2+ stimulates PA synthesishelease (Chou et al., '77b). Release of cellular PA into the culture medium has generally been referred to as a secretion phenomenon. Exocytosis serves as a general mechanism for secretion of substances contained within cytoplasmic secretory granules or vesicles (reviewed by Douglas, '74; Rubin, '74; Palade, '75). However, P A has been shown to be membrane associated (Unkeless et al., '74a; Christman et al., '75; Dvorak et al., '78) and more recently, PA has been found to be associated with a more purified plasma membrane-enriched fraction of the cell (Quigley, '76; Jaken and Black, '79). These results suggest that P A may be an integral membrane protein (Singer and Nicolson, '721, or a protein firmly associated with membranes (Quigley, '76). Thus, release of P A from the cell may more appropriately be referred to as a phenomenon of membrane protein shedding. Shedding of other membrane components has been described (Kapeller et al., '73; Doljanski and Kapeller, '76; LaMont e t al., '77). Several similarities as well as unique differences exist between exocytosis and PA shedding. Thus, both shedding and exocytosis require Ca2+and continuous energy production, may utilize Sr2+,but not Mg2+,in place of Ca2+,and are inhibited by an excess of Ca2+or Mg2+ (Douglas, '74; Rubin, '74). However, Ca2+-stimulatedshedding of PA requires both RNA and protein synthesis and commences after 10 to 12 hours stimulation whereas exocytosis of secretory products does not require protein synthesis (Jamieson and Palade, '71) and occurs soon after triggering with an appropriate stimulus (Douglas, '74; Rubin, '74; Palade, '75). The precise mechanism whereby Ca2+stimulates PA shedding is not known. However, the data presented herein indicate that Ca2+ stimulates PA synthesis. If PA is indeed an integral membrane protein, then Ca2+may play an essential role in the fusion event whereby PA may be inserted into the plasma membrane. Much current evidence indicates that integral plasma membrane proteins are inserted into the membrane by fusion of the vesicle or granule containing the membrane-

463

bound proteink) with the plasma membrane (reviewed by Leblond and Bennett, '77). Whether intracellular Ca-Pi precipitates, which have been visualized a t areas of close membrane contact where human erythrocytes ghosts are believed to undergo fusion in the presence of Ca2+and Pi (Zakai et al., '77), are present is not known; such precipitates have been thought to enhance membrane fusion of the ghosts (Zakai et al., '77). Extracellular Ca-Pi precipitates formed in the medium have been shown to stimulate DNA synthesis as well as sugar uptake (Rubin and Sanui, '77; Barnes and Colowick, '77). In these studies, maximal stimulation of DNA synthesis occurred a t 10-12 mM Ca2+,which was approximately 40-fold more effective than 4 mM Ca2+,in the presence of 1 mM Pi. Furthermore, stimulation of DNA synthesis was greatly enhanced when [Pil was raised from 1.0 mM to 5.0 mM in the presence of 4.0 mM Caz+(Rubin and Sanui, '77). Similarly, maximal stimulation of sugar uptake occurred a t 12-15 mM Ca2+plus 1 mM Pi or a t 15-20 mM Pi together with 2 mM Ca2+(Barnes and Colowick, '77). Large amounts of precipitates were formed a t these high concentrations of Ca2+or Pi. These results indicate that formation of extracellular Ca-Pi precipitates is correlated with the stimulatory activity of Ca2+or Pi, especially a t very high concentrations, on DNA synthesis and sugar uptake. In view of these findings, one may question whether the stimulatory effects described herein on P A synthesishelease are due to extracellular precipitates. In the studies described, however, optimal stimulation of PA synthesishelease occurred a t 4.3 mM Ca2+and Ca2+concentrations 2 5 mM were found to depress the optimal stimulation caused by 4.3 mM Ca2+(Chou et al., '77b). In addition, although optimal stimulation by 4.3 mM Ca2+ requires the presence of 1 mM Pi, an increase of [Pi1 from 1 mM to 2 and 3 mM resulted in formation of large amounts of precipitates as well as approximately 70% and 96%,respectively, inhibition of the maximal stimulatory activity (table 4, bottom and footnote). Additional experiments were carried out in which the medium containing various concentrations of Ca2+and 1.0 mM Pi were pre-incubated a t 37°C in a CO, incubator for 16 to 18 hours, filtered (Millipore Filter, 0.45 pm) and then used for stimulation of P A synthesis1 release. The results showed that the nonfiltered medium containing 6.8 mM Ca2+and

