(Schultz, 1949). In the rat it exerts a nephrotoxic effect (Artom et al., 1945).

Transcription

1 PANTOTHENATE STUDIES I. INTERFERENCE BY D-SERINE AND L-ASPARTIC ACID WITH PANTOTHENATE SYNTHESIS IN ESCHERICHIA COLI WERNER K. MAAS' AND BERNARD D. DAVIS U. S. Public Health Service, Tuberculosis Research Laboratory, Cornell University Medical College, New York 21, New York Received for publication August 17, 1950 In recent years growth inhibitors with simple chemical structures have been used to study biosynthetic pathways as well as mechanisms of antimicrobial action. This approach has been based on the concept of competitive inhibition of essential metabolites by structural analogues. The systematic application of this concept in the methodology of inhibition analysis (Shive and Macow, 1946; Shive, 1950) involves the assumptions that a competitive antagonist of the inhibitor is the substrate, or a precursor of the substrate, of the affected reaction, and that a noncompetitive antagonist is the product, or a derivative of the product, of this reaction. Auxotrophic (nutritionally deficient) mutants, like antimetabolites, have contributed to an understanding of biosynthetic pathways; in the present paper we propose to demonstrate that they also lend themselves to investigating the mode of action of growth inhibitors and to testing assumptions underlying inhibition analysis. The inhibition of Escherichia coli by D-serine and L-aspartic acid will be analyzed by a combination of the two methods, and it will be shown that the application of mutants strengthens the conclusions reached from inhibition analysis with the parent wild type alone. D-Serine produces toxic effects in various groups of organisms. Among bacteria it prevents the growth of Bacillus anthracis (Gladstone, 1939), Streptococcus faecalis R (Snell and Guirard, 1943), and E. coli (Davis and Maas, 1949) and inhibits the production of tetanus toxin by Clostridium tetani (Mueller and Miller, 1949). The growth of Mycobacterium tuberculosis (Dubos, 1949) and of Bacillus subtilis (Teas, 1950) is inhibited by DL-serine; in the case of the tubercle bacillus the inhibition has been traced to the D-isomer (Suter, 1950). In Drosophilia, D-serine inhibits development of the larvae grown on a synthetic medium (Schultz, 1949). In the rat it exerts a nephrotoxic effect (Artom et al., 1945). In a previous note (Davis and Maas, 1949) we have described the nature of D-serine inhibition of wild type E. coli and several factors associated with it. D-Serine was shown to reduce the growth rate of the bacteria without, however, affecting their viability. Optimal growth could be restored by a variety of amino acids, some of which had been shown to protect rats from the toxic action (Wachstein, 1947). The most effective antagonists were alanine and glycine, I Part of this work was done while the senior author was a staff member of the Naval Medical Research Institute, Bethesda, Maryland, on assignment to the Tuberculosis Research Laboratory. 733

2 734 WERNER K. MAAS AND BERNARD D. DAVIS [VOL. 60 which at equimolar concentrations completely reversed the inhibition. In contrast, L-aspartic acid in very low concentrations, in the presence of D-serine, increased the retardation of growth considerably; although by itself, even at much higher concentrations, it had no effect. In the present paper it will be shown that D-serine exerts its inhibitory effect by interfering with the conversion of 3-alanine to pantothenic acid. Furthermore, it will be demonstrated that L-aspartic acid alone, under appropriate conditions, inhibits the same reactions. These conclusions, derived from experiments with the wild type, were strengthened by the use of mutants blocked at various stages of pantothenate synthesis. METHODS The E. coli used in this study was the W strain, ATCC Two types of experiments were performed: with growing cells and with washed nongrowing cell suspensions. The auxotrophic mutants were obtained by the penicillin method (Davis, 1949). Growth experiments. The minimal medium employed has already been described (Davis, 1949). Growth responses were measured in agar pour plates with inocula of about 100 cells, the visually estimated values of colony size ranging from 0 to 4+. There are several reasons for preferring this semiquantitative procedure to the more widely accepted one of measuring growth turbidimetrically. D-Serine-resistant mutants occur with a relatively high spontaneous frequency and have such a strong selective advantage in the presence of D-serine that they rapidly outgrow the normal strain in liquid media. Only on solid media is it convenient to distinguish the growth of the normal strain, in response to various conditions, from the growth of the occasional resistant mutant. Furthermore, colony size can be followed over a period of several days, whereas turbidity advances rapidly once it has appeared. Finally, comparison of the number of colonies formed with the number of cells inoculated enables one to detect a