Transport across isolated bacterial cytoplasmic membranes

367

BIOCHIMICA ET BIOPHYSICA ACTA BBA 85110

TRANSPORT

ACROSS

ISOLATED

BACTERIAL

CYTOPLASMIC

MEMBRANES H. R. K A B A C K

The Roche Institute o f Molecular Biology, Nutley, N. J. 07110 (U.S.A.) (Received January 31st, 1972)

CONTENTS I.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

368

A. Definitions . . . . . 1. Passive diffusion . . 2. Facilitated diffusion 3. Active transport . . 4. Group translocation

. . . . .

368 368 369 369 369

Characterization and properties of membrane vesicles . . . . . . . . . . . . . . . A. Preparation and homogeneity . . . . . . . . . . . . . . . . . . . . . . . . B. Physical properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

369 369 373

III. Group translocation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. The P-enolpyruvate phosphotransferase system . . . . . . . . . . . . . . . . . B. Adenine phosphoribosyitransferase . . . . . . . . . . . . . . . . . . . . . .

375 375 377

IV. Active transport . A. Coupling of a transport in E. 1. Amino acid

377

II.

B. C. D. E. F.

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . . . . . . . . . . membrane-bound D-lactate coil membrane vesicles. . . transport . . . . . . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . . . . . . . . . dehydrogenase to amino . . . . . . . . . . . . . . . . . . . . . . . .

. . . . .

. . . . .

. . acid . . . .

. . . . .

. . . . .

. . . . .

. . . . and sugar . . . . . . . .

2. fl-Galactoside transport . . . . . . . . . . . . . . . . . . . . . . . . . 3. Coupling of other sugar transport systems to D-lactate dehydrogenase . . . . . 4. Activity of vesicles compared to whole cells . . . . . . . . . . . . . . . . 5. Source of D-lactate in E. coli . . . . . . . . . . . . . . . . . . . . . . . . Substrate oxidation by membrane vesicles . . . . . . . . . . . . . . . . . . Site of energy-coupling between D-lactate dehydrogenase and transport . . . . . . Mechanism of energy-coupling of D-lactate dehydrogenase to transport . . . . . . Coupling of ascorbate-phenazine methosulfate to transport . . . . . . . . . . A kinetic model for the redox transport mechanism . . . . . . . . . . . . . .

. . .

. .

377 377 378 380 382 382 382 385 386 391 392

Abbreviations: IPTG, isopropyl-fl-D-thiogalactopyranoside; TMG, methyl-l-thio-fl-D-galactopyranoside; TDG, fl-D-galactosyl-l-thio-fl-D-galactopyranoside;a-MG, methyl-a-D-glucopyranoside; a-MGP, methyl-a-D-glucopyranoside phosphate; HOQNO, 2-heptyl-4-hydroxyquinoline-N-oxide; PCMB, p-chloromercuribenzoate; CCCP, carbonyl cyanide-m-chlorophenylhydrazone; DCIP, dichloropbenolindophenol; PMS, phenazine methosulfate; DMO, 5,5-dimethyloxazolidine-2,4-dione; ANS, 1-anilino-8-naphthalene sulfonic acid.

Biochim, Biophys. Acta, 265 (1972) 367-416

368

H. R. KABACK G. H. I. J. K.

Solubilization and partial purification of "carriers"~ . . . . . . . . . . . . . . . General importance of dehydrogenase-coupled transport . . . . . . . . . . . . . Valinomycin-induced Rb ÷ transport . . . . . . . . . . . . . . . . . . . . . . Effect of valinomycin on respiration-coupled sugar and amino acid transport . . . Proton or potential gradients . . . . . . . . . . . . . . . . . . . . . . . . .

393 395 398 399 400

V. Applications to studies of membrane structure . . . . . . . . . . . . . . . . . . A. Functional separation of transport from barrier function . . . . . . . . . . . . B. Temperature transitions . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Structural and functional correlations . . . . . . . . . . . . . . . . . . . . . 1. X-ray diffraction studies . . . . . . . . . . . . . . . . . . . . . . . . . 2, Fluorescence depolarization studies. . . . . . . . . . . . . . . . . . . . . 3. Studies of fluorescence emission maximum (2Ema,) . . . . . . . . . . . . . .

402 402 406 409 410 410 410

VI. Conclusions and speculations . . . . . . . . . . . . . . . . . . . . . . . . . .

411

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

414

I. INTRODUCTION T w o m a j o r obstacles to u n d e r s t a n d i n g m e m b r a n e function have been the biochemical n a t u r e o f the m e c h a n i s m s involved in active t r a n s p o r t a n d the relationship o f these m e c h a n i s m s to the cell m e m b r a n e . I n 1960, it was r e p o r t e d that subcellular p r e p a r a t i o n s f r o m Escherichia coli W were able to catalyze the u p t a k e o f g l y c i n e l ; a n d it was subsequently shown that cell-free p r e p a r a t i o n s o f m e m b r a n e s , essentially d e v o i d o f c y t o p l a s m i c constituents, catalyzed the u p t a k e o f glycine 2,a, its u l t i m a t e c o n v e r s i o n to p h o s p h a t i d y l e t h a n o l a m i n e 3, a n d the concentrative u p t a k e o f p r o l i n e 4' 5. These studies established a m o d e l system which allowed an e x a m i n a t i o n o f the p r o b l e m s outlined above. W i t h the a c c u m u l a t i o n o f biochemical a n d genetic i n f o r m a tion f r o m this a n d other experimental systems, r a p i d d e v e l o p m e n t s have t a k e n place in this a r e a over a relatively short p e r i o d o f time. This discussion concerns a p o r t i o n o f these developments, p a r t i c u l a r l y those f r o m the a u t h o r ' s l a b o r a t o r y related to bacterial m e m b r a n e vesicles. O f necessity, relevant areas o f research which have been reviewed in detail will not be covered a n d the r e a d e r is referred to o t h e r reviews 6-12.

IA.

Definitions

Before progressing, several m e c h a n i s m s by which substances are t h o u g h t to cross cell m e m b r a n e s should be defined, since their clear distinction is i m p o r t a n t to this discussion. These mechanisms, which are based on kinetic a n d t h e r m o d y n a m i c considerations, are as follows: 1. Passive diffusion. By passive diffusion a substance crosses a m e m b r a n e as a result o f r a n d o m m o l e c u l a r m o t i o n . T h e t r a n s p o r t e d solute is t h o u g h t n o t to interact with a n y m o l e c u l a r species in the m e m b r a n e . Passive diffusion m e c h a n i s m s m a y be modified by solvent d r a g (in which the p e n e t r a t i n g substance is swept t h r o u g h a q u e o u s

Biochim. Biophys. Acta, 265 (1972) 367-416

BACTERIAL CYTOPLASMIC MEMBRANE TRANSPORT

369

pores in the cell membrane by bulk water flow), by membrane charge, and by the degree of hydrophobicity of the diffusion barrier. 2. Facilitated diffusion. In facilitated diffusion a transported solute is presumed to combine reversibly with a specific "carrier" in the membrane. The carrier or carrier-substrate complex oscillates between the inner and outer surfaces of the membrane, releasing and binding molecules on either side. Because of the short distances covered, it is thought that thermal energy and/or molecular deformation resulting from binding and release of substrate can account for the small amount of motion needed. Neither of these two mechanisms requires metabolic energy nor do they lead to concentration against a gradient. 3. Active transport. By active transport the solute is accumulated against an electrochemical or osmotic gradient. It is generally believed that this mechanism requires metabolic energy on the part of the cell, as well as a specific membrane carrier molecule. The classic model for this mechanism postulates that the penetrating species combines with a carrier and that the carrier or the carrier-substrate complex is then subjected to modification in the membrane. The carrier-substrate complex formed on the outside surface of the membrane crosses the membrane and is modified on the inside surface in such a way that the carrier has a lowered affinity for its substrate. The substrate is released into the interior of the cell and the carrier is free to cross back to the outside surface of the membrane where the cycle is repeated. 4. Group translocation. Another transport mechanism which is metabolically dependent is group translocation. In this process a covalent change is exerted upon the transported molecule such that the reaction itself results in the passage of the molecule through the diffusion barrier. Group translocation is not a "classic" active transport mechanism since the transported solute is modified chemically. One such group translocation mechanism, which is called vectorial phosphorylation 9, leads to the translocation and accumulation of a molecule within the cell by phosphorylation during passage through the membrane. This review is concerned primarily with active transport. Group translocation mechanisms are also discussed to some extent, and passive and facilitated diffusion are mentioned where pertinent. Mechanisms such as pinocytosis or phagocytosis, which may be involved in transport processes in cells of higher organisms, are not discussed since they are thought not to occur in bacterial cells.

II. CHARACTERIZATION OF MEMBRANE VESICLES

11.4. Preparation and homogeneity Although methods of preparation of bacterial cytoplasmic membrane vesicles, their homogeneity, composition, and physical properties have been reviewed in detail 13'14, a brief discussion is indicated. Biochim. Biophys. Acta, 265 (1972) 367-416

370

H. R. KABACK

For orientation, an electron micrograph of a longitudinal section through an intact E. coil cell is shown in Fig. I. The structure is rod shaped with two trilaminar membranes bordering its exterior. The outer membranous structure is the lipopolysaccharide layer of the cell wall: the inner, the plasma membrane. Located between these membranes, in the so-called pericytoplasmic space, is the peptidoglycan layer of the cell wall; this structure, however, cannot be seen in Fig. 1. It is the rigid peptidoglycan layer that confers the rodlike shape to the bacterial cell and prevents it from bursting in hypotonic environmentslL Finally, within the inner cell membrane are the ribosomes and " D N A - p l a s m " or nucleoplasm of the cell. When cells such as this are treated with enzymes (i.e., lysozyme with many organisms or lysostaphin with Staphylococci) that attack the rigid layer of the cell wall or grown in the presence of penicillin, the peptidoglycan layer is degraded or outgrown. The result is an osmotically fragile cell which is usually spherical in shape. When these osmotically sensitized

Fig. 1. Electron micrograph of a longitudinal section through an intact E. coil cell. LPS, lipopolysaccharide; PM, plasma membrane. This micrograph was taken by Dr. Samuel Silverstein of the Rockefeller University. Bioehim. Biophys. Acta, 265 (1972) 367-416

BACTERIAL CYTOPLASMIC MEMBRANE TRANSPORT

371

cells are subjected to osmotic lysis using EDTA, DNAase, RNAase, and large lysis ratios, followed by extensive homogenization, washing and differential centrifugation, the structures shown in Figs 2 and 3 are obtained. As shown in Fig. 2, when E. coli ML strain is subjected to this procedure, the structures obtained consist predominantly of intact "unit membrane"-bound sacs varying from 0.5 to 1.5#m in diameter. The great majority of these sacs are surrounded by a single trilaminar membrane layer which is 65-70/k thick. The sacs appear to be empty and without internal structure. It is also noteworthy that vesicles prepared using either the lysozyme-EDTA or penicillin methods appear to be morphologically identical 14.

Fig. 2. Electron micrograph of a membrane vesicle preparation from E. coil ML 308-225. The membranes were prepared from lysozyme-EDTA-induced spheroplastst3. The micrograph was taken by Dr. Vincent Marchesi of the National Institute of Arthritis and Metabolic Disease. Biochim. Biophys. Acta, 265 (1972) 367-416

372

H. R. KABACK

Fig. 3. Electron micrograph of a membrane vesicle preparation from E. coli W. Tile membranes were prepared from lysozyme-EDTA-induced sheroplasts ~3, The micrograph was taken by Dr. Vincent Marchesi of the National Institute of Arthritis and Metabolic Disease.

BACTERIALCYTOPLASMICMEMBRANETRANSPORT

373

When E. coli W was subjected to the same procedures (Fig. 3). the structures seen also consist of intact unit membrane-bound sacs but the vesicles are much more heterogeneous. The diameters of the sacs vary from 0.1 to 1.5 #m and they are surrounded by one to five or six trilaminar membrane layers. As in E. coli ML these membrane vesicles are also empty and devoid of internal structure, and each trilaminar membrane layer is also 65-70 A thick. Membranes prepared from other strains of E. coli (K-12, W2244, and K21t) and from Salmonella typhimurium are essentially indistiguishable from E. coli W. The reason for this striking morphological difference between ML membranes and membranes prepared from other strains of E. coli and S. typhimurium is probably related to the observation that ML membranes contain negligible quantities of lipopolysaccharide, whereas the other membrane preparations mentioned above contain significant quantities of this cell wall component (see below). In any case, there are no demonstrable physiological differences between ML membranes and the other membrane preparations as judged by any of the transport systems studied thus far. The purity and homogeneity of E. coli membrane preparations have also been established by a number of other criteria. Briefly, the membrane preparations contain less than 5 % of the DNA and RNA, 10-15 % of the protein, and at least 70 ~ of the phospholipids of the whole cells from which they were derived. Less than 1% of the activities of glutamine synthetase, fl-galactosidase, fatty acid synthetase, or leucineactivating enzyme and 2 % or less of each of the "periplasmic enzymes''6 are found in the preparations. Although 10-15% of the total cellular protein remains in the vesicles, almost all of the cytoplasmic proteins are lost, as demonstrated by acrylamide disc gel electrophoresis. As for contaminating peptidoglycan, vesicles prepared from penicillin-induced spheroplasts retain only 10 % or less of the diaminopimelic acid of the spheroplasts. Regarding lipopolysaccharide, vesicles prepared from the ML strains of E. coli have less than 3 % (by dry weight), whereas vesicles prepared from a number of other strains of E. coli and S. typhimurium have from 7 to 17 ~. Expressed as a function of dry weight, the vesicles are approximately 60-70 protein, 30-40 % phospholipid, and approximately 1% carbohydrate. liB. Physical properties One essential property of any system that is to be used as a model for transport is that it must have a continuous surface (i.e., it must be able to retain transported substrate). Although the sectioned material presented in Figs 2 and 3 gives the impression that vesicles are closed structures, it is only from techniques other than those in which there is a good possibility for sampling errors because of the thinness of the sections used that such conclusions may be drawn. The electron micrographs presented in Figs 4 and 5 were obtained using negative staining (Fig. 4) or freeze etching (Fig. 5) so that the surface of the vesicles could be observed. Both micrographs represent typical vesicles prepared from E. coli ML 308-225, and in both cases, there are no gross defects in the surface of the vesicles. Biochim. Biophys. Acta, 265 (1972) 367-416

374

H. R. KABACK

Fig. 4. Electron micrograph of E. coli ML 308-225 membrane vesicle negatively stained with phosphotungstic acid. Micrograph taken by Dr. Vincent Marchesi of the National Institute of Arthritis and Metabolic Disease. More convincing evidence for membrane continuity has been provided by experiments in which it was demonstrated that the vesicles are osmotically sensitive 5,14. That is, they shrink and swell appropriately when the osmolarity of the medium is increased or decreased. It has also been shown that there is a diffusion barrier to P-enolpyruvate which can be overcome by osmotically shocking the vesicles in the presence of this compound 14. Finally, it is noteworthy that a significant number of vesicles do not become inverted during lysis. This has been ascertained by freeze-etching microscopy (e.g. Fig. 5A and B) and by observing lysis under phase contrast*. Moreover, none of the vesicles prepared from Bacillus megaterium grown into the late logarithmic phase of growth completely loses its [~-hydroxybutyrate granules*. * H. R. Kaback, unpublished information. Biochim. Biophys. Acta, 265 (1972) 367-416

BACTERIAL CYTOPLASMIC MEMBRANE TRANSPORT

375

Fig. 5. Electron microscopy of freeze-etched membrane vesicles from E. coil ML 308-225. The micrographs were taken by Drs. Thomas Tillack and Vincent Marchesi of the National Institute of Arthritis and Metabolic Disease. (A) Outer surface; (B) inner surface. Approximately x 140 000.

