Cell surface action of thrombin is sufficient to initiate division of chick cells

Cell. Vol. 14. 811-823,

August

1978,

Copyright

0 1978 by MIT

Cell Surface Action of Thrombin Initiate Division of Chick Cells Darrell H. Carney and Dennis D. Cunningham Department of Medical Microbiology College of Medicine University of California, Irvine Irvine. California 92717 Summary Thrombin covalently linked to carboxylate-modified polystyrene beads initiated division of quiescent chick embryo (CE) cells either in medium containing low levels of serum or in serum-free medium. Release of thrombin was monitored by measuring acid-precipitable radioactivity released from 1251-thrombin beads into the medium during incubation with cells. Even if all of the acid-precipitable material released from the beads were active thrombin, it was not sufficient to account for any of the observed cell division, and was lo-30 fold less than the amount necessary to produce the increase in cell number caused by the thrombin beads. Two other kinds of experiments also showed that material released into the medium did not account for the observed initiation of cell division. First, medium taken from cultures incubated with thrombin beads did not initiate cell division when added to new quiescent cultures. Second, in coverslip experiments where populations of cells with and without thrombin beads shared the same medium, only bead-contacted cells divided. Several results suggested that the material which was released from the thrombin beads resulted from cell-associated proteolysis rather than from “leakage” of intact thrombin from the beads. For example, after incubating 1251-thrombin beads with or without CE cells, we were unable to detect any intact thrombin released into the medium. In addition, most of the material released from the beads was acid-soluble and was only released in the presence of CE cells. A few.thrombin beads were endocytosed by CE cells, but they were surrounded by an intact plasma membrane. Thus they did not directly interact with the cytoplasm. The close association of many of the beads with the cell surface and the presence of a few beads in endocytic vesicles made it important to consider the possibility that thrombin might be released from the beads directly into the cells. This possibility was explored using ultrastructural (EM) autoradiography. With this technique (where one grain represented 700900 thrombin molecules), we found that beads inside the cells had approximately the same number of grains as beads not in contact with cells. This suggested that little, if any, additional radioactive material had been released from the beads

is Sufficient

to

which were in contact with the cells. In addition, we were unable to detect any grains in the cytoplasm which could be attributed to released thrombin, even using an amount of 1251-thrombin beads which was 8 fold greater than the amount which produced maximal cell division. Taken together, these results provide direct evidence that thrombin action at the cell surface is sufficient to initiate division of CE cells. Introduction It is currently unclear whether the site of action of polypeptide hormones or growth factors is generally at the cell surface, or whether interaction with intracellular sites is required. This question was previously explored in several systems by the use of Sepharose-immobilized polypeptides (Cautrecasas, 1969; Blatt and Kimm, 1971; Oka and Topper, 1971; Anderson and Melchers, 1972; Greaves and Bauminger, 1972; Frazier, Boyd and Bradshaw, 1973). However, extensive release of active hormone, sometimes in a “superactive” form, has compromised the conclusions of some of these experiments (Davidson, Van Herle and Gerschenson, 1973; Garwin and Gelehrter, 1974; Bolander and Fellows, 1975; Kolb et al., 1975; Topper et al., 1976). Moreover, there is evidence that many peptides, including insulin, nerve growth factor, low density lipoprotein, and various toxins, enter cells by receptor-mediated uptake (for review, see Neville and Chang, 1978). There is also recent evidence that specific receptors exist in the nucleus for both insulin (Goldfine et al., 1977) and nerve growth factor (Andres, Jeng and Bradshaw, 1977). These observations have suggested that certain polypeptide factors might require uptake into cells for expression of their activity. Addition of trypsin or thrombin initiates division of quiescent chick embryo (CE) cells (Sefton and Rubin, 1970; Blumberg and Robbins, 1974; Vaheri, Ruoslahti and Hovi, 1974; Chen and Buchanan, 1975; Cunningham and Ho, 1975; Teng and Chen, 1975; Hovi and Vaheri, 1976; Zetter, Chen and Buchanan, 1976; Carney and Cunningham, 1977) or cells of certain mammalian lines and strains (Burger, 1970; Noonan and Burger, 1973; Kaplan and Bona, 1974; Noonan, 1976; Buchanan, Chen and Zetter, 1976; Chen, Teng and Buchanan, 1976; Pohjanpelto, 1977; Brown and Kiehn, 1977; Carney, Glenn and Cunningham, 1978). By attaching trypsin via a peptide bond to carboxylate-modified polystyrene beads, we have recently shown that trypsin action at the ceil surface is sufficient to initiate division of quiescent CE cells (Carney and Cunningham, 1977). There are, however, several indications that trypsin and thrombin initiate cell

Cell 812

division by somewhat different mechanisms (Carney et al., 1978). In addition, there is evidence that thrombin is internalized by CE cells, and that this internalization relates in some ways to the initiation of cell division. For example, studies on normal and Rous sarcoma virus-transformed CE cells have shown that internalization of thrombin correlates with initiation of cell division (Zetter, Chen and Buchanan, 1977). In addition, studies comparing thrombin and chymotrypsin have shown that chymotrypsin, which does not initiate cell division, is internalized by CE cells to a much smaller extent than is thrombin (Martin and Quigley, 1977). These considerations prompted us to determine whether the internalization of thrombin is a necessary event in the initiation of chick cell division, or whether thrombin action at the cell surface is sufficient to initiate. We now report that thrombin linked to carboxylate-modified polystyrene beads can initiate division of CE cells. This division cannot be explained by the release of thrombin from the beads either into the culture medium or directly into the cells. 1

Results Linkage of Thrombin to Different Bead Preparations In preliminary experiments, we attached lz51thrombin by carbodiimide condensation to carboxylate-modified polystyrene beads and to polystyrene beads substituted with N,N-bis(Paminopropyl)-1,3-propanediamine (nine carbon spacers with terminal amino groups). We also attached lz51thrombin to cyanogen bromide-activated Sepharose beads using standard techniques. After extensive washing, each of these preparations initiated division of CE cells. Analysis of the medium for radioactivity demonstrated that all of the bead preparations released some material into the medium. There was much less release, however, from beads with carboxylate-modified polystyrene thrombin attached by direct peptide linkage. We also found that after direct linkage to the carboxylate-modified polystyrene beads, the immobilized thrombin still retained approximately 40% of its esterase and fibrinogen-clotting specific activity. We therefore used this preparation (hereafter referred to as thrombin beads) in all the following experiments. Initiation of Cell Division by Thrombin Beads When added to cultures of quiescent CE cells in medium containing 0.03% chicken serum, 50 pg of lz51-thrombin beads (approximately 220 beads per cell containing a total of 1.1 pg of ‘*+thrombin) produced a 32% increase in cell number by 24 hr and a 54% increase by 36 hr (Figure 1A). The

