Ecto-enzymes of the guinea pig polymorphonuclear leukocyte. I. Evidence for an ecto-adenosine monophosphatase, adenosine triphosphatase, and -p-nitrophenyl phosphates.

Abstract A suspension of intact guinea pig polymorphonuclear leukocytes hydrolyzed added ATP, AMP, and p-nitrophenyl phosphate under physiologically appropriate conditions. These enzymatic activities were not due to artifacts such as breakage of the cells during the incubation period. It thus seemed possible that the hydrolyses were being catalyzed by ecto-enzymes, i.e. enzymes on the plasma membrane with their active sites facing the external medium. Three types of experiment were designed to test this hypothesis. First, the activities of intact cells were compared to those of homogenates, sonicates, and cells treated with detergent. Disruption of cells resulted in an approximately 2-fold increase in maximal ATPase and p-nitrophenyl phosphatase activities, suggesting that the plasma membrane was acting as a permeability barrier to the substrates involved. Disruption did not increase AMPase activity, leaving open the possibility that an ecto-enzyme is the only protein in polymorphonuclear leukocytes capable of hydrolyzing AMP. Second, the products of ATPase, AMPase, and p-nitrophenyl phosphatase activities of intact cells were localized by using radioactively labeled substrates. The concentration of inorganic phosphate produced by these reactions was 18 to 100 times greater in the extracellular medium than in the intracellular milieu. This suggests that the substrates are cleaved outside the cells, or that they are cleaved inside and the products are transported out. The latter possibility was militated against by the following experiment. Cells were loaded with inorganic [33P]phosphate, then allowed to hydrolyze substrates labeled with 32P. The distributions of the two isotopes were compared. Almost all of the inorganic [32P]phosphate was found outside of the cells, while 90% of the inorganic [33P]phosphate remained inside. Third, the cells were treated with the diazonium salt of sulfanilic acid, a reagent known not to penetrate into intact erythrocytes. This treatment rapidly and dramatically inhibited the intact-cell ATPase, AMPase, and p-nitrophenyl phosphatase, while lactate dehydrogenase, a soluble cytoplasmic enzyme, was unaffected. Control experiments demonstrated that in sonicates lactate dehydrogenase was as susceptible to inhibition by the diazonium salt as were the other three activities.


L. KARNOVSKY
From the Department of Biological Chemistry, Harvard Medical School, Boston, Massachusetts 02115 SUMMARY A suspension of intact guinea pig polymorphonuclear leukocytes hydrolyzed added ATP, AMP, and p-nitrophenyl phosphate under physiologically appropriate conditions. These enzymatic activities were not due to artifacts such as breakage of the cells during the incubation period. It thus seemed possible that the hydrolyses were being catalyzed by ecto-enzymes, i.e. enzymes on the plasma membrane with their active sites facing the external medium. Three types of experiment were designed to test this hypothesis.
First, the activities of intact cells were compared to those of homogenates, sonicates, and cells treated with detergent. Disruption of cells resulted in an approximately Z-fold increase in maximal ATPase and p-nitrophenyl phosphatase activities, suggesting that the plasma membrane was acting as a permeability barrier to the substrates involved. Disruption did not increase AMPase activity, leaving open the possibility that an ecto-enzyme is the only protein in polymorphonuclear leukocytes capable of hydrolyzing AMP.
Second, the products of ATPase, AMPase, and p-nitrophenyl phosphatase activities of intact cells were localized by using radioactively labeled substrates. The concentration of inorganic phosphate produced by these reactions was 18 to 100 times greater in the extracellular medium than in the intracellular milieu. This suggests that the substrates are cleaved outside the cells, or that they are cleaved inside and the products are transported out. The latter possibility was militated against by the following experiment. Cells were loaded with inorganic [ aaP]phosphate, then allowed to hydrolyze substrates labeled with azP. The distributions of the two isotopes were compared. Almost all of the inorganic [azP]phosphate was found outside of the cells, while 90% of the inorganic [ aaP]phosphate remained inside.