III. GROUP TRANSLOCATION IliA. The P-enolpyruvate phosphotransferase system In 196416, a bacterial phosphotransferase system was reported which catalyzes the transfer of phosphate from P-enolpyruvate to various carbohydrates according to the following reactions: Enzyme I, Mg 2÷ P-enoipyruvate + HPr ~-

~- pyruvate + P - H P r

(1)

-> sugar-P + H P r

(2)

Enzyme II, Mg 2÷ (Factor III) P - H P r + sugar P-enolpyruvate + sugar

Enzyme I ........... HPr, Mg 2+, Enzyme II (Factor III)

-+ sugar-P + pyruvate

(3)

HPr, a heat-stable, low molecular weight protein which has been purified to homogeneity 17, and Enzyme I are predominantly soluble proteins, whereas Enzyme II is membrane bound. Enzyme II is responsible for specificity with respect to the Biochim. Biophys. Acta, 265 (1972) 367-416

376

H. R. KABACK

various sugars studied. Recently, Enzyme II activity has been solubilized and partially purified 18, and it has been shown that at least two protein fractions (one of which is sugar specific) and phosphatidylglycerol are required for Enzyme II activity. There is also evidence in some systems for the involvement of yet a fourth protein (Factor II1) 19,2° whose function is unknown. Subsequent to the initial description of the phosphotransferase system, biochemical and genetic evidence was presented by a number of laboratories indicating that this system might be involved in bacterial carbohydrate metabolism and/or transport. For an extensive review of this work, the reader is directed to ref. 9. Studies utilizing isolated bacterial membrane vesicles have demonstrated that the phosphotransferase system catalyzes the vectorial phosphorylation of glucose and related monosaccharides in E. coli, S. typhimurium and Bacillus subtilis 9,14,21-23. By this means, free sugar in the medium is transported into the vesicles as sugar-P without the mediation of an intramembranal free sugar pool. The most direct evidence for this contention is derived from double isotope experiments with membrane vesicles in which the intramembranal pool was preloaded with [14C]-glucose under conditions in which there was no phosphorylation of the sugar. After removal of the external isotope, the preloaded vesicles were exposed to [aH]-glucose in the presence of P-enolpyruvate. The vesicles showed an almost absolute preference for the [3H]-glucose added to the outside simultaneously with P-enolpyruvate. Other lines of evidence which substantiate the interpretation that the phosphotransferase system catalyzes a vectorial reaction have been discussed in detail 9,14.21, and will not be covered in this discussion. It is noteworthy that the membrane vesicles take up little glucose-P when incubated with or without P-enolpyruvate 9,14,21 Moreover, the diffusion-limited rates of appearance of glucose-6-P or free glucose into the intramembranal pool are similar. In light of the previous observations, this finding implies that external sugar reaches a catalytic site within the membrane and is translocated as a result of phosphorylation. Also, these experiments indicate that the phosphotransferase system does not function as a trap since there would be little advantage in trapping sugar-P from the point of decreasing outward diffusion through the membrane. In Staphylococcus aureus, the phosphotransferase system is involved in the transport of many carbohydrates 24-28. In most other organisms, however, this does not appear to be the case. In E. coli, for instance, this system is involved in the transport of relatively few sugars, i.e. glucose, mannose, fructose, and mannito129. The phosphotransferase system is not involved in amino acid transport or in most inducible sugar transport systems in E. coli, as will be discussed subsequently, nor is it apparently involved in glucose transport in obligate aerobes 3°. A conceptual model which has been proposed as a possible mechanism for vectorial phosphorylation 9' 14, as well as experiments related to the regulation of the phosphotransferase system by sugar-P's 9' 14.31 and to the role of phosphatidylglycerol in phosphotransferase activity have been presented elsewhere 9,14,32 and will not be discussed here. Biochim. Biophys. Acta, 265 (1972) 367-416

BACTERIAL CYTOPLASMIC MEMBRANE TRANSPORT

377

IIIB. Adenine phosphoribosyltransferase Evidence for a group translocation mechanism for purine transport has recently been presented by Hochstadt-Ozer and Stadtman ~~9. These workers have demonstrated that the uptake of adenine by membrane vesicles of E. coli is accompanied by its conversion to AMP and is stimulated by 5-phosphoribosyl 1-pyrophosphate. The enzyme mediating this activity-- adenine phosphoribosyltransferase - - is required for uptake, and variations in enzyme activity are reflected by changes in adenine transport.

IV. ACTIVE TRANSPORT IVA. Coupling of a membrane-bound D-lactate dehydrogenase to amino acid and sugar transport in E. coli membrane vesicles 1. Amino acid transport. Recent studies aa.34 have defined the energetics of amino acid transport in detail. The addition of D-lactate to suspensions of vesicles markedly stimulates proline uptake with a 20-30-fold increase over baseline levels. Of all the other metabolic intermediates and cofactors tested, only succinate, L-lactate, D,L-~-hydroxybutyrate, and N A D H replace D-lactate to any extent whatsoever, and each is much less effective than D-lactate. Furthermore, N A D in the presence of D-lactate causes no additional stimulation of proline transport. D-[14C]Lactate or [14C]succinate are converted stoichiometrically to pyruvate or fumarate, respectively22,33,35. Moreover, L-[14C]lactate and D,L-[14C]lactate are also converted stoichiometrically to pyruvate, and [~4C]pyruvate is not metabolized to any significant extent by the vesicles. Neither pyruvate nor fumarate has any effect on proline transport as noted above. These results indicate that the concentrative uptake of proline involves electron transfer, and more specifically, that a membrane-bound lactate dehydrogenase with a high degree of specificity towards D-lactate is tightly coupled to proline transport. Since the rate and extent of conversion of lactate to pyruvate is much greater than can be accounted for by proline transport alone, and because this system is so specific for D-lactate, the effect of D-lactate on the transport of other amino acids was investigated. The results of these experiments 33 demonstrate that the conversion of D-lactate to pyruvate markedly stimulates the initial rates of uptake and the steadystate levels of accumulation of proline, glutamic acid, aspartic acid, tryptophan, serine, glycine, alanine, lysine, phenylalanine*, tyrosine*, cysteine**, leucine, isoleucine, valine, and histidine. The transport of glutamine, arginine, cystine, methionine, * In earlier experimentsa3, the vesicles took up phenylalanine and tyrosine to a lesser extent than some of the amino acids. However, subsequent experiments in which the pH was varied showed that these amino acids are transported very well. ** In earlier studies33, cysteine transport was studied at a concentration of this amino acid which is much below the K= of the transport system. Biochim. Biophys. Acta, 265 (1972) 367-416

378

H. R. KABACK

and ornithine is stimulated only marginally by D-lactate. Virtually all of the radioactivity taken up by the vesicles in the presence of D-lactate is recovered as the unchanged amino acid 33. Furthermore, under the conditions studied, the steadystate concentration of each of the amino acids taken up is many times higher than that of the medium assuming that all of the amino acids taken up are in free solution in the intravesicular pool. The observations discussed above have been extended to all of the amino acid uptake systems*. Primarily D-lactate and, to a much lesser extent, succinate, L-lactate, D,L-~-hydroxybutyrate, and N A D H are the only energy sources which initiate the uptake of any of the amino acids. It should be noted, moreover, that the relative effects of these compounds on a particular amino acid transport system varies 22. Aspartic acid uptake, for instance, is not stimulated by either L-lactate or succinate, whereas the uptake of glutamic acid, isoleucine, valine, and phenylalanine is moderately stimulated by succinate, but not by L-lactate. Generally, when effective, succinate is only one-third to one-half as active as D-lactate, and L-lactate only one-tenth to one-fifth as effective. It is noteworthy that when vesicles are prepared from cells grown on either glycerol or enriched media, a membrane-bound a-glycerolphosphate dehydrogenase is induced which is coupled to the amino acid transport systems about as effectively as succinate dehydrogenase 36,3v. Similar observations have also been made for formate dehydrogenase 36. By means of competition experiments in which the ability of a series of unlabeled amino acids to inhibit the uptake of a given radioactive amino acid was tested, it has been determined that there are at least nine amino acid transport systems, each one of which is specific for a structurally related group of amino acids. There is also genetic evidence in some cases which corroborates the assignment of a particular group of amino acids to one transport system. For instance, in E. coli W 157 vesicles, the defect is highly specific for proline transport 4'5. Similarly, in D-serine-resistant mutants of E. coli, the vesicles show a specific defect for glycine and L-alanine transport** (refs 1-3). The data presented in Table I are a summary of kinetic constants obtained for the transport of various amino acids. The amino acids listed have been grouped according to their affinity for a common transport system as determined by competition and/or genetic studies. The observation that the transport systems for histidine and leucine, isoleucine and valine have two sets of kinetic constants is consistent with studies carried out with whole cells 38'39. 2. fl-Galactoside transport. Although the fl-galactoside transport system in E. coli has been examined in great detail, the mechanism of the coupling of meta-

* F. J. Lombardi and H. R. Kaback, unpublished information. ** In vesicles prepared from another D-serine-resistant mutant of E. coli K-12 which also lacks D-serine deaminase activity ~2° (contributed by Sharon Cosloy, Department of Microbiology, New York University Medical Center) a similarly specific defect is observed. Biochim. Biophys. Acta, 265 (1972) 367-416

BACTERIAL CYTOPLASMIC MEMBRANE TRANSPORT

379

TABLE I [S]o.5 AND VVALUES FOR TRANSPORT OF VARIOUS AMINO ACIDS BY E. coli ML 308-225 MEMBRANE VESICLES IN THE PRESENCE OF 20 mM D-LACTATE L-Amino acid

[S]o.s

V

(ARM)

(nmoles/mg protein/min)

Proline Glutamic Acid Aspartic Acid Serine Threonine Glycine Alanine Phenylalanine Tyrosine Tryptophan Histidine

1.0 11.0 2.9 2.6 5.4 1.4 8.4 0.42 0.68 0.33 0.15 4.0

1.3 4.0 1.1 4.0 1.4 0.75 1.0 2.6 3.2 0.7 0.20 0.7

Leucine

1.1 18.0

0.25 1.1

Isoleucine

1.7 21.0

0.25 0.60

Valine

2.0 29.0

0.2 1.1

Cysteine

38.0

18.5

bolic energy to active galactoside transport was poorly understood. Scarborough, Rumley and Kennedy 4° suggested an involvement of ATP in the lactose transport system of E. coli; however, studies by Pavlasova and Harold 41 on anaerobic methyl1-thio-fl-D-galactopyranoside ( T M G ) uptake indicate that uncouplers of oxidative phosphorylation block T M G accumulation but do not alter ATP levels. Fox and Kennedy 42 demonstrated the existence of a "permease" protein (the M protein) which is a product of the y gene 42. The subsequent suggestion of a role for the P-enolpyruvate phosphotransferase system in T M G uptake in E. coli 4'* raised the possibility that the M protein might be an inactivated Enzyme II. This topic has been discussed in detail in previous reviews 9,14. Since much of the interest in this laboratory over the past few years had been directed towards the role of the phosphotransferase system in sugar transport, and since all attempts to implicate this system in the transport of galactosides by vesicles were uniformly negative, the effect of D-lactate on the uptake of fl-galactosides by the vesicles was investigated as. The conversion of D-lactate to pyruvate in membrane vesicles prepared from cells containing a functional y gene markedly stimulates the initial rate of transport of [3-galactosides, and in a short time, the vesicles accumulate these sugars to intravesicuBiochim. Biophys. Acat, 265 (1972) 367-416

380

H.R. KABACK

lar concentrations at least 100 times higher than that of the medium. Virtually all of the radioactivity accumulated in the membranes is recovered as the unchanged substrate, and there is no detectable galactoside-P at any of the times sampled. The effect of various metabolites and cofactors on lactose accumulation in membrane vesicles demonstrates that of all the compounds tested, only D-lactate and, to a lesser extent, D,L-a-hydroxybutyrate, succinate, and L-lactate increase lactose transport above endogenous levels. N A D H does not stimulate lactose transport. ~-Glycerol-P and formate also stimulate lactose transport in membranes prepared from cells grown on appropriate media as demonstrated for the amino acid systems 36,37. Vesicles prepared from E. coli GN-245, a mutant lacking Enzyme I of the phosphotransferase system, but constitutive for lac, transport fl-galactosides in the presence of D-lactate but are completely unable to vectorially phosphorylate a-methylglucoside 21. D-Lactate does not stimulate a-methylglucoside uptake by membrane vesicles containing an intact phosphotransferase system. Moreover, P-enolpyruvate does not stimulate lactose, isopropyl-fl-D-thiogalactopyranoside (IPTG) or T M G uptake, nor is lactose-P or TMG-P detected in these experiments. Finally, membranes which transport fl-galactosides fail to exhibit phosphatase activity towards TMG-P, and the addition of lactose to vesicles incubated in the presence of [32p]-phosphoenolpyruvate does not accelerate the appearance of 32p~, as might be expected if a lactose-P phosphohydrolase 9 were involved in this system. [3-Galactoside transport in E. coli clearly does not involve the P-enolpyruvate phosphotransferase system. 3. Coupling of other sugar transport systems to D-lactate dehydrogenase. The transport systems for galactose 46, arabinose*, glucuronic acid*, gluconic acid* and glucose-6-P** are coupled to D-lactate dehydrogenase in a manner identical to that described for fl-galactosides 34,47. The transport of these sugars by the vesicles requires induction of the parent cells, is coupled primarily to D-lactate dehydrogenase, does not involve the P-enolpyruvate phosphotransferase system, and is inhibited by the same conditions which affect amino acid and fl-galactoside transport (see below). Kinetic constants for the transport of most of these sugars by the membrane vesicles listed are given in Table II. D-Galactose transport in membrane vesicles46 is especially noteworthy. This system is present in lac y - vesicles prepared from induced E. coli ML 3 (i+z÷y-a ÷) and ML 35 (i-z+y-a+), and E. coli ML 32400 (gal K-) 4a and W3092cy- [F-lac (y-) gal K - ] 49 which transport galactose constitutively. Like the transport systems for other sugars and amino acids, galactose transport is coupled primarily to the membranebound D-lactate dehydrogenase and exhibits all of the other properties to be discussed which are common to these transport systems. Of particular interest is the observation that "galactose binding protein" is totally absent from the vesicles (Fig. 6), as is a high affinity, low V galactose transport system present in the whole cells. These findings, in * G. K. Kerwar and H. R. Kaback, unpublished observations. ** P. Bhattaeharyya, F. J. Lombardi and H. R. Kaback, manuscript in preparation.

Biochim. Biophys. Acta, 265 (1972) 367-416

BACTERIAL CYTOPLASMIC M E M B R A N E T R A N S P O R T

381

TABLE II KINETIC CONSTANTS FOR S U G A R T R A N S P O R T SYSTEMS IN M E M B R A N E VESICLES

Sugar

Vesicles

Electron donor

Km (mM)

V (nmoles/mg protein/min)

D-Lactate Ascorbate-PMS D-Lactate D-Lactate D-Lactate Ascorbate-PMS D-Lactate

0.2 0.2 0.25 0.03 0.05 0.05 0.14

78 240 50 83 12 18 12

(E. coli) Lactose

ML 308-225*

Glucose-6-P Glucuronic acid Galactose

GN-2* ML 30* ML 35

Arabinose

ML 30

* Membranes prepared under gentle conditions (Le. avoiding the use of tight-fitting homogenizer) and assayed prior to freezing in liquid nitrogen. See Section IVA4, Activity of vesicles compared

to whole cells.

Fig. 6. Galactose-binding protein determinations in membrane vesicles and whole cells. The photographs are from Ouchterlony immunodiffusion plates in which the reactivity of antiserum against purified galactose binding protein was tested against appropriately treated 46 membranes and whole cells of E. cob ML 3 (i+z+y-a+) (A) and ML 35 (i-z+y-a +) (B). In each case, the center well contained the antiserum; Well 1 contained purified galactose-binding protein; Well 2 contained whole cells; and Well 3 contained membrane vesicles. The binding protein and antiserum against it were graciously contributed by Dr. Winfried Boos of the Massachusetts General Hospital. Similar studies carded out with E. coli ML 32400 (gai K-) and W3092cy- [F-lac (y-) gal K-] gave identical results. From G. K. Kerwar, A. S. Gordon and H. R. Kaback, J. Biol. Chem., 247 0972) 298.

Biochim. Biophys. Acta, 265 (1972) 367-416

382

H.R. KABACK

addition to the observation that the vesicles do not transport fl-methylgalactoside, indicate that the galactose transport system retained by the vesicles is the so-called "gal permease" system s°. Moreover, it is obvious that this system does not require the presence of galactose-binding protein for activity. Similar studies are in progress using antisera to purified arabinose- and leucine-binding proteins. Recent experiments carried out in collaboration with Dr. M. Iaccarino of the International Institute of Genetics and Biophysics in Naples indicate that a similar situation may exist for the leucine, isoleucine, valine transport system in E. coli. Mutants defective in this transport system 121 contain normal amounts of "periplasmic binding protein" for these amino acids. It is important that vesicles prepared from one of the mutants manifest the transport defect present in the whole cells, i.e. the vesicles do not transport leucine, isoleucine, or valine in the presence of D-lactate or ascorbate-phenazine methosulfate (PMS)*. Thus, the presence of the binding protein alone is not sufficient for translocation of these amino acids across the membrane. 4. Activity of vesicles compared to whole cells. Although previous data suggest that the rates of sugar and amino acid transport by isolated membrane vesicles is 10-20% of that found in whole cells, recent experiments indicate that vesicles lose significant activity due to excessive manipulation during preparation. Vesicles prepared using gentle conditions (i.e., avoiding vigorous homogenization in tightfitting homogenizers) and assayed for lactose or proline transport prior to freezing have specific activities 3-5 times higher than whole cells (comparing initial rates of uptake per mg protein). It should be emphasized, however, that quantitative comparisons between vesicles and whole cells are extremely difficult to interpret, especially when the transport activity manifested by whole cells may be the composite of more than one system (e.g. galactose transport). 5. Source of D-lactate in E. coli. E. coli possesses two distinct D-lactate dehydrogenases, a soluble nucleotide-dependent enzyme which catalyzes the conversion of pyruvate to D-lactate 51,52, and a membrane-bound, flavin-linked enzyme which catalyzes the conversion of D-lactate to pyruvate **'s3. Thus, via the soluble enzyme, the cell can produce D-lactate which may then be utilized via the membrane-bound enzyme to drive many transport systems and perhaps other cellular processes.

