Role of the Membrane Potential in Bacterial Resistance to

Vol.20, No. 6 ANTIMICROBLAL AGENTS ANDCHEMOTHERAPY,Dec.1981,p.803-808

0066-4804/81/120808-06$02.00/0

Role of the Membrane Potential in Bacterial Resistance to Aminoglycoside Antibiotics

P. DIANNE DAMPERtANDWOLFGANG EPSTEIN*

Department ofBiochemistry, The University of Chicago, Chicago,Illinois60637 Received30July1981/Accepted15September1981

The electrical potential difference (A0) across the membrane ofEscherichia coliwasmeasuredby thedistributionoflipid-soluble cationsandcorrelatedwith resistancetodihydrostreptomycin,whereresistance is presumed due to reduced uptake ofthe drug. A good correlation between the two measured parameters wasfound under all conditions tested, which includedeffects of several mutations, inhibitors, changesin pH,and osmolarity. The most dramatic changes were seen whenpHwasvaried; in wild-type strainsresistance increased more than 100-fold, andAipfell by70mV when pH was reduced from 8.5 to 5.5. These results were interpreted as support for a model in which the uptake of the polycationic aminoglycosides is electrogenic and thereforedriven by

AO.

Thefactor common tomutations and conditionswhichincreaseresistance was areductioninA4+.A simple modelwasdeveloped whichrelatestheminiimal inhibitory concentration totherateof aminoglycoside uptake and therateof growth.

Resistancetoaminoglycoside antibiotics isas- sociated with a variety of metabolic lesionsin bacteria.Mutations producing resistanceinclude ones affecting the synthesis of heme (27, 29), menaquinone (26, 28), cytochromes (6, 28), and the proton-translocating adenosinetriphospha- tase(ATPase) (19). Because these mutations do notalter the ribosomal site ofaminoglycoside actionnorappear toresult in inactivation of the antibiotics, it has been assumed that resistance in suchmutantsisduetoreduced uptake ofthe drugs into the cells. Reduceduptake has been demonstrated insome cases(6,8,10,28),making this mechanismareasonableexplanation for this typeof resistance. Thismechanism is of wider interest, for it has beenreported insomeclini- cally resistant strains (4). The coupling between metabolic defects and resistancetoaminoglyco- sideswasthesubject of this study.

Mutations whichproduceresistance reduceor abolish the conversion ofenergy fromoxidation- reduction reactions in the membranetothe syn- thesis of ATPorotheruses.Theconnectionof resistance with oxidative metabolism is sup- ported bythe observation thatanaerobicgrowth isassociatedwith resistance toaminoglycosides (5, 7). Thistype of resistanceaffects allamino- glycosides regardless of specific structure. By contrast,resistanceinwhichR-factor-codeden- zymes inactivate thedrugorinwhichtheribo- somalsite of action is altered bymutation are tPresent address: DepartmentofMicrobiology and Im- munology,UniversityofIllinoisMedicalCenter,Chicago,IL 60612.

specific for those aminoglycosideswith the req- uisite structural features (3). Therefore, some relatively nonspecificproperty of the aminogly- cosidesmustbe involved. Onecommonproperty of all aminoglycosides is a large net positive charge at physiological pH due to multiple amino groups with a high pK. Any positively charged molecule experiences a strong driving force for entry into bacteria which maintain a large electricalpotential (A4),interior negative, acrosstheirmembrane (22). Areduction inA4 reduces this driving forceandcoulddecrease the rateofaminoglycoside uptaketomake thecells resistant.

Inthisstudy, weexamineda number of con- ditions that alter A4 in Escherichia coli and foundacorrelation between

AAp

andsusceptibil- itytoaminoglycosides.Theresults indicate that the magnitude of

AO

is an important determi- nantof resistancetoaminoglycosides.Aprelim- inary account of some of this work has been presented previously (11).

