Characterization of leukocyte enzymes involved in the release of amino acids in incubated blood cell lysates.

The neutral leukocyte proteases involved in the release of free amino acids in incubation mixtures of blood cell lysates and of human serum albumin with leukocyte lysates were characterized by several techniques including 1H nuclear magnetic resonance spectroscopy, electrophoresis, high performance liquid chromatography, gel filtration, amino acid analysis, and NH2-terminal analysis. The data suggested that the enzymes which contributed significantly to the extensive protein hydrolysis observed were two endopeptidases and three exopeptidases. Identification and analysis of the products obtained in incubation mixtures of human serum albumin with elastase and leucine aminopeptidase were compared with the products obtained in incubation mixtures of the protein with leukocyte lysates.

Classical studies of specific isolated enzyme activities have provided much information about metabolic machinery; however, metabolic maps do not constitute the final picture of metabolism. Understanding the whole demands that individual enzymes are integrated with metabolically related neighbors and with the cellular infrastructure. The construction of more global, and physiologically relevant, models thus requires methodologies with the potential to examine specific enzyme processes in complex, heterogeneous systems. In the preceding study (1) we outlined one such approach, which allowed protease activity to be monitored in incubated blood cell lysates. Similar procedures have been applied now to the characterization of the leukocyte enzymes involved in the proteolysis.
After identifying the resonances observed in the 'H NMR spectrum of incubated blood cell lysates as arising from methyl moieties of free aliphatic amino acids produced as products of putative neutral protease action in leukocytes (l), we investigated the enzyme or enzymes which take part in the hydrolysis. The extensive proteolysis suggested the action of endopeptidases and the fact that considerable amounts of free amino acids were released suggested the involvement of exopeptidases. Another result which needed investigation was the very small number of low molecular weight peptides present after incubations lasting between 15 and 20 h. In the complex mixture of proteins found in hemolysates, it was reasonable to expect that the action of leukocyte proteases would yield a wide gamut of peptides of many different sizes *This work was supported by grants from the National Health and Medical Research Council of Australia. 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.
$Recipient of an Australian Commonwealth Postgraduate Research Award. and compositions, but this was not the case. It could be argued that in blood cell lysates erythrocyte proteases would also contribute to the hydrolysis initiated by the leukocyte enzymes, providing thus an explanation for the thorough proteolysis. But since extensive hydrolysis was also observed in preparations containing only leukocyte lysates and large concentrations (180 mg ml-') of human serum albumin (l), the answer must be found in the properties of the leukocyte enzymes. In the present study we have attempted to establish which leukocyte neutral proteases were the main contributors to the hydrolysis observed in blood cell lysates, and to ascertain whether these enzymes are able to produce the observed proteolysis in the complex mixture of substrates that constitutes a hemolysate.

Methods
Erythrocyte Preparation-Human blood was collected by venipuncture and transferred into tubes containing 3 volumes of physiological saline (0.154 M). Cell suspensions were centrifuged for 10 min at 350 X g, 277 K. The platelet-rich supernatant was decanted, the cells were resuspended (5 volumes) in physiological saline, and centrifuged under the same conditions. After removing the supernatant, cells were washed four times with Krebs bicarbonate buffer (2) with antibiotics, pH 7.4, 277 K, 5 volumes, by centrifuging at 3600 X g for 5 min. A final wash was carried out in Krebs buffer constituted in 2H20 (99.96% 'H, Institute for Nuclear Science and Engineering, Lucas Heights, NSW, Australia) to provide an NMR lock signal for the spectrometer. Blood cell lysates, including the buffy coat, were prepared by twice freezing (78 K ) and thawing the samples. Leukocyte-free erythrocyte cell suspensions were obtained by the method of Beutler et al. (3), filtering whole blood through a column of microcrystalline Sigmagel Type 50 and or-cellulose fiber. The filtered cells were washed in Krebs bicarbonate buffer as above. Erythrocyte lysates were made by twice freezing and thawing the cells. White cell lysates were prepared from fresh concentrates supplied by the Blood Bank. Platelets were removed as above, and the buffy coat was collected and washed five times (room temperature, 5 volumes) in The abbreviations used are: TLCK, Nu-p-tosyl-L-lysine chloromethyl ketone; TPCK, N-tosyl-L-phenylalanine chloromethyl ketone; EGTA, [ethylenebis(oxyethylenenitrilo)]tetraacetic acid; HPLC, high pressure liquid chromatography; PMN, polymorphonuclear; dansyl, 5-dimethylaminonaphthalene-1-sulfonyl. physiological saline with 0.5 mM EDTA and 0.2 mM EGTA. The chelators were removed by twice washing in saline and twice in saline constituted in 'H20. The cells were suspended in a solution of human serum albumin (130 mg ml-') in saline 'Hz0 and lysed by twice freezing and thawing.