IVB. Substrate oxidation by membrane vesicles There is no relationship between the oxidase activity of the vesicles towards D-lactate, succinate, L-lactate, and N A D H , and the ability of these electron donors to stimulate transport a6. With vesicles prepared from succinate-grown cells, succinate is * G. K. Kerwar, H. R. Kaback, M. De Felice and M. Iaccarino, unpublished observations. ** In collaboration with Dr. Leonard D. Kohn of the Laboratory of Biochemical Pharmacology, National Institute of Arthritis and Metabolic Diseases, Bethesda, Md., the membrane-bound D-lactate dehydrogenase has been solubilizedand purified approximately 250-300-fold. At this stage, the preparation is approximately 80 % pure as judged by disc gel electrophoresis, and contains cytochrome bt but no phospholipid.

Biochim. Biophys. Acta, 265 (1972) 367-416

BACTERIAL CYTOPLASMIC M E M B R A N E TRANSPORT o x i d i z e d m u c h faster t h a n D - l a c t a t e a n d N A D H

383

is o x i d i z e d a p p r o x i m a t e l y as fast, yet

D - l a c t a t e is m a r k e d l y m o r e effective as a n e l e c t r o n d o n o r f o r s u g a r o r a m i n o a c i d transport. A s s h o w n in T a b l e III, D - l a c t a t e - d e p e n d e n t lactose t r a n s p o r t by vesicles is i n h i b i t e d 9 4 % by e x c l u s i o n o f oxygen. cyanide,

Moreover, the electron transfer inhibitors

2-heptyl-4-hydroxyquinoline-N-oxide

(HOQNO),

a m y t a l a n d the specific

D - l a c t a t e d e h y d r o g e n a s e i n h i b i t o r o x a m i c a c i d a6 also effectively b l o c k lactose a c c u m u l a t i o n . T h e effect o f t h e s e i n h i b i t o r s o n t h e r a t e o f D - l a c t a t e o x i d a t i o n is also s h o w n in T a b l e III, a n d it c a n be seen t h a t i n h i b i t i o n o f D - l a c t a t e o x i d a t i o n by these c o m p o u n d s is q u a l i t a t i v e l y s i m i l a r to the i n h i b i t i o n o f t r a n s p o r t .

Except for oxamate,

these i n h i b i t o r s also effectively b l o c k s u c c i n a t e a n d N A D H

o x i d a t i o n , as well as

the o t h e r t r a n s p o r t

systems discussed

previously.

R e c e n t i n v e s t i g a t i o n s o f the

r e s p i r a t o r y c h a i n o f E. coli 54 h a v e identified the a m y t a l - s e n s i t i v e site as a f l a v o p r o t e i n b e t w e e n D - l a c t a t e d e h y d r o g e n a s e a n d c y t o c h r o m e bi. I n a d d i t i o n , it has b e e n demonstrated that HOQNO

acts b e t w e e n c y t o c h r o m e bi a n d c y t o c h r o m e a2, p e r h a p s

TABLE III EFFECT OF VARIOUS C O M P O U N D S ON LACTOSE TRANSPORT A N D D-LACTATE OXIDATION BY ML 308-225 MEMBRANES Rates of oxygen uptake* by ML 308-225 membrane vesicles in the presence of D-Lactate were assayed with an oxygen electrode. [14C] Lactose transport was determined at 10 rain in the presence of D-lactate la. Inhibitors were incubated with membranes for 15 min at 25 o C in the absence of D-lactate. Reactions were then initiated by the addition of D-lactate and [14C]-lactate.

Condition during incubation 1. Aerobic 2. Anaerobic 3. Sodium cyanide 4.

HOQNO

5. Sodium amytal

6. Sodium oxamate 7. Sodium arsenate 8. Oligomycin 9. N-Ethylmaleimide 10. PCMB

Inhibitor concentration

Inhibition of lactose uptake

Inhibition of D-lactate oxidation

(M)

(%)

(%)

--10 -3

-94 76

-98 84

10 -2

98

--

29 70 37 -87 63 87 17 29 4 10 32 100 36 100

-52 -60 -70 91 -0 -0 -80 -90

2.10 -6 2"10-5 10-a 5"10-3 10 -2 10-3 5'10 -3 10-2 5'10 -2 2.10 -5 10-4 10-4 10 -a 10 -5 I0 "-4

* For the anaerobic samples (2), D-lactate oxidation was measured with D-[ 14C]-lactateaa'a s.

Biochim. Biophys. Acta, 265 (1972) 367-416

384

H.R. KABACK

at a quinone-containing component, and that cyanide blocks cytochrome a 2. Thus, each of the dehydrogenases studied is coupled to oxygen via a membrane-bound respiratory chain. D-Lactate-dependent lactose transport is inhibited by only 3 0 ~ despite concentrations of arsenate as high as 50 mM (Table III). Furthermore, the transport system is insensitive to oligomycin. Previously published observations aa demonstrate that the amino acid transport systems are not significantly inhibited by arsenate or oligomycin, and none of the other sugar transport systems studied are significantly inhibited by these compounds *'**'46. Vesicles have also been prepared with arsenate buffer rather than phosphate buffer throughout the procedure, and subsequently transport was assayed in arsenate buffer in the absence of inorganic phosphate. The results were the same as those discussed above - - little or no inhibition of transport. It is apparent from these studies that the effect of D-lactate on transport is not exerted through the production of stable high-energy phosphate compounds. This conclusion is supported by a large number of observations, among which are the absence of transport in the presence of ATP under conditions in which ATP is demonstrably accessible to reactive sites within the membrane sS, the failure of ADP or other nucleoside-diphosphates to stimulate transport in the presence of D-lactate or other electron donors, and the observation that similar membrane preparations do not carry out oxidative phosphorylation 56. In addition, the ATP content of the vesicles is below the limits of the luciferin-luciferase assay and does not increase when the vesicles are incubated with electron donors 73. Vesicles prepared from mutants which are "uncoupled" for oxidative phosphorylation 122 transport proline normally in the presence of D-lactate or ascorbate-PMS*. Respiratory particles prepared from Mycobacterium phlei transport proline in the absence of oxidative phosphorylation 123. Finally, lactose transport and D-lactate oxidation by the vesicles are inhibited by the sulfhydryl reagents N-ethylmaleimide and p-chloromercuribenzoate (PCMB). T h e inhibition observed with these reagents will be discussed in greater detail below. 2,4-Dinitrophenol, carbonyl cyanide m-chlorophenylhydrazone (CCCP), and azide do not significantly affect D-lactate oxidation, despite profound inhibition of lactose transport 34'a6. This finding is not surprising since most bacterial electron transfer systems are not subject to respiratory control. Numerous experiments have established that the addition of lactose and/or amino acids to the vesicles has no effect on D-lactate oxidation a6. These results are in opposition to the experiments of Kepes 57 who found that T M G stimulated oxygen uptake by whole cells. The observations, taken as a whole, indicate that the specificity of the transport systems for D-lactate dehydrogenase cannot be accounted for solely on the basis of its presence in the vesicles (to the exclusion of other dehydrogenases), and furthermore,

* G. K. Kerwar and H. R. Kaback, unpublished observations. ** P. Bhattacharyya, F. J. Lombardi and H. R. Kaback, manuscript in preparation.

Biochim. Biophys. Acta, 265 (1972) 36%416

BACTERIALCYTOPLASMICMEMBRANETRANSPORT

385

that the coupling of D-lactate dehydrogenases to transport involves the flow of electrons through a respiratory chain to oxygen as the terminal electron acceptor. IVC. Site of energy coupling between D-lactate dehydrogenase and transport Difference spectra between D-lactate-, succinate-, NADH-, L-lactate-, or dithionite-reduced samples and oxidized samples are indistinguishable quantitatively and qualitatively34'a6. Furthermore, difference spectra between D-lactate-reduced and NADH-, succinate-, L-lactate-, and dithionite-reduced samples show no absorption bands. Since the rate of reduction of cytochrome bl by these substrates is directly proportional to their rates of oxidation, these data indicate that each dehydrogenase is coupled to the same cytochrome chain, and that there is no unique cytochrome chain which couples D-lactate dehydrogenase to oxygen. Thus, the site of energy-coupling between D-lactate dehydrogenase and transport must lie between the dehydrogenase and cytochrome bx the first cytochrome in the E. coli respiratory chain. Further evidence for this conclusion is obtained from experiments in which the rate of proline transport was measured in the presence of saturating concentrations of D-lactate and increasing concentrations of succinate. In membrane preparations in which succinate dehydrogenase is more active than D-lactate dehydrogenase, the addition of succinate results in progressive inhibition of proline uptake. Moreover, using the radioactive assay for D-lactate oxidation32'3s, succinate inhibits the conversion of D-lactate to pyruvate. However, succinate does not inhibit the isolated, partially purified D-lactate dehydrogenase nor the membrane-bound enzyme when dichloroindophenol (DCIP, an artificial electron acceptor which accepts electrons directly from flavins) is used instead of oxygen as an electron acceptor 36. The only reasonable mechanism by which such inhibition could take place is that succinate dehydrogenase, when it is more active than D-lactate dehydrogenase, is able to saturate cytochrome bl kinetically, and thus inhibit D-lactate dehydrogenase. In order for transport to be inhibited by this means, the site of energy coupling must be proximal to cytochrome bl. Finally, direct evidence for this hypothesis is provided by experiments in which the effects of N-ethylmaleimide and PCMB on D-lactate and NADH oxidation were investigated36. Both of these sulfhydryl reagents produce marked inhibition of D-lactate oxidation at the same concentrations which inhibit transport. Moreover, PCMB inhibition of D-lactate oxidation is reversed by dithiothreitol, providing further evidence for sulfhydryl involvement in D-lactate oxidation. The effect of these thiol reagents on D-lactate oxidation does not appear, however, to be mediated at the level of the primary dehydrogenase. Neither the D-lactate: DCIP reductase activity of the intact vesicles nor that of the solubilized, partially purified preparation of this enzyme is sensitive to N-ethylmaleimide or PCMB. It is extremely important to note that NADH oxidation is not sensitive to N-ethylmaleimide or PCMB. Thus, neither the primary D-lactate dehydrogenase nor the cytochrome system contains an accessible sulfhydryl group. Therefore, the site of inhibition of D-lactate oxidation by NBiochim. Biophys. Acta, 265 (1972)367-416

386

H.R. KABACK

ethylmaleimide and PCMB must lie between D-lactate dehydrogenase and the cytochromes. A schematic representation of the sequence of events thought to occur is presented in Fig. 7. As shown, electrons from D-lactate flow through the "carriers" or something closely aligned with the "carriers" before entering the respiratory chain, after which they ultimately reduce oxygen to water. Each of the electron transfer inhibitors used in Table III (the sites of inhibition of which are also shown in Fig. 7) would then inhibit solute transport by interrupting the flow of electrons either above (oxamate), at (PCMB, N-ethylmaleimide), or below (amytal, HOQNO, cyanide, or anoxia) the site of energy-coupling between D-lactate dehydrogenase and transport.

D- Lactate dehydrogenase

="CARRIER"

• Cyt.ochrome b1 ~

i

Cytochrome o2

• 02

I

1 I I

OXAMATE

NEM PCMB

AMYTAL

HOQNO

CYANIDE

Fig. 7. Electron transfer pathway from D-lactate dehydrogenase to oxygen, showing sites of inhibition of electron transfer inhibitors. NEM, N-ethylmaleimide.

IVD. Mechanism of energy coupling of D-lactate dehydrogenase to transport The findings presented above, especially when considered in conjunction with those to be discussed, indicate that the transport-specific components of the D-lactate dehydrogenase-coupled transport systems reflect the redox state of the respiratory chain between D-lactate dehydrogenase and cytochrome bl. It should be emphasized at this point that although it cannot be concluded definitively that the "carriers" are electron transfer intermediates themselves, much of the evidence available at present is consistent with this probably oversimplified interpretation. D-Lactate dehydrogenase activity and the initial rates of sugar and amino acid transport respond identically to temperature; and furthermore, both phenomena have the same activation energy of 8400 cal/mole 47. A similar activation energy for transport is also obtained with an artificial electron donor system - - ascorbatephenazine methosulfate - - which reduces the respiratory chain after D-lactate dehydrogenase (el p. 391). On the other hand, the D-lactate: DCIP reductase activity of the vesicles exhibits a markedly different temperature profile, and has an activation energy of approximately 30000 cal/mole. Thus, some component of the membrane which is common to both D-lactate oxidation and transport apparently determines the activation energy for both processes. Evidence has been presente& 7 which demonstrates that the steady-state levels Biochira. Biophys. Acta, 265 (1972) 367-416

BACTERIAL CYTOPLASMIC MEMBRANE TRANSPORT

387

of sugar and amino acid accumulation at temperatures ranging from 0 to 53 ° C represent equilibrium states in which there is a balance between influx and efflux rates. Moreover, it has been shown47 that the efflux induced at 45 ° C is a saturable, "carrier"mediated phenomenon with a much lower affinity (i.e., higher Kin) for intramembranal substrate than the influx system has for external substrate, but a very similar maximum velocity. Anoxia, cyanide, HOQNO, and 2,4-dinitrophenol induce the efflux of intramembranal solute and the kinetics of cyanide-induced efflux manifest the same apparent affinity constant as temperature-induced ettlux and the same maximum velocity as that of the influx process. Thus, efflux occurs at rates which are comparable to those of influx, but much higher concentrations of substrate are required to saturate the "carrier" on the inside of the membrane. Data have also been presented47 which demonstrate that the efflux rate of intramembranal solute responds to temperature in a manner that is essentially the inverse of the response of the steady-state level of lactose accumulation to temperature. The fl-galactoside transport system in E. coli is inhibited by sulfhydryl reagents s8-6°, and the ability of two substrates of this transport system to protect against sulfhydryl inactivation has led to the identification and purification of the M protein, the product of the y gene42''*3'61'62. However, very little evidence has been presented which has any bearing on the mechanistic role of sulfhydryl groups in the galactoside transport system. Evidence discussed above demonstrates that Dlactate-induced respiration in the isolated membrane preparations is inhibited by PCMB and N-ethylmaleimide, and that the site(s) of action of these compounds is between the primary D-lactate dehydrogenase and cytochrome bl, i.e., at the site of energy coupling. Virtually every transport system in the membrane vesicles which is coupled to D-lactate dehydrogenase is inhibited by PCMB and N-ethylmaleimide, and the inhibition produced by PCMB is essentially completely reversed by dithiothreitoP 7. Moreover, all "carrier"-mediated aspects of [3-galactoside transport, even those that are independent of D-lactate oxidation, are inhibited by PCMB and N-ethylmaleimide; and in each case, the inhibition by PCMB is reversed by dithiothreitoP 7. Temperature-induced efflux, exchange of external lactose with [14C]-lactose present in the intramembranal pool, and efflux induced by the addition of 2,4dinitrophenol or cyanide are all blocked by N-ethylmaleimide and PCMB, and the inhibition by PCMB is completely reversed with dithiothreitol. Thus, D-lactate oxidation and "carrier" activity are both dependent upon functional sulfhydryl groups which are located in a specific region of the respiratory chain. A number of electron transfer inhibitors whose sites of action in the electron transfer chain of E. coli are well documented (see Fig. 7) were studied with respect to their effect on the ability of the vesicles to retain accumulated solute47. In the experiments shown in Fig. 8, data for the lac transport system are presented as a typical example. Under the conditions studied, each of the inhibitors used inhibits D-lactate oxidation and the initial rate of lactose transport by at least 70-80% (cf. Table III). Since only anoxia and those inhibitors which block electron transfer after the site of energy coupling (i.e., anoxia, cyanide, HOQNO, and amytal) cause Biochim. Biophys. Acta, 265 (1972) 367-416

388

H.R. KABACK

efflux, reduction of the electron transfer chain between D-lactate dehydrogenase a n d cytochrome bl must be responsible for efflux. Strikingly, oxamate which inhibits electron transfer before the site of energy coupling does n o t cause efftux, n o r does P C M B or N-ethylmaleimide. Thus, the redox state of the respiratory chain at the site of energy coupling determines the rate of efflux. The effect of anoxia a n d the same inhibitors on the time course of lactose uptake is also consistent with this interpretation (Fig. 9). Since the removal of oxygen or the addition of electron transfer inhibitors which inhibit after the site of energy coupling cause reduction of energy-coupling site, m e m b r a n e s incubated u n d e r these

.o~I

° 0

o 0

°

0XAUATE

~

ADD

\ z

\

° ~

A 2O

"~ 30

\

°

~o 25 AMYTAL

\\ ~

(-~)~oH20QNO

/ / / / =

I-

~

•

- KCN

0

L

5 TIME (MrN)

t

I0

F-

5 0

~

./

io

1

I0-o

~zo

0

OXAMATEI~,/ i f / '~'j AMYTAL t..0 ~ l

m t

t

•

/ f 2 <';..oo.o 5

I0

15

TIME(MIN)