MATERIALS AND METHODS Bacterial strains.The strains ofEscherichiacoli K-12 andtheirgenotypes are listedinTable 1. The hem-28 andunc-40wereobtainedinselections forlow- levelneomycin resistance after ultraviolet mutagene- sis. The hem-28 mutant has aphenotype typical of hemmutantsdescribedbyothers(27).Themutation inthis strainisclosetoand clockwisefrom cyanear min 85onthecurrentmapofE. coli(2),indicating^it

iseitherahemCorhemD mutation. StrainTK1207is unable togrowonsuccinate, has less than 2% of the 803

TABLE 1. Bacterialstrairns

Strain Genotype Originorreference FRAG-5 F-thi rha lacZgal (13)

kdpABC5

TK1207 F- thirha lacZ gal Ultraviolet mutant;

kdpABC5 unc-40 derived from FRAG-5 TK1208 F-thi rha lacZgal Spontaneous;from

kdpABC5unc-42 TK1207

TK1235 F- thi rhalacIgal Ultravioletmutantof kdpABC5 hem-28 lacIderivative of

FRAG-5

wild-type level of membrane ATPase (9), is defective inproline transport, and has amutation in theunc cluster near min 83 of the map. These properties indicate that it isaproton-leakyuncmutation. (12).

This mutant grows somewhat slowly on all media;

spontaneous mutations to more rapid growth arise frequently. Strain TK1208,which carriesoneof these mutations tomore rapidgrowth,isneither resistant toaminoglycosidesnordefectiveinprolinetransport, but remainsunabletogrowonsuccinate.Thegenetic lesion in strainTK1208, herecalledunc-42,isadouble mutation combining the original unc-40 mutations withasecond, closely linkedmutationwhichabolishes thehighprotonpermeability produced byunc-40.

Media. Minimal medium contained 10 mM (NH4)2804;0.4 mMMgSO4;0.5mMK2HPO4; 1mM Na citrate; 6 IMM FeSO4; and a buffer component, which was usually 100 mM K N-2-hydroxyethylpi- perazine-N'-2-ethanesulfonic acid (HEPES), pH 7.5.

To vary pH, 100 mM K-morpholineethanesulfonic acid (MES) wasused forpH 5.5to6.5, 100 mM K- HEPESwasused forpH7to8, and100mMtris(hy- droxymethyl)aminomethane (Tris)-chloridewasused forpH above 8. Glucose (50 mM) and 8gofCasa- mino Acidsperliter^wereused^ascarbon andenergy sources, and vitamins at 1 mg/liter were added as

needed.Solid medium contained15gofagar perliter.

Antibiotic sensitivity. The minimal inhibitory concentration(MIC) ofdihydrostreptomycin^wasde- terminedby growthonsolid medium.Mid-log phase

cells werediluted in minimal mediumtoaconcentra- tion of 104/ml, and 0.02 ml of this suspension was

spottedon aseries ofplates,eachdiffering from the nextbyat mostatwofolddifference indrugconcen-

tration. The MIC is the lowestconcentrationofdrug thatpreventedgrowth ofany coloniesvisible witha

low-powered (x20) microscope after 24 to 36 h of incubationat370C.

Electricalpotentialmeasurement. Themagni- tude of the membranepotentialwasdetermined from the distribution ratio of lipid-soluble phosphonium ions (25). Both [methyl-3H]triphenyl phosphonium (TPMP) (NewEngland Nuclear Corp.) and [3H]tetra- phenylphosphonium(TPP, kindly provided by H.R.

Kaback) were used with comparable results. Cells

growinginminialmediumwerecollected by filtra- tionon amembranefilter, washed with and suspended in0.1MTris-chloride (pH 8.0)at109 cellsperml and incubatedat370C for10min. ThenTris-ethylenedia- minetetraacetic acid(EDTA) (pH 8.0)wasaddedto2 mM, and5min laterMgCl2wasaddedto4mM.The

cells were collected by filtration, washed, and sus-

pended ⁱⁿ 20 mM morpholinepropanesulfomc acid (MOPS)-135 mM choline-chloride (pH 7.5)at2x109

cells perml andkeptonice forupto1.5huntilused.

For uptake assays, the cell suspension was diluted withanequalvolume of135mMNaCI-50mMglu- cose-20mMK-MOPS (pH 7.5)-2 MMphenylcarbaun- decarborane and incubated at230Cfor 10min with aeration. Uptakewasinitiated by the addition of la- beled TPMP+orTPP',usuallyto10 MuM. Samples of 0.5mlwereremovedatintervals, washed with5mlof 10 mM MOPS-200 mM choline-chloride (pH 7.5), dried, and counted in a Packard liquid scintillation counter.Inhibitorswereaddedtocell suspensions 10 min before initiation of uptake of the lipid-soluble cation. Formeasurements of Af at a differentpH, either K-MES(pH5.5to6.5)orTris-chloride (pH 8.0 to8.5)wassubstituted forK-MOPS (pH 7.0to7.5)in the post-EDTA suspension andwashing buffers.

