growth. and allowed a maximum number of fungi to grow on an agar (1944) that anisic and benzoic acids at a concentration of 150 ppm prevented

Transcription

1 THE CONTROL OF CONTAMINANTS IN PENICILLIN FERMENTATIONS BY ANTISEPTIC CHEMICALS' S. G. KNIGHT AND W. C. FRAZIER Department of Agricultural Bacteriology, University of Wisconsin, Madison, Wisconsin2 Received for publication July 6, 1945 The commercial production of penicillin requires aseptic procedures that often are difficult to follow in large-scale operation. Fermentations contaminated with penicillin-resistant bacteria not only yield little or no penicillin but also make the recovery process less efficient when mixed with normal fermentations. Therefore, a search was made for a chemical that could be added to the fermentation without depressing penicillin production, yet was capable of preventing or delaying the growth of contaminants. No reference was found to a chemical that could be used for this specific purpose. Tyner (1944) observed that 0.3 per cent boric acid suppressed bacterial growth. and allowed a maximum number of fungi to grow on an agar medium inoculated with a soil dilution. The observation was made by Eastwood (1944) that anisic and benzoic acids at a concentration of 150 ppm prevented the growth of bacteria but not of fungi in plant nutrient solutions. In the present study a number of chemicals were tested for their ability to prevent the growth of bacteria in shake flask penicillin fermentations. Representative data have been selected for presentation. METHODS The procedure for the experimental production of penicillin in shake flask fermentations was essentially that described by Koffier, Emerson, Perlman, and Burris (1945). A medium of the following composition was used: U Corn steeping liquor solids Crude lactose (one crystallization).40.0 Sodium nitrate.3.0 Monobasic potassium phosphate.0.50 Magnesium sulfate (7 H20).0.25 Distilled water to 1 liter. One hundred ml of the medium were put in a 500-ml Erlenmeyer flask and sterilized at 15 pounds' pressure for 20 minutes. The proper amount of the chemical to be tested was added to the medium from a sterile stock solution. In the case of borax, however, it was later shown that sterilizing the borax in the medium produced no evident harmful effects. Different strains of Penicillium notatum and Penicillium chrysogenum were propagated on a "sporulation" medium of the following composition: 1 Published with the approval of the Director of the Wisconsin Agricultural Experiment Station. 2 Supported in part by a grant from the Lederle Laboratories, Inc., in co-operation with the Office of Production Research and Development. 505

2 506 S. G. KNIGHT AND W. C. FRAZIER Agar Sugar beet molasses Peptone Sodium chloride Monobasic potassium phosphate Magnesium sulphate (7 H20) Distilled water to 1 liter. It was found that most of the strains grew faster and sporulated more readily on this medium than on one made by adding agar to the production medium. The inoculum, with one exception, consisted of 1 ml of a sterile water suspension of the appropriate mold spores. It was early noted that the number of spores in the inoculum did not affect the penicillin production of the fermentation. In the exception mentioned, the inoculuim consisted of 10 ml of a 30-hour fermentation per flask of fresh medium; this has been termed "a vegetative inoculum." The three contaminants were isolated from two different contaminated fermentations that developed no penicillin. Two of the organisms were gramnegative, nonsporeforming rods that were similar to Aerobacter aerogenes on IMViC tests, and the third organism was an unidentified micrococcus. The contaminants were maintained on separate nutrient agar slants and were added to the fermentations in a water suspension during the third day of the fermentation. When the contaminants were added at the start of the fermentation they were not so effective in depressing penicillin yields as when added during the third day of the fermentation. Each fermentation was sampled for assay daily, starting on or before the sixth day, through the eighth day. About 5 ml of the liquor were removed aseptically; a determination was made on a portion of this sample, and the remainder was used for penicillin assay according to the method of Schmidt and Moyer (1944). Pilot plant fermentations with Penicillium chrysogenum NRRL1951-B25 were carried out in 80-gallon tanks by a procedure to be described elsewhere. A medium of the following composition was used: Corn steeping liquor solids Crude lactose (one crystallization) Calcium carbonate (added before sterilization) Tap water to 1 liter. These fermentations were not contaminated, being designed only to determine whether 0.2 per cent borax was toxic to penicillin production when employed in large submerged fermentations. An attempt was made to acclimatize strain NRRL832 so that it would grow on the sporulation medium when it contained a normally lethal concentration of formaldehyde. Spores were transferred from a medium containing a level of formaldehyde that would just permit good growth and sporulation to a medium containing a level of formaldehyde that wouldslightly inhibit growth and sporulation, until a concentration of formaldehyde that would -not permit growth was 9