Fig. 8. Effect of various electron transport inhibitors on lactose efflux. ML 308-225 vesicles were incubated at 20 ° C in the presence of D-lactate and [1-1*C]lactose. After 15 min, one of the inhibitors shown was added at the following final concentrations: potassium cyanide (1), 0.01 M; HOQNO (A), 8"10-5 M; sodium amytal (A), 0.01 M; PCMB (v), 8.3"10-5 M; potassium oxamate (~), 6 mM. For the samples labeled (-) 02 (©), the incubations were carried out in tubes fitted with rubber stoppers through which a needle was inserted. During the initial 15 rain incubation, the tubes were gassed with air, and at 15 rain, the gas mixture was changed to argon which was continued for the remainder of the incubation. Control samples (No add. ; 0 ) were incubated under identical conditions with the exception that no inhibitor was added after the initial 15-rain incubation. From H. R. Kaback and E. M. Barnes, Jr, J. Biol. Chem., 246 (1971) 5523. Fig. 9. Effect of various electron transport inhibitors on the time course of lactose uptake by ML 308-225 membranes. Where indicated, reaction mixtures contained the inhibitor given in the figure at the concentration stated in the legend to Fig. 8. For the samples labeled (--)02 ([3- -~), the incubations were carried out in tubes fitted with rubber stoppers through which a needle was inserted. The samples were gassed with argon for 5 min before the addition of D-lactate and [14C]-lactose by injection through the stopper. The incubations were then continued under argon. The control samples ( 1 ) contained none of the inhibitors and were incubated under aerobic conditions (room air). From H. R. Kaback and E. M. Barnes, Jr, J. Biol. Chem., 246 (1971) 5523. Biochim. Biophys. Acta, 265 (1972) 36%416

BACTERIALCYTOPLASMICMEMBRANETRANSPORT

389

conditions should show profound inhibition of uptake throughout the time course of the experiment. On the other hand, inhibitors which work before or at the site of energy coupling prevent reduction of the energy-coupling site. Since the rate of reduction of this site by D-lactate dehydrogenase but not its rate of oxidation by cytochrome b~ is inhibited, vesicles incubated under these conditions exhibit markedly diminished initial rates of uptake but eventually accumulate significant quantities of solute. The experimental findings presented here are consistent with the conceptual working model presented in Fig. 10. It should be emphasized again that the evidence currently available indicates that all of the D-lactate dehydrogenase-coupled sugar and amino acid transport systems in E. cell membrane vesicles behave in the same manner. Simplistically, the "carriers" are depicted as electron transfer intermediates which undergo reversible oxidation-reduction. In the oxidized state, the "carrier" has a high affinity site for ligand which it binds on the exterior surface of the membrane. Electrons coming ultimately from D-lactate through one or possibly more flavoproteins reduce a critical disulfide in the "carrier" molecule resulting in a conformational change. Concomitant with this conformational change, the affinity of the "carrier" for its ligand is markedly reduced, and the ligand is released on the interior surface of the membrane. The reduced "sulfhydryl" form of the "carrier" is then oxidized by cytochrome b~ and the electrons then flow through the remainder of the cytochrome chain to reduce molecular oxygen to water. The reduced form of the "carrier" can also "vibrate" and catalyze a low affinity, "carrier"-mediated, non-energy dependent transport of ligand across the membrane. The proposed mechanism also implies that there are a number of functionally OUT

fP X ~ =

oxJOXIDIZED

IN

][Cyto~~

ox

OUT

OUT

I~PYR~RED

;:%~'~~:...... ~ ~ :.~

IN

:~:.~

~

~

fp' ~ H ~ H ~ y t o

--RE--HIGH

b .l ~

LJ\o

REDUCED

Km P Y R ~ R E D

~/I~ ~ ~.

fp~

~

~

--RE[~;~i~

~ ~I:~ ~ ~i~ii!

Cyto b]

o-L.ox i

REDUCED

============================

IN

Fig. L0. Conceptual working model for D-lactate dehydrogenase-coupled transport systems. D-LAC, D-lactate; PYR, pyruvate; fp, flavoprotein; cyto b~, cytochrome b,; ox, oxidized; red, reduced. OUT signifies the outside surface of the membrane; IN signifies the inside surface. The spheres located between fp and cyto b~ represent the "carrier"; DO , a high affinity binding site and ~ , a low affinity binding site. The remainder of the cytochrome chain from cytochrome b~ to oxygen has been omitted. From H. R. Kaback and E. M. Barnes, Jr, J. Biol. Chem., 246 (1971) 5523.

Biochim. Biophys. Acta, 265 (1972) 367-416

390

I-I. R. KABACK

heterogeneous electron transfer intermediates between different D-lactate dehydrogenase molecules and cytochrome bl. For each transport system, there should be an electron transfer intermediate which has a binding site that is specific for that particular transport substrate. Supportive evidence for this prediction is provided by the experiments in which lactose transport was studied in the presence of structurally unrelated solutes which are also transported by D-lactate dehydrogenase-coupled systems4L Little or no inhibition of either the initial rate of lactose uptake or the steady-state level of lactose accumulation is observed. Moreover, the sum of the V values of all of the known D-lactate dehydrogenase-coupled transport systems in a particular membrane preparation is less than the V of the D-lactate dehydrogenase activity in the same membrane preparation*. Although no direct evidence has been presented which demonstrates that the "carriers" are obligatory electron transfer intermediates, this formulation is consistent with most of the experimental observations available and is the simplest conception possible. D-Lactate oxidation and both the energy- and non-energy-dependent aspects of transport are inhibited by PCMB and N-ethylmaleimide. Furthermore, the inhibition observed with PCMB is essentially completely reversed with dithiothreitol. Since the site of energy coupling lies between the primary dehydrogenase and cytochrome b~, and the site(s) of inhibition by sulfhydryl reagents is also between D-lactate dehydrogenase and cytochrome bl, the proposed mechanism is supported by more than simplicity of conception. However, it should be emphasized that the model does not account for the mechanism of action of "uncoupling agents" (i.e., 2,4-dinitrophenol, CCCP) and azide on the system. These reagents abolish transport without inhibition of D-lactate oxidation 34-36, indicating that the "carriers" are probably not obligatory electron transfer intermediates. The observations, as well as the conceptual model to which they have led, do not conflict with studies carried out with whole cells 63-6s. Furthermore, the proposed mechanism would explain some apparent inconsistencies that have been reported. Only fl-D-galactosyl-l-thio-fl-D-galactopyranoside (TDG) and melibiose protect the galactoside transport system against inhibition by N-ethylmaleimide 61. Since many other galactoside analogues are transported via this system, it was proposed that the M protein has two sites and that T D G and melibiose bind to one site only. However, no evidence has been presented to substantiate this suggestion. According to the concept proposed here, the M protein has two sites, but one is involved in electron transfer and the other in binding substrate. Since presumably all galactosides bind to the M protein by virtue of their galactose moieties, it does not seem unlikely that T D G and melibiose, due to their size (TDG) or shape (melibiose is an a-galactoside), sterically protect a sulfhydryl group which is not in the binding site. A similar interpretation could also explain the so-called "energy-uncoupled" galactoside transport mutants 66'67. These mutants have increased concentrations

* F. J. Lombardi and H. R. Kaback, unpublished observations. Biochim. Biophys. Acta, 265 (1972) 367-416

BACTERIALCYTOPLASMICMEMBRANETRANSPORT

391

of the M protein, but do not carry out the concentrative uptake of galactosides as well as the parent. Concentrative uptake of TMG by membrane vesicles prepared from one of these mutants 66 is stimulated by D-lactate, but only about one-third as well as the parent preparation*. D-Lactate-dependent transport of proline and phosphotransferase-mediated uptake of a-methylglucoside, on the other hand, are identical to that of the parent membranes. Possibly the defect in these mutants is related to redox properties of M protein. Since the M protein, by our calculations, would account for no more than about 10% of the D-lactate dehydrogenase activity of the membrane preparations, it is not surprising that no significant difference in D-lactate oxidation is found between the mutant and the wild-type membrane preparations. IVE. Coupling o f ascorbate-PMS to transport

Konings, Barnes and Kaback 6s have demonstrated that an artificial electron donor system, ascorbate-PMS, can be coupled to sugar and amino acid transport in the vesicle preparations. The initial rate of lactose transport in E. coli membrane vesicles is stimulated about 3 times more effectively by ascorbate-PMS than D-lactate, the best physiological electron donor for this system. This finding is consistent with the observation that reduced PMS is oxidized much faster than D-lactate by the vesicles. It should be emphasized, however, that comparative initial rates of transport in the presence of ascorbate-PMS versus D-lactate differ, depending upon the transport system under investigation. The significance of this observation will be discussed below. The effect of ascorbate-PMS on transport is inhibited by anoxia, and by the addition of cyanide, HOQNO, PCMB, N-ethylmaleimide and amytal in a manner identical to that described for D-lactate. Significantly, oxamic acid has no effect, since this inhibitor acts at the level of the primary dehydrogenase for D-lactate. These studies indicate that ascorbate-PMS reduces the respiratory chain of the vesicles at a redox level below that of cytoehrome bl. This conclusion is substantiated by spectrophotometric data. Moreover, ascorbate-PMS reduces only 21% of the membranebound flavoprotein, whereas D-lactate reduces 63 %. Since PCMB and N-ethylmaleimide inhibit transport in the presence of ascorbate-PMS, the postulated sulfhydryl component(s) of the respiratory chain which is (are) common to both transport and D-lactate oxidation must reside within the segment of the respiratory chain reduced by ascorbate-PMS. Amytal inhibits the effect of ascorbate-PMS on transport, suggesting that there are flavoproteins between the "carriers" and cytochrome bl or possibly that the "carriers" are flavoproteins themselves. No evidence has been presented which can resolve these possibilities. Since the methods used in these experiments do not distinguish between flavoprotein and non-heme iron 4s, it is also possible that a nonheme iron protein is involved in this system. In this regard, recent experiments carried * H. R. Kaback, unpublishedinformation. Biochim. Bioph. Acta, 265 (1972) 367-416

392

H.R. KABACK

out with the iron chelator cyclic dihydroxybenzoyl serine triplex 69 demonstrate that this compound inhibits transport with either D-lactate or ascorbate-PMS as electron donors*. Moreover, cyclic dihydroxybenzoyl serine triplex does not cause the efflux of solute accumulated in the intramembranal pool. It is also noteworthy that Bragg 7° has demonstrated that ascorbate-PMS reduces non-heine iron in respiratory particles prepared from E. coli. These observations place non-heme iron near the site of ascorbate-PMS reduction either before or at the site of energy coupling. With regard to the conceptual model presented above (cf. Fig. 10), it is interesting that the relative effects of ascorbate-PMS or D-lactate differ depending upon which transport system is studied. Preliminary observations indicate that although most of the transport systems are stimulated more by ascorbate-PMS than by D-lactate, lysine transport is apparently stimulated more by D-lactate and serine, threonine, and proline transport are stimulated about equally by ascorbate-PMS and D-lactate. One reasonable explanation for these findings is that there are small differences in the redox potentials of various energy-coupling sites such that a few are not completely reduced by ascorbate-PMS. IVF. A kinetic mode~for the redox transport mechanism The mechanistic model presented in Fig. 10 may also be expressed kinetically as shown in Fig. 11. Using this formulation, Dr. F. J. Lombardi has derived the following differential equation**:

OUT

IN c°~-S tl

D-Loc-~

"'It'~ k'°D-Lac

S°

c'~;

pyr

Ik,,

--~SI

L ",,¢|

ACTIVE TRANSPORT

cyto x - ' ~ ' ~ ~. ................. C~ed--S[ .

.

.

c,

.

S° ~

! I

~

k k3

CIred-S

FACILITATED DIFFUSION ( EFFLUX, EXCHANGE, COUNTERFLOW)

Cred-S

Fig. 11. Kinetic scheme based on the mechanistic model presented in Fig. 10.

* G. K. Kerwar and H. R. Kaback, unpublished observations. ** F. J. Lombardi and H. R. Kaback, unpublished information. Biochim. Biophys. Acta, 265 (1972) 367-416

BACTERIAL CYTOPLASMIC MEMBRANE TRANSPORT

393

Vnetvn V°utEVCff ] SS°

Vnet=

1

~ m + S-° ( l ~ a )

--

S°

1+

K°m

St

a

+--

Kim

where v = velocity V ¢ = maximum velocity of energy-dependent active uptake K¢m= apparent Michaelis constant for energy-dependent active uptake S ° = external substrate concentration a = fraction of carriers in the reduced form under steady-state conditions Voff = maximum velocity of efflux Ktm = apparent Michaelis constant for efflux S 1 = internal substrate concentration K°m = apparent Michaelis constant for energy-independent influx Each term in this equation can be measured experimentally. With the exception of the term ~, the determination of the other terms in the equation should be obvious from the previous discussion. 0: is measured in the following manner: The rate of efflux at the steady state is a function of the amount of"reduced carrier". Thus, vesicles are incubated with a radioactive transport substrate in the presence of D-lactate. After substrate has been accumulated to steady-state levels, an excess of non-radioactive substrate is added, and the initial rate of efflux is measured. In this manner, the specific activity of the external transport substrate is markedly decreased, and unidirectional effiux is observed. Expressing the above equation in the form v = [f] (1-~)-[g]~, the rate of efltux under these conditions will be given by [g]0t. The same experiment is then carried out but cyanide is.added in addition to non-radioactive substrate in order to reduce the respiratory chain, and convert all of the "carrier" to the "reduced form". Under these conditions, ocwill be equal to 1, and according to the abbreviated equation written above, the rate of efflux is given by [g]. The ratio of these rates is equal to 0c. Using values derived from transport experiments carried out at a variety of temperatures with the equation shown above, a computer was programmed to generate uptake curves as a function of time. The computer-generated curves are indistinguishable from the actual experimental data 47. Although this type oftreatment does not discriminate between mechanisms in which the "carrier" is or is not an obligatory electron transfer intermediate, it provides strong evidence for the general type of redox mechanism presented in Fig. 10.

IVG. Solubilization and partial purification of"carriers'" Recently, Dr. Adrienne S. Gordon has succeeded in solubilizing and partially purifying a fraction from the vesicles which apparently contains many of the amino acid "carrier" components 71. Membrane vesicles were extracted with non-ionic detergents, and the extract (containing approximately 20 % of the total membrane protein) was subjected to gel filtration in the presence of detergent. The absorbance

Biochim. Biophys. Acta, 265 (1972) 367-416

394

H.R. KABACK

profile at 280 nm of the solubilized material after Sephadex G-100 chromatography is shown in Fig. 12. There are three absorbance peaks: (1)Fractions 15-17, excluded protein (Peak I); (2) Fractions 21-25, included protein of relatively high molecular weight (Peak II); and (3) Fractions 48-58, included protein of low molecular weight (Peak III). The relative negative absorbance in Fractions 18 and 19 is due to binding of detergent to protein in the void volume. The proline-binding activity of each fraction of the column as measured by ultrafiltration is also shown in Fig. 12. Although material in each of the three proteincontaining fractions binds proline, Peak III exhibits a much higher specific activity than the two other fractions containing high molecular weight material. It should be emphasized that [14C] proline is not altered chromatographically by 48 h incubation at room temperature with each of these fractions or with the unfractionated extract. When the eluent fractions from Sephadex G-100 were assayed for D-lactate dehydrogenase activity (Fig. 12), the activity was associated with Peaks I and II only. Peak III had no detectable D-lactate dehydrogenase activity. Succinate dehydrogenase and N A D H dehydrogenase activities were not detected in the column effluent.

A

,/i l'~

i I

/ ll.t" ,

I 1

•

I

I I

II

t

I

,.II

0.3

I

I

"l

I

I

I I

~. |

iI•

..

,i

i; c

1.0

• I

°

I

I I

t •

•

,. I

.

I

\

I

/

30 ;

.E

I.

I I

0.5

o.z o

*-

~. t

~~"

~. "~

-zo 8

,,-,

/

"

-~

~

E

c 0 ~)

to

r-

0 I0

20

30 TUBE

40

50

0

60

NO.