Theintracellularion concentration atequilibrium

was the plateau level reached within 25 min minus nonspecific uptake as determined by measuring the amountof cationremainingafter theaddition ofan

uncouplingconcentration of either1mMdinitrophe- nolor20 MMcarbonylcyanide m-chlorophenyl hydra-

zone.Intracellular concentrations werecalculated by the Nernst equation and avalue of 2.5 mlinternal volumeper g(dryweight) ofcells(14).

RESULTS

The membrane potential in bacteriasuch as

E. coliis generated by theextrusionofprotons coupledtooxidative reactions in the cellmem-

brane. This process generates a potential for protons,orprotonmotiveforce(PMF), theterm introduced by Mitchell (21). The PMF consists ofanelectrical component, representedbyA+, andaprotonconcentration componentgiven by the difference in pH. Because the interior of bacteria is maintainednearvH 7.5. thereislittle difference inpHacrossthe membranes ofcells in mediumat pH 7.5,so thatalmostall ofthe PMF isexpressedA4 (22).

Drugs commonlycalleduncouplersact toin-

creasethepermeability of protonsthrough the membrane, reducingthe PMFbydischarging it.

AtpH 7.5, uncouplers reduce

&4.

Table2shows the effect of1mMdinitrophenolontwosensitive strainsofE. coli.Thisconcentration ofuncou-

pler reducedthe growthrateby 20%andmod- estlyincreased resistancetodihydrostreptomy-

TABLE 2. Effect of2,4-dinitrophenol on&Pand MICfordihydrostreptomycin

Dinitro- MIC

Strain phenol b4 (mV)

(JAg/ml)

(MM)

FRAG-5 0 -142 0.5

1 -129 1

TK1208 0 -162 0.3

1 -133 0.5

VOL. 20, 1981

cin. Under the same conditions A4 showed a moderatedecrease in both strains, suggestinga correlation between

AO

anddrugresistance.

Areduced PMF isassociatedwithsome unc mutations (1). Strain TK1207 with the unc-40 mutation is presumedto have areducedPMF because it hasamarkedly reducedrateofproline uptake, a process known to be driven by the PMF(17). The reducedPMFinsuchuncstrains is due toabnormalleakage ofprotons through the membrane portionoftheproton-translocat- ingATPase. Thisleak can be blocked by dicy- clohexylcarbodiimide (DCCD) (1, 24). As seen in Table 3, strain TK1207 had a low

AO,

that increasedin the presence of0.5 mMDCCDby 30mVtovaluestypical of wild-type strains. At thesametime the MICwasreducedby afactor of 2. As control, we examined strain TK1208 with an unc mutation not associated with a proton leak. In this strain, treatment with DCCD caused no change in MIC and only a slight changein

AO',

^{suggestingthat the}effectof DCCDonstrain TK1207 is due to areduction intheprotonleak throughthemembranecom- ponent oftheATPase.

The additionofsalt is knowntoincreasere- sistanceto aminoglycosides (20, 30) andto in- hibit gentamicin uptake (10). We found that such resistancewasassociated withareduction inA4 in bothwild-typeandmutantstrains(Ta- ble4). Intwoof the strains0.5MNaCl allowed

TABLE 3. Effect of DCCDonAO\and MICfor dihydrostreptomycin

strain DCCD (nV) MICml)

TK1207 0 -120 12

0.5 -150 5

TK1208 0 -162 0.5

0.5 -155 0.5

TABLE 4. Effectof NaCI andsucrose on

A4b

and MICfor dihydrostreptomycin

Strain Solute Concn(M) Ai (mV) (ug/ml)MIC

FRAG-5 -a -142 0.5

NaCl 0.5 -120 7

Sucrose 0.3 -114 10

TK1208 - -162 0.3

NaCl 0.5 -114 5

Sucrose 0.3 -106 0.6

TK1207 - -120 12

NaCl 0.2 -96 28

Sucrose 0.3 -102 22

TK1235 - -106 5

NaCl 0.2 -55 24

Sucrose 0.3 ND" 15 a None.

bND, Not determined.

