JOURNAL OF BACTERIOLOGY, Jan., 1967, p. 71-79 Vol. 93, No. 1Copyright @ 1967 American Society for Microbiology Printed in U.S.A.

Mode of Action of Novobiocin inEscherichia coli

DAVID H. SMITH1 AND BERNARD D. DAVISDepartment of Bacteriology and Immunology, Harvard Medical School, Boston, Massachusetts

Received for publication 20 August 1966

ABSTRACTThe mechanism of action of novobiocin was studied in various strains of Esch-

erichia coli. In all strains tested except mutants of strain ML, the drug imme-diately and reversibly inhibited cell division, and later slowed cell growth. The pre-viously described impairment of membrane integrity, degradation of ribonucleicacid (RNA), and associated bactericidal effect were found to be peculiar to MLstrains. The earliest and greatest effect in all strains was an inhibition of deoxy-ribonucleic acid (DNA) synthesis; RNA synthesis was inhibited to a lesser extent,and cell wall and protein synthesis were affected later. The inhibition of nucleicacid synthesis was accompanied by an approximately threefold accumulation of alleight nucleoside triphosphates. Since novobiocin does not inhibit nucleoside tri-phosphate synthesis, degrade DNA, or immediately affect energy metabolism, itmust inhibit the synthesis of DNA and RNA by direct action on template-poly-merase complexes.

Novobiocin (NB), a clinically useful antibiotic,has been reported to affect several systems inbacteria, including the maintenance of membraneintegrity, nucleic acid synthesis, and cell wallsynthesis. Thus, it decreases the crypticity ofEscherichia coli ML-35 (5), a mutant that lacksthe 3-galactoside transport system (18), and itcauses loss of ribonucleic acid (RNA) in thisstrain (5). These observations have led to thesuggestion that NB acts primarily by damagingthe growing cell membrane. The resulting loss ofintracellular metabolites then leads to severalother effects. These include an increase in theintracellular uridylate pool and inhibition ofRNA synthesis in Streptococcus faecium (4),inhibition of cell division (20) and deoxyribo-nucleic acid (DNA) formation (2), production offilamentous forms (2), and induction of lambdaprophage (24) in E. coli. The last two effects areknown to be inducible by a number of agentsthat interfere with DNA synthesis, such as mito-mycin C, ultraviolet irradiation, and thyminedeprival. NB also causes the accumulation ofnucleotide precursors of cell wall, in Staphylo-coccus aureus (22, 23), and it inhibits the in vitrosynthesis of teichoic acid by extracts of Bacilluslicheniformis and Lactobacillus planatarum (6,10); however, inhibition of cell wall synthesis

1 Special Fellow, National Institute of Arthritis andMetabolic Diseases.

clearly cannot be the sole mechanism of actionof NB since it does not induce spheroplast forma-tion (23), and it inhibits the growth of proto-plasts (12, 19).With the aim of defining the primary action of

NB, we have studied the kinetics of several ofits effects in E. coli. The results have revealedthat the drug immediately and reversibly in-hibits DNA polymerization, and, less extensively,RNA polymerization, in a number of strains ofE. coli. The membrane damage, and an associatedlethal action, are a special response in the MLstrain.

Part of this work has been reported previously(21).

MATERIALS AND METHODS

Chemicals. The monosodium salt of novobiocin,lot SM 387, a gift of The Upjohn Co., Kalamazoo,Mich., was dissolved fresh daily in water at a concen-tration of 5 mg/ml. Uracil-2-C'4 (30 mc/mmole),methyl-C'4-thymine (3 mc/mmole), and H3-uracil(4 c/mmole) were purchased from New EnglandNuclear Corp., Boston, Mass.; L-leucine-C14 (20 mc/mmole) and L-phenylalanine-C14 (350 mc/mmole),from Schwarz Bio Research Inc., Orangeburg, N.Y.;and P3204-3, from Isoserve, Inc. H3-diaminopimelicacid (DAP; 201 mc/mmole) was purchased fromNuclear-Chicago Corp., Des Plaines, Ill., and waspurified by paper chromatography prior to use (14).MN-cellulose powder 300 was purchased from Brink-mann Instruments, Westbury, N.Y., and 50% poly-

71

on Novem

ber 24, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

SMITH AND DAVIS

merized ethyleneimine (PEI), from Chemirad Corp.Deoxyribonuclease, electrophoretically purified, waspurchased from Worthington Biochemical Corp.,Freehold, N.J., and polyuridylic acid (poly U) fromMiles Chemical Co.

Bacteria and media. Strains ML-3, ML-30,and ML-35 were a gift from J. Monod, strain 15TA-(thy-arg), from R. Rudner, and strain K12-300(Hfr, thi-), from L. Gorini. The low potassium modi-fication (AKo.3) of medium A (8) was used exceptwhen p3204-3 was employed. For those experiments,a low phosphate medium of the following compositionper liter was used: MgSO4-7H20, 0.1 g; Na2HPO4,0.04 g; (NH4)2SO4, 1 g; KCI, 2 g; maleic acid, 5. 8 g;tris(hydroxymethyl)aminomethane (Tris), 6.05 g; andNaOH, 1.92 g.

