Role of a Peptidase in Phagocyte Chemotaxis (N-Formylmethionyl Peptides/Neutrophils/Macrophages) S

Proc. Natl. Acad. Sci. USA Vol. 73, No. 7, pp. 2439-2442, July 1976 Cell Biology

Role of a peptidase in phagocyte chemotaxis (N-formylmethionyl peptides/neutrophils/macrophages) S. ASWANIKUMAR*, E. SCHIFFMANN*, B. A. CORCORAN*, AND S. M. WAHLO * Laboratory of Developmental Biology and Anomalies and t Laboratory of Microbiology and Immunology, National Institute of Dental Research, Bethesda, Maryland 20014 Communicated by DeWitt Stetten, Jr., April 30,1976

ABSTRACT The potencies of N-formylmethionyl (fMet) Mr. Henry Showell of the University of Connecticut Health peptides as chemotactic agents for phagocytes are related to the Center, Farniington, Conn., under Contract no. NO DE5 2477 rates at which they are hydrolyzed. Furthermore, chloromethyl with the National Institute of Dental Research and the National ketones inhibit chemotaxis as do the products of hydrolysis of fMet peptides. The directed migration of cells in response to Institutes of Health, Bethesda, Md.; all amino acid residues in such peptides is probably brought about by the binding of the these peptides were of the L configuration. L-(1-tosylamido- peptide to a cell receptor with subsequent cleavage by peptidase 2-phenyl)ethylchloromethyl ketone (Tos-PheCH2Cl), N-a- specific for aromatic residues, a process that allows the chemical p-tosyl-L-lysylchloromethyl ketone(Tos-LysCH2CI), and di- gradient to be detected. isopropylfluorophosphate were obtained from Calbiochem, Bethesda, Md., and carbobenzoxylamide-2-phenylalanyl Chemotaxis is the directed movement of cells along an in- chloromethane was obtained from Fox Chemical Co., Tucson, creasing chemical gradient. Recently we have observed that Ariz. The chemotactic complement derivative C5a and bac- certain N-formylmethionyl peptides attract phagocytes (1). We terial factor were prepared as described (8, 9). studied these peptides in an attempt to identify the chemotactic Assay for Chemotaxis. Neutrophils and macrophages were substances produced by bacteria. Since bacteria initiate protein taken from rabbit and guinea pig peritoneal exudates, respec- synthesis with N-formylmethionine, whereas animal cells use tively. Chemotaxis of each cell type was assayed as previously methionine, N-formylmethionyl peptides could be a selective described (10, 11) using a micropore filter to separate the chemical signal that phagocytes use to detect and guide their compartment in which the cells and chemotactic substances approach to bacteria. Our studies established that whereas the were placed. Briefly, in the case of rabbit neutrophils, the cells formylation of the a-amino group of certain methionyl di- were allowed to incubate in the modified Boyden chamber for peptides conferred chemotactic activity upon them, their rel- 2 hr at 370 (95% 02-5%CO2), after which the 5 gm Millipore ative potencies varied significantly depending upon the C- filter through which the cells had migrated were stained with terminal amino acid. It- was also found that the active fMet hematoxylin and the cells on the underside were counted. The peptides competed with CSa, a peptide of about 15,000 daltons results were expressed as average number of cells per 10 fields (2) produced from the C5 component of the complement sys- at a magnification of 800 for triplicate samples. Guinea pig tem, an observation suggesting similar receptors on the cells for macrophages were incubated for 90 min under similar condi- both materials. Since previous studies have indicated that tions in humidified air. The 5 ,um polycarbonate filters were neutrophils have esterase-like activity that may be involved in processed in a comparable manner, and 