Ultrafiltration-Separation of chemical species with different molecular weights was carried out by filtration through YM-2, PM-10, YM-30, or XM-50 membranes in an Amicon 8010 (10 ml) ultrafiltration cell.
Nuclear Magnetic Resonance Spectroscopy-For 'H NMR experiments, 0.65-ml samples were dispensed into 5-mm high-precision tubes (Wilmad). Spectra were observed at 400 MHz with a Varian XL-400 spectrometer operating in the Fourier transform mode. Spectral width was 4,800 Hz over 16,384 data points. The 90" pulse was approximately 15 ps long, and the relaxation delay was 3 s. Each spectrum was the average of 320 transients. Chemical shifts were referenced with respect to the resonance of internal dimethyl sulfoxide at 2.5 ppm. Spin-echo spectra were acquired at 310 K with the Hahn pulse sequence (4) with an echo time of 0.060 s. A standard curve was constructed by addition of valine and leucine amino acids to lysates. The integral of the low frequency methyl resonances of the incubated lysates, relative to the integral of the valine and leucine methyl resonances of the reference, was determined and compared with the standard curve.
Liquid Chromatography-HPLC separations were carried out with an LKB 2150/1/2 instrument (Bromma, Sweden), and CI-10 Milton Roy recording unit (Florida). Samples of 20 pl were injected onto a C-18 reverse-phase column (Lichrosorb, 10 pm, 25 X 0.4-cm, LKB). Separation was achieved by two linear gradients with mobile phases 0.01 M acetic acid, pH 5.2 (A), and acetonitrile made 2% in the same buffer (B). The first step was a 9-min gradient from 1 to 2% B, and the second a 10-min development to 80% B. The flow-rate was 1.0 ml min", and the absorbance was monitored at 210 nm with a 5-pl flow-cell and typical absorbance units full scale of 0.16. Amino acid standards and peptides were supplied by Sigma.
Amino Acid Analysis-A Waters Pico-Tag (Ryde, N. S. W.) system was employed to determine the amino acid composition of samples which had been derivatized with phenyl isothiocyanate.
Leukocyte Counting-Leukocyte counts were estimated electronically using a Coulter Counter (model ZF, Dunstable, United Kingdom). Three drops of Zaponin "Stromatolysing Agent" were added to 10 ml of the diluted sample (1 in 250 or 1 in 500 with Isoton diluent), and after repeated inversion the sample was counted in triplicate. Duplicate dilutions of each sample were counted, and coincidence correction was not used due to the low counts.
Concentration Measurements-Total concentrations of solutes were determined by measuring the osmolality of solutions with a Wescor vapor pressure osmometer model 5100C (Logan, Utah).
NH2-terminal Residue Analysis-Peptides were derivatized with dansyl chloride by the De Jong et al. (9) adaptation of the method of Tapuhi et al. (IO). Peptides were then hydrolyzed for 24 h at 383 K in 6.0 M HCl prior to HPLC. The samples were dissolved in 0.01 M acetic acid, pH 5.0, and applied to a reverse-phase CIS column (@Bondapak, 25 X 0.4-cm) which was developed essentially as described by Weiner and Tishbee (11). The elution of dansyl-amino acids was detected by monitoring the absorbance at 330 nm, and identification was made by comparison with dansyl-amino acid standards (Sigma).