I

12 Time Figure Beads

1. Time Course of initiation and by Soluble Thrombin

I

I

24 (hours) of Cell Division

1

I

36 by Thrombin

50 pg of 1*51-thrombin beads (containing 1 .l pg of ‘251-thrombin) or 4 pg of soluble thrombin were added to quiescent cultures of CE cells in 2 ml of DV medium with 0.03% chicken serum. After the indicated incubations, cell number and incorporation of 3Hthymidine were determined as described in Experimental Procedures. (A) Cell number; (B) 3H-thymidine incorporation. (04) soluble thrombin; (O-O) thrombin beads; (A-A) no addition.

addition of soluble thrombin to parallel cultures at a concentration of 2 /*g/ml caused a 60% increase in cell number by 24 hr and a 74% increase by 36 hr (Figure 1A). Adding more thrombin beads to parallel cultures did not significantly increase the amount of cell division. As shown in Figure 1 B, the peak of 3H-thymidine incorporation occurred 12 hr after the addition of either soluble thrombin or thrombin beads. Thus there was no lag in initiation by thrombin beads as might be expected if this initiation were dependent upon the gradual release of thrombin from the beads. The concentration of serum in the medium had little effect on the initiation of cell division produced by addition of these thrombin beads. As shown in Figures 2A and 2B. 200 pg of thrombin beads initiated cell division in cultures of CE cells which were rinsed and maintained for 48 hr in medium without serum. Although in this particular experiment, medium containing 0.03% serum somewhat enhanced the amount of division, in most of our studies, we found no difference be-

Immobilized

Thrombin

Initiates

Cell Division

an

tween the amount of initiation caused by thrombin beads in serum-free or low serum medium. Further increases in serum concentration (up to 1%) had little effect either on the increase in cell number produced by the thrombin beads (Figures 2A and 2B) or on the amount of material released from the beads into the medium (Figure 2C). These results showed that the thrombin beads were acting as mitogens independent of other serum growth factors present in the medium. Quantitation of Material Released from ‘251-Thrombin Beads and Analysis of Its Ability To Initiate Cell Division To determine whether initiation by the thrombin beads could have resulted from released material, we measured the amount of radioactivity released into the medium from 50 pg of ‘9-thrombin beads during 24 hr of incubation with CE cells. After 4 hr of incubation, a total of 27 ng of radioactive material was released, and by 8 hr, 54 ng of this material had been released (Figure 3A). These amounts were well below the levels of soluble thrombin necessary to initiate detectable division of CE cells (Figure 4C). Even by 24 hr, the total amount of radioactive material released into the medium was only 134 ng. This was still less than half the amount of soluble thrombin required to achieve an increase in cell number similar to that observed with this number of thrombin beads (compare Figures lA, 3A and 4C). It should also be noted that of the relatively small amount of material released from the 1251-thrombin beads, most was acid-soluble. For example, of the 134 ng of radioactive material released during 24 hr of incubation with CE cells, only 7 ng were acidprecipitable (Figure 3A). When the same number of 1Z51-thrombin beads were incubated without cells in conditioned medium taken from parallel cultures, there was little acid-soluble or acid-precipitable material released (Figure 38). This indicated that rather than “leakage” of intact thrombin from the beads, most of the material released during incubation with cells was cleaved from the beads by cell-associated proteases. More importantly, it suggested that most of this released material would not be mitogenically active. Indeed, we found that during autodigestion, 1z51-thrombin lost its mitogenie activity before it became acid-soluble. Thus a measure of the maximum mitogenic potential of released material could be obtained by measuring the acid-precipitable radioactivity. Figure 4 shows an experiment in which we measured the amount of initiation caused by ‘*9thrombin beads and the amount of acid-precipitable material released into the medium, and compared these with the amount of cell division initiated by soluble thrombin in parallel cultures of the

SOL

E

B

1;;;;

0

0.25

0.50

II 1 1 I I 0.75

I .oo

Serum (%I Figure 2. Effect of Chicken Serum on Initiation of CE Cell Division by ‘251-Thrombin Beads and on the Release of ‘7 Material from These Beads Cells were plated at a density of 6 x IO4 cells per cm*. After 4 hr, the cells were rinsed, and the medium was changed to DV medium containing the indicated serum concentration. 1251thrombin beads were added after 46 hr, and cell number was determined 24 hr later as described in Experimental Procedures. (A) Cell number afler addition of 200 pg of ‘251-thrombin beads (O-O); cell number on control plates with no addition (O+). (B) Data from (A) expressed as percentage of increase in cell number in ‘Z51-thrombin bead cultures over control cultures (C) Amount of acid-precipitable material released from the ‘Z51-thrombin beads during the 24 hr incubation with cells in DV medium containing the indicated serum concentrations. Error bars in this figure and in Figure 4 represent +- 1SD from the mean of triplicate and duplicate plates, respectively.

same cells. As shown in Figure 4A, addition of lz51thrombin beads to quiescent cultures of CE cells in medium containing 0.02% chicken serum resulted in cell number increases of from 26-30% by 24 hr. Beads to which ovalbumin had been attached under identical conditions did not initiate division, indicating that the initiation was specific for thrombin and not caused by the beads themselves. The amount of material released from the thrombin beads during incubation with cells was almost proportional to the amount of beads added with 50, 100 and 200 pg of thrombin beads releasing by

Cell 814

200 A ”

150

2 \ E 100 0 ‘5 z 5 .-: t i

8’

l -m

/’

/

50 /

5

2 v-

50 PP 100 JLg 200 Thrombin Beads

Jig

O

B 150

.a= ti,

Liz z 100 is g

330 I4 g:v&urnm

No Addition

-0-o-o

--

40'

OE

‘oe30, z"200.5 3! al

50 0

50,

0

Figure 3. Time Course of incubation of ‘V-Thrombin

0 Time

I6 1 hours)

8 0 24

IO *

cr

0:

0

6

12

16

24

Time (hours) I

Release of 1251 Radioactivity during Beads with or without CE Cells

(A) Radioactive material released into the medium during incubation of 50 rg of ‘251-thrombin beads with 5.2 x lo5 CE cells (see Figure 1). (B) Radioactive material released in the absence of cells in conditioned medium taken from parallel cultures of quiescent CE cells. (W+) total radioactivity; (Cl--Cl) acidprecipitable radioactivity.

24 hr 11, 16 and 25 ng of acid-precipitable material, respectively (Figure 48). Addition of up to 60 ng of soluble thrombin to the parallel cultures led to no detectable increase in cell number (Figure 4C). Moreover, 250-300 ng of soluble thrombin were required to equal the amount of cell division caused by the thrombin beads (Figure 4C). Thus even if all of the acid-precipitable material released from the beads were active thrombin, it was not sufficient to account for any of the observed cell division and was lo-30 fold less than the amount necessary to produce the increases in cell number caused by the thrombin beads. Furthermore, if released material were responsible for the observed initiation, then the amount of cell division would be proportional to the amount of material released. That this was not the case (see Figures 4A and 48) further indicated that thrombin on the beads was responsible for initiating cell division. Experiments similar to the one presented in Figure 4 have been carried out a total of 6 times using a wide concentration range of thrombin beads (from as low as 12.5 pg of beads with 0.2 pg of

SOL

t

I5

’

Soluble

60

’

’

’

250 Thrombin

’

’ ’ ’ ’ 4 1000 4000 (ng/Zml)