20 oil immersion fields chemotaxis (3-6), it was of interest to determine whether there were quantitated per filter in triplicate. is a relationship between the chemotactic activities of simple Peptidase Assay. Cells obtained from intraperitoneal lavage peptide attractants, and the rates of their hydrolysis by the re- were freed of erythrocytes by exposure to ACKt (12) lysing sponding cell. If this were shown to be the case, it might indicate buffer for 1 min at 00, and by subsequent washing in Gey's a general requirement for the metabolism of a variety of pep- balanced salt solution with 2% bovine serum albumin. Control tide chemotoxins during their stimulation of cell migration. In experiments showed that this treatment did not alter the che- this report we present evidence for the involvement of peptidase motactic responsiveness of leukocytes, and was necessary to activity in phagocyte chemotaxis. reduce the amount of contaminating peptidases from red blood cells. Two milliliters of cells (11 X 106 cells per ml) were incu- MATERIALS AND METHODS bated in Gey's solution together with peptides (1 mM), and were Commercially obtained dipeptides were formylated as de- shaken gently for 30 min at 37'. The suspension was centri- scribed by Sheehan and Yang (7). Met-Phe, Met-Trp, Met-Pro, fuged, and the supernatant was freed of proteins by dilution fMet, p-tosyl-L-arginine methyl ester (Tos-Arg-OMe), N- with alcohol (90% final concentration vol/vol). The supernatant benzoyl-L-tyrosine ethyl ester (Bz-Tyr-OEt), and diethylpy- was subsequently concentrated to a small volume to enable rocarbonate were obtained from the Sigma Chemical Com- characterization of the reaction products by thin-layer chro- pany, St. Louis, Mo.; Met-Asp, Met-Arg, Phe-Met, and Ala-Met matography on silica gel using n-butanol, acetic acid, and water from Research Plus Labs., Denville, N.J.; Met-Tyr, Met-Leu, (4:1:1 vol/vol/vol) as a mobile phase. The cleaved amino acids Met-Ala, Met-Val, and Met-His from Schwarz/Mann, Oran- were visualized after spraying with the cadmium ninhydrin geburg, N.Y. The peptides fMet-Leu-Phe and fMet-Leu-Arg reagent (13) and, in some cases, with the iodoplatinate reagent were prepared by Dr. Richard Freer, Dr. Elmer L. Becker, and for the detection of thioethers (14). The ninhydrin-treated plates were scanned densitometrically Abbreviations: Tos-Arg-OMe, p-tosyl-L-arginine methyl ester; Bz- in a Zeiss spectrophometric scanner at 550 nm. The absorbance Tyr-OEt, N-benzoyl-L-tyrosine ethyl ester; Tos-LysCH2CI, N-a-p- tosyl-L-lysylchloromethyl ketone; Tos-PheCH2CI, L-(1-tosylamido- t Contains per liter: NH4CL 8.29 g; KHCO3 I g; and disodium EDTA, 2-phenyl)ethylchloromethyl ketone. 37.2 mg. 2439 Downloaded by guest on September 25, 2021 2440 Cell Biology: Aswanikumar et al. Proc. Natl. Acad. Sci. USA 73 (1976) Table 1. Effects of protease inhibitors and substrates on chemotaxis Percent inhibition of chemotaxis to*: Cell Additions to cells C5a t Bacterial factor t 0.01 mM fMet-Phe § Neutrophils 0.1 mM Tos-PheCH2Cl 95 ± 2 100 100 ± 3 0.1 mM Tos-LysCH2Cl 75 ± 5 66 ± 1 12 ± 4 0.1mM Bz-Tyr-OEt 42 ± 5 60 ± 1 99 ± 2 0.lmMTos-Arg-OMe 14±6 0±5 18± 2 Macrophages 0.1 mM Tos-PheCH2Cl 91 ± 5 66 ± 3 91 ± 3 0.1 mM Tos-LysCH'Cl 4 ± 0 0 ± 14 12 ± 4 0.1 mM Bz-Tyr-OEt 98 ± 1 90 ± 1 99 ± 2 0.1 mMTos-Arg-OMe 9± 10 0± 30 18± 2 * Results are given as the average of triplicate samples k SEM. Negative control activity (in absence of attractants) was 3 for neutrophils and 7 for macrophages. t Positive control activity was 25 : 4 for neutrophils and 87 4 4 for macrophages. t Positive control activity was 44 i 11 for neutrophils and 52 + 16 for macrophages. § Positive control activity was 31 + 5 for neutrophils and 85 + 8 for macrophages.