RESULTS
An incubation mixture was prepared by adding white cell lysate (3.2 X lo6 cells ml-') to a red cell lysate (hematocrit 0.85), and adjusting the osmolality of the mixture of 350 mOsm kg". Chemical species were separated by size by filtering aliquots of the mixture through membranes with molecular weight cutoffs ranging from 2 X lo3 to 300 X lo3 nm.
Separations were carried out at 277 K. Under these conditions the amount of hemoglobin that appeared in the filtrates was very small, less than 2% in the fraction separated with the largest membrane pore. The retentate and the ultrafiltrate obtained in each separation were incubated for 20 h at 310 K.
The proteolytic activity was determined by measuring in the 'H NMR spin-echo spectrum of each fraction the integral of the methyl resonances of valine and leucine formed, relative to the integral of the internal dimethyl sulfoxide resonance. No free valine or leucine were detected in the incubated ultrafiltered fractions. The relative amount of product formed in the retentate fractions plotted as a function of membrane pore size is shown in Fig. 1. Full activity was observed after remixing the fractions separated by either the 100 x lo3 or 300 X lo3 M , cutoff membrane. Assuming that hemoglobin is the main substrate in the proteolytic reactions in which free valine and leucine are released, the 20-25% decrease in product formed in the retentate fractions obtained by filtration through membranes of M, cutoff larger than 50 X lo3, was interpreted as the result of separating one or several of the active proteases from the retentate mixture containing the substrate hemoglobin.
The effect of divalent cations on the enzymic action of the proteases was investigated by comparing the amounts of product formed in incubations carried out in the presence of different cations. Two stock incubation mixtures were prepared. One contained a solution of human serum albumin (180 mg m1-l) and white cell lysate (9.5 X lo6 cells ml-'), and the other was made of a blood cell lysate, without platelets, h at 310 K. Fig. 2 shows the 'H NMR spin-echo spectra of the incubated blood cell lysate used as control in these experiments and of lysates incubated in preparations containing MgClz and CaClZ. The resonances corresponding to the methyl moieties of alanine, threonine, valine, and leucine released as products are indicated on the figure. The product concentration in each incubation mixture was normalized by dividing the integral of the corresponding resonance by the integral of an internal dimethyl sulfoxide resonance. Typical product concentrations, obtained in four measurements, of free alanine, threonine, and valine and leucine, relative to that of a control preparation to which no divalent cation was added, are shown in Fig. 3. The "alkaline earths" do not appear to have much effect in the formation of proteolysis products in the incubations of leukocytes with human serum albumin; only the addition of M e produced a 15% increase in activity. In the incubations of blood cell lysates, on the other hand, the presence of Ca2+ stimulates the production of free alanine, and the presence of M F stimulates the production of threonine, valine, and leucine. The presence of Fez+, Co2+, Ni2+, and Cu2+ generally had an inhibitory effect. The strongest inhibitions were that of copper in incubation of leukocytes with human serum albumin and that of iron in the whole-blood cell lysate preparation. The presence of Zn2+ had opposite effects on either type of preparation; it produced a medium degree of inhibition of proteolysis in the white cell/human serum albumin incubation and a strong stimulation of the proteases in the blood cell lysates. CdZf and H e showed various degrees of protease inhibition in both types of preparations, with greater inhibition in leukocyte/human serum albumin preparations than in those of blood cell lysates. Sn2+ and Pb2+ showed a small and a medium degree of proteolysis inhibition, respectively, in leukocyte lysates incubated with human serum albumin.
The specificity of low molecular weight protease inhibitors was employed as an aid to identify the enzymes involved in the release of free amino acids from polypeptides. Preparations of leukocyte lysates (9.5 X lo6 cells ml-') with human serum albumin (130 mg ml-') were incubated with synthetic inhibitors (0.5 mM) for 18 h at 310 K. The effect of the inhibitors was quantified by measuring the amount of valine and leucine released relative to a control sample with no inhibitor present. Typical results obtained in four different measurements are given in Fig. 4, where the inhibitors have been grouped into four categories. The first three correspond to inhibitors of endopeptidases with specificity for thiol, carboxyl, and metallo proteinases, respectively (12). Bestatin, an inhibitor of exopeptidases, forms the fourth category (12). The strongest inhibition was effected by phenylmethanesulfonyl fluoride and by chymostatin, which reduced the release of amino acids by more than 50%. The amount of product formed in the presence of either EDTA or bestatin was about 70% of the product formed in the control preparation. Elastinal decreased proteolysis by about 20%.
