METABOLiTE REVERSAL OF ANTIBIOTIC INHIBITION, ESPECIALLYREVERSAL OF AUREOMYCIN INHIBITION BY RIBOFLAVIN'

J. W. FOSTER AND ROBERT F. PITTILLO

Department of Bacteriology, University of Texas, Auistin, Texas

Received for publication May 2, 1953

An approach to the mechanism of action ofantibiotics is to ascertain metabolites which areinvolved directly in the inhibition of the bacterialcell by a specific antibiotic. A simple means oftesting involvement of metabolites in antibioticinhibition derives from the demonstration (Fos-ter and Pittillo, 1953) that complex naturalmaterials, in small amounts, permit growth inotherwise inhibitory concentrations of penicillin,dihydrostreptomycin, aureomycin, chloramphen-icol, or teriamycin. Reversal of inhibition causedby certain antibiotics has been reported forspecific metabolites, e.g., actathiazic acid bybiotin or pimelic acid (Grundy et al., 1952),erythromycin by pantothenic acid, g-alanine orl-carnosine (Brown and Emerson, 1953), chlor-amphenicol by phenylalanine, tyrosine, tryp-tophan, or glycine (Woolley, 1950; Mentzeret al., 1950), and streptomycin by pantothenicacid (Lichstein and Gilfillan, 1951). However,in most cases the experiments were conductedin liquid media with relatively heavy inocula,and, from the data presented, proof of reversalis not unequivocal owing to lack of informationas to sensitivity of cells in the "reversed" cul-tures. Theoretical aspects of reversal relative toprocedures, with particular emphasis on selec-tion of resistors by complex supplements inliquid media, have been covered in our previouspaper (Foster and Pittillo, 1953).The present paper describes experiments on

each of the first mentioned five antibiotics listedabove in which reversal was tested with variousmetabolites. The degree of reversal obtainedranged from slight to very marked; even in thecase of the latter there was a limit to the re-versibility. Deducible from these results is thatdisturbance of the metabolism of the metabolite

I Supported by a grant from the Williams-Waterman Fund for the Combat of DietaryDiseases, and in part by a grant-in-aid from theAmerican Cancer Society upon recommendationof the Committee on Growth of the NationalResearch Council.

is at least part of the effect caused by an anti-biotic when it inhibits the growth of a suscepti-ble organism.

MATERIALS AND METHODS

The spread plate procedure using 1\1-9 agar (An-derson, 1946) and Escherichia coli, strain B, hasbeen employed exactly as described previously(Foster and Pittillo, 1953); repetition of thedetails is unnecessary. Briefly, the method in-volves preparation of replicate series of M-9agar petri plates containing graded doses of anantibiotic, differing by a factor of two. Someof the series received specified concentrationsof metabolites; others, none. The concentrationof antibiotic required for complete inhibition inthe presence of the metabolite, compared to thatin the absence, is called "reversal factor". Re-covery of fewer colonies than on control platesis regarded as "partial reversal". The inoculumconsisted of 50 to 150 twice-washed bacterialcells spread over the entire surface of the agar,and colony counts generally were made after36 hours at 37 C.The antibiotics and the metabolites were dis-

solved in sterile distilled water and added withoutfurther sterilization to cooled melted sterilizedMI-9 agar in sterile petri plates and the wholethoroughly mixed. Appearance of a colony,however small as compared to control colonies,was regarded as evidence of failure of the anti-biotic to inhibit. Absence of colonies visible tothe naked eye was the criterion of inhibition.Occasional contaminant colonies were recognizedeasily. Most of the experiments have been per-formed at least twice and with single, duplicate,oIr triplicate plates according to the requirementsfor accuracy.

RESULTS

Preliminary screening of metabolites for re-versing activity against the individual antibioticswas performed separately with the following

478

on March 6, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

METABOLITE REVERSAL OF ANTIBIOTIC INHIBITION

mixtures: casamino acids, 1 mg per ml agarmedium; a mixture of the vitamins2 in the Bcomplex; a mixture of purines and pyrimidines3;a mixture of ribosides.4 Virtually all the com-pounds were purchased from Nutritional Bio-chemical Company; their purity was not ascer-tained. A commercial yeast extract, basamine(Anheuser-Busch), 1 mg per ml agar medium,was always run separately as a positive reversalcontrol.

Mixtures displaying a reversal factor generallygreater than 2 were resolved and individualcomponents tested for reversing power at thesame concent-ation5 as they were present in themixture.

RESULTS OF SCREENING SURVEY

A brief summary will be given of the over-allscreening tests for reversal of the individualantibiotics.

