Mechanisms of active transport in isolated bacterial membrane vesicles. Further studies on amino acid transport in Staphylococcus aureus membrane vesicles.

Abstract Active amino acid transport in Staphylococcus aureus U-71 membrane vesicles is coupled to either α-glycerol phosphate dehydrogenase or l-lactate dehydrogenase depending upon the growth conditions of the parent cells. Vesicles prepared from cells grown on gluconate as a primary carbon source exhibit an absolute specificity for α-glycerol phosphate as a physiological electron donor for transport, whereas vesicles prepared from cells grown on glucose as a primary carbon source exhibit an absolute specificity for l-lactate as an electron donor for transport. Both preparations exhibit similar dehydrogenase activities qualitatively, indicating that the coupling between these dehydrogenases and transport is altered. l-Lactate oxidation, l-lactate:dichlorophenolindophenol reductase activity, and l-lactate-dependent amino acid transport exhibit similar apparent Michaelis constants with respect to l-lactate, indicating that l-lactate oxidation per se is the rate-limiting step for amino acid transport in the appropriate membrane preparation. Amino acid transport is dependent on electron transfer, and inhibition of l-lactate oxidation by anaerobiosis, cyanide, 2-heptyl-4-hydroxyquinoline-N-oxide, amytal, and oxalate is directly related to inhibition of amino acid transport. However, only anaerobiosis, cyanide, 2-heptyl-4-hydroxyquinoline-N-oxide, and amytal, each of which inhibits electron transfer after the site of energy coupling, cause efflux. Oxalate, a potent inhibitor of l-lactate dehydrogenase, does not cause efflux despite almost complete inhibition of l-lactate oxidation and amino acid transport. Moreover, oxalate blocks or inhibits efflux caused by each of the other inhibitors and by 2,4-dinitrophenol. These results provide further evidence that active transport is dependent on the oxidation-reduction potential of the respiratory chain at the site of energy coupling. Cyanide-induced efflux is a saturable process with an apparent affinity constant that is approximately 500 times higher than the affinity constant for active transport. The apparent maximum velocity of efflux, on the other hand, is the same as that of active transport. These findings suggest that one of the primary effects of energy coupling is to change the affinity of the carrier for substrate. Under anaerobic conditions serine uptake exhibits linear kinetics, indicating that the rate-limiting step for serine uptake under these conditions is a nonsaturable process with an infinite Km. Moreover, approximately 5 min is required for external serine to equilibrate with the intramembranal pool at a variety of concentrations. Thus, it is highly unlikely that facilitated diffusion is the rate-limiting step for active serine uptake.

that one of the primary effects of energy coupling is to change the affinity of the carrier for substrate.
Under anaerobic conditions serine uptake exhibits linear kinetics, indicating that the rate-limiting step for serine uptake under these conditions is a nonsaturable process with an infinite K,,,. Moreover, approximately 5 min is required for external serine to equilibrate with the intramembranal pool at a variety of concentrations.
Thus, it is highly unlikely that facilitated diffusion is the rate-limiting step for active serine uptake.
cr-Glycerol-P dehydrogenase-coupled amino acid uptake by Staphylococcus aureus membrane vesicles is catalyzed by 12 distinct and specific transport systems for structurally related amino acids, and the activity of the vesicles is comparable to the transport activity of intact cells (1, 2). Evidence has also been presented which demonstrates that except for differences in physiological electron donors, the characteristics of the transport system in S. uureus vesicles are very similar to those described for the Escherichia co2i vesicle system (l-4).
In vesicles prepared from both organisms, the site of energy coupling between respiration and active amino acid transport is localized in a segment of the respiratory chain between the primary dehydrogenases (i.e. a-glycerol-P dehydrogenase in S. aureus and n-lactate dehydrogenase in E. coli) and the cytochrome chain. In addition, evidence obtained with both systems indicates that the coupling between respiration and transport does not involve the generation or utilization of high energy phosphate or ATP (1, 3-5).
This paper deals with a further characterization of the amino acid transport systems in S. uureus membrane vesicles. 4276 scribed above are referred to as "gluconate" vesicles, while those prepared from cells grown on the glucose-containing medium are referred to as "glucose" vesicles. Measurement of Amino Acid Uptake and E&z-Transport of amino acids was assayed as described previously (1,2,8). For efflux experiments, vesicles were incubated in the presence of either a-glycerol-P or L-lactate and an appropriate radioactive amino acid until a steady state level of accumulation was achieved (5 to 10 min).
At this time, the indicated electron transfer inhibitor was added or the reaction mixtures were gassed with argon as described previously (8,9). At an appropriate time, the reactions were terminated and the samples were assayed as described previously (8,10 were reagent grade and were obtained from commercial sources.