Third, the cells were treated with the diazonium salt of sulfanilic acid, a reagent known not to penetrate into intact erythrocytes. This treatment rapidly and dramatically inhibited the intact-cell ATPase, AMPase, and p-nitrophenyl phosphatase, while lactate dehydrogenase, a soluble cytoplasmic enzyme, was unaffected. Control experiments demonstrated that in sonicates lactate dehydrogenase was as * This work was supported by United States Public Health Service Grant AI-03260 from the National Institute of Allergy and Infectious Diseases. Preliminary reports of parts of the work have appeared (1,2). susceptible to inhibition by the diazonium salt as were the other three activities.
The plasma membrane of cells may contain enzymes whose active sites face the external medium rather than the cytoplasm, and these enzymes are referred to as "ecto-enzymes. " We have recently reviewed the ecto-enzymes of a number of cell types (3), and shall therefore make citations below only as necessary.
As we have previously pointed out (3), the observed fact that particular intact cells act on an external substrate, even one that is commonly believed to be excluded from cells, does not provide sufficient evidence that the protein involved is an ectoenzyme.
The experiments described below provide evidence along rigorous lines of argument to establish the presence of three ectoenzymes on guinea pig polymorphonuclear leukocytes. An accompanying paper (4) discusses the properties of these ecto-enzymes and the possible uses to which they might be put. MATERIALS AND METHODS

Chemicals
Unless otherwise stipulated, all chemicals used were purchased from commercial sources and were of reagent grade. For chemicals of particular importance, the source is given where the substance is mentioned for the first time.
Cells were prepared from guinea pigs of either sex (Charles River Breeding Laboratories, Inc., Wilmington, Mass.) weighing 500 to 1000 g. The animals were fed chow pellets (containing vitamin C) and water ad lib&m.
The cells were quantified either by suspending an aliquot in 3% acetic acid (to lyse the red blood cells present) and counting under the microscope in a Spencer Bright-Line hemocytometer or by protein determination (see below).
Experiments with intact cells were always completed within 4 to 6 hours after harvesting. Polymorphonuclear leukocytes were obtained by a slight modification of the method reported previously from this laboratory (5). The animals were injected intraperitoneally with 30 ml of a sterile solution of 12% (w/v) sodium caseinate (Difco Laboratories, Inc., Detroit, Mich.) in normal saline (0.9% NaCl) and 15 to 20 hours later they were killed by exposure to ether. The peritoneal cavity was opened, the cell suspension was removed with a pipette, and the cavity was rinsed out once or twice with normal saline. After filtration through nylon gauze to remove any hair or fat globules, the cells were spun down at <lo0 X g for 10 min, washed once in Krebs

RESULTS
In considering possible ecto-enzymes of the polymorphonuclear leukocytes, acetylcholinesterase was the first to come to mind. This activity is present as an ecto-enzyme on red blood cells (20) and intact human leukocytes have also been reported to hydrolyze acetylcholine (21). '4s illustrated in Table II, the acetylcholinesterase detectable in guinea pig polymorphonuclear leukocyte preparations was probably due to the few contaminating red blood cells. No special procedures, such as density gradient centrifugation, were employed to reduce this contamination further, because the present purpose was not to ascertain whether polymorphonuclear leukocytes had any ecto-acetylcholinesterase at all, but whether they had high enough levels of such an activity for it to be useful as a plasma membrane marker.
It is clear from Table   II that if intact guinea pig polymorphonuclear leukocytes do hydrolyze acetylcholine, they do so at a rate not greater than, and probably substantially less than, 1 to 2% of the red blood cell activity. Contaminating red blood cells would thus seriously interfere with the use of any such weak acetylcholinesterase as a marker for the leukocyte plasma membrane (cf. Table I).
There seems to be a discrepancy between the present findings and those of Ross and Rosenbaum (21) who assayed acetylcholinesterase by allowing human white cells to hydrolyze acetylthiocholine and subsequently determined free sulfhydryl groups colorimetrically. They apparently did not inhibit the acetylcholinesterase with eserine in order to determine background levels of  gave a value for the guinea pig polymorphonuclear leukocyte that was 4% of the red blood cell activity, per cell; this compares favorably with the results in Table II. a This increased cellular AMPase is not simply due to lysis of the cells by the detergent because 100% of the total cell AMPase is detected when intact cells are assayed (see following paper, Ref. 4).
Upon further investigation it was discovered that intact guinea pig polymorphonuclear leukocytes hydrolyze exogenous ATP, AMP, and p-nitrophenyl phosphate. occurred.