of known amounts of peptides and the appropriate amino acids diisopropylfluorophosphate, caused a similar degree of inhi- were measured after chromatography under similar conditions. bition of hydrolysis. Macrophages also cleaved fMet-Leu to Radioactive peptides were incubated with cells as described fMet and Leu (not shown). above, the products were chromatographed in the same man- In Fig. 2 are plotted the minimal effective chemotactic ner, and the chromatograms were scanned for radioactivity concentrations of various fMet peptides against their rates of using a Packard radiochromatogram scanner. cleavage. While an exact correspondence is lacking, in general Assay for Cell Proteolytic Activity. Two milliliters of cells the more chemotactically potent the peptides were, the more (11 X 106 cells per ml) were incubated in phosphate-buffered rapidly they were hydrolyzed. A correlation coefficient of 0.72 saline with either Bz-Tyr-OEt or Tos-Arg-OMe (1 mM) at 370 was calculated for the data. The hydrolysis of nonformylated for 30 min. Substrate hydrolysis was determined essentially as peptides was also studied. Met-Phe was cleaved as rapidly as described by Hummel (15). fMet-Phe, even though it is not chemotactic. Met-Asp was readily cleaved, whereas fMet-Asp was not. Neither of these RESULTS peptides are chemotactic. Apparently the correlation is valid Cleavage of fMet peptides by neutrophils for fMet peptides only. Inhibition of chemotaxis by protease inhibitors and fMet-Phe was degraded at least 30% after incubation with substrates neutrophils to N-formylmethione and phenylalanine (Fig. 1). A similar pattern of cleavage by the cells was seen for other Since the protease inhibitors Tos-PheCH2Gl and Tos-LysCH2Cl peptides such as fMet-Trp, fMet-Leu, and fMet-Tyr. This hy- inhibited hydrolysis of chemotactic peptides by cells, it was also drolysis was inhibited at least 90% when the cells were pre- of interest to determine their effects upon chemotaxis. We treated with chloromethyl ketones, Tos-PheCH2Cl, and Tos- compared the effects of these irreversible inhibitors with dif- LysCH2Cl, inhibitors for chymotryptic and tryptic enzymes, ferent specificities on the chemotactic response of phagocytes respectively (not shown). A serine esterase inhibitor, 2.5 mM to C5a, fMet-Phe, and the bacterial chemotactic factor. The results show (Table 1) that in general the response of each type f-MET-PHE of cell to each of the three attractants was more markedly in-

1E -MET-PHE 5

o20000.2 PHE oU. 8- Z 8TRP LEU

0____510 Uj VAL

100 5 W ~ ARG =2i, 4- 04

ASP 0 PRO 0 0.1 0.2 DISTANCE FFron CLEAVAGE OF PEPTIDES (pM/30 MIN) FIG. 1. Separation of the products formed in the hydrolysis of the chemoattractant l4-omt-aee fMet-Phe (1 mM) by neutrophils. The developed FIG. 2. The relation between the potency of The abbreviations for amino acids chromatogramns were scanned for radioactivity (-), and after reaction and its hydrolysis of neutrophils. refer to the C-terminal residues of fMet dipeptides. The minimal with Cd-ninhydrin reagent for absorbance at 550 nm (--- -). Figure four times that in insert shows results in presence of cells; the other in their ab- chemotactic levels chosen produced a response at least sence. of negative controls. Downloaded by guest on September 25, 2021 Cell Biology: Aswanikumar et al. Proc. Nati. Acad. Sci. USA 73 (1976) 2441 hibited by preincubation for 10 min with Tos-PheCH2G1, an inhibitor of proteases with specificity for aromatic amino acids 00 (16), than by Tos-LysCH2ClI an inhibitor of trypsin-like proteases (17). Neither of these reagents is itself chemotactic, nor was any cytotoxicity observed (Trypan blue exclusion test) at 00o2 the levels employed (0.1 mM). In the neutrophil Tos-LysCH2Cl consistently caused