faecalis var. liquefaciens on Arginine and Its

Transcription

1 JOURNAL of BACmIoLY, Jan. 1969, p American Society for Microbiology Vol. 97, No. I Printed In U.S.A. Dependence of Protease Secretion by Streptococcus faecalis var. liquefaciens on Arginine and Its Possible Relation to Site of Synthesis1 IVAN A. CASAS AN LEONARD N. ZIMMERMAN Department of Microbiology, The Pennsylvania State University, University Park, Pennsylvania Received for publication 16 September 1968 Washed cells of Streptococcus faecalis var. liquefaciens, when harvested from media that supported protease biosynthesis, continued to release this extracellular enzyme in phosphate buffer. The addition of ethyleneiaminetetraacetic acid (EDTA) halted the secretion. If zinc ions were added to the EDTA-treated cells before 45 to 60 min had elapsed, a fraction of the anticipated enzyme activity was observed. After 60 min, arginine and phosphate, in addition to zinc, were necessary for the demonstration of proteolytic activity. The enzyme that was released was newly formed, because chloramphenicol, puromycin, or actinomycin D prevented its appearance. Energy for this synthetic reaction was obtained, apparently, from arginine; lactose could not be substituted for arginine. This last point is interpreted to mean that extracellular protease biosynthesis occurs in a localized area or cellular compartment into which the adenosine triphosphate derived from the fermentation of lactose cannot diffuse. No evidence was found for a protease zymogen, although this possibility is not completely precluded. The site of synthesis and method of release of extracellular enzymes of bacterial origin have commanded a large measure of attention in recent times (15). One aspect of the problem that requires explanation can best be illustrated with the extracellular protease of Streptococcus faecalis var. liquefaciens (6). This particular enzyme is capable of hydrolyzing cytoplasmic proteins (unpublished data). Under these circumstances, any active enzyme that is synthesized on ribosomes in the cytoplasm must be masked physically by being synthesized in some cellular compartment (as in the membrane or a mesosome) or must be masked functionally as a precursor. Shugart and Beck (22) could not demonstrate the presence of active protease in the cytoplasm of this organism. They concluded that protease was released from the wall or plasma membrane. This notion is consistent with the data of Hammel and Zimmerman (7), who showed that protoplasts could not make or release protease. On the other hand, Liu and Elliott (16) were able to show that the extracellular protease of group A streptococci is produced first as a zymogen, which is then converted to its X Authorized for publication on 13 May 1968 as paper no in the journal series of the Pennsylvania Agicultural Experiment Station. active form by reduction in the cell wall or by autocatalysis. The work described in this paper is concerned with the mechanism of secretion of protease by group D streptococci. MATERIAIS AND METHODS Organism. S. faecalis var. liquefaciens strain 31 was used throughout these studies. Stock cultures were maintained in litmus milk at -17 C. Temperature. All experiments, including incubation of cultures, were carried out at 37 C. AU centrifugations were at 5 C. Media. Cells were normally grown in AC broth (20), in which they produce little, if any, protease. When cells that were actively making protease were required, the modified B medium of Hammel and Zimmerman (7) was used. Both are complex media, with one sigdificant difference being the 0.4% arginine in modified B medium. Preliminary treatment of cells. A 2-liter amount of AC broth was inoculated with 20 ml of a culture grown in AC broth. After 14 to 16 hr of incubation, cells were harvested by centrifugation, washed twice in tris(hydroxymethyl)aminomethane (Tris)-chloride buffer (5 X 10- M, ph 7.0), and resuspended in 500 ml of modified B medium. After 2.5 hr of incubation, cooled ethylenediaminetetraacetic acid (EDTA) was added to the cell suspension to a final concentration of 307

