Effects of Membrane-Energy Mutatlns and Cations on

Transcription

1 ANTIMICROBIAL AGzNTS AND CHUmOTHzRAY, Aug. 1977, p Vol. 12, No. 2 Copyright 1977 American Society for Microbiology Printed in U.S.A. Effects of Membrane-Energy Mutatlns and Cations on Streptomycin and Gentamicin Accumulation by Bacteria: a Model for Entry of Streptomycin and Gentamicin in Susceptible and Resistant Bacteria L. E. BRYAN* AN H. M. VAN DEN ELZEN Department of Medical Bacteriology, University ofalberta, Edmonton, Alberta, Canada Received for publication 28 September 1976 Several mutants of Escherichia coli affecting aerobic energy generation and energization of the bacterial membrane have been examined for their effect on streptomycin and gentamicin accumulation and susceptibility. A heme-deficient mutant (K27) and two mutants (CJ-8 [colicin K insensitive] and NR-7) associated with defective aerobic active transport were associated with decreased transport of streptomycin and gentamicin and increased resistance to those antibiotics. These mutants also exhibited increased resistance to several other aminoglycoside antibiotics, but not the aminocyclitol spectinomycin. The same observations were made with a ubiquinone-deficient mutant, but a stra derivative of this mutant was shown additionally to be saturable for streptomycin accumulation at a concentration four or more times lower than that required for saturation of the parent. A mutant uncoupled for adenosine 5'-triphosphate synthesis from electron transport and membrane Mg-adenosine 5'-triphosphatase deficient was hypersensitive to those aminoglycosides tested and spectinomycin, and showed enhanced transport of streptomycin and gentamicin. A variety of compounds structurally related to streptomycin were examined at high concentrations for inhibition of streptomycin uptake in a stra mutant ofe. coli K-12 SA 136, but no evidence for competition was detected, suggesting the absence of a common transport carrier. Four different divalent cations were shown to inhibit streptomycin and gentamicin accumulation in E. coli K-12 SA 136. Divalent cations were shown to inhibit uptake of these two drugs in two bacterial species with distinct cell wall structures, Pseudomonas aeruginosa and Staphylococcus aureus, and to inhibit streptomycin uptake in spheroplasts of streptomycin-susceptible and -resistant E. coli. However, calcium had almost no inhibitory effect on streptomycin uptake by the ubiquinone-deficient mutant E. coli AN66. These and previous findings have been used to formulate a model for aminoglycoside entry into bacteria using a low-affinity membranous complex involved in membrane energization that includes respiratory quinones, which probably act to bind and transport aminoglycosides across the cell membrane. This phase of transport is associated with the lowest accumulation rate (termed energy-dependent phase I) that is rate limiting for susceptibility. It is further proposed that subsequent association of the membrane-bound aminoglycoside with higher-affinity binding sites on membrane-associated ribosomes carrying out a normal ribosomal cycle and protein synthesis results in a more rapid transport rate (termed energy-dependent phase II). The increased rate could result from a state of membrane energization analogous to that causing enhanced aminoglycoside transport rates seen in the uncoupled mutant, AN12. How this model explains the mechanism by which enzymatically modified aminoglycosides render cells resistant to unmodified aminoglycosides is also discussed. The action of the aminoglycoside antibiotics cumulation by bacteria is highly dependent on streptomycin and gentamicin on bacteria is de- conditions used to measure antibiotic uptake, pendent on a source of aerobically generated but in general three or more phases of accumuenergy (9, 1). Streptomycin or gentamicin ac- lation can be defined. The initial, very rapid 163

2 164 BRYAN AND VAN DEN ELZEN energy-independent phase is generally agreed to be due to ionic interaction of streptomycin or gentamicin with cell surface components. The next phase of uptake precedes the onset of loss of viability and inhibition of protein synthesis. In the case of streptomycin, it is the only additional phase detected when ribosomally streptomycin-resistant mutants are studied. However, in these mutants as well as in the streptomycin-susceptible parent, this second phase of streptomycin accumulation is inhibited by selected inhibitors of electron transport and oxidative phosphorylation (1). A second phase of gentamicin accumulation also occurs and has similar characteristics. However, well-defined ribosomally resistant gentamicin strains have not been reported. The third phase of streptomycin accumulation requires streptomycin-susceptible ribosomes, and for both streptomycin and gentamicin it has a much increased rate of accumulation compared with the second phase but retains identical energy requirements. Throughout this report, the phases of accumulation will be referred to as energy-independent (initial) and energy-dependent phase I (EDP-I, second phase) and phase II (EDP-ll, third phase). The onset of EDP-ll corresponds with the onset of inhibition of protein synthesis and loss of cell viability. We have proposed that the second phase (EDP-I) of uptake kinetics represents transport ofthe aminoglycoside across the cell membrane by a process that uses energy obtained from electron transport and oxidative phosphorylation but not from membrane Mg-adenosine 5'- triphosphatase (ATPase)-specified adenosine 5'-triphosphate (ATP) hydrolysis. To examine this proposal, we have examined various mutants ofmembrane energization for their effects on aminoglycoside susceptibility and accumulation and propose a model of streptomycin and gentamicin bacterial entry. This model also explains the mechanism by which R-factors specifying streptomycin or gentamicin inactivation achieve resistance to these drugs. MATERUILS AND METHODS Bacterial strains. Sources and characteristics of strains used are given in Table 1. Chemicals. 8-Aminolevulinic acid (8-ALA) hydrochloride was obtained from Sigma Chemical Co. All other chemicals were of reagent grade. Media. The medium used for growth and streptomycin or gentamicin accumulation studies, unless otherwise specified, was a peptone-supplemented low-phosphate, low-magnesium tris(hydroxymethyl)aminomethane (Tris)-buffered basal saltsglucose medium (aminoglycoside uptake medium). The medium contained, per liter: NH4Cl, 5 mg; ANTimicROB. AGZNTS CHEMOTHZR. NH4NO3, 1 mg; Na2SO4, 1 mg, MnSO4 4H2O,.1 mg; FeSO4 7H2,.5 mg; Tris, 12.1 g; gelysate peptone (BBL), 2.5 g; glucose, 2 g; and MgSO4. 7H2, 4 mg. Potassium phosphate buffer, ph 7., was added to a final concentration of.3 mm. E. coli strains AB2154 and AN66 also required.1% yeast extract (Difco) to support adequate growth. Nutrient broth was obtained from BBL. Streptomycin and gentamicin uptake. The uptake of [3H]dihydrostreptomycin or [3H]gentamicin by a membrane filter method was identical to that previously described (9, 1) except that in all cases, unless otherwise specified, aminoglycoside uptake medium was used for growth and uptake. All uptake values shown are total accumulation values except in Table 4, where zero time has been subtracted. Spheroplast preparation and streptomycin uptake. E. coli K-12 SA 136 was grown in Trypticase soy broth (BBL) to a density of.5 unit of absorbance at 6 nm (A wo), centrifuged at 7, x g for 1 min at 4C, and suspended in 2 ml of 15% (wt/vol) sucrose in 4 mm Tris-hydrochloride, ph 8.. bysozyme (2,ug/ml) was added, the mixture was incubated at 28 C for 2 min, and 4 ml of.2 M Trishydrochloride buffer (ph 8.) was added, followed by sodium ethylenediaminetetraacetic acid to a final concentration of 2 mm. Incubation was continued at 28C until about 99% of cells were spheroplasts as determined by phase-contrast microscopy. The suspension was chilled in an ice bath and centrifuged at 7, x g for 1 min at 4 C. The pellet was resuspended in 1 ml of 15% sucrose in 4 mm Tris-hydrochloride buffer (ph 7.6). This mixture was incubated with deoxyribonuclease I (2,ug/ml) and ribonuclease (4,g/ml) (Worthington Biochemicals Corp.) for 3 min at 28 C, centrifuged at 7, x g for 1 min at 4C, and suspended in 7.5 ml of 15% (wt/vol) sucrose in nutrient broth. Portions of each suspension, when diluted 1:1 in water, decrease in Awo more than 2 times that of the same suspension diluted 1:1 in 15% sucrose. Streptomycin uptake was carried out as for whole cells except that nutrient broth with 15% sucrose was the uptake medium, using an initial Awo of.5. Absorbance was monitored throughout uptake periods to assess cellular lysis or decrease in cell density, and cell mass determinations were made at the start and finish of each accumulation period. Determination of susceptibility of bacterial strains to aminoglycoside and aminocyclitol antibiotics. Minimal inhibitory concentration (MIC) values using a tube dilution method and growth responses were determined in aminoglycoside uptake medium. Disk susceptibilities were performed with commercial disks by the standard diffusion method of Bauer et al. (3), using Mueller-Hinton medium (BBL) with adjusted cation concentration (8). RESULTS Mutants defective in heme synthesis. Strain K27 requires 8-ALA to synthesize heme (17). In the absence of this compound, cytochrome apoproteins are formed but the heme compo-

