Plasminogen activator from cultured baby hamster kidney cells

A large number of cell clones were isolated which produced different amount of extracellular plasminogen activator. These clones were identified and selected by use of an overlay medium containing casein and plasminogen. All cells obtained by this method arose from a high passage cell line (BHK-21/N) which had been originally cloned from BKH-21/C13 cells. These cloned isolated from BHK-21/N were characterized with respect to their growth rates and efficiencies of plating in soft agar and on plastic surfaces. No correlations were found between any of these growth properties and the amounts of extracellular plasminogen activator as detected in serum-free cell culture media. A plasminogen activator which was produced by a high passage hamster kidney cell (Hpa-6) was characterized biochemically. The enzyme was found to be inhibited by diisopropylfluorophosphate and therefore was considered serine protease. Hpa-6 antivator was also found to react with plasminogen in a manner which was indistinguishable from the action of urokinase on the proenzyme. Evidence was obtained which suggests that the activator can hydrolyze either of the two molecular forms of plasminogen, i.e., plasminogen with glutamic acid amino terminus or plasminogen with lysine amino terminus at a single position to yield active plasmin. The cellular localization of the Hpa-6 cell plasminogen activator was examined and the results obtained indicate that this activity was largely associated with the plasma membrane.

PLASMINOGEN ACTIVATOR FROM CULTURED BABY HAMSTER KIDNEY CELLS by Robert Wickert Snyder A dissertation submitted to the faculty of the University of Utah in partial fulfillment of the requirements for the degree of Doctor of Philosophy Department of Microbiology University of Utah August 1977 THE UNIVERSITY OF UTAH GRADUATE SCHOOL SUPERVISOR Y COMM ITTEE APPROVAL of a dissertation submitted by Robert Wickert Snyder 1: ) 4�4�U!lLt�L-¥ I have read this dissertation and have found it to be of satisfactory quality for a p doctoral de ree. q/ '// . J D, 11 I • Douglas W· '(( ' Hill, Ph.D. Chairman, Supervisory Committee satisfactory quality for a I have read this di sertation and have found it to be of satisfactory quality for a doctoral degree. 1-,,;·' if Date Paul S. Lombardi, Ph.D. Member, Supervisory Committee I have read this dissertation and have found it to be of satisfactory quality for a pfal cfegrec. 1/i';!77 doct Date �tz- · Frank J. O'�l Member, Supervisory Committee I have read this dissertation and have found it to be of satisfactory quality for a doctoral degree. ' /30 / 'Yp Datil 7 Ph.D. ber, Supervisory Committee ,/',/ THE UNIVERSITY OF UTAH GRADUATE SCHOOL FINAL READING APPROVAL To the Graduate Council of The University of Utah: I have read the dissertation of Robert v.Jickert Snyder in its final form and have found that (I) its format, citations, and bibliographic style are consistent and acceptable; (2) its illustrative materials including figures, tables. and charts are in place: and (3) the final manuscript is satisfactory to the Supervisory Committee and is ready for submission to the Graduate School. Douglas w. Memher. Supcrvi,;irl' Committee the Major Department Donald F. Summers, Chairman, Dean M.D. Approved for the Graduate Council di /J , " /111 -<-�2. Itrt:;;:. -/'),1. Sterlj:ng M. , )/! )J./tA, l c· McMurrin, I<l'ean of The Graduate School ' f'/_yt./' Ph.D. ABSTRACT A large number of cell clones were isolated which produced different amounts of extracellular plasminogen activator. These clones were identified and selected by use of an overlay medium containing casein and plasminogen. All cells obtained by this method arose from a high passage cell line (BHK-2l/N) which had been originally cloned from BKH-2l/C13 cells. These cloned isolates from BHK-2l/N were characterized with respect to their growth rates and efficiencies of plating in soft agar and on plastic surfaces. No correlations were found between any of these growth properties and the amounts of extracellular plasminogen activator as detected in serum-free cell culture media. A plasminogen activator which was produced by a high passage hamster kidney cell line (Hpa-6) was characterized biochemically. The enzyme was found to be inhibited by diisopropylfluorophosphate and therefore was considered a serine protease. Hpa-6 cell anti- vator was also found to react with plasminogen in a manner which was indistinguishable from the action of urokinase on the proenzyme. Evidence was obtained which suggests that the activator can hydrolyze either of the two molecular forms of plasminogen, i.e., plasminogen with glutamic acid amino terminus or plasminogen with lysine amino terminus at a single position to yield active plasmin. The cellular localization of the Hpa-6 cell plasminogen activator was examined and the results obtained indicate that this activity was largely associated with the plasma membrane. v ACKNOWLEDGMENTS I would like to extend special thanks to my parents, Marjorie and Paul, for their support and encouragement since my earliest recollections. The help of Drs. Lombardi and O'Neill during this project is appreciated. Drs. Summers and Hill provided much help in the preparation of this manuscript and their advice, patience, and efforts on my behalf during my tenure in their laboratories will not be forgotten. Finally, but most importantly, Lynn and the kids, Kristen, Buggs and Moriah persevered and their contributions cannot be measured. TABLE OF CONTENTS Page ABSTRACT • . . . . iv ACKNOWLEDGMENTS . vi LIST OF TABLES • . • viii LIST OF FIGURES . ix CHAPTER I. LITERATURE REVIEW. INTRODUCTION . . . • . . . . . .• PHENOTYPIC PROPERTIES OF NORMAL AND TRANSFORMED CELLS INCLUDING EVIDENCE FOR OR AGAINST THE INVOLVEMENT OF PROTEASES IN THE EXPRESSION OF THESE CHARACTERISTICS . . . • . • . • . PLASMINOGEN ACTIVATOR IN CELL CULTURE FLUIDS . THE NATURE OF PLASMINOGEN ACTIVATOR. . . . .. REFERENCES . . . . . . . . . . . . . . II. ISOLATION AND CHARACTERIZATION OF CLONED BHK CELLS WHICH RELEASE DIFFERENT AMOUNTS OF PLASMINOGEN ACTIVATOR. . . • . SUMMARY. INTRODUCTION . . MATERIALS AND METHODS. RESULTS. . . DISCUSSION . REFERENCES • III. BIOCHEMICAL PROPERTIES OF A PLASMINOGEN ACTIVATOR FROM BHK CELLS . . SUMMARY. • • INTRODUCTION . . . . . MATERIALS AND METHODS. . RESULTS. . • DISCUSSION REFERENCES . VITA . . • 1 2 4 12 23 29 34 35 35 37 39 52 56 58 59 59 61 66 84 91 94 LIST OF TABLES Page Table 1. BHK-2l/N, Hpa, and Lpa cells: frequencies of activator positive colonies and relative amounts of activator in serum-free cell culture fluids . . . • . . • • . . • . . . . . . 40 Frequencies of activator positive colonies and levels of activator in serum-free supernates of co-cultivated cells. • . • . . • . . • . . 43 3. Intracellular plasminogen activator. 44 4. Plasminogen activator production and efficiencies of plating in soft agar and on plastic surfaces. 45 5. Generation times of Hpa and Lpa isolates . 49 6. Frequencies of activator positive colonies during extended periods of serial cell passages . . . • . . 50 7. Co-cultivation of Hpa and Lpa cells. . • • 51 8. Concentration of activator from Hpa-6 cell . culture fluid • . . • . . . . . . . . . . 67 High speed centrifugation of cell culture fluid with and without Triton X 100 . . • . . . . . . • . • . . 85 2. 9. LIST OF FIGURES Page Figure 1. Growth of Hpa and Lpa isolates . . . • 47 2. Polyacrylamide SDS gel electrophoresis of 1251 plasminogen and activation products . . • . • • 69 Kinetics of plasminogen activation by Hpa-6 cell activator. . • . . • . . • ... 72 Polyacrylamide SDS gel electrophoresis of 3H plasminogen activated by various concentrations of stage III Hpa-6 cell activator • . • . . . • . • . 75 Polyacrylamide SDS gel electrophoresis of plasminogen activated by DFP treated stage III activator . 77 Polyacrylamide SDS gel electrophoresis of 3H plasminogen activated by stage III activator in the presence of TAME • . • . • . • • • . . . . . . . . 80 Sucrose gradient fractionation of Hpa-6 cellular material . . . . . . . . . • . . . . . . . . . 83 3. 4. 5. 6. 7. CHAPTER I LITERATURE REVIEW 2 INTRODUCTION A large amount of the recent work in cell biology has been directed toward study of the neoplastic state at a cellular level and in this regard investigators have often examined the in vitro parameters of cell growth and/or behavior. In many of the reports cells have been classified as being "normal" or "transformed." Initially, the term transformed was used in reference to cells which had undergone a phenotypic change. For example, transformed was employed to designate cells which exhibited a loss of densitydependent inhibition of growth 11,2], increased plating efficiencies on plastic surfaces [3] or the abilities to grow in suspension [4,5,6]. On the contrary, normal cells were reported to exhibit density inhibition 1l,2J, low efficiency of plating 13] and the inabilities to grow in suspension 14,5,6J. More recently transfor- mation has also been reportedly accompanied by increased lectin agglutinability [7,8,9], lower levels of cyclic adenosine 3' 5' monophosphate [10], and the production and/or release of plasminogen activators [11,12]. The term transformed has often been used synonymously with the neoplastic state. In this regard, some of the early studies demonstrated that transformed cells were capable of proliferating in vivo or of forming metastatic lesions upon injection into suitable hosts. Such results led many investigators to assume that cells which were transformed were also neoplastic. However, recent reports have demonstrated that some normal cells may be tumorigenic under 3 certain conditions [13J and, likewise, some cells designated transformed have been found incapable of forming tumors in nude mice [14,15J. In view of these newer findings, the previously accepted relationships between cellular phentotypic properties and the neoplastic state have been obscured and this review will consider these parameters separately. The terms normal and transformed will be used in this text to describe cells designated as such by the authors cited, but these terms must be qualified. It should be noted that the criteria or phenotypic traits which have been employed to distinguish normal and transformed have often varied with different investigators. Numerous reports have also shown that considerable variation may occur among the phenotypic characteristics exhibited by transformed cells I15,16, 17,18] and such cells may possess only a few or even only one of the properties ascribed to transformed cells. Keeping in mind the possible problems inherent in the use of normal versus transformed these terms are of value to this text as employed to the extent that they merely designate cells which may possess differences in some of the parameters of in vitro growth or behavior. A number of the properties commonly ascribed to transformed cells have been shown to occur transiently following brief proteolytic treatment of normal cells. This finding has led to speculation that some cells may release proteases which can affect cell growth and/or behavior 19,19]. It has also been shown that proteolytic enzymes (plasminogen activators) are in fact present in the culture fluids of some cells and it was suggested that 4 plasminigen activators and/or the plasmin generated by their action could affect cell growth or phenotypic properties (11,12,20,21]. This hypothesis and relevant information concerning plasminogen activator will be the focal point of the following review. I will first briefly describe the phenotypic characteristics employed to differentiate normal and transformed cells and will also provide evidence for or against the involvement of prot eases with these properties. Secondly, those cells which have been described and the possible involvement of the activator in cell growth or behavioral alteration will be discussed. Finally, the nature of plasminogen activator will be considered with emphasis on the physical and biochemical properties of those enzymes which have been examined. PHENOTYPIC PROPERTIES OF NORMAL AND TRANSFORMED CELLS INCLUDING EVIDENCE FOR OR AGAINST THE INVOLVEMENT OF PROTEASES IN THE EXPRESSION OF THESE CHARACTERISTICS Density Inhibition The term "density inhibition" refers to the cessation of the in vitro growth of normal cells which have been propagated under conditions where virutally all cells are in contact with others [1,2]. This state has been examined by a number of methods of which the determination of saturation density (i.e., maximum number of cells per unit surface area) has been most commonly reported [6,14, 17]. Typically, normal cells have been shown to be density inhibited and reach finite saturation densities while transformed 5 " . cells do not respond to t h ose cond 1t10ns or . s1gna 1" s presen t 1n 11' a dense population of cells and consequently continue to grow and increase in number (eyen though cell-cell contact has been achieved). The treatment of confluent cultures of mouse 3T3 cells 122J or chick embryo fibroblasts 123,24,25] with low concentrations (1-10 ugm/ml) of trypsin or other proteolytic enzymes has been shown to stimulate their growth or temporarily release them from density inhibition. Thus, it was suggested that proteases may be involved in the loss of density inhibition and mechanisms by which proteases could affect this cellular property were proposed [26J. However, there have been many other so-called "serum factors,tt cellular products, and nutrient alterations which have been reported to release cells from density inhibition. Molecules such as pancreatic ribonuclease, digitonin, and hyaluronidase would seem to perform unrelated functions, yet all of these molecules have been reported to produce this effect [27J. In view of these observations events other than the release and action of proteolytic enzymes could account for the loss of density inhibition and the physiological basis for this phenomenon has yet to be determined. Efficiencies of Plating Efficiencies of plating have been commonly expressed as the percentage of cells which continue to grow and form colonies after sparse numbers of cells have been seeded into culture vessels [3]. Generally, normal cells such as hamster fibroblasts possess plating efficiencies of less than 1% while transformed cells or those 6 infected with oncogenic viruses have much higher efficiencies of plating. It has been suggested that cells release substances necesary to the continued growth of normal cells and that such molecules are diluted to ineffective levels when only a few cells are present in a relatively large quantity of growth medium 128J. The evidence which suggests an involvement of proteolytic enzymes in the above phenomena is scant. McIlhinney and Hogan 129] have shown that the proteolytic inhibitors, pepstatin, leupeptin, and