crowded plate method
A METHOD FOR INVESTIGATING LARGE MICROBIAL POPULATIONS FOR ANTIBIOTIC ACTIVITY'
ALBERT KELNER2.8 Laboratories of the Medical School, University of Pennsylvania, Philadelphia, Pennsylvania
Received for publication March 31, 1948
In their search for new antibiotics, investigators throughout the world have isolated from nature and tested thousands of microbial strains. Nevertheless only a few useful antibiotics have been found and, from this, it is obvious that only a minute fraction of the microflora of nature produce desirable antibiotics. A method that would drastically increase the number of strains screened without proportionate increase in labor would improve the chances of finding new antibiotics. It would also be useful for the selection of high-yielding variants in a pure culture of an antibiotic producer. The most serious drawback of earlier methods for testing large populations was that they detected only certain definite types of antagonists. Thus the "crowded plate" method (Waksman, 1945) detected only soil organisms inhibiting other soil organisms and producing an antibiotic after not more than one or two days of incubation. Foster and Woodruff's method (1946) detected only antagonists that produced antibiotics on media supporting the growth of saprophytic mycobacteria and that inhibited such bacteria. The development (Williston et al., 1947) of various media on which both pathogenic tubercle bacilli and antagonists will grow further extends the usefulness of Foster and Woodruff's method (1946) but does not remove its fundamental limitations. Stansly (1947) and Wilska (1947) described methods for spraying sensitive bacteria onto plates on which colonies of soil organisms had grown for several days. It was thus possible to test antibiotic action against a greater variety of bacteria than was possible before. Spray methods are still, however, restricted to media on which both antagonistic colonies and sensitive bacteria can grow. In addition, there is the ever-present possibility that inhibition zones would be caused by the exhaustion of nutrients around a colony, or by an inhibitory pH, rather than by true antibiotic production. The latter objection applies also to methods that add test bacteria by flooding the surface of the plate with a liquid suspension of sensitive organisms. Fleming (1942) suggested coating a plate on which an antagonist has grown with a layer of sterile agar, on which a sensitive bacterium could be streaked. He sought to test individual isolates, and did not describe methods for testing large populations. With Fleming's agar layer suggestion as a basis, a method has been developed whereby large populations can be tested for many types of
1 This work was aided by a grant from Smith, Kline, and French Laboratories. Present address: The Biological Laboratory, Cold Spring Harbor, Long Island, New York. 8 With the technical assistance of Betty Morgan. 157
antagonists capable of inhibiting various microorganisms, or even antagonists inhibiting nonmicrobial organisms.
The material serving as the source of antagonists is plated on agar so as to have 30 to 50 evenly distributed isolated colonies. After suitable incubation the colonies are covered with a layer of sterile agar, followed by a layer of agar containing the organism whose inhibition is to be tested. After further incubation clear zones over certain colonies disclose the antagonists. The addition of agar layers is a tecbnique remarkably well adapted to replication, and if each step in the procedure is made simple enough, 200 to 1,000 plates can be prepared in a few hours by two workers. From 6,000 to 50,000 colonies are screened in an afternoon. The agar layers, in the order to which they are added to the plate, are: (1) Foundation layer: a layer of about 15 ml sterile agar medium, of a composition suitable for the growth of the antagonist; e.g., peptone beef extract, corn meal, Czapek, wort agar, etc. (2) Seeding layer: a layer of 0.5 ml soft agar medium (0.25 per cent agar), in which is diluted the microflora to be tested. The dilution is calculated to give 30 to 50 colonies per plate.. The layer of soft agar soon dries down to such a thinness that all colonies are surface colonies. If it is not necessary to have all the colonies absolutely on the surface, 5 ml of a nutrient agar containing 1.5 per cent agar can be substituted for the softer medium; this is even better adapted for mass replication. (3) Diffusion, or barrier, layer: 5 to 10 ml sterile agar of special composition, whose use is illustrated in problem 3. (4) Test layer: 3 or more ml of agar medium (0.5 to 1.5 per cent agar) containing a suspension of bacteria whose inhibition is to be tested. The actual working out of the method and its flexibility will be illustrated by describing its use in four diverse problems in antibiotic screening. Problem 1. To test soil for antagonists against Escherichia coli-the antagonists in the soil to grow on nutrient agar. A suspension of soil in distilled water was centrifuged lightly to remove coarse particles. The supernatant was asayed on nutrient agar for viable cells, then stored at 5 C. The colony count was determined after incubation for 2 to 4 days at 28 C. Foundation layers of 15 ml nutrient agar were added to 200 plates. Since it was not imperative to have the soil colonies strictly surface colonies, a seed layer of 5 ml nutrient agar (1.5 per cent agar) was used. The soil suspension was diluted so that 5 ml contained 50 viable cells. The last step in the dilution was made into 500-ml portions of melted nutrient agar, kept in a 45 C water bath. Five-ml seed layers were pipetted onto the foundation layer, and the plates were then incubated at room temperature for 1 week. There were now 200 plates containing about 50 colonies each, or a total of 10,000 colonies. The diffusion layer was omitted, and a test layer was added directly over the soil colonies. To prepare the test layer, 20 ml of a 24-hour broth culture of Escherichia coli were added to 1,000 ml nutrient crystal violet agar kept in a 45 0 water bath. Five ml of this material were pipetted to each of 200 plates to form