AMINOGLYCOSIDE RESISTANCE 805 reasonableratesofgrowthandproducedover a 10-fold increase in MIC andreduction inAi\ of 22 to 48 mV. Inthe two mutants inwhich the high salt concentration virtually abolished growth, the lower concentration of0.2 M pro- duced two- to fivefold increases in MIC and sizablereductions inA+.Weexamined the effect ofanon-ionicsolute,sucrose,and foundsimilar effects (Table 4). The concentration ofsucrose used here, 0.3 M, is the osmotic equivalent of 0.17MNaCl,indicating thatosmoticallyequiv- alent concentrations ofsucrose are at least as effectiveasNaCl inproducingresistance. Thus the important parameter here appears to be osmolarity, notionic strength. The mechanism whereby Ai\ is reduced when osmolarity is in- creased isnotknown, but thecorrelation further supports a connection between resistance and

Alp.

Awiderrangeof values ofA4 wasachievedby varying extemal pH. As external pH was re- duced, the pH difference increased, resultingin acompensatoryfallinAA such that thesumof the two, thePMF, remainedvirtuallyconstant.

Our results(Fig. 1)aresimilartothosereported by others(31).Itisknown thatreducingexternal pH markedlyincreases resistance to aminogly- cosides (30),aneffect illustratedbyourdatain Fig. 2. The effects of external pH are another example, anda very strikingone, of the corre-

-1501

-140k

-120

E .J -100

-80k

TK1208

__-

0o- -

,'FRAG-5

i I

I1d

I C /

/'

6 7 8 9

pH

FIG. 1. Variationof

Au4

with externalpH. Thetwo

strainsexaminedweregrown inmedium at the in- dicatedpH, andA#wasmeasuredatthesamepH^as described in thetext. Eachpointis the averageof

severalmeasurementsperformedina singleexperi-

ment.

-601

300

It0 79 100

CP ~

'N

Z 30 A..

5

so 10- %\

o 0

:r w0.

I-C,) 0 a I

cr0.

u 0.13

0035 .

5 6

1L _t-1

7 8 9

pH

FIG. 2. Effect ofexternal pH on MIC for dihydro- streptomycin in thefourstrains used in thisstudy.

Datafor strain TK1207 do notextend belowpH6 because this straindoesnotgrowatmoreacidic pH.

lation between

AO

and resistance to aminogly- cosides.

To simplify the quantitative analysis of our results,wecombined data in Fig. 1and 2 and in Tables2through4for the singleplot of Fig. 3.

Thisfigure shows the relationship of the loga- rithm ofthe MICtoA+p.Inspite of somescatter, there isareasonably goodcorrelation between these two parameters. Thesolid line in Fig. 3, drawn by the method of least squares, is log MIC (ug/ml) = 3.3 - 0.025

AO

(mV). The cor- relationcoefficient,r, is-0.84.

Resistancetotwo otheraminoglycoside anti- biotics, kanamycin and neomycin, was also measured under many of theconditions studied here. Ineachcasethe relative changesin MIC wereidentical withinexperimentalerror tothose observed fordihydrostreptomycin.Wethusinfer thatthe correlation between resistanceand

AO

istrue forthethreeaminoglycosideswe studied andisprobablytrueforall aminoglycosides.

DISCUSSION

The mechanism ofaminoglycoside uptake is notwell understood. Holtjehas suggestedthat uptakeisbyatransport system forpolyamines (18). Thismodel is not readilyreconciled with somefindingsofBryan andcolleagues (6), nor withthediscordant effectsofpolyaminestarva- tiononpolyamine and aminoglycosidetransport (10). Nor is this model supported by the fact that mutants resistant from loss of a specific

transport system have not been reported (8).