The decrease in activity observed in the presence of inhibitors specific for endopeptidases and for exopeptidases provided further evidence that the release of free amino acids was catalyzed by several proteases. We concluded in the previous study (1) that the neutral enzymes involved in proteolysis are localized in the white cell membrane and/or were released into the extracellular medium. This conclusion was explored further by measuring the release of amino acids in incubations of human serum albumin with intact and lysed leukocytes. The results, given in Table I, indicate a greater extent of hydrolysis in incubations with lysates than with intact cells.
Pancreatic elastase was incubated with human serum albumin and with hemolysates to obtain information on the products released from these substrates by the endopeptidase. Data on the products formed from the same substrates by exopeptidases were obtained from incubations with kidney microsomal and cytosolic leucine aminopeptidases. Fig. 5 shows the 'H NMR spin-echo spectra of incubation mixtures of human serum albumin or hemolysates with added microsomal leucine aminopeptidase, with elastase, and with both enzymes together. In the incubations with human serum albumin, leucine aminopeptidase released almost no products. The proteolytic action of elastase yielded more products, mostly in the form of polypeptides, characterized by broad resonances in the 'H NMR spectrum. Narrow spectral lines with low intensity observed at chemical shifts corresponding to free amino acids suggested that small amounts of these molecules were also released as products of elastase. Larger quantities of free amino acids appeared as products in preparations containing both enzymes, but the broader resonances  also observed at slightly different chemical shifts in the NMR spectrum suggested the presence of polypeptides. Amino acids were released in incubations of hemolysates with either leucine aminopeptidase, elastase, or with both enzymes together. The quantitative results of the incubations with hemolysates are given in Table 11. The 'H NMR spin-echo spectrum of the preparation containing leucyl aminopeptidase showed only spectral lines at the chemical shifts corresponding to amino acids, suggesting that the proteins from which the amino acids had been cleaved remained otherwise largely intact. The spectrum of the hemolysate incubated with elastase showed a number of overlapping resonances with linewidths and chemical shifts corresponding to free amino acids and to residues in polypeptides, indicating a more extensive proteolysis which had resulted in cleavage of protein chains into smaller polypeptides, as well as the release of larger amounts of free amino acids than in the presence of only leucine aminopeptidase. In hemolysates incubated with both enzymes, proteins were thoroughly hydrolyzed with the products almost completely in the form of free amino acids. Perusal of the amount of amino acids released in the control hemolysate incubation (Table II) indicated that erythrocyte proteases did not initiate proteolytic processes in these preparations. On the other hand, the difference between the spectra of preparations with human serum albumin and with hemolysates in each incubation could be ascribed to the action of red cell proteases.