partial bactericidal effect. For the recording of the biochemically significant facts the method is sufficiently sensitive. In order to study the effect of D-serine on various portions of the growth curve, it was necessary to measure growth more accurately in liquid cultures. The results of these experiments will be published separately. Since L-serine does not antagonize growth inhibition by D-serine, the DLmixture was employed. In experiments with nongrowing cell suspensions, however, the D-isomer was used. We wish to thank Dr. J. P. Greenstein for a generous sample of D-serine. Washed cell suspensions. Bacteria were grown overnight at 37 C in NY medium (minimal medium enriched with 0.2 per cent enzymatically hydrolyzed casein- Sheffield N-Z-case--and 0.2 per cent Difco yeast extract). The organisms were harvested by centrifugation, washed twice with distilled water, and concentrated into the desired volume of liquid. Experiments on pantothenate production were carried out in sterile test tubes. Samples taken from the tubes at the beginning and end of the experiment were heated at 100 C for 10 minutes in

3 1950] PANTOTHENATE STUDIES 735 order to inactivate the pantothenate-producing enzyme system; they were then centrifuged, and the supernatants were autoclaved. No significant amount of intracellular pathothenate was released into the medium during the heating, as shown by the pantothenate values at zero time. Appropriate dilutions of these samples were assayed with E. coli mutant strain 99-1, which responds specifically to pantothenate and not to any of the precursors. The medium (medium A) has been described (Davis and Mingioli, 1950). This method has an advantage over the conventional one employing Lactobacillus arabinosus because of the simple assay medium, consisting of inorganic salts and glucose; in our experience the results were equally sensitive and reliable. The assay organism was grown for 20 hours in NY medium, washed twice with distilled water, and suspended in the original volume of water. Colorimeter tubes (i inches in diameter) containing medium A and the sample to be tested (total volume = 10 ml) were inoculated with 0.2 ml of 10 per cent glucose ' PANrOTHENATE m/#g/tuj8e Figure 1. Response of mutant 99-1 to pantothenate. Curve A: medium A (see text). Curve B: medium A, enriched with 0.5 per cent hydrolyzed casein (enzymatic, Nutritional Biochemicals) and the following vitamins in/ag per L: inositol 5,000, thiamine 20, pyridoxine 20, nicotinic acid 100, PABA 20. solution containing 10-2 ml of the washed culture. Turbidities were measured after approximately 18 and 24 hours of incubation at 37 C with an Evelyn photoelectric colorimeter, filter 515 m,u. In each experiment a control assay was run, and some of the tubes were tested for the presence of back mutants. Only rarely were appreciable numbers of prototroph cells found in any tube, and these unsatisfactory assays could be recognized by the development of abnormally high turbidity during the 24-hour incubation period. The response of the organism to added pantothenate in medium A, and also in medium A enriched with casein hydrolyzate and a mixture of vitamins, is shown in figure 1. It can be seen that the enrichments exert a slight sparing effect on pantothenate; one must therefore be cautious in using this assay for the estimation of pantothenate in complex biological materials. The W strain of E. coli, however, excretes no amino acids, and among the vitamins only PABA, desthiobiotin, and pantothenate (Davis, 1950a; unpublished observations). Since

4 736 WERNER K. MAAS AND BERNARD D. DAVIS [VOL. 60 PABA and desthiobiotin in the concentrations excreted produce no change in response to pantothenate, the assay is satisfactory under the conditions of our experiments. RESULTS Growth experiments with wild type. Reversal of D-serine inhibition by several amino acids has been reported (Davis and Maas, 1949). When all the available vitamins were subsequently tested, only pantothenate completely reversed D- serine inhibition. Thiamine had a slight antagonistic effect that could not be enhanced by increasing its concentration. (In this connection a mutant should be mentioned in which the pantothenate requirement can be replaced partially by thiamine. This mutant will be described in a separate publication.) TABLE 1 Inhibition by DL-serine of growth of the wild type strain and its reversal by Ca-pantothenate and fi-alanine as measured by effect on colony size in agar medium DL-SERINE SUPPLEMENT None or 50 pg/ml + DL- DL-asp. 50 Ag/ml 200 pg/ml 1,000 pg/mi aspartic acid 10 pg/mi 10 pag/mi None... 1+, 4+ 0, 2+, 4+ 0, i+, 3+ 0, 0, 1+, 3+ 0, 0, 0, 0 Ca-pantothenate 3 mpg/ml , 4+ 0, 3+, 4+ 0, 2+, 4+ 0, 4+ Ca-pantothenate 10 mpag/ml... - m, 4+ m, 4+ 0, 2+, 4+ m, 4+ Ca-pantothenate 30 mpg/ml... _ 1+, 4+ 1+, 4+ m, 3+, 4+ 1+, 4+ Ca-pantothenate 100 mpag/ml , 4+ 1+, 4+ m, 3+, 4+ 1+, 4+ Ca-pantothenate 300 mpg/ml , 4+ 1+, 4+ m, 3+, 4+ 1+, 4+ Ca-pantothenate 1,000 mpug/ml - 1+, 4+ 1+, 4+ m, 3+, 4+ 1+, 