3 VOL. 12, 1977 E. coli K-12 AB2154 E. coli K-12 AN66 E. coli K-12 strain 7 E. coli K-12 NR-7 E. coli K-12 CJ-7 E. coli K-12 CJ-8 E. coli K-12 CJ-11 Pseudomonas aeruginosa 28 P. aeruginosa 28 Met-Strr Staphylococcus aureus UAl STREPTOMYCIN AND GENTAMICIN ACCUMULATION 165 TABLz 1. Bacterial strains, sources, and characteristics Strain Source Characteristics Escherichia coli K-12 SA 136 K. E. Sanderson, University of Pro- Met- Nalr Calgary, Calgary, Alberta, Canada E. coli K Str' Obtained by selection for resistance to 2, gg of streptomycin per ml E. coli K-12 K27 P. Overath, Max-Planck-Institut fur Biologie, Freiburg, West Germany E. coli K-12 AN18 F. Gibson, Australian National arge3 thi-1 stra University, Canberra E. coli K-12 AN12 F. Gibson F. Gibson F. Gibson B. P. Rosen, University of Maryland, Baltimore B. P. Rosen Charles Plate, Massachusetts Institute of Technology, Cambridge Charles Plate Charles Plate Our laboratory Obtained by selection for resistance to 2,,ug of streptomycin per ml Strain collection, Dept. of Medical Bacteriology, University of Alberta, Edmonton, Alberta, Canada Unpurified ribosomes do not bind streptomycin above background, stra 8-ALA-requiring mutant arge3 thi-1 stra unca41, (Unc-, membrane Mg-ATPase deficient) derived from AN18 mete leu-6 thr-1 thi- stra leu-6 thr-1 ubid stra thi-, derived from AN2154 Cavalli Hfr ATPase- Unc-, defective in aerobic transport of several compounds; mutant of strain 7 Thi- Suc+ colicin K sensitive (Suc+ transductant of CJ-7 with B51 as donor) Thi- Suc- colicin K insensitive (Suc- transductant of CJ-7 with B51 as donor) Thi- Suc& (Su& transductant of CJ-7 with B51-7 as donor) Met- Unpurified ribosomes do not bind streptomycin Wild type nent is absent. MIC values of gentamicin and streptomycin for strain K27 with and without 8-ALA are shown in Table 2. Measurement of growth response and of viability after treatment of strain K27 with and without 8-ALA and with either streptomycin or gentamicin (not shown) demonstrated about a four- to eightfold increase in resistance without 8-ALA for streptomycin and a fourfold increase for gentamicin. Resistance to the other aminoglycosides, kanamycin, amikacin, tobramycin, and netilmicin, was also increased in the cytochrome-deficient state as determined by disk diffusion testing. However, interestingly, the opposite effect, an increase in susceptibility to the aminocyclitol spectinomycin, occurred. This drug differs from the aminoglycosides listed in being bacteriostatic and more active on anaerobic bacteria. The increased resistance of 8-ALA-requiring mutants was associated with reduced accumulation of [3H]dihydrostreptomycin and [3H]gentamicin when strain K27 was grown in the presence of limiting amounts of 8-ALA (Fig. 1B and 2). Accumulation of both drugs was markedly increased by the addition of 8-ALA to the medium. Strain K27 does utilize some oxygen in the absence of 8-ALA, but this does not apparently involve cytochromes (17). These findings agree with previous results obtained with various inhibitors of electron transport (1). They confirm that electron transport is an important source of energy for streptomycin or gentamicin entry. However, agents that uncouple electron transport and oxidative phosphorylation but do not inhibit electron transport also prevent streptomycin or gentamicin entry. Thus, electron transport does not appear to act directly as, for example, a carrier system, but rather only as an energy source.

