The Adenylosuccinate Synthetase from the Hyperthermophilic ArchaeonPyrococcusSpecies Displays Unusual Structural Features

The first example of a hyperthermophilic adenylosuccinate synthetase is reported, which is an enzyme that must maintain its folded structure at temperatures as high as 102°C. The amino acid sequence of this key enzyme has been determined after cloning and sequencing thepurA-like gene from the archaealPyrococcussp. strain ST700. The corresponding protein displays two unexpected features: (1) it is 21% shorter than the homologous mesophilic enzymes and this shortening corresponds to the loss of two α-helices and three β-strands present in theEscherichia colienzyme; (2) surprisingly, the archaeal adenylosuccinate synthetase has a significant number of substitutions in residues that are conserved in all other homologous enzymes from bacteria to man. InE. coli, the conserved residues have been described as essential for catalytic activity and/or for maintaining the folded structure of the homodimer. Despite these drastic differences, thepurA-like archaeal gene seems to be normally expressed and its product functionsin vivoin bacteria, since it complemented anE. coli purAauxotroph. The archaeal adenylosuccinate synthetase appears to be a good example of abona fideorthologous protein. Reconstruction of phylogenetic trees showed that the archaeal gene is equally distantly related to both eukaryotes and bacteria, independently of the numerous substitutions observed at critical positions.     

Overexpression and Purification of Asparagine Synthetase from Escherichia coli

Overexpression of the asnA gene from Escherichia coli K-12 coding for asparagine synthetase (EC 6.3.1.1) was achieved with a plasmid, pUNAd37, a derivative of pUCI8, in E. coli. The plasmid was constructed by optimizing a DNA sequence between the promoter and the ribosome binding region. The enzyme, comprising ca. 15%, of the total soluble protein in the E. coli cell, was readily purified to apparent homogeneity by DEAE-Cellulofine and Blue-Cellulofine column chromatographies. The amino-terminal sequence, amino acid composition, and molecular weight of the purified protein agreed with the predicted values based on the DNA sequence of the gene. Furthermore the native molecular weight measured by gel filtration confirmed that asparagine synthetase exists as a dimer of identical subunits.     

Prototrophic Regulatory Mutants of Adenylosuccinate Synthetase in Yeast

MUTANTS of the genes ade12 and ade13 in Saccharomyces cerevisiae require adenine, which suggests that they are defective in the conversion of inosine monophosphate (IMP) to adenosine monophosphate (AMP). Two enzymes are required for this conversion: adenylosuccinate lyase and adenylosuccinate synthetase. The former has been shown to be absent in mutants of ade13 (ref. 1) so that one can infer that ade12 specifies the latter. Mutants of ade12 have been isolated on the basis of an adenine insensitive pigment accumulation by strains carrying ade1 or ade2 (refs. 1 and 2). They seem to synthesize purines constitutively and excrete hypoxanthine¹. These observations suggest that in addition to its catalytic function adenylosuccinate synthetase has an important function in the regulation of purine synthesis.     

Regulation of S-Adenosylmethionine Synthetase in Escherichia coli

Addition of methionine to the growth medium of Escherichia coli K-12 leads to a reduction in the specific activity of S-adenosylmethionine (SAM) synthetase. Thus the enzyme appears to be repressible rather than inducible. Mutant strains (probably metJ(-)) are constitutive for SAM synthetase as well as for the methionine biosynthetic enzymes, suggesting that the regulatory systems for these enzymes have at least some elements in common. Cells grown to stationary phase in complete medium, which have low specific activities of the enzymes, were routinely used for derepression experiments. The lag in growth and derepression when these cells are incubated in minimal medium is shortened by threonine. Ethionine, norleucine, and alpha-methylmethionine are poor substrates or nonsubstrates for SAM synthetase and are ineffective repressors. Selenomethionine, a better substrate for SAM synthetase than methionine, is also slightly more effective at repression than methionine. Although SAM is considered to be a likely candidate for the corepressor in the control of the methionine biosynthetic enzymes, addition of SAM to the growth medium does not cause repression. Measurement of SAM uptake shows that too little is taken into the cells to have a significant effect, even if it were active in the control system.     

Overexpression of Glutathione Synthetase in Escherichia coli

Kohamaic acid A is a potent DNA polymerase inhibitor isolated from the Okinawan marine sponge Ircinia sp. A series of structurally simplified analogs of kohamaic acid A were synthesized with the aim of evaluating structure-activity relationships.     