2 308 CASAS AND ZIMMERMAN J. BACrERIOL. 5 X 10-3 M. The cells were cooled in an ice bath, harvested, and washed twice in Tris-chloride buffer containing S X 10-4 M EDTA. These cells were then suspended as specified for further utilization. Release cells. After some experimentation, a procedure was developed in which the bacteria secreted protease under controlled conditions. These cells were called release cells and were obtained by suspending all the cells (prepared as described in the preceding paragraph) in 75 ml of Tris-chloride (5 X 10-' M, ph 7.0) and EDTA (5 X 10-4 M), and incubating them for 1.5 hr. Cells were harvested by centrifugation and washed twice in basal solution (7.5 X 10-2 M KH2PO4, SX 104 M EDTA, and 5 X 10-4 M ZnSO4 ; ph 7.0 as adjusted with 5 N NaOH). The cell yield (from the original 2 liters of AC broth) at this point was about 500 to 600 mg on a dry weight basis. Test samples were made to contain 35 to 45 mg (dry weight) of these release cells suspended in 5 ml of basal solution and other chemicals as indicated. After 1.5 hr of incubation (unless otherwise stated), the bacteria were removed by centrifugation and the supernatant fluid was assayed. Protease assay. Anson's method (1) was used to estimate proteolytic activity. To a 2% vitamin-free casein solution (6) was added an equal volume (normally 1.5 ml) of the supernatant fluid to be assayed. The tubes were incubated for 1.5 hr, and the reaction was stopped with 0.6 M trichloroacetic acid; 5 ml of trichloroacetic acid was used for each 3 ml of enzymesubstrate mixture. The contents of the tubes were filtered through Whatman no. 2 filter paper, and the solubilized aromatic amino acids in the filtrates were assayed colorimetrically (3). Appropriate zero-time controls were run by adding the trichloroacetic acid to the substrate prior to the addition of the enzyme. Activity is expressed in terms of micrograms of tyrosine solubilized per 1.5 hr per milliliter of filtrate. Lactose assay. Residual lactose in the supernatant fluid of test samples was determined with anthrone reagent (17). Sonic disruption. Cells prepared according to the preliminary treatment procedure were suspended in 10 ml of Tris-chloride buffer and exposed to a sound field of 20,000 cycles/sec for 40 min at 2 to 4 C in an MSE sonic oscillator. Warburg technique. The ability of the cells to utilize arginine, or its intermediates citrulline and carbamyl phosphate (11), as measured by COa evolution, was determined by standard techniques of Warburg respirometry (24). To the main compartment of the flask was added 2.7 ml of basal solution which contained S X 10' M arginine, citrulline, or carbamyl phosphate. A 0.3-ml sample of release cells was put in the side arm. Another sample of release cells was boiled and used as a control with carbamyl phosphate because of the relative instability of this compound (12). Endogenous CO2 was subtracted from all readings. RESULTS Initial experiments were designed to detect the possible presence of a protease precursor. Bleiweis and Zimmerman (3) had shown that streptococcal group D protease is inactivated by EDTA and reactivated by Zn. If zymogen was being made by cells that were actively making protease, the addition of EDTA should have prevented or slowed down the autocatalytic conversion of enzyme from precursor. Experimentally, such a condition existed among cells prepared according to the preliminary treatment. Figure 1 shows that such cells (treated with EDTA) continued to secrete protease for almost 90 min if Zn+ was added immediately. In the absence of Zn++ (Znlf is added to these samples just prior to assay), only small amounts of enzyme were produced. (It is important to note that no enzyme was elaborated under either circumstance if the same experiment was performed on cells that were harvested from AC broth, a medium in which the cells do not produce enzyme.) The addition of 5 X 10- M Zn+ at 90 min did not cause recovery or activation of enzyme; likewise, the addition of 250 units of purified protease (3) with or without Zn had no effect. Even when cells were sonically disrupted and exposed to ZnH and protease, singly or in combination, no increase in active enzyme over control values could be detected. The possibility that an enzyme precursor precedes protease formation thus appeared unlikely, although it was not completely disproved. bi z 0 ry TIME (min) FiG. 1. Secretion ofprotease by EDTA-treated cells. Cells were incubated in the presence of Zn+ (0) or the absence ofzn (@). At each time interval, a sample was removed and centrifuged. Zn+ was added to the supernatantfluid ofthe cells incubated in its absence.