4 166 BRYAN AND VAN DEN ELZEN TABLsz 2. ANTimcROB - MICs of streptomycin, gentamicin, and spectinomycin for strains ofe. coli defective in various aspects of membrane energization and for control strains MIC (pg/ml) Strain Description Streptomycin Gentamicin Spectinomycin K27 8-ALA requiring, -8-ALA K27 8-ALA requiring, +8-ALA AN12 Unc-, membrane Mg-ATPase AN18 Parent of AN AN66 Ubiquinone deficient >4, AN2154 Parent of AN66 2, CJ-8 Suc- transductant of CJ-7, colicin K in- 2, sensitive CJ NR-7 Unc-, membrane ATPase-, deficient aerobic transport Strain 7 Parent of NR AGENTS CHZMOTHZR. Uncoupled (unc-) and membrane Mg-ATPase-deficient mutants. Facultative bacteria may apparantly supply energy for pyridine nucleotide transhydrogenase and some active transport processes by ATP hydrolysis using the membrane Mg-ATPase as well as by electron transport. Several mutants have been isolated in which the coupling of the above functions and oxidative phosphorylation to electron transport has been affected. Mutant AN12 is unable to couple oxidative phosphorylation to electron transport and is defective in membrane Mg-ATPase- and ATP-dependent pyridine nucleotide transhydrogenase functions (14). Mutant AN12 was hypersensitive to gentamicin in that growth was inhibited with lower concentrations than that needed to inhibit the parent strain (Table 2). Although this mutant carries a stra mutation, the level of resistance to streptomycin. was also significantly reduced (Table 2). These observations are associated with increased accumulation of either aminoglycoside at similar drug concentrations (Fig. 3 and 4B). Susceptibility to kanamycin, amikacin, tobramycin, neomycin, and netilmicin was also increased, as was susceptibility to the aminocyclitol spectinomycin. Mutant AN12 demonstrates that possession of a functioning Mg-ATPase normally coupled with an energized membrane is not required for cellular entry of streptomycin or gentamicin. In fact, loss of these functions results in increased accumulation of these drugs. This increase in accumulation could result from increase either in quantity or utilization of a membrane component responsible for aminoglycoside transport. A mutant isolated by Rosen (33), NR-7, is membrane Mg-ATPase deficient and unable to couple ATP synthesis to electron transport, but in contrast to AN12 shows heightened resistance to several aminoglycosides, including streptomycin and gentamicin but not spectinomycin (Table 2), and reduced transport of streptomycin and gentamicin (Fig. 1A and 2A). This mutant differs significantly from AN12 in that aerobic transport of several solutes is impaired with glucose as an energy source. It has been proposed by Simoni and Postma (35) that the defect present in NR-7 results in a membrane that is leaky to protons due to a defective ATPase and is unable to maintain the high energy state generated by respiration. Insofar as streptomycin and gentamicin transport are concerned, NR-7 appears to have some defect in addition to that present in AN12 and which is intimately concerned with aminoglycoside entry. It is probable that NR-7 involves a genetic deletion, since revertants have not been obtained. Some of the properties seen with NR-7 are similar phenotypically to ubiquinone-deficient (ubi-) mutants in that aerobic transport of some solutes with glucose as an energy source is impaired even though ubiquinone-deficient strains have functional membrane Mg- ATPase activity. These strains act as if it were difficult to generate or maintain a high energy state from either respiration or ATP hydrolysis. However, electron transport is defective in ubi - mutants, which is not true of NR-7. Colicin K-insensitive mutant. Another mutant, isolated by Plate (32), which, like NR-7, exhibits a defect in aerobic active transport, is strain CJ-8. This mutation also confers resistance to colicin K, but the strain has normal membrane Mg-ATPase activity and electron transport. Thus the mutant is distinct from NR-7, ubid, unca, and uncb mutants. Strain CJ-8 compared with the parent, CJ-7, showed about a fourfold increase in the MICs of gentamicin and streptomycin (Table 2) and a significant increase in resistance to other aminoglyco-

5 VOL. 12, 1977 STREPTOMYCIN AND GENTAMICIN ACCUMULATION 167 sides as determined by disk susceptibility testing. Spectinomycin resistance was very slightly lowered in CJ-8 relative to CJ-7 (Table 2). The increased aminoglycoside resistance was associated with reduced gentamicin (Fig. 2B) and streptomycin accumulation (not shown). The exact nature of this mutation is not known, but it has phenotypic similarities with NR-7 and ubi- mutants. The mutation presumably affects a protein, since a temperature-conditional mutant, CJ-11, has been prepared. This last mutant shows the CJ-8 phenotype at 42 C and the CJ-7 phenotype at 27C (32). CJ-11 exhibits increased aminoglycoside resistance at 42C, but has the susceptibility pattern of CJ-7 at 27C. This protein presumably functions primarily in respiratory-linked active transport, as do ubiquinones and some of the structure or structures affected in mutant NR-7. ubi- mutants. Strains of E. coli defective in ubiquinone synthesis have been isolated by Cox et al. (15, 3). We have studied mutant ubid (AN66), which synthesizes about 2% of the normal quantity of ubiquinone. This mutant exhibits increased resistance to all aminoglycosides we have tested and reduced capability to A. t. C, 323 Stroin 7.- o NR -7o - * i 1 16 v1-2 ty I. *& * qk TIME IN MINUTES FiG. 1. Accumulation of streptomycin at 37 C with time. (A) E. coli strains 7 and NR-7 at 2 Pg of streptomycin per ml; (B) E. coli K27 in aminoglycoside uptake medium with 2 x 1-5 M 6-ALA added or with no 8-ALA added at 4 pg of streptomycin per ml. 11 S. CJ7 CJ.E z EUt _----*----, z 2, K 2oo * ALA - - ALA 12 IOC so ff C 2 TIME IN MINUTES FIG. 2. Accumulation ofgentamicin at 37 C with time at 1.5 pg ofgentamicin per ml. (A) E. coli strains 7 and NR-7; (B) E. coli CJ-7 and CJ-8; (C) E. colik27 in aminoglycoside uptake medium with 2 x 1-5M & ALA added or with no -ALA.

6 ._ 168 BRYAN AND VAN DEN ElLZEN ANTIMICROB. AGENTS CHEMOTHER. accumulate gentamicin and streptaomycin. The underestimate the effect of the ubiquinone-dei its parent fective state in that AN66 required at least results for the ubid mutant an' strain, AB2154, are presented in Table 2 and eight times the concentration of streptomycin Fig. 4A and 5. The MIC values showvn in Table 2 or gentamicin (and of kanamycin, amikacin, tobramycin, netilmicin, or neomycin) to inhibit growth within 2 or 3 h at 37C compared with AB2154. A similar effect was seen in disk testing in that two zones of inhibition for each of these aminoglycosides were detected. A significant inner zone of growth was nearly as prominent as the outer growth after 8 to 18 h of incubation at 37 C. With longer incubation, the inner growth zone did not significantly increase in density and was clearly inhibited, whereas the outer growth zone continued uninhibited ) -E E ) z 2.41,,,--''; growth. The outer growth began for each disk equidistant to that of the parent strain, V AB2154, and the parent never showed the inner L) 2.1 growth zones, which usually occupied about one-half the distance between the disk and the outer zone for strain AN66. Therefore, growth 1.61 ~~ ~ ~ ~ ~ ~ ~ ~. of AN66 appears to be inhibited much more slowly than that of the parent, but after a 6~~~~~~~~~~~~~~~~~. Lu longer period the mutant's growth is inhibited The reason for this observation is seen in Fig. 4A and 5. The rate of EDP-I gentamicin accumulation for AN66 was reduced below that of I.-.8 4,) AB2154, and the duration of that phase was very much prolonged (16 min versus 4 min).4 (Fig. 4A). Thus, at equivalent concentrations AN66 transports gentamicin more slowly. The effect for streptomycin is even more obvious AN66 and AB2154 carry stra mutations and thus show only EDP-I transport. At all concen- MINUTES trations in Fig. 5, the initial rates of EDP-I FIG. 3. Accumulation of streptomycin at 37C accumulation of streptomycin were greater for with time by E. coli AN18 (@) and. AN12 (V) at strain AB2154. In addition, total accumulation 125 pg of streptomycin per ml. was greater in strain AB2154, especially at -.2 AB B z Z TIME IN MINUTES FIG. 4. Accumulation ofgentamicin at 37 C with time at 1.5 pg ofgentamicin per ml. (A) E. coli AB2154 (a) and AN66 (V). Aminoglycoside uptake medium was supplemented with 1 %yeast extract (Difco). (B) E. coli AN18 (c) and AN12 (V).