Figure 4. Initiation of CE Cell Division by ‘ZSI-Thrombin and Analysis of Mitogenic Potential of Released Material

Beads

beads on quies(A) Effect of ovalbumin beads and 12JI-thrombin cent cultures prepared as described in Experimental Procedures (pg quantities represent pg of polystyrene beads). Ovalbumin beads and ‘251-thrombin beads were prepared as described for thrombin beads in Experimental Procedures. ‘2*l-thrombin beads contained 1.2 pg of ‘251-thrombin per 50 pg of polystyrene beads (1.14 x lOa beads). (B) Amount of acid-precipitable material released into the medium from indicated amounts of ‘**I-thrombin beads during the incubation with chick cells: (O+) 50 pg; (O-O) 100 pg; (A-A) 200 pg. (C) Effect of various concentrations of soluble thrombin on parallel cultures of quiescent CE cells.

thrombin to 400 pg of beads with 8.0 pg of thrombin). In these experiments, the amount of acidprecipitable material released from the beads ranged from lo-60 fold less than the amount of

Immobilized 015

Thrombin

Initiates

Cell Division

soluble thrombin required in parallel cultures to initiate the amount of cell division caused by the thrombin beads. In some of these experiments, we also examined material released from the beads into the medium to determine whether any of the released material was actually intact thrombin. Figure 5 shows such a comparison between the Sephadex G-100 elution profile of soluble 1251thrombin and the profiles of material released from ‘*%thrombin beads during incubation with CE cells or during incubation with conditioned medium alone. As shown, the majority of the soluble 1251-thrombin migrated as a single peak (a-thrombin) with a small amount of higher molecular weight dimer and some P-thrombin which represents the first autolytic breakdown product of athrombin (Fenton et al., 1977). Previous studies have shown that as cr-thrombin is degraded to pthrombin, there is a loss of both clotting activity (Fenton et al., 1977) and mitogenic activity (D. H. Carney, J. W. Fenton, and D. D. Ctinningham, unpublished observations). Figure 5 shows that material released during incubation without cells eluted from the column with p-thrombin, and material released during incubation with cells came off as both p- and y-thrombin (the final breakdown product). In each case, there was no detectable intact Lu-thrombin released from the beads. Thus even though we have used total acid-precipitable material as a measure of the mitogenic potential of the released material, it is probable that virtually none of this material was mitogenically active. Coverslip Experiments To insure that the initiation served with thrombin beads

of cell division obwas not a result of

“superactive” peptides released from the beads (as has been described for Sepharose-immobilized peptides) (Topper et al., 1976) or from released material that had lost its radioactive iodine, we also performed experiments which did not rely on measurement of released radioactive material. Some of these experiments simply involved adding medium containing released material to new cultures. Medium was removed from cultures at various times (up to 24 hr) after the addition of thrombin beads. Any beads remaining in this medium were removed by centrifugation, and the medium was then added to cultures of quiescent CE cells. In all cases, the medium without beads was unable to initiate cell division (data not presented) We also examined this problem using experiments in which two populations of cells, one with and one without thrombin beads, were cultured together in the same medium (Table 1). For these experiments, we cultured CE cells on small plastic coverslips and then added thrombin beads to some of the cultures. After 3 hr, the beads adhered firmly to the cells and the coverslips could be relocated into new dishes. As shown in Table 1, coverslips with cells in contact with thrombin beads had cell number increases after 36 hr of at least 60% over control coverslip cultures in the same or separate dishes. Thus cells not directly contacted by thrombin beads were not initiated by material released from the beads. If initiation on the coverslips with thrombin beads were due to released material, then diluting the concentration of the released material should have resulted in less initiation. As shown in Table 1, this was not the case. By placing a single coverslip with thrombin bead-contacted cells into a 60 mm dish, Table 1. Effect of Thrombin Quiescent CE Cells”

Beads

on Coverslip

Cultures

of

0

602

50 Number

Figure 5. Fractionation of ‘Z51-Thrombin from ‘251-Thrombin Beads

Material

8.0 2 1 .a 4.1 * 0.7

Five control coverslips per 60 mm dish

-

5.3 k

One thrombin bead coverslio oer 60 mm dish

+

8.9 2 0.9

Four thrombin bead coverslips and one control coverslip per 60 mm dish

1

do and

+

9 %

$0 Froctlon

Number of Cells per cm* (x lo-“)

Conditions

40 ,"

20

Plus or Minus Thrombin Beads

L

Released

Radioactive profiles of 1 ml fractions from Sephadex G-100 columns (44 cm x 1 cm equilibrated with serum-free DV medium; ‘*Y-thrombin; (O-O) material fractionation at 4°C). (O+) released from 25 mg of ‘Z51-thrombin beads during a 3 hr incubation in DV medium conditioned for 48 hr on CE cells; (A-A) material released from 25 mg of ‘251-thrombin beads during a 36 hr incubation in DV medium with chick cells.

of Culture

0.1

a Secondary CE cells were plated at a density of 5.0 x 10” cells per cm* on 13 mm plastic coverslips (Lux) in 60 mm dishes with 6.0 ml of DV medium containing 2.0% chicken serum and 2.0% tryptose phosphate broth. After 4 hr. the cells were rinsed, and the medium was replaced with DV medium containing 0.03% chicken serum. After 48 hr, 500 pg of thrombin beads were added to some of the dishes. 3 hr later, coverslips with bead-contacted cells and coverslips with cells only were transferred to new 60 mm dishes with fresh DV medium. Cell number was determined 36 hr later as described in Experimental Procedures.

Cell 616

the bead to cell ratio was the same as in the above experiments, but the concentration of any released material was decreased by 4 fold relative to the other coverslip experiments and by 25 fold relative to the previous experiments with cells in culture dishes. As shown, there was no decrease in the amount of initiation by the thrombin beads on these single coverslips. Similar experiments were also performed with dishes placed on rocker platforms to insure adequate diffusion of released material. In these experiments as well, there was no decrease in initiation when single coverslips were cultured in 60 mm dishes and no significant increase in cell number on control coverslips sharing medium with coverslips covered with thrombin beads.

Could Direct Release of Thrombin into CE Cells Account for the Observed Initiation? The above experiments demonstrated that initiation of CE cell division by thrombin beads did not result from material released into the medium. Another possibility, however, was that material could be released from the beads directly into the cells, where it could then interact with intracellular sites and initiate cell division. This possibility was especially important to consider because our previous EM studies with trypsin beads showed that many of the beads were tightly associated with the cell surface, and that after 9 hr of incubation, approximately 4 of 225 beads added per cell were endocytosed by these cells (Carney and Cunningham, 1977). In the present experiments, we found a similar association and uptake of thrombin beads by the CE cells. We also found that like the endocytosed trypsin beads, all of the endocytosed thrombin beads appeared to be completely surrounded by an intact membrane. Thus if the thrombin on these beads were involved in initiation, its action was still on the plasma membrane and not on the cytoplasm. It should also be mentioned that ovalbumin beads, which did not initiate cell division, were also taken up by these cells. Thus the process of bead endocytosis itself did not lead to initiation of cell division. The close association of the beads with the cell surface and the presence of a few beads inside the cells, however, made it important to determine whether material was being released from the beads directly into the cytoplasm. Before looking for direct release, we had to determine whether material released inside the cells would remain there long enough to be detected. We therefore incubated CE cells with soluble 1251-thrombin for various periods of time and monitored the amount of trypsin-insensitive radioactivity remaining in the cells as incubation continued (Figure 6). As shown, after removing the ‘?