Cultural conditions. Inocula were grown overnightat 37 C with limiting glucose (0.1%), and were thendiluted 25- to 50-fold in the same medium with 0.2%glucose. Overnight cultures of 15TA- were supple-mented with 100 ,gg/ml of arginine and 5 jg/ml ofthymine, and were diluted in the morning in mediumsupplemented with 50 jug/ml of arginine, 1 ,g/ml ofthymine, and 0.2% glucose. Cultures of strain K12-3000 were supplemented with 0.1 ug/ml of thiamine.

Cultures were grown at 37 C, either in Erlenmeyerflasks on a rotary shaker or in tubes aerated by filteredair. Growth was exponential for at least two genera-tions before the drug was added. When the constitu-ents of the medium were to be changed, bacteria wereharvested on membrane filters (type HA, 0.45-,g poresize, 47-mm diameter; Millipore Filter Corp., Bed-ford, Mass.) and were washed with 2 volumes of pre-warmed salts solution (AKo.3). Care was taken toprevent the filters from sucking dry during the filtra-tion and washing. The filter was placed in the originalvolume of prewarmed medium, and the cells were re-suspended by pipetting back and forth.

Cell growth was measured by determining theabsorbancy of cultures at 490 mrp in a Lumetron color-imeter; an optical density (OD) of 0.1 was equivalentto 0.09 mg (dry weight) and 108 viable cells (ex-ponentially growing) per milliliter. Viability countswere determined on samples diluted in 0.9% NaCland were plated in duplicate in melted tryptic digest-agar.

Membrane integrity. Damage to the cell membranewas measured in terms of the leakage of intracellularmaterial absorbing at 260 m,>. Samples of 3 ml weretreated with sodium azide (final concentration, 10-2M) and centrifuged at room temperature at 10,000 Xg for 10 min, and the supernatant fluids were shakentwice for 10 min on a wrist-action shaker with anequal volume of n-amyl acetate. This extraction re-moved more than 99% of the NB, which absorbs at260 m, (adsorption maximum, 305 m,u). Samplesfrom an untreated culture were extracted in parallelwith samples from the drug-treated culture, and theOD at 260 m,u was determined.

Synthesis of cellular macromolecules. To detectchanges in the rate of cellular macromolecule synthe-sis, radioactive precursors were added at the sametime as NB. Protein synthesis was followed by measur-ing the incorporation of C14-leucine; 1-ml samples of

the culture were added to an equal volume of 10%trichloroacetic acid, heated at 95 C for 20 min, col-lected on membrane filters, and washed three timeswith 3 ml of 5% trichloroacetic acid. RNA and DNAsynthesis were followed by measuring the incorpora-tion of C14- or H3-uracil and C14-thymine intocold trichloroacetic acid-precipitable material. Cellwall synthesis was followed by measuring the in-corporation of H3-DAP into trichloroacetic acid-pre-cipitable material in cells growing in excess lysine,after growth overnight in lysine. The lysine repressesand inhibits DAP decarboxylase, thus minimizingconversion of labeled DAP to lysine and insuringmaximally selective incorporation into cell wall glyco-peptide.

Incorporation data for all experiments, except thatillustrated in Fig. 6 and 7, were plotted against a timescale which yields linear curves for a culture in bal-anced exponential growth; this is a scale proportionalto the increase in turbidity of a culture, measuredfrom the time of addition of NB and radioactive pre-cursors (8).

Synthesis ofprotein in vitro. The preparation of ex-tracts of E. coli B, and poly U-directed polypeptidesynthesis, were performed as described previously (7).Polypeptide synthesis stimulated by natural messengerRNA (mRNA) was performed with extracts preparedby slower centrifugation, without deoxyribonucleasetreatment and supplemented by nucleoside triphos-phates as indicated.

Nucleotide pool. To compare the uptake and theincorporation of C14-thymine, 1-ml samples were takenat intervals and filtered on membranes (to determinetotal uptake) or they were added to cold 10% tri-chloroacetic acid (to determine incorporation). Thefiltered specimens were washed three times with 4 mlof a salts solution (AKo.3) containing 50 ,ug/ml ofthymine; those precipitated with trichloroacetic acidwere prepared as above.The cellular pools of nucleoside triphosphates were