3 CONTROL OF CONTAMINANTS IN PENICILLIN FERMENTATIONS reached. A similar method was used to acclimatize strain NRRL1951-B25 to boric acid. Spores of the mold that had been so acclimatized were used to inoculate fermentations containing normally lethal concentrations of either formaldehyde or boric acid. RESULTS The following chemicals were tested for their ability to prevent the growth of the contaminants and to allow penicillin production by Penicillium notatum NRRL832: acriflavin anisic acid benzoic acid borax boric acid calcium dimethyldithiocarbamate calcium lactate catechol copper sulfate diphenyl "dowicides" A, B, C, G, F formaldehyde furfural gallic acid hexylresorcinol hydroquinone 2,4,hydroxybenzoic acid iodoacetic acid mandelic acid orcinol "phemerol" phenol resorcinol salicylic acid silver nitrate sodium azide sodium chloride sodium fluoride succinic acid sulfanilic acid tartaric acid thymol "zepharin" With the exception of formaldehyde, boric acid, and borax, all of the chemicals were toxic to the mold at a level capable of preventing the growth of the contaminants. When formaldehyde was employed at a level capable of slowing the growth of the contaminants, it exerted a delaying and depressing effect on penicillin production by P. notatum NRRL832 (table 1). Levels of formaldehyde higher than per cent stopped the growth of the mold. Table 2 shows data typical of a limited number of fermentations by strain NRRL832 when boric acid was used to delay the growth of the contaminants. Three-tenths per cent boric acid prevented the growth of the contaminants and allowed penicillin production equal to that in the uncontaminated control; 0.4 per cent boric acid depressed penicillin production. Because boric acid seemed to offer some promise, tests were made with borax (Na2B4O7.7H20) which is less expensive and yields approximately 65 per cent boric acid upon solution. Table 3 presents data showing that 0.2 per cent borax delayed the activity of the contaminants and allowed penicillin production almost equal to that in the uncontaminated control. There is some evidence that strain NRRL832 was sensitive to 0.1 per cent or more of borax. While this work was being done, many different strains of Penicillium notatum and chrysogenum were being tested for penicillin production at this laboratory. The results of testing a limited number of these strains in shake flask fermenta- 507

4 50n8 S. G. KNIGHT AND W. C. FRAZIER tions containing borax are shown in table 4. As may be seen, strains NRRL832, J347, and 1982 produced less penicillin when 0.1 per cent or more borax was TABLE 1 The effect of formaldehyde on contaminants and on penicillin production by Penicillium notatum NRRL83S in shake flask fermentations Each figure is the average of two flasks PENICILLIN PtODUCTION FORMAIDE- Days Days per cent Oxford units per ml * * * * * Flasks contaminated with two gram-negative rods and one micrococcus; 24,000 bacteria per flask. TABLE 2 The effect of boric acid on contaminants and on penicillin production of Penicillium notaum NRRL8SS in shake flask fermentations Each figure is the average of three flasks PENICIN PRODUCTION BORIC ACID Days Days per Cent Oxford units per xa * * * *Flasks contaminated with two gram-negative rods and one micrococcus; 2,600,000 bacteria per flask. added to the medium; strains R38, NRRL1951, and were indifferent to the presence of borax; and strain NRRL1951-B25 seemed to be stimulated by

5 CONTROL OF CONTAMINANTS IN PENICILLIN FERMENTATIONS borax. Because strain NRRL1951-B25 was replacing strain NRRL832 in industrial fermentations further work was confined to NRRL1951-B25. Since the pure cultures that usually were used as contaminants might not be typical of industrial contamination, a fermentation was contaminated with 1 ml of a suspension of incubator dust in water. The dust suspension contained at least 9,000 microorganisms per ml, mostly of the aerobic, sporeforming type. Table 5 compares the penicillin-destroying activity of the mixture of pure cultures and of the organisms in the dust, and again illustrates the ability of borax to delay and even prevent the destruction of penicillin by contaminants. A number of similar fermentations (table 6) gave similar data: 0.2 per cent borax delayed TABLE 3 The effect of borax on contaminants and on penicillin production by Penicillium notatum NRRL885 in shake flask fermentations Each figure is the average of three flasks PENICILLIN PRODUCTION BORAX Days Days j 7 8 per cent Oxford units per ml * S * * * * Flasks contaminated with two gram-negative rods and one micrococcus; 16,000,000 bacteria per flask. penicillin destruction by contaminants through the sixth and seventh day and often through the eighth day of the fermentation. In industry it might be desired to add borax to a fermentation after a mishap had increased the likelihood of contamination. With this thought in mind, borax was added to flasks at the start, after 4 days, and half at the start and the remainder after 4 days of the fermentation. Table 6 shows that 0.2 and 0.3 per cent borax added at the start of the fermentation or half at the start and the remainder after 4 days had no depressing effect on penicillin production. However, when either 0.2 or 0.3 per cent borax was added after 4 days of the fermentation, the penicillin yield was markedly suppressed. This indicates that mold grown in a medium containing borax was more resistant to later additions of the chemical than mold not so acclimatized. Since in plant operation large fermenters are inoculated with a vegetative 509