N端的改变对大肠杆菌精氨酰—tRNA合成酶的影响

得到了缺失Asn2 的大肠杆菌 (E .coli)精氨酰 tRNA合成酶 (ArgRS)的变种和在其N端添加酵母ArgRS的N端 2 3个氨基酸残基的嵌合变种。它们的基因在大肠杆菌中表达时 ,可能由于蛋白质误折叠 ,大部分产生了包涵体。与天然酶相比 ,缺失变种保留了全部的氨基酸活化活力 ,但氨基酰化活力下降了 2 6 % ;嵌合变种的以上两种活力下降了 90 %以上 ,不能氨基酰化酵母tRNAArg。缺失Asn2 和Ile3 的变种在E .coli中虽被表达 ,但不稳定。与天然酶相比 ,嵌合变种的荧光光谱的最大发射波长向长波移动 ,强度减小。表明变种酶的构象和天然酶不同 ,色氨酸更暴露。用远紫外CD光谱预测变种酶的二级结构表明 ,嵌合酶的α螺旋更少 ,β折叠更多 ,无规卷曲稍多。E .coliArgRS的N端结构域对活力和正确折叠是重要的     

Determinants of L-aspartate and IMP recognition in Escherichia coli adenylosuccinate synthetase.

Adenylosuccinate synthetase governs the first committed step in the de novo synthesis of AMP. Mutations of conserved residues in the synthetase from Escherichia coli reveal significant roles for Val(273) and Thr(300) in the recognition of l-aspartate, even though these residues do not or cannot hydrogen bond with the substrate. The mutation of Thr(300) to alanine increases the K(m) for l-aspartate by 30-fold. In contrast, its mutation to valine causes no more than a 4-fold increase in the K(m) for l-aspartate, while increasing k(cat) by 3-fold. Mutations of Val(273) to alanine, threonine, or asparagine increase the K(m) for l-aspartate from 15- to 40-fold, and concomitantly decrease the K(i) for dicarboxylate analogues of l-aspartate by up to 40-fold. The above perturbations are comparable with those resulting from the elimination of a hydrogen bond between the enzyme and substrate: alanine mutations of Thr(128) and Thr(129) increase the K(m) for IMP by up to 30-fold and the alanine mutation of Thr(301) abolishes catalysis supported by l-aspartate, but has no effect on catalysis supported by hydroxylamine. Structure-based mechanisms, by which the above residues influence substrate recognition, are presented.     

大肠杆菌亮氨酰－tRNA合成酶基因的克隆和鉴定

本文根据遗传互补原理,利用大肠杆菌亮氨酰－ｔＲＮＡ合成酶基因（ｌｅｎＳ）的温度敏感突变株ＫＬ２３１,从大肠杆菌基因文库一克隆（λ１５Ｄ７）中筛选出带完整ｌｅｕＳ基因的ＤＮＡ片段。该片段长度为３．２ｋｂ。对此片段做了１４种限制性内切酶图谱分析和部分ＤＮＡ序列鉴定,并与文献报道的ｌｅｎＳ基因序列进行了比较。发现在编码区和３’端非编码区各有一对碱基发生了转换。另外在３’端非编码区有一对碱基缺失。编码区的碱基对转换导致编码的氨基酸由组氨酸变成了精氨酸。带有ｌｅｎＳ基因的质粒（ｐＬｅｕＳ９１）转入大肠杆菌ＴＧＩ菌株中,测得转化子的亮氨酰－ｔＲＮＡ合成酶比活力是ＴＧ１菌株的１０倍以上。     

Escherichia coli Mutants Overproducing Phenylalanyl- and Threonyl-tRNA Synthetase

