EFFECTS OF POTASSIUM AND SODIUM IN MICROBIOLOGICAL

EFFECTS OF POTASSIUM AND SODIUM IN MICROBIOLOGICAL ASSAY MEDIA' R. J. SIRNY,' 0. R. BRAEKKAN,3 M. KLUNGS0YR,4 AND C. A. ELVEHJEM Department of Biochemistry, College of Agriculture, University of Wisconsin, Madison, Wisconsin Received for publication January 11, 1954 Studies on the effects of potassium and sodium in media used for microbiological determination of amino acids and vitamins were initiated as a result of the report by MacLeod and Snell (1948) that a competitive relationship existed between sodium (the antagonist) and potasium (the essential metabolite) in several lactic acid bacteria. Inspection of the composition of various amino acid assay media listed in the review by DuInn (1949) reveals that most media contain relatively high amounts of sodium. A similar observation can be made with respect to media used for microbiological determination of vitamins (Snell, 1950). With ratios between sodium and potasium as high as 50 to 1 in some of these media, it was of interest to determine if sodium contributed any deleterious effect under practical assay conditions. The present report deals with general as well as certain specific differences noted in the response of several asay organisms to amino acids and vitamins on media containing varying amounts of sodium and potasium. Because it was observed early in this work that effects in any one organism seemed to vary with the particular amino acid or vitamin which was limiting in the medium, the scope of the studies was expanded to include a survey of the potsiumsodium effects on seven lactic acid bacteria in response to amino acids or vitamins for which they possessed a requirement or for the asay of which they were commonly used. I Published with the approval of the Director of the Wisconsin Agricultural Experiment Station. Supported in part by a grant from the National Live Stock and Meat Board. ' Present address: Department of Agricultural Chemistry Research, Oklahoma A. and M. College, Stillwater, Oklahoma. 3 Present address: Norwegian Fishery Research Institute, Bergen, Norway. Present address: Sildolje OG Sildemelindustriens, Forskningsinstitutt, Damsgftrd, Bergen, Norway. One observation from these studies has been published previously by Chitre et al. (1951). This concerned the finding that D-tyrosine inhibits Leunstoc me8enteroide8, strain P-60, and that this inhibition is almost completely relieved by elimination of sodium from the medium. An increased acid production on such a medium also was indicated. EXPERIMENTAL METHODS Cultures. The organisms used in these studies were LactobaciUus arabinosus, strain 8014,' Ladobacillus casei, strain 7469, Lactobacilus leichmannii, strain 4797, Leuconostoc mesenteroides, strain 8042, Leuconostoc citrovorum, strain 8081, Streptocc faecalis, strain 8043, and Lactobacius delbhkii, strain 3. All except L. l&hmannii were carried as stab cultures on yeast-glucose agar and were transferred bimonthly. L. leichmannii was carried in litmus milk as described by Thompson et al. (1950). Test procedures and media. The microbiological procedures employed in all these studies were essentially those of Henderson and Snell (1948). Their uniform assay medium, with modification and certain exceptions to be described later, was used as a basis for most of this work. For any one assay, three variations with respect to potassium and sodium content were generally used, and these will be termed as follows: (1) the all-k medium, in which all sodium salts were replaced with equimolar quantities of the corresponding potassium salts, (2) the K-Na medium, which was the unchanged medium as customarily used, and (3) the all-na medium, in which all potassium salts were replaced with sodium salts. It is to be emphasized that (1) and (3) are not considered free of either sodium or potassium, respectively, since no attempts were made to remove contaminants from the salts used. However, reagent grade salts were used throughout. A fourth variation, to be termed the ' American Type Culture Collection numbers. 103

104 R. J. SIRNY ET AL. [VOL. 68 low-na medium, was used also in some of the later amino acid work. In this, the sodium citrate of the Henderson-Snell medium (1948) was replaced with potassium citrate, leaving sodium acetate as the only (low) source of sodium. Sodium chloride, which is used in many microbiological media in negligible amounts, was omitted from this and all other media used in these studies. For studies of these effects in vitamin assays, it was convenient to use media in which all except the following were the same as those in the Henderson-Snell amino acid assay medium: (1) sodium acetate (or an equimolar amount of potassium acetate) was substituted for sodium citrate at 4.0 g per 100 ml double strength medium; (2) an acid hydrolyzate of casein, with cystine and tryptophan added back, was used in niacin and pantothenic acid assay media; (3) a tryptic digest of casein, prepared as described by Roberts and Snell (1946) with KHCO3 replacing NaHCO3 as the buffer, was used for the riboflavin medium. The studies on the vitamin B12 