12 96-

0

h/ !/mx. -~-~---A*-‘*---,--- ---*-I 6

I 12 Time

Figure 6. Uptake Trypsin-Insensitive

--------. -----_ *m----------~ 1 I8

--. I 24

(hours)

and Retention of Soluble Region by CE Cells

‘251-Thrombin

into a

Soluble Y-thrombin (6.5 x lo5 cpm/rg) was added to cultures of quiescent CE cells in serum-free medium to a concentration of 2 fig/ml. At the indicated times, incubation was interrupted, the cells were rinsed and incubation in serum-free medium lacking ‘V-thrombin was continued. The amount of trypsin-insensitive radioactivity was determined as described in Experimental Procedures. It should be noted that trypsin-insensitive ‘251-thrombin radioactivity has been correlated with the amount of internalized ‘V-thrombin visualized by EM autoradiography (Zetter et al., 1977). Arrows indicate points at which medium was removed and replaced with fresh medium to continue incubation. Dotted lines indicate incubation with soluble ‘251-thrombin.

soluble thrombin at 2, 4 and 6 hr, there was a slight drop in the amount of trypsin-insensitive material, but in each case, there was little further decrease in trypsin-insensitive material with incubations up to 24 hr. Other studies have also shown that internalized (trypsin-insensitive) thrombin can be recovered from CE cells as intact active thrombin after 10 hr (Zetter et al., 1977) or even after 20 hr of incubation (Martin and Quigley, 1977). Thus unlike factors such as epidermal growth factor, which are rapidly degraded and released from cells (Carpenter et al., 1975), thrombin appears to be retained by the CE cells for longer times once it is taken up and the lz51 label is not rapidly released. This indicated that if material were released from the beads directly into the cells, we could expect to find the radioactive peptides or peptide fragments still inside the cells. To look for possible direct release from the beads, we utilized ultrastructural (EM) autoradiography. In these experiments, we incubated CE cells with 400 pg of 1251-thrombin beads for 10 hr and then fixed the cells and prepared them for EM autoradiography. We chose a 10 hr incubation to allow time for several beads to be endocytosed by each cell and also because this time just preceded the peak of DNA synthesis (Figure 18). Thus if thrombin were acting internally to initiate cell division, it should have been inside the cells by this time.

Immobilized 617

Thrombin

Initiates

Cell Division

Figure 7 shows an EM autoradiograph of a CE cell IO hr after the addition of ‘*+thrombin beads. As shown, most of the emulsion tracks or grains were associated with the beads. There were, however, a few grains which appeared to be in the cytoplasm, and some above or below the cells which did not appear to be bead-associated. It was therefore. important to determine how far away from the beads grains could be which still represented bead-bound thrombin, and whether grains observed in the cytoplasm represented released material or whether they could be attributed to random background. Because ‘251-thrombin binds only to the outer edge of the polystyrene beads, we could determine the resolution of this technique by measuring the distance from the center of grains to the edge of beads above cells where released material would be washed away. After measuring the distance from the edges of beads to the center of over 1000 grains, we found that 95% of the ;_

Figure

7. Ultrastructural

._

--:.‘_

Autoradiography

grains were ~0.2 pm from beads and 99% were within 0.3 km (Figure 8). Thus any grains in the cytoplasm of cells farther than 0.3 pm from the edge of a bead should represent released material or background. If the grains in the cytoplasm represented material released directly from the beads, then we might expect beads in contact with or inside the cells to have fewer grains per bead than those which had not interacted with the cells. Table 2, however, shows that beads above the cells averaged 10.61 grains per section, and beads inside endocytic vesicles within the cells averaged 10.16 grains per section. Thus there was little, if any, difference in the number of grains associated with these beads. This measurement therefore suggested that the grains observed in the cytoplasm were not released directly from the beads. To determine more rigorously whether any of the grains observed in the cytoplasm could be attrib-

,

of ‘251-Thrombin

Beads

10 fir after Addition

to Quiescent

CE Cells

It should be noted that some of these beads were endocytosed by the cells, but in all cases, such beads appeared to be completely surrounded by an intact membrane. Preparation of samples and EM autoradiography were carried out as described in Experimental Procedures. In control experiments, we found that after 6-10 hr of incubation with soluble ‘Z51-thrombin. -93% of the cell-associated ‘*? radioactivity was retained during fixation and dehydration. Most of the material that was lost came off during early fixation and was therefore most probably released from the cell surface. Bar = 0.5 pm.

Table 2. Autoradiographic Beads inside and outside 0 e .-C

30 Location

Outside of cells (not contacting plasma membrane)

20

Inside of cells (enclosed in endocytic vesicles)

IO

Figure

6. Resolution

of Bead

0.1

0.2

0.3

Distance

from

Beads (pm)

Grains of Cell9

Associated

with

‘ZSI-Thrombin

Number of Bead CrossSections

Total Number Grains

Average Number of Grains per CrossSection

271

2874

10.61

86

a72

10.16

p 400 pg of ‘Y-thrombin beads (1.76 pg of ‘251-thrombin per 100 pg beads; 5.4 x lo6 cpmlpg thrombin) were added to 35 mm cultures of CE cells brought to quiescence in DV medium with 0.03% chicken serum as described in Experimental Procedures. After IO hr, the cells were fixed and embedded, and thin sections (600-600 A thick) were cut perpendicular to the monolayer from random locations in the dish. These sections were then prepared for EM autoradiography and examined as described in Experimental Procedures.

of EM Autoradiography

Distances from the centers of over 1000 grains to the edge of the nearest bead were measured on micrographs similar to the one shown in Figure 7. The only grains which were scored were those on beads located above cells where any released material would have been washed away.

uted to released material, we quantitated the number of grains in the cytoplasm farther than 0.3 pm from the edge of visible beads and compared this value with the number of background grains in the areas above the cells with approximately the same number of visible beads. As shown in Table 3, after examining random cross-sections of more than 225 cells, we found no difference between the number of grains in the cytoplasm and the number of background grains in comparable areas above the ceils. In addition, we noted that the distribution of grains in the cytoplasm was completely random. There was no localization near beads or in any specific cellular organelles. Thus all of the grains observed in the cytoplasm of these cells could be attributed to random background. It should be noted that in this experiment, we added a large number of 1251-thrombin beads (400 pg) to increase the probability of detecting any released material. In similar experiments with 50 pg of beads (the concentration which produced maximal cell division), we also could not detect any grains in the cytoplasm which could be attributed to released material. Sensitivity of the EM Autoradiography To evaluate the results obtained with the EM autoradiography, it was important to determine how many molecules were represented by each grain. Based on the number of ‘*Y-thrombin molecules per bead and the number of grains per section, we calculated that each grain represented from 700-