appreciable inhibition of chemotaxis to C5a, 80 in contrast to its effect on the cell response to fMet-Phe, which L +f-MET was much less. The effect of these inhibitors on chemotaxis to 01U (5pM) the bacterial factor appeared to be intermediate to that of the 0 0 140 other two attractants. However, even chemotaxis to fMet-Phe 6 20 was progressively inhibited if the neutrophils were preincu- U. "Is bated with Tos-LysCH2CI for periods longer than 15 min, al- w 20 though the cells were still viable. We also found that Tos- 0. PheCH2Gl, but not Tos-LysCH2Cl, strongly inhibited neutro- 0 phil response to fMet-Arg and fMet-Leu-Arg. In the macro- 00 5 4 3 phage the pattern of inhibition was more consistent. Tos- -LOG [PHE] PheCH2Cl was markedly effective in inhibiting chemotaxis to all the three attractants, whereas had Tos-LysCH2Cl practically 100 no effect. Other protease active site reagents, which have ~.1 specificity for imidazole residues such as carbobenzoxylam- 0 90 b ide-2-phenylalanyl chloromethane and diethylpyrocarbonate, 8 also inhibited chemotaxis to all three substrates. A serine pro- 0 tease inhibitor, 5 mM diisopropylfluorophosphate inhibited tf0 f-MET-PHE chemotaxis to all three attractants by at least 80%. The effects of two model substrates for proteases were 60 1 studied. Neither Bz-Tyr-OEt (12) (specific for chymotryptic enzymes) nor Tos-Arg-OMe (15) (a substrate for tryptic activity) 50 ~ C~ is chemotactic, but Bz-Tyr-OEt in the case of both neutrophils and macrophages was an effective inhibitor of chemotaxis, 0 20 1~ whereas Tos-Arg-OMe had little or no effect on the response I 10 to each of 10 the three attractants (Table 1). In a related study, we +~~~~~~~~11Bobceilfatr4b ;acdtofet-hr(0ial55+5 have found that the whole cells of each type of phagocyte hy- 5 4 3 drolyze Bz-Tyr-OEt (at least 25% in 30 min) but not Tos-Arg- -LOG H4-PHE-MET] OMe. FIG. 3. (a) The inhibition of neutrophil chemotaxis to Met-Phe (10fM) by the products of its hydrolysis by the cells. Control response Effect on chemotaxis to fMet peptides by their initial to Met-Phe was 39t 2. The SEM for each point is given by the bar. cleavage products fMet (5 X 10a M) was a nonchemotactic level (b) The inhibition by phe-Met of neutrophil chemotaxis. Control response to C5a was 43 It is well-known that the products of hydrolysis can reduce the p 11; to bacterial factor, 44d 2; and to fMet-Phe (10tM),c 55 5. rate of hydrolysis by a proteolytic enzyme. Phe alone and, more Negative control was 4 2. r effectively, the combination of fMet and Phe were found to inhibit the response of neutrophils to fMet-Phe (Fig. Sa). Similar results were obtained with the other attractants, fMet-Trp, DISCUSSION fMet-Ala, fMet-Leu, and their respective cleavage products. If chemoattractants were metabolized it would be easier to Formate, methionine, and nonformylated peptides were understand how their receptors on cells are freed to detect the without effect when tested at the same levels. Neither of these chemical gradient. However, bacterial chemotaxis does not latter substances nor Phe (1 mM) affected the chemotactic re- require metabolism of the chemotactic substance (18), whereas sponse to CSa or bacterial factor. We have also found that the in the slime mold hydrolysis of the attractant, cAMP, may be hydrolysis of fMet-Phe by cells was inhibited in the presence a part of the chemotactic response (19). Our results with of 1 mM Phe to an extent of 75%. phagocytes indicate that directed movement of these cells also requires metabolism of the attractant. Prior treatment of cells Effect of the positional isomer of fMet-Phe upon with chloromethyl ketone reagents markedly inhibited both chemotaxis thefr capacities to respond to, and to hydrolyze, chemotactic fMet