7 ) ) E ) I- z Lu I.- LI) A. STREPTOMYCIN 125,ug/ml -.,. - t---- T A I 1I Minutes B. STREPTOMYCIN 25,ug/ml C. STREPTOMYCIN 5 ug/ml D. STREPTOMYCIN 16 1 ug/ml 2 ~~~~~~ 14 E 12 Z!- D 8- <6 z Minutes FIG. 5. Accumulation of streptomycin at 37 C by E. coli AB2154 (a) and AN66 (V) at four streptomycin concentrations. 169

8 17 BRYAN AND VAN DEN ELZEN 1,,ug/ml. In the last case, AB2154 showed a significant increase in total accumulation of streptomycin from that at 5,ug/ml, whereas this did not happen for mutant AN66. Therefore, mutant AN66, unlike strain AB2154, exhibits saturation for streptomycin accumulation at 5,ug/ml, implying that ubiquinones are directly responsible for streptomycin accumulation, most likely as direct binding sites. Alternatively, ubiquinones could act stoichiometrically as direct energy sources to a transport molecule. Mutant AN66 did not show a reduction in susceptibility to spectinomycin (Table 2). This is an interesting observation because, as noted, spectinomycin is more active on obligately anaerobic bacteria than other aminoglycosides and, unlike most other aminoglycosides, is bacteriostatic. The parent strain or spontaneously arising revertants (the revertants showed the same aerobic growth yield as the parent) regained general aminoglycoside susceptibility (but no change for spectinomycin) and showed enhanced accumulation of streptomycin and gentamicin. Carrier-mediated streptomycin-gentamicin entry. Results obtained using various mutants affecting aerobic membrane energization clearly demonstrate, in E. coli at least, that this function plays a central role in streptomycin and gentamicin bacterial entry. It would not be unreasonable, however, to assume that the components defective in these mutants act as intermediates in energizing a carrier mediating these drugs' transport rather than as direct binding sites. It would seem likely that such a carrier or carriers would function primarily to mediate transport of some other solute and not exist specifically for aminoglycoside transport. Under these conditions, the transport of streptomycin and gentamicin would be fortuitous. We have examined the effect of various compounds structurally related to the two drugs on uptake of [3H]dihydrostreptomycin, using a stra mutant of E. coli K-12 SA 136. The compounds tested were: glucose, a-methyl- D-glucoside, N-acetylglucosamine, glucosamine, xylose, ribose, arabinose, 3-N-acetylsisomicin, 6-N-acetylsisomicin, 1-N-acetylsisomicin, 2,3,3,6-tetra-N-acetylsisomicin (streptomycin sulfate or modified sisomicins [as sulfates] at 2 mg/ml; other compounds at 1 mm), inositol, lactose, galactose, putrescine, arginine (at a concentration of 2 mm arginine a 3% reduction in streptomycin accumulation occurred, but at 1 mm only a 1% reduction was ANnmicRoB. AGzNmm CHZMOTHZR. detected; similar reductions were readily obtained nonspecifically by a variety of monovalent ions at concentrations of 5 to 1 mm), thiomethyl-d-galactoside, and 2-deoxystreptamine. We used 2,,g of streptomycin per ml in uptake mixtures because the uptake of streptomycin at this concentration was no longer linearly related to streptomycin concentration. Therefore, under these conditions we considered it probable that the streptomycin transport sites were nearing saturation. Each of the solutes examined for competition with streptomycin was used at a concentration of 1 mm, which we know in some cases and expected in all cases was a concentration that saturated transport carriers specific for these compounds. In no case was a significant reduction of streptomycin uptake detected. Our conclusion from these studies is that streptomycin is not transported by a transport system specific for the compounds examined. Although it is very difficult to exclude all possible transport carriers, we feel that the most probable transport carriers have been shown not to be used for streptomycin transport. Cations and streptomycin-gentamicin accumulation. Cations inhibited accumulation of streptomycin and gentamicin. Ca2+, Mg2+, Co2+, or Mn2+ inhibited streptomycin accumulation at concentrations as low as.2 mm (Table 3) in E. coli K-12 SA 136. Similar results were obtained for gentamicin accumulation. These observations strongly suggest that a specific transport system for one of the cations is not involved, since no system has been described which transports all four cations, and Ca2+ appears to be actively transported out of bacteria. The decrease in streptomycin and gentamicin accumulation due to cations resulted in decreased susceptibility to these drugs. This effect was more marked with Pseudomonas aeruginosa than E. coli. As a result, the effects of cations have been associated with cell wall structure, but from the results which follow we believe that this is not the major site of cation antagonism. Cations antagonize streptomycin susceptibility and accumulation in Mycobacterium smegmnatis. Beggs and Andrews (5) have shown that magnesium will reverse the killing effect of streptomycin during the time period between initial drug exposure and the onset of killing. This suggests that cations will displace streptomycin from some cellular site. Using a stra mutant of E. coli K-12 SA 136 in which the ribosome-dependent phase of streptomycin accunmulation (EDP-II) is not seen, calcium or magnesium inhibited secondphase streptomycin accumulation (Fig. 6). We believe that this accumulation phase involves the transport of streptomycin across the cell