The combined effect of endo-and exopeptidases and inhibitors was investigated by measuring the amount of various products released in incubations with human serum albumin (130 mg ml-l). The results after 20 h at 310 K are given in Table 111. The amount of amino acids released was measured from the 'H NMR spin-echo spectra of the incubated samples as the ratio of the integral of the peaks of the methyl moieties of alanine, threonine, and valine and leucine, to the integral of the internal dimethyl sulfoxide resonance. Chymostatin

Relative amounts of product formed (+5%) in incubation of human serum albumin (130 mg ml-I) with intact and lysed leukocyte suspensions (40 X lo6 cells rn1-I) and of red cell lysate (hematocrit 0.85)
with intact and lysed buffy coat suspenswm (6.5 X lo6 cells m1-I) Buffy coats did not include platelets. Cell suspensions were lysed by twice freezing at 78 K and thawing. Samples were incubated for 15 h at 310 K. The amount of amino acids released was measured as the ratio of the integral of the resonance of the methyl moieties of alanine, threonine, and valine and leucine, to the integral of the internal dimethyl sulfoxide resonance. The values were normalized with respect to the number of cells in each suspension before lysis. and Nu-tosyl-L-arginine methyl ester produced between 74 and 86% inhibition of elastase. The release of free amino acids in preparations containing elastase and leucyl aminopeptidase was reduced to 11-25% in incubations with elastinal, and to 41-51% in incubations with bestatin. The presence of both inhibitors together reduced the amount of products formed to 5-12% of the amount released in control preparations. Several techniques were employed to characterize more completely the size and composition of the proteolysis products obtained in these incubations. Each sample was filtered through a membrane that retained molecular species over 50 kDa. The ultrafiltrates were analyzed in 12% Laemmli sodium dodecyl sulfate gels. The products, polypeptides, and amino-  Fig 6 (top); the elution profiles of the products from both incubations are different. The chromatogram of the products of incubations with elastase and microsomal leucine aminopeptidase showed similarities with the chromatogram of incubation products of human serum albumin with white cell lysates (Fig. 5, Ref. 1). The elution profiles of these preparations showed two groups of peaks, one eluting before 8 min and the other between 12 and 17 min. Free amino acids were eluted in the first group Amount of amino acids released in incubations of hemolysates with endoand exopeptidases Hemolysates were prepared by twice freeze-thawing erythrocyte suspensions of 0.85 hematocrit. No enzyme was added to the control sample; the following amounts of enzymes were added to the other preparations: 30 pl carboxypeptidase A (1250 units ml-'), 0.57 mg ml" porcine pancreatic elastase (80 units mg"); 0.86 mg ml" microsomal leucine aminopeptidase (9 units mg"); 0.86 mg ml" cytosolic leucine aminopeptidase (90 units mg"). Samples were incubated for 20 h at 310 K. The amount of amino acids released was measured as the ratio of the integral of the resonance of the methyl moieties of alanine, threonine, and valine and leucine, to the integral of the dimethyl sulfoxide resonance. TABLE I11 Amount of amino acids released in incubations of human serum albumin (130 mg ml") with endoand exopeptidases, and with inhibitors No enzyme or inhibitor was added to the control sample. Enzyme amounts in each preparation were the same as for Table 11; inhibitor concentrations were 0.5 mM. Samples were incubated for 20 h at 310 K. Relative amounts of amino acids were measured as for Table 11. In the spectra of these incubations there was a substantial overlap between resonances arising from amino acids and from peptide residues; consequently the errors in determining the content of free amino acids were as high as 40%. TAME, Nu-p-tosyl-L-arginine methyl ester. of peaks (Fig. 5, Ref. 1). In the second group the two major peaks, O1 and Oz, corresponding to peptides produced in the incubation, appeared a t elution times close to those of the two major peaks, 0 ; and O:, observed in the incubation of human serum albumin with elastase and leucine aminopeptidase. The amino acid compositions of the peptides eluted under peaks 06, O;, and 04, are given in Table IV.
Further characterization of the peptides eluted in the HPLC peaks was carried out by gel filtration. The elution profiles of peak 06, and of peaks 0; and 0 4 are shown in Fig. 6 (bottom).
Separation through a Sephadex G-25 column resolved peak 06 into several main components of molecular masses 5.3, 4.9, 3.7, and 2.05 kDa. The size of the peptide eluted about fraction 54,3.7 kDa, is similar to the size of the largest peptide fragment obtained in incubations with white cell lysates (

DISCUSSION
The methyl resonances observed in the 'H NMR spectra of incubated blood cell lysates and in preparations of intact lysed leukocytes with human serum albumin were identified as arising from products of leukocyte neutral proteases (1). The extensive hydrolysis observed and the variety of amino acids released suggested that several enzymes take part in the proteolysis. The data obtained for retentates separated from blood cell lysates (Fig. 1) indicated that proteases of at least two different sizes participated in the release of amino acids in the incubated samples; one size appeared to be in the range 20-40 kDa and the other over 100 kDa. Endopeptidases in erythrocytes and leukocytes usually have a size range between 25 and 65 kDa (13), whereas exopeptidases are much more complex proteins consisting of several subunits with sizes Leukocyte neutral endopeptidases can be grouped into two types according to the range of substrates that they hydrolyze.