Fig. 12. Absorption profile, proline binding activity and D-lactate: DCIP reductase activity of Brij 36-T extract from E. coli ML 308-225 fractionated on Sephadex G-i00. From A. S. Gordon, F. J. Lombardi and H. R. Kaback, Proc. Natl Acad. Sci. U.S., 69 (1972) 358. Biochim. Biophys. Aeta, 265 (1972) 367-416

BACTERIAL CYTOPLASMIC MEMBRANE TRANSPORT

395

Judging from the ultraviolet absorption spectrum of Peak III, this fraction is composed predominantly of protein. Moreover, when extracts from membranes prepared from cells grown in the presence of 32p1 were chromatographed on Sephadex G-100, no significant radioactivity appeared in Peak III. Thus, very little, if any, phospholipid is associated with this fraction. The proline-binding activity exhibited by Peak III is highly specific. Of all the amino acids tested, only proline itself inhibits the binding of the radioactive amino acid. In contrast, proline binding is not affected by the presence of a mixture of structurally unrelated amino acids. The proline-binding activity of Peak III is inhibited by N-ethylmaleimide and PCMB, and the inhibition exhibited by PCMB is reversed by dithiothreitol. These results are analogous to the behavior of the proline transport system as studied in the intact membrane vesicles47'71. It should be re-emphasized that these compounds affect both D-lactate-dependent and D-lactate-independent aspects of transport. In contrast to the inhibition of proline binding by sulfhydryl reagents, electron transfer and general metabolic inhibitors such as amytal, cyclic dihydroxybenzoyl serine triplex, HOQNO, CCCP, and 2,4-dinitrophenol at concentrations which inhibit lactose and proline transport, do not inhibit proline binding by Peak III. In addition to proline, Peak III also exhibits binding activity for lysine, serine, tyrosine, and glycine. These amino acids, like proline, are transported by systems with low apparent Km values (cf. Table I), suggesting that the amino acid-specific components have high affinities. Like the proline binding activity of Peak III, binding activity for each of these amino acids is inhibited reversibly by PCMB. Moreover, binding activity for each of these amino acids is not affected by the presence of structurally unrelated amino acids. Binding activity for amino acids other than those mentioned has not yet been tested. It should be emphasized that the solubilized membrane components discussed here are distinctly different from the "binding proteins" released by the cold osmotic shock method of Neu and Heppe172 (see also refs 6, 7, 11, 12). The latter are water soluble and localized in the periplasmic space, whereas the components presented here are associated with cytoplasmic membrane and are soluble only in detergent. In addition, as discussed previously, the membrane vesicles contain no "galactosebinding protein" (cf. Fig. 6). Finally, the soluble binding proteins obtained by osmotic shock are insensitive to sulfhydryl reagents and contain no cysteine. Although one might predict that specific components of other D-lactate dehydrogenase-coupled transport systems are also present in Peak III, many of them would be difficult to detect by direct binding assay. It is unlikely, for instance, that the M protein could be detected by these techniques, since the fl-galactoside transport system has a relatively high Km (cf. Table II).

IVH. General importance of dehydrogenase-coupled transport The ascorbate-PMS system has extended the study of respiration-coupled transport systems to membrane vesicles prepared from a wide variety of bacteria 6a. Biochim. Biophys. Acta, 265 (1972) 367-416

H. R, KABACK

396

3.0

AI E. c01i ML 308-225

1.5 E) M. denitrificans I

2.0

hO

=

_=

/

1.0

/

0.5 0

0

o 1.5

GLU-N

B) S. typhimurium

F) S.aureus U-71

0.9

PRO

0. c 0

,~ i.o

0.6

E Q

E

0.3

c

o

0

a 1.5 C) Ps. putida tlJ PRO I--

°

Q.

/

0.6

/

1.0

0.4

n~

I--

¢~

-0,5

/

0.2

o z

0 :5.0 - D) R mirobilis

2.0

0.75

/

0.5

1.0

0

0.25

I

2 3 4 TIME (MIN)

5

H) B. ~ubtili$ 6 0 - 0 0 9

7-

/

I

2

:5

4

5

TIME (MIN)

Fig. 13. Amino acid uptake by membrane vesicles prepared from various Gram-positive and Gramnegative organisms. Membrane vesicles prepared from E. coli ML 308-225 (A), S. typhimurium SB 102 (B), Ps. putida (C), P. mirabilis ATCC 9240 (D), M. denitrificans ATCC 13543 (E), S. aureus U-71 (F), B. megaterium (G), and B. subtilis 60-009 (H) were assayed for the uptake of the radioactive amino acids shown at 25 ° C under oxygen. O, ascorbate and PMS were added; A, ascorbate was added; v , PMS was added; and O, none of the above added. From W. N. Konings, E. M. Barnes, Jr and H. R. Kaback, J. Biol. Chem., 246 (1971) 5857.

Biochim. Biophys. Acta, 265 (1972) 367-416

B A C T E R I A L CYTOPLASMIC M E M B R A N E T R A N S P O R T

397

The data presented in Fig. 13 demonstrate that ascorbate-PMS markedly stimulates the uptake of proline by membrane vesicles prepared from E. coli (A), S. typhimurium (B), Pseudomonas putida (C), Proteus mirabilis (D), B. megaterium (G), and B. subtilis (H). In addition, the uptake of glutamine by membranes prepared from Microeoecus denitrificans (E) and lysine by S. aureus membrane vesicles (F) is similarly stimulated by ascorbate-PMS. In each case, ascorbate plus PMS, but not ascorbate or PMS alone, markedly stimulates both the initial rate of uptake and the steady-state levels of accumulation of the appropriate amino acid in each membrane preparation studied. Although the physiological electron donor(s) for the transport systems in many of the membrane vesicles is (are) not known at present, the following observations are noteworthy: (I) Short, White and Kaback 73 have demonstrated that the transport of 16 amino acids by membrane vesicles prepared from S. aureus is coupled exclusively to a membrane-bound a-glycerol phosphate dehydrogenase. With this sole exception, amino acid transport in S. aureus vesicles appears to be catalyzed by mechanisms which are very similar to those described in the E. coli system. Thus, transport is coupled to a specific dehydrogenase, is dependent on electron transfer but independent of oxidative phosphorylation, the site of energy coupling occurs between the primary dehydrogenase and the cytochrome chain, and there appear to be one or more sulfhydryl components in the respiratory chain between a-glycerolphosphate dehydrogenase and the cytochrome chain which are essential for transport and a-glycerol-P oxidation. (2) Konings and Freese* have shown that the concentrative uptake of a variety of amino acids, in addition to L-serine TM, by B. subtilis membranes is coupled primarily to a-glycerolphosphate and NADH dehydrogenases, and also, to some extent, to L-lactate dehydrogenase. In many other respects, this system is also similar to the E. cell system. (3) Isolated membrane vesicles from M. denitrificans accumulate glycine, alanine, glutamine, and asparagine in the presence of D-lactate**. (4) Barnes*** has recently observed that glucose transport by membrane vesicles prepared from the obligate aerobe Azotobaeter rinlandii is markedly stimulated by ascorbate-PMS, D-lactate, and malate (in the presence of flavin adenine dinucleotide). Despite the preliminary nature of some of these observations, the data indicate that active transport systems which are basically similar to the D-lactate dehydrogenase-coupled systems described in E. cell vesicles are present in membrane preparations from a variety of other organisms. The coupling of particular dehydrogenases to transport systems may be important with regard to the ecology of various bacterial species.

* W. N. Konings and E. Freese, manuscript submitted for publication. ** H. R. Kaback, unpublished information. *** E. M. Barnes, Jr, personal communication.

Biochim. Biophys. Acta, 265 (1972) 367-416

398

H . R . KABACK

IVI. Valinomycin-induced Rb + transport The cyclic depsispeptide antibiotic valinomycin 75.76 greatly increases membrane permeability, specifically to K ÷, Rb ÷, and Cs ÷ (refs. 77, 79). Membranes whose permeability is altered by this ionophore include bacterial77,s 0. s 1, erythrocyte7 s,s 2.s 3, mitochondrial s4, and artificial black lipid films 79. Studies with black lipid films indicate that valinomycin forms one-to-one complexes with K ÷ (ref. 83) which then diffuse across the film. Be' this means, K ÷ is sheltered from the hydrophobic interior of the film within the cyclic valinomycin molecule 76, the exterior of which is soluble in the matrix of the film. Valinomycin facilitates the net movement of K ÷ in mitochondria s* which has led to the hypothesis that this ionophore may be a prototype for natural K + carriers. Recently, Bhattacharyya, Epstein and Silver 8s demonstrated that the addition of valinomycin to isolated bacterial membrane vesicles prepared from E. coli results in the accumulation of K ÷ or Rb ÷ by a temperature- and energy-dependent process. Moreover, these workers demonstrated that membrane vesicles prepared from mutants of E. coli altered in K + transport show defects in the valinomycin-induced accumulation of K ÷ that are related to the defects manifested by the intact cells. Since the vesicles establish what appears to be a proton gradient 91, it seemed likely that the extrusion of protons might be the driving force for this transport system and possibly others. In collaboration with Dr. F. J. Lombardi of this laboratory and Dr. John P. Reeves of Rutgers University, it has been demonstrated that isolated membrane vesicles prepared from E. coli, M. denitrificans, and S. aureus, prepared in the absence of K ÷, are unable to accumulate Rb ÷. When valinomycin is added to the vesicles, however, the membranes catalyze the active transport of Rb ÷ by an extremely active mechanism which appears to be identical to the mechanisms for sugar and amino acid transport described above. Valinomycin-induced Rb ÷ transport in E. coli membrane vesicles is at least 5 times more active than any of the sugar or amino acid transport systems. With E. coli and M. denitrificans vesicles, valinomycin-induced Rb ÷ uptake is markedly stimulated by D-lactate and ascorbate-PMS. Succinate, D,L-a-hydroxybutyrate, L-lactate, and N A D H partially replace D-lactate or ascorbate-PMS, but are much less effective. With S. aureus membranes, only a-glycerol-P and ascorbatePMS stimulate valinomycin-induced Rb + uptake. ATP and P-enolpyruvate, in addition to a number of other metabolites and cofactors, do not stimulate Rb ÷ uptake in the presence of valinomycin with any of the membrane vesicles studied. Rb ÷ uptake by E. coli vesicles in the presence of D-lactate and valinomycin requires oxygen, and is blocked by the electron transfer inhibitors and uncoupling agents mentioned above. Initial rates of D-lactate dehydrogenase and Rb + uptake respond identically to temperature and both processes have the same activation energy as that of the sugar and amino acid transport systems. Steady-state levels of Rb ÷ accumulation at a variety of temperatures represent equilibrium states in which there is a balance between influx and efltux. Biochim. Biophys. Acta, 265 (1972) 367-416

BACTERIAL CYTOPLASMIC MEMBRANE TRANSPORT

399

Initial rates of uptake at various concentrations of Rb + in the presence of D-lactate or ascorbate-PMS and saturating concentrations of valinomycin exhibit saturation kinetics with an apparent Km of approximately 2 mM for Rb ÷. Rate studies carried out at various concentrations of valinomycin also exhibit saturation kinetics for the antibiotic. In this case, an apparent Km of approximately 3"10-7 M was obtained. Preliminary attempts to study the affinity of valinomycin for the membrane indicate that the compound is tightly bound. Using Rb ÷ uptake in the presence of D-lactate as an indication of valinomycin binding by the membranes, vesicles were treated with saturating concentrations of valinomycin, washed, and assayed for Rb ÷ uptake. After diluting and washing the preparations twice, at least 50 % of the capacity for Rb + uptake was retained. As demonstrated previously with the sugar and amino acid transport systems 47, reduction of the respiratory chain between D-lactate dehydrogenase and cytochrome b~ causes Rb ÷ efflux. Thus, anoxia, cyanide, HOQNO, and amytal, each of which inhibit selectron transfer after the site of energy coupling, all cause marked rates of Rb ÷ efflux. Oxamate does not initiate Rb ÷ efflux, despite marked inhibition of the initial rate of valinomycin-induced Rb ÷ transport. This observation provides strong evidence that valinomycin does not simply enhance the passive flux of Rb ÷ across the membrane since the membranes maintain a large Rb ÷ concentration gradient in the presence of oxamate despite inhibition of D-lactate dehydrogenase and Rb ÷ influx.

IVJ. Effect of valinomycin on respiration-coupled sugar and amino acid transport Previous work a4,aS,as demonstrated that valinomycin inhibits the transport of many sugars and amino acids that are transported via D-lactate-dependent mechanisms. Moreover, inhibition by valinomycin is completely dependent on the presence of K + (refs. 34, 85)*. When transport by the vesicles is assayed in media lacking K + valinomycin exhibits no inhibitory effect. In addition, valinomycin has no effect on D-lactate oxidation in the presence or absence of K +. The following experimental observations** indicate that the inhibitory effects of valinomycin are not related primarily to either changes in K + concentration across the membrane or to increased K + flux across the membrane: (1) The half-maximal inhibitory concentration of valinomycin (Ki) differs depending upon which transport system is studied. For instance, the K~ for valinomycin with respect to lactose transport is approximately 10-7 M (i.e., similar to the Km for valinomycin with respect to Rb + transport). However, the K~ values for proline, serine, lysine, and glutamic acid transport are 10-8, 6" 10-9, 10-8, and 10-7 M, respectively. Finally, glucuronic acid transport is not inhibited by valinomycin. Since Rb + uptake as a function of valinomycin concentration is a smooth function, suggesting a single reaction mechanism for valinomycin-induced Rb + transport, it is * E. M. Barnes, Jr., F. J. Lombardi and H. R. Kaback, unpublished information. ** F. J. Lombardi, J. P. Reeves and H. R. Kaback, manuscript in preparation.

Biochim. Biophys. Acta, 265 (1972) 367-416

400

I-t. R. KABACK

unlikely that the inhibitory effect of valinomycin is related to K + movements per se. (2) In the presence of K ÷, valinomycin induces the efflux of lactose and proline from the intramembranal pool subsequent to accumulation in the presence of Dlactate. This effect of valinomycin is not blocked by oxamate at concentrations which essentially completely inhibit Rb ÷ transport in the presence of valinomycin. These results indicate that K+-valinomycin is able to inhibit sugar and amino acid transport by means of specific interactions with component(s) in the membrane. Possibly, by competing with sugar and amino acid "carriers" for an electron source, valinomycin is able to inhibit other transport systems and simultaneously channel much of the electron flow from D-lactate dehydrogenase to Rb ÷ transport. Such a mechanism would be consistent with both the high activity of the valinomycin-induced Rb ÷ transport system and with the inhibitory effect of valinomycin. IVK. Proton or potential gradients Chemiosmotic coupling has been suggested by Mitchell a6'av (see also ref. 88) as a possible mechanism for oxidative phosphorylation in mitochondria, and Harold and Baarda 89 and West 9° have applied this theory to active transport in bacteria. According to the theory, oxidative phosphorylation or active transport is coupled to the movement of protons or cations as the primary driving force. For instance, valinomycin-induced Rb + transport could be coupled to the extrusion of protons or another positively charged species from the vesicles. In this way, K + or Rb + would move into the vesicles in order to maintain electroneutrality, and valinomycin would simply facilitate the movement of these cations. One of the attractive aspects of this hypothesis is that it attributes the primary inhibitory mechanism of ionophores and proton-conducting "uncoupling agents" to alterations in hypothetical proton or potential gradients. Although the pH of the suspension decreases on addition of D-lactate and other electron donors 91, a proton gradient is probably not established under most conditions. The evidence for this conclusion is as follows*: (1) Vesicles do not take up the lipid-soluble weak acid 5,5-dimethyloxazolidine-2,4-dione (DMO) in the presence of D-lactate or other electron donors. D M O is a weak acid, metabolically inert, which diffuses passively across many biological membranes. The permeability coefficient of the uncharged acid is generally much greater than that of the anion. The distribution of DMO is therefore a function of pH 89,92, and these observations indicate that the intramembranal pH is not alkaline with respect to the external medium. (2) The addition of transport substrates (i.e., lactose plus a mixture of amino acids) has no effect on the rate or absolute amount of acidification. (3) Membrane vesicles treated with mouse duodenal phospholipase A-B 93'**

* F. J. Lombardi, J. P. Reeves, and H. R. Kaback, manuscript in preparation. ** Mouse duodenal phospholipase A-B was graciously contributed by Dr. A. Ottolenghi of the Department of Pharmacology, Duke University. Biochim. Biophys. Acta, 265 (1972) 367-416