9 VOL. 12, 1977 STREPTOMYCIN AND GENTAMICIN ACCUMULATION TABLE 3. Streptomycin uptake by streptomycin-susceptible strains of bacteria in uptake mixtures with and without cations Strain Cation Concn (mm) Streptomycin up- Streptomycin take (nlmg dry concn wt] (jugim1) per 15 min) Escherichia coli None 3,954 1 K-12 SA 136 Magnesiumb Manganese Cobalt Calcium Magnesium.2 2,825 1 Manganese.2 2,91 1 Cobalt.2 3,544 1 Calcium.2 3,7 1 Staphylococcus None 14 1 aureus UA1 Calcium 1 <1 1 None 15 5 Magnesium Pseudomonas None aeruginosa 28 Calcium None Calcium a Uptake was done in nutrient broth (BBL) with defined cation concentrations; zero-time values have been subtracted. b All cations used as chloride salts. 171 membrane. These last results show that cations antagonize the transport of streptomycin and not simply ribosomal binding. To determine whether the site of action of cations was the bacterial cell wall, two additional studies were performed. Calcium antagonized streptomycin uptake in Staphylococcus aureus, P. aeruginosa StrS (Table 4), and P. aeruginosa Strr (Fig. 6) in addition to E. coli. Calcium (or magnesium) also antagonized gentamicin accumulation in P. aeruginosa and S. aureus. These three bacteria have markedly different cell wall structures, yet all showed the effects of calcium. As noted, M. smegmatis also shows this effect (5). Magnesium and calcium inhibited streptomycin accumulation by spheroplasts of both E. coli K-12 SA 136 streptomycin-susceptible and ribosomally streptomycin-resistant strains (Fig. 7), again supporting the conclusion that the membrane is the major site of action of the cation effect, although cation-mediated impairment of streptomycin cell wall penetration may also occur. E. coli K-12 SA 136 cells grown in a nutrient broth with calcium added to a final concentration of 1 mm were resuspended in nutrient broth either with or without added calcium, and the effect on accumulation of streptomycin or gentamicin was determined (Table 4). The first sample, taken 2.5 min after the addition of the antibiotic, showed that the cells without added calcium were already accumulating more streptomycin or gentamicin. In view of the fact that the generation time in this medium was 75 min, it seems unlikely that the effect of calcium restriction is to cause a structural change of the cell wall or cell membrane that facilitates streptomycin or gentamicin accumulation. These studies agree with those of Beggs and Andrews using M. smegmatis (5) in that cations seem to compete with streptomycin (or gentamicin) for a membrane site that is essential to streptomycin transport. Studies with the E. coli ubid mutant AN66 and the parent, AB2154, showed that calcium was much less effective at inhibiting streptomycin or gentamicin uptake in the ubid stra mutant (Table 5). These observations support the view that the membrane is the major site of antagonism of streptomycin accumulation by cations and further support a direct involvement of ubiquinones in streptomycin and gentamicin cellular entry. DISCUSSION Previous investigations have demonstrated that energy is required for the entry of streptomycin and gentamicin (2, 9, 1, 19, 2, 27, 37) into bacterial cells. Studies have demonstrated that aerobically generated energy will support streptomycin (2, 1, 19) or gentamicin (9) transport, but anaerobically generated energy is not capable of or much less effective at doing so (1, 27). The Mg-ATPase-deficient and uncoupled mutant AN12 supports these conclusions in that the mutant exhibits an actual increase in

10 - 172 BRYAN AND VAN DEN ELZEN 4 E~ ~ ~ ~~ ~lo62 s 3 z <v TIME IN MINUTES uj In FIG. 6. Accumulation ofstreptomycin at 1 pg/ml at37 C in nutrient broth (BBL) by ribosomally streptomycin-resistant strains of E. coli K-12 SA 136 with no added cations () or 1 mm added Ca2+ (U) and ofp. aeruginosa 28 with no added cations () or 1 mm added Ca2+ (). Ca2+ ions were added as CaC12- both streptomycin and gentamicin susceptibility and transport in spite of ATPase deficiency. AN12 is unable to energize various membrane functions anaerobically (14). Agents uncoupling oxidative phosphorylation significantly reduce streptomycin (2, 1) and gentamicin (9) uptake, but apparently do not impair electron transport. These observations, coupled with the results observed with heme mutants, demonstrate that electron transport alone is not adequate to cause streptomycin or gentamicin cellular entry, but acts as the major source of energy. Ubiquinones in E. coli represent a special case in that they do not seem to act only as carriers in electron transport and will be discussed below. The requirement of energy for aminoglycoside transport could be due to several possibilities. These aminoglycosides could fortuitously bind to one or more carriers involved in active transport of other solutes. This seems an unlikely probability for several reasons, including the apparent inability to use ATP hydrolysis for transport, a property not seen with most active transport systems (35). Additionally, we have been unable to demonstrate transport competition between streptomycin and a wide variety of structurally related compounds by using very high concentrations of both streptomycin and the related compounds. We are also unaware of mutations in any specific transport system resulting in a simultaneous change in streptomycin and gentamicin susceptibility in spite of an extensive search of the literature. Mutants selected for changed susceptibility to aminoglycosides have been shown to involve either the ribosome (6) or the membrane energization process (26) rather than any specific transport system. Although we cannot absolutely exclude a fortuitous carrier, we feel that it is highly unlikely for the above reasons. We think it more likely that aminoglycosides are transported by binding to a membrane complex involved in maintaining the high-energy membrane state used as an intermediate in coupling electron transport with active transport, ATP synthesis, and pyridine nucleotide transhydrogenase activities (for reviews, see references 14 and 35). The kinetics of either streptomycin or gentamicin accumulation show two energy-dependent phases of drug accumulation. The second of these is not detected until evidence of intracel- _ 3 P) z ANTimicRoB. AGzNTs CHEMOTHBR. z TIME IN MINUTES FIG. 7. Accumulation of streptomycin at 2.5 pgl ml by spheroplasts ofe. coli K-12 SA 136 in nutrient broth (BBL) containing 15% (wtlvol) sucrose at 37 C. E. coli K-12 SA 136, streptomycin sensitive, with no added cations () or with 1 mamca2+ (added as chloride 8alt) (A) or 1 mm Mg2+ (added as chloride salt) (); E. coli K-12 SA 136, streptomycin resistant, with no added cations (U) or with 1 mm mg2+ (as chloride) ().