Althoughit ispossiblethat the transportsystem involved is essential, we prefer the view that aminoglycosidesenterbymanydifferent trans- port systems,which haveslight affinityfor ami-

noglycosides. Candidates for such uptake in- clude systems for sugars, aminosugars, polya- mines, and amino acids. Ifuptake is bymany differentsystems, anycompetitiveeffectbysub- stratesofonesystemwould probably besmall, accounting for the failure to observecompetition by a large number of substrates each tested singly (8). We assume that uptake of thepoly- cationicaminoglycosidesby all systems is elec- trogenic so that changes in the driving force haveaprofound effect on the rate of uptake.For the PMF-driven proline and lactose transport systems of E. coli,itis known thatareduction in thedrivingforcemarkedly reduces the rate of uptake (23).

Bryanand colleagues have proposed that ami- noglycosides enter cells bytransport on quinones orothercomponents of the electron transport chain (8). More recently, they have modified the model to includeAA as thedrivingforce forentry via electron transport components (6). In our view, the available data do not support any direct role of electron transport, only anindirect roleincreatingandmaintaining A+,which serves as a driving force. The role of AA is strongly supportedby our data. Every condition we stud- ied, whether amutation,addition of aninhibitor, change in external pH, dr change in osmolarity,

100

--O

z 0 cr2 0 0

2c 0n Ir

1o0

0.1

'0

%\4 0 0

\% ^{o \}_4.0⁰ _4. 0

04. 4

\%O o\o

0

-40 -80 -120 -160

t&*(mV)

FIG. 3. Compositeplotofalldata obtainedinthis study, showing the relationship between the loga- rithm of the MIC fordihydrostreptomycin andA4+.

The solidline, drawn by the method of least squares, islog MIC(8g/ml) = 3.3 - 0.025AA (mV). Dashed linesrepresent the rangeof variation expected from random errorsasmentioned in the text.

ANTIMICROB. AGENTS CHEMOTHER.

VOL. 20, 1981

showed corresponding changes in resistance and

AO.

Ifanelectron transport component, for ex- ample,acytochrome,wereinvolvedintransport, then the hem mutant we studied would have muchhigher resistancethanwouldbepredicted onthe basis ofA+, becauseit would haveboth alower driving force and losttheuptakesystem.

Datafor the hem mutant fit well the data for other conditions included in Fig. 3, indicating that such unusual resistance is not seen. The increased sensitivitytoand uptakeofaminogly- cosides insomemutants thatoverproduce ter- minalrespiratory enzymesarereadilyexplained by the evidence ofagreaterdriving force(6)and therefore donotrequireanyspecificrole ofthese enzymesintransport.

Thevalidity of using lipid-soluble cationsto measure

A0

in bacteria hasrecently been dem- onstrated ingiantE.coli, in whichdirectmeas- urements of the potential can be made with microelectrodes (15). Our measurements of

AO

aresubjecttoan errorof about± 10mVbased onvariations we sawin values obtained under thesameconditionsondifferentdays.Themeas- urementofMICbytwofold dilutionassayresults in a standarddeviation factor of1.2 (16). The combination of these random errors, indicated by the dashedline inFig.3,accountsforagood deal ofthescatterinourdata.We thus conclude that thevalue ofAA is themajordeterminant of aminoglycoside resistance in the types ofmu- tants and under the conditions studied here where reduceduptakeappearstobe the mech- anism of resistance. However, not all of our scatterislikelytobe accounted forbyrandom errors ofmeasurement. Some of the more ex- tremevariations, suchasthe 10-fold difference in MIC for verysimilar values ofAA (Table3, data withDCCD)suggestthat other factors also play arole, albeit aminorone, in thistype of resistance.

Toobtainaquantitativerelationshipbetween therateofuptakeofadrugand thesensitivity

to the drug, we developed a simple model. In growing cells, the concentration ofdrug inside represents a balance between uptake and the dilutionproduced bycellgrowth. Forouranal- ysis,weassumethe accumulateddrugdoesnot exit and isbiochemicallyunchanged, that cells are inbalancedexponential growthsothatthe ratio of area to volume is constant, and that uptake undergivenconditionsisproportionalto the surface area ofthe bacteria. Let A be the areainsquarecentimeters,V bevolumeincubic centimeters, J the influx rate per unit area in micromolesperminuteper squarecentimeter,k be the growth rate constant per minute, and subscripts0andtrefertotimeofdrugaddition

AMINOGLYCOSIDE RESISTANCE 807 andalatertime, respectively.The totalamount accumulatedattimetis:

JAdt =

JAoeAddt

=JAo(ekt -1)/k Cell volumeattimetis

Voek',

sotheconcentra- tion C ofdrug at time t is C =

JAo(ekt,-

1)/

kVoekt.