FRACTION NUMBER
Elastase, cathepsin G, and PMN-chymotrypsin are proteases able to digest a wide range of proteins, whereas the action of the other endopeptidases is directed to specific proteins like casein, fibrin, proteoglycans, collagen, gelatin, and elastin. It was shown in the previous study (1) that the leukocyte proteases involved in the release of free aliphatic amino acids in incubations of blood cell lysates also hydrolyzed other proteins like concanavalin A, cytochrome c, hemoglobin, human serum albumin, and lysozyme. Thus, it is unlikely that any of the more specific white cell neutral endopeptidases contribute significantly to the release of free amino acids in incubations with red blood cell lysates or with human serum albumin. In the group of endopeptidases with broad specificity, leukocyte elastase digests elastin and proteins generally, and it hydro-

TABLE IV Amino acid composition of peptides isolated by HPLC
The compositions of peptides in peaks 00, 01, and 02 are given as the molar ratios of residues normalized to either cysteine content, measured as cysteic acid, or relative to Asx. Samples were prepared by incubating different enzymes with human serum albumin or with red blood cell lysates for 20 h at 310 K, and were filtered through Amicon YM-30 membranes. The ultrafiltrates were analyzed by HPLC, and the isolated peaks were lyophilized prior to gas-phase hydrolysis in 6 M HCl at 383 K for 24 h. No correction was made for serine and threonine destruction and tryptophan content was not determined. The values given in parentheses are the rounded figures. Elastase especially cleaves substrates containing valine, alanine, or serine residues (24, 25). Cathepsin G cleaves preferentially at phenylalanyl-, leucyl-, and tyrosyl-X bonds (26), and hydrolyzes a number of proteins including hemoglobin, but not albumin. PMN-chymotrypsin shows preference for aromatic amino acid residue bonds, but also exhibits considerable activity against methionine, leucine, and asparagine esters. It hydrolyzes hemoglobin, and with lesser activity, albumin (16). The data suggest therefore that elastase and PMN-chymotrypsin were the principal endopeptidases involved in the release of free amino acids.
Incubations with divalent cations (Fig. 3) and with low molecular weight inhibitors (Fig. 4) supported this conclusion. Leukocyte elastase has been shown to have enhanced activity at ionic strengths up to 1.0 M NaCl. Its activity is stimulated by MgC12, it is not affected by CaC12, and is inhibited by ZnSOI (27). The activity of cathepsin G is not greatly affected by ionic strength, is mildly enhanced by MgC12, is slightly inhibited by CaCI2, and is unaffected by ZnSOd (28). The enzymatic activity of PMN-chymotrypsin is dependent on ionic strength. In incubations with cations it was observed that Ca2+ ions produced a small activation of the enzyme, and Cuz+ and Zn2+ strongly inhibited it (16). The properties of leukocyte elastase and of PMN-chymotrypsin, but not those of cathepsin G , agree with the observed dependence on ionic strength (1) and with the data for incubations with human serum albumin in the presence of different divalent cations given in Fig. 3.