The structural genes for threonyl-tRNA synthetase (ThrRS) and phenylalanyl-tRNA synthetase (PheRS) are closely linked on the Escherichia coli chromosome. To study whether these enzymes share a common regulatory element, we have investigated their synthesis in mutants which were selected for overproduction of either ThrRS or PheRS. It was found that mutants isolated previously for overproduction of ThrRS as strains resistant to the antibiotic borrelidin (strains Bor Res 3 and Bor Res 15) did not show an elevated level of PheRS. PheRS-overproducing strains were then isolated as revertants of strains with structurally altered enzymes. Strain S1 is a temperature-resistant derivative of a temperature-sensitive PheRS mutant, and strain G118 is a prototrophic derivative of a PheRS mutant which shows phenylalanine auxotrophy as a consequence of an altered K(m) of this enzyme for the amino acid. In both kinds of revertants, S1 and G118, the concentration of PheRS and ThrRS was increased by factors of about 2.5 and 1.8, respectively, whereas the level of other aminoacyl-tRNA synthetases was not affected by the mutations. Genetic studies showed that the simultaneous overproduction of PheRS and ThrRS in revertants G118 and S1 is based upon gene amplification, since this property was easily lost after growing the cells in the absence of the selective stimulus, and since this loss could be prevented by the presence of the recA allele. By similar criteria, the four- and eightfold overproduction of ThrRS in strains Bor Res 3 and Bor Res 15, respectively, was very stable genetically, indicating that it is caused by a mutational event other than gene amplification. From these results, we conclude that the concomitant increase of PheRS and ThrRS in strains G118 and S1 is an expression of gene duplication and not of a joint regulation of these two aminoacyl-tRNA synthetases. This conclusion is further supported by the result that, in mutant G118 as well as in its parental strain G1, growth in minimal medium lacking phenylalanine led to an additional twofold increase of their PheRS concentration. This increase was restricted to the PheRS, since the level of other aminoacyl-tRNA synthetases, including the ThrRS, stayed unchanged.     

Asymmetric Amino Acid Activation by Class II Histidyl-tRNA Synthetase from Escherichia coli

Aminoacyl-tRNA synthetases (ARSs) join amino acids to their cognate tRNAs to initiate protein synthesis. Class II ARS possess a unique catalytic domain fold, possess active site signature sequences, and are dimers or tetramers. The dimeric class I enzymes, notably TyrRS, exhibit half-of-sites reactivity, but its mechanistic basis is unclear. In class II histidyl-tRNA synthetase (HisRS), amino acid activation occurs at different rates in the two active sites when tRNA is absent, but half-of-sites reactivity has not been observed. To investigate the mechanistic basis of the asymmetry, and explore the relationship between adenylate formation and conformational events in HisRS, a fluorescently labeled version of the enzyme was developed by conjugating 7-diethylamino-3-((((2-maleimidyl)ethyl)amino)carbonyl)coumarin (MDCC) to a cysteine introduced at residue 212, located in the insertion domain. The binding of the substrates histidine, ATP, and 5′-O-[N-(l-histidyl)sulfamoyl]adenosine to MDCC-HisRS produced fluorescence quenches on the order of 6–15%, allowing equilibrium dissociation constants to be measured. The rates of adenylate formation measured by rapid quench and domain closure as measured by stopped-flow fluorescence were similar and asymmetric with respect to the two active sites of the dimer, indicating that conformational change may be rate-limiting for product formation. Fluorescence resonance energy transfer experiments employing differential labeling of the two monomers in the dimer suggested that rigid body rotation of the insertion domain accompanies adenylate formation. The results support an alternating site model for catalysis in HisRS that may prove to be common to other class II aminoacyl-tRNA synthetases.The aminoacyl-tRNA synthetases (ARSs)² comprise two distinct classes of enzymes, all of which catalyze a two-step reaction to generate aminoacyl-tRNA for protein synthesis (1, 2) (Reactions 1 and 2). During the first of two partial reactions in aminoacylation, the cognate amino acid is condensed with ATP to form an aminoacyl-adenylate. This half reaction proceeds by an associative mechanism in which the stereochemistry of the α-phosphate undergoes inversion (3). The adenylate then undergoes a subsequent attack by the cognate tRNA, with the amino acid undergoing transfer to the 3′-terminal adenosine. Aminoacyl transfer requires the activation of 2′ or 3′ of the terminal hydroxyl, and its rate may be accelerated by a number of different mechanisms, including proton transfer to the adenylate, and proton shuttling to the 2′-OH and then to neighboring active site residues (4, 5). Many ARSs can activate their cognate amino acids in the absence of tRNA, allowing the two partial reactions to be studied individually. Notably, there are significant gaps in our understanding of how the adenylation and aminoacyl transfer half reactions are integrated into the overall reaction schemes of ARSs.Class I and class II enzymes can be broadly distinguished by their oligomeric structure. The former are generally monomeric, whereas the latter are typically dimeric or tetrameric (6). Notable exceptions to this pattern are the class Ic tyrosyl- and tryptophanyl-tRNA synthetases, both of which form obligatory dimers (7, 8). Both have been described as possessing half-of-sites reactivity (9, 10), but the picture is more complex. Consistent with half-of-sites reactivity, TyrRS binds one mole of tyrosine per dimer and retains a single mole of adenylate per mole of dimers when the E·Ade complex is purified away from unreacted substrates by size-exclusion chromatography (11). However, the steady-state kinetics of TyrRS show no evidence of cooperativity, the second binding site becomes accessible to substrates when the first site is occupied by adenylate, and TyrRS clearly binds 2 mol of tRNA in the crystal (7, 12).On the basis of these and other observations involving the rate of hydrolysis of the on-enzyme adenylate, Fersht (13) proposed that the second site of TyrRS possesses weak catalytic activity and that TyrRS is asymmetric in solution. The impact of this potential asymmetry in the activation reaction on the complete TyrRS catalytic cycle remains to be explored. TrpRS also exhibits half-of-sites reactivity, and a detailed analysis of the aminoacyl transfer reaction by pre-steady state kinetics proposed both random and ordered versions of alternating site catalysis as models of the enzyme (14). In the class II ARSs, the tetrameric SepRS represents the single example where half-of-sites reactivity has been demonstrated experimentally (15).Despite this apparent class distinction, recent work on HisRS, a class IIa ARS that is well characterized with respect to structure (16–19), tRNA recognition, and reaction kinetics (4, 20), highlighted several functional attributes that are reminiscent of class I TyrRS. Like TyrRS, HisRS retains only 1 mol of adenylate per dimer when subjected to size-exclusion chromatography (4). A detailed pre-steady-state analysis of mutants of tRNA^His or HisRS compromised with respect to tRNA identity suggested that, in the complete aminoacylation reaction, formation of aminoacyl adenylate in the second active site is contingent upon a productive aminoacyl transfer reaction in the first (20). These and other data led to the proposal of an alternating site model for HisRS (20) that is analogous to the “flip flop” catalysis suggested for class II PheRS (21, 22) and class Ic TrpRS (14). This raises the possibility that the catalytic cycles of dimeric class II enzymes and dimeric class Ic enzymes share some common feature.Alternating catalysis requires a mechanism for coupling events between active sites, presumably through conformational changes propagated between these active sites. To investigate these events, a version of HisRS was developed that featured the site-specific incorporation of extrinsic environmentally sensitive fluorescent probes, allowing the adenylation reaction to be followed by stopped-flow fluorometry. Comparison of the kinetics of substrate-induced fluorescence changes to the kinetics of product formation determined by rapid quench suggests that adenylation rates are asymmetric with respect to the two active sites of the dimer, and that conformational changes linked to the insertion domain may be rate-limiting for product formation. The implications of these results for a previous model (20) of alternating site catalysis in HisRS are discussed.     