11 VOL. 12, 1977 TABLz 4. Gentamicin or streptomycin accumulation by E. coli K-12 SA 136 in nutrient broth with or without added 1 mm Ca2+ after growth in nutrient broth containing 1 mm Ca2+a Antibiotic accumulation (ng/mg [dry wt]) Time Gentamicin (.75 &g/ Streptomycin (1 gg/ (min) ml) ml) +Ca2+ -Ca2+ +Ca2+ -Ca , , , , ,794 a Cells were grown in nutrient broth (NB) (BBL) containing a final concentration of 1 mm Ca2+ (as CaCl2) at 37C for 4 h, rapidly cooled to 4C, divided into two equal volumes, and centrifuged at 7, x g for 1 min. The supernatant was carefully removed, and cells were resuspended in equal volumes of either NB or NB plus 1 mm Ca2+ and vigorously resuspended for 2 min at 37 C. Labeled gentamicin or streptomycin was added, and uptake was measured. TABLE 5. Uptake of streptomycin by E. coli K-12 strains AB2154 and AN66 (ubid) a Strain Calcium concn Streptomycin accu- (mm) mulation (ng/mg per 15 min) AB2154 <.1 1,551 AB AN66 < AN a Both strains are stra mutants. Uptake was in aminoglycoside uptake medium supplemented with.1% yeast extract (BBL). Streptomycin concentration: 125 jug/ml. Zero-time values for AB2154 and AN66 without and with calcium, respectively, were 25, 24, 15, and 155 ng/mg. lular events is also seen, and it requires affinity of streptomycin or gentamicin for ribosomal structures. EDP-II accumulation began almost simultaneously with the first evidence of inhibition of protein synthesis, and this corresponded to the onset of eventual loss of cell viability (Fig. 8). EDP-II accumulation is not seen as part of streptomycin accumulation kinetics in stra mutants (1). It is also absent with gentamicin or streptomycin accumulation in bacteria containing R-factors (R+) that specify gentamicin- or streptomycin-modifying enzymes at streptomycin or gentamicin concentrations below those inhibiting growth of the strains (11) (L. E. Bryan and H. M. Van Den Elzen, Prog. Abstr. Intersci. Conf. Antimicrob. STREPTOMYCIN AND GENTAMICIN ACCUMULATION 173 Agents Chemother., 15th, Washington, D.C., Abstr. 168, 1975). Thus, clearly the interaction of streptomycin or gentamicin and ribosomes is necessary for this phase of accumulation. However, the preceding energy-dependent phase, EDP-I, is seen in stra mutants and in R+ bacteria and precedes detectable intracellular events. This phase (EDP-I), in our opinion, represents transport of the highly charged, lipidinsoluble aminoglycoside across the cytoplasmic membrane barrier to contact the ribosome. This conclusion is supported by several observations made with cations, including those of Beggs and Andrews (5) that cations will prevent streptomycin-induced lethality of M. snegmatis if cells are treated during the lag before induction of lethality (EDP-I in our terminology). Results reported in this study demonstrate that cations appear to nonspecifically inhibit streptomycin or gentamicin accumulation in three bacterial species and in spheroplasts. This suggests that the cell wall is less important as a site for cation antagonism than the cell membrane. Finally, the rapidity of cationic effects makes a structural effect less likely than a direct effect on aminoglycoside binding. We have also shown that ubiquinone-deficient mutants exhibit almost no cation antagonism I 4.' Z 3 :E v) TIME IN MINUTES FIG. 8. Simultaneous incorporation of ['4C]proline into acid-insoluble materials and accumulation of [3H]streptomycin at 1 pg of streptomycin per ml with time in E. coli K-12 SA 136 at 37 C and incorporation of [P4C]proline into acid-insoluble materials with time at 37 C in a control preparation of E. coli K-12 SA 136 not exposed to streptomycin. Symbols: (-) Proline incorporation with streptomycin added; (A) with no streptomycin; (@) streptomycin accumulation. Uptake was performed as given for streptomycin and in an identical fashion for [14C]proline (Amersham/Searle; 5 mci/mmol) except that the washing fluid was 1% trichloroacetic acid. b.l n E E. z 2.6

12 174 BRYAN AND VAN DEN ELZEN of streptomycin or gentamicin accumulation, supporting a role of these structures in the effects of cations. Cationic antagonism ofedp-i streptomycin uptake is also seen in stra mutants, and magnesium enhances streptomycin binding to the ribosome at similar cation concentrations (1, 13). Thus the ribosome is a very unlikely site for cationic antagonism. The summary of these data clearly demonstrates that cellular entry of streptomycin and gentamicin involves energy-dependent transport across the membrane to the ribosome, during which cations may antagonize transport, presumably by competing with these antibiotics for binding to some membrane component. This membrane component is not a specific cation transport system, since at least four cations are equally effective, including Ca2+, which is tansported out of bacteria (34). Monovalent ions are also effective, although at higher concentrations. From results reported here, from our previous studies, and from studies of other investigators, a model for the entry of aminoglycosides into bacteria can be proposed as follows. The initial interaction of the bacterial cell and streptomycin or gentamicin is on an electrostatic basis, and results in association of these agents with the cell wall and the external surface of the cytoplasmic membrane (energy-independent phase of accumulation). The binding site in the cell membrane involved in transport across the membrane must be lipid soluble, yet able to bind charged streptomycin or gentamicin molecules. We propose that this binding site is composed ofmembranous structures involved in aerobic energy transfer between electron transport and such fimctions as active transport, ATP synthesis, and pyridine nucleotide transhydrogenase. We feel that transport by this complex across the membrane represents EDP-I. These membranous structures must include respiratory quinones, a presumed protein defective in the mutant described by Plate (32), and probably some additional structure defective in mutant NR-7 (33). Respiratory quinones are particularly attractive as binding sites since they are mobile and lipid soluble, undergo oxidation-reduction (and thus have a polar component), and are present in large amounts in the membrane (3). In addition, our results clearly demonstrate that the ubiquinone deficiency in E. coli produces a significant decrease in streptomycin and gentamicin trans- ANTrmcROB. AGZNTS CHEMOTHBR. port and susceptibility. A survey of bacteria possessing respiratory quinones of either the ubiquinone or menaquinone type shows that these bacteria could be classified as generally susceptible to aminoglycosides (7, 22, 25, 28, 4, 41). On the other hand strictly anaerobic bacteria like Clostridium sporogenes and Clostridium perfringens are resistant to aminoglycosides and do not have detectable respiratory quinones. Organisms of intermediate susceptibility such as Streptococcus faecalis do carry out limited oxidative phosphorylation (21) and possess demethylmenaquinone (4). The microaerophilic organism Eikenella corrodens (24) is sensitive to gentamicin and possesses ubiquinone (22). Taber and Halfenger have also shown increased aminoglycoside resistance in a menaquinone-deficient mutant of Bacillus subtilis (36). Thus there is very good general correlation between possession of respiratory quinones and aminoglycoside susceptibility. Cox and Gibson (14) have proposed that ubiquinones in E. coli act as more than simple carriers in electron transport and appear to have a more general role in energy transfer that fits well with our proposal. In electron transport, respiratory quinones are generally placed to precede the cytochromes, although Cox and Gibson also propose an association following cytochrome b,. The inhibitor 2-n-heptyl4.hydroxyquinoline-N-oxide (HQNO) has been long recognized as an effective antagonist of streptomycin (19). This inhibitor has been shown by Cox and Gibson to cause E. coli membranes to act as if they were ubiquinone deficient and thus seems to act on respiratory quinones or on closely associated structures (14). We have also shown that HQNO will partially inhibit streptomycin and gentamicin uptake in E. coli spheroplasts of normal but not ubiquinone-deficient cells (whole cells are not susceptible to HQNO) (L. E. Bryan, unpublished data). These results again support a central role for respiratory quinones but do not necessarily mean that quinones act directly in the binding of streptomycin, although the saturation of the ubid (AN66) mutant at a reduced streptomycin concentration strongly supports this role. Quinones could act indirectly to establish a proper receptor state in an associated structure. This structure could be another part of the aerobic membrane energization complex or, much less likely, a transport protein for some other compound. Streptomycin or gentamicin molecules are proposed to cross the membrane surface to sites of higher affinity, which we suggest are membrane-associated ribosomes functioning in protein synthesis. When binding to ribosomal components participating in a normal ribosomal cycle occurs, the next phase of accumulation kinetics, EDP-II, is seen (see below). The affinity of streptomycin or gentamicin for the membranous transport site is low as far as we can determine. Attempts to determine a "ikm" have yielded values in excess of 1 mm,