464

IIH-NAN (GEORGE) CHOU, ROBERT COX AND PAUL H. BLACK

1.0 mM Pi was much less stimulatory than the non-filtered medium containing 4.3 mM Ca2+ and 1.0 mM Pi. After filtration, however, the medium originally containing 6.8 mM Ca" and 1.0 mM Pi showed a stimulatory effect comparable t o that of filtered medium containing 4.3 mM Ca2+and 1.0 mM Pi (data not shown); in fact, the filtered medium containing 4.3 mM Ca2+and 1.0 mM Pi was slightly more stimulatory than the non-filtered medium. This may be due t o removal of a small amount of precipitate formed under these conditions (data not shown). Furthermore, Ba2+ a t 0.8-1.6mM normally stimulated P A release by 2- to 3-fold without visible precipitates; however, when [Ba2+lwas raised to 2 2 . 0 mM, it resulted in precipitation of very fine particles and complete loss of the Ba2+stimulatory effect (table 2). Thus, the formation of such precipitates extracellularly appears to be inhibitory t o stimulation of PA synthesis and release caused by various agents. Nevertheless, our studies cannot rule out the possibility that precipitate particles smaller than 0.45 pm in diameter could be internalized by the cells and may play a role in stimulation of PA synthesishelease. It is of interest to note that substances which form complexes with Ca2+, such as ADP, ATP and sodium tripolyphosphate (all from Sigma), also stimulate PA synthesis/ release in the presence of 1.8 mM or 4.3 mM Ca2+.Thus, addition of ADP (0.5-2.0 mM), or ATP (0.5-2.0mM) or sodium tripolyphosphate (0.25-0.5mM) to the serum-free medium containing 1.8 mM Ca2+resulted in an increase in PA synthesislrelease. However, in the presence of 4.3 mM Ca2+,the effective concentrations of ATP and sodium tripolyphosphate which stimulated PA synthesis/release were 0.25-1.0 mM and 0.125-0.25 mM, respectively (unpublished data). The mechanismk) whereby these substances stimulate PA synthesis/release is not known. In contrast, Rubin and Sanui ('77) reported that none of these agents stimulated DNA synthesis since they did not form precipitates with Ca'+. We propose the following working hypothesis to explain the Ca2+and Pi stimulation of PA synthesishelease. When extracellular [Ca2+lor [Pi1 is raised, it results in an increased CaZ+ influx, thereby increasing the intracellular Ca content. Subsequently, both transcriptional and translational activities are enhanced by an as yet undetermined mechanism with a resultant increase in syn-

thesis of new PA. With continued new synthesis of PA, a fraction of the total PA is shed from the cells after translocation and incorporation into the plasma membrane presumably by a membrane-fusion reaction. It is likely that the increased synthesis as well as the enhanced fusion of PA containing vesicles or granules with the plasma membrane results in the enhanced shedding of soluble PA molecules. The mechanism of cleavage of PA from the membrane during shedding and whether proteolysis is involved is not known. Enhanced release of PA also occurs in growing 3T3 as well as transformed 3T3 cells compared with confluent 3T3 cells (Chou e t al., '77a). Whether the mechanisms postulated here are operative in these cells is under investigation. ACKNOWLEDGMENTS