Aureomycin. (a) Amino acids. Glycine gave a4-fold reversal. Aureomycin was unique in thatpartial reversal over a 2-fold range was obtainedwith virtually all the amino acids tested, glutamicacid perhaps being most prominent in this respect.

(b) Vitamins. Riboflavin in different experi-ments showed reversal factors ranging from 8 to32. Biotin, thiamin, pyridoxine, and pyridoxa-mine gave up to 8-fold reversal. p-Aminobenzoicacid, folic acid, and Ca pantothenate reversedover a 2-fold range.

(c) Purines and pyrimidines. Two- to 4-foldreversal by the mixture. Uracil showed 4-foldreversal, adenine, guanine, or hypoxanthineshowing none to slight.

2 Final concentration, ,ug per ml agar medium:calcium pantothenate, 2.0; nicotinic acid, 2.0;riboflavin, 2.0; thiamin, 1.0; p-aminobenzoic acid,1.0; pyridoxamine, 1.0; folic acid, 0.3; biotin, 0.01;pyridoxine, 16.6; vitamin B12, 0.001.

3 Final concentration, 5.0,g per ml agar mediumfor each of the following: uracil, guanine, thymine,adenine, hypoxanthine, cytosine, and xanthine.

4 Final concentration, 5.0,ug per ml agar mediumfor each of the following: adenosine, xanthosine,inosine, guanosine, and cytidine.

5 Pure amino acids were tested at a final con-centration of 80 Ag per ml agar medium; those usedwere: glycine, serine, alanine, glutamic acid,aspartic acid, leucine, isoleucine, lysine, valine,isovaline, tryptophan, proline, histidine, methio-nine, threonine, norvaline, arginine, phenyl-alanine, cystine, cysteine.

(d) Ribosides. Inosine and guanosine gaveappreciable reversal.

(e) Basamine. One mg per ml resulted in areversal factor of 8 to 16.

Detailed studies on aureomycin reversal aredescribed below.

Chloramphenicol. (a) Amino acids. Tyrosineand phenylalanine showed complete reversalover a 2-fold range of the antibiotic. This confirmsWoolley's (1950) experiments with E. coli. Higherconcentrations of these amino acids did notresult in a higher reversal factor, making it im-possible to decide if the effect is competitive.Hydrolyzed casein gave only a 2-fold reversal.

(b) Vitamins. The vitamin mixture and thi-amin alone each gave a 2-fold reversal.

(c) Purine and pyrimidines. Two- to 4-foldfactor for the mixture. Individual compoundswere not tested.

(d) Ribosides. The mixture was inactive.(e) Basamine. One mg per ml resulted in a

2- to 4-fold reversal factor.Penicillin G. (a) Amino acids. Hydrolyzed

casein gave a 2-fold reversal. Leucine, tyrosine,histidine, tryptophan, and norvaline showedpartial reversal over a 2-fold concentration ofpenicillin.

(b) Vitamins. The mixture gave a 2-fold re-versal, but the individual vitamins were inactive.

(c) Purines and pyrimidines. The mixturegave a 2-fold reversal. Individual compoundswere not tested.

(d) Ribosides. The mixture was inactive.(e) Basamine. One mg per ml resulted in a

2- to 4-fold reversal factor.Dihydrostreptomycin. (a) Amino acids. Phenyl-

alanine gave partial reversal over an 8-foldconcentration of this antibiotic. Arginine andglutamic acid sporadically gave partial reversalover a small range of antibiotic.

(b) Vitamins. The mixture and the individualvitamins were inactive.

(c) Purines and pyrimidines. The mixturegave a 2-fold reversal. The individuals were nottested.

(d) Ribosides. The mixture was inactive.(e) Basamine. One mg per ml resulted in a

4- to 8-fold reversal factor.Terramycin. (a) Amino acids. Hydrolyzed

casein showed a reversal factor of 4. Arginine,glycine, glutamic acid, leucine, and methionine

47919531

on March 6, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

J. W. FOSTER AND ROBERT F. PITTILLO

showed partial reversal over a 2-fold range ofthis antibiotic.

(b) Vitamins. The mixture and the individualvitamins were inactive.

(c) Purines and pyrimidines. The mixture wasinactive. Individuals were not tested.

(d) Ribosides. The mixture showed no revers-ing activity.