tosylphenylchloromethylketone (TPCK) reduced the efficiency of plating of hamster kidney cells. The interpretation of these results was obscured by the fact that leupeptin and pepstatin inhibit trypsin-like enzymes while TPCK inhibits chymotrypsin. However, the concentrations of TPCK used in these studies were also found to decrease the rates of cellular protein synthesis. A further problem with the interpretation of these results stems from McIlhinney and Hogan's observation that the addition of leupeptin and pepstatin also resulted in an apparent decrease in cellular growth rates, but from the information they provided it was not possible to ascertain whether the inhibitors actually slowed the growth of individual cells or were toxic and killed cells. Cell Growth in Semisolid Suspension Normal cells will not grow when suspended in a medium made semisolid by the addition of agar or methylcellulose while a high percentage of transformed cells are able to form colonies under these conditions I4,5,6J. Previous results have suggested that the ability 7 to grow in suspension (loss of anchorage dependence) may correlate with a change in the internal organization or structure of actin containing microfilaments or microfilament sheaths 130]. Since the addition of trypsin or other proteases has also been reported to result in a change in the internal organization of microfilaments or cytoskeletal elements 131] it could be supposed that proteases also affect the abilities of cells to grow in suspension. Though this has not been tested directly, Ossowski et al. 132] reported that serum which had been depleted of plasminogen was less effective in supporting the growth of SV40 virus infected rat cells when suspended in medium which contained agar. Lectin Agglutinability As reviewed by Rapin and Burger [33], various plant lectins such as concanavalin A (Con A) and wheat germ agglutinin ac" capable of agglutinating cells, and some transformed cells have been shown to be aggregated by concentrations of lectin which are much lower than the amounts of lectin required to agglutinate norma.;. t:ells [7,8,9]. Nicholson [34] and Sachs et al. [35J have sugge:.;ted that the increased agglutinability of SV40 infected 3T3 cells or of hamster cells derived from tumors was due to an increased mobility of molecules within the plasma membrane. This membrane mobility was determined by the propensity of multivalent lectin molecules which were wound to surface moieties to be sequestered or form patches on the plasma membrane. However, other investigators employing mouse fibroblast, 3T3, SV40-3T3, or lymphoma cells reported that an increased mobility of molecules which bind lectin was not always 8 accompanied by increased agglutinability [36]. Thus, it does not appear that differences in lectin agglutinability can be explained by the differences in the mobility of membrane molecules, and other mechanisms or biological processes which could control the phenotypic expression of lectin agglutinability remain to be found. The treatment of a variety of cells with low concentrations of trypsin or other proteases has been shown to increase their agglutinability with some lectins [7,9,37]. Borek et al. [8] have observed that mouse 3T3 and rat cells incubated in the presence of cycloheximide exhibited increased agglutinability with Con A or wheat germ agglutinin. However, the addition of both cycloheximide and a proteolytic inhibitor, tosyllysylchloromethylketone (TLCK), to growth media did not result in an increase in lectin agglutinability. These and further experiments were interpreted by the authors to suggest that proteases are involved in membrane glycoprotein turnover and that the pertubation of this turnover could effect the organization or integrity of the cell surface as reflected by lectin agglutinability [8J. Intercellular Cyclic Adenosine 3' 5' Monophosphate As reviewed by Pastan and Johnson [10] many authors have suggested that the intracellular levels of cyclic adenosine 3' 5' monophosphate (c-AMP) may influence vitro cell growth and phenotype. Properties which have been reported to be influenced by cellular c-AMP levels include cell growth rate [38], density inhibition [38], progression through the cell cycle [39], growth in soft agar [10], and agglutinability by Con A [40]. Some of these characteris- 9 tics are normally ascribed to transformed cells and in general such cells have been reported to possess lower levels of intracellular c-AMP [lQJ. Of possible significance to this review is the fact that intracellular levels of c-AMP have been shown to be decreased following the treatment of a variety of cells with trypsin I4l,42, 43]. Since adenyl cyclase and 3' 5' phosphodiesterase may be membrane bound I44,45J it is possible that the proteolytic perturbation of the cell membrane could also affect the activities of these enzymes. Yet, it is also possible that proteolytic action could affect the integrity of the membrane in such a way as to facilitate the leakage of the c-AMP into the culture fluid. Adhesion of Cells to the Surfaces on Which They Grow Work has also been completed which has shown that transformed cells may be more easily removed than normal cells from surfaces on which they are growing by either trypsin, chelating agents, or mechanical means I46,47,48,49]. It has been proposed that cells may adhere to a thin layer of serum molecules [50,51] or cell secreted molecules [52,53,54,55] both of which could coat the surfaces of Petri dishes. However, mechanisms by which cells control the tenacity of their attachment have not been clearly elucidated. It is a well-known fact that trypsin and other proteolytic enzymes can decrease the adherence of cells and it may be that proteases remove or cleave those molecules which function to "cement" cells to their substratum. However, Revel et al. [31] have ob- served that the treatment of cells with proteolytic enzymes affected 10 the endoskeleton of these cells and this along with other findings led them to conclude that the decreased adherence following proteolytic treatment was primarily due to the disruption or disorganization of cytoskeletal microfilament or microtubular elements. Differences in the Surface Molecules of Normal and Transformed Cells A number of investigators employing a variety of techniques have compared the external surface glycoproteins of some normal and transformed cells. Originally, Hynes et al. 156] employed lactoper- oxidase catalyzed surface iodination and identified a "large external transformation sensitive" (LETS) glycoprotein on the surface of hamster cells. Surface glycoproteins of similar size and which appear to be absent on transformed cells have also been detected by others employing different techniques 157,58,59,60,61,62,63,64]. The presence of these large external glycoproteins has been shown to be temperature-dependent when chick cells were infected with a temperature-sensitive mutant of Rous sarcoma virus 159,60,61,62] or when hamster cells were infected with temperature-sensitive polyoma virus [57]. Recent evidence has suggested that these large glyco- proteins may be cross-linked to form a network or skeleton on the exterior surfaces of chick cells [65] but the biological function of these molecules remains to be established. At this point in time, the presence or absence of LETS glycoprotein or similar molecules in the cell surface has simply provided further evidence suggesting that structural differences may exist in the exterior plasma membranes of normal and transformed cells. 11 LETS or similar glycoproteins have been shown to be removed from cell surfaces by proteases such as trypsin I19,56,66], plasmin I19,66J, collagenase, and a-chymotrypsin I66J. It is possible that the decreased amounts of LETS on some transformed cells may result from the action of proteases in culture fluids. Hynes et ale [67J have reported that the loss of LETS glycoprotein from the surfaces of chick cells transformed by temperature-sensitive mutants of Rous sarcoma virus was blocked when these cells were shifted to a temperature permissive for transformation and maintained in the presence of TLCK. However, the concentrations of TLCK which pro- duced this effect were also found to inhibit protein synthesis in these cells by 28% and a number of other proteolytic inhibitors did not prevent the loss of LETS glycoprotein following the appropriate temperature shift. Other Phenotypic Properties Besides those in vitro growth or behavioral characteristics previously described, increased glucose transport [68,69], continued nuclear division in the presence of cytochalasin B [70], decreased serum requirement for growth 16], abilities to migrate into a wound or denuded area of the cell monolayer [71], increased growth rate [72], and the loss of the actin filament bundles [30] are some additional properties ascribed to transformed cells. These charac- teristics have not been examined with a large number of cell types and for the most part, cellular mechanisms which are involved in their expression are not understood. For these reasons, these 12 properties will not be considered in more detail in this discussion. PLASMINOGEN ACTIVATOR IN CELL CULTURE FLUIDS Preface Observations of plasminogen activator production by cells cultured in vitro are not new to the field of cell biology. For example, Fisher [73] had described a fibrinolytic activity in sarcoma cell cultures in 1925. The aforementioned evidence and subsequent speculation concerning a possible involvement of proteases in producing or maintaining some of the phenotypic properties of transformed cells has renewed much interest and investigation into the role that plasminogen activators may play in the growth or expression of phenotypic properties of cultured cells. The term plasminogen activator has been used to describe those enzymes or molecules which proteolytically convert the proenzyme plasminogen to the active protease, plasmin. The use of this term should in no way be taken as an inference that the proteolytic actions of these activators are limited to the substrate plasmin and other yet unidentified proteins could be the de Facto substrates for these enzymes. It is also important to note that most authors have not distinguished between the occurrence of extracellular plasminogen activator, the release of activator, or its production. The terms, production and release, invoke two different possibilities to explain differences in amounts of extracellular activator. Since appropriate information usually was not provided, production and/or release will be used in combination in this section to prevent any 13 erroneous conclusions about the occurrence of extracellular activator. Is Increased Extracellular Activator A Property of Transformed Cells? Taylor et ala IIlJ have reported plasminogen activator activity in the culture fluids of high passage BHK-21/C13 cells. Those cells which produced and/or released increased amounts of the enzyme also possessed a number of properties commonly ascribed to transformed cells, namely, higher growth rates, increased plating efficiencies, increased lectin agglutinabilities and growth in soft agar. Work by Unkeless et ala [12] and Ossowski et ala [21] demonstrated that primary chick or mammalian fibroblasts which were infected and transformed with either DNA or RNA oncogenic viruses produced and/or released activators of plasminogen, while their normal non-infected counterparts or normal counterparts infected with cytopathic viruses did not. The occurrence of extracellular activator was found to be temperature-dependent when chick cells were infected with a temperature-sensitive mutant of Rous sarcoma virus [12]. These authors suggested that the occurrence of extracellular plasminogen activator was a qualitative and distinct difference observed with transformed cells. Similar findings were also reported by Goldberg [74] when he observed that chick fibroblasts transformed by RSV and Balb C/3T3 cells transformed by mouse sarcoma virus, RSV, or SV40 virus all produced and/or released plasminogen activator. However, in his studies primary chick fibroblasts were found to produce and/or release detectable, albeit, low levels of the enzyme 174]. The 14 detection of plasminogen activator in primary chick cell cultures indicates that the presence of the enzyme was not an exclusive property of transformed cells. Additional evidence suggesting a lack of correlation between plasminogen activator production and cell transformation has been reported by Mott et ala 175] when it was shown that normal mouse 3T3, human WI-38, and WI-26 cell culture fluids possessed more enzyme than their SV40 transformed counterparts. Equivalent amounts of the activator were measured in culture fluids from established and virus transformed Chinese hamster kidney cell lines. The production and/or release of plasminogen activator from Balb/C 3T3, SV40-transformed Balb/C 3T3 (clone SVT2) and transformation revertants (density inhibited cells) of SVT2 cells have also been examined (R. Snyder, unpublished results). Amounts of activity in cell cul- ture fluids were determined by use of an 125 1 fibrin released assay [12] or alternately cultured cells were directly examined for extracellular enzyme by overlaying pseudoclones (a defined number of cells on a small area of the Petri dish) with a mixture containing Eagle's Medium, plasminogen, denatured casein, and agarose. As determined by either method, the Balb/C 3T3 and SVT 2 revertant clones, D3 and E2 , released equivalent, if not greater amounts of activity than did transformed SVT2 cells. Though studies considered up to this point indicate that plasminogen activator production and/or release does not correlate with transformation the evidence reported is not conclusive. In this regard, the preceding reports with the exception of Taylor 15 et al. Ill] did not include phenotypic characterizations of those cells examined. More importantly, determinations of extracellular plasminogen activator were usually performed with cells which had never, or at least, not recently, been cloned and these cell populations could have consisted of individuals which differed greatly in their abilities to produce and/or release plasminogen activators. In further attempts to clarify a possible role for plasminogen activator in cell growth or behavioral modification, a number of investigators have attempted to correlate the presence of plasminogen activator with various phenotypic properties commonly ascribed to transformed cells. Pollack et al. 117] have studied cloned popula- tions of SV40 transformed rat cells and concluded that those cells which released the largest amounts of activator also possessed the greatest abilities to grow in a semi-solid medium containing methylcellulose. However, no differences in the growth rates or saturation densities