Unusual Valyl-Transfer Ribonucleic Acid Synthetase Mutant of Escherichia coli

Escherichia coli strain NP2907 was isolated as a spontaneous mutant of strain NP29, which possesses a thermolabile valyl-transfer ribonucleic acid (tRNA) synthetase. The valyl-tRNA synthetase of the new mutant, unlike that of its immediate parent, retains enzymatic activity in vitro but differs from the wild-type enzyme in stability and apparent K(m) for adenosine triphosphate. The new mutant locus, valS-102, cotransduces with pyrB at the same frequency as does the parental locus, valS-1. Cultures of strain NP29 cease growth immediately in any medium when shifted from 30 to 40 C. The new mutant grows normally at 30 C, and upon a shift to 40 C growth quickly accelerates exactly as for normal cells. Exponential growth, however, cannot be sustained at 40 C. At a point characteristic for each medium, growth becomes linear with time. This transition occurs almost immediately in rich media and after 1.5 generations in glucose minimal medium. Net synthesis of valyl-tRNA synthetase ceases in the new mutant as soon as the temperature is raised to 40 C, irrespective of the growth medium. We conclude that it is the amount of valyl-tRNA synthetase activity that limits the rate of growth in the linear phase at 40 C. This property of the mutant makes it possible to evaluate the in vivo efficiency of this enzyme at different growth rates and thereby to determine the concentration that is necessary for a given rate of protein synthesis. The results of our measurements indicate that cells of E. coli growing in minimal medium normally possess a functional excess of valyl-tRNA synthetase with respect to protein synthesis and to repression of threonine deaminase.     