We thank Drs. E. W. Schroder and S. Jaken for their critical comments on the manuscript, Dr. R. Neer for determining [Ca2+lby atomic adsorption spectrophotometry, C. Dietrich and for excellent assistance. I.N.C. is supported by a Cancer Research Scholar Award from the American Cancer Society, Massachusetts Division. This research was supported by Grant CA 10126 from the National Institutes of Health. LITERATURE CITED Barnes, D. W., and S . P.Colowick 1977 Stimulation of sugar uptake and thymidine incorporation in mouse 3T3 cells by calcium phosphate and other extracellular particles. Proc. Nat. Acad. Sci. (U.S.A.), 74: 5593-5597. Borle, A. B. 1972 Kinetic analyses of calcium movements in cell cultures. J. Mem. Biol., 10: 45-66. Boyton, A. L., and J. F. Whitfield 1976 Different actions of normal and supranormal calcium concentrations on the proliferation of BALB/c 3T3 mouse cells. In Vitro, 7: 479-484. Chou, I. N., P. H. Black and R. 0. Roblin 1974a Suppression of fibrinolysin T activity fails to restore density-dependent growth inhibition to SV3T3 cells. Nature, 250: 739-741. 1974b Non-selective inhibition of transformed cell growth by a protease inhibitor. Proc. Nat. Acad. Sci. (U.S.A.), 71: 1748-1752. Chou, I. N., S. P. O h n n e l l , P. H. Black and R. 0. Roblin 1977a Cell-density dependent secretion of plasminogen activator by 3T3 cells. J. Cell. Physiol., 91: 31-38. Chou, I. N., R. 0. Roblin and P. H. Black 197% Calcium stimulation of plasminogen activator secretioniproduction by Swiss 3T3 cells. J. Biol. Chem., 252: 6256-6259. Christman, J. K., and G. Acs 1974 Purification and characterization of a cellular plasminogen activator associated with oncogenic-transformation:the plasminogen activator from SV40-transformed hamster cells. Biochim. Biophys. Acta, 340: 339-351. Christman, J. K., G . Acs, S. Silagi and S. C. Silverstein 1975 Plasminogen activator: Biochemical characterization and correlation with tumorigenicity. In: Proteases and

MECHANISM OF CA2+ STIMULATION OF PLASMINOGEN ACTIVATOR Biological Control. E. Reich, D. B. Rifkin and E. Shaw, eds. Cold Spring Harbor Laboratory, Cold Spring Harbor, pp. 827-839. Culp, L. A,, and P. H. Black 1972 Contact-inhibited revertant cell lines isolated from SV40-transformed cells. J. Virol., 9: 611-620. Deutsch, D. G., and E. T. Mertz 1970 Plasminogen: purification from human plasma by affinity chromatography. Science, 170: 1095-1096. Doljanski, F., and M. Kapeller 1976 Cell surface shedding-the phenomenon and its possible significance. J. Theor. Biol., 62: 253-270. Douglas, W. W. 1974 Involvement of calcium in exocytosis and the exocytosis-vesiculation sequence. Biochem. Soc. symp., 39: 1-28. Dulbecco, R., and J. Elkington 1975 Induction of growth in resting fibroblastic cell cultures by Caz+.Proc. Nat. Acad. Sci. (U.S.A.), 72: 1584-1588. Dvorak, H. F., N. S. Orenstein, J. Rypysc, R. B. Colvin and A. M. Dvorak 1978 Plasminogen activator of guinea pig basophilic leukocytes: probable location to the plasma membrane. J. Immunol., 120: 766-773. Goldberg, A. R. 1974 Increased protease levels in transformed cells: a casein overlay assay for the detection of plasminogen activator production. Cell, 2: 95-102. Jaken, S.,and P. H. Black 1979 Differences in intracellular distribution of plasminogen activator in growing, confluent, and transformed 3T3 cells. Proc. Nat. Acad. Sci. (U.S.A.), 76: 246-250, 1979. Jamieson, J. C., and G. Palade 1971 Synthesis, intracellular transport, and discharge of secretory proteins in stimulated pancreatic exocrine cells. J. Cell Biol., 50: 135-158. Kapeller, M., R. Gal-Oz, N. B. Grover and F. Doljanski 1973 Natural shedding of carbohydrate containing macromolecules from cell surfaces. Exp. Cell Res., 79: 152-158. LaMont, J. T., M. T. Gammon and K. J. Isselbacher 1977 Cell surface glycosyltransferases in cultured fibroblasts: increased activity and release during serum stimulation of growth. Proc. Nat. Acad. Sci. (U.S.A.), 74: 1086-1090. Leblond, C. P., and G. Bennett 1977 Role of t h e Golgi apparatus in terminal glycosylation. In: International Cell Biology, 1976-1977.R. Brinkley and K. R. Porter, eds. The Rockefeller University Press, pp. 326-340. Moore, L., and I. Pastan 1977 Energy-dependent calcium uptake activity in cultured mouse fibroblast microsomes. J. Biol. Chem., 252: 6304-6309. Ossowski, L., J. C. Unkeless, A. Tobia, J. P. Quigley, D. B. Rifkin and E. Reich 1973 An enzymatic function associated with transformation of fibroblasts by oncogenic viruses. J. Exp. Med., 137: 112-126.