of those cloned populations were found. In contrast to the observation of Pollack et al., the amounts of extracellular activator detected with either cloned populations of human fibrosarcoma cells [76] or cloned isolates of RSV transformed chick cells [77] have been shown not to correlate with their abilities to grow while suspended in semisolid media. In this regard, some cloned populations from both. cell types produced and/or released relatively low levels of activator yet grew very well in suspension. Further examinations of the cloned fibrosarcoma cells revealed that the saturation densities of those cell populations which possessed low levels of extracellular activator did not differ from those of cells 16 possessing high levels of extracellular enzyme [76J. Recently, Gallimore et al. 115J have studied cloned isolates of rat cells transformed by adenovirus and found that all of these transformed clones produced and/or released amounts of activator which were much greater than the amounts of extracellular enzyme possessed by cultures of non-infected rat cells. However, the transformed cells varied considerably with respect to their growth in methylcellulose suspension, saturation densities in 0.5% versus 5% calf serum, and growth rates in 0.5% versus 5% serum. The aforementioned results strongly suggest that extracellular plasminogen activator may not be involved in the expression of phenotypic properties commonly ascribed to transformed cells. However, all in vitro cell cultures may possess at least small amounts of plasminogen activator and it is possible that some yet unidentified conditions or events along with the action of plasminogen activator may be necessary for the expression of some of the phenotypic alterations which have been described for transformed cells. Is Increased Extracellular Activator a Property of Neoplastic Cells? Since the original observation by Fisher [73], enzymes capable of converting plasminogen to plasmin have been measured in the culture fluids of a variety of cells derived from tumor tissue. Supernates from cultures of human sarcoma, mesothelioma, malignant melanoma [78J and neuroblastoma cells 179] all contained significant amounts of the activator. However, results from studies which have 17 investigated cloned populations of cells suggest that plasminogen activator production is not a general phenomena common to all cells derived from tumor tissue. In one case, only three out of six clones of cells derived from rat ovarian tumors were reported to possess detectable amounts of extracellular activator 180J. Like- wise, as previously described, a variability in the amounts of extracellular activator among clones of cells isolated from a human fibrosarcoma cell line has also been reported l76]. In a different approach researchers have attempted to relate the amounts of plasminogen activator produced and/or released by cultured cells with their abilities to form tumors upon injection into suitable hosts. Jones et al. l76] have shown that all cloned isolates of human fibrosarcoma cells which possessed large amounts of extracellular activator were also capable of proliferating in vivo and forming tumors upon injection into immunosuppressed hamsters. However, the cloned isolates which possessed relatively low levels of extracellular activator also formed tumors under these conditions. It is important to note that the tumors formed by the fibrosarcoma cells grew rapidly, eventually killed their hosts, possessed human karyotypes and were determined to be fibrosarcomas by histological examinations. Work reported with cloned rat cells which had been transformed with adenovirus has shown that some cells which release increased amounts of activator are incapable of forming metastasizing tumors after injection into syngeneic rats or nude mice lIS]. In other experiments, Laug et al. [81] have demonstrated that HeLa cell cultures contain small amounts of 18 activator yet these cells were shown to be capable of forming tumors in immunosuppressed hamsters. The results of these tttumori- genicity" experiments suggest that extracellular plasminogen activator may not be an absolute correlate to the neoplastic state, yet such studies are difficult to compare and evaluate since different test animals were employed for tumorigenicity determinations and the criteria upon which a growth was determined to be a tumor differed. With yet other experimental approaches, authors have provided evidence supporting the hypotheses that plasminogen activator may be involved in neoplasia. Jones et al. £76] observed that cells ob- tained from fibrosarcoma explants released large amounts of activator even though the cloned cells which had been introduced to form these tumors had previously possessed small amounts of extracellular activity. These authors concluded that the in vivo growth of these cells induced the release of increased amounts of plasminogen activator. Their conclusion was not unreasonable in that the ex- planted cells which released large amounts of activator were cloned and found to be of human origin and not of the host animal, i.e., hamsters. Secondly, experiments were performed which tended to rule out the possibility that tumor growth selected for cells which released large amounts of activator and were already present in the inoculum. Mak et al. £82J have shown that dihydrotestosterone specifically induced the release of activator from an androgen dependent mammary carcinoma (Sionogi SC-115) cell line. Christman et al. I83] have also demonstrated that a loss of tumorigenic 19 potential occurred concomitant with a decrease in activator release when a Brdu (5-bromo-2-deoxyuridine) sensitive clone of IDouse melanoma cells was cultured in the presence of Brdu. This decrease in tumorigenicity could not he explained by a loss of viability in the presence of Brdu and the alterations due to Brdu could be reversed upon its removal from the medium. Christman et al. I83] have not offered any suggestions as to how Brdu specifically affects the production of plasminogen activator or the ability of these cells to form tumors in host animals. A General Phenomena of Cells In Vitro As mentioned previously, a variety of normal cells which were neither virus transformed nor obtained from tumor tissues have also exhibited release of plasminogen activators and this list can be extended. Primary cultures of human embryonic kidney [84], monkey kidney [85], rat embryo I17] and chick fibroblasts [74] all have been reported to release and/or produce detectable amounts of the protease. Secondary cultures of human diploid kidney, heteroploid kid- ney, and lung cells have also been shown to possess supernate activity [84]. Since the presence of this protease in cell culture fluid may be a ubiquitous phenomena it is appropriate to discuss briefly some possibilities for its biological function. There are changes in the cell membrane which correspond to the cell's phase in the growth cycle and some of these changes may be a result of a proteolytic action. Balb C/3T3 cells have been found to become highly agglutinahle with Con A or wheat germ agglutinin 20 during mitosis f37,86J. The 250,000 dalton LETS glycoprotein, as described by Hynes and Humphryes 156J, was not present in the membrane of normal cells' which were treated with colchicine and arrested in mitosis f191. Cells may round up and become less adherent during mitosis 187J. Lower levels of c-AMP have also been shown to occur with 3T3 188J or Chinese hamster ovary (CRO) cells f89] during mitosis and into early Gl phase and this may also reflect proteolytic surface alterations. If plasminogen activator was responsible for some of these cell cycle dependent surface changes it would be reasonable to suppose that the protease should only act during specific periods in the cell cycle. However, this is probably not the case since the activator appears to be continuously present in the fluids of many unsynchronized cell cultures. A second means by which the action of plasminogen activator could be involved in transient cell surface alteration can also be proposed. It is possible that the activator continuously reacts with surface material and that the cell could regulate membrane alterations by controlling the synthesis and replacement of those surface molecules which serve as substrate for the protease. Whether plasminogen activator can act directly on the cell surface is not known. Attempts to detect loss of cell surface material following the addition of crude supernates which contained plasminogen activator have been unsuccessful I66,83J. If the activa- tor does function on a cell surface component, it is possible that such substrate molecules would competitively inhibit the activity of 21 plasminogen activator. No inhibition of the protease was observed when supernates of RSV transformed chick cells were mixed with homogenates of primary chick cells and assayed oy l25I_fibrin release I90J. When nigh passage BHK-2l/C13 cells (which possessed large amounts of extracellular activator) were co-cultivated with low passage cells (which possessed small amounts of extracellular activator) very little plasminogen activator could be measured Ill]. This was interpreted to suggest that the enzyme was binding to some substrate on the surfaces of low plasminogen activator producing cells and that this interaction resulted in an apparent decrease in measurable enzyme. Plasmin and Its Possible Effects Since plasminogen activators convert plasminogen to the active enzyme, plasmin, the possible action of plasmin in cell cultures should also be considered. The chromatography of serum on lysine- substituted Sephorose has been employed to free serum of plasminogen [91] and these sera have been utilized in attempts to identify those growth and behavioral characteristics which may require the action of plasmin [32,92]. The ability of SV40 transformed hamster cells to grow in soft agar was reported to be decreased when serum depleted of plasminogen was added to the media I32]. Sera which were more efficient at promoting fibrinolysis with added cellular activator also were found to increase the efficiencies of colony formation when added to SV4Q transformed rat cells suspended in Methocel I17J. These results strongly suggest that the action of 22 plasmin may be involved in the expression of this characteristic. However, as previously mentioned, some cloned RSV infected chick [77] and human fibrosarcoma cells {76] which exhibited very little extracellular activator activity grew relatively pended in semisolid media. wel~ while sus- It is possible that the low levels of activator released by these cells could activate a quantity of plasmin which would be sufficient to produce some required alteration. Studies with SV40 transformed hamster cells have shown no differences in growth rates in media supplemented with plasminogen depleted sera {32]. Similarly, no differences in growth rates, saturation densities or glucose transport were observed when RSV transformed chick cells were grown in plasminogen depleted sera {92]. It has also been reported that the morphological properties of SV40 transformed hamster fibroblasts were dependent upon the sera in which these cells were grown {93]. These morphological alterations, which consisted of an initial rounding of cells and their subsequent detachment were promoted by the addition of sera which would produce relatively high levels of fibrinolysis when mixed with activator from these cells (93J. The morphological changes which were noted for SV40-transformed hamster cells were prevented when sera depleted of plasminogen were employed. Likewise, a decrease in adherence which has been reported for RSV-CEF cells was also reversed by the addition of a high concentration of soy bean trypsin inhibitor (2.5 mg/ml) and other inhibitors of plasmin 194J. It is possible that the action of plasmin may alter cell morphology and/or adherence. 23 THE NATURE OF PLASMINOGEN ACTIVATOR A Group of Enzymes With a Similar Action Primary and established cultures of human embryonic lung and kidney cells produce at least two antigenically distinguishable activators [84]. Some of the activity released by these cells was neutralized by antibody to human urokinase (a protease originally isolated from human urine). The percent of activity which was neutralized with antibody to urokinase increased when kidney cell cultures were obtained from progressively older embryos [84]. These results could suggest that cells from the same organ and possibly of similar type may produce more than one activator. However, the cells employed in these studies were not cloned and it is likely that different cell types may have existed in the populations and that these different cell types each produced a different activator. The activator from a human melanoma cell line has been examined and the activity was found to correspond to two proteins which were reported to have molecular weights of 60,000 and 48,000 daltons as determined by SDS polyacrylamide gel electrophoresis 178]. However, it was not clear from the information present in this report whether these cells were a heterogeneous population or had been recently cloned. Aside from this consideration, it is also possible that the activators from human melanoma cells may have been antigenically related to urokinase and/or may have been enzymatically active peptides resulting from a partial degradation of urokinase. In this regard, the molecular weight of urokinase as purified from human 24 urine has been reported to be 54,000 daltons and treatment of this enzyme with trypsin or subtilisin has resulted in an active fragment of 36,000 daltons [95]. Plasminogen activators from cells other than those of human origin have also been characterized and multiple molecular forms of the enzyme have been found from cells of the same species. The activity from rat ovarian tumor cells has been reported to correspond to two species of large molecular weight (615,000 and 175,000 daltons) as determined by sucrose density gradient centriguation [80]. More recently, Chrisman et ale reported their results from immunological studies which employed antibody directed against a plasminogen activator released by SV40 transformed hamster cells [96]. The anti- body was found to inhibit the enzyme released by primary or established hamster lung cells, but had no effect when mixed with activator from either established heteroploid or primary hamster kidney cells. The antibody was also without effect when mixed with activators released by a number of other hamster cell lines which had been transformed by either RSV or chemical agents. It is not known whether the activators produced by cells from different animal species are similar or identical in structure. Christman et ale [96] have shown that antibody directed against a hamster plasminogen activator