BACTERIAL CYTOPLASMIC MEMBRANE TRANSPORT

401

such that they retain the catalytic activities associated with transport (i.e., P-enolpyruvate-dependent phosphorylation of a-methylglucoside and D-lactate oxidation) but are unable to accumulate solute (cf., Section VA, Separation of transport from barrier function, below) exhibit similar p H changes on addition of electron donors. Moreover, so-called proton conductors such as 2,4-dinitrophenol and CCCP have the same effect on pH changes observed in normal and phospholipase-treated vesicles. Since these preparations are devoid of barrier function, it is highly unlikely that the observed pH changes can be due to proton gradients. (4) Vesicles containing high intramembranal concentrations of K ÷ or Na + incubated in the presence of D-lactate exhibit no change in the rate of acidification on addition of valinomycin. (5) When vesicles are loaded with 22Na÷, addition of valinomycin, D-lactate, and RbCI results in the movement of 22Na+ out of the intravesicular space. 22Na+ efliux is obliterated by the omission of any one of these components. Moreover, D-lactate alone, or D-lactate plus a mixture of lactose and amino acids does not induce 22Na÷ efflux, nor does acidification of the external medium. Finally, addition of D-lactate in addition to tetraphenylarsonium chloride, a lipophilic cation, has no effect on 22Na÷ movements. (6) When vesicles contain low Na ÷ concentrations, addition of valinomycin and K ÷ doubles the rate and extent of acidification over that observed with D-lactate alone. These results indicate that the extrusion of Na + or protons from the vesicles is the result of (rather than the cause of) the active accumulation of Rb ÷ or K ÷. Regarding the role of potential gradients, a number of experimental approaches were used: (1) The transport of each ionic species in the reaction mixtures was measured in the presence of D-lactate or ascorbate-PMS, and it was established that the vesicles do not accumulate Mg 2+, 5042-, PO4 a-, K ÷ (in the absence of valinomycin), Na ÷, CI-, or pyruvate. (2) An intensive study of l-anilino-8-naphthalene sulfonic acid (ANS) fluorescence was initiated. This compound is an anion which becomes highly fluorescent when it interacts with certain proteins and membranes 94-1°°. The behavior of this fluorescent probe has been postulated to reflect membrane potential changes in mitochondrial°L When D-lactate is added to membrane vesicles in the presence of ANS, there is a rapid decrease in fluorescence which is maintained until the reaction mixture becomes anaerobic at which time there is a large increase in fluorescence. The rate at which the fluorescence decrease progresses is most marked with D-lactate, and much slower with succinate or L-lactate. Addition of electron transfer inhibitors prior to D-lactate markedly inhibits the fluorescence changes. When added subsequent to D-lactate, electron transfer inhibitors whose site of action is after the site of energy coupling (i.e., cyanide, HOQNO, or amytal) reverse the initial decrease in fluorescence, whereas oxamate, N-ethylmaleimide, and PCMB which inhibit before or at the site of energy coupling have little or no effect. These effects are obviously analogous to certain aspects of the transport systems. As demonstrated with the pH measurements,

Biochim. Biophys. Acta, 265 (1972) 367-416

402

H.R. KABACK

the behavior of phospholipase-treated membranes is similar to that of untreated membranes with regard to each of the fluorescent studies mentioned. Thus, it is unlikely that membrane potentials are the primary driving force for transport. More definitive evidence that the fluorescence changes described are due to structural changes in the membrane has recently been obtained from preliminary studies of ANS fluorescence in the energy transfer mode. If a fluorescent probe like ANS can absorb energy from another fluorescing species in its environment, the ANS molecule becomes excited and emits photons. This phenomenon is called energy transfer 124, and by this means, information can be obtained with regard to the primary fluorescent source. In this case, the tryptophan residues in the membrane proteins are excited by light at 292 mn, and the fluorescence emission of ANS is recorded at 480 nm. Since tryptophan emission and ANS absorption overlap, it is possible to study ANS in an energy transfer mode. It has been demonstrated directly that the ANS fluorescence is due to absorption from tryptophan by the decrease in tryptophan emission. Moreover, the ratio of the 290 to 370 nm excitation maxima of ANS is increased when energy transfer is the mode of excitation. In any case, the effects of various electron donors and electron transfer inhibitors on ANS fluorescence in the energy transfer mode are qualitatively similar to the findings described above, and phospholipase-treated membrane vesicles exhibit similar phenomena. These studies provide strong support for the contention that conformational changes in components of the membrane are intimately involved in the mechanism of active transport.

V. APPLICATIONS TO STUDIES OF MEMBRANE STRUCTURE

VA. Functional separation of transport from barrier function Due to the unique features of the P-enolpyruvate phosphotransferase system in the vesicles, this vectorial phosphorylation mechanism may be used to study certain functional aspects of membrane structure. As opposed to the "carrier"-mediated active transport systems discussed above, sugars transported via the phosphotransferase system appear in the intravesicular pool as phosphorylated derivatives. Since the vesicles do not dephosphorylate sugar phosphates nor can they transport sugar phosphates 9,14,21,3~ either by active transport or by facilitated diffusion unless a specific sugar-P transport system is induced in the parent cells (cf Section IVA3, Coupling of other sugar transport systems to D-lactate dehydrogenase), the retention of substrates transported via the phosphotransferase system is a reflection of the passive permeability of the membrane. By means of a few simple experimental manipulations, the permeability barrier of the vesicles can be studied as an isolated functional entity. The data presented in Fig. 14 show the time course of methyl-a-D-glucopyranoside (a-MG) uptake and phosphorylation at 46 ° C in the presence of P-enolpyruvate. The data are expressed as a percentage of the concentration of a - M G added to the reaction mixtures, so that the amounts of a-MG and methyl-a-D-glycopyranoside Bioehim. Biophys. Acta, 265 (1972) 367-416

BACTERIAL CYTOPLASMIC M E M B R A N E TRANSPORT

403

phosphate (a-MGP) recovered from the vesicles can be directly compared with that recovered from the medium. As shown, uptake (a-MGP~,) increases rapidly for about 5 min, when it stops abruptly and begins to decrease. By 60 min, the vesicles lose over 60 ~ of the a-MGP accumulated in 5 min. The loss of a-MGP from the vesicles appears to result from increased membrane permeability at 46 ° C since vesicles incubated at lower temperatures are able to retain high intravesicular concentrations for at least 30 min 2~. As shown by the curve labeled "a-MGPo~t", a-MGP in the medium increases rapidly for the first 5 min of the incubation, slows down, and finally achieves an inverse relationship to a-MGP~. I00' I00'

90 80

80

70 A ,:x

60 ~-MG~t

50 40

ILl

70

~

60

~

50

\ \,,

./:

,,

u.l

46 ---270C

90

,~-ot-MGout

"o ~,~,,..,=- ~ ~ , / . ,,.

~

\x

z o_ 4 0 UPTAKE

30

""--

(~

\x \x A xx

.~ so

20

.

-*46°¢

27

zo

....

I0' .

I

I0

I

I

.

I

20 30 40 TIME (min)

.

.

.

o

I

I

50

60

,o

0

I

I

I

I

I

I

I0

20

:50

40

50

60

TIME (min)

Fig. 14. Time course of ct-MG uptake at 46 ° C by E. coil ML 308-225 membrane vesicles. Total uptake ( O - O ) , uptake of a-lZ4C]-methylglucoside by vesicles incubated in the presence of P-enol. pyruvate; a-MGPj~ Ok- - -Ak), a-MGP recovered from vesicles; a-MGln ( I I - III), a-MG recovered from vesicles; a-MGPout (~ -- - -- ~), a - M G P recovered from the medium in which the vesicles were suspended; a-MGout (D . . . . []), free a-methylglucoside recovered from the medium in which the vesicles were suspended. The experiment was carried out as described by Kaback 21. Fig. 15. Effect of temperature shifts on a-MG taken up by E. coil ML 308-225 membrane vesicles at 46 or 27 ° C. 46 ~ 27 o C, membrane vesicles were incubated with [~4C]a-methylglucoside in the presence of P-enolpyruvate at 46 ° C for 5 min (see Fig. 14). The samples were then immediately placed in a 27 ° C water bath, and the incubations were continued for the times indicated. A control sample left at 46 ° C for 60 min contained approximately 30% of the zero time activity (see Fig. 14). 27 -* 46 ° C, membrane vesicles were incubated with [~C]~-methylglucoside in the presence of P-enolpyruvate at 27 ° C for 30 rain. The sample was then centrifuged in the cold, and the supernatant was aspirated and discarded. The pellet was resuspended in fresh reaction mixture without a-[t'C]methyiglucoside, and samples were incubated at 46 ° C for the times shown. A control sample incubated at 27 o C for 60 rain lost less than 5 % of the activity of the zero time sample. From H. R. Kaback, J. BioL Chem., 243 (1968) 3711.

Biochim. Biophys. Acta, 265 (1972) 367-416

404

H.R. KABACK

In other words, a-MGPjn has a precursor-product relationship to a-MGPout. By 5 min, about 95 % of the free sugar added to the reaction mixtures has been phosphorylated (a-MGou0. Thus, phosphorylation (and transport) stops at 5 min because substrate becomes limiting; simultaneously, a - M G P in the vesicles is redistributed from a higher concentration in the intravesicular pool to a lower concentration in the medium. Despite the large quantity of a - M G P that appears in the medium at 60 min, the concentration of a - M G P is still about 25 times higher inside of the vesicles than it is in the medium. If the incubations are continued, the two compartments eventually equilibrate. The induction and reversal of the leak occur extremely rapidly (Fig. 15). When the vesicles are incubated with a - M G at 27 ° C and subsequently transferred to 46 ° C, there is an immediate decrease in the a - M G P retained by the vesicles. On the other hand, when the vesicles are incubated with a - M G at 46 ° C for 5 min and then shifted to 27 ° C, there is essentially no loss of a - M G P from the intravesicular pool. It should be emphasized that the temperature-induced leak described here is not "carrier" mediated as is the case for the D-lactate dehydrogenase-coupled systems 47. The rate of leakage of a - M G P as a function of intramembranal concentration exhibits nonsaturation kinetics with an infinite maximum velocity. These experimental manipulations allow a functional study of the barrier properties of the vesicles in the absence of transport. That is, the vesicles can be pre-loaded with a - M G P at 46 ° C until all of the transport substrate has been phosphorylated. At this point, the concen7.0

6.0

c

_~ 5.0 o o.

4.0 r, ~ 3.0

~

'

=~ 2.o,

I.O

I 0.02

I 0.04

I I ~ ' ; 0.06 0.08 0.10 0.12 PHOSPHOLIPASE A - B ( m g / r n l )

0.14

0.16

0.18

Fig. 16. a-MG uptake and a - M G P leakage as a function of mouse duodenal phospholipase A-B concentration. Membrane vesicles prepared from E. coli ML 308-225 were assayed for a-MG uptake (aMGPt.) as described previously t3'2~'3t'32. Vesicles were incubated at 46 ° C in the presence of P-enolpyruvate (0.1 M) and phospholipase A-B in the concentrations shown for 15 min before the addition of a-[t4C]-methylglucoside, a-MGP leakage (a-MGPout) was measured as described previously I a' 21.

Biochim. Biophys. Acta, 265 (1972) 367-416

BACTERIAL CYTOPLASMIC MEMBRANE TRANSPORT

405

tration of sugar-P in the intravesicular pool is at least 50 times higher than that of the medium 21. By lowering the temperature, leakage can be abolished, and the effect of various perturbants on the membrane can be studied. In the experiment shown in Fig. 16, a - M G uptake (a-MGPt,) and a - M G P leakage (a-MGPout) were studied as a function of phospholipase A - B 9a concentration. The membranes were incubated at 46 ° C for 15 min in the presence of the phospholipase concentrations shown, a - M G was added, and the incubations were continued for 7 min. As shown, the uptake of a - M G (a-MGPi.) decreases markedly with increasing phospholipase concentration and is essentially 0 at 0.I mg/ml. The appearance of a - M G P in the medium (a-MGPout), on the other hand, increases in a manner almost stoichiometrically equal to the loss in transport activity. The experiment shown in Fig. 17 demonstrates that phospholipase A - B specifically affects the membranes' ability to act as a barrier to a - M G P efflux. The data shown on the curve labeled " a - M G P efflux" was obtained by allowing the membranes to accumulate a - M G P at 46 ° C for 7 min, at which time the reaction mixtures were transferred to 30 ° C and phospholipase was added in the concentrations shown. There is a marked loss of a - M G P from the membranes as a function of increasing phospholipase concentration. The curve labeled " a - M G uptake" demonstrates that the effect of phospholipase A - B on a - M G uptake (in this case, as opposed to the experiment shown in Fig. 16, the reaction was carried out at 30 ° C, i.e., the same temperature at which a - M G P efflux was studied) and a - M G P efflux are very similar, if not identical. Finally, the data shown by the curve labeled "PE" demonstrate that the effect of

I00 \\

90

~

qx-MGP EFFLUX _~ A

80 70

8Q~ ~ I--

§

5o 40

50

/

2O

I0 0

I 0

I

I

I

I

I

0.1 0.2 0.5 PHOSPHOLIPASE A-B ( r n g / m t )

I

I

0.4

Fig. 17. a-MG uptake, a-MGP effiux, and hydrolysis of phosphatidylethanolaminc (PE) as a function of mouse phospholipase A-B concentration. The experiment shown was carried out at 30 ° C as described in the text.

Biochim. Biophys. Acta, 265 (1972) 367-416

406

H. R. KABACK

phospholipase A-B is directly related to the hydrolysis of membrane phosphatidylethanolamine (which constitutes approximately 75 ~ of the membrane phospholipid3,14.32). In the experiment shown, membranes prepared from cells grown in the presence of 32pi were incubated with increasing concentrations of the phospholipase under the same conditions as those used to study transport and efflux, and chromatographed directly 3'3z. It can be seen that there is a direct relationship between the loss of transport activity, the increase in a - M G P efflux, and the hydrolysis of phosphatidylethanolamine. These results not only demonstrate that the barrier functions of the membrane can be altered without affecting the activity of the P-enolpyruvate phosphotransferase system, but also imply that the phospholipids of the membrane, phosphatidylethanolamine in particular, are intimately related to the ability of the membrane to act as a barrier. Previous experiments 32 demonstrated that phosphatidylglycerol can be degraded without any effect on membrane barrier function. It is also noteworthy that D-lactate oxidation is not inhibited by treating vesicles with phospholipase A-B. A variety of other agents which are thought to have their primary effects on phospholipids (e.g., snake venom phospholipase A, surfactin - - a bacteriolytic substance produced by B. subtilis ~°2 - - acetone extraction, and a variety of detergents) also result in marked increases in membrane permeability with little or no effect on vectorial phosphorylation 14. Proteases (e.g., chymotrypsin, trypsin, or pronase) and protein fixatives (e.g., glutaraldehyde), on the other hand, have little or no effect on the leakage of a-MGP, despite the fact that these agents completely inhibit sugar and amino acid transport and cause marked changes in the membrane proteins as demonstrated by sodium dodecyl sulfate gel electrophoresis 9. ~4,zz. VB. Temperature transitions

The experiments presented in Fig. 18A demonstrate the effect of temperature on the initial rate of uptake of a-MG and the maximal level of accumulation of a - M G P by membrane vesicles incubated in the presence of P-enolpyruvate. The membranes were derived from E. coli M L 308-225 grown on minimal medium with glucose(Panel 1), glycerol (Panel 2), or succinate (Panel 3) as carbon sources. Although the data are not shown, membranes prepared from E. coli ML 308-225 grown on minimal medium with tryptone as a carbon source behave similarly to "succinate" membranes. With increasing temperature, the maximal level of a - M G P accumulated (uptake in 30 min) increases until it reaches an optimum at approximately 30, 40, or 46 ° C with "glucose", "glycerol", or "succinate" membranes, respectively. At temperatures exceeding these optima, the intramembranal level of a - M G P declines markedly. On the other hand, the initial rates of uptake (uptake in 3 or 5 min) in all three preparations display sharp optima at 46 ° C. It is also noteworthy that the increase in the initial rates of transport as a function of temperature in all three membrane preparations is very slow from 0 to approximately 20 ° C. Above 20 ° C, the initial rates of uptake increase much more Biochim. Biophys. Acta, 265 (1972) 367-416

407

BACTERIAL CYTOPLASMIC M E M B R A N E T R A N S P O R T

30 Min/ ,A' I

4 [ A. ~-MGPIN

/ I

'GLYCEROL'

_cos,

~

|

M;

30

30 Min/!~

5 Min

///

,./ 4 r B. ~-MGPou T

i

~.

I

~

o

30 Mln oOIo

3~i

=

"

#,go o

/ 30 MIn/=,

A

,i

3

-

/\

-

02/ 5Mt.~

2

=

0

I

0

.in~

C. o.-MGP LEAKAGE

=i :E

3o

I0 20 30 4 0 50

0

I0

i:"

-~.

/ I

I

20 30 40

50

0

I

I0

~1_1/

20 ~

~

50

TEMPERATURE (°C) Fig. 18. Effect of temperature on the initial rate of u-MG uptake and the level of a - M G P accumulation (A), the appearance of a - M G P in the medium (B), and the initial rate of a - M G P leakage (C) in E. coli M L 308-225 membrane vesicles. A • a-MGPI., nmoles a - M G taken up per mg of membrane protein in 3 rain ( • - - • ) , 5 rain ( • - - - • ) , or 30 rain ( O - - - © ) . B. a-MGPout, nmoles ct-MGP per mg membrane protein recovered from the filtrates of the samples shown in A. C. Leakage ( • -- 0 ) , m o l e s a - M G P lost per mg membrane protein compared to baseline samples, u - M G P in the filtrates (aMGPout) from the samples assayed in A (Fig. 18B) was measured chromatographically. For "Leakage", samples of "glucose", "glycerol", or "succinate" membrane vesicles were incubated in the presence of P-enolpyruvate and a-[t4C]-methylglucoside for 7, 10 and 25 rain, respectively. At these times, essentially all of the free a-[14C]-methyglucoside in each reaction mixture had been phosphorylated, but most of it was concentrated in the vesicles (see Fig. 16). The samples were then transferred to the temperatures shown for 15 rain and assayed. Baseline samples were assayed immediately after the 46 ° C incubations. The membrane vesicles used in these experiments were prepared from E. coli M L 308-225 grown on minimal medium with glucose, glycerol, or succinate as carbon sources. Although not shown, membrane vesicles prepared from cells grown on minimal medium with tryptone as carbon source behave qualitatively in a manner similar to that shown for "succinate" membrane vesicles. From E. Sbechter, T. Gulik-Krzywicki and H. R. Kaback, Biochim. Biophys. Acta, 274 (1972) 466.