assay were limited to a comparison of acid production on the medium described by Thompson et al. (1950) to that on the same medium in which the acid hydrolyzed casein was neutralized with KOH instead of NaOH and in which potassium acetate was substituted for sodium acetate. All constituents of the media were adjusted carefully to ph 6.8 with the aid of a ph meter. KOH was used for neutralization of all-k or K-Na media and NaOH for the all-na modifications. Assays were autoclaved then for 5 minutes at 121 C. Inoculum, prepared by resuspending cells in sterile distilled water if all-na media were being used in the comparisons or in sterile 0.9 per cent KCl in other assays, was added to each assay tube with a hypodermic syringe to increase uniformity of drop size. Incubation time was approximately 72 hours at 37 C, after which the acid produced was titrated with the aid of a quinhydrone electrode and the Cannon titrator.6 All titration values to be presented are expressed in titration counts, with 100 counts approximately equal to 2 ml 0.1 N NaOH. It is also noteworthy that all comparisons to be presented are based on data obtained within single assay racks. This precaution was taken to eliminate 6 International Instrument Co., Los Angeles, California. variation which has been found to occur occasionally between different racks of a single assay. A minimum of 2 tubes was used for establishing each point on the standard curves, and 3 to 4 or more tubes were the customary replication. RESUITS The approach in these studies consisted essentially of a comparison of the standard curves of the response of the organism to graded amounts of the test amino acid or vitamin in media containing varying ratios of potassium and sodium. Since a total of 45 separate assays were so studied, it appears impractical to present all the results obtained. Accordingly, for each organism, standard curves considered typical of the general findings or illustrative of a special observation will be presented. The results with 5 different ratios of potassium and sodium are shown for two organisms in figure 1. A progressive increase in acid production with increasing proportions of potassium is seen for both L. delbrueckii and L. mesenteroides in their response to arginine and proline, respectively. It is seen also that increasing the ratio of potassium to sodium reduces the lag which occurs in the proline standard curve under usual conditions. It should be pointed out that with the exception of the medium with a potassium-sodium ratio of 0:1, all media were richer in potassium than the Henderson-Snell medium. Therefore in subsequent experiments, comparisons of the all-k medium and the all-na medium were made directly to the Henderson-Snell (K-Na) medium. The standard curves for different amino acids for three of the organisms are shown in figure 2. An increased acid production, similar to that seen for proline, is seen for isoleucine and valine with L. mesenteroides on the all-k medium. Results with other amino acids (aspartic acid, cystine, glycine, histidine, leucine, lysine, methionine, threonine, and tryptophan) were of similar magnitude, varying from a slightly increased production of acid to a marked increase such as previously reported (Chitre et al., 1951) for tyrosine with this organism. It is seen also that the all-na medium failed to support growth, an expected finding since MacLeod and Snell (1947) have shown that L. mesenteroides requires potassium under their conditions. A slightly increased activity of L. delbrueckii

19541 POTASSIUM AND SODIUM IN ASSAY MEDIA 105 10O LEUCONOSTOCi, '80 MESENTEROIDES ( X.2L UC I N%<T~~~ESENT s Z 100 LACTOBACILLUS,LACT08ACILLWS - D DEoRECI DELB BRUCKI JJG L-AMINO ACID PER ML Figure 1. Effect of different ratios of potassium and sodium in the medium on the response of Lactobacillus delbrueckii, strain 3, and Leuconostoc mesenteroides to arginine and proline, respectively. One hundred titration counts equal approximately 2 ml 0.1 N NaOH. on the all-k medium is seen also in figure 2 in the standard curves for isoleucine and leucine. Slight increases such as these also were found in response to histidine, phenylalanine, and valine. Substantial growth was promoted by the all-na medium in these instances, perhaps indicative of sufficient potassium contamination in the medium to meet the requirement of this organism. Also presented in figure 2 are results with S. faecalis for threonine and leucine, chosen as typical of the responses obtained with this organism under the different conditions. Other amino acids studied included histidine, isoleucine, methionine, and tryptophan. The slight increase in acid production promoted by the all-k medium is noted again. However, the results on the all-na medium are surprising in that response to the lower levels of the amino acids is greater on the all-na medium than on either the all-k or the K-Na medium. At higher concentrations, the curves level off considerably below those on the two K containing media, but this may conceivably be due to the possibility that potassium has become limiting. The effect of a small addition of potassium to the all-na medium has not been determined for this organism though it would appear to be of definite interest. The curves shown in figure 3 illustrate the potassium-sodium effects on L. citrovorum and L. arabinosus. The response of the former to methionine and the latter to glutamic acid is presented as typical of the results found for ISOLEUCINE c F g.100 FAECLI a FSATECALI F LEUCINE.4 THREONINE - L EUCINE '0 2.4.6. 