determined by thin-layer chromatography of extractsof cells labeled with P3204-3 in low phosphate medium(specific activity, 13.2 mc/mmole) for at least onedoubling time to insure intracellular equilibration.Samples (25 ml) were added to trichloroacetic acid(final concentration, 5%) in tubes chilled in ice; after30 min, the precipitates were filtered on membranefilters (47-mm diameter), and were washed threetimes with 5 ml of ice-cold water. The combinedfiltrate and washes were mixed with 100 mg of acti-vated charcoal, which was collected on a membranefilter and extracted three times with 5 ml of 50%ethyl alcohol (pH raised to 8.2 with ammonia). Theextracts were lyophilized and were taken up in 5 ml ofwater, and 5-pliter samples were applied to cellulose-PEI thin layers and were developed according toNeuhard et al. (16). This technique separates the in-dividual nucleoside triphosphates, after eliminatingmost of the other acid-soluble phosphorus-containingcompounds. The nucleoside triphosphates were lo-cated by radioautography, and the spots were cut out,placed in scintillation vials, and counted.

Radioactivity determinations. Precipitates contain-ing C14 were collected on membrane filters, glued to

72 J. BAcTERioL.

on Novem

ber 24, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

MODE OF ACTION OF NOVOBIOCIN

aluminum planchets, dried under a lamp, and countedin a Nuclear-Chicago thin-window gas-flow counter.When the precipitates contained H3, the filters wereplaced in scintillation vials, dried under a lamp,covered with a solution of toluene containing 4 mgper liter of 2,5-diphenyloxazole and 50 mg per literof 1,4 - bis - 2 - (4 - methyl - 5 - phenyloxazolyl) -

benzene (Pilot Chemical Co.), and counted in a Nu-clear-Chicago scintillation spectrometer. Chromatog-raphy specimens containing p3204-3 were counted inthe same solution.

RESULTS

Effects on growth, viability, and permeability.In growing cultures of E. coli ML-35 (and otherderivatives of strain ML), the addition of NBat 300 ,ug/ml produced an immediate, progres-sive decrease in viable count and a later de-crease in optical density (Fig. 1). In addition,as previously observed (2) there was loss ofintracellular 260 m,-absorbing material (Table

0.3.

0

C~ ~ ~ ~ ~~o

° 0.2 C-,ssl

Q 02W~300/gfmI

0-

0

40

Coni,ol20

0

10

a)5

> \300040/I2.5

'5 30 45 50

Minutes

FIl. 1. Effect ofnovobiocin on turbidity and viabilityofEscherichia coli ML-35. Exponentially growing cellswere inoculated at zero-time into growth medium with or

without novobiocin (300 ,ug/ml). The turbidity andviable-cell counts were assayed at intervals as describedin Materials and Methods.

1), degradation of RNA (Table 2), and decreasein crypticity of ML-35 for g-galactosides (un-published data).

With several other strains of E. coli, in con-trast, the inhibition of growth was entirelyreversible. With strain 15TA-, for example, NBat 100 ,ug/ml immediately stopped cell division,the number of viable cells remaining constantfor at least 2 hr (Fig. 2). Cell growth (measuredturbidimetrically), however, was slowed onlyafter approximately 20 min (Fig. 2). Even at300 ,ug/ml (Fig. 2) and 500 ,ug/ml, the drug wasonly slightly more inhibitory, and it did not kill

TABLE 1. Effect of novobiocin on the excretion of260 m,u-absorbing material from

Escherichia colia

Strain Increase in ODStrain ~~~~at260 mp

l5TA-........................... .018ML-35...................... .......... .438B................................... .009W ......................... .036K12-3000 ............................. .015

a Exponentially growing cultures of each strainof E. coli were inoculated into growth mediumwith or without novobiovin (300 ,ug/ml). After 60min of incubation, samples were removed andtreated as described under Materials and Methods.The OD of samples of the treated cultures (cor-rected for the contribution by novobiovin) wasapproximately 4.0; of untreated cultures, 5.5. Theresults are expressed as the difference in OD at 260mru between the supernatant fluids of the drug-treated and the untreated cultures of each strain.

TABLE 2. Effect ofnovobiocin on the degradation ofRNAa

Counts per min per ml

Min after ML-35 15TA-resuspension ML51A

+NB -NB +NB -NB

0 4,675 4,810 4,870 4,65015 4,703 - 5,020 4,76830 4,706 5,094 4,740 4,77045 4,579 - 5,060 4,77960 3,332 4,806 5,070 4,788

a Cells of Escherichia coli ML-35 and 15TA-,grown for two doubling times in the presence ofC14-uracil (10 yg/ml), were harvested in exponen-tial growth, filtered, washed, and resuspended ingrowth medium with or without novobiocin (300,ug/ml) and with unlabeled uracil (20 ,ug/ml).Samples of 1 ml were removed at intervals, andacid-insoluble radioactivity was determined asdescribed in Materials and Methods.