4+,9-Alanine 3 mpg/ml , 2+, 4+ 0, 1+, 4+ 0, 0, 2+, 3+ 0, 0, 0 1-Alanine 10 mpg/ml... _ 0, 3+, 4+ 0, 3+, 4+ 0, m, 2+, 4+ 0, 0, Alanine 30 mpg/ml... _ 0, 4+ 0, 3+, 4+ 0, 1+, 3+, 4+ 0, 0, 1-2+ fl-alanine 100 mpg/ml... - m, 4+ 0, 4+ 0, 2+, 3+, 4+ 0, 1-3+, 4+ f-alanine 300 mug/ml , 4+ n1, 4+ 0, 2+, 4+ m-1+, 4+,6-Alanine 1,000 mpg/ml , 4+ m, 4+ m, 3+, 4+ 1+, 4+ Agar pour plates of minimal medium supplemented as indicated. Inoculum of 50 to 100 cells per plate, taken from 24-hour culture in minimal meditum. Colony sizes recorded at successive 24-hour intervals. m = microscopic, visible under binocular microscope (10 X magnification); 1+ = visible to unaided eye; 4+ = large. Inhibition of the growth of the wild type strain and its reversal by Ca-pantothenate and 1-alanine are shown in table 1. DL-Serine produces a delay in the appearance of colonies the duration of which depends upon the concentration used. Pantothenate at a level of 30 m,ug per ml, but not at 10 m,ug per ml, completely reverses the inhibition at several concentrations of DL-serine. With,3- alanine, on the other hand, the concentration required to reverse the inhibition completely is proportional to the D-serine concentration. This competitive relationship and the noncompetitive reversal by pantothenate suggest interference by D-serine with the conversion of,b-alanine to pantothenate. At the highest level of DL-serine tested (1 mg per ml), partial inhibition appears even in the presence of pantothenate and cannot be overcome by raising the pantothenate concentration (table 1); apparently a sufficiently high concen-

5 1950] PANTOTHENATE STUDIES 737 tration of D-serine affects metabolic processes other than pantothenate synthesis. In this connection it may be noted that glycine, which has been shown to reverse inhibition by D-serine (Davis and Maas, 1949), does so even at a level of DLserine as high as 5 mg per ml, at which pantothenate has no antagonistic effect. The addition of L-aspartic acid has been shown to prolong the retarding effect of DL-serine (Davis and Maas, 1949); the inactivity of D-aspartic acid, as shown by comparison of the L-form with the DL-mixture, permitted the use of the racemate. As shown in table 1, DL-aspartic acid does not change the noncompetitive reversal of the inhibition by pantothenate but does increase the concentration of f3-alanine needed for reversal. Apparently it acts together with D-serine in retarding synthesis of pantothenate from,b-alanine. The effect of D-serine was tested on several other wild type strains of E. coli. Inhibition was observed with strain P.A. 14 of Dr. van Niel's collection whereas strains K-12 and "Texas" were not affected. The latter two strains behave like the D-serine-resistant mutants obtained in the W strain, which in one mutational step develop complete resistance to the inhibitor in concentrations as high as 10 mg per ml. Inhibition by D-serine in another wild type strain has recently been reported to be reversed by the same amino acids that are effective in the W strain (Mentzer et al., 1950). During this study the constituents of the medium were tested for their effect on the inhibition and its antagonism. It was found that, in the absence of chloride from the medium, pantothenate only partly reverses inhibition by even low concentrations of DL-serine (50,ug per ml). The complete antagonistic activity of pantothenate could be restored by the addition of either sodium or potassium chloride. It therefore appears that the absence of chloride lowers the threshold at which reactions not involving pantothenate are inhibited by D-serine. Chloride is not necessary for the growth of the wild type and has no effect on the quantitative response of a pantothenate-requiring mutant to various concentrations of pantothenate. Growth experiments with mutants. Studies on the effect of DL-serine in a mutant blocked in the synthesis of jl-alanine support the hypothesis of interference by D-serine with conversion of 3-alanine to pantothenate (table 2). It can be seen that, at a pantothenate concentration that allows optimal growth, DL-serine produces no inhibition at all. In the presence of a concentration of g-alanine sufficient to support optimal growth, however, D-serine can produce marked inhibition, which is overcome by increasing the j3-alanine concentration. As in the wild type, the concentration of the antagonist required for complete reversal depends on the concentration of the inhibitor used, thus confirming the conclusion that D-serine interferes with jb-alanine conversion. In the,b-alanineless mutant, as in the wild type, L-aspartic acid enhances the inhibition. Furthermore, in contrast to the behavior of the wild type, this mutant is inhibited by aspartic acid alone at a concentration of 10,ug per ml in the presence of,b-alanine concentrations suboptimal for growth; it is not inhibited, however, in the presence of pantothenate. In this strain, therefore, aspartic acid, like D-serine, interferes with the conversion of fb-alanine to pantothenate.