The following considerations make it unlikely that the intact-cell enzymes are simply adsorbed soluble enzymes.

Examination
of Possibility that Enzymatic Activities of Intact Cells Arise from Artijacts or from Contamination-The most obvious artifact that could lead to the hydrolysis of substrates added to intact cells is the presence of broken cells or the breakage of cells under the assay conditions.
The following observations rule out such artifacts in the present study.
1. As will be discussed in more detail below, comparison of intact cell activities with the activities of homogenates or sonicates revealed that about 50% of the total cell ATPase, 55% of the total p-nitrophenyl phosphatase, and 100c% of the total AMPase could be measured with intact cells. If these results are to be attributed to breakage artifacts, a very large percentage of the cells would have to be disrupted.
2. Ninety-nine per cent of the cells, however, excluded trypan blue before being used in an enzyme assay, and 95% of them still excluded this dye after the assay (6-8).
As depicted in Table III, washing the exudate cells 10 times with 50 to 100 volumes of either an ionic (KRP) or chiefly nonionic (0.34 M sucrose containing 10 mM Tris, pH 7.4) buffer solution did not remove significant amounts of the three enzymes. In the case of the ATPase and the p-nitrophenyl phosphatase, the removal of adsorbed enzyme might be disguised by a growing leakiness of the cells caused by repeated centrifugation and resuspension; such leakiness might allow the expression of intracellular enzymes catalyzing the same hydrolyses.
This was ruled out by comparing total cell ATPase and p-nitrophenyl phosphatase activities after disrupting the cells with a detergent (saponin, 1 mg per ml, was found to reveal all of these activities).
Washing 10 times did not alter the total activities.
As illustrated in Table  IV, certain effecters could be used to distinguish between the cellular and exudate fluid enzymes.
In addition, in exudates containing large numbers of cells, as much as 80 to 90% of the total ATPase, AMPase, and p-nitrophenyl phosphatase activities of the exudate were associated with the cells. 3. Catalase is a soluble cytoplasmic enzyme in guinea pig polymorphonuclear leukocytes (10, 25). If cells are disrupted under the assay conditions, collection of the cells by centrifugation after the assay should leave catalase behind in the supernatant.
When this experiment was performed and the catalase of the supernatant fluid was determined (lo), 5y0 or less of the total catalase was found.
Evidence that Hydrolyses of ATP, AMP, and p-Nitrophenyl Phosphate by Intact Polymorphonuclear Leukocytes Are Catalyzed by Ecto-enzymes--It remains to be demonstrated that the substrates added to the assay media are actually cleaved by ectoenzymes on the intact cells and do not simply enter the cells to be hydrolyzed inside. Three types of experiments have been performed. 4. The measured ATPase, AMPase, and p-nitrophenyl phosphatase activities were all linear with time, suggesting that. cell breakage during the course of incubation was not adding appreciably to the activit,ies.
5. The possibility that the intact cell activities studied here are actually due to lysosomal enzymes extruded during the assay period without actual cell breakage (26, 27) is excluded by Item 4 above.
Furthermore, direct measurement revealed that, under the assay conditions for intact cell ATPase, AMPase, and p-nitrophenyl phosphatase, only 3.2% of the total acid phosphatase and 5.8% of the alkaline phosphatase of the leukocytes were released into the medium.
The ATPase, AMPase, and p-nitrophenyl phosphatase activities were also shown not to be associated with the cell nucleus (28). Fig. 1, the 1 mM concentrations of ATP, AMP, or p-nitrophenyl phosphate routinely used in the assays represent virtually saturating levels of substrate for the three intact cell activities.

As shown in
Thus, if disrupted cells were shown to have higher ATPase, AMPase, and p-nitrophenyl phosphatase activities, this would be evidence for a permeability barrier in intact cells which the substrates could not cross by any nonsaturable process, e.g. diffusion.
If the method of disrupting the cells fragmented the plasma membrane but did not seriously damage the membranes of intracellular organelles, this approach would suggest that t.he permeability barrier in question was the plasma membrane.