73VOL. 93, 1967

on Novem

ber 24, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

SMITH AND DAVIS

0.3L

0-

E0

0

a

0.0

0.21

0.1

40

0

x

0)

a)

20~

0

0~,,

I~~~~~

I/

0-I

0Ozi

0 Control

5-O0 NB 100 pg/ml* NB 300Eg/ml

2.515 30 45 60

Minutes

FIG. 2. Effect ofnovobiocin on turbidity and viabilityofEscherichia coli JSTA-. Exponentially growing cellswere inoculated at zero-time into growth medium with orwithout various concentrations of novobiocinl. The tur-bidity and viable-cell counts were assayed at intervals asdescribed in Materials and Methods.

the cells. Moreover, the drug did not degradeRNA prelabeled with C'4-uracil before treatment(Table 2), and it did not promote the loss ofintracellular 260 m,u-absorbing material (Table1). With E. coli strains B, W, and K12-3000,the drug affected growth, viability, RNA stabil-ity, and permeability in a manner similar to thatdescribed for strain 15TA-.These results indicated that the effect of novo-

biocin on permeability is not its usual actionbut is associated with a lethal action unique toML strains, among those strains tested.

Effects on macromolecule synthesis. Since theinhibition of growth of most strains of E. coliis clearly not due to an effect on permeability,we investigated the kinetics of the inhibition ofthe synthesis of various cellular macromoleculesin strain 15TA-.

Nucleic acids. At 100 ,ug/ml, NB immediatelyinhibited the synthesis of both DNA and RNA,as measured by the incorporation of C14-thymine

and C'4-uracil, respectively. DNA synthesis wasthe more extensively affected: during 60 min,it was inhibited by 80% and RNA by 50% (Fig.3). At 300 and 500 ,g/ml, NB inhibited DNAsynthesis duting 60 min by 90 and 95%, respec-tively, but it did not quite completely stop DNAsynthesis even up to 5 hr. At these concentrations,the drug inhibited RNA synthesis by 65 and 75%,respectively, during 60 min.To differentiate the effects on DNA and RNA

synthesis more clearly, the effects of lower con-centrations were studied. At 5 ,ug/ml, NB imme-diately inhibited DNA synthesis by approxi-mately 10% but did not affect RNA synthesisdetectably during a 30-min incubation (Fig. 3).

Protein. At 100 ,ug/ml, the drug did not affectthe incorporation of C'4-leucine into protein for15 to 20 min, and it inhibited synthesis during 60min by approximately 30% (Fig. 4). This delaysuggests that NB affects protein synthesis in-directly, by its action on RNA synthesis. More-over, NB did not inhibit protein synthesis indeoxyribonuclease-treated extracts in vitro withpoly U or endogenous RNA as messenger(Tables 3 and 4), but it did inhibit protein syn-thesis in an extract of E. coli which was nottreated with deoxyribonuclease and was supple-mented with nucleoside triphosphates (Table 4).The continued cellular synthesis of protein,

and the unimpaired increase in optical density(Fig. 2), indicate that NB does not have anyearly effect on energy supply.

Cell wall. The synthesis of cell wall, measuredby the incorporation of H3-DAP in the presenceof unlabeled lysine, was not affected until approxi-mately 15 min after addition of the drug (at 100,ug/ml), and the amount of radioactivity incor-porated in 60 min in the treated culture was 70%of that in the untreated culture. At 5 ,ug/ml, noeffect was seen (Fig. 5).

Thus, the earliest and most pronounced effectof NB was an inhibition of DNA synthesis; RNAsynthesis was also affected early, but less ex-tensively, whereas protein synthesis and cellwall synthesis were affected only later. Similarresults have been obtained with several otherstrains of E. coli, including mutants of the MLstrain. There seems little doubt that the primaryaction of NB in E. coli is an inhibition of DNAsynthesis.Mechanism of inhibition of DNA synthesis.

In the above experiments, DNA synthesis wasfollowed by measuring the incorporation ofradioactive thymine into acid-insoluble materialin a thymine auxotroph; hence, the observedeffects could result from several possible actions:(i) interference with transport of the added pe-

74 J. BACrERIOL.

on Novem

ber 24, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

MODE OF ACTION OF NOVOBIOCIN

Control 0

'4C-Thyminie, oUracil

./ _- NB 100g/ml

_~~ '4C-Urtcil

le,,' , NB lOOug/mi

10 20 30 45 60Minutes Minutes

FIG. 3. Effect of novobiocin (100 ,ug/ml or 5 j.g/ml) on nucleic acid synthesis in Escherichia coli JSTA-. Ex-ponentially growing cells were inoculated at time zero into growth medium containing (a) C'4-thymine (I jig/ml)or C'4-uracil (10 j.g/ml), and (b) C04-thymine (I jug/ml) and H3-uracil (10J ,g/ml), with or without novobiocin(100 ,ug/ml in a, 5 ,ug/ml in b). Incorporated radioactivity was determined as described in Materials and Methods.The very similar results obtained for C'4-uracil and C"4-thymine incorporation in the untreated culture in (a) aredue to a fortuitous choice ofspecific activities of respective radioisotopes.