465

Palade, G. 1975 Intracellular aspects of the process of protein synthesis. Science, 189: 347-358. Quigley, J. P. 1976 Association of a protease (plasminogen activator) with a specific membrane fraction isolated from transformed cells. J. Cell Biol., 71: 472-486. Rifkin, D. B., J. N. Loeb, G. Moore and E. Reich 1974 Properties of plasminogen activators formed by neoplastic human cell cultures. J. Exp. Med., 139: 1317-1328. Rifkin, D. B., and E. R. Pollack 1977 Production of plasminogen activator by established cell lines of mouse oripin. J. Cell Biol., 73: 47-55. Roblin, R. O., I. N. Chou and P. H. Black 1975a Proteolytic enzymes, cell surface changes, and viral transformation. Adv. Cancer Res., 22: 203-260. 1975b Secretion of plasminogen activator by normal mouse cells in vitro. J. Cell Biol., 67: 366a. 1975c Role of fibrinolysin T activity in properties of 3T3 and SV3T3 cells. In: Proteases and Biological Control. E. Reich, D. B. Rifkin and E. Shaw, eds. Cold Spring Harbor Laboratory, Cold Spring Harbor, New York, pp. 869-884. Rohrlich, S. T., and D. B. Rifkin 1977 Patterns of plasminogen activator production in cultured normal embryonic cells. J. Cell Biol., 75: 31-42. Rubin, H., and H. Sanui 1977 Complexes of inorganic pyrophosphate, orthophosphate and calcium as stimulants of 3T3 cell multiplication. Proc. Nat. Acad. Sci. (U.S.A.), 74: 5026-5030. Rubin, R. P. 1974 Calcium and the Secretory Processes. R. P. Rubin, ed. Plenum Press, New York, pp. 1-150. Singer, S. J., and G. L. Nicolson 1972 The fluid mosaic model of the structure of cell membranes. Science, 175: 720-731. Tupper, J. T., and F. Zorgniotti 1977 Calcium content and distribution as a function of growth and transformation in the mouse 3T3 cells. J. Cell Biol., 75: 12-22. Unkeless, J., K. Dano, G. M. Kellerman and E. Reich 1974a Fibrinolysis associated with oncogenic transformation. J. Biol. Chem., 249: 4295-4305. Unkeless, J. C., S. Gordon and E. Reicb 1974b Secretion of plasminogen activator by stimulated macrophages. J. Exp. Med., 139: 834-850. Unkeless, J. C., A. Tobia, L. Ossowski, J. R. Quigley, D. B. Rifkin and E. Reich 1973 An enzymatic function associated with transformation of fibroblasts by oncogenic viruses. J. Exp. Med., 137: 85-111. Zakai, N., R. G. Kulka and A. Loyter 1977 Membrane ultrastructural changes during calcium phosphate-induced fusion of human erythrocyte ghosts. Proc. Nat. Acad. Sci. (U.S.A.), 74: 2417-2421.