1.0 0 2 4 WL MILLILITERS OF STANDARD Figure 2. The response of Leuconostoc mesenteroides, Lactobacillus delbrueckii, strain 3, and Streptococcus faecalis to various amino acids with different levels of potassium and sodium in the medium. Solid line: all-k medium (see text); dash line: K-Na medium; dot-dash line: all-na medium. Concentration of amino acids in,g per ml of standard: DL-Isoleucine, 30; DL-valine, 30; L-leucine, 15; DL-threonine, 30. most amino acids required by these two organisms. However, the responses of L. citrovorum to glutamic acid and of L. arabinosus to valine demonstrate a further improvement which has been noted consistently in these cases, that of a marked increase in acid production appearing concomitantly with elimination or alleviation of lags in the curves by use of an all-k medium. Brickson et al. (1948) have shown that the lag in the valine standard curve of L. arabinoms is due to a competitive inhibition by isoleucine, leucine, and methionine. The cause of the lag in the glutamic acid standard curve of L. citrovorum has not been determined but probably involves inhibition by aspartic acid, which is known to inhibit glutamic acid utilization by L. arabinosus. However, while the valine lag by this latter organism has been eliminated repeatedly by using the all-k medium, the same has not been true for all lags attributed to competitive inhibitions. An example can be seen in a previous paper (Klungsoyr et al., 1951) in which all work

Cl 106 R. J. SIRNY ET AL. [VOL. 68 LACTOBACILLUS =; LACTOBACILWS 180 ARABINOSUS ARABINOSUS,,-~.160 N 140 /.120 / BNIACIN / PNTOTHENIC p '20 ACID LACTOBACILLUS LACTOBACILLUS 140 CASEI.8 LEICHMANNII -.20 cr100. ' MILLILITERS OF STANDARD Figure 3. The response of Leuconostoc citrovorum and Lactobacillus arabirosus to various amino acids with different levels of potassium and sodium in the medium. Solid line: all-k medium; dash line: K-Na medium; dot-dash line: all-na medium. Concentration of amino acids in pg per ml of standard: L-Glutamic acid, 50; L-methionine, 8; DL-valine, 30. was done with an all-k medium; the lag in the isoleucine standard curve is still present. The extension of these studies to vitamin asays permitted observation of the potasium and sodium effects under somewhat different conditions since acetate was used as buffer and a natural protein hydrolyzate served as the amino acid source in the vitamin assay media. Even under these different conditions, results similar to those in the amino acid studies were observed for three of the vitamins. These are shown in figure 4, in which a higher acid production can be seen on the all-k media for riboflavin with L. casei and vitamin B1, with L. leichmannii. The increase in acid production on the all-k medium also is seen for L. arabinosus in response to niacin, but a surprising reversal of this effect is obtained in its response to pantothenic acid. Whereas in response to limiting amino acids (figure 3) this organism grew only slightly on the all-na medium, this latter medium, free of any added potassium, supports the highest growth of the organism when pantothenic acid is the limiting nutrient. This observation, an antithesis of the general finding that an all-k medium supports best acid production by L. arabinosus as 8 / RIBOFLAVIN 7 VI T. B12 20 0.2.4.6.8.2.4.6.8 L0 MILLILITERS OF STANDARD Figure 4. The response of Lactobacillus arabinosus to niacin and pantothenic acid, Lactobacillus casei to riboflavin, and Lactobacillw leichmannii to vitamin B1,2. Solid line: all-k medium; dashdot line: K-Na medium. Concentration of vitamins per ml of standard: Niacin, 0.2 pg; pantothenic acid, 0.1 pg; riboflavin, 0.06 pg; vitamin B12, 0.1 pp. well as the other organisms, is based on several experiments involving response to pantothenic acid.7 Other experiments conducted during the course of these studies confirm the general finding stated above. These experiments included a limited number in which growth of 5 of the organism was measured turbidimetrically after 24 hours' incubation; simila trends were observed, with slightly higher growth being indicated on the all- K medium. In two other experiments, the effect of potasium and sodium on maximum growth of these organisms, i.e., with all nutrients present in excess, was determined after 72 hours' incubation. A significantly higher acid production (209 versus 151 titration counts) was noted on the all-k medium as compared to the K-Na one in 7The authors have learned in a private communication from Dr. E. E. Snell and co-workers that a similar beneficial effect of sodium has been observed in the medium for Lactobacillus bulgaricus in response to the pantothenic acid conjugate, pantethine.