did not inhibit the activation of plasminogen when it was mixed with activators from murine or human cells. As noted above, these studies are not conclusive in that the antibody employed only reacted with one of the activators released by hamster cells. It is possible that there are a number of different 25 groups or types of plasminogen activators some or all of which are produced by cells from a variety of species. It is also important to emphasize that since there may be a large number of prot eases produced by cells which possess in common, the capabilities to activate plasminogen, in the future authors should attempt to characterize or identify those particular enzymes which they are studying. Physical and Biochemical Properties The activator released by RSV infected chick embryo fibroblasts has been partially purified and characterized [90J 'and was reported to correspond to a single protein of 39,000 daltons as determined by SDS gel electrophoresis. The enzyme was found to be inhibited by p-tosyl-L-arginine methyl ester (TAME) and by high concentrations of L-arginine. The activator was also irreversibly inhibited by diisopropylflurophosphate (OFP) [90] and this was considered to be sufficient evidence to warrant its classification as a serine protease [97J. The use of OFP has also provided evidence that the acti- vators produced and/or released by transformed hamster cells (50,000 daltons) [98] and human melanoma cells (60,000 and 48,000 daltons) [78] are also serine proteases. The enzymatic specificities of some activators have also been examined. The proteolytic conversion of plasminogen to plasmin by the chick cell activator was assessed by SOS gel electrophoresis of the reaction mixture and the actions of the chick cell activator resulted in cleavage products which were indistinguishable from those resulting from the action of urokinase or streptokinase on plasminogen 26 [90]. Since urokinase and streptokinase (an activator produced by streptococcus) have been reported to hydrolyze an arginyl-valine peptide bond in the proenzyme [99J and the chick cell activator is inhibited by TAME and L-arginine it is possible that both of these enzymes have similar specificities and split plasminogen at the same position [90J. Since plasminogen contains a number of arginyl residues and the activators show a strict preference for only one of these, it is likely that the molecular conformation near the arginylvaline sequence is extremely important for substrate recognition by the activator. Christman et ale [83J have also reported that the activator released by hamster cells is specific for plasminogen and does not activate other zymogens such as trypsinogen, chymotrypsinogen or pepsinogen. It should also be pointed out that protein substrates other than plasminogen have not been reported for any of the activators described in this review. Intracellular Activator Plasminogen activator has been detected in Triton X 100 extracts of RSV transformed chick cells while normal (uninfected) cells did not contain measurable amounts of this intracellular activity [90]. These findings were interpreted by these authors to suggest that the increased amounts of extracellular enzyme present in cultures of RSV infected chick cells was the result of increased activator production rather than an increase in its release. However, it is also possible that the differences in intracellular activator observed with chick 27 and RSV-chick cells could be accounted for by differences in enzyme turnover in these two cell types. Jones et a1. 176] have found that the amounts of Triton X 100 extractable or intracellular levels of plasminogen activator did not vary among cloned isolates of human fibrosarcoma cells even though there was considerable variation among these clones in the amounts of extracellular activator. These findings may suggest that the differ- ences in extracellular activator observed among some cells could be accounted for by differences in its release and not by differences in its production. Unkeless et al. {90] reported that the activator produced by transformed chick cells was associated with crude lysosomal or microsomal fractions. However, a previous communication by this group had concluded that the occurrence of extracellular activator did not coincide with any increase in lysosomal enzymes in the cell culture fluid lI2]. If both observations were correct this paradox would be diffi- cult to reconcile. But the methods employed to determine the intra- cellular localization were not very precise and consisted of a few steps of differential centrifugation. More recently, Quigley [100] has reported that the cellular activator from RSV transformed chick fibroblasts was associated with plasma membrane or plasma membranelike elements. In his communication he demonstrated that an activator from transformed chick cells co-purified with material which appeared to be membranous in electron micrographs and with molecules commonly reported to be associated with the plasma membrane. Christman et ale [83] have also investigated the localization of the activator in SV40- 28 transformed hamster cells. The enzymatic activity could be localized in fractions enriched for plasma membranes following equilibrium sedimentation of plasma membrane ghosts in sucrose gradients. These authors also noted that plasminogen was rapidly activated when added to previously washed monolayers of transformed hamster cells, and this was interpreted as evidence for a cell surface localization of the enzyme. REFERENCES 1. Stoker, MCP & Rubin, H. Nature 215 (1967) 171. 2. Nilausen, K & Green, H, Exp cell res 40 (1965) 166. 3. Stoker, MCP & MacPherson, I, Virology 14 (1961) 359. 4. MacPherson, I & Montagnier, L, Virology 23 (1964) 291. 5. Montagnier, L, Growth control in cell culture (eds GEW Wolstenholme & J Knight) p. 33 Churchill Livingstone, Edinburgh and London (1971). 6. Risser, R & Pollack, R, Virology 59 (1974) 474. 7. Burger, MM, Fed proc 32 (1973) 91. 8. Borek, C, Grob, M, 9. Burger, 11M, Proc nat acad sci US 62 (1969) 994. & Burger, l1M, Exp cell res 77 (1973) 207. Johnson, GS, Adv can res 19 (1974) 303. 10. Pastan, I 11. Taylor, JC, PhD Dissertation, University of Utah (1973). 12. Unkeless, JC, Tobia, A, Ossowski, L, Quigley, JP, Rifkin, DB, & Reich, E, J exp med 137 (1973) 85. 13. Boone, CW, Science (1975) 68. 14. Stiles, CD, Desmond, W, Sato, G, & Saler, H, Proc nat acad sci US 72 (1975) 4971. 15. Gallimore, PH, McDougall, JK, & Chen, LB, Cell 10 (1977) 669. 16. Freedman, VH & Shin, S, Cell 3 (1974) 355. 17. Pollack, R, Risser, R, Conlon, S, & Rifkin, P, Proc nat acad sci US 71 (1974) 4792. 18. Ukena, TE, Goldman, E, Benjamin, TL, & Karnovsky, 7 (1976) 213. 19. Hynes, RO, CellI (1974) 147. & ~U, Cell 30 20. Taylor, JC, Hill, DW, & Rogolsky, M, Exp cell res 73 (1972) 422. 21. Ossowski, L, Unkeless, JC, Tobia, A, Quigley, JP, Rifkin, DB, & Reich, E, J exp med 137 (1973) 112. 22. Burger, MM, Growth control in cell cultures (eds GEW Wolstenholme & J Knight) p. 45 Churchill Livingstone, Edinburgh and London (1971). 23. Sefton, BM & Rubin, H, Nature 227 (1970) 843. 24. Cunningham, DD & Ho, TS, Proteases and biological control (eds E Reich, D Rifkin, & E Shaw) p. 795 Cold Spring Harbor Lab, Cold Spring Harbor, New York (1975). 25. Blumberg, PM & Robbins, PW, Cell 6 (1975) 137. 26. Rubin, H, Growth in cell cultures (eds GEW Wolstenholme & J Knight) p. 127 Churchill Livingstone, Edinburgh and London (1971). 27. Vasielev, JM, Gelfand, 1M, Guelstein, VI, & Fetisova, EK, J cell phys 75 (1970) 305. 28. MacPherson, I, Adv cancer res 13 (1970) 169. 29. McIlhinney, A & Hogan, 30. Pollack, R, Osborn, M, & Weber, K, Proc nat acad sci US 72 (1975) 994. 31. Revel, JP, Hoch, P, & Ho, D, Exp cell res 84 (1974) 207. 32. Ossowski, L, Quigley, JP, Kellerman, GM, & Reich, E, J exp med 138 (1973) 1056. 33. Rapin, AMC & Burger, MM, Adv can res 20 (1974) 1. 34. Nicolson, G, Nature new bioI 243 (1973) 218. 35. Sachs, L, Inbar, M, & Shinitzky, M, Control of proliferation in animal cells (eds B Clarksen & R Beserga) p. 283 Cold Spring Harbor Lab, Cold Spring Harbor, New York (1974). 36. Raff, MC, de Petris, S, & Mallucci, L, Control of proliferation in animal cells (eds B Clarksen & R Beserga) p. 271 Cold Spring Harbor Lab, Cold Spring Harbor, New York (1974). 37. Shoman, J & Sachs, L, Control of proliferation in animal cells (eds B Clarksen & R Beserga) p. 297 Cold Spring Harbor Lab, Cold Spring Harbor, New York (1974). BL~, Bioch bioph res com 60 (1974) 348. 31 38. Otten, J, Johnson, GS, & Pastan, I, Bioch bioph res com 44 (1971) 1192. 39. Bombik, BM & Burger, MM, Exp cell res 80 (1973) 88. 40. Willingham, M & Pastan, I, J cell bioI 63 (1974) 288. 41. Sheppard, JR, Nature new bioI 236 (1972) 14. 42. Otten, J, Johnson, GS, & Pastan, I, J bioI chern 247 (1972) 7082. 43. Burger, MM, Bombik, BM, Brekenridge, B, & Sheppard, J, Nature new bioI 239 (1972) 161. 44. Russel, T & Pastan, I, J bioI chern 248 (1973) 5835. 45. Roblin, R. Chou, IN, & Black, PH, Adv can res 22 (1975) 203. 46. Culp, LA & Black, PH, Biochemistry 11 (1972) 2161. 47. Shields, R & Pollock, K, Cell 3 (1974) 31. 48. Gail, I1H & Boone, CW, Exp cell res 70 (1972) 33. 49. Weber, MJ, Hale, AH, & Roll, PE, Proteases in biological control (eds E Reich, D Rifkin, & E Shaw) p. 915 Cold Spring Harbor Labs, Cold Spring Harbor, New York (1975). 50. Yaoi, Y & Kanaski, T, Nature 237 (1972) 238. 51. Revel, JP & Wolken, K, Exp cell res 78 (1973) 1. 52. Poste, G, Greenham, LW, Mallucci, L, Reeve, P, & Alexander, DJ. Exp cell res 78 (1973) 303. 53. Culp, LA, J cell bioI 63 (1974) 71. 54. Terry, AH & Culp, LA, Biochemistry 13 (1974) 414. 55. Roblin, R, Albert, SO, Gelb, NA, & Black, PH, Biochemistry 14 (1975) 347. 56. Hynes, RO & Humphreyes, KC, J cell bioI 62 (1972) 438. 57. Gahmberg, CG & Hakomori, S, Proc nat acad sci US 70 (1973) 3329. 58. Critchley, DR, Cell 3 (1974) 121. 59. Wikus, CG, Bronton, PE, & Robbins, PW, Control of proliferation in animal cells (eds B Clarksen & R Baserga) p. 541 Cold Spring Harbor Lab, Cold Spring Harbor (1974). 32 60. Stone, KR, Smith, RE, & Joklik, WK, Virology 58 (1974) 86. 61. Robbins, PW, Wickus, GG, Branton, PE, Gaffney, BJ, Hershberg, CB, Fuchs, P, & Blumberg, PM, Cold Spring Harbor Symp on quant bioI 39 (1974) 1173. 62. Blumberg, PM & Robbins, PW, Proteases in biological control (eds E Reich, D Rifkin, & E Shaw) p. 945 Cold Spring Harbor Labs, Cold Spring Harbor, New York (1975). 63. Vaheri, A & Ruoslahti, E, lnt j cancer 13 (1974) 579. 64. Hogg, NM, Proc nat acad sci 71 (1974) 489. 65. Heski-Oja, J, Mosher, DF, & Vaheri, A, Cell 9 (1976) 29. 66. Blumberg, PM & Robbins, PW, Cell 6 (1975) 137. 67. Hynes, RO, Wyke, JA, Bye, 3M, Humpheyes, KC, & Pearlstein, ES, Proteases in biological control (eds E Reich, D Rifkin, & E Shaw), p. 931 Cold Spring Harbor Labs, Cold Spring Harbor, New York (1975). 68. Hatanaka, M. Bioch bioph acta 355 (1974) 77. 69. Plageman, PGW & Richey, DO, Bioch bioph acta 344 (1974) 263. 70. O'Neill, FJ, J nat cancer ins 52 (1974) 653. 71. Dulbecco, R, Nature 227 (1970) 802. 72. Diamondopoulos, GT & Enders, JF, Proc nat acad sci US 54 (1965) 1092. 73. Fischer, A, Arch entwicklungsmech org 104 (1925) 210. 74. Goldberg, AR, Cell 2 (1974) 95. 75. Mott, DM, Fabisch, PH, Brahma, P. Sorof, S, & Sorof, S, Bioch bioph res com 61 (1974) 621. 76. Jones, PA, Laug, WE, & Benedict, WF, Cell 6 (1975) 245. 77. Wolfe, BA & Goldberg, AR, Proc nat acad sci US 73 (1976) 3613. 78. Rifkin, DB, Loeb, J, Moore, G, & Reich, E, J exp rned 139 (1974) 1317. 79. Wasksrnan, J & Biedler, JL, Exp cell res 86 (1974) 264. 80. Yunis, AA, Schultz, DR, & Sato, GH, Bioch bioph res com 52 (1973) 1003. 33 81. Laug, W, Jones, P, & Benedict, W, J nat cancer inst 54 (1975) 173. 82. Mak, TW, Rutledge, G, & Sutherland, DJA, Cell 7 (1976) 223. 83. Christman, JK, Acs, G, Si1agi, S, & Silverstein, S, Proteases in biological control (eds E Reich, D Rifkin, & E Shaw) p. 827 Cold Spring Harbor Labs, Cold Spring Harbor, New York (1975). 84. Bernik, M & Kwaan, H, J c1in invest 48 (1969) 1740. 85. Bernett, EV, Proc soc exp bioI med 102 (1959) 308. 86. Fox, T, Sheppard, JR, & Burger, MM, Proc nat acad sci US 68 (1971) 244. 87. Terasirna, T & To1mach, LJ, Exp cell res 30 (1963) 344. 88. Burger, MM, Bombik, BM, Breckenridge, BM, & Sheppard, JR, Nature new bioI 239 (1972) 161. 89. Sheppard, JR & Prescott, DM, Exp cell res 75 (1972) 293. 90. Unke1ess, J, Dano, K, Kellerman, G, & Reich, E, J bioI chem 249 (1974) 4295. 91. Deutsch, DG & Mertz, ET, Science 170 (1970) 1095. 92. Chen, LB & Buchanan, JM, Proc nat acad sci US 72 (1975) 1132. 93. Ossowski, L, Quigley, JP, & Reich, E, J bioI chern 249 (1974) 4312. 94. Weber, MJ, Cel15 (1975) 253. 95. Lesuk, A, Terminie11e, L, Traner, JH, & Groff, J, Fed proc 26 (1967) 647. 96. Christman, JK, Silverstein, SC, & Acs, G, J exp med 142 (1975) 419. 97. Hartley, BS, Ann rev bioch 29 (1960) 45. 98. Christman, JK & Acs, G, Bioph bioch acta 340 (1974) 339. 99. Summaria, L, Hsieh, B, & Robbins, KC, J bioI chern 242 (1967) 4279. 