The exudate fluid itself was found to contain ATPase, AMPase, and p-nitrophenyl phosphatase activities. The levels of these activities and the dilution of the exudate fluid during harvesting, washing, and resuspension of the cells are such that less than 0.1 Ye of the intact cell activities could be attributed to exudate fluid in the assay sample. However, adsorption to the cells might have As shown in Table V, assay of the homogenates prepared by a method that does not seriously damage the permeability barriers of intracellular organelles (10, 28) revealed levels of ATPase and of p-nitrophenyl phosphatase that were about twice the corresponding activities of intact cells*. As also illustrated, these * This increased ATPase was not due to a nonspecific phosphatase, since the 5 mM p-nitrophenyl phosphate used in the assay mixture was sufficient to prevent attack on ATP by nonspecific phosphatases in both disrupted and intact cell preparations.   for fewer values, the average deviation is presented. * Prepared as described under "Materials and Methods." c Saponin, 0.5% (w/v), was found to give maximum activities of Ah4Pase and p-nitrophenyl phosphatase, while 1.0% was optimal for the ATPaae.
homogenate enzymes were not latent, i.e. their action was not significantly increased by sonication or by the addition of detergent, procedures designed to break down remaining permeability barriers.
This lack of latency suggests that these activities cannot be inside lysosomes, mitochondria, or nuclei. It, may be concluded that a permeability barrier to ATP and p-nitrophenyl phosphate exists and that this barrier is probably the plasma membrane.
The data above do not eliminate the possibility that these substrates cross the plasma membrane by a saturable process such as facilitated diffusion or active transport. Active transport of ATP and p-nitrophenyl phosphate would require energy. Iodoacetate at a concentration of 3 x lo-' M has been found to be a very effective energy poison for polymorphonuclear leukocytes (29), and levels of iodoacetate as high as low2 M have been found to have no significant effect on the hydrolysis of ATP and p-nitrophenyl phosphate by intact cells. Finally, Table V demonstrates that there is no significant difference in the AMPase activity of intact cells and homogenates. (The stimulation of AMPase by detergent is discussed in the following paper, Ref. 4.) This result does not provide evidence for or against the existence of an ecto-AMl'ase but leaves open the provocative possibility that an e&o-enzyme may be the only protein in polymorphonuclear leukocytes capable of catalyzing the hydrolysis of AMP.
2. The second series of experiments was directed at localizing the products of ATP, AMP, and p-nitrophenyl phosphate hydrolysis by intact cells. The possible distributions of the products formed by hydrolysis of ATP, AMP, and p-nitrophenyl phosphate by intact cells are as follows.
If the substrate is cleaved inside the cells, then (a) the products may remain inside the cells, (b) the products may passively diffuse out of the cells, or (c) the products may be actively transported out of the cells. If the substrate is cleaved outside the cells, then the situations obverse to (a), (b), and (c) above would pertain. Of course, the products may distribute independently of each other, so that they may end up on different sides of the plasma membrane.
Another possibility, that a product is bound to the cell membrane, is suggested by findings that the y-phosphateof ATP added to intact Ehrlich ascites carcinoma cells can be transferred to serine or to threonine (30, 31). Product bound to the plasma membrane is equivalent in the following experiments to product inside the cells. A final possibility is that the substrate is cleaved while crossing the plasma membrane.
The experimental design was as follows.
A thick cell suspension in KRP was incubated in the presence of 0.5 mM ATP, AMP, or p-nitrophenyl phosphate at 37" for long enough to hydrolyze all of the substrate.
A concentration of 0.5 mM is still nearly saturating for all three enzymes (see Fig. 1) ; but because it, is one-half the normal concentration, the incubation time required for total hydrolysis was substantially reduced.
There was less risk of movement of products into or out, of cells. The substrates were labeled with aP (as usual, only the y-phosphate of ATP was radioactive).
In addition, several experiments were done with ATP and AMP labeled with tritium in the adenosine moiety. The usual concentrations of p-nitrophenyl phosphate were used to prevent hydrolysis of ATP and AMP by the p-nitrophenyl phosphatase.
[14C]Inulin also was present. After an incubation period, an aliquot was removed for the usual determination of the amount, of substrate hydrolyzed. The remaining cells were pipetted into a finely calibrated hematocrit tube and centrifuged for 1 min. The total volume of the hematocrit tube contents was noted. Subsequently, the supernatant was removed and the volume of the pellet was noted. Aliquots of the supernatant and of the pellet (resuspended in a known volume of KRP) were digested in NCS (see "Materials (1 This value was calculated from the figures in the third and fourth columns as follows: ('% s2Pi outside cells/% volume outside cells)/ (ye 32P i inside cells& volume inside cells). and Methods") ; 3H, r4C, and rzP were determined in these aliquots by differential scintillation counting. Inulin was included in the incubation mixture to serve as a measure of the extracellular space present in the pellet (32). The appropriateness of this method was checked as follows. The volume of an individual cell was calculated by subtracting the volume of the extracellular fluid (inulin space) from the total volume of the pellet to give the volume of the cells present.