3 VOL. 97, 1969 PROTEASE SECRETION BY STREPTOCOCCUS FAECALIS 309 The ability to secrete protease wds lost on a time-dependent basis. Figure 2 shows that twothirds of the ability of zinc to rescue active protease in cells prepared according to the preliminary treatment was lost in about 15 min. Figure 2 also may be interpreted as indicating the decay or diminution of some critical material necessary for protease formation. The assumption that release cells might have become deficient in some specific nutrient was easy to test. Although the suspension of release cells in half-strength AC broth (a good growth medium) yielded only 147 units of activity, a similar suspension in halfstrength modified B medium gave 772 units. The various components of modified B medium (Sheffield N-Z Case, yeast extract, gelatin, lactose, arginine, and K2HPO4) were all tested for their ability to stimulate enzyme production. Only arginine and phosphate had this ability, provided zinc was also present. Figure 3 shows this dependence when release cells were suspended in phosphate or arginine at various concentrations in the presence of 5 X 104 M Zn++ and EDTA. Arginine and K2HPO4 concentrations of 5 x 10-' M and 7.5 X 10-2 M, respectively, seemed to be optimal for protease secretion; these concentrations were used in subsequent experiments. The need for zinc can be explained easily in terms of its presence in protease (3); the phos- NJl J-1 D -J 0 U) uji z o 0) TIME OF Zn** ADDITION (min) FiG. 2. Dependence ofproiease secretion on the time of addition of Zn ". At indicated times, Zn++ was added, and tubes were incubated an additional 1.5 hr before the supernatant fluid was assayed for protease. >- 100 U 80 LL >- 60 N zll 40- < 20- ARGININE 102 M) 30 4L PHOSPHATE (10-2 M) FiG. 3. Secretion ofprotease as a function ofarginine concentration (0) and phosphate concentration (0). When the arginine concentration was varied, the concentration ofphosphate was 7.5 X 1072 m. When the phosphate concentration was varied, the concentration ofarginine was kept at 5 X 10 2M; all tubes were made to contain S X 107' M Tris-chloride (ph 7.0). phate and arginine requirement required further experimentation. Although it was shown that lactose (in the experiment testing the constituents in modified B medium) had no effect on enzyme appearance, it was possible to assume that the energy derived by these organisms from arginine (11) breakdown via the arginine dihydrolase system (10) could be specifically needed for protease secretion. In this connection, release cells were incubated with two inhibitors of energy metabolism that were known to prevent the formation of adenosine triphosphate (ATP) from arginine (13, 14) in S. faecalis; both NaF and Na2HAsO4 exercised such an effect (Table 1). Unfortunately, neither citrulline nor carbamyl phosphate, intermediates of arginine breakdown (11) in this organism, could replace arginine with regard to protease elaboration in the release solution. Warburg respirometry showed that no CO2 was produced from citrulline, as had been shown for whole cells by previous workers (19), nor was CO2 produced by whole cells from carbamyl phosphate. Obviously, these intermediates of arginine catabolism in this organism had difficulty in reaching the arginine dihydrolase system. Since energy appeared to be required for secretion of enzyme, it was of interest to determine whether synthesis of new protein was occurring also. Table 1 shows that three of four inhibitors (actinomycin D, puromycin, chloramphenicol, and streptomycin) whose net effect is to prevent protein synthesis all curtailed the appearance of protease. Penicillin, the control antibiotic to

4 310 CASAS AND ZIMMERMAN J. BAcrERioL. TABLE 1. Effect ofinhibitors on secretion ofprotease Inhibition of Compound Concn enzyme secretion NaF M 89 Na2HAsO M 69 Chloramphenicol Jg/ml 100 Puromycin ,ug/ml 100 Actinomycin D... 4,Ag/ml 76 Penicillin G ,ug/ml 0 Streptomycin jag/ml 8 whose action these cells are very susceptible, had very little effect. The possibility that ph might affect the secretion of enzyme, as is the case with penicillinase in Staphylococcus aureus (4), was also examined. A series of tubes containing basal solution (no arginine) adjusted from ph 5.0 to 8.5 at intervals of 0.5 ph units was inoculated with release cells; no enzyme was elaborated under these circumstances. The same experiment performed with arginine gave the greatest amount of enzyme when the initial ph was 7.0 (and the final ph went to 7.5). At least 50% of the optimal yield was obtained in the presence of arginine at initial ph values of 5.5 and 7.5 and at final ph values of 6.6 and 8.2, respectively. With lactose added to the basal solution at a concentration of 10-2 M instead of arginine, the ph dropped instead of rising. As a result, exact comparisons were difficult. In no case was protease produced when lactose was substituted for arginine. In the best comparisons, which were made with 1.5 X 10-1 M KH2PO4 in the basal solution to increase its buffering capacity, the ph of the lactose-containing tubes changed from 7.0 and 7.5 to 6.7 and 6.95, respectively, without elaborating enzyme, whereas the addition of arginine resulted in protease secretion. Further evidence to show that the lactose was being utilized at a rate comparable to that of