ANTIBIOTIC ACTIVITY IN MICROBIAL POPULATIONS
the test layer. The small concentration of crystal violet, 1:750,000, prevented soil fungi, actinomycetes, and gram-positive bacteria from growing up into the test layer, but it did not appreciably inhibit the growth of E. coli. After incubation at 37 C overnight numerous clear zones were found (figure 1). Problem 2. To test soil for antagonists against Staphylococcus aureus the antagonists in the soil to grow on corn meal agar. A suspension of soil wvas prepared as described for problem 1 and assayed on corn meal agar. Use of this medium would promote the growth of colonies that may have been suppressed or overgrown on nutrient agar, and, in addition, a different nutritive medium for antibiotic production would probably reveal antagonists missed on nutrient agar. The foundation and seed-layer medium was corn meal agar. Twenty ml of a 24-hour broth culture of S. aureus w-ere added to 1,000 ml melted nutrient agar at
on Figure 1 Typical plates of soil organisms growin1 nutrient agar and tested by the addition of a test layer of crystal violet nutrient agar containing Escherichia coti. Left: Plate showing large inhibitioni zone. In most cases the active colony is the, one at the mathematical center of the zone. Right: Typical plate with no active colonies.
45 C, and test layers of 5 ml each were added to the plates. Since crystal violet inhibits S. aureus it could not be uised; howiever, by incuibating the test layers for 6 hours instead of over-night, the soil colonies did not grow up into the test layer to any appreciable extent, whereas the S. aureits colonies in this period grew sufficiently to make inhibition zones clearly visible. It was convenient to add test layers in the afternoon, store the plates overnight at 5 C, then incubate at 37 C for 6 to 8 hours. This experiment illustrates the use of different media in the foundation and test layers. Problem 3. To study the variation in antibiotic activity of strains of a puire cultutre of an actinomycete (A-13) which was a good antibiotic producer. If the spores of actinomycete A-13 were plated on nutrient agar and tested against E. coli, the size of the zones of inhibition around individual colonies could be taken as a measure of the antibiotic activity of the organisms in each colony. Those colonies with exceptionally large zones of inhibition could be isolated for further study.
A preliminary experiment showed (1) that all colonies had to be on the surface, for even partly buried colonies had distinctly smaller zones than the surface colonies, and (2) when a plate contained more than 2 or 3 colonies the large (60-mm radius) inhibition zones overlapped, making impossible the comparison of zone size. The use of a 0.5-ml seed layer of soft (0.25 per cent agar) nutrient agar readily ensured surface colonies. The spores were diluted in the melted agar kept at 45 C in a water bath. Aided by tilting and shaking of the plates, the 0.5-ml inoculum spread easily, especially over freshly poured, solidified but still warm foundation layers. Two people working as a team could inoculate 200 to 300 plates in an hour. The large zones of inhibition caused a more serious difficulty, for unless plates containing 30 to 50 colonies could be tested, the aim of testing large populations was defeated. Smaller average zone size might be obtained by (1) using a medium less favorable for antibiotic production than nutrient agar, (2) growing all of the colonies subsurface, or (3) testing the colonies after only 2 or 3 days of incubation, before the peak of activity was reached. Such measures were discarded because it was feared that suboptimum conditions for antibiotic production would result in a nonsignificant increase in variability of zone sizes. The use of organisms more resistant to the antibiotic than E. coli might solve the problem. Serratia marcescens, Eberthella typhosa, Alcaligenes faecalis, Escherichia communior, and Bacillus mycoides were tried, but none gave suitable small zones. Continued research with other bacterial species might have disclosed strains more suitable than those tried. A simple physical method was discovered, however, by which one could at will decrease the zone size to any desired degree. A 10-ml sterile diffusion layer was interposed between the A-13 colonies and the test layer. If the diffusion layer was simple nutrient agar, the additional distance the antibiotic had to diffuse resulted in a decrease in zone size of a few millimeters at most. But incorporation into the diffusion layer of substances that would partially adsorb the antibiotic would result in a sharp decrease in zone size. Several adsorbents were tried, and decolorizing carbon (norit A) was found satisfactory. The zone size varied inversely with the concentration of norit A in the diffusion layer (table 1). The decrease in zone size varied from experiment to experiment, and in actual practice 0.75 per cent norit A in a 10-ml diffusion layer of nutrient agar gave satisfactorily small zones. Norit A was added to the nutrient agar before autoclaving. Since the heavy powder settled rapidly in melted agar, it was necessary to shake the suspension just before use. The use of a diffusion layer as outlined is capable of widespread application, since most antibiotics can be adsorbed with one substance or another. The procedure finally adopted for study of this active actinomycete A-13, was as follows: A suspension of conidia of A-13 was filtered six times through absorbent cotton, assayed on nutrient agar, and then stored in the refrigerator until the
ANTrIBIOTIC ACTIVITY IN MICROBIAL POPULATIONS
assay colonies had grown. To petri dishes were then added foundation layers of nutrient agar, and 0.5 ml of soft (0.25 per cent agar) nutrient agar suspension of conidia calculated to contain 40 viable spores per 0.5 ml. The plates were incubated at 28 C for 5 to 7 days. A diffusion layer of 10 ml nutrient agar containing 0.75 per cent norit A was added, followed by a test layer of 3 ml nutrient agar (0.5 per cent agar) containing 1 to 2 ml of 24-hour broth culture of E. coli per 100 ml agar. After incubation overnight at 37 C, zone sizes of the actinomycete colonies were measured. Problem 4. To test soil for antagonists against a green plant. By substituting small plant seeds for the bacteria in the test layer, the agar overlay method might be used for detecting microorganisms inhibiting (or stimulating, or affecting in another specific way) germinating seeds. The inhibition of germinating seeds by javanicin and oxyjavanicin (Anstein et al., 1946), polyporin (Bose et at., 1948), and penicillin (Ribeiro, 1946; Smith, 1946) has been reported. TABLE 1 Effect of norit A on zone size
CONCENRAION OF NORIT A IN DIPMSION LAYZR
AVERAGE ZONE SIZE (RADIUS)
% 0 0.3
0.5 0.7 0.8 1.0
16 13 8
The several days necessary for seed germination (as compared to the 6- to 18-hour incubation period for bacteria) complicated the method by allowing time for soil organisms to grow up into the test layer; and no simple means for testing large populations was found. Nevertheless, the success obtained by the following procedures illustrates the adaptability of the method. Sterile filter paper was placed in the bottom of petri dishes, followed by a 20-ml foundation layer of corn meal agar, and a 3-ml seed layer of a soil suspension diluted in corn meal agar to give 40 colonies per plate. The plates were incubated at room temperature for 1 week. Before adding the seeds, the entire agar layer was carefully overturned into a sterile petri dish, so that the underside of the foundation layer was now on top and the soil colonies on the bottom. It helped to use a spatula inserted between the petri dish rim and the agar layer, and under the filter paper. By peeling off the filter paper a clean sterile surface was uncovered on which to add the test layer, with lessened danger of contamination by soil organisms. Redtop grass seeds (Agrostis alba) were surface-sterilized by being soaked in a solution of sodium hypochlorite for 10 minutes, then washed in sterile distilled water. The test layer medium consisted of 7 ml of half-strength Shive solution (Miller, 1931) in melted agar at 45 C, to which had been added about 1 part in 5
of the seeds. The test layer was pipetted onto the foundation layer, and the seeds were spread uniformly over the surface by gentle tilting of the plate. The plates were incubated in diffuse light at room temperature for 3 days. Despite some contamination around the edge of the plates, most of the seeds germinated cleanly. Clear-cut circular zones of inhibition were seen around some soil colonies.
A general method is described whereby large numbers of microbial colonies from diverse sources can be tested for antibiotic activity. Use of the method is illustrated by descriptions of experiments in which (1) a soil flora is tested for strains producing antibiotics against Escherichia coli, or Staphylococcus aureus; (2) a pure culture of a highly active actinomycete is tested for variants with increased activity; and (3) a soil flora is tested for strains inhibiting the germination of redtop seeds.
REFERENCES ARNSTEIN, H. R. V., COOK, A. H., AND LACEY, M. S. 1946 Production of antibiotics by fungi. II. Production by Fusarium javanicum and other Fusaria. Brit. J. Exptl. Path., 27, 349-355. BOSE, S. R., BOSE, A. B., AND DEY, K. L. 1948 Effect of crude polyporin on seed germination and root growth: a preliminary study. Science, 107, 63. FLEMING, A. 1942 In vitro tests of penicillin potency. Lancet, I, 732-733. FOSTER, J. W., AND WOODRUFF, H. B. 1946 Bacillin, a new antibiotic substance from a soil isolate of Bacillus subtilis. J. Bact., 51, 363-369. MILLER, E. C. 1931 Plant physiology. McGraw-Hill Book Co., New York. RIBEIRO, D. F. 1946 Penicillin action on germination of seeds. Science, 104, 18. SMITH, W. J. 1946 Effect of penicillin on seed germination. Science, 104, 411-412. STANSLY, P. G. 1947 A bacterial spray apparatus useful in searching for antibiotic-producing microorganisms. J. Bact., 54, 443-446. WAKSMAN, S. A. 1945 Microbial antagonisms and antibiotic substances. The Commonwealth Fund, New York. WILLISTON, E. H., ZIA-WALRATH, P., AND YoumANs, G. P. 1947 Plate methods for testing antibiotic activity of actinomycetes against virulent human type tubercle bacilli. J. Bact., 54, 563-568. WILSKA, A. 1947 Spray inoculation of plates in the detection of antagonistic microorganisms. J. Gen. Microbiol., 1, 368.