peptides. Furthermore, a correlation was observed be- To determine the requirements for the location of methionine tween the activities of these peptides and the rates of their hy- in chemotactic dipeptides, we studied the chemotactic prop- drolysis by whole cells. The inhibition of chemotaxis to fMet- erties of fPhe-Met and fAla-Met, positional isomers of the active Phe by its primary cleavage products, Met and Phe, lends compounds fMet-Phe and fMet-Ala, respectively. Both were additional support for the involvement of a peptidase in di- cleaved when incubated with neutrophils, but neither peptide rected cell migration. This inhibition of chemotaxis could be was chemotactic. However, fPhe-Met was found to inhibit the explained by the increase in local levels of attractants, which chemotactic response of neutrophils to fMet-Phe, the bacterial would tend to decrease the gradient or, alternatively, by the factor, and CSa (Fig. Sb), whereas fAla-Met inhibited chemo- possibility that thie products might compete with the attractant taxis to fMet-Ala (not shown). On the other hand, the non- for the chemotactic receptor site. formylated Phe-Met did not have any effect upon the neutro- The peptidase probably has more specificity for aromatic phil response to fMet-Phe. amino acids than for other acids. Peptides with such residues Downloaded by guest on September 25, 2021 2442 Cell Biology: Aswanikumar et al. Proc. Natl. Acad. Sci. USA 73 (1976)

are hydrolyzed most rapidly. Inhibitors and substrates specific would cleave the bound peptide, free the receptor for further for chymotrypsin-like enzymes are markedly more effective reaction, and allow the polarity of the gradient to be detected. than those specific for trypsin-like activity in inhibiting che- Obvious analogies to the acetylcholine-cholinesterase system motaxis to fMet peptides, the bacterial factor, and C5a. This exist. The cleavage of peptides at the receptor may also occur specificity is apparent even in the inhibition of chemotaxis to in other systems, such as in hormonal actions. peptides containing a basic residue such as fMet-Arg and We are indebted to David W. Towne for generous and expert assist- fMet-Leu-Arg. While an increased inhibition of chemotaxis by ance in the densitometric assay of thin-layer chromatograms using the Tos-LysCH2Cl was observed as a function of its time of prein- Zeiss scanning device, and to Dr. Elemer Mihalyi for making available cubation with cells, this could be a reflection of the effects of such facilities in his laboratory at the Laboratory of Biochemistry, chloromethyl ketones upon other cellular functions (20, 21). National Heart and Lung Institute, National Institutes of Health, Be- Goetzl (22) has recently reported that both Tos-PheCH2Cl and thesda, Md. We wish to thank Mr. Steven Schaale, a volunteer at the Tos-LysCH2Cl inhibited chemotaxis to c5a and C3a in human National Institute of Allergy and Infectious Diseases, National Institutes neutrophils. Our results with complement-derived attractant of Health, Bethesda, Md., for his valuable assistance during the summer accord with this, as well as with the early work of Ward and of 1975. Becker (23), who found that esters containing aromatic residues 1. Schiffmann, E., Corcoran, B. A. & Wahl, S. M. (1975) Proc. Nati. were inhibitors of leucotaxis. In a related finding, we were able Acad. Sci. USA 72,1059-1062. to show that diisopropylfluorophosphate inhibited both the 2. Snyderman, R., Phillips, J. & Mergenhagen, S. E. (1970) Infect. chemotactic response to an fMet peptide and its hydrolysis by Immun. 1,521-525. the neutrophil. Although the levels of reagent used were high, 3. Ward, P. A. & Becker, E. L. (1967) . Exp. Med. 125,1001-1020. 