arginine was obtained by measuring the amount of residual lactose in the supernatant fluid of the test solutions. Over 95% of the lactose had disappeared from the solution after 1 hr of incubation with the cells; Warburg studies showed that about 83% of the total CO2 was evolved from arginine during this same 1-hr incubation period. Since the initial and final ph values of the lactose tubes fell into the range where protease is produced if arginine is used, and since the lactose in the basal medium was being depleted at a rate commensurate with arginine breakdown, it was concluded that lactose could not support protease secretion under these experimental conditions. DISCUSSION The data show that the addition of EDTA to the modified B medium in which S. faecalis var. liquefaciens was actively making extracellular protease did not stop secretion of the enzyme (see upper curve, Fig. 1), even after the cells had been washed and suspended in a Tris-chloride buffer containing zinc. Without zinc, only a small amount of enzyme appeared in the supernatant fluid. As a matter of fact, Fig. 2 shows that the zinc must be added before 15 min have elapsed if any significant amount of protease is to be demonstrated. After this period, phosphate and arginine must be added (in addition to zinc) for secretion of protease to occur. It may be assumed that the secretory process during the first 15 min derives its necessary arginine and phosphate from an internal pool. Secretion of protease does not appear to be explainable in terms of the autocatalytic conversion of zymogen to protease. Evidence for this is that EDTA-treated cells did not liberate active protease when zinc alone was added 15 min after incubation. Based on the finding by Bleiweis and Zimmerman (3) that EDTA-inactivated protease may be reactivated by zinc, one would have anticipated that zymogen would accumulate during EDTA treatment, with subsequent activation by the later addition of zinc or active protease, or both. The observation that three of four antibiotics known to curtail protein synthesis prevented the secretion of protease, whereas one control compound (pencillin) had very little effect, adds credence to the notion that elaboration of protease results from new synthesis. It would appear that, for this metalloenzyme, zinc may be essential for the production of an active protein moiety during the translation process. Of course, one variation of the precursor hypothesis that cannot be eliminated is that protease secretion represents the assembling of already synthesized peptide strands, in the presence of zinc. Presumably such an assembly would require an enzyme with a short half-life (whose synthesis is interfered with by chloramphenicol, puromycin, or actinomycin D) and energy. If the elaboration of protease includes a biosynthetic step, certainly amino acids must be involved. Since the streptococci have complex nutritive requirements [several amino acids (18) are necessary for the growth of S. faecalis var. liquefaciens] and only a single amino acid (arginine) is supplied in our experiments, protease formation must be dependent on an internal amino acid pool or on protein turnover. A possible explanation, therefore, for the arginine requirement is that its addition is necessary to fill

5 VOL. 97, 1969 PROTEASE SECRETION BY STREPTOCOCCUS FAECALIS 311 the depleted arginine reserve. To put it another way, of all the amino acids present in the pool, arginine is the one that is limiting. Were this the case, then the addition of arginine and zinc (and no phosphate) should have restored secretion of enzyme. Gale has shown (5) in these bacteria that the uptake of arginine itself requires no additional source of energy. Interestingly, he found that the energy obtained from arginine could be used to bring other amino acids into the cell. Our results appear to be consistent with the notion that arginine is essential to the process of enzyme secretion because it is the sole source of energy. Although arsenate and fluoride ions have broad inhibitory effects, both compounds have been shown to inhibit specifically the arginine dihydrolase system; Knivett (13) showed that arsenate uncouples the energy production from arginine degradation and that fluoride inhibits the degradation of citrulline (14). Since arginine could be expected to enter the cell in the presence of both fluoride and arsenate, its availability should have led to protease synthesis were this its function in the protease secretion system. An additional argument supporting the role of arginine as an energy source comes from the work of Harold et al. (8, 9). These workers have shown that ATP is necessary for uptake of phosphate (or arsenate) by cells of S. faecalis. In the absence of an energy source, minute amounts of phosphate are taken up, indicating that the ATP pool of resting-cell suspensions is extremely low. This finding that freshly harvested cells have little residual ATP reenforces the idea that the ATP derived from arginine is the only energy available for protease secretion in our experiments. Another interesting facet of the work of Harold and his collaborators is that the ATP produced from arginine cannot serve as the energy source to pump phosphate into the cell; however, ATP derived from fermentable carbohydrate works very well. One