^Whentislarge,

ek'

islargerelative to 1;

thus this expressionbecomes C=JD/k, where Dis the ratio of area tovolume,A/V.

Thisresultstatesthattheconcentration ulti- mately reached in the cell is directly propor- tionaltothe influxrateJand inversely propor- tional to the growth rate k. The MIC is the minimal externalconcentration atwhich inter- nal concentrationultimately reaches the lethal concentration. Thedependence of C on growth rate explains the well-known fact that rapidly growing bacteriaaremoreresistant to a number of antibiotics, including the aminoglycosides (30).The effect ofgrowth ratecontributes only a part of the variance in the data of Fig. 3, because growth rates varied by only slightly morethanafactorof 2.The majoreffectseems tobe thedependence of JonA+~. This depend- enceisvery steep; theMIC doublesforeach12 mVreductioninAA(Fig. 3).Asteepdependence ofinflux on A4A is consistent with uptake of a polycationic compound by electrogenic trans- port.

ACKNOWLEDGMENT

This work was supported by grant GM22323 from the Institute of General Medical Sciences of the National Insti- tutesofHealth andbygrantPCM7904641 from the National ScienceFoundation.

LITERATURE CITED

1. Altendorf, K.,F.MLHarold, and R. D.Simoni. 1974.

Impairment and restoration of the energized state in membrane vesicles of a mutant of Escherichia coli lacking adenosinetriphosphatase. J. Biol. Chem. 249:

4587-4593.

2. Bachmsnn,B.J., and K. B. Low. 1980.Linkage map of Escherichia coli K-12, edition 6. Microbiol. Rev. 44:1- 56.

3. Benveniste, R., and J.Davies. 1973. Mechanisms of antibiotic resistance inbacteria. Annu. Rev. Biochem.

42:471-506.

4. Bryan, L. E., R.Haraphongse,and H. M. Van Den Elzen. 1976. Gentamicin resistance in clinical-isolates of Pseudomonas aeruginosaassociated with dimin- ished gentamicin accumulation and no detectable en- zymatic modification. J. Antibiot.(Tokyo) 29:743-753.

5. Bryan, L. E., S. K. Kowand, and H. M. Van Den Elzen. 1979. Mechanism ofaminoglycosideantibiotic resistance in anaerobic bacteria: Clostridiumperfrin- gens and Bacteroides fragilis. Antimicrob. Agents Chemother. 15:7-13.

6. Bryan, L. E., T. Nicas, B. W. Holloway, and C.

Crowther. 1980.Aminoglycoside-resistantmutation of Pseudomonasaeruginosa defective incytochrome c552

and nitrate reductase. Antimicrob. Agents Chemother.

17:71-79.

7. Bryan,L. E., and H. M. Van Den Elzen. 1976. Strep- tomycin accumulation in susceptible and resistant strainsof Escherichiacoli and Pseudononas aerugi- nosa. Antimicrob. AgentsChemother.9:928-938.

8. Bryan, L. E., andH. M. Van Den Elzen.1977.Effects of membrane-energy mutations andcationa on strep- tomycin and gentamicin accumulation by bacteria: a model for entry of streptomycin and gentamicin in susceptible and resistant bacteria. Antimicrob. Agents Chemother. 12:162-177.

9. Butlin, J. D., G. B. Cox, and F.Gibson.1971.Oxidative phosphorylation in Escherichia coli K-12. Mutations affecting magnesium ion- or calcium ion-stimulated adenosinetriphosphatase. Biochem. J. 124:75-81.

10. Campbell,B.D.,and R. J. Kadner. 1980. Relation of anaerobiosis and ionicstrength to the uptake of dihy- drostreptomycin in Escherichia coli.Biochim. Biophys.

Acta593:1-10.

11.Damper, P. D.,and W.Epstein. 1980.Mechanism of low-levelaminoglycoside resistance in Escherichia coli, p. 706-708. InJ. D.Nelson and C. Grassi (ed.), Current chemotherapy and infectiousdiaease,vol. 1. American SocietyforMicrobiology, Washington,D.C.