100. Quigley, JP, J cell bioI 71 (1976) 472. CHAPTER II ISOLATION AND CHARACTERIZATION OF CLONED BHK CELLS WHICH RELEASE DIFFERENT AMOUNTS OF PLASMINOGEN ACTIVATOR by Robert W. Snyder and Douglas W. Hill 35 SUMMARY A large number of cell clones were isolated which produced different amounts of extracellular plasminogen activator. These clones were identified and selected by use of an overlay medium containing casein and plasminogen. All cells obtained by this method arose from a high passage cell line (BHK-2l/N) which had been originally cloned from BKH-2l/C13 cells. These cloned isolates from BHK-2l/N were characterized with respect to their growth rates and efficiencies of plating in soft agar and on plastic surfaces. No correlations were found between any of these growth properties and the amounts of extracellular plasminogen activator as detected in serum-free cell culture media. IHTRODUCTION Many previous experimental findings have implicated the involvement of proteolytic enzymes in the alteration of cell growth or morphological properties. These have been reviewed by Hynes [1] and more recently by Roblin, Chou, and Black [2]. In light of this, the presence of a plasminogen activator (PA) in cell culture fluids had received much attention since activators of plasminogen have been shown to be proteases [3] and the action of the activator could generate plasmin from the zymogen which is present in serum. An early report suggested that in the in vitro production and/or release of the activator was a unique property of transformed cells [4]. It was proposed that the action of the enzyme may be responsible for some cell alterations necessary for the aberrant morphological or growth properties of transformed cells [1]. A number of investigators 36 have attempted to relate the ability to produce and/or release PA with various cell growth or physical properties. Taylor [5] has detected activators of plasminogen in serum-free culture fluids from high passage BHK-21-CI3 cells. Those cells which released increased amounts of the enzyme also possessed increased growth rates and plating efficiencies, the abilities to grow in soft agar suspension, and higher levels of lectin agglutinability. A correlation between high levels of extracellular activator and the abilities of cells to grow while suspended in methyl cellulose has also been described for clonal isolates of SV40-transformed rat embryo fibroblasts [6]. In other studies, Ossowski et ale [7], demonstrated a dependence upon the presence of plasminogen for the growth of SV40-transformed hamster cells in soft agar suspension. In this study, a large number of cell clones were isolated which released different amounts of PA. These cells were characterized in an attempt to correlate increased extracellular PA with a number of properties normally ascribed to transformed cells. The clones of cells were originally isolated by employing overlays containing plasminogen and casein [8,9]. This method allows detection of PA production through visualization of the caseinolysis which results from plasmin's action on denatured casein. The cell line utilized in these studies was originally cloned from BHK-21/CI3 cells. The BHK-21/N cells have previously been shown to possess high growth rates and high efficiencies of plating on plastic surfaces or in soft agar suspension. Furthermore, these cells can also be agglutin- ated by relatively low levels of concanavalin A [5]. These cells 37 have been extensively "passaged" and were designated BHK-2l/N to distinguish them from BHK-2l/C13 cells. MATERIALS AND METHODS The BHK-2l/N cells employed in these studies have been previously described by Taylor IS]. These cells are high passage cells and were originally cloned from BHK-2l/C13. All cells were propagated in Eagles Minimal Essential Medium (MEM) (GIBCO) supplemented with either 5% or 10% calf serum (CS) (Flow Labs). Cells were routinely transferred one day after confluency. Isolation of Hpa and Lpa Cells and Frequencies of Activator Positive Colonies Cells were trypsinized, suspended in cold MEM + 10% CS, counted by aid of a hemocytometer, and the appropriate dilutions were made in culture media. Approximately 200 cells were added per 60 mm Falcon plastic Petri dish in 5 ml of medium. Following seven days of incuba- tion at 37°C the macroscopically visible colonies were washed three times with calcium and magnesium deficient phosphate buffered saline (PD). The colonies were then overlayed with 3 ml of a mixture containing MEM, 0.4% w/v of agarose, 0.5% denatured skim milk and approximately 0.01 mg/ml of plasminogen. The human plasminogen had been purified by the methods of Deutsch and Mertz [10]. Following 24 hours at 37°C the plates were examined and either caseinolytic or noncaseinolytic colonies were picked from the plates. This procedure was repeated at least once for all those Hpa and Lpa isolates used. The same protocol was also followed to determine the frequencies of 38 activator positive colonies. Following a 24-hour incubation with the plasminogen casein overlays the number of colonies demonstrating caseinolysis was determined for each plate. The overlays were then removed and cell colonies were stained and counted. Efficiencies of Plating on Plastic Surfaces Five ml of MEM + 10% CS which contained 200 cells were added per 60 rom Falcon dish. The cultures were placed in a 37°C incubator and great care was taken to avoid disturbing them. Seven days later the medium was removed and cell colonies stained and counted. Efficiencies of Plating in Soft Agar The procedure was essential that of Macpherson and Montagnier [11]. The soft agar (0.3% agar in MEM + 10% CS) contained 200 cells per 60 rom Falcon Petri dish. Following eight days at 37°C the number of macroscopically visible colonies was determined for each culture plate. Determination of Plasminogen Activator in Cell Culture Fluids Confluent monolayers of cells were washed three times with PD. A 4.5 ml volume of MEM without calf serum was added per 25 cm 2 culture flask and removed after 18-22 hours of incubation at 37°C. Cell debris was removed by low-speed centrifugation and the supernates frozen. The activator was quantitated by use of an l25I-fibrin release assay or by fibrin clot lysis. The l25I-fibrin release assay was similar to that described by Unkeless et ale I4J with the following modifications: (1) the l25I-fibrin was clotted by addition 39 of 0.1 ml of 1 u/ml bovine thrombin (Park-Davis) in the presence of phosphate buffered saline for 2 hours at 37°C; (2) the assays were carried out in Costar multiwelled dishes (16 rom) with 0.1 ml of plasminogen (approximately 0.1 mg/ml), 0.1 ml of 0.1 M Tris (pH 8.1), and 0.1 ml of activator preparation in each well. lysis assay was essentially that of Taylor modifications: The fibrin clot I5J with the following (1) plasminogen was added to 0.4% human fibrinogen for a final concentration of approximately 0.01 mg/ml; (2) the cell supernates or dilutions thereof in PD were added to equal volumes of fibrinogen, clotted, and incubated at 37°C. The greatest dilution which yielded clot lysis was then determined after 24 hours of incubat ion. RESULTS Isolation of Cells Producing Different Amounts of Activator When colonies of BHK-2l/N cells were overlayed with casein and plasminogen, both activator producing (positive) and non-producing (negative) colonies were detected. The differences observed for plasminogen activator production did not appear to correlate with colony size. However, only large negative colonies or small positive colonies were selected from these plates. Cells chosen from positive colonies were designated Hpa (high plasminogen activator) while those from negative colonies were designated Lpa (low plasminogen act ivator). As shown in table 1, the Hpa isolates contained a high fre- quency of cells which were capable of forming activator positive Table 1. BHK-21/N, Hpa, and Lpa cells: frequencies of activator positive colonies and relative amounts of activator in serum-free cell culture fluids. Activator positive per total colony number Frequency of activator positives colonies Reciprocal of greatest dilution demonstrating Cell isolate Serial passage level BHK-21/N 39 210/938 0.224 1 ND Hpa-1 16 1476/1476 1.000 32 ND Hpa-1 20 ND ND 16 78 Hpa-1 27 1017/1024 0.993 16 72 Hpa-2 4 1056/1056 1.000 32 ND Hpa-4 9 1327/1327 1.000 32 ND Hpa-6 1 480/486 0.988 32 ND Lpa-1 1 192/1206 0.159 No Lysis ND Lpa-2 1 185/917 0.202 Lpa-4 11 10/1440 0.007 No Lysis 0 Lpa-5 11 51/1375 0.037 No Lysis 0 Lpa-6 8 11/1085 0.010 No Lysis ND fibrino1~sis 1 % of 125I-fibrin released ND ..p.. 0 Table 1 (continued) r'?'" a. NO signifies that the data were not obtained. b. Passage levels for BHK-2l/N represents the number of serial passes since the initiation of these studies. c. Passage levels for Hpa and Lpa isolates represent the number of serial passes since the time of their isolation. ,"'. J:'j--I 42 colonies. Likewise, the Lpa populations were enriched with cells which yielded activator negative colonies. However, populations of cells which would consistently yield only positive or negative colonies during extended serial passage were never isolated. The amounts of plasminogen activator present in serum-free culture fluids from either Hpa or Lpa cells are also shown in table 1. The supernates of Hpa cell cultures were found to possess relatively high levels of the enzyme, while many of the Lpa cells tested did not possess detectable amounts of the activator in their serum-free culture media. The results presented in table 1 establish a correlation between the observed frequency of positive colonies in a casein plasminogen overlay assay and the amount of enzyme present in serum-free cell culture fluids. These results also validate the use of the casein plasminogen overlay technique as an efficient method for isolating populations of cells which produce and/or release different amounts of PA. It is presumed that the amounts of plasminogen activator detected in supernatant fluids also reflect the amounts of activator present under normal growth conditions. Calf serum is known to contain inhibitors of plasmin and this has prevented the detection of plasminogen activation in the normal growth media [5]. It is unlikely that the low levels of PA as found in Lpa cell culture fluids were entirely due to the presence of a cell associated or soluble inhibitors. observations. This conclusion is supported by a number of A short incubation (10 min.) of Hpa culture fluid with monolayers of Lpa cells did not result in a loss of activator as 43 measured by l25 I -fibrin release. In another experimental approach, equal numbers of Hpa and Lpa cells were mixed and the amounts of activator present in serum-free culture fluids were determined. In the experiment reported in table 2, the amount of activator was determined after one serial passage of the cell mixture. The fre- quencies of cells yielding activator positive colonies represent those of various cell populations at the time at which serum-free supernates were prepared. The fluids from the co-cultivated cells contained detectable amounts of the enzyme and the frequency of activator positive colonies which arose from the population suggests that Lpa cells were present. The decrease of activator in mixed cell culture fluid correlates with that expected from the decrease in positive colonies. Table 2. Frequencies of activator positive colonies and levels of activator in serum-free supernates of co-cultivated cells. Cell isolate Frequency of activator positive colonies Hpa-6 0.913 Lpa-6 0.009 Hpa-6 x Lpa-6 0.335 Reciprocal of greatest dilution of supernate yielding total clot lysis 8 No Lysis 2 It is possible that differences in extracellular PA result from different levels of PA production. To test this, intracellular extracts of Hpa-l and Lpa-4 cells were obtained by treating confluent 44 cell monolayers with Q.5% V/V Triton X 100 as described previously by Unkeless et al. ll2J. As shown in table 3 PA was not detectable in extracts from Lpa-4 cells. It should also be noted that treatment of BHK-2l/C13 cells with Triton X 100 failed to release PA and that they also fail to produce detectable amounts of extracellular activator. These results suggest that differences in extracellular PA can be accounted for by differences in activator production. However, intracellular determinations were not made for all Lpa isolates and it is also possible that some Lpa populations may produce high levels of PA but release relatively small amounts of the activator. Table 3. Cell isolate Intracellular plasminogen activator. PA Units per 0.1 ml cell lysate ug Protein per 0.1 ml cell lysate Hpa-l 37 220 Lpa-4 o 300 a. One unit of PA correspond~to the amount of activator necessary for release of 5% of the 125I-fibrin. b. Protein determinations were performed according to Lowry et al. (1951). Efficiencies of Plating and Growth in Agar Suspension In examining various cell isolates, no apparent correlation between the amounts of plasminogen activator and efficiencies of plating on Falcon Petri dishes or abilities to grow in soft agar Table 4. Plasminogen activator production and efficiencies of plating in soft agar and on plastic surfaces. Cell isolate Frequency of activator positive colonies BHK-2l/N 0.224 Hpa-l Reciprocal of greatest dilution of supernate yielding visible clot lysis Cells forming colonies in soft agar Efficiency of plating on plastic surfaces 1 54.5 ± 3.5% 69.5 ± 4.0% 1.000 32 50.8 ± 9.3% 81.5 ± 9.0% Hpa-2 1.000 32 53.0 ± 7.8% 66.0 ± 3.3% Hpa-3 1.000 32 58.0 ± 5.0% 91.0 ± 3.6% Hpa-4 1.000 32 58.5 ± 4.9% 82.5 ± 3.5% Lpa-l 0.159 No Lysis 73.8 ± 6.9% 67.0 ± 3.9% Lpa-2 0.202 1 65.0 ± 6.7% 65.0±15.0% Lpa-3 0.032 ND 50.5 ± 4.8% 62.0 ± 3.9% a. The values expressed for efficiencies of plating represent the mean with 95% confidence limits. b. ND signifies that the datum was not determined. .p. V1 46 suspension were found. As shown in table 4, BHK-2l/N and isolates derived from them shared similar growth properties. These cells all possessed high efficiencies of plating on plastic surfaces and abilities to grow in soft agar. These characteristics were judged to be independent of the amounts of plasminogen activator produced. For example, Lpa-l had the highest efficiency of plating in soft agar, while Lpa-2 and Hpa-l shared the lowest efficiencies of plating in soft agar. Likewise, the efficiency of plating of Hpa-2 cells on plastic surfaces was similar to that observed with Lpa isolates. Growth Rates Growth rates were determined for a number of Hpa and Lpa isolates and these results are depicted in figure 1 and table 5. There was no apparent correlation between activator production (table 1) and the respective growth rates of those cells examined. However, the growth rates for Lpa-4, Lpa-5 and Hpa-l were determined at different time periods in the investigation than those for Lpa-6 and Hpa-6 cells. In comparing those growth rates determined at the same time, the Lpa cells demonstrated faster growth than their Hpa counterparts. This finding is consistent with the observation that Lpa colonies were larger than Hpa colonies when plated under identical conditions. It must also be noted that Lpa cells were initially selected as larger colonies. Aside from these considerations, it is apparent that those cells capable of releasing and/or producing increased levels of plasminogen activator do not possess increased growth rates. 47 Fig 1. Abscissa: Time in hours; Ordinate cell number XlO- 5 • Growth of Hpa and Lpa isolates. (0-0-0), Lpa-6 ( Hpa-6 (~--~--~) D-- D-- D), Lpa-4 (e __ e__ e), Lpa-5 Hpa-l ( I --- D- D), and cells were grown in MEM + 5% CS, the points on the growth curves represent multiple cell counts from duplicate plates. 