13 VOL. V9STRENOMYCIN 12, 1977 AND GENTAMICIN ACCUMULATION 175 and we have observed that EDP-I accumulation has a very low rate (1). It is the lowest of the three rates observed in accumulation kinetics and represents the rate-limiting step in aminoglycoside susceptibility (1). In keeping with a low-affinity system, the rate of EDP-I accumulation is directly dependent on the external concentration ofthe drug. Although affinity is low, as would be expected of a fortuitous binding such as that proposed, the number of binding sites is high (1), s only a small fraction of potential sites are occupied at any one time. In summary, we propose that EDP-I represents streptomycin or gentamicin transport across the membrane using a low-affinity, rate-limiting (for susceptibility) membrane energization complex involving respiratory quinones. Streptomycin or gentamicin is subsequently discharged to higher-affinity ribosomal binding sites, which is associated with the onset of EDP-ll. This proposal is consistent with the observation that EDP-ll does not appear in stra mutants where ribosomal affinity is decreased for streptomycin (1) or, conversely, where streptomycin affinity is reduced for the ribosome (11) (Bryan and Van Den Elzen, Prog. Abstr. Intersci. Conf. Antimicrob. Agents Chemother., 15th, Washington, D.C., Abstr. 168, 1975). The latter situation occurs in R+ bacteria that specify enzymatic streptomycin phosphorylation or adenylylation. Both of these processes reduce affinity for the ribosome (6). The observation that R+ bacteria modifying gentamicin or streptomycin in addition to the markedly reduced accumulation of these drugs associated with the absence of EDP-II also fail to accumulate modified drug in the extracellular medium (Bryan and Van Den Elzen, Progr. Abstr. Intersci. Conf. Antimicrob. Agents Chemother., 15th, Washington, D.C., Abstr. 168, 1975; P. Dickie and L. E. Bryan, unpublished data) suggests that these bacteria are resistant because of inactivation of a small number of critically important molecules of streptomycin or gentamicin. Evidence suggests that modifying enzymes are located in or on the cellular membrane (16) and seem ideally located to modify such critically important molecules occupying membrane transport sites. Under conditions where the rate of EDP-I was less than that of enzymatic inactivation, cells would be resistant since those molecules being transported would also be inactivated. Such modified molecules have reduced ribosomal affinity and would be unlikely to bind to the ribosome, thus resulting in no EDP-II being observed. Incidentally, a prediction of this model is that R+ bacteria that modify gentamicin but not streptomycin would, when exposed to gentamicin, still be susceptible to streptomycin, since the rate of EDP-I is concentration dependent and since streptomycin at appropriate concentrations would be transported (by newly created binding sites as part of energy transduction in metabolizing cells) but not modified. This observation has been made repeatedly in our laboratory. Several agents that inhibit protein synthesis are known to be antagonistic to streptomycin (1, 19, 42), including chloramphenicol, erythromycin, and tetracycline. Ifthese agents prevent the normal ribo;omal cycle, EDP-II, according to our model, would be reduced or eliminated. Although this observation is so (1, 19, 23), chloramphenicol also reduces EDP-I for unknown reasons. However, chloramphenicol is lipid soluble and could reduce the association of streptomycin or gentamicin with the proposed aerobic energization complex by its own interaction with membrane lipids. Chloramphenicol is known to produce respiratory disturbances in mitochondria (31). EDP-II of streptomycin or gentamicin accumulation is characterized by a significant increase in rate of their transport and is correlated with the onset of inhibition of protein synthesis. We suggest that the interaction of membrane-bound aminoglycosides with ribosomal binding sites causes a change in the state of membrane energy analogous to that seen with mutant AN12 and which facilitates cellular entry of aminoglycosides. This could result from a functional or architectural interference with energy transduction so that the normal rate of utilization of the membrane proton gradient would be reduced. Under these conditions, more binding sites (as reduced respiratory quinones?) would be available for aminoglycoside binding. Alternatively, internalization of the aminoglycoside when bound to a ribosomal site could enhance the protonmotive force due to the electrochemical characteristics of aminoglycosides. An increased rate of streptomycin and gentamicin entry has been shown here to occur with AN12 and apparently in a mutant described by Turnock et al. (38, 39), so this seems a feasible proposal. Mutants like AN12 and Turnock's mutant should be hypersensitive to aminoglycosides as a group if our model is correct, which is definitely true for AN12 at least. Mutants like this are probably hypersensitive because the binding complex is still energized, but this energy (as a proton gradient?) is not transferred to usual functions and may maintain a larger fraction of binding sites available for the aminoglycoside at a specific drug concentration. Streptomycin, in particular, has recognized effects on membrane fimction such as decreased oxygen utilization (18), especially with succinate as a substrate