6 510 S. G. CNlTGHT AND W. C. FRAZIER TABLE 4 0 Borax tolerance of strains of Penicillium notatum and Penicillium chrysogenum in shake flask fermentations Each figure is the average of three flasks PENICILLIN PRODUCTION STRAIN OF MOID BORAX Days Days per cent Oxford units per ml NRRL NRRL1951-B R NRRL NRRL U J

7 CONTROL OF CONTAMINANTS IN PENICILLIN FERMENTATIONS 511 or actively growing mold instead of spores, a comparison was made of penicillin production by strain NRRL1951-B25 with spore, plain vegetative, and acclimatized vegetative inoculum. The results of this comparison are presented in table 7. The spore-inoculated fermentations containing borax produced more penicillin than the other fermentations, including those containing no borax. Generally the fermentations that received the plain vegetative inoculum produced slightly more penicillin than those receiving the acclimatized vegetative inoculum. In this fermentation, as in the one illustrated in table 4, strain NRRL1951-B25 produced considerably more penicillin in a medium containing TABLE 5 The effect of borax on contaminants and on penicillin production by Penicillium chrysogenum NRRL1951-B25 in shake flask fermentations Each figure is the average of three flasks PENICILLIN PRODUCTION BORAX CONTAMINANT Days Days per cent j8 Oxford units per ml 0.0 Control *Cultures tdust Control Cultures Dust Control Cultures Dust Control Cultures Dust * Flasks contaminated with two gram-negative rods and one micrococcus; 1,100,000 bacteria per flask. t Flasks contaminated with incubator dust; 9,000 aerobic sporeformers per flask. 0.2 or 0.3 per cent borax than in a medium without borax. This stimulation was often noted and will be discussed in another paper. Since 0.2 per cent borax was effective in preventing penicillin destruction by contaminants and did not depress penicillin yields in shake flask fermentations, the next logical step was to see what effect borax would have on penicillin yields by strain NRRL1951-B25 in 80-gallon pilot fermentations. The results of two such fermentations are shown in table 8. The inoculum for the fermentation represented in table 8a contained no borax or was not acclimatized, whereas the inoculum for the fermentation represented in table 8b was raised in 0.2 per cent borax or was acclimatized. In the fermentations with the unacclimatized

8 11 01 TABLE 6 The effect of the time of the addition of borax on contaminants and on penicillin production by Penicillium chrysogenum NRRL1961-B1 in shake flask fermentations Each figure is the average of three flasks BORX ADDITION PENICILI PRODUCnON Days Days Start 4 days 1 _ ~~ S 6 j 7 j 8 pr can per cent Oxford units per nd * * * * * * Flasks contaminated with two gram-negative rods and one micrococcus; 1,200,000 bacteria per flask. The effect of the TABLE 7 type of inoculum on penicillin production by Penicillium chrysogenum NRRL1951-BR5 in shake flask fermentations Each figure is the average of three flasks PEMICILLIN PRODUCTION BORAX TYPE OF INOCULUM Days Days J per cent Oxford units per ml 0.0 Spore Vegetative Spore Vegetative Acclimatized Spore Vegetative Acclimatized Spore Vegetative Acclimatized