6 738 WERNER K. MAAS AND BERNARD D. DAVIS [VOL. 60 The role of L-aspartic acid in these reactions will be considered further in the discussion. The foregoing interpretation of the site of action of D-serine was confirmed by the study of two other pantothenateless mutants (table 3). One, blocked in the TABLE 2 Inhibition by DL-serine of,b-alanineless mutant 99-2 and its reversal by Ca-pantothenate and 3-alanine DL-SERINE SUPPLEMENT 500 pg/ml + None + None 50 pg/ml 200 pg/ml 30 pg/mi L-acartic Nn+g p"g/milojgm None... 0, O Ca-pantothenate 10 mpg/ml... 2+, 4+ 2+, 4+ 1+, 4+ 1+, 4+ 1+, 4+ 3+, 4+ Ca-pantothenate 30 mpg/ml... 2+,4+ 2+,4+ 2+,4+ 2+,4+ 2+,4+ 3+,4+ Ca-pantothenate 100 mpg/ml.. 2+, 4+ 2+, 4+ 2+, 4+ 2+, 4+ 3+, 4+ 3+, 4+ Ca-pantothenate 300 mpg/ml.. 2+, 4+ 2+, 4+ 2+, 4+ 2+, 4+ 3+, 4+ 3+, 4+ B-Alanine 10 mp4g/ml...0, 2+,0 0,O0 0, 0 0, 0 0, 0 P-Alanine 30 mpg/ml... 2+,4+ 0,0 0,0 0,0 0,0 0,3+ fl-alanine 100 mag/ml... 2+, 4+ 0, 3+ O, 3+ 0, 2+ 0, 0 2+, Alanine 300 m,ug/ml... 2+, 4+ 1+, 4+ m, 4+ m, 3+ m, 4+ 3+, 4+ fl-alanine 1,000 mqg/ml.. 2+,4+ 2+,4+ 1+,4+ m,4+ 2+,4+ 3+, 4+ Same technique as that described in table 1. Bacteria grown for 24 hours in NY medium before inoculation into plates. MUTANT NO. TABLE 3 Effect of D-serine on pantoicless mutant 99-4 and pantothenateless mutant 99-1 in the presence of their growth factors DL-SERINE Supplement None 50 pg/ml 1,000 g/mil 0 5,gm apartic Oglml L None 0, 0, Pantoic acid 10 pag/ml 2+, 4+ 0, 2+, 3+, 4+ 0, 1+, 3+, 4+ 0, 0, 0, 0 Pantoic acid 100 pg/ml 2+, 4+ 0, 2+, 3+, 4+ 0, 1+, 3+, 4+ 0, 0, 0, 0 Ca-pantothenate 10 mpg/mi m, 2+, 3+, 3+ m, 2+, 3+, 3+ 0, 1+, 3+, 3+ m, 2+, 3+, 3+ Ca-pantothenate 100 mpg/ml 2+, 4+ 2+, 4+ 1+, 4+ 2+, Ca-pantothenate 10 mpg/mi m, 2+, 3+, 3+ m, 2+, 3+, 3+ 0, 1+, 3+, 3+ m, 2+, 3+, 3+ Ca-pantothenate 100 mpg/mi 2+, 4+ 2+, 4+ 1+, 4+ 2+, 4+ Same technique as that deecribed in table 1. Bacteria grown for 24 hours in NY medium before inoculation in plates. synthesis of the pantoic acid moiety of the pantothenate molecule, was unable to grow in the presence of D-serine when pantoic acid was used as the growth factor; in the presence of pantothenate, however, the inhibition was reversed noncompetitively, as in the other strains tested. The second, a specific pantothenicless mutant, which does not respond to either of the precursors, was not

7 1950] PANTOTHENATE STUDIES 739 inhibited by D-serine. In both of these mutants the combination of D-serine and L-aspartic acid, like D-serine alone, produces no inhibition in the presence of pantothenate. Cell suspensions. It was found by McIlwain and Hughes (1944) and confirmed by us, that resting cell suspensions of E. coli synthesize pantothenate in the presence of 13-alanine, and that a considerable portion of the pantothenate formed is excreted into the medium. One can therefore study the effect of D-serine on the pantothenate synthesizing system more directly, without the complicating effect of growth. The results of a typical experiment with a,-alanineless mutant are shown in table 4. Similar results have been obtained with the wild type. Because of the growth-stimulating effect of L-serine (see next section), D-serine was used instead of the DL-mixture. The amount of nitrogen in the jl-alanine added in this experiment is too small to permit measurable growth. It can be seen that D-serine TABLE 4 Inhibition by D-serine of pantothenate production in resting cell suspensions of j8-alanineless mutant 99-9 SUPPLEMENT PANTOTENATE. PRODUCED pg/mi P-Alanine Pantoyl lactone No D-serine D-Serine 1,000 pg/ml ug/mi pg/mi (<0.1) All tubes contain 1.0 ml of 10 X concentrated cell suspension; 0.2 ml of 10 per cent glucose; 0.3 ml of 0.1 M sodium phosphate buffer, ph 7.0. Total volume = 2.0 ml (adjusted with distilled water). Incubated for 4 hours at 25 C. markedly inhibits pantothenate excretion; as in growing cells, the inhibition is overcome by,3-alanine. At a g-alanine level of 3,ug per ml, a D-serine:#-alanine ratio of 300:1 produces almost complete inhibition. Partial reversal is obtained with a 3-fold increase, and complete reversal with a 30-fold increase in the j3- alanine concentration. It should be noted in table 4 that,b-alanine is required by the mutant for pantothenate excretion; it is similarly required by suspensions of the wild type even though this strain can synthesize i3-alanine during growth. In contrast, the other precursor, pantoyl lactone, is not required for pantothenate excretion; addition of this compound in excess, however, considerably increases pantothenate output. Apparently,B-alanine cannot be synthesized in the absence of a nitrogen source, whereas pantoyl lactone (or pantoic acid), which has no nitrogen, can be formed, though at a limited rate, from intracellular resources. In a mutant blocked in the synthesis of pantoyl lactone we have observed, as expected, that both B-alanine and pantoyl lactone must be added.