48 40 30 20 10 I I I / I 8 I I I I I 6 / / 7 I I 4 I 3 I I / I 2 I ,?, If! t' t/ I I 'I I I I I / 1/ /I ; /1 / I 0 20 40 60 80 100 49 Table 5. a. Generation times of Hpa and Lpa isolates. Cell isolate Estimated (a) generation time Lpa-4 13.2 hr Lpa-5 16.9 hr Hpa-l 17.6 hr Lpa-6 12.6 hr Hpa-6 15.5 hr The generation times were estimated from the slopes of the curves depicted in fig 1. A further argument against a relationship between extracellular activator and cell growth stems from the observation that the frequencies of activator positive colonies remained rather stable during the serial passage of a number of isolates and examples of this are shown in table 6. If the action of plasminogen activator did affect cell growth rate this should result in a selection for either an increase or decrease in the frequency of activator positive colonies during the serial passage of cells. This did not occur and is interpreted as further evidence against any relationship between the levels of extracellular enzyme and cell growth rate. To further test this point, cell co-cultivation was employed. Equal numbers of Hpa and Lpa cells were mixed and the frequencies of activator positive colonies were determined at various times during their serial passage. shown in table 7. The results of this experiment are The frequencies of activator positive colonies 50 Table 6. Frequencies of activator positive colonies during extended periods of serial cell passage. Number of positive colonies per total colony number Frequency of positive colonies Cell isolate Passage level BHK-2l/N 6 102/719 0.142 BHK-2l/N 25 152/1323 0.115 BHK-21/N 39 210/938 0.224 Hpa-l 5 1462/1462 1.000 Hpa-l 16 1056/1056 1.000 Hpa-l 27 1017/1024 0.993 Lpa-4 1 8/1245 0.006 Lpa-4 11 10/1440 0.007 Lpa-6 1 7/710 0.01 Lpa-6 8 11/1085 0.010 a. Passage levels for BHK-21/N represent the number of serial passes since the initiation of these studies. b. Passage levels for Hpa and Lpa isolates represent the numbers of serial passes since the time of their isolation. Table 7. Co-cultivation of Hpa and Lpa cells. Frequency of activator positive colonies ( Positive Colonies ) (Total Colony Number ) Passage level 1 Cell isolate 2 3 5 8 Hpa-6 0.987 480 486 0.913 304 333 1.00 70 70 1.000 195 195 0.891 291 327 Hpa-6 x Lpa-6 0.456 308 675 0.355 309 871 0.20 20 98 0.08 43 562 0.008 7 883 Lpa-6 0.01 0.06 15 250 0.010 11 1085 710 0.009 667 ND ND indicates that the datum was not obtained. I..n ...... 52 decreased during serial passage. A similar finding was also obtained when Lpa-4 and Lpa-5 cells were mixed with Hpa-l cells and the frequency of activator positive colonies determined after 11 serial passages. Growth rates for the cells employed in co-cultivation were given in table 5. The observed decrease in activator positive colonies is reasonably attributed to the higher growth rates of the Lpa cells. These findings support the postulate that cells capable of releasing increased extracellular plasminogen activator do not exhibit enhanced cell growth under the conditions utilized for cell propagation. DISCUSSION The use of plasminogen casein overlays provides a rapid and effective means of cloning cells which release different amounts of plasminogen activator into serum-free media. The differences in activator production can be observed by determining either the frequencies of activator positive colonies arising from various cell populations or by measuring the amount of activator in serumfree culture fluids. Cells isolated from BHK-2l/N exhibited relatively large differences in amounts of extracellular plasminogen activator. Further characterization of these cells revealed a lack of correlation between amounts of extracellular PA and their growth rates or efficiencies of plating in soft agar or on plastic surface. Such findings are not supportive of previous conclusions of Pollack et al. [6] and Taylor 15]. However, Wolf and Goldberg I13] have 53 recently reported a lack of correlation between either the growth of RSV transformed chick cells in soft agar or their rate of glucose transport and the levels of extracellular activator in cell culture fluids. Jones, Laug and Benedict 114] were also unable to find a relationship between either growth in suspension, saturation densities, or tumorigenicity, and levels of extracellular activator in cultures of human fibrosarcoma cells. contradictions are unclear. Reasons for such apparent Yet, such findings rule out any general- ization concerning the relationships of plasminogen activator to the growth of cells suspended in semi-solid media. It is presumed that the amounts of activator which are present in serum-free supernates reflect the amounts of activator present in normal growth media. With the methods employed to assay PA this could not be confirmed due to the presence of large amounts of plasmin inhibitors in normal calf serum I5J. Nonetheless, it is concluded that cells which have relatively large amounts of activator in their serum-free culture media do not possess any enhanced growth properties under the conditions employed for cell propagation. conclusion stems from a number of experimental observations. This As shown in table 6, the frequency of activator-producing colonies remained stable during extended cell passage. Fig 1 and table 5 show that the growth rates of Hpa cells were not consistently greater than those of Lpa cells. Further supportive evidence came from the cell mixing experiment. When Hpa cells were co-cultivated and then followed during serial passage, a decrease in the number of activator positive colonies was observed. This is in direct 54 contrast to what would be expected if the presence of extracellular enzyme did favor cell growth. Work of Mott et ale 115], Goldberg [9J, and also unpublished data from this laboratory confirm that plasminogen activator production is not a unique quality of transformed cells. Many normal cells including primary fibroblasts have been shown to produce the enzyme. Although large differences in amounts of extracellular activator may be observed in culture fluids of various cells, it is plausible to suggest that all cells produce enough plasminogen activator to provide a required physiological function. However, it is unclear whether extracellular levels of the enzyme are in fact physiologically important. It is possible that the enzyme may func- tion intracellularly or while membrane associated. With the exception of chick cell plasminogen activator [12], little is known about the specific action of activators found in cell culture fluids on plasminogen. Furthermore, protein substrates other than plasminogen have not been reported and attempts to identify susceptible surface components have been unsuccessful [16, l7,J. It is also possible that cells release a number of molecules with activator activity. Poste [18J has shown that more than one species of activator can be separated from SVT2 culture fluids by exclusion chromatography while Christman, Silverstein and Acs ll9J have reported that different cells from the same animal species are capable of releasing antigenically distinguishable activators. The specificity of the activator released by BHK-2l/N cells and its intracellular localization is currently being investigated. 55 The elucidation of such properties for the activator produced by BHK-21/N as well as other cell lines may well preclude the discovery of its physiological function and importance in cell transformation. REFERENCES 1. Hynes, RO, CellI (1974) 147. 2. Roblin, R. Chou, I. Black, PH, Adv cancer res 22 (1975) 203. 3. Sumaria, L, Hsieh, B, & Robbins, KC, J bioI chern 242 (1967) 4279. 4. Unkeless, JC, Tobia, A, Ossowski, L, Quigley, JP, Rifkin, DB, & Reich, E, J exp med 137 (1973) 85. 5. Taylor, JC, PhD dissertation, University of Utah (1973). 6. Pollack, R, Rissen, R, Conlon, S, & Rifkin, P, Proc nat acad sci US 71 (1974) 4792. 7. Ossowski, L, Quigley, JP, Kellerman, Q1, & Reich, E, J exp med 138 (1973) 1056. 8. Taylor, JC, Hill, DW, & Rogolsky, M, Exp cell res 73 (1972) 442. 9. Goldberg, AR, Cell 2 (1974) 95. 10. Deutsh, DG & Mertz, ET, Science 170 (1970) 1095. 11. MacPherson, I & Montagnier, L, Virology 23 (1964) 291. 12. Unkeless, JC, Keld, D, Kellerman, G, & Reich, E, J bioI chern 249 (1974) 4295. 13. Wolf, BA & Goldberg, AR, Proc nat acad sci US 72 (1976) 3613. 14. Jones, PA, Loug, WE, & Benedict, WF, Cell 6 (1975) 245. 15. Mott, DM, Fabisch, PH, Brahma, P, Sorof, S, & Sorof, S, Bioch bioph res com 61 (1974) 621. 16. Christman, JK, Acs, G, Silagi, S, & Silverstein, SC, Proteases and biological control (eds E Reich, EB Rifkin, & E Shaw) p. 827 Cold Spring Harbor Lab, Cold Spring Harbor, New York (1975). 17. Blumberg, PM & Robbins, PW, Cell 6 (1975) 137. 18. Poste, G, Cancer res, 35 (1975) 2558. 57 19. Christman, JK, Silverstein, SC, & Acs, G, J exp med 142 (1975) 419. 20. Lowry, OR, Rosborough, NJ, Farr, AL, & Randall, RJ, J bioI chern 193 (1951) 265. CHAPTER III BIOCHEMICAL PROPERTIES OF A PLASMINOGEN ACTIVATOR FROM BHK CELLS by Robert W. Snyder and Douglas W. Hill 59 SUMMARY A plasminogen activator which was produced by a high passage hamster kidney cell line (Hpa-6) was characterized biochemically. The enzyme was found to be inhibited by diisopropylfluorophosphate and therefore was considered to be a serine protease. Hpa-6 cell activator was also found to react with plasminogen in a manner which was indistinguishable from the action of urokinase on the proenzyme. Evidence was obtained which suggests that the activator can hydrolyze either of the two molecular forms of plasminogen, i.e., plasminogen with glutamic acid amino terminus or plasminogen with lysine amino terminus at a single position to yield active plasmin. The cellular localization of the Hpa-6 cell plasminogen activator was examined and the results obtained indicate that this activity was largely associated with the plasma membrane. INTRODUCTION A large number of cell types grown in vitro have been shown to release enzymes which are capable of activating plasminogen Il,2,3,4,5,6,7,8,9]. This phenomena has generated much interest since investigators had previously suggested that the action of proteolytic enzymes may be involved in alteration of the in phenotypic properties of cells flO,llJ. --- There have been a number of reports which have attempted to relate the presence of amounts of extracellular activator with alterations of cellular phenotypic properties 12,8,12,13,14,15J and generally their conclusions have not been in agreement. Some of the confusion which has resulted from these studies may 60 be ascribed to the use of a variety of cell lines which may produce activators with different enzymatic capabilities and/or biological functions. Cells from different species and similar cells obtained from the same species have been shown to release activator molecules which differ in molecular weight [6,16,17J or are antigenically unrelated [18,19J. Unkeless et ale [17J and Christman et al. [20] have also provided evidence that the cellular localization of plasminogen activator differs in chick cells transformed by Rous sarcoma virus versus hamster cells transformed by SV40 virus. Since researchers will continue to employ a variety of cells to investigate the possible effects of extracellular proteases on cell growth or behavior it is of obvious importance that those proteolytic enzymes dealt with be examined and described. In this regard, biochemical information would not only be useful toward elucidating the biological functions of such enzymes but would allow for more accurate comparisons of results obtained with different cell lines. In our previous studies cells which differed in the amount of extracellular plasminogen activator were cloned from a high passage hamster cell line which had originated from BHK-2l/C13 [21], and populations of these cloned cells designated Hpa (possessing high levels of extracellular activator) and Lpa (possessing low levels of extracellular activator) were examined with regard to a number of growth properties. This report is directed toward presenting data concerning some of the biochemical and enzymatic properties of the activator released by Hpa cells. More specifically, the following results will provide information concerning the action of this enzyme 61 on plasminogen, the effects of various inhibitors on its activity, and the intracellular localization of this protease. MATERIALS AND METHODS The cells used in these studies (Hpa-6) were selected for their ability to release large amounts of plasminogen activator [21). These cells were isolated from BKH-21/N cells which had been serially passaged at least 200 times in culture and that had originated from BHK-21/C13 cells. The Hpa-6 cells were grown in Eagle's medium (MEM) (Gibco) supplemented with 5% calf serum (CS) (Flow Labs). Concentration of Extracellular Plasminogen Activator Hpa-6 cells were seeded into Corning plastic roller bottles and incubated at 37°C until the cells were nearly confluent. The MEM with 5% CS was removed and cells were gently washed with three changes of phosphate buffered saline without added calcium and magnesium ions (PD). A 50 ml volume of MEM without calf serum was added to each bottle and the cells were incubated for approximately 18 hours at 37°C. The serum-free supernates to be used for activator concentra- tion were then removed and MEM with 5% CS was added back to the roller bottles. This entire procedure was repeated once more after the cells had been incubated for 5 additional hours in the presence of normal growth medium. The method employed for plasminogen activa- tor concentration was essentially that of Christman and Acs [16). The ser~free culture fluids were centrifuged for 10 minutes at 62 ~OOO g in a refrigerated centrifuge immediately following their har- vest. Supernates were carefully removed and the pellet containing cellular debris was discarded. These fluids were then filtered through Whatman #50 filter paper and adjusted to pH 3.0 by the addition of LM HCl. The acidic solution was placed in an ice bath and brought to 50% saturation by the addition of solid ammonium sulfate. Following 30 minutes at 4°C, this solution was centrifuged for 15 minutes at 7,000 g in a refrigerated centrifuge. The pellet was resuspended in 0.0125 times the original volume with a 0.05M glycine buffer at pH 3.0. This preparation (stage 1) was frozen at -70°C. The entire procedure was repeated for the second batch of culture fluids and the stage 1 fractions from both harvests were pooled, and a precipitate was removed following its centrifugation at 11,000 g at 4°C for 15 minutes. The resulting supernate (stage II) was dialyzed twice against 100 volumes of O.005M phosphate buffer at pH 7.4 and aliquots of the dialysate (stage III fluid) were then frozen at -70°C. Protein determinations of crude harvest fluids and stage II fractions were performed according to Lowry et ale [22]. Radioactive Labeling of Plasminogen Plasminogen was purified according to the method of Deutch and Mertz [23] and it was then reacted with [3H] acetic anhydride (2Ci/m mole) (New England Nuclear) as described by Montelaro and Reuckert [24]. The amount of [3H] acetic anhydride added was in 5 molar ex- cess of the amount of plasminogen present. Alternately, the plasminogen was labeled with l25Iodine. In this 63 case, the plasminogen used for iodination was first rendered free of most residual plasmin by adding a plasminogen solution to Affigel-IO beads (Bio Rad) to which soy bean trypsin inhibitor had been previously coupled. Following a 60 minute incubation at 4°C, the gel beads were removed by low speed centrifugation. The resulting plas- minogen solution was iodinated by the method of Helmkamp et ale [2S] and l2S 1 and iodine chloride (ICI) were added to give a ratio of approximately 3 moles ICI to 1 mole plasminogen. Both [3H] acetic anhydride and l2S1 labeled