15 30 45

Minutes

FIG. 4. Effect of novobiocin on protein synthesis inEscherichia coli 1STA-. Exponentially growing cellswere inoculated at time zero into growth medium con-taining either Cl4-thymine (I ,ug/ml) or C'4-leucine (40jig/mi), with or without novobiocin (100 jig/ml). Ra-dioactivity incorporated into DNA or protein was deter-mined as described in Materials and Methods. The re-sults of C'4-thymine incorporation were identical tothose in Fig. 3, and therefore are not illustrated.

cursor, (ii) interference with its conversion tothymidine triphosphate (TTP), (iii) inhibitionof the synthesis of one of the other deoxynucleo-side triphosphates, (iv) degradation of templateDNA, or (v) direct inhibition of the polymerizingsystem.

Degradation of the DNA template. This possi-bility is excluded by the reversibility of the action

TABLE 3. Effect of novobiocin on in vitro proteinsynthesis stimulated by poly Ua

Novobiocin Counts/min

'gg/ml0 4,74525 4,69550 4,765150 4,795300 4,640450 4,720600 4,520

a Incubation mixtures contained deoxyribo-nuclease-treated, preincubated, dialyzed extractsof Escherichia coli B. Incubation with poly U andC14-phenylalanine, with various concentrationsof novobiocin, was performed as described in

60 Materials and Methods.

of the drug: not only was NB purely bacterio-static for all strains tested except the ML strains,but when it was removed from a culture of 1 5TA-the synthesis of DNA immediately resumed at a

rate comparable to that of an untreated culture(Fig. 6). Moreover, when the DNA of the cells

was prelabeled with Cu4-thymine, subsequentexposure to NB in unlabeled growth mediumcaused no loss of acid-insoluble radioactivity(Table 5).

Transport. To test the possibility that NB was

inhibiting the transport of thymine into the cell,the total uptake of C'4-thymine, as well as itsconversion into acid-insoluble material, were

measured at close intervals in growing cells ofstrain 15TA-. At 100 ,tg/ml, the drug almost im-mediately inhibited the incorporation of the

. 3

2

'E

a._E

c7

E1

b

Is

0

2

a.-)

J

VoL.93,1967 75

on Novem

ber 24, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

SMITH AND DAVIS

TABLE 4. Effect of novobiocin on in vitro proteinsynthesis stimulated by the presence ofDNAa

Novobiocin Counts/min

,ug/ml

0 58862 471125 444250 400500 310

1,000 168

a Extracts of Escherichia coli B were preparedas described in Materials and Methods but werenot treated with deoxyribonuclease and werecentrifuged twice for 5 min at 7,000 X g. Two setsof incubation mixtures containing extract, C14-leucine, supplemented with 100 mjAmoles of uridinetriphosphate and cytidine triphosphate, andvarious concentrations of novobiocin were pre-pared as in Table 3; deoxyribonuclease (10 pAg/ml)was added to 1 set of tubes. The mixtures weretreated as described in Materials and Methods.Novobiocin had no effect on the acid-insolubleradioactivity in the tubes treated with deoxy-ribonuclease. The difference in acid-insolubleradioactivity between tubes treated with deoxy-ribonuclease or not treated is expressed as countsper minute.

400

30C

0.

0IL

cx

20C

10C

FIG. 5. Effetion into cell15TA-. Cells i,lysine were inotcontaining H3-1with or withouIncorporated rc

in Materials at

thymine, butsignificant dreflects entrythe later rate

,|ID21 Control/ 0

20 40 60 80Mi;nUteS

FIG. 6. Effect of removal of novobiocin on DNAsynthesis. Exponentially growinig ISTA- cells were in-oculated at time zero into growth medium containingC14-thymine (I lAglml) with or without novobiocin (100.uglml). After 30 min of incubation, the drug-treatedcells were filtered, washed, and resuspended in the samevolume ofdrug-free growth medium containing C14-thy-mine. Incorporated radioactivity was determined asdescribed in Materials and Methods.

TABLE 5. Effiect of novobiocin on the stability ofDNA in Escherichia coli 15TA-

Counts per min per mlM in after

resuspensionControl Drug

0 2,031 1,97110 2,019 1,99820 2,032 1,95130 2,086 1,99845 2,101 2,03160 2,040 2,040

15 30 45 60 a Exponentially growing 15TA- cells were grownMinutes for three doubling times in medium containing

C14-thymine (2 ;sg/ml). The cells were then col-Vct of novobiocin on H3-DAP incorpora- lected on a membrane filter, washed with 3 vol-wall glycopeptide in Escherichia coli umes of buffered salts solution, and resuspended inn exponential growth in the presence of growth medium with or without novobiocin (100culated at zero-time into growth medium ;eg/ml). Radioactivity in acid-insoluble materialDAP (10 ,ug/ml), and lysine (100 ,ug/ml), was determined as described in Materials andr various concentrations of novobiocin. Methods.2dioactivity was determined as describedid Methods.

incorporation into DNA; hence, the drug inter-uptake was slowed only after a feres with a later step rather than with the trans-

elay (Fig. 7). The initial uptake port of thymine. Moreover, NB increased ap-into the metabolite pool, whereas proximately two- to threefold the size of theof uptake is limited by the rate of pool of thymine and its nucleotides, as revealed

76 J. BACTERIOL.

/

a 0 Control) NB 5fLg/ml

aZ * NB l00,g/mI

v

on Novem

ber 24, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

MODE OF ACTION OF NOVOBIOCIN

by the difference between total and incorporatedradioactivity after 120 sec.