900 molecules of thrombin. We also calculated, based on the amount of area examined and the volume of a CE cell, that in the experiment presented in Table 3, we had about a 50% chance of detecting a single grain if 800 molecules had been released inside each cell. Thus if fewer than a few thousand molecules of internalized thrombin could initiate cell division, they would not be detected with this technique. Because of this limit in sensitivity, it was important to determine how many intracellular thrombin molecules would be required to initiate cell division, assuming that intracellular thrombin were capable of causing cell division. Previous studies have shown that after a 1 hr exposure to mitogenic levels of soluble 1251-thrombin, approximately 1 x lo5 molecules are taken up by each CE cell, but this exposure does not lead to a detectable increase in cell number (Martin and Quigley, 1977; Zetter et al., 1977). In our experiments, we incubated CE cells with various concentrations of soluble I*?-thrombin and then compared the increase in cell number after 24 hr with the amount of trypsin-insensitive thrombin internalized by each cell by 10 hr. The 10 hr exposure was chosen because it was just prior to the peak of DNA synthesis (Figure 1B) and it was at the same time that we performed EM autoradiography (Figure 7). As shown in Figure 9, at concentrations of thrombin where 3 x lo5 molecules per cell by 10 hr. Thus even assum-

Immobilized 619

Thrombin

Table 3. Ultrastructural Hr after the Addition CeW

Initiates

Cell Division

Location of Autoradiographic Grains 10 of ‘Y-Thrombin Beads to Quiescent CE

Total Grains Counted

Grains per lOa pm*

5200

1130

Inside of cells farther than 0.3 pm from visible beads (4.6 x lo3 prnz)

243

53

Background in area above (17.3 x lo3 wmz)

976

56

Location

of Grains

In contact

with

bead??

cells

a See Table 2 and Figure 7. b Only beads in contact with or endocytosed’by cells were scored. This number insured that regions of cells examined had the potential of demonstrating released material. ’ Background was determined in areas above cells with numbers of beads comparable to those on or in cells. The background in areas with no visible beads was not, however, significantly different.

ing that intracellular thrombin could cause cell division, these results indicate that the sensitivity of the EM autoradiography used in these studies would allow us to detect levels of thrombin inside the cells which were not able to produce significant cell division. We did not detect any grains in the cytoplasm which could be attributed to released material, even under conditions where we would expect 8 times more release than the amount released from the maximally mitogenic concentration of thrombin beads. Direct release of thrombin from the beads into the cytoplasm therefore did not account for any of the observed increase in cell number. Discussion In this study, we have immobilized thrombin on polystyrene beads and determined that the cell division initiated by this preparation cannot be explained by material released from the beads. These studies therefore provide direct evidence that thrombin action at the cell surface is sufficient to initiate cell division. Moreover, they provide a new approach to address the more general question of whether peptide hormones or other growth factors can elicit their biological activity by action only at the cell surface, or whether interaction with intracellular sites is required. Previous studies attempting to demonstrate the cell surface as the site of action of peptide hormones or growth factors have largely utilized peptides linked by cyanogen bromide to Sepharose beads (Cuatrecasas, 1969; Blatt and Kimm, 1971; Oka and Topper, 1971; Anderson and Melchers, 1972; Greaves and Bauminger, 1972; Frazier et al., 1973). Although in most cases these preparations were biologically active, several laboratories have

I 01 1 ’ IO3 IO4 Trypsin-Insensitive (molecules

I I IO5 IO6 Thrombin per cell 1

Figure 9. Comparison between Initiation of Cell Division by Soluble ‘2sl-Thrombin and the Amount of ‘251-Thrombin Internalized by CE Cells Soluble ‘Z51-thrombin was added to quiescent CE cells in DV medium containing 0.03% chicken serum at concentrations from 15 rig/ml to 2 pg/ml. After 10 hr of incubation, the amount of trypsin-insensitive radioactivity was determined as described in Experimental Procedures. After 24 hr, cell number was determined and expressed as percentage of increase in cell number over controls with no addition.

reported that the biological responses could often be explained by the amount of soluble material released from the beads during incubation with cells (Davidson et al., 1973; Garwin and Gelehrter, 1974; Bolander and Fellows, 1975; Kolb et al., 1975; Carney and Cunningham, 1977). In some cases, Sepharose-immobilized peptide preparations displayed specificities for target tissues which were different from the soluble hormone (Oka and Topper, 1971; Anderson and Melchers, 1972; Greaves and Bauminger, 1972). This suggested that the immobilized factors acting outside the cells were responsible for the activity, since released material would not be expected to show the altered specificity. Topper et al. (1976), however, have now shown that incubating certain Sepharose bead-hormone preparations with other proteins causes the release of “superactive” peptides with altered specificity for their target tissues. Thus both the altered specificity and the observed biological responses to these Sepharose-peptide preparations appeared to result from soluble released material. In our preliminary experiments with Sepharosetrypsin and Sepharose-thrombin, we also found that released material accounted for most of the biological activity of these preparations. In contrast, we found that attachment of these proteases by direct peptide linkage to carboxylate-modified polystyrene beads resulted in a much more stable preparation. This allowed us to demonstrate that trypsin action at the cell surface was sufficient to

Cell 820

initiate division of quiescent CE cells (Carney and Cunningham, 1977). It is noteworthy that Heatley et al. (1977) have recently cross-linked galactose oxidase to latex beads using glutaraldehyde. Although this linkage was not as stable as the direct peptide linkage reported here, the amount of material released accounted for only about 30% of the lymphocyte stimulation caused by a 20 min treatment with the bead-enzyme preparation. In our present experiments, we found that the addition of 50 pg of thrombin beads to cultures of quiescent CE cells in medium containing 0.03% serum caused a 32% increase in cell number over control populations by 24 hr and a 54% increase by 36 hr. The peak of DNA synthesis in these beadtreated cultures occurred 12 hr after bead addition, coinciding with the peak of DNA synthesis in parallel cultures treated with soluble thrombin. A similar initiation of cell division by thrombin beads was observed in populations of CE cells which had been rinsed and incubated for 48 hr in serum-free medium. In addition, increasing the serum concentration in the medium had little effect on the amount of initiation by the thrombin beads. Initiation by the thrombin beads was therefore not dependent upon the presence of serum growth factors. By 8 hr, the total amount of radioactive material released from 50 pg of 1251-thrombin beads into the medium was not enough to account for any of the observed initiation. Even by 24 hr, the total amount of radioactive material released was about 2 fold less than the amount of soluble thrombin required to initiate the same amount of cell division. Thus even if all of the released material were intact thrombin, it could not account for the observed initiation. In these experiments, however, there was no detectable intact thrombin released into the medium during incubation with or without cells, and most of the material released from the beads in the presence of cells (90-95%) was acid-soluble. In the absence of cells, little acid-soluble or acidprecipitable material was released. Thus these results indicate that rather than “leakage” of intact thrombin, released material was cleaved from the beads by cell-associated proteolysis. Since both the clotting activity (Fenton et al., 1977) and the mitogenic activity (D. H. Carney, J. W. Fenton, and D. D. Cunningham, unpublished observations) of thrombin are lost after partial proteolysis, most of this released material should not be mitogenically active. Even if we assume that all of the acidprecipitable material released were active thrombin, the amount released from 50 pg of 12Y-thrombin beads would be 30 fold less than the amount of soluble thrombin required to produce the amount of cell division caused by these thrombin beads. The results from other experiments also indi-