8 740 WERNER K. MAAS AND BERNARD D. DAVIS [VOL. 60 Table 5 shows the effect of aspartic acid, alone and in combination with D- serine, on pantothenate formation in nongrowing cell suspensions of the wild type. In contrast to its failure to inhibit growth, aspartic acid alone inhibits pantothenate excretion; this inhibition, as well as the combined one with D- serine, is reversed by increasing the 3-alanine concentration, lending additional support to the conclusion that aspartic acid interferes with f-alanine utilization in pantothenate synthesis. Possible reasons for the difference between growing and nongrowing cells of the wild type in their behavior toward L-aspartic acid will be considered in the discussion. Is the inhibition produced by D-serzne itself or by a conversion product of it? Chargaff and Sprinson (1943) have demonstrated the production of pyruvate from DL-serine in E. coli by isolation of its 2,4-dinitrophenylhydrazone. They postulated, on the basis of the amounts of ammonia produced by washed cell TABLE 5 Inhibition by D-serine and DL-aspartic acid of pantothenate resting cell suspensions of wild type production in SUPPLEMENT 8-Alanine D-Serine DL-Aspartic acid PANTOTHENATE PRODUCED.Ug/ml,ug/ml /Ag/Ml,ug/ml Q , , , , All tubes contain 1.0 ml of 15 X concentrated cell suspension; 0.2 ml of 10 per cent glucose; 0.2 ml of 1 M sodium phosphate buffer pl1 7.0; 0.3 ml of 0.1 per cent pantoyl lactone. Total volume = 3.0 ml (adjusted with distilled water). Incubated for 4 hours at 25 C. suspensions in the presence of DL-serine, that both isomers are converted into pyruvate. Since D-serine has now become available, production of pyruvate from both DL- and D-serine was tested in our strain of E. coli, in collaboration with Dr. D. Sprinson. The conditions were those described by Chargaff and Sprinson. Although the 2,4-dinitrophenylhydrazone of pyruvate was readily obtained following incubation of cell suspensions with the DL-mixture, none could be isolated following incubation with the D-isomer only. Thus in our strain of E. coli the enzyme system responsible for the conversion of serine to pyruvate is not demonstrably active toward the D-isomer. The same authors have demonstrated the production of hydroxypyruvic acid from D-serine by rat kidney slices (Sprinson and Chargaff, 1946). This compound was prepared by their method from bromopyruvic acid kindly furnished by Drs. D. Sprinson, E. May, and E. Mosettig. It was impossible to test for an

9 1950] PANTOTHENATE STUDIES 741 inhibitory effect of hydroxypyruvate on growing cells since traces of bromopyruvate, itself strongly inhibitory, could not be removed. In nongrowing cell suspensions, however, this preparation of hydroxypyruvate was found to inhibit pantothenate excretion only slightly at a molar concentration at which D-serine inhibited excretion completely. Pyruvate had the same activity as hydroxypyruvate. It is clear from these experiments that D-serine does not produce its effect via either of these substances. The possibility of the formation of a conversion product of D-serine was also investigated by testing D-serine as the sole nitrogen source in growth experiments, using a D-serine-resistant mutant that grows as well in the presence of D-serine as in its absence. It was found that growth was very slight on plates containing 5 mg per ml of D-serine and no other N source, although DL-serine at the same concentration was as effective as the (NH4)2SO4 ordinarily used. Since these experiments indicate that D-serine is not readily deaminated, it is unlikely that the inhibition observed is due to a deamination product. Furthermore, the product of its