plasminogen were then freed of nonreacted radioactivity by column chromatography with Bio Rad PIO equilibrated with O.OOSM phosphate buffer at pH 7.4 and appropriate aliquots of radiolabeled plasminogen were frozen at -70°C. Activation of Radioactive Plasminogen All reactions were carried out in a final concentration of approximately 0.1M Tris buffer at pH 8.1 or 8.S. Appropriate volumes of stage III activator were added to aliquots of radiolabeled plasminogen and incubated at 37°C. The volumes of each reactant varied in different experiments and are indicated in the text. The reac- tions were stopped at desired times by the addition of equal volumes of 10% glycerol, S% S-mercapto-ethanol, 3% sodium dodecyl sulfate (SDS) and 0.062SM Tris at pH 6.8 (sample buffer used for SDS polyacrylamide gel electrophoresis) followed by immediate freezing at -20°C. 64 SDS Polyacrylamide Gel Electrophoretic Analysis of Radiolabeled Proteins The electrophoresis of plasminogen or activated plasminogen preparations were carried out in the presence of SDS as previously scribed by O'Farrel et al. [26]. de~ The plasminogen plus activator reaction mixtures which were in sample buffer were thawed immediately by placing them in a boiling water bath for approximately 2 minutes and appropriate sample volumes were then electrophoresed. Gels which contained 3H labeled plasminogen were prepared for fluorography [27] and RPX o-mat X-ray film was employed to locate and identify plasminogen or split products derived from the proenzyme plasminogen. Gels which contained l25I-plasminogen were simply fixed with 12% acetic acid, dried under vacuum, and used to expose the X-ray film. These resultant autoradiographs were placed above the original gel and the appropriate regions were identified and cut from the gels and radioactivity was assayed in a Beckman Biogamma counter. Fluorescamine Labeling and Isolation of Plasma Membrane Enriched Fractions The preparation of plasma membrane enriched fractions was accomplished by use of a method similar to that reported by Atkinson and Summers [28]. Hpa-6 cells were grown in 15 rom Lux Petri dishes to approximately 50% confluence, washed twice with 15 ml volumes of PD, and 5 ml volumes of PD were added. One-half of these cell cul- tures were reacted with fluorescamine by adding 0.02 ml volumes of 10 mg/ml Fluoram (Roche Diagnostics) in acetone. The PD and PD plus hydrolyzed fluorescamine were removed, cells were washed twice more 65 with 15 ml volumes of PD, and 5 ml volumes of 0.005M EDTA plus 1 mg/ml bovine serum albumin in PD were added. Following 5 minutes at room temperature the loosely adherent cells were removed by pipetting the fluid from each dish onto patches of the monolayer. The suspensions of cells were pooled and placed in 12 ml conical centrifuge tubes and pelleted by low speed centrifugation. The cells were washed in PD, repelleted, and suspended in a volume of swelling buffer (O.OIM Tris plus 0.1 roM MgC1 2 at pH 7.4) which was at least 20 times greater than that of the cell pellet. Following 4 or 5 minutes in an ice bath, the cells were ruptured with a stainless steel Dounce homogenizer (0.002 inch clearance). The cell homogenate was brought to 3mM MgC12 and O.OIM NaCl and then centrifuged to remove whole cells and nuclei. The resulting supernate was held on ice and appropriate volumes were added to the top of a discontinuous sucrose gradient which contained 3.3 ml 50%, 3.3 m1 39%, and 3.3 ml 20% sucrose in ET buffer (O.1M Tris and 0.001 M EDTA at pH 7.4). These gradients were centrifuged in a Beckman SW41 rotor at 37,000 rpm for 3 hours at 4°C. Fractions (0.55 ml) from the gradient were collected and aliquots of each frozen. Determination of Plasminogen Activator Activity Two methods were employed to quantitate amounts of plasminogen activator. An l25I-fibrin release assay was performed in a manner similar to that described by Unkeless et ale [3) with the following modifications: (1) l25I -fibrin was clotted by addition of 0.1 u/ml 66 bovine thrombin (Parke-Davis) in the presence of phosphate buffered saline for 2 hours at 37°C; (2) the assays were carried out in Costar multiwelled dishes (16 mm) with 0.1 ml of plasminogen (approximately 0.1 mg/ml), 0.1 ml of 0.1M Tris (pH 8.1) and 0.1 ml of activator preparation in each well. The plasminogen had been purified from outdated human plasma by the method of Deutch and Mertz [23]. Plasminogen activator was also assayed by use of casein-plasminogen containing overlays. Petri dishes were filled to a depth of 2-3 rom with 0.4% agarase, 0.4% denatured skim milk and 0.01 mg/ml plasminogen in PD. Wells of approximately 4 rom in diameter were cut out and filled with 0.02 ml of an activator preparation. The diame- ter of the zone of clearing was measured after the plates had been incubated at 37°C for 24 hours. The log of activator concentration was plotted versus the log of the diameter of visible lysis to yield a standard line. RESULTS Table 8 shows a typical preparation of the stage III Hpa-6 cell activator contained approximately 92% of the original activity in about 0.0125 times the original volume. The mechanism of activation of l25I-labeled plasminogen by stage III Hpa-6 cell activator and urokinase (Cal-Biochem) was examined. An autoradiograph of the l25I-labeled plasminogen which had been reduced in S-mercaptoethanol and electrophoresed in an SDS polyacrylamide gel is shown in fig 2b. This particular preparation appears relatively free of other Table 8. Concentration of activator from Hpa-6 cell culture field. Specific Activity Volume Protein Activator a Stage I 700 ml 2.1 mg 5.6 x 104U 2.66 x 10 4 uimg 100% Stage III 8.6 ml 0.8 mg 5.16 x 104U 6.45 x 10 4 U fmg 92% Yieldb ~. Units of activator are relative to the amount present in a standard pool of cell culture fluid and 0.6U of activator results in the release of approximately 5% of the 1251 fibrin in the assay system employed. b. The percent yield refers to percent of activator activity remaining. 0"1 -.....J 68 contaminating proteins and the only visible bands corresponded to the two molecular forms of plasminogen. It should be noted that the smaller peptide corresponding to the plasmin light chain is not visible and the amount of radioactivity at this position was at least 100 times less than the amount present in the plasminogen bands. In figs 2c and 2d the products resulting from the proteolytic activation of l25I-labeled plasminogen by Hpa-6 cell activator and urokinase respecitvely are shown. These products were compared to 35 S_ methionine-labeled T4 phage marker proteins supplied kindly by G. Stetler and shown in fig 2a. The apparent molecular weights for Glu- PIn (plasminogen with flutamic acid at the amino terminal position), Lys-Pln (plasminogen with lysine at the amino terminal end), Glu-H (heavy plasmin chain derived from Glu-Pln), and Lys-H (heavy plasmin chain derived from Lys-Pln) are larger than those commonly reported [29,30,31,32] but were comparable to those depicted by Dano and Reich [33] who employed a similar electrophoresis system. It should be pointed out that the use of the designations Gly-Pln and Lys-Pln for the two moledular forms of plasminogen was based on data communicated by other authors who had examined the terminal amino groups of purified human plasminogen I34,35]. As seen in figs 2c and 2d the prod- ucts resulting from plasminogen activation by urokinase or Hpa-6 cell activator appear to be indistinguishable. Both molecular forms of plasminogen (Glu-Pln and Lys-Pln) were activated. This conclusion is based on the presence of a single plasmin light chain (PI) and two distinguishable plasmin heavy chains (Glu-H and Lys-H) which have an apparent molecular weight difference which was consistent with that 69 Fig 2. Polacrylamide SDS gel electrophoresis of 125 1 plasmino- gen and activation products. All reaction mixtures were incubated for 1 hour at 37°C, diluted 2-fold in sample buffer and reduced by boiling, and appropriate samples were run in the wells corresponding to the following reaction mixtures: b: 20 ~1 l2SI plasminogen (z20~gm) + 5~1 0.4M Tris (pH 8.S) + 20~1 O.SmM sodium phosphate (pH 7.4) + 5~1 H20. c; 40~1 l2SI plasminogen (z40~gm) + 10~1 0.4M Tris (pH 8.S) + stage III activator + u/ml). + 20~1 40~1 10~1 Traysylol (10,000 d; 20~1 l2SI plasminogen (z20~gm) + 5~1 0.4M Tris (pH 8.S) urokinase (35u/ml) + S~l Traysy10l (lO,OOOu/ml). e; 20~1 125 1 plasminogen (z20~gm) + 5111 0.4M Tris (pH 8.5) + 20111 streptokinase (l,OOOu/ml) + 5~1 H20. f; 20~1 1251 plasminogen (~20lJgm) + S1-11 0.4M Tris (pH 8.S) + 20lll urokinase (3SlJlm1) + S1-1l H20. The proteins in well a are from 3SS-methionine-labeled T4 phage and their estimated molecular weights in da1tons are indicated. a b G1u- PlnLy s- Pln 112 ,000Gl u- H Lys- H - 67, 000- 46 ,000 36,00024,000- PI- ed e f 71 expected if the heavy plasmin chains were derived from the amino terminal portions of Glu-Pln and Lys-Pln. The activation of Gly-Pln Lys-Pln by Hpa-6 cell enzyme seemed to be equally as efficient. In this regard, the regions of the gel corresponding to each form of the zymogen and each of heavy plasmin chains were cut out of some gels in which they were well separated and the ratio of radioactivity of Glu-Pln/Lys-Pln was compared to Glu-H/Lys-H. These ratios (Glu-Pln/Lys-Pln=0.684 and Glu-H/Lys-H = 0.686) were found to be nearly identical as would be expected if both forms were cleaved at the same rate. The activation of plasminogen was examined as a function of time and activator concentration. In fig 3 the amounts of radioactivity in the positions corresponding to plasmin light and heavy chains increased with time when l25I-plasminogen was reacted with a constant amount of Hpa-6 cell activator. When 3H plasminogen + Hpa-6 cell activator reaction mixtures were electrophoresed and fluorographed the amount of plasmin light chain (see fig 4) was found to be dependent upon the concentration of activator added. The effects of some inhibitors on Hpa-6 cell activator were also examined. As shown in fig 5 a 1 roM concentration of diisopropyl- fluorophosphate (DFP) was very effective in inhibiting the enzyme. The amount of plasmin light chain present in the DFP containing reaction mixture was determined by cutting the corresponding band out of the gel and determining radioactivity. The radioactivity in this plasm band was only 20% of the amount obtained from a control reaction mixture which contained 3% propanol and 1 roM toysyllysylchloro- 72 Fig 3_ Abscissa: Time in minutes; ordinate (left) % of radioactivity in plasmin heavy chains versus total radioactivity in all bands (.------) (right) % of radioactivity in plasmin ·light chain versus total radioactivity in all bands (.-.-.). Kinetics of plasminogen activation by Hpa-6 cell activator_ The following reaction mixture was incubated at 37°C, aliquots were removed at appropriate times and diluted 2-fold with gel electrophoresis sample buffer, reduced by boiling, and electrophoresed. Appropriate bands were cut from the gels and the radioactivity determined. + l80~1 + 9~1 Reaction mixture: l80~1 1251 plasminogen (~18~gm) Hpa-6 cell stage III activator + Traysylo1 (10,000 U/m1). 45~1 0.4M Tris at pH 8.5 73 25 20 15 10 10 5 5 a 20 40 60 80 100 120 74 methylketone (TLCK). This decrease in the activity of Hpa-6 cell activator resulted from a 30 minute pre-incubation of the enzyme with DFP. Though the kinetics of inactivation by DFP were not examined, the data which were obtained strongly suggest activator is a serine protease [36]. that the A 1 roM concentration of tosylphenylchloromethylketone (TPCK) was without effect on the enzyme. In further studies which employed 3H plasminogen as sub- strate a decrease in activator activity was noted in the presence of tosylarginylmethyl ester (TAME). This finding was based on the visualization of decreased amounts of plasmin light chain in fluorographed reaction mixtures (fig 6) which contained 0.5 roM TAME or 0.05 roM TAME. However, this decrease in amount of plasmin light chain was not quantitated. Plasminogen activator from Triton X 100 extracts of Hpa cells has been obtained and further studies were performed to identify its cellular localization. As described in the methods section, cell surface molecules were labeled with fluorescamine, fractions enriched for plasma membrane were prepared, and the material was centrifuged to equilibrium in sucrose gradients. results of this analysis. Fig 7 shows the Fractions 12, 13, and 14, which contained approximately 50% of the detectable activity, were the only fractions which possessed visible fluorescence and contained fluorescent plasma membrane ghosts as determined by microscopy. This strongly suggests that much of the cellular activator may be associated with the plasma membrane. It should also be noted that greater than 50% of the enzymatic activity which was added to the gradient could be 75 Fig 4. Polyacrylamide SDS gel electrophoresis of 3H plasminogen activated by various concentrations of stage III Hpa-6 cell activator. The reaction mixtures contained TLCK as an inhibitor of plas- min and they were incubated at 37°C for one hour before addition of electrophoresis sample buffer. All reaction mixtures contained 20~1 3H plasminogen (~lO~gm), 5~1 O.35M Tris (pH 8.5) with O.07M TLCK, and a through e also contained the following: phosphate (pH 7.4). b; 20~1 a·, 25~1 5 roM sodium 5 mM sodium phosphate (pH 7.4) + 5~1 stage III activator. c; 15111 5mM sodium phosphate (pH 7.4) + lO~l stage III activator. d-, lO~l 5mM socium phosphate (pH 7.4) + l5~1 stage III activator. e·, 25~1 stage III activator. ab GI u - Pin L'y s -Pin PI ~- c d e 77 Fig 5. Polyacrylamide SDS gel electrophor~sis activated by DFP treated stage III activator. of plasminogen The reaction mixtures represented in a and b each contained 2~1 1251 plasminogen (~2~gm), l6~1 stage III activator, and l5~1 0.35M Tris (pH 8.5) with 3.5 mm TLCK and were incubated at 37°C for 1 hour prior to the addition of electrophoresis sample buffer. The stage III activator in well a had been pre-incubated for 30 minutes at 25°C with a final concentration of 1 roM DFP and 3% propanol, the stage III activator in well b was pre-incubated for 30 minutes at 25°C with a final concentration of 3% propanol. Glu-Pln Lys-Pln PI • ., • 79 accounted for. The apparent activity of plasminogen activator could be enhanced when Triton X 100 was added to the 1251 fibrin release assay reaction mixture. In view of the plasma membrane association of Hpa-6 cell activator it seemed possible that some of the cell culture fluid activity could still be associated with plasma membrane vesicles or fragments and that this association could account for the enhancement by Triton X 100. It was found that when 0.166% Triton X 100 was present in the l25 I -fibrin assay the amount of 125 1 fibrin released by Hpa-6 cell activator plus plasminogen was nearly doubled. However, when streptokinase (a presumably soluble activator of bacterial origin) and plasminogen were preincubated and then assayed by l25I-fibrin release a similar enhancement of fibrin release by added Triton X 100 was observed and Triton X 100 also increased the amount of l25I-fibrin solubilized in control reactions which did not contain activator and/or plasminogen. Since the l25I-fibrin which serves as substrate in these assays was first insolubilized on the surface of Petri dishes it is possible that the detergent (Triton X 100) could increase the solubility of 1251 fibrin substrate or l25I-fibrin degradation products and thereby raise the total 125 1 released. A second approach was then employed to investigate whether the activator present in cell culture fluids could be associated with membranous material. Serumrfree Hpa-6 cell culture fluids were ob- tained after 18 hours of incubation. The harvested fluids were 80 Fig 6. Polyacrylamide SDS gel electrophoresis of 3H plasmino- gen activated by stage III activator in the presence of TAME. The reaction mixtures contained TLCK as an inhibitor of plasmin and were incubated at 37°C for three hours before the addition of electrophoresis sample buffer. ogen (~7.5pgm), IO~1 20~1 stage III activator. They all contained 15~1 3H plasmin- O.45M Tris (pH 8.2) with O.045M TLCK, and Additionally, mixture b contained 5 roM TAME and mixture c contained lpl 0.5 roM TAME. l~l Glu-Pln Lys-~' In PI ..