19541 POTASSIUM AND SODIUM IN ASSAY MEDIA 107 TABLE 1 Effect of omission of sodium from an acetate buffered medium* Data are in terms of titration counts measuring the acid production of Lactobacillus arabinosus in response to leucine. Each value is the average of 5 tubes. L-LUCINEI K-Na MEDIUM ALL-K -E mm pg 0 4 7 3 42 50 6 68 79 9 89 100 12 111 118 15 126 129 * Medium (described in text) similar to that reported by Schweigert et al. (1944). the case of L. mesenteroid, but maximum acid production was essentially the same on the two different media with the other 4 organisms. Another experiment to determine whether or not the improvements were associated with the citrate buffer in the Henderson-Snell medium was conducted with an all-k modification of the acetate buffered medium reported by Schweigert et al. (1944). The data obtained are shown in table 1, from which a significantly higher acid production on the all-k medium can be seen. DISCUSSION More information is needed about the specific roles of sodium and potassium in bacteriological media before adequate explanations for the effects described here can be offered. It must be remembered that these effects are observed under conditions in which some required nutrient is present in such low concentration that it limits the growth of the organism. Since some of the sodium-potassium effects vary with the growth limiting nutrient present, it is evident that some kind of interrelationship between the cations and the limiting nutrient is involved. The nature of this interrelationship cannot be ascertained from these studies though an effect of the cations on the absorption of the nutrient from the external medium appears as one possibility. The ion antagonism between potassium and sodium reported by MacLeod and Snell (1948, 1950a, b) also may be involved in the explanation of these effects, but since these studies were made with other nutrients limiting, direct comparison to their work cannot be made. It is possible that the slight but consistent decrease in acid production on K-Na media represents the magnitude of the ion antagonism. This would not account, however, for the more striking examples of higher growth with all-k media, such as in response to valine by L. arabinosus, tyrosine by L. mesenteroides, or riboflavin by L. casei. Furthermore, the growth improvement on the all-na medium by L. arabinomss in response to pantothenic acid actually precludes consideration of sodium as an antagonist in this case. In fact, the possibility that sodium plays a favorable role in the pantothenic acid metabolism of this organism merits further investigation. These studies have indicated, in general, certain advantages which may be gained by use of all-k media in microbiological assays for vitamins and amino acids. However, unless an assay medium actually yields more reliable results, very little is gained by changing to a medium in which the only advantage would be a slightly increased acid production. Though direct comparisons of asay results on all-k media and K-NA media have not been extensive, several instances may be cited in which actual improvement in assay results has been achieved by use of all-k media. It was seen in an earlier paper (Chitre et at., 1951) that in the determination of tyrosine in alkaline hydrolyzates, the inhibitory effect of D-tyrosine noted on a K-Na medium with L. mesenteroides was almost completely eliminated on an all-k medium. Therefore, assay values on the latter medium more closely approached the correct value. In other studies (unpublished) the absence of a lag in the valine standard curve, together with the markedly increased acid production, by L. arabinosus also appears to improve asay results. Analysis of beef round hydrolyzed for 5 hours with 3 N HCl yielded a value of 6.6 per cent valine on the K-Na medium as compared to a value of 5.5 per cent on the all-k medium. The latter value was in agreement with values obtained on samples hydrolyzed for longer periods of time and with other assay organisms; it was considered to be a more correct value. The high, incorrect value on the K-Na medium may be due to a less efficient utilization of free valine than of peptides remaining in the 5 hour hydrolyzate, i.e., peptides would appear to be stimulatory (Klungs0yr et al., 1951). This may account also for the statistically significant difference noted in the average valine content of 32