decarboxylation, ethanolamine, did not inhibit growth at a concentration of 200 jig per ml. These several experiments indicate that D-serine itself, rather than a conversion product, is probably responsible for the inhibition, although the possibility of more complex conversion products cannot be excluded. DISC'USSION The experiments presented illustrate the use of auxotrophic mutants in conjunction with inhibition analysis in determining the site of action of an antimetabolite. The combined methodologies enable one to draw conclusions with greater certainty than is possible with the wild type alone since the mutants permit experimental test of certain assumptions implicit in inhibition studies. For instance, the assumption that a reversing agent is a metabolite can be experimentally proved by the demonstration of growth response of a mutant. Furthermore, quantitative growth requirements show whether or not an antagonist acts at the "physiological" concentration range ordinarily present in the cell; this information is especially pertinent with a noncompetitive antagonist, for which the same concentration should be required to overcome the inhibition and to support optimal growth of the mutant. Comparisons of this kind are, of course, much more significant when made between the wild type and a mutant of the same species than between different species. Finally, and most important, an effective competitive reversing agent may be a known metabolite, and yet be shown by means of mutants not to be a true precursor of the inhibited reaction. This effect will be illustrated below by the example of glycine antagonism of D-serine inhibition. Mutants have the further advantage of sometimes revealing an inhibition that is obscured by compensatory mechanisms in the wild type. For example, in the present work inhibition by aspartic acid alone could not be detected in the wild type, presumably because of compensatory reactions, but could be detected in a mutant unable to synthesize,3-alanine. Furthermore, comparison between

10 742 WERNER K. MAAS AND BERNARD D. DAVIS [VOL. 60 tables 1 and 2 shows that in a,b-alanineless mutant in the presence of a,b-alanine concentration optimal for growth (30 mpg per ml) the inhibition is more prolonged than it is in the presence of the same concentration of f-alanine in the wild type, indicating that synthesis of j3-alanine by the latter strain is at least partially responsible for the temporary nature of the inhibition. A few examples will illustrate the danger of erroneous conclusions from the application of inhibition analysis alone. In Lactobacillus casei bromouracil behaves like the product of a reaction inhibited by nitrouracil; yet bromouracil is not a normal metabolite (Hitchings et at., 1948). In E. coli as the result of inhibition analysis, tryptophan has been postulated as a precursor of phenylalanine (Beerstecher and Shive, 1946), and the latter as a precursor of tyrosine (Beerstecher and Shive, 1947); studies with biochemical mutants, however, indicate that these three aromatic acids arise from a common precursor and are not interconvertible (Davis, 1950b). In the present case the following possible misinterpretation might have been given to the action of D-serine if one did not have the data provided by the mutants. Glycine has been shown to overcome D-serine inhibition in a competitive manner over a 20-fold range of concentration (Davis and Maas, 1949), including the range in which the inhibition is reversed competitively by,balanine and noncompetitively by pantothenate. According to the principles of inhibition analysis, glycine would be interpreted as a precursor of the blocked reaction and therefore of pantothenate; since,b-alanine is found in pantothenate, glycine would be interpreted as a precursor of j-alanine. Yet glycine cannot replace 3-alanine in promoting pantothenate excretion in nongrowing cell suspensions of wild type. Furthermore, glycine reverses D-serine inhibition at a concentration of the inhibitor at which even an excess of 3-alanine does not reverse it. Finally, though glycine cannot satisfy the growth requirement of a mutant blocked in the synthesis of,b-alanine, it does reverse the inhibitory effect of D-serine on the growth of this mutant in the presence of a moderate