12.Downie, J. A., F. Gibson, and G. B.Con.1979.Mem- brane adenosinetriphosphatases of prokaryotic cells.

Annu.Rev. Biochem. 48:103-131.

13. Epstein, W.,and B.S.Kim.1971.Potassium transport loci inEscherichia coli K-12. J. Bacteriol. 108:639-644.

14. Epstein, W.,and S. G. Schultz. 1965. Cationtransport inEscherichia coli. V. Regulation of cation content. J.

Gen.Physiol.49:221-234.

15. Felle, H.,J. S. Porter, C. L. Slayman, and H. R.

Kaback. 1980. Quantitative measurements of mem- branepotential in Escherichia coli.Biochemistry 19:

3585-3589.

16. Finney,D. J.1952.Statisticalmethods in biological assay.

C.Griffin,London.

17.Hirata, H.,K.Altendorf,andF. M. Harold.1973.Role of an electricalpotential in the coupling ofmetabolic energy to active transport by membrane vesicles of Escherichia coli.Proc.Natl. Acad. Sci. U.S.A. 70:1904- 1908.

18.Holtje,J.-V. 1978.Streptomycin uptake via an inducible polyamine transport system in Escherichia coli. Eur. J.

Biochem. 86:345-351.

19.Kanner,B.I., and D. L. Gutnick.1972.Use of neomycin in theisolation of mutants blocked in energy conserva-

tion in Escherichiacoli. J. Bacteriol. 111:287-289.

20. Medeiros, A.A., T. F.O'Brien, W. E. C. Wacker, and N. F. Yulug. 1971. Effect of salt concentration on the apparent in vitro susceptibility of Pseudomonas and othergram-negativebacilli to gentamicin. J. Infect. Dis.

124(Suppl.):S59-S64.

21. Mitchell,P. 1966. Chemiosmotic coupling in oxidative and photosynthetic phosphorylation.Biol. Rev. 41:445- 502.

22.Padan, E., D.Zilberstein, and H.Rottenberg. 1976.

The proton electrochemical gradient in Escherichia coli cells. Eur. J. Biochem. 63:533-541.

23. Robertson,D.E., G.J.Kaczorowski,M.-IL, Garcia, and H. R. Kaback. 1980. Active transport in mem- branevesicles of Escherichia coli: the electrochemical protongradient alters the distribution of the lac carrier between twodifferent kinetic states.Biochemistry19:

5692-5702.

24.Rosen, B. P. 1973./8-galactosidetransport and proton movements in an adenosine triphosphatase-deficient mutant of Escherichia coli. Biochem. Biophys. Res.

Commun.53:1289-1296.

25. Rottenberg, H. 1979. The measurement of membrane potential and ApH in cells, organelles, and vesicles.

MethodsEnzymol.55:547-569.

26. Sisirman, A.,M. Surdeanu,V.Portelance, R. Do- bardzic, and S. Sonea.1969. Vitamin K-deficient mu- tantsof bacteria. Nature(London)224:272.

27. Sis&rman, A., M. Surdeanu, G.Szegli,T. Horodni- ceanu, V. Greceanu, and A. Dumitrescu. 1968.

Hemindeficient mutants of Escherichia coli K-12. J.

Bacteriol. 96:570-572.

28. Taber, H.,and G. M. Halfenger. 1976.Multiple-ami- noglycoside-resistant mutants of Bacillus subtilis defi- cient in accumulation of kanamycin. Antimicrob.

Agents Chemother. 9:251-259.

29. Tien, W., andD. C. White. 1968. Linear sequential arrangement of genes for thebiosynthetic pathwayof protohemeinStaphylococcusaureus. Proc. Natl.Acad.

Sci. U.S.A. 61:1392-1398.

30. Youmnans, G. P., and M. W.Fisher. 1949. Action of streptomycin onmicroorganismsinvitro, p. 91-111. In S.A.Waksman(ed.),Streptomycin,natureand practi- calapplications.Williams&WilkinsCo., Baltimore.

31. Zilberstein, D.,S. Schuldiner, and E. Padan. 1979.

Proton electrochemical gradient in Escherichia coli cells and its relation to active transport of lactose.

Biochemistry18:669-673.