possible explanation for this seemingly anomalous situation is that the ATP produced from arginine is not available in the same area of the cell as is ATP from carbohydrate. By analogy, in the system reported here, ATP produced from lactose does not appear to be available for protease secretion, as is the ATP from arginine. In this connection, Trentini and Chesbro (23) have shown that, of the two enzymes connected with arginine breakdown in this organism, the one related to energy transfer (the carbamyl transferase) is associated with a particulate fraction of the cells, whereas the arginine deiminase is a soluble enzyme, presumably from the cytoplasm. We would propose, then, that the energy required for protease synthesis can come only from arginine, because the energy-yielding reaction is compartmentalized in an area of the cell where protease is also synthesized and into or through which ATP generated from lactose cannot diffuse, or at best, diffuses very slowly. Streptomycin, but not puromycin, would also have difficulty in reaching the site of protease biosynthesis. In a recent article, Richmond (21) suggested that the extracellular penicillinase of Staphylococcus aureus is synthesized on ribosomes that are not available for general protein synthesis by the cell. Presumably, these ribosomes would also be located in this or a similar area or compartment of the cell. In a related paper, Beaton (2) showed that morphological changes occur in mesosomes of Staphylococcus aureus during penicillinase liberation. It is unfortunate that neither citrulline nor carbamyl phosphate (the intermediates of arginine breakdown in S. faecalis) can cross the permeability barrier presented by the plasma membrane. As a result, we have been unsuccessful in showing that either of these two compounds can replace arginine in our system, and the possibility remains that a very small percentage of the large quantity of arginine used during release may be required for incorporation rather than energy. ACKNOWLEDGMENTS This investigation was supported by the Public Health Service grant A from the National Institute of Allergy and Infectious Diseases. During part of this work, Ivan Casas held a fellowship from the Food and Agriculture Organization of the United Nations, Project 80. LITERATURE CITED 1. Anson, M. L The estimation of pepsin, trypsin, papain, and cathepsin with hemoglobin. J. Gen. Physiol. 22: Beaton, C. D An electron microscope study of the mesosomes of a penicillinase-producing Staphylococcus. J. Gen. Microbiol. 50: Bleiweis, A. S., and L. N. Zimmerman Properties of proteinase from Streptococcus faecalis var. liquefaciens. J. Bacteriol. 88: Coles, N. W., and R. Gross Liberation of surfacelocated penicillinase from Staphylococcus aureus. Biochem. J. 102: Gale, E. F The assimilation of amino acids by bacteria. I. The passage of certain amino acids across the cell wall and their concentration in the internal environment of Streptococcus faecalis. J. Gen. Microbiol. 1: Grutter, F. H., and L. N. Zimmerman A proteolytic enzyme of Streptococcus zymogenes. J. Bacteriol. 69: Hammel, J. M., and L. N. Zimmerman The dependence of proteinase biosynthesis on the cell wall in Streptococcus faecalis var. liquefaciens. Biochim. Biophys. Acta 129: Harold, F. M., and J. R. Baarda Interaction of arsenate with phosphate-transport systems in wild-type and mutant Streptococcus faecalis. J. Bacteriol. 91: Harold, F. M., R. L. Harold, and A. Abrams A mutant of Streptococcus faecalts defective in phosphate uptake. J. Biol. Chem. 240: Hills, G. M Ammonia production by pathogenic bacteria. Biochem. J. 34:

6 312 CASAS AND ZIMMERMAN J. BACrERIOL. 11. Jones, M. E Carbamyl phosphate. Science 140: Jones, M. E., and F. Lipmann Chemical and enzymatic synthesi of carbamyl phosphate. Proc. Natl. Acad. Sci. U.S. 46: Knivett, V. A The effect of arsenate on bacterial citrulline breakdown. Biochem. J. 56: Korzenovsky, M Metabolism of arginine and citrulline by bacteria, p In W. D. McElroy and H. B. Glass (ed.), Amino acid metabolism. The Johns Hopkins Press, Baltimore. 15. Lampen, J Secretion of enzymes by microorganisms. Symp. Soc. Gen. Microbiol. 15: Liu, T. Y., and S. D. Elliott Activation of streptococcal proteinase and its zymogen by bacterial cell walls. Nature 206: Morris, D. L Quantitative determination of carbohydrates with Dreywood's anthrone reagent. Science 107: Niven, C. F., Jr., and J. M. Sherman Nutrition of the enterococci. J. Bacteriol. 47: Oginsky, E. L., and R. F. Gehrig The arginine dihydrolase system of Streptococcus faecalls. I. Identification of citrulline as an intermediate. J. Biol. Chem. 198: Rabin, R., and L. N. Zimmerman Proteinase biosynthesis by Streptococcus liquefacens. I. The effect of carbon and nitrogen sources, ph, and inhibitors. Can. J. Microbiol. 2: Richmond, M. H New type of restriction to the expression of a structural gene in bacteria. Nature 216: Shugart, L. R., and R. W. Beck Occurrence and distribution of proteinase of Streptococcus faecalls var. liquefaclens. J. Bacteriol. 92: Trentini, W., and W. Chesbro Localization of the "arginine dihydrolase system" in Streptococcus faecfum. Biochim. Biophys. Acta 67: Umbreit, W. W., R. H. Burris, and J. F. Stauffer Manometric techniques. Burgess Publishing r,o., Minneapolis, Minn. Downloaded from on September 18, 2018 by guest