concentration of f-alanine. These observations indicate that the mechanism of antagonism of this amino acid, which is at present obscure, differs from that of,b-alanine and does not involve entrance of glycine into the normal pathway leading to pantothenate synthesis. Details of the effect of glycine and alanine on D-serine inhibition will be reported in a separate publication. Although it has been shown that D-serine interferes with the utilization of j-alanine, the more intimate mechanism of its action is as yet obscure. The similarity in chemical structure of the two compounds suggests competition for a specific site on an enzyme surface. This hypothesis is strengthened by the reported interference with fl-alanine utilization produced by several other compounds with similar chemical structures. These include a-aminobutyric acid, aspartic acid, isoserine (a-hydroxy-3-alanine) and,3-phenyl-j-alanine in yeast (Nielsen et al., 1944), and propionic acid in E. coli (King and Cheldelin, 1948). Another mechanism of inhibition, suggested by Hartelius (1946), is the formation of a peptide bond between the carboxyl group of,b-alanine and the amino group of the inhibitor. This hypothesis would not account for inhibition

11 1950] PANTOTHENATE STUDIES 743 by propionic acid; furthermore, one would not expect peptide formation to be restricted to amino acids with a 3- or 4-carbon chain. In the present paper, one of these structurally related compounds, L-aspartic acid, has been shown to produce several effects: it inhibits conversion of fbalanine to pantothenate in growing cells of a,3-alanineless mutant and in nongrowing cell suspensions of the wild type; it enhances the inhibitory action of D-serine on pantothenate synthesis in the growing wild type but does not by itself inhibit growing cells of the wild type. The results reported suggest that its synergistic effect with D-serine on the growth of the wild type, like its other effects, involves interference with f3-alanine conversion to pantothenate. However, it is not clear whether or not L-aspartic acid and D-serine interfere with the same reaction step. Lack of inhibition by L-aspartic acid alone in growing wild type cells may depend on the formation of sufficient 3-alanine from added aspartic acid to overcome its own inhibition. On the basis of cell suspension experiments (Billen and Lichstein, 1949; Maas, unpublished) and inhibition analysis (Ravel and Shive, 1946), aspartic acid has been postulated as the natural precursor of,3- alanine. Synthesis of 3-alanine from L-aspartic acid in resting cells, however, is confined to acid conditions of ph 6 or below; inhibition of,8-alanine conversion by L-aspartic acid, on the other hand, was studied at ph 7. This limitation of,3-alanine synthesis to an acid ph range may explain why L-aspartic acid cannot reverse its own inhibition in resting cells; yet growing cells may behave differently and permit g-alanine formation from L-aspartic acid even at neutrality. Experiments are in progress to clarify the double role of L-aspartic acid as a precursor of,3-alanine and at the same time an inhibitor of,8-alanine conversion. It is a great pleasure to acknowledge the excellent technical assistance of Margaret C. Sanderson. BMMARY D-Serine inhibits growth of the W strain of Escherichia coli by interfering with the conversion of,b-alanine to pantothenate. The evidence for this mode of action, obtained from experiments with the wild type, has been strengthened by study of appropriately selected mutants blocked at different stages of pantothenate formation. The effect of D-serine on pantothenate formation has also been demonstrated in nongrowing cell suspensions. L-Aspartic acid enhances growth inhibition of the wild type by D-serine but does not by itself inhibit growth. In a,b-alanine-requiring mutant, however, and in nongrowing cell suspensions of the wild type, aspartic acid interferes with,balanine conversion to pantothenate. Reasons for the lack of growth inhibition of the wild type are discussed. Evidence is presented that D-serine itself rather than a conversion product is the effective inhibitory agent. The usefulness of biochemical mutants combined with inhibition analysis in the study of the mechanisms of antimicrobial action is discussed.