Pseudouridylate Synthetase of Escherichia coli: a Catabolite-Repressible Enzyme

The growth on pseudouridine of two pyrimidine auxotrophs of Escherichia coli (Bu(-) and W63-86) was markedly enhanced when glycerol replaced glucose as a carbon source or when adenosine 3':5'-cyclic monophosphoric acid was added to medium containing glucose. These results indicated that an enzyme catalyzing a reaction in the pathway of pseudouridine conversion to uracil was sensitive to catabolite repression. The following pathway is proposed for pseudouridine utilization: [Formula: see text] [Formula: see text] Pseudouridylate synthetase was sensitive to catabolite repression in strains Bu(-) and W63-86. In contrast, strains B5RU and W5RU, mutants of Bu(-) and W63-86 which were selected for their ability to grow rapidly on pseudouridine in the presence of glucose, had high levels of pseudouridylate synthetase in the presence of glucose. In the case of B5RU but not W5RU, synthetase activity was greater in cells grown on glycerol or on glucose plus adenosine 3':5-cyclic monophosphoric acid than on glucose.     

Adenosylcobalamin-dependent methylmalonyl-CoA mutase from Propionibacterium shermanii. Active holoenzyme produced from Escherichia coli.

The linked structural genes coding for both subunits of adenosylcobalamin-dependent methylmalonyl-CoA mutase from the Gram-positive bacterium Propionibacterium shermanii have been altered by site-directed mutagenesis and placed under the control of an inducible phage-T7-specific plasmid promoter in Escherichia coli. Conditions have been found under which both alpha- and beta-subunits are produced in soluble form, in near 1:1 ratio, and assemble to form apo-mutase totalling about 5% of the total cellular protein. Methylmalonyl-CoA mutase purified from these cells could be readily converted into the holoenzyme by addition of adenosylcobalamin. The active holoenzyme apparently crystallizes in the same space group as an inactive corrinoid-containing form of the enzyme obtained previously.     

Active Dimeric Form of Inorganic Pyrophosphatase from Escherichia coli

A dimeric form can be obtained from native hexameric Escherichia coli inorganic pyrophosphatase (E-PPase) by destroying the hydrophobic intersubunit contacts, and it has been shown earlier to consist of the subunits of different trimers. The present paper is devoted to the kinetic characterization of such a "double-decked" dimer obtained by the dissociation of either the native enzyme or the mutant variant Glu145Gln. The dimeric form of the native inorganic pyrophosphatase was shown to retain high catalytic efficiency that is in sharp contrast to the dimers obtained as a result of the mutations at the intertrimeric interface. The dimeric enzymes described in the present paper, however, have lost the regulatory properties, in contrast to the hexameric and trimeric forms of the enzyme.     

Asp338 controls hydride transfer in Escherichia coli IMP dehydrogenase

IMP dehydrogenase (IMPDH) catalyzes the oxidation of IMP to XMP with the concomitant reduction of NAD(+). This reaction involves the formation of a covalent adduct with an active site Cys. This intermediate, E-XMP, hydrolyzes to produce XMP. The mutation of Asp338 to Ala severely impairs the activity of Escherichia coli IMPDH, decreasing the value of k(cat) by 650-fold. No (D)V(m) or (D)V/K(m) isotope effects are observed when 2-(2)H-IMP is the substrate for wild-type IMPDH. Values of (D)V(m) = 2.6 and (D)V/K(m) (IMP) = 3.4 are observed for Asp338Ala. Moreover, while a burst of NADH production is observed for wild-type IMPDH, no burst is observed for Asp338Ala. These observations indicate that the mutation has decreased the rate of hydride transfer by at least 5 x 10(3)-fold. In contrast, k(cat) for the hydrolysis of 2-chloroinosine-5'-monophosphate is decreased by only 8-fold. In addition, the rate constant for inactivation by 6-chloropurine riboside 5'-monophosphate is increased by 3-fold. These observations suggest that the mutation has little effect on the nucleophilicity of the active site Cys residue. These results are consistent with a recent crystal structure that shows a hydrogen bonding network between Asp338, the 2'-OH of IMP, and the amide group of NAD(+) [Colby, T. D., Vanderveen, K., Strickler, M. D., Markham, G. D., and Goldstein, B. M. (1999) Proc. Natl. Acad. Sci. U.S.A. 96, 3531-3536].     