(e) Basamine. One mg per ml resulted in a4-fold reversal factor.Aureomycin was the only antibiotic of the 5

tested in which reversal by metabolites was strik-ing (see below). However, penicillin G, dihydro-streptomycin, chloramphenicol, and terramycinwere definitely reversible over a small concentra-tion range by specific metabolites. That basaminealways showed a higher reversal than any in-dividual compound may indicate that the revers-ing activity of the latter was due to untestedknown or unknown components or to the knownreversing components in combination.

AUREOMYCIN REVERSAL

Reversal separately by vitamins, by ribosides,and by glycine. Tables 1 and 2 illustrate the datafrom experiments in which individual vitamins,individual ribosides, and glycine were shown toreverse otherwise inhibitory concentrations ofaureomycin. In table 1, 1.0 ,ug aureomycin suf-ficed for complete inhibition in the absence ofadded vitamins, whereas in the presence of thelarger concentration of the vitamin supplementsthe level of aureomycin required for completeinhibition was considerably higher. Riboflavinwas outstanding in this respect. In its presence(2 Mug) even 8 MAg of aureomycin were incapableof causing any inhibition. Biotin, thiamin,pyridoxine, and pyridoxamine each increased8-fold the aureomycin level required for inhibi-tion; p-aminobenzoic acid, folic acid, and calciumpantothenate increased it 4-fold; vitamin B12,slightly; and nicotinic acid, none.Each of the ribosides tested had some revers-

ing activity (table 2), inosine being superior inthis respect with an 8-fold reversing factor.Inosine proved to be less effective a reversingagent than riboflavin, less than 1 Mug of the lattercausing a degree of reversal obtainable only witha considerably larger amount of inosine.

Glycine also was a less effective reversingagent than riboflavin (table 2), a reversal factorof 8 being obtained with 80 ,ug of the amino

acid. The reversing power of glycine increasedwith concentration, suggesting a competitiverelation.The proportion between aureomycin level and

concentration of riboflavin required for reversalis clear-cut (table 2), also suggesting a com-petitive relation. The highest level of riboflavintested, 2.0 ,g, was somewhat toxic by itself.

Terramycin can be distinguished readily fromaureomycin because under the conditions de-scribed here inhibition of E. coli, strain B, byterramycin was completely irreversible by ribo-flavin.There appears to be no chemical interaction

between riboflavin and aureomycin, one in-activating the other. Equal parts of these twocompounds in microgram quantities per mlwere allowed to stand 4 hours, autoclaved todestroy the aureomycin, and the riboflavincontent assayed with Lactobacillus casei (Barton-Wright, 1952). There was no loss in riboflavinactivity as compared to a control.

Reversal by combinations of metabolites. Table3 indicates that riboflavin and glycine actsynergistically in reversal of an otherwise in-hibitory concentration of aureomycin. In otherexperiments the same thing has been found forthe combinations riboflavin-inosine and glycine-inosine. Evident in table 3 is that 60 Mg glycinealone had a reversal power allowing 22 out of apossible 82 colonies to develop, and that 0.5 MAgriboflavin alone allowed 38 to develop. However,these two supplements combined gave completerecovery of all the colonies (87).

Similarly, in another experiment, 0.5 ,g inosinealone reversed the inhibition of 21 of a possible56 colonies, and 0.6 Mug riboflavin alone reversedthe inhibition of only 8 colonies. Combined, theactivity of the two supplements was more thanadditive, reversing the inhibition of 46 colonies.Colony recovery with 0.7 MAg riboflavin alone wassubmaximal; in combination with 0.5 MAg inosinethe vitamin gave maximal colony recovery.

Also, in a third experiment, 0.5 ug inosinealone recovered 18 of a possible 56 colonies, and70 MAg glycine alone recovered 37. These concen-trations combined resulted in maximal recovery.Interpretation of these synergism experiments ismade in the discussion.

Reversal of aureomycin inhibition of other bac-teria. Twelve different organisms which grewwell in M-9 medium and which were sensitive to

480 [Vol,. 66

on March 6, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

METABOLITE REVERSAL OF ANTIBIOTIC INHIBITION 4

TABLE 1Reversal by B vitamins of inhibition of Escherichia coli, strain B, caused by aureomycin

The figures represent colonies per plate.

AUR}50^1YCIN, /G PER ML AGAR MDIUM REVERSAL

VITAMIN SUPPLE5MNT AND AMOUNT PER L , FACTORt-(CALCU

0 0.5 1.0 2.0 4.0 8.0 LATED)

None* 128 31 0 0 0 0116 64 2 0 0 0 -

Basamine,* 100 jig 130 107 131 64 2 0 4138 116 120 81 0 0 4

Vitamin mixture, 0.lxt 120 126 60 0 0 0 2Xt 134 114 126 106 111 96 >8

Biotin, 0.001 pIg 120 125 131 67 22 0 80.01 pg 131 126 133 130 97 0 8

Thiamin, 0.1 ,ug 124 131 0 0 0 0 01.0 pg 128 130 117 116 128 0 8

Riboflavin, 0.2 pg 124 118 106 42 0 0 42.0,g 119 134 129 117 130 128 >8

Pyridoxine, 1.7 jug 124 118 130 0 0 0 216.6 pg 107 121 126 131 90 0 8

Pyridoxamine hydrochloride 0.1 pag 137 122 80 21 0 0 41.0 pg 127 133 128 140 129 0 8

p-Aminobenzoic acid, 0.1 pug 120 119 44 0 0 0 21.0 pAg 126 124 127 84 0 0 4

Folic acid, 0.03 pg 127 116 0 0 0 0 00.3,ug 133 130 123 80 0 0 4

Vitamin B12, 0.0001 pAg 127 98 0 0 0 0 00.001 ,lg 120 117 64 13 0 0 2

Nicotinic acid, 0.2 pug 116 0 0 0 0 0 02.0,g 130 13 0 0 0 0 0

Ca pantothenate, 0.2 pg 130 126 17 0 0 0 02.0ug 113 130 121 47 0 0 4

* Duplicate plates in these treatments.t x = final concentrations as listed in footnote 1.t See heading of table 2.

aureomycin were tested for reversal by riboflavin,by glycine, and by inosine.

In addition to the five organisms which arediscussed in the next paragraph, the followingwere tested with results in each case comparableto the other five: Bacillus mesentericus, Bacillussubtilis var. nger, Bacillus subtilis, strain H,

Bacillus cereus, Bacillus mycoides, BaciUus brevis,and Pseudomonas aeruginosa. The species ofBacillus were tested on glucose-glutamic acid-salts medium (Foster and Heiligman, 1949).

Riboflavin was uniformly the most potentreversing compound. In the case of Serratiamarcescens, glycine was almost as potent as

4811953]

on March 6, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

J. W. FOSTER AND ROBERT F. PITTILLO

riboflavin, whereas for BaciUus globigii neitherglycine nor inosine showed significant reversalof aureomycin inhibition. Similarly, inosine was

half as active as riboflavin and glycine in revers-

ing inhibition of BaciUus megaterium but was

considerably less active than those metabolitesin Aerobacter aerogenes and S. marcescene; it hadno significant activity in reversing the inhibitionof B. globigii cells.

TABLE 2Reversal factors for several metabolites tested on

Escherichia coli, strain B, inhibited byaureomycin

The reversal factors are calculated and repre-

sent the ratio of concentration of aureomycinrequired for complete inhibition of colony forma-tion in the presence of the metabolite, to that re-

quired in the absence of metabolite.

MOUNT OPr META11- METABOLITE AND THE CORRESPONDING

OLITE, JAG PER ML REVERSAL FACTORS01r AGAR MEDIUM

Adenosine Xanthosine5 <1 <1

50 2 <1Inosine Guanosine

5 8 250 8 4

Cytidine Riboside mix-ture*

5 2 450 2 4

Glycine RiboflavinjAg per ml

10 2 0.125 020 3 0.25 440 4 0.5 880 8 1.0 16

2.Ot 8

* All five, final concentration of each 5 and 50pg per ml, respectively.t This concentration of riboflavin was some-

what inhibitory in the absence of aureomycin.To ascertain how high a reversal factor is

obtainable with riboflavin where the experimentis not limited by the inhibitory action of highlevels of riboflavin itself, as it was in table 2,search was made for an organism sensitive toaureomycin and insensitive to 10 jig riboflavinper ml of M-9 agar. B. subtilis, strain ATCC6633, satisfied these requirements. In the presenceof 10 ,ug riboflavin per ml the concentration ofaureomycin required for complete inhibition was

128 times greater than that otherwise adequatefor complete inhibition. This is the largest re-versal factor yet obtained.

Effect of pH on riboflavin reversal of aureomycininhibition. In M-9 medium adjusted to differentpH values, optimum reversal of E. coli inhibitionwas obtained over a narrow range, pH 7.2 to

TABLE 3Synergism in reversal activities of riboflavin and

glycine against aureomycinAll plates contained 1.0 pAg aureomycin per ml

agar medium.* The figures represent coloniesper plate, in triplicate.

GLYCINEI Is0 PER ML AGAR MXDIUMpG PZR ML

0 40 50 60 70 80

None 0 0 0 21 52 890 0 0 29 44 790 0 0 17 46 86

22 avg 47 avg 85 avg

0.5 32 49 54 89 86 9137 51 47 90 85 7644 42 56 82 79 90

38 avg 47 avg 52 avg 87 avg 83 avg 86 avg

0.6 62 84 91 86 90 8670 86 87 72 92 8063 89 90 79 88 91

64 avg 86 avg 89 avg 79 avg 90 avg 86 avg

0.7 83 84 76 88 74 7686 80 80 88 82 7779 77 83 81 83 89

83 avg18O avg 80 avg 86 avg 80 avg 81 avg

No aureomycin, 87, 76, 82 = 82 avg.* Inhibition controls, with no riboflavin or

glycine:0.25 pg aureomycin per ml, 41, 67 = 54 avg0.50 ug aureomycin per ml, 0, 11 = 6 avg1.0 ,g aureomycin per ml, 0, 0 = 0 avg2.0 ,ug aureomycin per ml, 0, 0 = 0 avg.

7.4 being demonstrably more conducive to re-versal than pH 7.0 or less or pH 7.6 or above.

Reversal by riboflavin analogues. Galactoflavin,lyxoflavin, and sorbitylisoalloxazine were suppliedgenerously by Drs. Karl Pfister and Karl Folkersof Merck & Co., Inc. No aureomycin reversal

482 (VOL. 66

on March 6, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

METABOLITE REVERSAL OF ANTIBIOTIC INHIBITION

activity could be demonstrated for these com-pounds. Lumichrome, 6,7-dimethylisoalloxazine,a purified preparation made by bacterial oxida-tion of riboflavin (Foster, 1944), had reversalactivity approximately equivalent to that ofriboflavin.

COMPETITIVE RELATION BETW EN AUREOMYCINAND RIBOFLAVIN

Riboflavinless Bacillus subtilis. The competitivenature of the data in table 4 is consistent withthe idea that aureomycin inhibits utilization ofthe riboflavin, which is synthesized, or a pre-cursor thereof (Williams et al., 1950). It becomespossible to put this thesis to a more direct testby using an organism which cannot synthesizeriboflavin, requiring an exogenous supply of thisvitamin. A riboflavinless strain of B. subtilis,strain 137, kindly furnished by Dr. Paul Burk-holder of Yale University was selected becauseof its ability to grow in M-9 medium supple-mented with riboflavin. A cross concentrationspread plate experiment was performed usingB. subtilis, strain 137, as inoculum with syste-matically varied levels of aureomycin and ribo-flavin. The results (table 4) show that the growthinhibition caused by aureomycin at any specificlevel of riboflavin was overcome by a higher levelof the vitamin. This was generally true for levelsof aureomycin and riboflavin up to toxic con-centrations of the latter (64 and 125 ,g per ml).The proportionation evident in table 8 supportsthe mechanism of a competitive relation betweenaureomycin and riboflavin.

Riboflavinless Aspergillus niger. Comparableexperiments with this mutant (obtained in thislaboratory by J. 0. Norman by ultraviolet irra-diation) gave under some conditions indicationsof an interference in riboflavin utilization byaureomycin, but the effects were not clear-cut.

Riboflavin oxidizing bacterium. Pseudomona8riboflavina oxidizes riboflavin as a substrate tolumichrome (Foster, 1944). Aureomycin wasfound to be without effect on the oxidation ofriboflavin by washed cellular suspensions of thisorganism. The oxidation was followed by fluor-imetric determination of residual riboflavin.

PROTECTED INHIBITION

The foregoing results, especially those demon-strating a competitive inhibition of utilizationof one or more metabolites by aureomycin, imply

that however effective aureomycin may be as aninhibitor of growth its intrinsic effectivenessmust be masked by those metabolites presentin the test bacterial cell. Whether of exogenousor endogenous origin, presumably they arepresent intracellularly in finite amounts, and it isreasonable to assume that to some unpredictabledegree they antagonize a concentration of aureo-mycin which otherwise would be inhibitory.Where the reversing metabolites are of endoge-nous origin, as in these experiments, it followsthat inhibition by aureomycin will occur only ata concentration sufficient to overcome the re-versing potential of the intracellular metabolites,particularly riboflavin.

TABLE 4Cross concentration testing of aureomycin and ribo-flavin with a riboflavinless strain of Bacillus

subtilisThe figures represent colonies per plate.

AURZO- RIBOFIAVIN, uG PER ML AGAR DuxKYCIN,

pG PER ML0 0.25 0.5 1.0 2.0 4.0 8.0 16.0132.0 64.0 125.0

None 0 67 79 86 76 87 69 71 81 26 00.0625 0 54 64 77 81 73 77 62 72 13 00.1250 0 40 37 69 83 76 72 70 70 0 00.25 0 38 45 59 63 68 59 69 63 0 00.50 0 11 36 41 50 59 63 67 67 0 01.0 0 0 9 27 21 28 30 56 62 0 02.0 0 0 2 3 9 13 21 49 68 0 04.0 0 000 0 1 0 15 27 0 08.0 0 00 0 00 2 21 3 0 0

Expected therefore would be an apparentpotentiation of aureomycin activity, i.e., inhi-bition by a lower level than otherwise required,if the concentration of the specific metabolitesis reduced. One way to achieve this would be toinhibit the synthesis of excess metabolite by thecell. It is possible to test this idea in the case ofriboflavin synthesis. Woolley (1950) has shownthat 1,2-dichloro4,5-diaminobenzene inhibitsriboflavin synthesis in B. megaterium and thatthe inhibition was reversed by 1,2-dimethyl-4, 5-diaminobenzene. He interpreted these resultsas meaning that 1,2-dimethyl44,5-diaminoben-zene or a compound closely related to it is a nor-mal precursor in the biosynthesis of riboflavin,and that the analogue 1 ,2-dichloro4,5-diamino-benzene is a competitive inhibitor of the utiliza-tion of 1 ,2-dimethyl4,5-diaminobenzene. It re-

1953] 483

on March 6, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

J. W. FOSTER AND ROBERT F. PITTILLO

mained to test whether the analogue can be usedto effectively lower the riboflavin concentration inthe cell, and by so doing "protect" the intrinsicinhibitory action of aureomycin. 1,2-Dichloro-4,5-diaminobenzene was kindly furnished byDr. Karl Pfister of Merck & Co., Inc.

B. subtilis, strain ATCC 6633, was found to beinhibited by 1 ,2-dichloro4,5-diaminobenzene,presumably in part at least, by interference withits riboflavin synthesis. Riboflavin reverses

aureomycin inhibition of this organism to a

marked extent (see above). Cross concentrationexperiments using 1 ,2-dichloro4,5-diaminoben-zene and aureomycin to inhibit this organism

showed (table 5) that by itself 0.125 jig aureo-

mycin per ml was required for complete inhibi-tion, and by itself, 12.0 ,ug 1,2-dichloro4,5-diaminobenzene. Six mg 1,2-dichloro4,5-diam-

TABLE 5

Protection of aureomycin inhibition by 1, B-dichloro-4, 5-diaminobenzene

The figures represent colonies per plate.

1, 2-DICHLoRo-4, 5-DIu[NoBENzENz, pGAUREOMYCIN, ;G PER IL AGAR MEDIUK

PER mL

None 1.5 3.0 6.0 12.0

None 78 77 71 69 00.0075 61 89 54 0 00.015 80 62 12 0 0

0.03125 75 47 0 0 0

0.0625 49 12 0 0 0

0.125 0 0 0 0 0

inobenzene per ml caused no significant inhibition.When combined with 6 jig 1,2-dichloro44,5-diaminobenzene, 0.0075 Mg aureomycin was com-

pletely inhibitory. Thus, a noninhibitory concen-

tration of 1,2-dichloro4,5-diaminobenzene hasapparently potentiated aureomycin activity 16-fold. In all probability, however, 1,2-dichloro-4,5-diaminobenzene has protected the aureomy-

cin from reversal by riboflavin by inhibitingsynthesis of the latter. The term "protected in-hibition" is proposed to describe this phenome-non. In a second experiment the apparent po-

tentiation of aureomycin was 8-fold.Inhibition of riboflavin synthesis by 1,2-dichloro-

4 ,6-diaminobenene. B. #ubtilis, strain ATCC6633, was cultivated in the presence of gradeddoses of 1,2-dichloro44,5-diaminobenzene for 22hours in M-9 broth on a shaking machine. In-

hibition of growth was definite, but not com-plete, in the flasks containing 200 and 400 ,g1,2-dichloro44,5-diaminobenzene per ml, re-spectively. Aliquots were removed for dry weightdeterminations of the cells present in each flask,after washing by centrifugation. Other portionsof the whole cultures were acidified with aceticacid and autoclaved to release bound riboflavin.After neutralization the riboflavin content ofthe clear supernatant was determined by meansof the L. casei assay (Barton-Wright, 1952).Riboflavin synthesis per mg of dry bacterial

TABLE 6Inhibition of riboflavin synthesis in bacteria by

various inhibitors

BACTERIAL GROWTH, RIBOFLAVIN SYNTEX-INIBITOR,MG MG DRY WEIGHT SIZED, pG PER MG DRYPER ML CULTUIR WEIGHT OF CELLS

DCDAB*0 14.7 0.0032

200 10.8 0.0016400 7.9 0.0016

Aureomycint0 29.4 0.00412 18.4 0.00394 4.3 0.0030

Terramycint0 29.4 0.00412 16.3 0.00284 3.5 0.0016

* 1,2-Dichloro-4,5-diaminobenzene. The organ-ism tested with this compound was Bacillus sub-tilis, strain 6633.

t The organism tested with both these com-pounds was Aerobacter aerogenes, incubated for16 hours.

cells was reduced by 50 per cent in the presenceof 1,2-dichloro44,5-diaminobensene (table 6).This confirms Woolley's (1950) discovery rela-tive to 1,2-dichloro4,5-diaminobenzene andimplements the interpretation of the results pre-sented in the preceding paragraph.

Inhibition of riboflavin synthesis by aureomyctnand by terramycin. An experiment identical withthe preceding one was done with aureomycinand terramycin as inhibitors of Aerobacter aero-genes. This organism is used currently in prefer-ence to E. coli because it grows much faster inM-9 medium. The results (table 6) show that

484 (voL. M1

on March 6, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

METABOLITE REVERSAL OF ANTIBIOTIC INHIBITION

both antibiotics inhibit riboflavin synthesis,terramycin being clearly more effective in thisrespect.

DISCUSSION

These data were obtainable because the methodallows experiments to be done at threshold con-centrations of inhibitors in which selection ofresistors and physiological factors consequentto higher cellular concentrations are rigorouslyexcluded. Basic to this objective is the use ofvery small inocula for relative uniformity in thepopulation and the all or none criterion of growthwith respect to individuals in the inoculum. It isbelieved this kind of test most accurately demon-strates the intrinsic sensitivity of a given bacterialstrain to an inhibitor.

It is probable that the reversing metabolites,especially those active at very low levels, repre-sent the most sensitive sites of inhibition in cellsgrowing under these conditions. That severalmetabolites can reverse an antibiotic may betaken as further evidence, already implicit inenzyme studies (Smith, 1953), that multiplemechanisms may be inhibited simultaneously ina given cell by any given concentration of anti-biotic. Reversibility is not achieved at relativelyhigh antibiotic levels, indicating that additionalmechanisms, inhibition of which apparently isirreversible individually or collectively, are nowreached by the antibiotic.The experiments summarized in table 3 may

be used as evidence that at any given level ofantibiotic the threshold inhibition of differentindividuals in a given population is the resultof different mechanisms. In other words, anypopulation is heterogeneous with respect tomechanisms most sensitive to a given level of anantibiotic, and at 100 per cent inhibition, differ-ent cells fail to grow for different reasons. Indifferent species the relative reversal patternsdiffer.Aureomycin evidently inhibits synthesis of

riboflavin (table 6) as well as utilization of thepreformed vitamin (table 4). Presumably theantibiotic is a structural analogue of both thevitamin and a precursor of the vitamin. Terra-mycin has not been tested as a competitive in-hibitor of riboflavin utilization, but such a func-tion would not be surprising on the basis ofstructural analogy with aureomycin. That terra-mycin is not reversible by riboflavin, we in-terpret as the result of concomitant irreversible

inhibition of other mechanisms at threshold in-hibitory levels of the antibiotic. In the case ofaureomycin, at the levels employed, the ribo-flavin metabolism apparently was inhibitedselectively since the vitamin alone can reversethe inhibition.The concept of protected inhibition is similar

to the synergistic experiments of Beerstecher andShive (1947) and the sequential blocking of Potter(1951), but is distinct in the implication that onefinds empirically by a method such as we havedescribed a reversing metabolite for any in-hibitor, then combines with the inhibitor in abacterial culture a compound inhibiting the syn-thesis of that metabolite. Also implied here isthat one automatically selects perhaps the mostsusceptible mechanisms in the cell to any givenconcentration of inhibitor. The methodology ofprotected inhibition may well provide a rationalefor new drug therapy, by potentiation of anyinhibitors. This might reduce the curative dose,thereby minimizing host toxicity and increasingtolerance. Protected inhibition also may be ofinterest in connection with development of re-sistance. These matters are under investigation.

ACKNOWLEDGMENT

We are grateful to Dr. Tom Jukes of LederleLaboratories for a generous gift of aureomycinused in the latter part of this work.

SUMMARY

Numerous pure metabolites were found toreverse the inhibition of growth of Escherichiacoli, strain B, caused by certain levels of peni-cillin G, aureomycin, dihydrostreptomycin, terra-mycin, or chloramphenicol. The tests wereconducted by the spread plate method in glucose-salts agar using about 100 washed cells as inocu-lum per plate. The best reversal was obtained inthe case of aureomycin. Riboflavin, lumichrome,inosine, and glycine were superior reversingmetabolites of aureomycin in 12 different bac-teria. Concentration experiments with riboflavinindicate the aureomycin-riboflavin relationshipis competitive. Aureomycin competitively in-hibited utilization of riboflavin by a riboflavinlessmutant of Bacillus subtilis. Dichlorodiaminoben-zene, a known inhibitor of riboflavin synthesis,aureomycin and terramycin all were found toinhibit riboflavin synthesis in the test bacteriaused here. In the presence of a subinhibitory

1953] 485

on March 6, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

J. W. FOSTER AND ROBERT F. PITTILLO

level of dichlorodiaminobenzene only one-six-teenth the amount of aureomycin was requiredto inhibit B. subtilis, strain ATCC 6633, thanwas required in the absence of the analogue. Theconcept of protected inhibition is developed andproposed as the mechanism of the observed po-tentiation. The generalized significance of pro-tected inhibition as a rationale arid tool for newdrug therapy is discussed.

REFERENCESANDERSON, E. H. 1946 Growth requirements of

virus-resistant mutants of E. coli strain B.Proc. Natl. Acad. Sci. U. B., 32, 120-128.

BARTON-WRIGHT, E. C. 1952 The microbiologicalassay of the vitamin B-complex and amino acids.Pitman Publishing Corp., New York, p. 35-45.

BEERSTECHER, E., JR., AND SHIVE, W. 1947 Pre-vention of phenylalanine synthesis bytyrosine. J. Biol. Chem., 167, 527-534.

BROWN, R. G., AND EMERSON, G. A. 1953 Invitro antagonism of pantothenic acid, ,-alanine and l-carnosine to inhibitory effects oferythromycin on growth of Corynebacteriumdiphtheriae. Federation Proc., 12, 304-305.

FOSTER, J. W. 1944 Microbiological aspects ofriboflavin. II. Bacterial oxidation of ribo-flavin to lumichrome. J. Bact., 47, 27-41.

FOSTER, J. W., AND HEILIGMAN, F. 1949 Bio-chemical factors influencing sporulation in a

strain of Bacillus cereus. J. Bact., 57,639-646.

FOSTER, J. W., AND PIrTLLO, R. F. 1953 Re-versal by complex natural materials of growthinhibition caused by antibiotics. J. Bact.,65, 361-367.

GRUNDY, W. E., WHITMAN, A. L., RDZOK, E. G.,RDZOEK, E. J., HANES, M. E., AND SYLVESTER,J. C. 1952 Actathiazic acid. I. Microbio-logical studies. Antibiotics and Chemo-therapy, 2, 399-408.

LICHSTEIN, H. C., AND GILFILLAN, R. F. 1951Inhibition of pantothenate synthesis by strep-tomycin. Proc. Soc. Exptl. Biol. Med., 77,459-461.

MENTZER, C., MEUNIER, P., AND MOLHO-LA CROIX, L. 1950 Faits de synergie etd'antagonisme entre la chloromycetine etdivers amino-acids vis-i-vis de cultures d'E. coli. Compt. rend., 230, 241-243.

POTTER, V. R. 1951 Sequential blocking ofmetabolic pathways in vivo. Proc. Soc.Exptl. Biol. Med., 76, 41-46.

SMITH, G. N. 1953 The possible modes of actionof chloromycetin. Bact. Revs., 17, 19-29.

WILLIAMS, R. J., EAKIN, R., BEERSTECHER,E., JR., AND SHIVE, W. 1950 BiochemistryofB vitamins. Reinhold Publishing Corpora-tion, New York.

WOOLLEY, D. W. 1950 A study of non-competi-tive antagonism with chloromycetin and re-lated analogues of phenylalanine. J. Biol.Chem., 185, 293-305.

486 (voL. 66

on March 6, 2021 by guest

http://jb.asm.org/

Dow

nloaded from