Nucleotide synthesis. Direct measurements ofthe intracellular nucleotide pool in strain 15TA-

40 120 200 280

Seconds

FIG. 7. Effect of novobiocin on thymine uptake andincorporation. Exponentially growing 15TA- cells were

filtered, washed to remove nonradioactive thymine asdescribed in Materials and Methods, and resuspendedin growth medium containing C'4-thymine (I lg/ml),with or without novobiocin (100 ug/ml). Samples were

removed at intervals for determination of total cellularradioactivity and acid-insoluble radioactivity, as de-scribed in Materials and Methods.

TABLE 6. Effect ofnovobiocin on the distribution ofP3204-3 in nucleoside triphosphates in

Escherichia coli 15TA-

Nucleotideb

dATP ....

dTTP......dGTP.....dCTP.....Total......

ATP.......UTP ......

GTP ......

CTP.......Total......

Counts/specimen

Control

180340186273979

2,4741 ,320909822

5,525

NB

8121,016

502744

3,074

10,0084,4724,0882,505

21,073

Percentage of totalcounts

Control I NB

18.534.719.027.8100

45.023.816.414.8

100

26.233.316.324.2100

47.521.219.411.9

100

aCells of E. coli 15TA- were grown in lowphosphate growth medium with P320O47 and werethen treated or not treated with novobiocin (100,g/ml) for 5 min. Acid-soluble extracts were pre-pared and chromatographed, and the radio-activity in each nucleoside triphosphate wasassayed, as described in Materials and Methods.

b Deoxyadenosine, deoxythymidine, deoxy-guanosine, deoxycytidine, adenosine, uridine,guanosine, and cytidine triphosphates.

showed that NB did not inhibit the synthesis ofTTP from thymine, or the synthesis of any ofthe other deoxynucleoside triphosphates fromglucose. Indeed, the inhibition of DNA syn-thesis was accompanied, within 5 min, by a three-to fourfold increase in the content of all theprecursors of DNA (Table 5), as had been shownin Fig. 7 for the total pool of thymine nucleo-tides. A similar increase was observed in theribonucleoside triphosphate precursors of RNA(Table 5). It is noteworthy that in both expandedprecursor pools the proportions of the variouscompounds remained essentially unchanged.

Since NB thus does not inhibit DNA synthe-sis by depleting any substrate of polymerization,by degrading template DNA, or by interferingwith the energy supply (see discussion of proteinsynthesis above), it must directly inhibit thepolymerizing enzyme-template complex.

DISCUSSION

The experiments reported here define the se-quence and the extent of several effects of NB inE. coli. The drug initially inhibits DNA synthesisand, to a lesser degree, RNA synthesis; cellwall and protein synthesis are inhibited later.The previously described impairment of theintegrity of the cell membrane, degradation ofRNA, and associated bactericidal action wereconfirmed, but were found to be an additionalresponse peculiar to the ML strains; these effectswill be discussed in a subsequent paper.The present results also show that NB must

inhibit DNA synthesis by inhibiting the enzyme-template system responsible for DNA polymeri-zation: the synthesis of all four deoxynucleosidetriphosphates is not inhibited, and templateDNA is not degraded. The bacteriostatic actionof the drug (except for ML strains), and the im-mediate resumption of DNA synthesis afterremoval of the drug (Fig. 6), exclude suchmechanisms as the formation of interstrand orintrastrand cross links in DNA, as observed withmitomycin and other alkylating agents. More-over, since NB produces an immediate and ex-tensive inhibition ofDNA synthesis (95% at highdrug concentrations), it must affect the poly-merization process and not merely its initiation.Studies on the effect of NB on the in vitro ac-tivity of a purified E. coli DNA polymerase (21),to be reported in detail in a later communication,further support this conclusion.