The protein in an aliquot of the resuspended pellet was compared to the protein content of a known number of cells in order to determine the number of cells in the pellet.
The total volume of the cells was divided by their total number.
Using this result and the formula for the volume of a sphere, the diameter of a single cell was approximated as 9.97 f 0.46 Frn (mean and standard deviation of nine determinations).
This calculated diameter agrees well with values arrived at through morphological studies (33, 34) ; since it depended upon inulin space regarded as extracellular volume, it would appear that inulin is in fact excluded.
From calculations using the percentage of the total incubation mixture and the percentage of the total 32P that was inside cells, Table VI was constructed. It is clear from the table that the substrates are either hydrolyzed outside the cell and the products remain there, or they are hydrolyzed inside the cell and the products are actively transported out.
In an attempt to eliminate the latter possibility the following experiment was performed.
Cells in KRP were allowed to take up inorganic phosphate labeled with 33P (New England Nuclear Corp.) for 30 to 45 minutes at 0". aaP can easily be distinguished from Y' by differential scintillation counting. The loaded cells then were washed free of extracellular inorganic [""PI phosphate and used to repeat the experiments described above. One change was the use of inulin labeled with 3H instead of i4C so that this compound could be distinguished from both @P and 33P. The ATP and AMP used in these experiments were labeled only with "P. If the inorganic [3*P] Table VII arise because the 32P-labeled products of ATP, AMP, and p-nitrophenyl phosphate hydrolysis have never been inside the cells and so have not mixed with the intracellular pool of inorganic phosphate. Thus, localization of the inorganic phosphate produced by hydrolysis of ATP, AMP, and p-nitrophenyl phosphate by intact cells strongly suggests that these reactions are catalyzed by ectoenzymes. The results were less definitive when the localization of the other products was examined.
Since p-nitrophenol has a pK of about 7.15, a substantial amount of this product would be present in the unionized form in KRP (pH 7.4) and would thus be expect.ed to enter cells readily; indeed, most of the released p-nitrophenol was found inside the cells. The use of tritiated ATP and AMP in the above experiments generally gave values for the ratio of the concentration of 3H outside the cells to the concentration inside of 3 or less. The easiest explanation for this finding is that the ADP produced is subsequently hydrolyzed to AMP (the intact cells do exhibit an ADPase activity; see below) \vhich then is hydrolyzed to adenosine, and adenosine is taken up relatively rapidly by the cells (36).
The enzymatic identification and "trapping" of the products of the intact cell ATPase and AMPase (discussed in the next paper, Ref. 4) provide better evidence that the adenosine-containing products of these enzymes are originally localized outside the cells.
3. The third set of experiments designed to show that the intact cell ATPase, AMPase, and p-nitrophenyl phosphatase are ecto-enzymes was based on the reasoning that if a substance u hich does not penetrate into the intact cells can inhibit these enzymes, they are probably located on the plasma membrane.
An initial attempt was made to inactivate the intact cell activities by adding a protease (trypsin, pronase, or papain) to the medium. This approach has provided evidence that erythrocyte acetylcholinesterase is an ecto-enzyme (37). If the polymorphonuclear leukocytes were in suspension, they formed large clumps shortly after the protease was added; so the experiment was performed with cell monolayers, a procedure which effectively eliminated clumping. Table VIII shows that none of the proteases had an impressive effect on the intact cell hydrolases.
These results are apparently not in accord with cytochemical studies on glutaraldehyde-fixed granulocytes reported by North (38). The diazonium salt of sulfanilic acid is well suited to be a nonpenetrating reagent for use with intact cells (18, 39). The diazonium group is highly reactive and forms covalent bonds with many of the functional groups in proteins, including sulfhydryl, amino, and phenolic groups (e.g. Ref. 40) ; the highly charged sulfonic acid group presumably prevents the molecule from entering intact cells. Experiments were performed with monolayers of polymorphonuclear leukocytes because of the extreme speed and ease with which reagents can be removed from these preparations (17).