108 R. J. SIRNY ET AL. [VOL. 68 fish samples (Neilands et al., 1949) for which a value of 4.7 per cent was obtained with L. delbruecii as compared to 5.3 per cent with L. arabinosue. It appears likely that the high value with the latter organism is the less correct one, inasmuch as the analyses were on 5 hour hydrolyzates using a K-Na medium. The fact that the all-k media used in these studies are not absolutely free of sodium does not permit the prediction of results which would be found with truly sodium-free media. Furthermore, the apparent beneficial effects of sodium in certain instances, i.e., with limiting pantothenic acid for L. arabinous and at the lower portions of amino acid standard curves with S. faecalis, make it undesirable to suggest the general use of an all-k medium. However, the elimination of nearly all sodium from most media, including that for pantothenic acid determination, can be strongly recommended. A convenient modification of the Henderson and Snell medium that achieves this end involves replacement of the sodium citrate with potassium citrate, leaving the sodium acetate as the only source of the sodium ion. This "low-na" medium should possess all the improved aspects of the all-k medium noted here without possessing the possibility that undesirable effects might be found in a truly sodium-free medium; preliminary studies in this direction have shown it to be equally as effective as the all-k medium. Such a low-na medium was employed by Alexander et al. (1953). ACKNOWLEDGMENT The authors wish to gratefully acknowledge the assistance of Mrs. Joyce M. Thompson and Miss Alice J. Violante in some phases of these studies. SUNMMRY The use of high potassium-low sodium (or sodium-free) media for microbiological assays has been shown to promote increased acid production by seven different assay organisms. Though generally only slight, this increase was very marked in certain instances and, in some of these, was accompanied by improvement in the shape of the standard curves and in the accuracy of assay results. Certain exceptions to the above were noted, the most striking being with Lactobacilus arabinosws in response to pantothenic acid. In this assay, removal of sodium was shown to be very detrimental, and the possibility of a metabolic role for sodium in this organism is suggested. Excluding the pantothenic acid assay, the use of "low-na" media for most microbiological assay work is recommended. REFERENCES ALEXANDER, J. C., BECKNER, C. W., AND EL- VEHJEM, C. A. 1953 The alanine, cystine, glycine and serine content of meat. J. Nutrition, 51, 319-328. BPICKSON, W. L., HENDERSON, L. M., SOLHJELL, I., AND ELVEHJEM, C. A. 1948 Antagonism of amino acids in the growth of lactic acid bacteria. J. Biol. Chem., 176, 517-528. CHITRE, R. G., SIRNY, R. J., AND ELVEHJEM, C. A. 1951 Inhibition of Leuconostoc mesenteroide$ P-60 by D-tyrosine and its relationship with sodium and potassium content of the medium. Arch. Biochem. and Biophys., 31, 398-405. DUNN, M. S. 1949 Determination of amino acids by microbiological assay. Physiol. Revs., 29, 219-259. HENDERSON, L. M., AND SNELL, E. E. 1948 A uniform medium for determination of amino acids with various microorganisms. J. Biol. Chem., 172, 15-29. KLUNGBSYR, M., SIRNY, R. J., AND ELVEHJEM, C. A. 1951 Effect of incomplete hydrolysis on microbiological determination of amino acids. J. Biol. Chem., 189, 557-63. MAcLEOD, R. A., AND SNELL, E. E. 1947 Some mineral requirements of the lactic acid bacteria. J. Biol. Chem., 170, 351-365. MAcLEOD, R. A., AND SNELL, E. E. 1948 The effect of related ions on the potassium requirement of lactic acid bacteria. J. Biol. Chem., 176, 39-52. MAcLEOD, R. A., AND SNELL, E. E. 1950a The relation of ion antagonism to the inorganic nutrition of lactic acid bacteria. J. Bact., 59, 783-792. MACLEOD, R. A., AND SNELL, E. E. 1950b Ion antagonism in bacteria as related to antimetabolites. Ann. N. Y. Acad. Sci., 52, 1249-1259. NEILANDS, J. B., SIRNY, R. J., SOLHJELL, I., STRONG, F. M., AND ELVEHJEM, C. A. 1949 The nutritive value of canned foods. II. Amino acid content of fish and meat products. J. Nutrition, 39, 187-202. ROBERTS, E. C., AND SNELL, E. E. 1946 An

19541 POTASSIUM AND SODIUM IN ASSAY MEDIA 109 improved medium for microbiological assays with Lactobacillus casei. J. Biol. Chem., 163, 499-509. SCHWEIGERT, B. S., MCINTIRE, J. M., ELVEHJEM, C. A., AND STRONG, F. M. 1944 The direct determination of valine and leucine in fresh animal tissues. J. Biol. Chem., 155, 183-191. SNELL, E. E. 1950 Microbiological methods in vitamin research, pp. 327-05. In Vitamin Methods, Vol. I. Edited by Gyorgy, P. Academic Press, Inc., New York, N. Y. THOMPSON, H. T., DIETRICH, L. S., AND ELVEHJEM, C. A. 1950 The use of Lactobacillus leichmannii in the estimation of vitamin B12 activity. J. Biol. Chem., 184, 175-180.