It was originally suggested that NB induceslambda prophage by damaging the cell mem-brane, thereby promoting the loss of the repressorfor lambda (G. R. L. Worthington and H.Wolochow, Bacteriol. Proc., p. 136, 1964). Thepresent data suggest that NB, like other agents

140

cgEcl

100

* 0

- vo//~~~~~

o 00 //

/ /, O NB

/ ,0 Control

/ o , - Total cellular* / ,-O radioactivity- --- Acid-insoluble

/ z^'0 radioactivity/I-

60

20

77VOL. 93, 1967

on Novem

ber 24, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

SMITH AND DAVIS

that induce lysogenic phage, promotes inductionas a consequence of its inhibition of DNA syn-thesis. Certain nucleotides accumulate in E. colitreated with ultraviolet light (11) or deprived ofthymine (17), conditions which induce prophage,raising the possibility that some nucleotide(s)plays a role in the induction of lysogenic phage(11). Since NB promotes the accumulation ofmany nucleotides, it might be useful in furtherstudies of this relationship.The technique of thin-layer chromatography

provides a convenient and accurate method forevaluating intracellular nucleotide pools; indeed,the pools of most deoxynucleoside triphosphatescould not be defined by previous techniques.When this method was employed here to studythe effect of NB on the synthesis of nucleic acidprecursors, the results indicated that a rathersubstantial (three- to fourfold) increase in theintracellular concentration of all eight nucleosidetriphosphates accompanies the inhibition ofnucleic acid synthesis. This phenomenon is notunique to NB-treated cells, as ribonucleosidetriphosphates have been found to accumulatewhen RNA synthesis is slowed by a step-downin energy source (0. Maale, personal communica-tion). Since the nucleoside triphosphates appearto be the immediate effectors in the regulation ofpurine and pyrimidine metabolism (9, 13), thesefindings indicate a rather loose control of thesepathways.As NB does not inhibit the synthesis of any of

the four ribonucleoside triphosphates, it pre-sumably affects RNA synthesis much like DNAsynthesis, by directly inhibiting the RNA poly-merase-DNA template complex. This conclusionis supported by the effect of the drug on proteinsynthesis in vitro (Tables 3 and 4), and by theinhibition of the in vitro activity of a purifiedE. coli RNA polymerase (unpublished data).

Although NB promotes the accumulation ofglycopeptide precursor nucleotides in Staphylo-coccus aureus (22, 23), it inhibits the incorporationof P204-3 into nucleic acids to a greater extentthan that incorporated into cell wall (23), and itdoes not inhibit the in vitro synthesis of glyco-peptide by extracts of staphylococci (1, 15).Furthermore, the present results, and those ofGlaser (10), indicate that NB promotes the in-tracellular accumulation of many types of nucleo-tides, in addition to glycopeptide precursors.The effect of NB on the synthesis of cell wallcould, therefore, be presumed to be a secondaryevent (23); the delayed effect of the drug onH3-DAP incorporation (Fig. 5) confirms thisconclusion. The observed effect of the drug onthe in vitro synthesis of teichoic acid (6, 10) can-

not be a primary action of the drug in vivo, sinceNB inhibits Bacillus megaterium, an organismwhich does not contain detectable concentrationsof teichoic acid (L. Glaser, personal communica-tion); preliminary studies indicate that NBinhibits growth and division of this organism bya primary inhibition of DNA synthesis (unpub-lished data).Novobiocin has two negative charges, and it

binds magnesium and inhibits certain mag-nesium-dependent enzyme systems in vitro (4).Magnesium deprivation of ML-35 cells leadseventually (approximately 4 to 5 hr) to impairedmembrane integrity and RNA degradation (3).On these grounds, Brock proposed that the effectsof NB are the consequence of binding intracellu-lar magnesium (3, 4). Since these effects havenow been found to be unique to ML strains ofE. coli, magnesium binding seems an unlikelyexplanation for the action of NB. Furthermore,the effect of NB on DNA polymerization invitro is independent of the magnesium concen-tration (unpublished data).A striking and somewhat unexpected finding

has been the immediate and complete cessationof cell division (Fig. 2), associated with extensiveinhibition of DNA synthesis. Electron micro-graphs from many laboratories have revealedthat E. coli cells are usually multinucleate, andoften contain an incomplete septum. Since NBdoes not appear to interfere early with the forma-tion of cell wall, and since it does not impair thegrowth rate [as reflected in turbidity (Fig. 2)]for some time, one might have expected a sub-stantial number of cells to complete the processof division. The absence of this completion sug-gests two possibilities: that NB inhibits withoutdelay some unknown biosynthetic or hydrolyticstep essential for the completion and cleavage ofa septum; or that the process of continuing DNAsynthesis is intimately connected with the finalstages in cell division.

ACKNOWLEDGMENT

This investigation was supported by NationalScience Foundation grant GB-1307.

LITERATURE CITED1. ANDERSON, J. S., M. MATSUHASHI, M. A. HASKIN,

AND J. L. STROMINGER. 1965. Lipid-phospho-acetyl muramyl-pentapeptide and lipid phos-phodisaccharide-pentapeptide: presumed mem-brane transport intermediates in cell wall syn-thesis. Proc. Natl. Acad. Sci. U.S. 53:881-889.