After incubation with the diazonium salt at 37", the supernatant was decanted from the monolayer plates and they were washed nine times with KRP; the whole procedure took 30 s or less (17). Plates which received the reagent and were then immediately washed had the same ATPase, AMPase, and p-nitrophenyl phosphatase activities as untreated plates; thus, the washing was effective. In addition to being rapid and convenient, this washing procedure almost certainly harms the cells less than repeated centrifugation and resuspension.
Figs. 2 and 3 illustrate the effects of incubating the cell monolayers with 3.5 mM diazonium salt for various lengths of time and with different concentrations of diazonium salt for 30 min. Inhibition of the ATPase, AMPase, and p-nitrophenyl phosphatase of intact cells was rapid and dramatic.
After a 5-min incubation with 3.5 mM reagent, the ATPase and AMPase were inhibited more than 70%, but inhibition of the p-nitrophenyl phosphatase was not as severe at this early time point.
Using a 30-min incubation the concentration of diazonium salt that brought about The specific activity of the diazonium salt could be used to convert counts per min of 8% into numbers of molecules.
The procedure with sonicates was similar, except that the reaction was terminated by addition of 10 mg of bovine serum albumin and 10 ml of 10% trichloroacetic acid. The bovine serum albumin served as carrier protein for the precipitation.
The precipitate was centrifuged down and washed once with 10 ml of 10% trichloroacetic acid before being dissolved in NCS for counting.
a 50% inhibition was about. 0.01 mM for the ATPase and the AMPase and about 0.1 mM for the p-nitrophenyl phosphatase. Incubation for 30 or 45 miu with 3.5 or 0.75 mM diazonium salt turned the cells orange, not an unexpected effect (41), but even these most severe treatments caused only 10 to 20% of the cells on the monolayer to come off during the washing procedure following the incubation.
The precursors of the diazonium salt, sulfanilic acid and NaNOz, had no effect on the enzymes of the intact cells. Fig. 4 represents data concerning penetration of the diazonium salt int,o the cells. During the first 15 min of incubation the reagent was found to react much more rapidly with a sonicate than with a cell monolayer.
Between 30 and 40 min of incubation the rate of reaction with intact cells approached that with the sonicate. The easiest explanation of these results is that the plasma membrane of intact cells acts as a permeability barrier to the reagent molecules; thus, most of the groups capable of reacting with the diazonium salt are initially inaccessible.
After a time changes in the membrane brought about by the attachment of reagent molecules cause this permeability barrier to break down. Berg (18) used a similar analysis to demonstrate that human FIG. 5 (left). Time course of the effect of the diazonium salt of sulfanilic acid on enzymes of sonicates. Sonicates prepared from suspensions of IO8 cells per ml were treated with 3.5 mM diazonium salt for the time period indicated on the abscissa. When the phosphatases were to be assayed, no convenient procedure for terminating the reaction could be found. Instead, short (1 min) assays were performed with the diazonium salt still present. The assays were begun 30 s before and terminated 30 s after the time point being studied. Because of the much higher activity of lactate dehydrogenase in these sonicates, the reaction with diazonium salt could be terminated by a 1200-fold dilution before assay of this enzyme. pNPPase, p-nitrophenyl phosphatase. FIG. 6 (right). Concentration dependence of the effect of the diazonium salt of sulfanilic acid on enzymes of sonicates. Sonicates prepared from suspensions of lo* cells per ml were treated for 30 min with diazonium salt at the concentration indicated on the abscissa. When the phosphatases were to be assayed, no convenient procedure for terminating the reaction could be found. Instead, short (2 to 3 min) assays were performed with the diazonium salt still present.
(See legend to Fig. 5.) erythrocytes remain impermeable to the diazonium salt of sulfanilic acid for about 60 min at 37".
However, it is apparent from Fig. 4 that even though the diazonium salt reacts more slowly with intact cells, the extent of t.his reaction is substantial even at early incubation times. FUIthermore, when the cells were scraped from the monolayer plates, sonicated, and centrifuged at 100,000 x g for 1 hour, about 507, of the total reagent molecules bound to the cells were found in the supernatant fluid.