The low molecular weight inhibitors antipain and leupeptin did not have any effect on the amount of product formed (Fig.   4). Since they inhibit plasmin, trypsin, and cathepsin B (22), which are not found in leukocytes, these results were in agreement with the general conclusion that the proteases involved in the release of amino acids observed in the 'H NMR spectrum of incubated blood cells are located in leukocytes. Similarly, the very small inhibition observed in incubations with pepstatin, a specific inhibitor of some acid proteases (29), is consistent with the conclusion that the enzymes are neutral proteases. The strong inhibition observed in the presence of phenylmethanesulfonyl fluoride (Fig. 4) indicated the participation of serine proteases in the hydrolysis. The lack of inhibition in the incubation mixtures with the halomethyl ketones of lysine (TLCK), and of phenylalanine (TPCK) helped to elucidate which endopeptidases have a significant role in the proteolysis. Cathepsin G is inhibited by both TLCK and TPCK (26). Elastase is inhibited by TLCK but not by TPCK (27), and PMN-chymotrypsin is inhibited by TPCK but not TLCK (16). Consequently, the data suggested that there was no significant contribution by cathepsin G, and the results were consistent with the presence of both neutrophil elastase and PMN-chymotrypsin. The proteolytic action of a chymotrypsin-like enzyme was also supported by the strong effect observed in the presence of the inhibitor chymostatin. The loss of activity measured in the presence of EDTA did not correspond to the catalytic properties of either proteinase. However, a strong inhibition of PMN-chymotrypsin has been reported when Ca2+ and EDTA are both present in the incubation medium (16). Although EDTA inhibits many metalloenzymes, and a collagenolytic metalloproteinase is found in the specific granules of human polymorphonuclear leukocytes (19), it is unlikely that this enzyme is a major contributor to the release of amino acids observed in this work because the presence of phosporamidon, a metalloendopeptidase inhibitor (28), did not affect the amount of products formed. The small degree of inhibition by elastinal indicated the presence of leukocyte elastase. Finally, the lack of inhibition observed in incubations with iodoacetamide and with dithiothreitol was consistent with the proteolytic activity of .
Leucine aminopeptidase is activated by M$+ (30), but Ca2+ and M$+ have little effect on the activity of the other common exopeptidases present in white cells. Thus, the increase in product formation observed in preparations incubated with 2 mM M$+ (Figs. 2 and 3 ) supported the conclusion that leucine aminopeptidase was one of the enzymes catalyzing the release of free amino acids. Soluble arginine aminopeptidase and proline dipeptidase are inhibited by Cu2+, Zn2+, Cd2+, and H e (31, 32). Since proteolysis was reduced in incubations with these ions, the data in Fig. 3 are consistent with the presence of both enzymes. Moreover, soluble arginine aminopeptidase is also inhibited by Pb2+ (31), and a 30% decrease in the amount of product formed was observed in incubations with this divalent cation. The presence of microsomal alanine aminopeptidase was suggested by the inhibition measured in samples with Co2+, which inhibits this enzyme even at very low concentrations (33). Thus, the data from incubations with divalent cations provided evidence for the involvements of the four common leukocyte exopeptidases.
EDTA inhibits soluble arginine aminopeptidase, microsomal alanine aminopeptidase, and proline dipeptidase (31-33) but does not inhibit native (Zn-containing) leucine aminopeptidase (25). The native enzyme is activated and stabilized by M P (2 mM), and the Me-activated leucine aminopeptidase is inhibited by EDTA. The incubations with EDTA were carried out with preparations in which leucine aminopeptidase was in the Mg2+-activated form. Bestatin does not affect proline dipeptidase (32) but is a potent inhibitor of the other three aminopeptidases (30-33). Hence, the data obtained from incubation mixtures with low molecular weight inhibitors supported the presence of three exopeptidases which would preferentially release NHz-terminal leucine, alanine, and arginine residues, but which are also capable of cleaving NH2terminal valine, threonine, phenylalanine, and tyrosine residues.
Information on the products released by the action of endoand exopeptidases was obtained from data of incubation mixtures of porcine pancreatic elastase and/or porcine kidney leucine aminopeptidases with either human serum albumin or hemolysate. Considering that there are important catalytic differences between the porcine and the human enzymes, it was necessary in the first place to establish whether this modeling of the reactions taking place in blood cell lysates would yield results of interest. The 'H NMR spin-echo spectra measured from incubation mixtures of hemolysates with microsomal leucine aminopeptidase, with elastase, and with both enzymes together are shown in Fig. 5. The spectrum of the sample containing the hemolysate and both enzymes (Fig. 5, F ) is qualitatively very similar to the spectrum measured in blood cell lysates (Fig. 2) containing hemolysate and leukocyte lysate. Since it was shown that erythrocyte proteases were not able to initiate the reactions observed in hemolysates containing leukocytes, the results of the incubation of hemolysate with elastase and microsomal leucine aminopeptidase demonstrated that the presence of these enzymes was sufficient to initiate proteolysis. This activity together with that of the erythrocyte proteases, yielded a very similar release of free amino acids as that observed in incubation mixtures with the human leukocyte proteases. The effect of low molecular weight inhibitors on the formation of free amino acids in incubations of human serum albumin with the porcine enzymes was employed as another means of comparing their action with that of the human enzymes. The effect of chymostatin, elastinal, and bestatin on the release of free amino acids summarized in Table 111, indicated that the porcine enzymes were sensitive to the inhibitors affecting the leukocyte proteases.