9 CONIROL OF CONTAMINANTS IN PENICILLIN FERMENTATIONS inocula, the borax-containing tank yielded 11.7 per cent more penicillin than the control tank after 37 hours; whereas in the tanks receiving an acclimatized inoculum, the borax-containing fermentation was 30 per cent ahead of the control after 34 hours. After 44 to 45 hours of fermentation, the borax-containing tank that received the unacclimatized inoculum yielded 24.6 per cent less penicillin than its control, whereas the borax-containing, acclimatized-inoculum tank yielded 5.6 per cent less penicillin than its control. After 61 to 64 hours of fermentation, the borax-containing unacclimatized-inoculum tank yielded 35.6 per cent less penicillin than its control, whereas the borax-containing acclimatized-inoculum tank yielded 15.5 per cent less penicillin than its control. The table shows that in the early stages of both fermentations, borax stimulated penicillin production, but that later the fermentation without borax yielded the TABLE 8 Penicillin production by Penicillium chrysogenum NRRL1951-BR5 in 80-gallon pilot fermentations containing 0.2 per cent borax a. Inoculum grown in medium containing no borax b. Inoculum grown in medium containing 0.2 per cent borax PENICLLIN PRtODUCTION BORAX Hours Hours j per cent Oxford units per ml more penicillin. The latter difference was not nearly so marked when the inoculum was raised in 0.2 per cent borax. It is evident that, although an acclimatized vegetative inoculum did not affect penicillin yields in shake flask fermentation, it was an important factor in the 80-gallon pilot fermentations. The most striking and consistent observation in regard to was that in contaminated flasks which contained no borax, the either failed to rise or dropped sharply. Fermentations containing 0.2 per cent borax usually had a about 0.1 to 0.3 units higher than those containing no borax. It was found that Penicillium notatum NRRL832 would grow and sporulate normally on a sporulation medium that contained per cent formaldehyde but that growth and sporulation were delayed when the medium contained per cent formaldehyde. By successively transferring spores from a medium,513

10 G. KNIGHT AND W. C. FRAZIER containing a nonlethal level of formaldehyde to one containing per cent more formaldehyde, it was eventually possible to obtain growth and sporulation on a medium containing per cent formaldehyde. However, spores that were produced on a sporulation medium that contained per cent formaldehyde would not germinate in a fermentation medium that contained per cent formaldehyde and produced poor growth and no penicillin in a medium containing per cent formaldehyde. Penicillium chrysogenum NRRL1951-B25 was acclimatized to grow and sporulate on a medium containing 3.2 per cent boric acid, whereas the unacclimatized mold grew and sporulated poorly on 1.4 per cent boric acid. The unacclimatized mold grew readily but produced a maximum of 11 units of penicillin in shake flask fermentations containing 0.6 per cent boric acid. When spores that were produced on a sporulation medium that contained 3.2 per cent boric acid were used to inoculate shake flask fermentations that contained 0.6 per cent boric acid, good growth was obtained, but the maximum yield of penicillin was 19 units per ml. DISCUSSION In conducting this study it was understood that much of the data obtained in shake flask fermentations might not be directly applicable to large industrial fermentations. However, it was thought that the discrepancy would be at a minimum in this work, since a chemical capable of allowing penicillin production in a contaminated shake flask would have the same effect on the contaminants, and probably the mold, in a larger fermentation. An endless number of compounds might be studied in an attempt to find one that would suppress the growth of bacteria and still permit the mold to produce near its maximum amount of penicillin. Among the chemicals that were tested were common antiseptics as well as enzyme inhibitors. The observation that borax (or boric acid) might be the sought-for chemical was encouraging because it is unlimited in amount, inexpensive, chemically stable, and apparently nontoxic (Frost and Richards, 1945). No statement can be made regarding the effect of boric acid or borax on the recovery of penicillin from the fermentation liquor. The microorganism or group of microorganisms that could be used in determining the value of the chemicals are unlimited in number. Certainly the type of contaminant varies with the location and design of each penicillin plant. The gram-negative, nonsporeforming rods and the micrococcus used in this work were isolated from contaninanted fermentations and represent bacteria often found in air, dust, and cooling water. Not only were the contaminants added in large numbers (usually over a million per ml), but they also were added when actively growing and at a time when the fermentation was at a suitable for their growth. Most of the bacteria added with the dust suspension were of the sporeforming type, but the results were no different than with the nonsporeforming cultures. Undoubtedly there are bacteria that will grow and destroy penicillin in a