• .. 82 Fig 7_ Abscissa: Fraction number from gradient; Ordinate (left) units of plasminogen activator activity, (right) density of sucrose gm/cm 3 ( - - - - - ) ; top: visible fluorescence (+). Sucrose gradient fractionation of Hpa-6 cellular material. Aliquots from each fraction were diluted 2-fold in 0.5% Triton X 100 and the amount of activator in a 20~1 volume was measured by means of casein and plasminogen containing overlays as described in Materials and Methods. The units of activator activity were deter- mined relative to a standard cell culture fluid and are the same as those in table 8. 83 I I I I ., ~ 450 400 350 1.25 300 1.20 250 -1.15 200 1.10 150 1.05 100 80 5 13 17 21 25 84 centrifuged twice at low speed in conical tubes to remove any whole cells or large debris and Triton X 100 was added to some of the aliquots to yield a final concentration of 0.5%. Samples with and without the Triton X 100 were centrifuged at 100,000 g for one hour at 4°C and the supernates and sedimented material were carefully obtained. All preparations were brought to 0.5% Triton X 100 and assayed for l25I-fibrin release. As shown in table 9, 31% of the supernate activity was pelleted in the absence of Triton X 100, while only 3% of the activity was sedimented in the presence of 0.5% Triton X 100. The evidence suggests that some of the acti- vator is particulate and possibly associated with membranous material. However, this is not conclusive and further experiments have not been performed. DISCUSSION The results presented (see fig 2) show that the products resulting from the activation of plasminogen by Hpa-6 cell activator or urokinase were indistinguishable. Early reports had suggested that a preactivation peptide was first removed from Glu-Pln to yield Lys-Pln [37,38,39] and the subsequent activation of Lys-Pln was accomplished by the hydrolysis of a single arginyl-valine bond in the Lys-Pln proenzyme [39]. However, the experiments which led to these conclusions were difficult to interpret in that it was also possible that the autoproteolytic action of plasmin could have accounted for the conversion of Glu-Pln to Lys-Pln. Recently, investigators have employed the proteolytic inhibitor Traysylol to Table 9. High speed centrifugation of cell culture fluids with and without Triton X 100 Activator Preparation 100,000 G supernatant 100,000 G pellet 100,000 G Triton X 100 supernatant 100,000 G Triton X 100 pellet a. Activity (c) in Units % of Total 210 69 96 31 317 93 11 3 Plasminogen activator was measured by 125 1 fibrin release and units are as defined in Table 8. 00 l/l 86 examine plasminogen activation and found that the cleavage of the preactivation peptide did not occur when plasmin was inhibited [40] and the results reported here confirm these findings. Evidence was presented which suggests that Hpa-6 cell activator and urokinase can activate either the Glu-Pln or Lys-Pln form of the zymogen through the splitting of a single peptide bond. Further evidence was also presented that suggests that the Hpa-6 cell activator reacts with Glu-Pln or Lys-Pln with equal efficiency. Our findings are different from those reported by Unkeless et ale [17] following their investigation of the activation of plasminogen with the activator from RSV transformed chick cells in the presence of Traysylol. These authors had concluded that the rate of activation of Lys-Pln by chick cell activator was greater than the conversion of Glu-Pln. It is possible that the reaction mixtures which they examined contained active plasmin. The presence of plasmin in their reaction mixtures was evidenced by the presence of "large degradation material" which migrated slightly further than Lys-H in their electrophoretic system and which was clearly distinguishable in the autoradiographs presented. These "large degradation peptides" were also noticed in the present studies in autoradiographs (fig 2e and 2f) of those electrophoresed reaction mixtures which did not contain Traysylol or other suitable plasmin inhibitors. Aside from providing information on the mechanism of plasminogen activation by Hpa-6 cell activator our results also suggest that the Hpa-6 cell activator has a relatively specific action on plasminogen and may possess a rather narrow range of substrates. 87 In this regard only one peptide bond in plasminogen was apparently cleaved by the activator. In contrast, plasmin's action on plasminogen (fig 2e and 2f) was found to result in not only the removal of the preactivation peptide but numerous other peptide fragments were observed. It should also be noted that crude cell culture fluids which contained plasminogen activator did not contain any distinguishable proteolytic activity when 3H-hemoglobin or 3H- human serum albumin were employed as substrates and the resulting reaction mixtures were electrophoresed and analyzed by fluorography (results not shown). Christman et ale [20] have re- ported that the plasminogen activator released by hamster cells did not activate trypsinogen, chymotrypsinogen, or pepsinogen, while other investigators have been unable to demonstrate any loss of radiolabeled cell surface proteins when cells were incubated in the presence of fluids containing plasminogen activators [20,41]. The effects of some inhibitors on the activity of Hpa-6 cell activator was also examined. A 1 mM concentration of TPCK was found to be without effect when added to a reaction mixture containing Hpa-6 cell activator. In contrast, a large reduction in activator activity was observed when it was incubated with 1 mM DFP for 30 minutes. This result strongly suggests that the activator is a serine protease [36]. Similar conclusions have also been reported for activators released by Rous sarcoma virus transformed chick cells [17], cells obtained from human melanomas, [6] and SV40 transformed hamster cells [16]. Unkeless et aL [17] have also reported 88 that the chick cell activator was inhibited by TAME and similar findings were also observed for the Hpa-6 cell activator. Examination of the cellular localization of Hpa-6 cell activator has suggested that this activity may be associated with the plasma membrane. ,Hpa-6 cells were fluorescently labeled with fluorescamine a compound previously shown to react specifically with surface molecules [42]. When plasma membrane ghosts [28] which were prepared from the labeled Hpa-6 cells were centrifuged to equilibrium in sucrose gradients, much of the activator activity cosedimented with fluorescent membrane ghosts and all of the visible fluorescence (see fig 7). Since these materials co- sedimented to a density (~1.16 gm/cm 3 ) previously described as isopycnic for plasma membrane ghosts [28] it is likely that much of the plasminogen activator activity was associated with the cell membrane. Since the Hpa-6 cell plasminogen activator appears to be associated with plasma membrane the presence of extracellular activator could reflect its active release from the cell surface, the budding of membranous vesicles containing the activator, or membrane perturbation and subsequent loss of surface material. Though these possi- bilities have not been examined directly, a limited amount of evidence was obtained which may indicate that some of the extracellular enzymatic activity may be associated with membranous material. Approximately 1/3 of the culture fluid activator activity could be sedimented by a 1 hour centrifugation at 100,000 g and the sedimentation of this material was prevented by the addition of 0.5% 89 Triton X 100. The distruptive action of detergents on cell membranes is well known and this could account for the altered sedimentation properties of extracellular activator when Triton X 100 was added. Christman et al. [20] have also reported that the cellular activator from SV4D transformed hamster fibroblasts was associated with the plasma membrane. However, Unkeless et aL [17] have em- ployed differential centrifugation of intracellular material and suggested that the activator present in Rous sarcoma virus transformed chick cells was localized in lysomal or microsomal fractions. These findings were somewhat surprising since these authors had previously demonstrated that no increases in the extracellular concentration of various lysosmal enzymes were detected in the culture fluids of transformed cells which produced and/or released increased amounts of activator [3]. More recently, Quigley [43] has re- ported that the cellular activator from RSV transformed chick fibroblasts was associated with plasma membrane or plasma membrane-like elements. In his communication he demonstrated that an activator from transformed chick cells co-purified with material which appeared to be membranous in electron micrographs and with molecules commonly reported to be associated with the plasma membrane. It is not known whether the activators from Hpa-6, Rous sarcoma virus transformed chick, and SV40 transformed hamster cells are associated with the surface membrane or other internal and membranous organelles. Authors have suggested that proteases could be involved in the alterations of some surface properties which are ascribed to transformed cells and it is possible that the actions of plasminogen 90 activators could be responsible for some surface alterations (see reviews by Roblin et al. [44] and Hynes [11]). Yet, attempts to detect loss of cell surface material following the addition of fluids containing plasminogen activator have been unsuccessful [20,41] and protein substrates other than plasmin have not been identified for any of the activators described. Since the activator is produced by a wide variety of cells it is also possible that it is involved in some common yet undefined cellular process, i.e., glycoprotein secretion, protein turnover, or modification of the newly synthesized proteins. Certainly, the membrane association could be studied in more detail and the results of such investigation could help to elucidate the biological function of some plasminogen activators. REFERENCES 1. Taylor, JC, Hill, DW, & Rogolsky, M, Exp cell res 73 (1973) 422. 2. Taylor, JC, PhD Dissertation, University of Utah (1973). 3. Unkeless, JC, Tobia, A, Ossowski, L, Quigley, JP, Rifkin, DB, & Reich, E, J exp med 137 (1973) 85. 4. Ossowski, L, Unkeless, Je, Tobia, A, Quigley, JP, Rifkin, DB, & Reich, E, J exp med 137 (1973) 112. 5. Goldberg, AR, Cell 2 (1975) 95. 6. Rifkin, DB, Loeb, J, Moore, G, & Reich, E, J exp med 139 (1974) 1317. 7. Wacksman, J, & Biedler, JL, 8. Pollack, R, Risser, R, Conlon, S, & Rifkin, R, Proc nat acad sci US 71 (1974) 4792. 9. Mott, DM, Fabisch, PH, Brahma, P, Sorof, S, & Sorof, S, Bioch bioph res com 61 (1974) 621. Exp cell res 86 (1974) 264. US 62 (1969) 994. 10. Burger, MM, Proc nat acad sci 11. Hynes, RO, 12. Ossowski, L, Quigley, JP, Kellerman, GM, & Reich, E, CellI (1974) 147. J exp med 138 (1973) 1056. 13. Gallimore, PH, McDougall, JK & Chen, LB, Cell 10 (1977) 669. 14. Jones, PA, Laug, WE, & Benedict, WF, 15. Wolfe, BA & Goldberg, AR, 16. Christman, JK & Acs, G bio9h bioph acta 340 (1974) 339. 17. Unkeless, J, Dano, K, Kellerman, G, & Reich, E, Cell 6 (1975) 245. Proc nat acad sci US 73 (1976) 3613. J bioI chem 249 (1974) 4295. 18. Christman, JK, Silverstein, SC, & Acs, G, (1975) 419. J exp med 142 92 19. Bernik, M & Kwaan, H, 20. Christman, JK, Acs, G, Silagi, S, & Silverstein, SC, Proteases in biological control (eds E Reich, DB Rifkin, & E Shaw) p. 827 Cold Spring Harbor lab, Cold Spring Harbor, New York (1975). 21. Snyder, RW & Hill, DW (Manuscript Submitted Exp cell res). 22. Lowry, OH, Rosenbrough, NJ, Farr, AL, & Randall, RJ, chem 193 (1961) 265. 23. Duetsch, DG & Mertz, ET, Science 170 (1970) 1095. 24. Montelaro, RC & Reuckert, RR, 25. Helmkamp, RH, Goodland, RL, Bale, WF, Spar, IL, & Mutschler, LE, Cancer Res 20 (1960) 1945. 26. O'Farrel, PZ, Gold, LA, & Huang, WM, 5499. 27. Bonner, WM & Laskey, RA, Eur j bioch 46 (1974) 83. 28. Atkinson, DH, & Summers, DF, 29. Summaria, L, Arzadon, L, Bernabe, P, & Robbins, KC, J bioI chem 247 (1972) 4691. 30. Gonzalez-Gronow, M, Violand, BN, & Caslettino, FJ, J bioI chem 252 (1976) 2175. 31. Robbins, KC, Summaria, L & Barlow, GH, Proteases in biological control (eds E Reich, DB Rifkin, and E Shaw) P 305 Cold Spring Harbor labs, Cold Spring Harbor New York (1975). 32. Wallen, P & Wiman, B, Proteases in biological control (eds E Reich, DB Rifkin, and E Shaw) p. 291 Cold Spring Harbor lab, Cold Spring Harbor New York (1975). 33. Dano, K & Reich, E, Proteases in biological control (eds EReich, DB Rifkin, and E Shaw) p. 357 Cold Spring Harbor Lab, Cold Spring Harbor, New York (1975). 34. Robbins, KC, Bernabe, P, Arzadon, L, & Summaria, L, J bioI chem, 248 (1973) 7242. 35. Summaria, L, Arzadon, L, Bernbe, P, & Robbins, KC, J bioI chem, 248 (1973) 2984. J clin invest 48 (1969) 1740. J bioI J bioI chem 250 (1975) 1413. J bioI chem 248 (1973) J bioI chern 246 (1971) 5162. 93 36. Hartley, BS, Ann rev bioch 29 (1960) 45. 37. Wiman, B & Wallen, P, 1973 Eur j bioch 36 (1973) 25. 38. Wiman, B, Eur j bioch 39 (1973) 19. 39. Walthen, PJ, Steinman, HM, Hill, RL, & McKee, PA, J bioI chem 249 (1974) 1173. 40. Summaria, L, Arzadon, L, Bernabe, P, Robbins, KC, J bioI chem 250 (1975) 3988. 41. Blumberg, PM, & Robbins, PW, Cell 6 (1975) 137. 42. Hawkes, SP, J supra mol struc 4, sup 1 (1976) 149. 43. Quigley, JP, J cell bioI 71 (1976) 472. 44. Roblin, R, Chou, IN, & Black, PH, Adv can res 22 (1975) 203.