When human serum albumin was incubated with porcine microsomal leucine aminopeptidase and porcine elastase (Fig.  5, C ) , free amino acids were released, but the spectrum also showed broad resonances indicating the presence of substan-tial amounts of polypeptides. No broad lines, however, were observed in the spectra of human serum albumin samples incubated with leukocytes. These results suggested that in preparations containing leukocytes more than one endopeptidase and/or exopeptidase were involved in proteolysis; this is in good agreement with the conclusions derived from the experiments with divalent cations and inhibitors. The linewidth of 'H NMR resonances is an indicator of the rotational mobility of the chemical groups from which they arise. In the particular case of polypeptides in solution, linewidths depend on internal molecular motions and on the overall molecular tumbling time. Other physical factors like binding to macromolecules, exchange between chemically different environments, or magnetic field inhomogeneities created by the presence of cells or tissue, also affect the width of NMR lines. The sensitivity of resonances to physicochemical conditions has been successfully exploited to obtain information about many biological systems; but it also creates problems in the interpretation of NMR spectra. In this study, HPLC and gel filtration have been employed to provide information on molecular sizes, complementing the unique detailed picture given by NMR measurements.
In the identification of methyl resonances in the 'H NMR spectrum of incubated blood cell lysates (1) it was found that the largest products of incubation of human serum albumin with leukocyte lysates were peptides with molecular masses of 3.7, 3.6, and 2.05 kDa. The extent of the proteolysis, and the number and low molecular weight of the peptide products were unexpected results. Analysis of the incubation products of human serum albumin with the porcine proteases provided an indirect means to verify whether the action of the leukocyte enzymes was sufficient to account for the observed strong protein degradation. The peptides resulting from incubations of human serum albumin with elastase and with elastase and leucine aminopeptidase were separated by HPLC (Fig. 6, top) and their size determined by gel chromatography (Fig. 6,  bottom).
The amino acid analyses of each of the major peptide peaks (Oh, O:, and 0;) obtained in the chromatograms of both incubations are given in Table IV. Comparison of the compositions of the peptides in peaks 01, and 0 2 , obtained in incubations with leukocyte lysates (Table 11, Ref. 11, with the composition of the peptides in peaks 0; and 04, obtained in incubations with elastase and leucine aminopeptidase (Table  IV), revealed a remarkable similarity; only the content of glycine, threonine, alanine, isoleucine, and leucine was significantly different. A final characterization of these peptides was made by determining the first residue in each peptide by NHn-terminal analysis ( Table V). The most important result was that in the incubations with elastase and leucine aminopeptidase only five different peptides were identified with molecular masses between 3.5 and 5.9 kDa, one of which had the same size and NHn-terminal residue as a peptide obtained in incubation with leukocyte lysates. These comparisons suggested that, in all probability, the free amino acids and the limited number of peptides detected in incubations of leukocytes with human serum albumin and with hemolysates resulted from an extensive proteolysis by more than one endopeptidase and one exopeptidase; the endopeptidases were most probably neutrophil elastase and PMN-chymotrypsin, and the exopeptidases were most probably leucine aminopeptidase, soluble arginyl aminopeptidase, and microsomal alanine aminopeptidase.
These conclusions are consistent with the data presented and with the interpretation of results that has been followed. However, it should be made clear that considering the com-plexity of the systems studied, and the qualitative nature of some of the data presented, the interpretation given is probably not unique, and the participation of other enzymes in the proteolysis has not been ruled out.