11 CONTROL OF CONTAMINANTS IN PENICILLIN FERMENTATIONS fermentation containing 0.2 per cent borax. It might be expected, however, that 0.2 per cent borax would delay their growth long enough, 24 to 36 hours, sb that considerably more penicillin would be obtained than if borax were not present. This might be very important in industrial plants where the fermentation is complete in 40 to 60 hours. Limited time and the need for other experiments curtailed the number of fermentations possible in 80-gallon pilot fermenters. The t"wo fermentations in pilot tanks show that in 40 hours 0.2 per cent borax did not significantly decrease the penicillin yield, especially when the inoculum had been raised in a medium that contained borax. With a longer fermentation, over 40 hours, the borax-containing tanks yielded less penicillin than the controls; here again, however, the difference was less when the inoculum had been raised in a medium containing borax. The logical procedure in industry would be to start the inoculum in 0.2 per cent borax and thus minimize accidental contamination that might occur in the inoculum tanks. Experiments with shake flask fermentations indicate that the addition of borax to tank fermentations after penicilli production has started probably would cause low penicillin yields. The fact that different strains of Penicillium notatum and Penicillium chrysogenum vary in their tolerance to borax should be kept in mind when changing from one strain to another. It is possible, however, that the differences in tolerance to borax would not be so marked in large fermentations as they were in shake flasks. The experiment on acclimatization seems to indicate that strain NRRL1951-B25 can be acclimatized so that it will grow in a medium that contains a normally lethal concentration of boric acid, but that the penicillin-producing mechanism will remain inhibited. From the results cited above it is believed that the following general statements are warranted: 0.2 per cent borax in a shake flask fermentation is capable of suppressing the growth of added contaminants long enough to permit penicillin production equal to the controls; different strains of Penicillium notatum and Penicillium chrysogenum vary in their tolerance to 0.2 per cent borax; penicillin production by Peniciltium chrysogenum NNRL1951-B25 in shake flask fermentations was not depressed when 0.2 to 0.3 per cent borax was added at the start of the fermentation; penicillin production by strain NRRL1951-B25 in 80-gallon pilot fermenters was first stimulated and later decreased by 0.2 per cent borax; in the later stages of pilot fermentations, borax did not decrease penicillin yields so much if the inoculum had been grown in 0.2 per cent borax. SUMMARY In all, 37 different chemicals were tested for their ability to prevent the growth of contaminants and still allow penicillin production in contaminated shake flask fermentations. Of the chemicals tested, only borax and boric acid could be used at a level high enough to delay the growth of contaminants and still not interfere with penicillin production. 515

12 516 s. G. KNIGHIT AND W. C. FRAZIER Two-tenths per cent borax in the fermentation medium delayed the growth of contaminants and allowed normal penicillin yields by Penicillium chrysogenrum NRRL1951-B25. Of 10 different strains of Penicillium notatum and Penicillium chrysogenum tested to determine their tolerance to borax, strains NRRL1951-B25, NRRL- 1951, R38, and were tolerant to 0.2 to 0.3 per cent borax. Penicillin production by strain NRRL1951-B25 was sometimes greatly stimulated in fermentations that contained 0.2 to 0.3 per cent borax. Strain NRRL1951-B25 was acclimatized so that it grew in a medium that contained a normally lethal concentration of boric acid, but penicillin production remained inhibited. Eighty-gallon pilot fermentations with strain NRRL1951-B25 in a medium containing 0.2 per cent borax yielded more penicillin at 30 hours than similar fermentations without borax;, later the fermentations without borax yielded more penicillin than those with borax. ACKNOWLEDGMENT The work reported here is paxt of a large co-operative project on penicillin that has been done under government contract at the University of Wisconsin. The authors of this paper are indebted to Drs. M. P. Backus, R. H. Burns, M. J. Johnson, E. McCoy, W. H. Peterson, and J. F. Stauffer of the university staff for counsel in the planning and execution of the work. The experiments involving tank fermentations were performed by J. J. Stefaniak, B. H. Olson, and F. B. Gailey under the direction of Dr. M. J. Johnson, and assays were made by Mrs. Margaret Dodson. REFERENCES EASTWOOD, T. M Bacteriostatic and fungistatic acqion of some organic chemicals. Science, N. S., 100, 10. FROST, D. V., AND RICHARDS, R. K The low toxicity in animals of boric acid as a preservative agent. J. Lab. Clin. Med., 30, KOFFLER, H., EMERSON, R. L., PERLMAN, D., AND BURRIS, R. H Chemical changes in submerged penicillin fermentations. J. Bact., 50, SCHMIDT,W. H., AND MOYER, A. J Penicillin. I. Methods of assay. J. Bact., 47, TYNER, L. E Effect of media composition on the numbers of bacterial and fungal colonies developing in petri dishes. Soil Sci., 57,