4. E. L. & Ward, P. A. (1967) Exp. Med. 125,1021-1030. the viability of the cells was unaffected. This result agrees with Becker, J. 5. Becker, E. L. & Ward, P. A. (1969) J. Exp. Med. 129, 569- the previous studies of Ward and Becker (3) in implicating a 584. serine esterase in leucotaxis. 6. Ward, P. A. & Becker, R L. (1970) J. Immunol. 105,1057-1067. Other evidence for the involvement of a specific peptidase 7. Sheehan, J. C. & Yang,. D. H. (1958) J. Am. Chem. Soc. 80, in phagocyte chemotaxis is that the nonchemotactic fPhe-Met 1154-1158. inhibits taxis to fMet-Phe, its positional isomer, whereas the 8. Shin, H. S., Gewurz, H. & Snyderman, R. (1970) Proc. Soc. Exp. nonformylated Met-Phe and Phe-Met are neither attractants Biol. Med. 131, 203-207. nor inhibitors. Since all four compounds are hydrolyzed by the 9. Schiffmann, E., Showell, H. V., Corcoran, B. A., Ward, P. A., cells, but only fMet-Phe and fPhe-Met affect the chemotactic Smith, E. & Becker, E. L. (1975) J. Immunol. 114, 1831- response, this supports the suggestion that the formylated 1837. peptides interact with chemotactic receptor-associated pepti- 10. Tempel, T. R., Snyderman, R., Jordan, H. V. & Mergenhagen, S. E. (1970) J. Periodontol. 41,3/71-12/80. dase activity, whereas the nonformylated peptides are cleaved 11. Wahl, S. M., Altman, L. C., Lppenheim, J. J. & Mergenhagen, by receptor-independent hydrolytic activity. Presumably, S. E. (1974) Int. Arch. Allergy Appl. Immunol. 46,768-784. fPhe-Met inhibits chemotaxis to fMet-Phe by competitively 12. Roos, D. & Loos, J. A. (1970) Biochim. Blophys. Acta 222, inhibiting hydrolysis of the latter. Also the similar inhibitory 575-582. effect of fPhe-Met upon cell response to fMet-Phe, bacterial 13. Blackburn, S. (1965) in Methods of Biochemical Analysis, ed. factor, or C5a suggests a common site of action for all three Glick, D. (Inter-Science, New York), Vol. XIII, p. 39. attractants in accordance with the previous findings (1) in which 14. Tonnies, G. & Kolb, J. J. (1951) Anal. Chem. 23,823-826. it was shown that these attractants competed with each 15. Hummel, B. C. W. (1959) Can. J. Biochem. Physiol. 37, 1393-1399. other. 16. Ong, E. B., Shaw, E. & Schoellmann, G. (1964) J. Am. Chem. Soc. Our current evidence indicates that formylation of the N- 86, 1271-1272. terminal methionine is required for maximal activity [(24), in 17. Shaw, E., Mares-Guia, M. & Cohen, W. (1965) Biochemistry 4, this report it has been verified that fMet peptides are true at- 2219-2224. tractants]. Hydrophobic and especially aromatic residues in the 18. Adler, J. (1969) Science 166, 1588-1597. next position greatly enhance activity. 19. Malchow, D., Fuchila, J. & Nanjundiah, V. (1975) Biochim. Although not conclusive, we feel that the convergence of the Biophys. Acta 385,421-428. independent lines of evidence developed here strongly suggests 20. Rossman, T., Norris, C. & Troll, W. (1975) J. Biol. Chem. 299, the existence of a stereospecific receptor located in a lipid 3412-3417. 21. Pong, S., Nuss, D. L. & Koch, G. (1975) J. Biol. Chem. 250, membrane. In support of this we have recently found (un- 240-245. published observations) that compounds consisting of fMet-Phe 22. Goetzl, E. J. (1975) Immunology 29, 163-174. coupled to inulin and to a spin-labeled, hydrophobic molecule 23. Ward, P. A. & Becker, E. L. (1968) J. Exp. Med. 127, 693- (the latter synthesized by Drs. Robert Katz and Kenneth Fichter 709. of the Mid-Atlantic Research Institute, Bethesda, Md.) are good 24. Showell, H. J., Freer, R., Zigmond, S., Schiffmann, E., Aswani- attractants for neutrophils. Interaction, then, of peptide and kumar, S., Corcoran, B. A. & Becker, E. L. (1976) J. Exp. Med., receptor initiates movement. The peptidase discussed here 143, 1154-1169. Downloaded by guest on September 25, 2021