Biochim. Biophys. Acta, 265 (1972) 367-416

408

H.R. KABACK

rapidly up to 46 ° C. A discontinuity at approximately 15 ° C has also been observed with D-lactate dehydrogenase activity and with the D-lactate dehydrogenase-coupled transport of lactose 47. Moreover, similar discontinuities have been observed more recently with the D-lactate dehydrogenase-coupled transport systems for sugars and amino acids*'**'***. The data presented in Fig. 18B represent a - M G P recovered from the filtrates of the reaction mixtures shown in Fig. 18A. The appearance of a - M G P in the medium bears an inverse relationship to the ability of the membranes to retain transported a-MGP, i.e., "glucose" membranes compared to "succinate" membranes accumulate less a-MGP, but more a - M G P is found in the external medium. "Glycerol" membranes represent an intermediate situation. Also, it should be clear from the data that the initial rates of phosphorylation, regardless of transport, are optimal at approximately 46 ° C in all three membrane preparations. a - M G P efflux (Fig. 18C) was also studied as a function of temperature by the methods outlined above (cf. Section VA, Functional separation of transportfrom barrier function). Thus, vesicles were pre-loaded with a - M G P at 46 ° C and the rate of a - M G P ettlux was measured at the temperatures shown. There is no significant leakage of a - M G P from 0 to just below 30, 40, or 46 ° C with "glucose", "glycerol", or "succinate" membranes, respectively. Above each of these temperatures, the rate of etttux in each preparation increases markedly and becomes maximal at about 55 ° C. The effects of incubation at temperatures up to and including 46 o C on accumulation and phosphorylation are completely reversible. Above 48 ° C, there is an irreversible inactivation of the phosphotransferase activity of the system. Therefore, the ability of the vesicles to accumulate high intramembranal concentrations of a - M G P as a function of temperature is determined by the temperature at which the leakage transition occurs in each preparation. Despite leakage at 46 ° C the rate of influx of a - M G P is considerably faster than its passive efflux, resulting in progressive accumulation of a - M G P only until phosphorylation (and transport) stops because free a - M G , the transport substrate, becomes rate-limiting (cf Fig. 15). Since P-enolpyruvate is present in excess, if additional free a - M G is added at this time, uptake resumes immediately (Fig. 19). The results shown in Figs 18A, B, and C are identical when glucose, fructose, or mannose, rather than a - M G , are used as transport substrates. It is also noteworthy that studies carried out with vesicles prepared from B. subtilis exhibit similar phenomena, and preliminary observations indicate that the critical leakage temperature of these membrane vesicles is correlated with the average chain length of the fatty acids in the membrane phospholipids 12' 14. With E. coli vesicles, on the other hand, there appears to be an increase in the amount of diphosphatidylglycerol (cardiolipin) which is correlated with increased leakage temperatures****. * F. J. Lombardi and H. R. Kaback, unpublished information. ** G. K. Kerwar and H. R. Kaback, unpublished observations. *** P. Bhattacharyya, F. J. Lombardi and H. R. Kaback, manuscript in preparation. **** J. Dittmer and H. R. Kaback, unpublished information.

Biochim. Biophys. Acta, 265 (1972) 367-416

B A C T E R I A L CYTOPLASMIC M E M B R A N E T R A N S P O R T

409

o'

0 0

I

1

I

I0

20

30

I

I

I

40 50 60 TIME (rain)

I

I

I

70

80

90

Fig. 19. a - M G uptake by E. coli ML 308-225 membrane vesicles incubated at 46 o C in the presence of P-enolpyruvate. Membrane vesicles were incubated at 46 ° C in the presence of 0.1 M P-enolpyruvate for 15 rain. a-[14C]-Methylgiucoside was then added (1) to give a final concentration of 3.64'10 -5 M, and samples were assayed at the times indicated. At 30 rain (2) and at 60 min (3), additional a-[14C]-methylglucoside was added to the reaction mixtures. Assays were carded out as described previously13,2 l,a z.32.

VC. Structural and functional correlations Over the past few years, various spectroscopic techniques have been applied to the study of biological membrane structure 9~'1°3-111. In most cases, however, little or no attempt was made to correlate these studies with biochemical processes in a well-defined biological membrane system. In view of the ability of highly purified membrane vesicles to catalyze the active transport of a large number of solutes, and the demonstration of at least two types of temperature-induced transitions in transportrelated functions, this system seems ideal for such structure-function investigations. The studies to be described have been initiated recently by Drs. E. Shechter and T. Gulik-Krzywicki of the Centre de G6n&ique Mol6culaire at Gif-sur-Yvette in collaboration with the author*. Using vesicles into which dansyl phosphatidylethanolamine was incorporated, changes in depolarization of fluorescence and fluorescence emission maxima were studied as functions of temperature and compared with changes in X-ray diffraction, transport, and passive permeability. The fluorescent probe was incorporated int o the vesicles by means of surface exchange via dansyl phosphatidylethanolamine-containing liposomes. Subsequently, the liposomes were separated from the membranes by means of discontinuous sucrose density centrifugation, and it was demonstrated that the fluorescence emission spectrum of dansyl phosphatidylethanolamine in the vesicles is markedly different from that of the dansyl phosphatidylethanolamine-containing

* E. Shcchter, T. Gulik-Krzywicki and H. R. Kaback, Biochim. Biophys. Acta, 274 (1972) 466.

Biochim. Biophys. Acta, 265 (1972) 367-416

410

H.R. KABACK

liposomes. Thus, the dansyl phosphatidylethanolamine is incorporated into the fabric of the membrane, and the results to be discussed cannot be due to the trivial attachment of dansyl phosphatidylethanolamine-containing liposomes to the vesicles. Moreover, it should be emphasized that membrane vesicles containing dansyl phosphatidylethanolamine manifest no significant differences in the transport activities studied. The P-enolpyruvate-dependent vectorial phosphorylation of a-MG, as well as the D-lactate-dependent transport of lactose and proline are essentially the same as that of control preparations. 1. X-ray diffraction studies. X-ray diffraction measurements carried out with centrifugally "stacked" vesicles at low temperatures reveal a sharp reflexion at 4.2~ superimposed on a broad band at 4.5 A. The 4.2-;~ reflexion is characteristic of rigid lipid paraffin chains, while the 4.5-,~ band is characteristic of "liquid-like" lipid paraffin chains 112'113. From 0 to approximately 15 ° C, the intensity of the 4.2-/~ reflexion remains unchanged; above 15 ° C, it begins to diminish in intensity; at 30 ° C, it can no longer be detected. Identical results are obtained with each of the membrane preparations shown in Fig. 18, indicating that the activity transitions observed at around 15 ° C (cf. Kaback and Barnes, Fig. 347, in addition to Fig. 18A, "5 min") may be related to the "melting" of the lipid paraffin chains of the membrane phospholipids. This conclusion is consistent with recent studies carried out with fatty acid auxotrophs of E. coli in other laboratories a 14-117. 2. Fluorescence depolarization studies. Since fluorescence depolarization reflects primarily the mobility of a fluorescent chromophore, variations in this parameter as a function of temperature using dansyl phosphatidylethanolamine were studied. Experiments were carried out initially with an artificial lipid-water system in order to demonstrate that fluorescence depolarization of dansyl phosphatidylethanolamine in a defined system could detect the "melt" transition discussed above. Thus, dansyl phosphatidylethanolamine was incorporated into a lecithin-water system which had been previously characterized by X-ray diffraction, and it was demonstrated that the transition observed in depolarization of fluorescence when studied as a function of temperature correlated with the X-ray diffraction changes. Subsequently, fluorescence depolarization measurements were made with dansyl phosphatidylethanolaminelabeled vesicles, and it was observed that "glucose" and "succinate" membranes both exhibit transitions at 15-20 ° C. These studies indicate that the "melting" of lipid paraffin chains of the membrane phospholipids as demonstrated by X-ray diffraction is reflected in the mobility of dansyl phosphatidylethanolamine incorporated into the vesicle membrane. 3. Studies of fluorescence emission maximum (2E,,ax). The fluorescence of dansyl chromophores is sensitive to the polarity of the micro-environment around the probe. The lower the polarity, the greater the displacement of the emission spectrum to the blue ~8. Moreover, previous studies have shown that the emission maximum of dansyl phosphatidylethanolamine incorporated into protein-lipid-water model systems is dependent on the type of lipid-protein interaction (i.e., hydrophobic versus electrostatic) in the system 1°9' 1lo Biochim. Biophys. Acta, 265 (1972) 367-416

BACTERIAL CYTOPLASMIC MEMBRANE TRANSPORT

411

The fluorescence of dansyl phosphatidylethanolamine incorporated into protein-lipid-water- systems displaying pure electrostatic interactions or electrostatic and hydrophobic interactions was studied. One system, m a d e up of lysozymephosphatidylinositol-water.exhibits electrostatic interactions at temperatures below 35-40 o C, and hydrophobic interactions at temperatures above 40 ° C, as demonstrated by X-ray diffraction 1°9'11o. This transition is accompanied by a change in ;tEmaxat approximately 40 ° C. A second system consisting of lysozyme--cardiolipin exhibits hydrophobic interactions over the temperature range studied 1°9'11°. No discontinuity is observed with the later system; moreover, the change in ;tEmax as a function of temperature is similar to that observed after the transition from electrostatic to hydrophobic interactions in the lysozyme-phosphatidylinositol-water system. With dansyl phosphatidylethanolamine-labeled membrane vesicles, studies of 2Emax as a function of temperature exhibit behavior that is remarkably similar to that of the lysozyme-phosphatidylinositol-water system described previously, that is, two linear functions with different slopes as a function of temperature yielding a sharp discontinuity at a particular temperature. It is important that the temperatures at which the discontinuities occur vary from one vesicle preparation to another (i.e., 30 ° C for "glucose" vesicles, 38 ° C for "succinate" vesicles, and 40 ° C for "tryptone" vesicles), and correlate reasonably well with the temperatures at which these membrane vesicles exhibit leakage transitions (cf. Fig. 18C). In view of the similarity between these results and those obtained with the artificial model systems, it is tempting to attribute the leakage transitions to temperature-dependent changes in the nature of the lipid-protein interactions within the membrane.

VI. CONCLUSIONS AND SPECULATIONS It has been the intent of this review to discuss some recent observations regarding mechanisms of bacterial solute transport and their relevance to studies of membrane structure. In particular, much of the discussion has dwelled on respiration-coupled active transport systems as studied in isolated bacterial membrane vesicle preparations. Since a number of these observations are surprising, and in some aspects may be contrary to established doctrine, a few of the more important general conclusions will be re-emphasized. Regarding the P-enolpyruvate phosphotransferase system, evidence has been presented which indicates that this system, although of great importance, is not of such wide-spread significance as was originally thought s. A short time ago, it seemed possible that such a mechanism, coupled possibly with specific sugar phosphate phosphohydrolases, could be of unique importance s.9,6 s; however, this view no longer seems tenable. In some bacterial species (e.g. Staphylococci), the phosphotransferase system appears to be involved in the transport of many sugars; however, these observations cannot be generalized. In E. coli, relatively few, albeit important, sugars are transported via this mechanism. On the other hand, most of the inducible sugar

Biochim. Biophys. Acta, 265 (1972) 367-416

412

H. R. KABACK

transport systems and almost all of the amino acid transport systems in E. coli are coupled to respiration by means of D-lactate dehydrogenase. Moreover, evidence has been cited which suggests that in many other bacterial species, respirationcoupled active transport as opposed to vectorial phosphorylation is the prevalent mechanism. Neither the generation nor utilization of high-energy phosphate compounds appears to be involved in respiration-coupled active transport. This conclusion is based upon a large number of experimental observations and appears to be well substantiated, at least in those organisms which have been studied in detail. Alternative means of redox coupling to transport such as proton or potential gradients have been considered and appear inconsistent with the data presented, although it should be pointed out that it is impossible to refute a negative hypothesis with positive experiments. Bhattacharyya et al. s5 demonstrated that the addition of valinomycin to membrane vesicles induces a highly specific active transport system for K ÷ or Rb +. Since the vesicles establish what appeared to be a proton gradient 9 i, it seemed likely that valinomycin-induced Rb ÷ transport was coupled to a proton gradient, and that a detailed study of this system might provide definitive evidence with regard to this question. Surprisingly, the valinomycin-induced transport of Rb ÷ behaves, with few exceptions, like the sugar and amino acid transport systems previously studied. Moreover, detailed ion transport studies as well as studies of pH and ANS fluorescence indicate that proton or potential gradients are not the primary driving force for transport, even for Rb ÷ transport induced by valinomycin. Although this possibility still cannot be excluded definitively, it should be obvious that if proton or potential gradients were the driving force for active transport in this system, the proton or ion "carriers" must have the same properties as the postulated solute-specific electron transfer intermediates in Fig. 10. That is, the proton or ion pumping mechanism must be coupled primarily to a specific dehydrogenase, and the proton or potential gradient that is established must respond to the same perturbants in the same manner as the sugar, amino acid, and valinomycin-induced transport systems (i.e., the generation of the proton or potential gradient must respond to temperature in the same way, and the proton or potential gradient that is established must collapse only in the presence of inhibitors which block electron transfer after the site of energy coupling). In other words, such a mechanism would appear to provide an unnecessary duplication in the system. On the other hand, a proton gradient mechanism would provide a convenient explanation for the action of "uncoupling agents ''aT. Although evidence has been presented which indicates that the "carriers" reflect the redox potential of the respiratory chain between the primary dehydrogenase and the cytochrome chain, the conceptual model presented in Fig. 10 is merely the simplest interpretation of the data which is presently available. No definitive evidence has been presented which demonstrates that the "carriers" are obligatory electron transfer intermediates. Moreover, the mechanism presented provides no immediate explanation for the action of "uncoupling agents" like 2,4-dinitrophenol, CCCP or azide. However, the ANS fluorescence studies discussed above indicate that these Biochim. Biophys. Acta, 265 (1972) 367-416

BACTERIAL CYTOPLASMIC M E M B R A N E T R A N S P O R T

413

compounds may interact directly with components in the membrane, and this interaction may ultimately explain their mechanism(s) of action. The observation that respiration-coupled transport is coupled to specific dehydrogenases is noteworthy if only because it was completely unexpected. Aside from the surprise element and from the obvious technical importance of this observation, this aspect of the mechanism places certain restrictions upon the means by which transport may be regulated. One can envisage at least three levels at which control could be exerted: at the level of the primary dehydrogenase; at the site of energy coupling; and at the level of the "carriers". Although little is known at the present time about the regulation of these transport systems, an important level of control might be at the primary dehydrogenase. A free living organism like a bacterium rarely finds itself in an environment containing a single sugar or amino acid. Thus, on a teleological basis, it might be expected that a primary mechanism would exist for regulating transport en bloc. By controlling the specific dehydrogenase involved with transport, this could be accomplished. Control at other levels of transport might be superimposed upon this basic means of regulation. In this regard, preliminary experiments with E. coli membrane vesicles indicate that many glycolytic intermediates, especially 2,3-diphosphoglycerate, inhibit D-lactate dehydrogenase. Thus, a cell growing on glucose as a carbon source might not transport other nutrients as effectively as a cell growing under different metabolic conditions. Another aspect of the respiration-coupled systems that has not been studied in detail thus far is their possible relevance to transport under anaerobic conditions. Obligate anaerobes or facultative anaerobes transport nutrients; moreover, &aminolevulinic acid- or heme-requiring mutants of E. coli do not apparently manifest transport defects although they have not been studied in detail from this point of view. Possibly anaerobically growing cells or cells lacking cytochromes utilize unique transport systems. However, the possibility should be considered that they may use the same general type of transport systems as those used aerobically, with the exception that an alternative electron acceptor is used rather than the cytochrome chain and oxygen. Recent experiments carried out in this laboratory demonstrate that lactose or proline transport by aerobically grown E. coli is markedly inhibited by anoxia. On the other hand, anaerobically grown E. coli transport equally well under anoxic or aerobic conditions. Thus, something appears to be induced under anaerobic growth conditions which allows the cells to transport in the absence of oxygen. It is of considerable interest that preliminary experiments with membrane vesicles prepared from these cells partially reflect the transport properties of the intact cells. Moreover, although the fumarate reductase or nitrate reductase systems would appear to be possible candidates for such a role, neither of these systens is used for this purpose. It is also noteworthy that all attempts to drive transport in the vesicle system under anoxic conditions with ATP have been uniformly negative. Rather than terminate this review on a purely speculative note, it seems pertinent to discuss some very recent experiments which may have significance for future progress in this area. Dr. Jen-Shiang Hong, who recently joined the author's laboraBiochim. Biophys. Acta, 265 (1972) 367-416

414

H. R. KABACK

tory, has isolated a n d characterized a n u m b e r o f p r i m a r y d e h y d r o g e n a s e m u t a n t s which manifest interesting properties. P r i m a r y D-lactate d e h y d r o g e n a s e m u t a n t s were selected based on their ability to grow on p y r u v a t e but n o t D-lactate. The whole cells are able to t r a n s p o r t lactose or proline normally, as might be expected since o t h e r d e h y d r o g e n a s e s can drive t r a n s p o r t to s o m e extent. M e m b r a n e vesicles p r e p a r e d f r o m these mutants, however, exhibit some unexpected properties. In the presence o f a s c o r b a t e - P M S , vesicles f r o m the m u t a n t t r a n s p o r t p r o l i n e or lactose at the same rate a n d to the same extent as vesicles p r e p a r e d from the wild type. D-lactate, as expected, does n o t stimulate t r a n s p o r t by m u t a n t vesicles. L-Lactate a n d / o r succinate (depending u p o n the c a r b o n source used for the g r o w t h o f the cells) stimulates t r a n s p o r t in m u t a n t vesicles as well as D-lactate in the wild type. Even more striking is the o b servation that the rate o f o x i d a t i o n o f L-lactate or succinate is similar to that observed in vesicles p r e p a r e d from the wild type. In other words, in the absence o f D-lactate d e h y d r o g e n a s e activity, the coupling o f L-lactate d e h y d r o g e n a s e or succinate dehydrogenase to t r a n s p o r t is enhanced. W i t h vesicles p r e p a r e d from a m u t a n t defective in b o t h D - l a c t a t e d e h y d r o g e n a s e and succinate dehydrogenase, only L-lactate dehydrogenase is c o u p l e d to transport. A p p a r e n t l y the cell will go to great lengths to protect this aspect o f its metabolism. Hopefully, this genetic a p p r o a c h , c o m b i n e d with enzymologic a n d physical techniques currently in use will extend the study o f m e m b r a n e t r a n s p o r t to the m o l e c u l a r level.