12 744 WERNER K. MAAS AND BERNARD D. DAVIS [VOL. 60 REFERENCES ARTOM, C., FISHMAN, W. H., AND MOREHEAD, R. P The relative toxicity of 1- and dl-serine in rats. Proc. Soc. Exptl. Biol. Med., 60, 284. BEERSTECHER, E., JR., AND SHIVE, W Biochemical transformations as determined by competitive analogue-metabolite growth inhibitions. III. A transformation involving phenylalanine. J. Biol. Chem., 146, BEERSTECHER, E., JR., AND SHIVE, W Biochemical transformations as determined by competitive analogue-metabolite growth inhibitions. V. Prevention of tyrosine synthesis by j3-2-thienylalanine. J. Biol. Chem., 167, BILLEN, D., AND LICHSTEIN, H. E Studies on the aspartic acid decarboxylase of Rhizobium trifolii. J. Bact., 58, CHARGAFF, E., AND SPRINSON, D. B Studies on the mechanism of deamination of serine and threonine in biological systems. J. Biol. Chem., 151, DAVIS, B. D The isolation of biochemically deficient mutants of bacteria by means of penicillin. Proc. Natl. Acad. Sci. U. S., 35, DAVIS, B. D. 1950a Studies on nutritionally deficient bacterial mutants isolated by means of penicillin. Experientia, 6, DAVIS, B. D. 1950b Manuscript in preparation. DAVIS, B. D., AND MIAAS, W. K Inhibition of E. coli by D-serine and the production of serine-resistant mutants. J. Am. Chem. Soc., 71, DAVIS, B. D., AND MINGIOLI, E. S Mlutants of Escherichia coli requiring methionine or vitamin B12. J. Bact., 60, DUBOS, R. J Toxic effects of dt-serine on virulent human tubercle bacilli. Am. Rev. Tuberc., 60, 385. GLADSTONE, G. P Interrelationships between amino acids in the nutrition of B. anthracis. Brit. J. Exptl. Path., 20, HARTELIUS, V Glutamic acid, aspartic acid and glutanmine as antigrowth substances for B-alanine. Compt. rend. trav. lab. Carlsberg, s6r. physiol., 24, HITCHINGS, G. H., ELION, G. B., AND VANDERWERFF, H The limitations of inhibition analysis. J. Biol. Chem., 174, KING, T. E., AND CHELDELIN, V. H Pantothenic acid studies. IV. Propionic acid and,b-alanine utilization. J. Biol. Chem., 174, MCILWAIN, H., AND HUGHES, D. E Biochemical characterization of the actions of chemotherapeutic agents. 2. A reaction of haemolytic streptococci, involving pantothenate-usage, inhibited by pantoyltaurine, and associated with carbohydrate metabolism. Biochem. J., 38, MENTZER, C., MEUNIER, P., AND MOLHO-LACROIX, L Faits de synergie et d'antagonisme entre la chloromyc6tine et divers amino-acides vis-a-vis de cultures d'e. coli. Compt. rend., 230, MUELLER, J. H., AND MIILLER, P. A Inhibition of tetanus toxin formation by D- serine. J. Am. Chem. Soc., 71, NIELSEN, N., HARTELIUS, V., AND JOHANSEN, G Uber die Antiwuchsstoffe der Pantothensaure und des f-alanin. Coimpt. rend. trav. lab. Carlsberg, ser. physiol., 24, RAVEL, J. M., AND SIIIVE, W Prevention of pantothenic acid synthesis by cysteic acid. J. Biol. Chem., 166, SCHULTZ, J Personal communication. SHIVE, W The utilization of antimetabolites in the study of biochemical processes in living organisms. Ann. N. Y. Acad. Sci., 52, SHIVE, W., AND MACOwV, J Biochemical transformations as determined by competitive analogue-metabolite growth inhibitions. I. Some transformations involving aspartic acid. J. Biol. Chem., 162, SNELL, E. E., AND GUIRARD, B. M Some interrelationsliips of pyridoxine, alanine,

13 1950] PANTOTHENATE STUDIES 745 and glycine in their effect on certain lactic acid bacteria. Proc. Natl. Acad. Sci. U. S., 29, SPRINSON, D. B., AND CHARGAFF, E The occurrence of hydroxypyruvic acid in biological systems. J. Biol. Chem., 164, SUTER, E Personal communication. TEAS, H. J Mutants of Bacillus subtilis that require threonine or threonine plus methionine. J. Bact., 59, WACHSTEIN, M Nephrotoxic action of dl-serine in the rat. Arch. Path., 43,