14 176 BRYAN AND VAN DEN ELZEN (F. L. Jackson, unpublished data) (uncoupled bacteria utilize succinate poorly as an energy source), and loss of permeability control for amino acids and nucleotides (1, 18). These rather profound changes occur once EDP-ll is underway and suggest that severe derangement of membrane energy transfer may occur. This aspect of our proposal is currently under study in our laboratory. In our view, the accumulated data support direct binding of aminoglycosides by a membranous aerobic energization complex. However, another interpretation of these data might be that aminoglycoside transport rates reflect the cross-mnembrane protonmotive force. NR-7 especially, and perhaps ubid, may be leaky for protons and therefore may have a reduced protonmotive force. AN12 is not proton leaky and, in view of its uncoupling could conceivably have an enhanced proton-motive force. Therefore, these mutants could reflect the state of membrane energy, and changes in transport of aminoglycosides, like changes in transport of some other solutes, would occur because of the altered driving force supplied to membrane transport carriers. However, in the absence of demonstrable conventional carriers and in the presence of saturation of ubid mutants at decreased streptomycin concentrations, we feel that the aminoglycoside carrier is the proposed membrane complex involving respiratory quinones. It is also of interest to note that a change in susceptibility to bacteriocins may be simultaneously associated with a change in susceptibility to several aminoglycosides. Two good examples of this correlation are seen with the mutant of E. coli K-12 described by Plate (32) and an aeruginocin-intolerant mutant of P. aeruginosa described by Mills and Holloway (29). In the latter instance, aeruginocin tolerance is associated with hypersensitivity to many aminoglycosides. Colicin K is considered to act on structures involved in membrane energization and the aeruginocin used by Mills and Holloway has a similar mechanism of action. The preceding model serves to explain many previously uninterpretable observations made with aminoglycosides and is consistent with the data associated with streptomycin or gentamicin transport. Like all scientific models, it is open to further testing and is presented as a basis for further study. ACKNOWLEDGMENTS This research was supported by Medical Research Council of Canada grant MT435. We gratefully acknowledge the technical assistance of Neil Harris and Richard Sherburne and assistance from Bunnie Rakowski in preparing the manu6cript. We also ANTimICROB. AGENTS CHEMOTHICR. thank Peter Dickie and F. L. Jackson for allowing us to quote unpublished data and for many useful discussions of the results presented. LITERATURE CITED 1. Anand, N., and B. D. Davis Effect of streptomycin on Escherichia coli: damage by streptomycin to the cell membrane of Escherichia coli. Nature (London) 185: Andry, K., and R. C. Bockrath Dihydrostreptomycin accumulation in E. coli. Nature (London) 251: Bauer, A. W., W. M. Kirby, J. C. Sherris, and M. Turck Antibiotic susceptibility testing by a standardized single disc method. Am. J. Clin. Pathol. 45: Baum, R. H., and M. I. Dolin Isolation of a new naphthoquinone from Streptococcus faecalis 1Cl. J. Biol. Chem. 238:PC419-PC Beggs, W. H., and F. A. Andrews Inhibition of dihydrostreptomycin binding to Mycobacterium smegmatis by monovalent and divalent cation salts. Antimicrob. Agents Chemother. 9: Benveniste, R., and J. Davies Mechanisms of antibiotic resistance in bacteria. Annu. Rev. Biochem. 42: Bishop, D. H. L., K. P. Pandya, and H. K. King Ubiquinone and vitamin K in bacteria. Biochem. J. 83: Bryan, L. E., R. Haraphongse, and H. M. Van Den Elzen Gentamicin resistance in clinical-isolates of Pseudomonas aeruginosa associated with diminished gentamicin accumulation and no detectable enzymatic modification. J. Antibiot. 29: Bryan, L. E., and H. M. Van Den Elzen Gentamicin accumulation by sensitive strains of Escherichia coli and Pseudomonas aeruginosa. J. Antibiot. 28: Bryan, L. E., and H. M. Van Den Elzen Streptomycin accumulation in susceptible and resistant strains of Escherichia coli and Pseudomonas aeruginosa. Antimicrob. Agents Chemother. 9: Bryan, L. E., H. M. Van Den Elzen, and M. S. Shahrabadi The relationship of aminoglycoside permeability to streptomycin and gentamicin susceptibility of Pseudomonas aeruginosa, p In S. Mitsohashi and H. Hashimoto (ed.), Microbial drug resistance. University of Tokyo Press, Tokyo. 12. Butlin, J. D., G. B. Cox, and F. Gibson Ozidative phosphorylation in Escherichia coli K12. Biochem. J. 124: Chang, F. N., and J. G. Flaks Binding of dihydrostreptomycin to Escherichia coli ribosomes: characteristics and equilibrium of the reaction. Antimicrob. Agents Chemother. 2: Cox, G. B., and F. Gibson Studies on electron transport and energy-linked reactions using mutants of Escherichia coli. Biochim. Biophys. Acta 346: Cox, G. B., I. G. Young, L. M. McCann, and F. Gibson Biosynthesis of ubiquinone in Escherichia coli K12: location of genes affecting the metabolism of 3- octaprenyl4-hydroxybenzoic acid and 2-octaprenylphenol. J. Bacteriol. 99: Davies, J. E., and R. E. Benveniste Enzymes that inactivate antibiotics in transit to their targets. Ann. N. Y. Acad. Sci. 235: Devor, K. A., H. U. Schairer, D. Renz, and P. Overath Active transport of 8-galactosides by a mutant of Escherichia coli defective in heme synthesis. Eur. J. Biochem. 45: Dubin, D. T., R. Hancock, and B. D. Davies The sequence of some effects of streptomycin in Escherichia coli. Biochim. Biophys. Acta 74:

15 VOL. 12, 1977 SDRENPOMYCIN AND GENTAMICIN ACCUMULATION Hancock, R Uptake of '4C-streptomycin by Bacillus megaterium. J. Gen. Microbiol. 28: Hancock, R Uptake of "4C-streptomycin by some micro-organisms and its relation to their streptomycin sensitivity. J. Gen. Microbiol. 28: Harold, F. M Conservation and transformation of energy by bacterial membranes. Bacteriol. Rev. 36: Hollander, R., and W. Mannheim Characterization of hemophilic and related bacteria by their respiratory quinones and cytochromes. Int. J. Syst. Bacteriol. 25: Hurwitz, C., and C. L. Rosano Accumulation of label from C'4-streptomycin by Escherichia coli. J. Bacteriol. 83: Jackson, F. L., and Y. E. Goodman Transfer of the facultatively anaerobic organism Bacteroides corrodens Eiken to a new genus, Eikenella. Int. J. Syst. Bacteriol. 22: Jacobsen, B. K., and H. Dam Vitamin K in bacteria. Biochim. Biophys. Acta 4: Kanner, B. I., and D. L. Gutnick Use of neomycin in the isolation of mutants blocked in energy conservation in Escherichia coli. J. Bacteriol. 111: Kogut, M., J. W. Lightbown, and P. Isaacson Streptomycin action and anaerobiosis. J. Gen. Microbiol. 39: Lester, R. L., and F. L. Crane The natural occurrence of coenzyme Q and related compounds. J. Biol. Chem. 234: Mills, B. J., and B. W. Holloway Mutants of Pseudomonas aeruginosa that show specific hypersensitivity to aminoglycosides. Antimicrob. Agents Chemother. 1: Newton, N. A., G. B. Cox, and F. Gibson Function of ubiquinone in Escherichia coli: a mutant strain forming a low level of ubiquinone. J. Bacteriol. 19: Pestka, S Chloramphenicol, p In J. W. Corcoran and F. E. Hahn (ed.), Antibiotics III: mechanism of action of antimicrobial and antitumor drugs. Springer-Verlag, New York. 32. Plate, C. A Mutant of Escherichia coli defective in response to colicin K and in active transport. J. Bacteriol. 125: Rosen, B. P Restoration of active transport in an Mg2+-adenosine triphosphatase-deficient mutant of Eacherichia coli. J. Bacteriol. 116: Rosen, B. P., and J. S. McClees Active transport of calcium in inverted membrane vesicles of Escherichia coli. Proc. Natl. Acad. Sci. U.S.A. 71: Simoni, R. D., and P. W. Postma The energetics of bacterial active transport. Annu. Rev. Biochem. 44: Taber, H., and G. M. Halfenger Multiple-aminoglycoside-resistant mutants of Bacillus subtilis deficient in accumulation of kanamycin. Antimicrob. Agents Chemother. 9: Tseng, J. T., L. E. Bryan, and H. M. Van Den Elzen Mechanisms and spectrum of streptomycin resistance in a natural population of Pseudomonas aeruginosa. Antimicrob. Agents Chemother. 2: Turnock, G The action of streptomycin in a mutant of Escherichia coli with increased sensitivity to the antibiotic. Biochem. J. 118: Turnock, G., S. K. Erickson, B. A. A. Ackrell, and B. Birch A mutant of Escherichia coli with a defect in energy metabolism. J. Gen. Microbiol. 7: Whistance, G. R., J. F. Dillon, and D. R. Threlfall The nature, intergeneric distribution and biosynthesis of isoprenoid quinones and phenol in Gramnegative bacteria. Biochem. J. 111: White, D. C The function of 2-demethyl vitamin K2 in the electron transport system of Hemophilus parainfluenzae. J. Biol. Chem. 24: Yamaki, H., and N. Tanaka Effects of protein synthesis inhibitors on the lethal action of kanamycin and streptomycin. J. Antibiot. 16: