Countercurrent Distribution Studies on Hamycin

APPuED MICROBIOLOY, Jan.. 1967, p 43-48 Copyright ) 1967 American Society for Microbiology Vol. 15, No. I Printed in U.S.A. Countercurrent Distribution Studies on Hamycin B. N. GANGULI Am V. M. DOCTOR' Department of Biochemistry, Research Laboratory, Hindustan Antibiotics Ltd., Pimpri, Poona, India Received for publication 1 July 1966 ABSTRACT Hamycin, a polyene antifungal antibiotic, was isolated by use of countercurrent distribution. A pattern was obtained by plotting the absorption at 383 m,u of the contents of the various tubes against the tube numbers. The results indicated that the antibiotic contained three fractions, a major fraction (peak 2) comprising 48% of the total activity and two minor fractions (peak 1 and peak 3) comprising 3.62 and 11.32%, respectively, of the total activity. The solid material was isolated by pooling the contents of the tubes containing the major fraction, concentrating this in vacuo, and cooling. The antibiotic activities of the three peaks were evaluated by use of a cup-plate assay method with Paecilomyces varioti as test organism. All three components showed antibiotic activity; however, the preparation obtained from the major fraction showed approximately a 7-fold increase in antibiotic activity, a doubling of the E1`,m value at 383 m,i, and approximately a 2.5-fold decrease in the amino acid content in comparison with the starting material. There was an apparent correlation obtained by plotting the curves of the absorption at 383 m, of the different tubes comprising the major fraction and their biological activities. The importance of countercurrent distribution (3) in the isolation and purification of new antibiotic preparations is well established. The technique has been used to separate endomycin A and B (9), to identify the neomycin complex (7), to separate Trichomycin A and B, and to crystallize Trichomycin A (4, 6). Hamycin, a new polyene antifungal antibiotic, was discovered in 1961 by Thirumalachar, Menon, and Bhatt (8). Its isolation and chemical properties were subsequently reported by Bhate et al. (2). Preliminary studies were conducted by Bhate and Acharya (1) on the purification of this antibiotic using a 5-tube Craig (3) countercurrent apparatus. The present report is concerned with a detailed study on the separation and purification of hamycin, by use of a 2-tube Craig countercurrent apparatus. MATERIALS AND METHODS A 2-tube semiautomatic Craig countercurrent apparatus (Quickfit) was used, and the volume of each phase present in individual tubes was 25 ml. Agitation was carried out for 3 min, followed by a settling time of 1 min to allow the phases to separate completely. All operations were carried out at 24 C in the dark. Solvent systems. Three solvent systems were tried: (A) the system used for the separation of endomycin A and B (1), i.e., ethyl acetate-.1 M ammonium acetate (ph 7.)-n-propanol (3:3:1); (B) the system used by Nakano (6) for the purification of Trichomycin, i.e., chloroform-methanol-.1 M sodium acetate (2:2:1); and (C) methyl isobutylketone-.1 M potassium phosphate buffer (ph 6.8)-methanol (14:1:1). Of the above three solvent systems, the one that was used by Nakano (6) was found to be the most suitable because of the greater solubility of the antibiotic in that solvent system and a rapid phase-breaking property. Procedure. The antibiotic was dissolved in a suitable volume of both phases and was filtered. The solution was distributed in the first few tubes of the Craig apparatus. The number of tubes filled depended on the quantity of the antibiotic dissolved, but this number did not exceed 5% of the total number of tubes. For a 3-transfer experiment (Fig. 1), 1 g of the antibiotic was dissolved in 1 ml of each phase, and 25 ml of each phase was then added to the first four tubes. The fundamental technique (3) was used to apply the first 2 transfers, followed by the singlewithdrawal technique (4) for the next 1 transfers. Chemical assay. The distribution of the antibiotic in the tubes was determined by measuring the absorption at 383 m,u of both phases combined, by use of a Beckman DU spectrophotometer. Biological activity. Determinations of the antibiotic activity were carried out on the contents of each tube 1 Present address: Chemistry Department, University of Houston, Houston, Tex. varioti as the test organism. The procedure used was by use of the cup-plate method with Paecilomyces 43

44 GANGULI AND DOCTOR APPL. MICROBIOL. 8-7 6.9- -5- to) r')!4 *4 tfl z -3 w a -J 4 o.2 a. -1 PEAK I FUNDAMENTAL PEAK 2 SINGLE z 16. w WITHDRAWAL X. en Z 14 D > 12 V 2 4 6 1 12 14 16 16 2 25 23 21 TUBE NUMBE R FiG. 1. Countercurrent distribution pattern obtained after a 3-transfer experiment by use of solvent system B The fundamental procedure (3) was followed to apply the first 2 transfers, and the single-withdrawal technique (4) was followed to apply the next 1 transfers. Symbols:, curve obtained by plotting the absorption at 383 my of the contents ofeach tube (only those tubes were selectedfor estimations which were known, from earlier experiments, to contain the material absorbing at 383 mu); straight line, amino acid percentage; A, antibiotic activity of the contents of the individual tubes measured against Paecilomyces varioti. The activity of tubes 15 and 184 at the base ofpeak 2 was taken as 1 units per ml, and the activities ofall the other tubes were calculated in comparison with this. essentially the same as reported by A. P. Rahalkar (M.S. Thesis, Univ. Bombay, Bombay, India, 1964). The organism was allowed to sporulate in barley flasks, which were prepared according to the following procedure. An aqueous spore suspension was obtained from the barley flasks by suspending the barley grains in sterile, distilled water. The suspension was diluted to give 6% transmission with filter 66 on a Klett- Summerson colorimeter; 2% of this was used to inoculate M2 agar medium (Rahalkar, M.S. Thesis, 1964). A 5-mi amount of the inoculated agar was poured into each petri dish. Samples were plated by means of fish spine beads, and the zones of inhibition were measured after 24 hr of incubation at 37 C. The antibiotic activities of the compounds obtained from the different fractions were also measured by a turbidimetric method (Rahalkar, M.S. Thesis, 1964), with Candida albicans as the test organism. The procedure was as follows. A 9-ml amount of M2 medium broth (8) was placed in a test tube (15 by 19 mm) and was sterilized. Aqueous solutions of the samples were prepared and diluted in a suitable manner. A 1-ml amount of each dilution was added to the above tubes to give a series of concentrations. C. albicans HA 4463 was grown fresh for 24 hr on M2 agar slants (8), and an aqueous suspension was prepared to give 6% transmission with ifiter 66 of a Klett-Summerson colorimeter. One drop of this suspension was added aseptically to the M2 tubes, which were then incubated at 28 C for 24 hr. The growth was measured in a Klett colorimeter with filter 66, and the minimal inhibitory concentration was noted. The activities of all the samples were referred to the activity of the starting material, which was taken as 1 İsolation of compounds. The contents of the tubes forming peak 2 (Fig. 1, tubes 121-155) were pooled and were concentrated under vacuum at 45 C. By cooling overnight at -5 C, a precipitate was obtained. This was filtered, washed with chilled water, dry acetone, and ether, and dried in vacuo. The contents of the tubes comprising peaks 1 and 3 (tubes 1-2 and tubes 21-225, respectively), when processed in a similar manner, did not yield solid materials. Therefore, all studies of the comparison of the properties of peaks 1, 2, and 3 (Fig. 1) were made on the liquid material. 4 z w -3 Q co CL Ln -2 I z 4r

VOL. 15, 1967 COUNTERCURRENT DISTRIBUTION STUDIES ON HAMYCIN 45 1.9-8.7 Z.6 O 5 4 I.- a. 3 Z2 o 1. 3 31 32 33 34 35 36 37 38 39 4 41 42 WAVELENGTH ma FIG. 2. Absorption spectra of the different fractions obtained from a 3-transfer countercurrent experiment (Fig. 1). Allfractions were diluted suitably in methanol. Symbols: O, spectra ofthe starting material; *, spectrum of the contents of tube 138, peak 2, Fig. 1; X, spectrum of the contents of tube 16, peak 1, Fig. 1; A, spectrum of the contents of tube 218, peak 3, Fig. 1. Chromatography and amino acid analysis. The compounds isolated from the peak fractions were hydrolyzed in sealed Pyrex tubes with 6 N HCl as reported by Bhate and Acharva (1), except that the hydrolysis was carried out in vacuo. The hydrolysates were chromatographed in a descending manner on Whatman no. 1 ifiter paper, by use of butanol-acetic acid-water (4:1:5) as the solvent system. The amino acids were detected by spraying the papers with.25% ninhydrin in ethyl alcohol. Quantitative elution and estimation of the amino acid content of the developed spots was carried out by the method of Moore and Stein (5). All amino acids were referred to a graph based on leucine as the standard. RESULTS AND DIscussIoN Figure 1 shows the results obtained for a 3- transfer experiment with solvent system B. The pattern indicated the presence of three fractions (peak 1, 2, and 3). The fraction of peak 3 was highly soluble in the aqueous phase and thus moved along with the solvent (K = 2.65, Fig. 1). The fraction of peak 1, on the other hand, was highly soluble in the organic phase and remained in the first few tubes of the apparatus (K =.56, Fig. 1). However, the fraction of peak 2 was equally distributed in the two phases (K.85, Fig. 1), and = was present in the middle tubes of the apparatus. Peak 2 formed 48%, peak 1, 3.62%, and peak 3, 11.32% of the starting material, when measured as the area under each peak. The results presented in Fig. 2 and Table 1 show that the absorption spectra of the three peaks have four maxima. The starting material showed maxima at 345, 363, 383, and 46 m,u. This slight shift noted above in the maxima at 363, 383, and 46 m,u disappeared when the solid material was isolated from the contents of the pooled tubes comprising peak 2 (tubes 121-155, Fig. 1) and when its absorption spectra were determined. The results presented in Fig. 2 also show that there were significant differences in the ratios of the absorption at 46 min compared with that obtained at 383 m,u; the ratio was higher than 1 with

46 GANGULI AND DOCTOR APPL. MICROBIOL. TABLE 1. Antibiotic activity and absorption spectra measurements of the different fractions of a 3-transfer countercurrent experiment (Fig. 1) Antibiotic Spe- Prepn Absrp ioatot activity against varioti mpm." 1 Original starting ma- 345 1a terial 363 383 46 2 Peak 1, Fig. 1, contents 345 122b of tube 16 361 381 44 3 Peak 2, Fig. 1, contents 345 1,65b of tube 138 361 381 44 4 Peak 3, Fig. 1, contents 345 167b of tube 218 361 381 44 5 Material isolated from - 341a the contents of tubes 11-12, Fig. 1 6 Material isolated from 345 7a the contents of tubes 363 121-155, Fig. 1 383 46 7 Material isolated from - 351a the contents of tubes 156-172, Fig. 1 a The figures for the antibiotic activities of the solid materials (samples 1, 5, 6, and 7) are based on a comparison with the figure for the starting material, which was taken as 1 per milligram. b In the case of liquids (samples 2, 3, and 4), the activities are compared with the activities of tubes 15 and 184 at the base of peak 2 (Fig. 1), which were taken as 1 units per ml. peak 2 (1.59), whereas the ratios were appreciably lower than 1 with peak 1 (.93), peak 3 (.816), or the starting material (.9). The contents of all three peaks were found to possess antibiotic activity against P. varioti. However, the contents of peak 2 were nearly 1 times more active than the contents of peak 1 and peak 3 (Table 1, Fig. 2). Also, the contents of the peak tube (tube 138, Fig. 1) of the peak 2 fraction were 17 times more active than those of the base tubes (tubes 15 and 184, Fig. 1). The antibiotic activity curve (against P. varioti) obtained by measuring the activities of the different tubes comprising peak 2 (Fig 1) paralleled the curve obtained by measuring the absorption at 383 m,u of the same tubes. A similar relationship between the FIG. 3. Zones of inhibition shown against Paecilomyces varioti by the contents of tubes 136, 138, and 14, Fig. 1. The experiment was run in duplicate, and the two larger zones of inhibition were yielded by the contents of tube 138 in contrast to the four smaller zones which were yielded by the contents of tubes 136 and 14. Refer to the text for details of the procedure. antibiotic activity and absorption values was reported by Hattori et al. (4) for Trichomycin. Figure 3 shows an increase in the zones of inhibition of P. varioti produced in tube 138 (Fig. 1) in comparison with those produced in tubes 136 and 14 (Fig. 1). The parallel relationship between the antibiotic activity and absorption curves at 383 my obtained for peak 2 (Fig. 1) was confirmed by measuring the activities of the solid materials isolated from different fractions in the region of peak 2 (from tubes 11-12, 121-155, and 156-172, Fig. 1). Antibiotic activities of these fractions as measured against P. varioti are presented in Table 1. The results showed that the solid material isolated from peak 2 (Fig. 1) was seven times more active when tested against P. varioti. Although the amino acid curve of Fig. 1 showed a slight increase in the amino acid content of the peak tubes of peak 2, there was a 2.5-fold decrease in the amino acid content of Peak 2 as compared with that of the original sample (Table 2). Direct amino acid estimations were also carried out on the liquid contents of peaks 1, 2, and 3 (tubes 16, 138, and 218, respectively, of Fig. 1) according to the procedure of Bhate and Acharya (1). The results showed that tubes 16, 138, and 218 contained 18, 122, and 52,ug/ml of amino acid, respectively. Figure 4 shows the amino acids detected in the hydrolysates of the compound iso-

... VOL. 15, 1967 COUNTERCURRENT DISTRIBUTION STUDIES ON HAMYCIN 47 TABLE 2. Analysis of the compounds isolated from the major fractions (peak 2) of three countercurrent experiments E1m at Biological Sample Prepn 383 M Activity against Microanalysis acid in 6% Cansdidaalbicanscotn methanol (per mg)cotn 1 Starting material 4 1 C, 69.93; H, 9.29; N, 3.78; 9.28 residue, 36.5 2 Material isolated from peak 2 8 4 C, 6.84; H, 8.28; N, 2.61; 2.43 of a 2-transfer experiment residue,.77 3 Material isolated from peak 2, 862 8-3.84 Fig. 1, of a 3-transfer experiment 4 Material isolated from peak 2 of 962 4 C, 6.8; H, 8.5; N, 2.2; 8.9a a double run residue, nil athe data reported for sample 4 are from studies of Bhate and Acharya (1), who carried out amino acid analysis directly on the hydrolyzed material. 4 tamic acid, alanine, valine, tyrosine, and, in some cases, phenylalanine. Table 2 shows the results obtained from the :.. analyses of the fractions isolated from peak 2 of a 2-transfer experiment, peak 2 of a 3-transfer experiment (Fig. 1), and peak 2 of a double run. In the 3-transfer experiment, the solid material isolated from peak 2 gave a 2.5-fold decrease in amino acid content, an 8-fold increase in antibiotic activity, and a 2-fold increase in the Eljm value at 383 ma,u whereas the results obtained for the 2-transfer experiment showed a 4-fold decrease in amino acid content, a 4-fold increase in antibiotic activity, and a 2-fold increase in the Eljm value at 383 m,u. ACKNOWLEDGMENTS..; :. We wish to thank M. J. Thirumalachar for supply-... ing the hamycin used in this investigation, and for his continued interest in the work. The assistance of...: :: K. A. Tendulkar, S. B. Thadani, S. K. Pilgaonkar, :.::. :.. Y. V. Ashtekar, and M. G. Dange in running the s; :.: countercurrent apparatus is gratefully acknowledged. :: i:::: LITERATURE CITED 1. BHATE, D. S., AND S. P. ACHARYA. 1962. Purification of hamycin. Hindustan Antibiot. Bull. 5: FIG. 4. Amino acid pattern obtained after 2 hr of descendingpaper chromatography in butanol-acetic acidwater (4:1:5). (1, 2, 3) Amino acids present in the 2. BHATE, D. C., G. R. AMBEKAR, K. K. BHATNAGAR, 161-168. compound isolatedfrom tubes 121-155 ofpeak 2, Fig. 1. AND R. K. HULYAKAR. 1961. Hamycin. I. Isolation and chemical properties. Hindustan Anti- (4) Amino acid present in the compound isolated from tubes 11-12 of Fig. 1. (5) Amino acids present in the biot. Bull. 3:133-14. compound isolated from tubes 156-172 of Fig. 1. Refer 3. CRAIG, L. C., AND D. CRAIG. 1956. Extraction and to the text for details of the procedure. distribution, p. 171-311. In A. Weissberger [ed.], Technique of organic chemistry, vol. 3. Interscience Publishers Inc., New York. lated from peak 2 (tubes 121-155, Fig. 1) and the 4. HArrORI, K., N. NAKANO, M. SEKI, AND Y. HITATA. compound isolated from tubes 11-12 and 1956. Studies on trichomycin. IV. J. Antibiotics tubes 156-172 of Fig. 1. The results indicated the (Tokyo) Ser. A 9:176-181. presence of lysine, arginine, aspartic acid, glu- 5. MooRE, S., AND W. H. STEIN. 1948. Photometric

48 ninhydrin method for use in the chromatography of amino acids. J. Biol. Chem. 176:367-388. 6. NAKANo, H. 1961. Studies of trichomycin. V. J. Antibiotics (Tokyo) Ser. A 14:68-71. 7. SWART, E. A., H. A. LECHEVALIER, AND S. A. WAKsMAN. 1953. The identity of the neomycin complex as measured by countercurrent distribution and microbiological analyses. J. Am. Chem. Soc. 73:3253-3255. GANGULI AND DOCIOR APPL. MICRoBIoL. 8. THIRUMALACHAR, M. 3., S. K. MENON, AND V. V. BHArr. 1961. Hamycin. I. Discovery and biological studies. Hindustan Antibiot. Bull. 3:136-142. 9. VmoNG. L. C., AND W. A. TABER. 1957. Separation of endomycins A and B and their identification as members of the polyene groups of antifungal antibiotics. Can. J. Chem. 35:1461-1466.