REFERENCES 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22

H. R. Kaback, Fed. Proc. 19 (1960) 130. H. R. Kaback and A. B. Kostellow, J. Biol. Chem., 243 (1968) 1384. H. R. Kaback and E. R. Stadtman, J. Biol. Chem., 243 (1968) 1390. H. R. Kaback and E. R. Stadtman, Proc. Natl. Acad. Sci. U.S., 55 (1966) 920. H. R. Kaback and T. F. Deuel, Arch. Biochem. Biophys., 132 (1969) 118. L. A. Heppel, Science, 156 (1967) 1451. A. B. Pardee, Science, 162 (1968) 632. S. Roseman, J. Gen. Physiol., 54 (1969) 1385. H. R. Kaback, Annu. Rev. Biochem., 39 (1970) 561. E. C. C. Lin, Annu. Rev. Genetics, 4 (1970) 225. D. L. Oxender, Metabolic Pathways, Vol. 6, Academic Press, New York, in the press. D. L. Oxender, Biomembranes, Vol. 3, Plenum Press, New York, in the press. H. R. Kaback, in W. B. Jakoby, Methods in Enzymology, Vol. 22, Academic Press, New York, 1971, 19.99. H. R. Kaback, in A. Kleinzeller and F. Brormer, Current Topics in Membranes and Transport, Vol. 1, Academic Press, New York, 1970, p. 35. M. R. J. Salton, The Bacterial Cell Wall, Elsevier, Amsterdam, 1964. W. Kundig, S. Ghosh and S. Roseman, Proc. Natl Acad. Sci.U.S., 52 (1964) 1067. B. Anderson, N, Weigel, W. Kundig and S. Roseman, J. Biol. Chem., 246 (1971) 7023. W. Kundig and S. Roseman, J. BioL Chem., 246 (1971) 1407. R. D. Simoni, M. F. Smith and S. Roseman, Biochim. Biophys. Res. Commun., 31 (1968)804. W. Hengstenberg, W. K. Penberthy, K. L. Hill and M. L. Morse, J. Bacteriol., 96 (1968) 2187. H. R. Kaback, J. BioL Chem., 243 (1968) 3711. H R. Kaback, E. M. Barnes, Jr and L. S. Milner, in A. P6rez-Miravete and D. Pel6az, Recent Advances in Microbiology, Asociacion Mexicana de Microbiologia, Mexico City, 1971, p. 171.

Biochim. Biophys. Acta, 265 (1972) 367-416

BACTERIAL CYTOPLASMIC M E M B R A N E T R A N S P O R T

415

23 H. R. Kaback, in D. C. Tosteson, The Molecular Basis of Membrane Transport, Brentice-Hall, Engle~vood Cliffs, N. J., 1969, p. 421. 24 J. B. Egan and M. L. Morse, Biochim. Biophys. Acta, 97 (1965) 310. 25 J. B. Egan and M. L. Morse, Biochim. Biophys. Acta, 109 (1965) 172. 26 J. B. Egan and M. L. Morse, Biochim. Biophys. Acta, 112 (1966) 63. 27 W. Henstenberg, J. B. Egan and M. L. Morse, J. BioL Chem., 243 (1968) 1881. 28 D. Laue and R. E. MacDonald, Biochim. Biophys. Acta, 165 (1968) 410. 29 S. Tanaka, S. A. Lerner and E. C. C. Lin, J. Bacteriol., 93 (1967) 64. 30 A. H. Romano, S. J. Eberhard, S. L. Kingle and T. D. McDowell, J. Bacteriol., 104 (1970) 808. 31 H. R. Kaback, Proc. Natl. Acad. Sci. U.S., 63 (1969) 724. 32 L. S. Milner and H. R. Kaback, Proc. Natl Acad. Sci. U.S., 65 (1970) 683. 33 H. R. Kaback and L. S. Milner, Proc. Natl Acad. Sci. U.S., 66 (1970) 1008. 34 H. R. Kaback, E. M. Barnes, Jr ,A. S. Gordon, F. J. Lombardi and G. K. Kerwar, 1st Meeting Pan Am. Biochem. Soc., Caracas, 1971, in the press. 35 E. M. Barnes, Jr and H. R. Kaback, Proc. Natl Acad. Sci. U.S., 66 (1970) 1190. 36 E. M. Barnes, Jr and H. R. Kaback, J. Biol. Chem., 246 (1971) 5518. 37 G. W. Dietz, Fed. Proc., 30 (1971) 52. 38 G. F. Ames, Arch. Biochem. Biophys., 105 (1964) 1. 39 J. Piperno and D. L. Oxender, J. Biol. Chem., 243 (1968) 5914. 40 G. A. Scarborough, M. K. Rumley and E. P. Kennedy, Proc. NatlAcad. Sci.U.S., 60 (1968) 951. 41 E. Pavlasova and F. M. Harold, J. Bacterk~l., 98 (1969) 198. 42 C. F. Fox and E. P. Kennedy, Proc. Natl Acad. Sci. U.S., 54 (1965) 891. 43 C. F. Fox, J. R. Carter and E. P. Kennedy, Proc. Natl Acad. Sci. U.S., 57 (1967) 891. 44 W. Kundig, F. D. Kundig, B. E. Andersen and S. Roseman, J. Biol. Chem., 241 (1966) 3243. 45 S. Tanaka, P. G. Fraenkel and E. C. C. Lin, Biochem. Biophys. Res. Commun., 27 (1967) 63. 46 G. K. Kerwar, A. S. Gordon and H. R. Kaback, J. Biol. Chem., 247 (1972) 298. 47 H. R. Kaback and E. M. Barnes, Jr, J. Biol. Chem., 246 (1971) 5523. 48 B. L. Horecker, J. Thomas a n d J . Monod, J. Biol. Chem., 235 (1960) 1580. 49 H. C. P. Wu, J. Mol. Biol., 24 (1967) 213. 50 A. K. Ganesan and B. Rotman, J. Mol. Biol., 16 (1966) 42. 51 E. M. Tarmy and N. O. Kaplan, J. Biol. Chem., 243 (1968) 2597. 52 E. M. Tarmy and N. O. Kaplan, J. Biol. Chem., 243 (1968). 2587. 53 E. S. Kline and H. R. Mahler, Ann. N.Y. Acad. Sci., 119 (1965) 105. 54 G. B. Cox, N. A. Newton, F. Gibson, A. M. Snoswell and J. A. Hamilton, Biochem. J., 117 (1970) 551. 55 H. Weissbach, E. L. Thomas and H. R. Kaback, Arch. Biochem. Biophys., 147 (1971) 249. 56 W. L. Klein, A. S. Dhams and P. D. Boyer, Fed. Proc., 29 (1970) 341. 57 A. Kepes, C. R. Acad. Sci. Paris, 244 (1957) 1550. 58 A. Kepes and G. N. Cohen, in I. C. Gunsalis and R. Stanier, The Bacteria, Vol. 4, Academic Press, New York, 1962, p. 179. 59 A. Kepes, in A. Kleinzeller and F. Bronner, Current Topics in Membranes and Transport, Vol.1, Academic Press, New York, 1970, p. 101. 60 A. Kepes, Biochim. Biophys. Acta, 40 (1960) 70. 61 J. R. Carter, C. F. Fox and E. P. Kennedy, Proc. Natl Acad. Sci. U.S., 60 (1968) 725. 62 T. H. D. Jones and E. P. Kennedy, J. Biol. Chem., 244 (1969) 5981. 63 D. Schachter and A. J. Mindlin, J. BioL Chem., 244 (1969) 1808. 64 J. A. Manno and D. Schachter, J. Biol. Chem., 245 (1970) 1217. 65 A. L. Koch, J. Mol. Biol., 59 (1971) 447. 66 P. T. S. Wong, E. R. Kashket and T. H. Wilson, Proc. Natl Acad. ScL U.S., 65 (1970) 63. 67 T. H. Wilson, M. Kusch and E. R. Kashket, Biochem. Biophys. Res. Commun., 40 (1970) 1409. 68 W. N. Konings, E. M. Barnes, Jr and H. R. Kaback, J. Biol. Chem., 246 (1971) 5857. 69 I. G. O'Brien and F. Gibson, Biochim. Biophys. Acta, 215 (1970) 393. 70 P. D. Bragg, Can. J. Biochem., 48 (1970) 777. 71 A. S. Gordon, F. J. Lombardi and H. R. Kaback Proc. Natl Acad. Sci. U.S., 69 (1972) 358. 72 H. C. Neu and L. A. Heppel, J. Biol. Chem., 240 (1965) 3685. 73 S. A. Short, D. C. White and H. R. Kaback, J. Biol. Chem., 247 (1972) 298. 74 W. N. Konings and E. Freese, FEBS Lett., 14 (1971) 65.

Biochim. Biophys. Acta, 265 (1972)367-416

416

la. R. KABACK

75 M. M. Shemyakin, N. A. Aldanova, E. I. Vinogradova and M. Yu Feigina, Tetrahedron Lett., issue 28 (1963) 1921. 76 M. M. Shemyakin, Yu A. Ovichinnikov, V. T. Ivanov, V. K. Antonov, E. I. Vinogradova, A. M. Shkrob, G. G. Malenkow, A. V. Eustatov, I. A. Laine, E. I. Melnik and I. D. Ryabova, J. Membrane Biol. 1 (1969) 402. 77 F. M. Harold, Adv. Microbiol. PhysioL, 4 (1970) 45. 78 P. F. J. Henderson, J. D. McGivan and J. B. Chappell, Biochem. J., 111 (1969) 521. 79 P. Mueller and D. O. Rudin, Biochem. Biophys. Res. Commun., 26 (1967) 398. 80 F. M. Harold and J. R. Baarda, J. Bacteriol., 94 (1967) 53. 81 F. M. Harold and J. 1~. Baarda, J. Bacteriol., 95 (1968) 816. 82 J. B. Chappell and A. R. Crofts, Biochem. J., 95 (1965) 393. 83 D. C. Tosteson, T. E. Andreoli, M. Tiefenberg and P. Cook, J. Gen. PhysioL, 51 (1968) 3735. 84 B. C. Pressman, E. J. Harris, W. S. Jagger and T. H. Johnson, Proc. Natl Acad. Sci. U.S., 58 (1967) 1949. 85 P. Bhattacharyya, W. Epstein and S. Silver, Proc. Nail Acad. Sci. U.S., 68 (1971) 488. 86 P. Mitchell, Fed. Proc., 26 (1967) 1370. 87 P. Mitchell, Biol. Rev., 41 (1966) 445. 88 G. D. Greville, Current Topics in Bioenergetics, 3 (1971) 1. 89 F. M. Harold and J. R. Baarda, J. Bacteriol., 98 (1968) 2025. 90 I. C. West, Biochem. Biophys. Res. Commun., 41 (1970) 655. 91 J. P. Reeves, Biochem. Biophys. Res. Commun., 45 (1971) 931. 92 W. J. Waddel and R. G. Bates, Physiol. Rev., 49 (1969) 285. 93 A. Ottolenghi, Fed. Proc., 28 (1969) 885. 94 E. Daniel and G. Weber, Biochemistry, 5 (1966) 1893. 95 G. M. Edelman and W. O. McClure, Acc. Chem. Res., 1 (1968) 65. 96 B. Rubalcava, D. Martinez de Mufioz and C. Gitler, Biochemistry, 8 (1969) 2742. 97 B. Chance, G. K. Radda and C. P. Lee, Proc. Natl. Acad. Sci. U.S., 62 (1969) 612. 98 R. B. Freedman and G. K. Radda, FEBSLett., 3 (1969) 150. 99 J. Vanderkooi and A. Martinosi, Arch. Biochem. Biophys., 133 (1969) 153. 100 J. R. Brokelhurst, R. B. Freedman, D. J. Hancock and G. K. Radda, Biochem. J., 116 (1970) 721. 101 A. Azzi, P. Gherardini and H. Santato, J. BioL Chem., 246 (1971) 2035. 102 K. Aruma, N. Tsukagoshi and G. Tamura, Biochim. Biophys. Acta, 163 (1968) 121. 103 J. Lenard and S. J. Singer, Proc. Natl. Acad. Sci. U.S., 56 (1966) 1828. 104 D. F. H. Wallach and P. J. Zahler, Proc. Natl. Acad. Sci. U.S., 56 (1966) 1552. 105 W. L. Hubbel and H. M. McConnel, Proc. NatL Acad. Sci. U.S., 61 (1968) 12. 106 D. Chapman and D. F. H. Wallach, in D. Chapman, Biological Membranes, Academic Press, New York, 1968, p. 125. 107 D. F. H. WaUach, J. M. G r a h a m and B. R. Fernbach, Arch. Biochem. Biophys., 131 (1969) 322. 108 T. Gulik-Krzywicki, E. Shechter, M. Iwatsubo, J. L. Ranck and V. Luzzati, Biochim. Biophys. Acta, 219 (1970) 1. 109 T. Gulik-Krzywicki, E. Shechter, V. Luzzati and M. Faure, Nature, 223 (1969) 1116. 110 E. Shechter, T. Gulik-Krzywicki, R. Azerad and C. Gros, Biochim. Biophys. Acta, 241 (1971) 431. 111 T. Gulik-Krzywicki, E. Shechter, V. Luzzati and M. Faure, Proc. Int. Symp. Biochem. Biophys. Mitochondrial Membranes, Academic Press, New York, in the press. 112 V. Luzzati, in D. Chapman, Biological Membranes, Academic Press, New York, 1968, p. 71. 113 M. H. F. Wilkens, A. E. Blaurock and D. M. Engelman, Nature, 230 (1971) 42. 114 H. U. Schairer and P. Overath, J. MoL Biol., 44 (1969) 209. 115 C. F. Fox, Proc, Natl. Acad. Sci. U.S., 63 (1969) 850. 116 U. Henning, G. Dennert, K. Rehn and G. Deppe, J. Bacteriol., 98 (1969) 784. 117 M. Esfahani, E. M. Barnes, Jr and S. Wakil, Proc. NatL Acad. Sci. U.S., 64 (1969) 1057. 118 R. F. Chen, Arch. Biochem. Biophys., 120 (1967) 609. 119 J. Hochstadt-Ozer and E. R. Stadtman, J. BioL Chem., 246 (1971) 5304. 120 S. D. Cosloy and E. McFall, BacterioL Proc., P220 (1971) 160. 121 J. Guardiola and M. Iaccarino, J. BacterioL, 108 (1971) 1034. 122 J. D. Butlin, G. B. Cox and F. Gibson, Biochem. J., 124 (1971) 75. 123 H. Hirata, A. Asano and A. F. Brodie, Biochem. Biophys. Res. Commun., 44 (1971) 368. 124 L. Stryer, Science, 162 (1968) 526.

Biochim. Biophys. Acta, 265 (1972) 367-416