cated that released material could not account for the observed initiation. For example, adding medium from thrombin bead-treated cultures back to new cultures of quiescent cells did not initiate cell division. Moreover, in coverslip experiments in which populations of cells with and without thrombin beads shared the same medium, only beadcontacted cells divided. These studies allowed us to make two conclusions which could not be made from the radioactive data alone. First, there was no detectable release of active material which might have lost its I251 label. Second, release of “superactive” peptides [as reported for Sepharose-immobilized peptides by Topper et al. (1976)] did not account for the observed initiation of cell division. The polystyrene beads used in these studies were 0.93 pm in diameter, small enough to be endocytosed by the cells. Our previous studies with trypsin linked to these beads showed that many of the beads were tightly associated with the cell surface and that of 225 beads added per cell, by 9 hr four were endocytosed by each cell (Carney and Cunningham, 1977). In the present experiments, we found a similar association of thrombin beads with the cell surface and a similar uptake of these beads by the cells. It is important to emphasize that careful examination of cross-sections of internalized thrombin beads by transmission electron microscopy revealed that all the beads appeared to be surrounded by an intact plasma membrane. Thus even if these few internalized beads were involved in the initiation, their action was on the plasma membrane and not on the cytoplasm. The close association of thrombin beads with the cell surface and the presence of a few of these beads in endocytic vesicles made it important to consider the possibility that thrombin might be released directly from the beads into the cells without passing into the culture medium. By incubating CE cells with soluble 1251-thrombin, we found that internalized thrombin was not rapidly degraded and released from the CE cells. Other investigators have also found that after incubation of soluble thrombin with CE cells for 10 hr (Zetter et al., 1977) and 20 hr (Martin and Quigley, 1977), up to 70% of the internalized thrombin was still intact. Thus if material were released directly into the cytoplasm, it should have remained in the cells long enough to be detected. To determine whether there was any direct release, we used ultrastructural autoradiography of cells treated with 12Vthrombin beads. In these experiments, we found that beads inside cells had approximately the same number of grains as those which were not in contact with the cells. This suggested that little, if any, material could be released from beads directly into the cytoplasm. In addition, we counted grains inside cells and grains in equal areas above the

Immobilized 821

Thrombin

Initiates

Cell Division

cells to determine whether any of the grains in the cytoplasm represented released material. We could not detect any autoradiographic grains in the cytoplasm of the CE cells which could not be attributed to random background. This was true even when 8 fold more beads were used than required for maximal thrombin bead initiation. This indicated that there was no direct release of 1251-thrombinderived material into the cytoplasm, and that little, if any, of the material which had been released from the beads into the medium was taken up by these cells. In these EM autoradiographs, each grain represented from 700-900 thrombin molecules. Thus if only a few thousand molecules of internalized thrombin could initiate cell division, we would not detect them. Other studies, however, have indicated that a large number of thrombin molecules can be internalized without initiating cell division. For example, after 1 hr of exposure to mitogenic levels of soluble thrombin, approximately 1 x lo5 molecules are taken up by each cell (Martin and Quigley, 1977; Zetter et al., 1977). Such an exposure, however, is not sufficient to initiate cell division. In our studies, we also found that 1 x lo5 internalized thrombin molecules did not cause significant cell division. It therefore appears that the sensitivity of the EM autoradiography (700-900 molecules per grain) was sufficient to rule out the possibility that the observed initiation resulted from release of thrombin from the beads into the cells. Because our present results show that cell surface action of thrombin is sufficient to initiate cell division, we are examining a number of cell surface proteins for their possible involvement in this process. Several cell surface proteins have been considered as possible mediators of protease-initiated cell division. These include the LETS protein (230,000 daltons) and proteins of 205,000 and 45,000 daltons (Hynes, 1973; Blumberg and Robbins, 1975; Teng and Chen, 1975,1976; Chen et al., 1976; Zetter et al., 1976; Buchanan, Chen and Zetter, 1977). Both the LETS and 205,000 dalton proteins have been shown to be removed by proteases which both do and do not initiate cell division (Teng and Chen, 1975, 1976; Zetter et al., 1976). It therefore appears that these proteins are not negative effector molecules which prevent cell division. We have recently found that soluble ‘*Ythrombin binds specifically to a single affinity class of receptors on the surface of mouse embryo cells (D. H. Carney and D. D. Cunningham, manuscript in preparation). This suggests that initiation of cell division by thrombin may involve interaction with specific cell surface receptors. Experimental

Procedures

Materials Highly purified human thrombin (about was supplied by Dr. John W. Fenton,

3000 NIH units per mg) II. Carboxylate-modified

polystyrene beads (0.93 pm diameter) were purchased from Dow Diagnostics. I-cyclohexyl-3(2-morpholinoethyl) carbodiimide metho-p-toluene sulfonate was purchased from Sigma. Dulbecco-Vogt’s modified Eagle’s medium (DV medium) and tryptose phosphate broth were obtained from Flow Laboratories. Trypsin solution, glutamine, antibiotics and chicken serum were obtained from Gibco. Chemicals for electron microscopy fixation and embedding were obtained from Polysciences and from Ted Pella, Inc. Illford Autoradiographic emulsion L4 was obtained from Polysciences. Cell culture Chick embryo cells were prepared from the body walls of 9 day old chick embryos as described by Rein and Rubin (1968). Primary cells were grown in DV medium supplemented with 2.0% chicken serum and 2.0% tryptose phosphate broth. Penicillin (100 units per ml) and streptomycin (100 fig/ml) were added to all media used in this study. Nonproliferating secondary cultures of these cells were prepared as follows. Primary cultures were subcultured into 35 mm diameter dishes at a density of 6.2 x IO’cells per cm* in DV medium containing 2.0% chicken serum and 2.0% tryptose phosphate broth. Most of the cells attached by 4 hr. The cultures were then rinsed with serum-free DV medium and cultured either in serum-free DV medium or in DV medium containing 0.02% chicken serum. After 48 hr, the cultures were quiescent as judged by cell counts using a Coulter electronic particle counter. The indicated additions were made at this time, and cell number was measured 24 hr later. It should be noted that the small polystyrene beads used for these studies were not counted by the Coulter counter. Measurement of DNA Synthesis DNA synthesis was measured as incorporation of 3H-thymidine into cells as an acid-insoluble product. ‘H-thymidine (Schwartz/ Mann) was added to cultures of CE cells to a final concentration of 5.7 x lo-’ M (2.5 &i/ml). The cultures were incubated for 30 min, the medium was removed and the cells were rinsed 3 times with cold PBS. Cold 10% trichloroacetic acid (TCA) was added for 5 min. followed by four rinses with cold 10% TCA. The cells were then dissolved in 1 N KOH, and ‘H radioactivity was determined in a liquid scintillation counter. Thrombin Bead Preparation Thrombin (500 fig) was added to approximately 25 mg of washed carboxylate-modified polystyrene beads in 50 mM NaPO,, 200 mM NaCl at pH 7.0 (total volume of 0.75 ml) and linked by the addition of water-soluble carbodiimide [I-cyclohexyl-3(2-morpholinoethyl) carbodiimide metho-p-toluene sulfonate] to a final concentration of 0.1 M. After 4 hr of stirring at 4”C, the beads were washed 5 times by centrifugation (10,000 x g for 15 min) through PBS, preincubated for 3 hr at 37°C in conditioned DV medium (medium taken from 46 hr secondary cultures) and washed twice and resuspended in 2.5 ml of PBS. Turbidity measurements were used to determine the number of beads in the final bead preparations. Based on experiments in which ‘x51-thrombin was attached to beads by this procedure, we found that an average of 0.9 pg of thrombin was linked per 50 pg of carboxylate-modified polystyrene beads. The amount of acid-precipitable material released from ‘Ythrombin beads was determined as follows. Plates with medium were rocked back and forth, and all of the medium was removed. Thrombin beads were removed by sedimentation at 10,000 x g for 15 min at 4°C. Aliquots were added to an equal volume of 10% trichloroacetic acid with 0.05% BSA carrier. This procedure maximized the possibility of detecting released material that might be close to the cell surface which had not diffused into the bulk medium. After 16-24 hr at 4°C. precipitates ware collected by centrifugation, dissolved with 1 N NaOH and counted in a gamma counter. lodination Human thrombin

was

iodinated

using

lactoperoxidase

as de-

Cell 822

scribed by Martin et al. (1976). This labeling procedure routinely yielded thrombin with approximately 2.8 x 1Oe cpm/pg, with an iodine to thrombin molar ratio of 0.6:1 .O and with nearly 100% of its mitogenic activity. For certain experiments (including EM autoradiography), the specific activity of thrombin was increased by doubling the amount of ‘?. Measurement of Internalized Thrombin The amount of soluble ‘251-thrombin internalized by CE cells was measured by determining the amount of ttypsin-insensitive radioactivity as described by Zetter et al. (1977). Briefly, cells were incubated at 37°C for 10 min in PBS containing 10 @g/ml trypsin (3X recrystalized, Worthington). They were then removed from the plates and incubated an additional 15 min. The cells were then sedimented at 8730 x g for 2 min in 1.5 ml microfuge tubes. lz51 radioactivity associated with the pelleted cells was measured in a gamma counter. The amount of trypsin-insensitive radioactivity determined by this technique corresponds to the amount of ‘*+thrombin observed inside CE cells by EM autoradiography (Zetter et al., 1977). EM Autoradiography For electron microscopy and EM autoradiography, cells were fixed in situ for 2 min with 2.5% glutaraldehyde in 0.2 M NaPO, (pH 7.4). followed by combined glutaraldehyde (1.25%) and osmium tetroxide (1%) in 0.1 M NaPO, (Trump and Bulger, 1966) for 30 min at 4°C and for 30 min in aqueous uranyl acetate. Cells were then dehydrated in ethanol and embedded in Epon 812 (Luft. 1961). Plastic dishes were removed, and thin sections (600800 A) were cut perpendicular to the cell monolayer. To increase autoradiographic resolution, sections were stained with uranyl acetate and lead citrate prior to overlaying emulsion (Salpeter, Fertuck and Salpeter, 1977). lllford L4 emulsion was applied as a thin film (-0.1 pm thick) by the loop method to sections on formvar-coated grids. Exposure was carried out for 3 weeks at 4°C. The emulsion was then developed in Kodak Microdol. Sections were restained, and then studied and photographed with a Philips EM 300 electron microscope. Pictures magnified 30,00048,000x were used for determining the number of grains per bead and for determining the resolution of this procedure. Grains in background and cytoplasmic areas were scored on the viewing screen as the sections were examined at a magnification of 12.500x. Acknowledgments We would like to thank Dr. John W. Fenton, II, for gifts of highly purified human thrombin, and Mr. Tom Ho for technical assistance. This work was supported by a research grant from the National Cancer Institute. D.D.C. is a recipient of a Research Career Development Award from the National Cancer Institute. D.H.C. was supported by a postdoctoral training grant from the USPHS. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. Received

February

27,1978;

revised

May 22, 1978

.References Anderson, J. and Melchers, F. (1972). Induction of immunoglobulin M synthesis and secretion in bone-marrow-derived lymphocytes by locally concentrated concanavalin A. Proc. Nat. Acad. Sci. USA 70, 416-420. Andres, Ft. Y., Jeng, I. and Bradshaw, R. A. (1977). Nerve growth factor receptors: identification of distinct classes in plasma membranes and nuclei of embryonic dorsal root neurons. Proc. Nat. Acad. Sci. USA 74, 2785-2789.

Blatt, L. M. and Kimm, K. H. (1971). Regulation of hepatic glycogen synthetase: stimulation of glycogen synthetase in an in vitro liver system by insulin bound to Sepharose. J. Biol. Chem. 246, 4895-4898. Blumberg, P. M. and Robbins, P. W. (1975). Effect on activation of resting chick embryo fibroblasts surface proteins. Cell 6, 137-147.

of proteases and on cell

Bolander, F. F.. II and Fellows, R. E. (1975). Growth hormone covalently bound to Sepharose or glass. Analysis of ligand release rates and characterization of soluble radiolabeled products. Biochemistry 14, 2938-2943. Brown, M. and Kiehn, D. (1977). Protease effects growth properties of normal and transformed baby ney cells. Proc. Nat. Acad. Sci USA 74, 2874-2878. Buchanan, J. M., Chen, L. B. and Zetter, related effects in normal and transformed mology, J. Schultz and F. Ahmad, eds. York), pp. l-24.

on specific hamster kid-

B. R. (1976). Proteasecells. In Cancer Enzy (Academic Press: New

Buchanan, J. M.. Chen, L. B. and Zetter. 8. R. (1977). Reaction of thrombin with fibroblasts and mouse splenocytes. In Chemistry and Biology of Thrombin, R. L. Lundblad. J. W. Fenton, II and K. G. Mann, eds. (Ann Arbor, Michigan: Ann Arbor Science), pp. 519-530. Burger, M. M. (1970). Proteolytic enzymes initiating cell division and escape from contact inhibition of growth. Nature 227, 170171. Carney, D. H. and Cunningham, D. D. (1977). cell division by trypsin action at the cell surface. 606.

Initiation of chick Nature 268, 602-

Carney, D. H.. Glenn, K. C. and Cunningham, D. D. (1978). Conditions which affect initiation of animal ccl! division by trypsin and thrombin. J. Cell Physiol. 95, 13-22. Carpenter, G., Lembach, (1975). Characterization growth factor to human 4304. Chen, blood Acad.

K. J., Morrison, M. M. and Cohen, S. of the binding of ‘*‘l-labeled epidermal fibroblasts. J. Biol. Chem. 250, 4297-

L. B. and Buchanan, J. M. (1975). MitJgenic components. I. Thrombin and prothrombin. Sci. USA 72, 131-135.

Chen, L. B., Teng. N. N. H. and Buchanan, icity of thrombin and surface alterations Exp. Cell Res. 101, 41-46. Cuatrecasas, P. (1969). Interaction brane: the primary action of insulin. 450-457.

activity of Proc. Nat.

J. M. (1976). Mitogenon mouse splenocytes.

of insulin with the cell memProc. Nat. Acad. Sci. USA63,

Cunningham, D. D. and Ho, T.-S. (1975). Effects of added proteases on concanavalin A-specific agglutinability and proliferation of quiescent fibroblasts. In Proteases and Biological Control, E. Reich, D. B. Rifkin and E. Shaw. eds. (New York: Cold Spring Harbor Laboratory), pp. 795-806. Davidson, M. 8.. Van Herle, A. J. and Gerschenson, Insulin and Sepharose-insulin effects on tyrosine levels in cultured rat liver Cells. Endocrinology92,

L. E. (1973). transaminase 1442-1446.

Fenton, J. W.. II, Fasco, M. J., Stackrow, A. B., Aronson, D. L., Young, A. M. and Finlayson, J. S. (1977). Human thrombins. Production, evaluation, and properties of a-thrombin. J. Biol. Chem. 252,3587-3598. Frazier, Of nerve petence Sci. USA

W. A., Boyd, L. F. and Bradshaw. R. A. (1973). Interaction growth factor with surface membranes: biological comof insolubilized nerve growth factor. Proc. Nat. Acad. 70, 2931-2935.

Garwin, J. L. and Gelehrter, T. D. (1974). Induction of tyrosine aminotransferase by Sepharose-insulin. Arch. Biochem. Biophys. 164, 52-59. Goldfine. I. D., Smith, G. J., Wong, Cellular uptake and nuclear binding

K. Y. and Jones, A. L. (1977). of insulin in human cultured

Immobilized 823

Thrombin

lymphocytes: action. Proc.

evidence Nat. Acad.

Initiates

Cell Division

for potential intracellular Sci. USA 74, 1368-1372.

Greaves, M. F. and Bauminger, S. (1972). lymphocytes by insoluble phytomitogens. 67-70.

site

of insulin

Activation of T and B Nature New Biol. 235,

Heatley,S. A., Gordon, I. L., O’Brien, Ft. A., Rembaum. A. and Parker, J. W. (1977). Lymphocyte transformation initiated by galactose oxidase coupled to Latex microspheres and formalinized erythrocytes. Exp. Cell Res. 108. 139-149. Hovi. T. and Vaheri. A. (1976). Reversible release of chick embryo fibroblast cultures from density dependent inhibition of growth. J. Cell Physiol. 87, 245-252. Hynes. R. 0. transformation 3170-3174.

(1973). Alteration and by proteolysis.

Kaplan, J. G. and Bona, Cell Res. 88. 388-394.

of cell-surface proteins by viral Proc. Nat. Acad. Sci. USA 70.

C. (1974).

Proteases

as mitogens.

Exp.

Kolb, J. H., Renner, R., Hepp. K. D.. Weiss, L. and Wieland. 0. H. (1975). Re-evaluation of Sepharose-insulin as a tool for the study of insulin action. Proc. Nat. Acad. Sci. USA 74, 248-252. Luft, J. H. (1961). Improvements ods. J. Biophys. Biochem. Cytol.

in epoxy resin 9, 409-414.

embedding

meth-

Martin, 8. M. and Quigley, J. P. (1977). Binding and uptake of thrombin: possible role in the thrombin-induced mitogenesis of chick embryo fibroblasts. In Chemistry and Biology of Thrombin. R. L. Lundblad. J. W. Fenton, II and K. G. Mann, eds. (Ann Arbor, Michigan: Ann Arbor Science), pp. 531-544. Martin, B. M.. Wasiewski. W. W., Fenton, J. W., II and Detwiler, T. C. (1976). Equilibrium binding of thrombin to platelets. Biochemistry 15,4886-4893. Neville, D. tein transport antibodies, and carrier port, 10, A. Press), pp.

M. and Chang. T.-M. (1978). Receptor-mediated prointo cells. Entry mechanisms for toxins, hormones, viruses, lysosomal hydrolases. asialo-glycoproteins, proteins. In Current Topics in Membranes and TransKleinzeller and F. Bronner. eds. (New York: Academic 65-150.

Noonan, K. D. (1976). Role of serum in protease-induced stimulation of 3T3 cell division past the monolayer stage. Nature 259, 573-576. Noonan, division

K. D. and Burger, M. M. (1973). Induction of 3T3 cell at the monolayer stage. Exp. Cell Res. 80, 405-414.

Oka, T. and Topper, Y. J. (1971). Insulin-Sepharose and the dynamics of insulin action. Proc. Nat. Acad. Sci. USA 68, 20662066. Pohjanpelto. P. (1977). Proteases stimulate man fibroblasts. J. Cell Physiol. 91, 387-392.

proliferation

of hu-

Rein, A. and Rubin, H. (1968). Effects of local cell concentrations upon the growth of chick embryo cells in tissue culture. Exp. Cell Res. 49, 666-678. Salpeter, M. M.. Fertuck, H. C. and Salpeter. E. E. (1977). Resolution in electron microscope autoradiography. Ill. lodine125. the effect of heavy metal staining, and a reassessment of critical parameters. J. Cell Biol. 72, 161-173. Sefton. B. M. and Rubin, H. (1970). Release from density dependent inhibition by proteolytic enzymes. Nature 227, 843-845. Teng, N. N. H. and Chen. L. B. (1975). The role of surface proteins in cell proliferation as studied with thrombin and other proteases. Proc. Nat. Acad. Sci. USA 72, 413-417. Teng, N. N. H. and Chen. L. 8. (1976). Thrombin-sensitive surface protein of cultured chick embryo cells. Nature 259, 578-580. Topper, Y. J., Oka. T.. Vonderhaar. B. K. and Wilcheck, M. (1976). An insulin derivative with biological activity greater than that of native insulin. J. Cell Physiol. 89, 647-650. Trump, B. F. and Bulger, R. E. (1966). New ultrastructural characteristics of cells fixed in a glutaraldehyde-osmium tetroxide mixture. Lab. Invest. 15, 368-379.

Vaheri, A., Ruoslahti, E. and Hovi, T. (1974). Cell surface and growth control of chick embryo fibroblasts in culture. In Control of Proliferation in Animal Cells, B. Clarkson and R. Baserga. eds. (New York: Cold Spring Harbor Laboratory), pp. 305-312. Zetter. B. R.. Chen, L. B. and Buchanan, J. M. (1976). Effects of protease treatment on growth, morphology, adhesion, and cell surface proteins of secondary chick embryo fibroblasts. Cell 7, 407-412. Zetter, B. R., Chen, L. B. and Buchanan, J. M. (1977). and internalization of thrombin by normal and transformed cells. Proc. Nat. Acad. Sci. USA 74. 596-600.

Binding chick