This was true even at the earliest incubation times, which suggests that some diazonium salt molecules were reacting with soluble cytoplasmic proteins from the very beginning of the incubation and therefore must have entered the cells. It therefore must be shown that the molecules that penetrate into the cells are not the ones responsible for the observed inhibitions of the ATPase, AMPase, and p-nitrophenyl phosphatase exhibited by intact cells.
An intracellular enzyme could be used as an indicator for the extent of reaction between the diazonium reagent and proteins inside intact cells. Lactate dehydrogenase, which is generally a cytoplasmic enzyme, was examined. When guinea pig polymorphonuclear leukocytes were homogenized so as to cause minimal damage to intracellular organelles (40) and the homogenate centrifuged at 100,000 x g for 1 hour, all of the lactate dehydrogenase remained in the supernatant.
Thus, this enzyme is a soluble cytoplasmic prot,ein in these cells.
Figs. 5 and 6 indicate that in sonicates lactate dehydrogenase, ATPase, AMPase, and p-nitrophenyl phosphatase were all about equally susceptible to inhibition by the diazonium salt. There may be some problem in interpreting the results for ATPase and p-nitrophenyl phosphatase since only about 50% of these total sonicate activities are actually due to the corresponding activities exhibited by intact cells (see Table V). However, there seems to be little difference between the inhibition profiles of the various enzymes when sonicates are used. On the other hand, Figs. 2 and 3 demonstrate that in intact cells the soluble cytoplasmic enzyme was much less susceptible to inhibition by the diazonium salt than were the three potential ecto-enzymes.
Thus, after as much as 30-min incubation of cell monolayers with 3.5 mM diazonium salt, la&ate dehydrogenase was only slightly (6.8%) inhibited.
This treatment should have inhibited the ATPase, AMPase, and p-nitrophenyl phosphatase of intact cells to about the same extent if these activities were due to intracellular proteins; however, it inhibited t,hem more than 80%. Similarly, treatment of cell monolayers with 0.35 mM diazonium salt inhibited the phosphatase activities shown by intact cells 60 to 90%, but did not affect lactate dehydrogenase at all. The simplest explanation of these results is that the permeability barrier of the plasma membrane of intact cells restricts the entry of diazonium salt molecules and thereby severely limits the rate of inhibition of intracellular enzymes by this reagent.
The proteins involved in ATPase, AMPase, and pnitrophenyl phosphatase activities of intact cells must have functional groups located outside the permeability barrier of the plasma membrane.
It is possible that the diazonium salt of sulfanilic acid inhibits the hydrolysis of ATP, AMP, and p-nitrophenyl phosphate by intact polymorphonuclear leukocytes simply by preventing the entry of these substrates into the cells. The functional groups of the outer cell surface that are attacked by the reagent might be involved in determining the permeability of the plasma membrane or in transporting the three substrates in. If this is the case, the prediction would be that sonicates prepared from cells that had been incubated with the diazonium salt would have the same levels of ATPase, AMPase, and p-nitrophenyl phosphatase activities as sonicates of untreated cells. Table IX shows that this prediction was not fulfilled.
The data with nonpenetrating reagent do not provide absolute evidence that the active sites of the three enzymes studied are outside the permeability barrier of the plasma membrane. It seems entirely possible that a plasma membrane enzyme with its active site facing the cytoplasm and some other part of the molecule exposed to the external medium could be inhibited by attack on the externally exposed groups.
In fact, this seems to have occurred in a study on human erythrocytes performed by Ohta and his co-workers (42; for review see Ref. 3). The best evidence that the active sites of the ATPase, AMPase, and pnitrophenyl phosphatase of intact polymorphonuclear leukocytes are outside the permeability barrier of the plasma membrane comes from the localization of the inorganic phosphate produced, but in fact the evidence adduced in each of the contents discussed should be taken together. DISCUSSION It is clear from the experiments reported here that guinea pig polymorphonuclear leukocytes have an ecto-ATPase, an ecto-AMPase, and an e&o-p-nitrophenyl phosphatase. It is also clear, from the evidence presented, that virtually all the hydrolysis of exogenous ATP, AMP, and p-nitrophenyl phosphate by intact polymorphonuclear leukocytes is catalyzed by these ectoenzymes.