青霉素酰化酶活性中心的定点突变

对Ｅ．ｃｏｌｅ ＡＴＣＣ １１１０５的青霉素酰化酶（ｐｅｎｉｃｉｌｌｉｎ Ｇ ａｃｙｌａｓｅ,ＰＧＡ）活性中心的Ｓｅｒ２９０分别进行了定点突变和和化学修饰,将活性中心的Ｓｅｒ２９０改变成Ｃｙｓ或Ｓｅｃｙｓ,ＰＧＡ水解活力均大为下降,但仍保留部分活性。其对底物６－硝基（３－苯乙酰氨基）苯甲酸（６－ｎｉｔｒｏ－３－ｐｈｅｎｙｌａｃｅｔａｍｉｄｏ ｂｅｎｚｏｉｃ ａｃｉｄ,ＮＩＰＡＢ）的ｋｃａｔ分别从１     

Biosynthesis and Characterization of 4-Fluorotryptophan-Labeled Escherichia coli Arginyl-tRNA Synthetase

Escherichia coli 4-fluorotryptophan-substituted arginyl-tRNA synthetase was biosynthetically prepared and purified from a tryptophan auxotroph which could overproduce this enzyme. A method was developed to separate 4-fluorotryptophan from tryptophan and to determine accurately their contents in the 4-fluorotryptophan-containing proteins. It was confirmed that more than 95% of the tryptophan residues in the purified 4-fluorotryptophan-substituted arginyl-tRNA synthetase were replaced by 4-fluorotryptophan. Studies on the effect of the 4-fluorotryptophan replacement on properties of the enzyme showed that, when compared with the native enzyme, both the specific activity and the first-order rate constant of the fluorinated enzyme decreased by approximately 20% with just slightly higher K _m values. CD studies, however, did not reveal any difference between the secondary structure of the native and fluorinated enzymes. In addition, thermal unfolding studies showed that the 4-fluorotryptophan replacement did not significantly affect the thermal stability of the enzyme. We may conclude that the substitution of 4-fluorotryptophan in arginyl-tRNA synthetase had no substantial effect on the structure and function of the enzyme. Finally, a preliminary study of ¹⁹F nuclear magnetic resonance spectroscopy of the fluorinated enzyme has shown promising prospect for further investigation of its structure and function with NMR.     

Regulation of Glutamine Synthetase X. Effect of Growth Conditions on the Susceptibility of Escherichia coli Glutamine Synthetase to Feedback Inhibition

The kinetic properties of Escherichia coli glutamine synthetase are markedly influenced by the manner in which the organism is grown. Enzyme obtained from stationary-phase cells grown on glycerol and glutamate is strongely inhibited by each of the eight feedback effectors known to influence this enzyme; however, the enzyme from log-phase cells grown on glucose and growth-limiting concentrations of NH(4)Cl is stimulated by some of these effectors. Of the growth variables examined, nitrogen source and time of harvest were the most important; carbon source and aeration seemed to have no effect. Two purified enzyme preparations have been obtained from cells grown under two different conditions, designated enzymes I and II for convenience. Enzyme I is stimulated by adenosine 5'-monophosphate, histidine, and tryptophan in the transfer assay, whereas enzyme II is strongly inhibited by all effectors tested. Enzyme I has a higher specific activity in the forward assay in the presence of Mg(++) or Co(++), whereas enzyme II is more active in the presence of Mn(++).     

Active efflux of bile salts by Escherichia coli.

Enteric bacteria such as Escherichia coli must tolerate high levels of bile salts, powerful detergents that disrupt biological membranes. The outer membrane barrier of gram-negative bacteria plays an important role in this resistance, but ultimately it can only retard the influx of bile salts. We therefore examined whether E. coli possessed an energy-dependent efflux mechanism for these compounds. Intact cells of E. coli K-12 appeared to pump out chenodeoxycholate, since its intracellular accumulation increased more than twofold upon deenergization of the cytoplasmic membrane by a proton conductor. Growth inhibition by bile salts and accumulation levels of chenodeoxycholate increased when mutations inactivating the acrAB and emrAB gene clusters were introduced. The AcrAB system especially appeared to play a significant role in bile acid efflux. However, another efflux system(s) also plays an important role, since the accumulation level of chenodeoxycholate increased strongly upon deenergization of acrA emrB double mutant cells. Everted membrane vesicles accumulated taurocholate in an energy-dependent manner, apparently consuming delta pH without affecting delta psi. The efflux thus appears to be catalyzed by a proton antiporter. Accumulation by the everted membrane vesicles was not decreased by mutations in acr and emrB genes and presumably reflects activity of the unknown system seen in intact cells. It followed saturation kinetics with Vmax and Km values in the neighborhood of 0.3 nmol min(-1) mg of protein(-1) and 50 microM, respectively.