2. BROCK, T. D. 1956. Studies on the mode of actionof novobiocin. J. Bacteriol. 72:320-323.

3. BROCK, T. D. 1962. Effects of magnesium iondeficiency on Escherichia coli and possible rela-

78 J. BACTERIOL.

on Novem

ber 24, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

MODE OF ACTION OF NOVOBIOCIN

tion to the mode of action of novobiocin. J.Bacteriol. 84:679-682.

4. BROCK, T. D., 1962. Magnesium binding as anexplanation of the mode of action of novo-biocin. Science 136 :316-317.

5. BROCK, T. D., AND M. L. BROCK, 1959. Effect ofnovobiocin on permeability of Escherichia coli.Arch. Biochem. Biophys. 85:176-185.

6. BURGER, M. M., AND L. GLASER. 1964. The syn-thesis of teichoic acid. I. Polyglycerophosphate.J. Biol. Chem. 239:3168-3177.

7. DAVIES, J., L. GORINI, AND B. D. DAVIs. 1965.Misreading of RNA codewords induted byamino glycoside antibiotics. Mol. Pharmacol.1:93-106.

8. DUBIN, D., R. HANCOCK, AND B. D. DAVIS. 1963.The sequence of some effects of streptomycinin E. coli. Biochim. Biophys. Acta 74:476-489.

9. GERHART, J. C., AND A. B. PARDEE. 1962. Theenzymology of control by feedback inhibition.J. Biol. Chem. 237:891-896.

10. GLASER, L. 1964. The synthesis of teichoic acids.II. Polyribitol phosphate. J. Biol. Chem. 239:3178-3186.

11. GOLDTHWAIT, D., AND F. JACOB. 1964. Sur lemecanisme de l'induction du developement duprophage chez les bacteries lysogenes. Compt.Rend. 259:661.

12. HANCOCK, R., AND P. C. FITZ-JAMES. 1964. Somedifferences in the action of penicillin, bacitracin,and vancomycin on Bacillus megaterium. J.Bacteriol. 87:1044-1050.

13. LARSSON, A., AND P. REICHARD. 1966. Enzymaticsynthesis of deoxyribonucleotides. X. Reduc-tion of purine ribonucleotides; allosteric be-havior and substrate specificity of the enzymesystem from E. coli B. J. Biol. Chem. 241:2540-2549.

14. LEIVE, L., AND B. D. DAVIS, 1965. The transportof diaminopimelate and cystine in E. coli. J.Biol. Chem. 240:4362-4369.

15. MATSUHASHI, M., C. P. DIETRICH, AND J. L.STROMINGER. 1965. Incorporation of glycineinto cell wall glycopeptide in Staphylococcusaureus: role of sRNA and lipid intermediates.Proc. Natl. Acad. Sci. U.S. 54:587-594.

16. NEUHARD, J., E. RANDERATH, AND K. RANDER-ATH. 1965. Ion-exchange thin-layer chromatog-raphy. XIII. Resolution of complex nucleosidetriphosphate mixtures. Anal. Biochem. 13:211-222.

17. NEUHARD, J., AND A. MUNCH-PETERSEN. 1966.Studies on the acid-soluble nucleotide pool inthymine-requiring mutants of E. coli duringthymine starvation. IL. Changes in the amountsof deoxycytidine triphosphate and deoxy-adenosine triphosphate in E. coli 15T-A-U-.Biochim. Biophys. Acta 114:61-71.

18. RICKENBERG, H. V., G. N. COHEN, G. BUTTIN,AND J. MONOD. 1956. La galactoside-permeased'Escherichia coli. Ann. Inst. Pasteur 91:829-857.

19. SHOCKMAN, G. D., AND J. 0. LAMPEN. 1962. In-hibition by antibiotics of the growth of bacterialand yeast protoplasts. J. Bacteriol. 84:508-512.

20. SMITH, C. G., A. DIETZ, W. T. SOKOLSKI, AND G.M. SAVAGE. 1956. Streptonivicin, a new anti-biotic. I. Discovery and biological studies.Antibiot. Chemotherapy 6:135-142.

21. SMITH, D. H., AND B. D. DAVIS. 1965. Inhibitionof nucleic synthesis by novobiocin. Biochim.Biophys. Res. Commun. 18:796-800.

22. STROMINGER, J., AND R. H. THRENN. 1959. Theo-retical configuration of the alanine residues ina uridine nucleotide and in the cell wall ofStaphylococcus aureus. Biochim. Biophys. Acta33:280-281.

23. WISHNOW, R. M., J. L. STROMINGER, C. H. BINGE,AND R. H. THRENN. 1965. Biochemical effectsof novobiocin on Staphyloccus aureus J. Bac-teriol. 89:1117-1123.

VOL.93, 1967 79

on Novem

ber 24, 2020 by guesthttp://jb.asm

.org/D

ownloaded from