Before applying the criteria we have used to define ecto-enzymes it had to be demonstrated that the intact cell activities were not simply due to soluble contaminants or to the leakage of Monolayers were incubated with 3.5 mM diaeonium salt for 30 min at 37". The reaction was terminated as described in the text. Control monolayers were of two sorts: one set of controls received no reagent, while the other set was washed immediately after the diazonium salt was pipetted onto them. Some of the monolayers subsequently were assayed for the ATPase, AMPase, and p-nitrophenyl phosphatase of intact cells. Cells were scraped off other monolayers and sonicated; these sonicates were also assayed for ATPase, AMPase, and p-nitrophenyl phosphatase.
I (1 All activities are expressed as a percentage of the activity of untreated cells (Column 1). intracellular enzymes. The importance of eliminating the possibility of such artifacts is well recognized.
In their investigation of ATP hydrolysis by intact Ehrlich ascites carcinoma cells, Wallach and Ullrey (43) showed that there was no leakage of ATPase from the cells into the medium during incubation.
Simlar evidence was provided by Mustafa and his co-workers (44) in their investigation of ATP hydrolysis by intact alveolar macrophages and by Ronquist (45) in his study of the enzymatic activities of intact erythrocytes.
The first indication of the presence of ecto-enzymes on guinea pig polymorphonuclear leukocytes was the detection of AMPase, ATPase, and p-nitrophenyl phosphatase activities with intact cells, and increased activities with the latter two enzymes upon disruption of the cells. Of course, as seen with the ecto-AMPase, no such increase will occur if there is no intracellular enzyme catalyzing the same reaction.
Several other groups of investigators have also used this approach.
Studying Ehrlich ascites carcinoma cells, Wallach and Ullrey (43) found that homogenates had 2.5 times the ATPase activity of intact cells. Sonication of alveolar macrophages resulted in a 60% increase in ATPase activity over that of the intact cells (44). The simplest explanation of such findings, and the first criterion for an ecto-enzyme, is that some cellular structure presents a permeability barrier to substrate molecules and that active sites capable of acting on substrate added to the medium of intact cells lie outside of this barrier. This first criterion was also made use of by Patton and Trams (46) in their investigation of bovine milk fat globules. These globules are surrounded by the plasma membrane of mammary gland cells and were found to exhibit 5'-nucleotidase activity.
There was no increase in this activity if the globules were disrupted.
This suggests that bovine mammary gland cells also have an ecto-5'-nucleotidase that is the only enzyme catalyzing this reaction in these cells as appears true of guinea pig granulocytes. Care was taken in the present study to disrupt the cells in a manner that caused relatively little damage to membranes other than the plasma membrane (28)) and this would strengthen the conclusion that it is the plasma membrane that acts as a permeability barrier to the substrates.
The second criterion involves localization of the products of intact cell enzymatic activity, an approach that also has been used by other investigators (3). ATP added to the medium of Ehrlich ascites carcinoma cells was hydrolyzed but did not affect intracellular levels of inorganic phosphate or acid-labile phosphate (43), suggesting that neither substrate nor product was entering the cells. Ninety-five to ninety-six per cent of the products of intact erythrocyte enzymatic activities were found in the extracellular medium (45). In an older study (47), 96% of the inorganic phosphate resulting from hydrolysis of glucose l-phosphate by rat intestinal loops was recovered in the medium. Furthermore, it was demonstrated that the intra-and extracellular pools of inorganic phosphate had not mixed.
A similar demonstration that the inorganic phosphate produced by ecto-enzyme activity was never inside the cells is provided by the present studies.
The third criterion utilized here was the inhibition of intact cell enzymes by a "nonpenetrating" reagent.
In a similar fashion, Vansteveninck and his co-workers (48) inhibited glucose transport by erythrocytes with a nonpenetrating sulfhydryl reagent. Bender et al. (39) discovered that treatment of intact erythrocytes with the diazonium salt of sulfanilic acid inhibited the ecto-acetylcholinesterase, as well as the facilitated diffusion of glucose. An additional approach introduced by the present study was the use of a soluble cytoplasmic enzyme as an internal indicator for the penetration of reagent molecules. Taken together, the results of the application of these criteria to guinea pig polymorphonuclear leukocytes establish the existence of ectoenzymes in this cell type, one of which the 5'-nucleotidase, appears to be the only enzyme responsible for hydrolysis of 5'-mononucleotides present in this cell.