Amino
Acid Transport by S. aureus Jlembrane Vesicles-Although amino acid transport in S. uureus has been reported to be coupled specifically to the oxidation of a-glycerol-P to dihydroxyacetone-P, the specificity of transport for this dehydrogenase may be completely altered by varying the growth conditions of the parent cells (Fig. 1). As shown previously (1, 2) and in Fig. lA, when vesicles are prepared from cells grown on a medium with gluconate as the primary carbon source (i.e. gluconate membranes), addition of cY-glycerol-P to the reaction mixtures results in a dramatic increase in the initial rate and steady state level of serine accumulation, while addition of L-lactate has no effect.
On the other hand, with vesicles prepared from cells grown on glucose as the primary carbon source (i.e. glucose membranes), the reverse is obtained (Fig. 1B). Addition of L-lactate, but not cr-glycerol-P, results in marked stimulation of both the initial rate and steady stat.e level of serine accumulation. Data qualitatively identical with those shown for serine were also obtained for alanine, leucine, threonine, lysine, glutamic acid, and proline. As shown previously, serine and the other amino acids accumulated by S. aureus membrane vesicles can be quantitatively recovered in a structurally unaltered form (1) (16), the acetylenic hydroxy acid, nc2-hydroxy-3-butynoic acid, irreversibly inactivates n-lactate-dependent oxygen uptake, L-1actate:dichlorophenolindophenol reductase activity, and n-lactate-dependent amino acid transport in S. aureus glucose vesicles.
Dependence of gluconate membranes on cr-glycerol-P as a physiological electron donor for amino acid transport is highly specific, and other electron donors and energy sources do not stimulate amino acid transport (1). Similarly, glucose membranes exhibit a high degree of specificity for n-lactate (Table I). The requirement for a specific dehydrogenase substrate in each of these vesicle preparations does not reflect the absence of other dehydrogenases.
As shown in Table II, gluconate and glucose vesicles exhibit the same dehydrogenase activities qualitatively, although their relative specific activities vary to some extent.
In particular, glucose vesicles catalyze the oxidation of a-glycerol-P at 30 to 50% the rate of gluconate membranes, although o-glycerol-P does not stimulate amino acid uptake. Moreover, the rate of n-lactate oxidation is almost identical in both preparations, yet L-lactate is an effective electron donor for transport in glucose membranes only (cf. Fig. 1). It is also noteworthy that NADH, although a better electron donor than either a-glycerol-P or L-lactate in both vesicle preparations, does not support transport in this system (cf . Tables I and II  and Ref. 1).
The serine-threonine, alanine-glycine, leucine-isoleucine-valine, lysine, glutamate-aspartate, and proline transport systems (2)  A difference is observed, however, between gluconate and glucose membranes in the stimulation of transport by the artificial electron donor system, ascorbate-phenaeine methosulfate (8,9)) relative to the physiological electron donors, a-glycerol-P and L-lactate. As shown in Table III, amino acid uptake by gluconate vesicles incubated in the presence of ascorbate-phenazine methosulfate is much less than that observed in the presence of a-glycerol-P.
On the other hand, amino acid uptake in glucose membranes incubated in the presence of ascorbate-phenazine methosulfate is considerably greater than that observed in the presence of L-lactate. Relationship between Electron Transfer and Amino Acid Uptake and E@Zuz-n-Lactate-dependent serine uptake by glucose membranes is markedly inhibited by anaerobiosis and by the electron transfer inhibitors oxalate, amytal, 2-heptyl-4-hydroxyquinoline-N-oxide, and cyanide (Fig. 3). Moreover, as shown in Table  IV, inhibition of transport by these inhibitors is directly related to inhibition of n-lactate oxidation.
Regarding the sites of action of the inhibitors in the respiratory chain, previous studies (17, 18) have shown that cyanide inhibits cytochrome a, 2-heptyl-4-hydroxyquinoline-N-oxide inhibits between cytochromes b and a, and amytal inhibits at the flavin level.' O&ate is a potent competitive inhibitor of n-lactate dehydrogenase with an apparent Ki of approximately 20 pM (data not shown). As demonstrated by the spectrophotometric measurements presented in Fig. 4, addition of oxalate to anaerobic vesicle suspensions prior to aeration results in re-oxidation of at least 85% of the reduced cytochrome b + o and reduced cytochrome a (compare Spectrum ZZZ to II).
These results dem0nstrat.e that the respiratory chain is maintained in an oxidized state in the presence of oxalate.
With the exception of oxalate, the same electron transfer inhibitors and anaerobiosis produce similar degrees of inhibition of o-glycerol-P-dependent amino acid uptake in gluconate vesicles.
Oxalate, as expected, has no effect on o-glycerol-P-de-' The studies referred to were carried out with E. coli; however, similar experiments with S. aureus U-71 membrane vesicles indicate that these respiratory inhibitors act at similar sites in this system.
Thus, addition of amytal to anaerobic vesicle suspensions prior to aeration results in re-oxidation of approximately 85yo of the reduced cytochrome 6 + o and cytochrome a; addition of 2-heptyl4-hydroxyquinoline-N-oxide results in re-oxidation of reduced cytochrome a only; and addition of cyanide does not result in re-oxidation of any of the cytochrome pigments. Identical results were obtained with gluconate vesicles except that potassium oxalate did not inhibit amino acid uptake by these vesicles.
FIG . 4 (center). Difference spectra of StaphyZococcus aureus glucose membrane vesicles. Membrane vesicle suspensions containing about 1 mg of membrane protein per ml in l-cm cuvettes were examined in the Cary 15 spectrophotometer at 25". L-Lactate was added to one cuvette, and after the anaerobic steady state was achieved, difference spectra were recorded.
I, difference spectrum of two suspensions in the oxidized state; II, difference spec-  or a-glycerol-P oxidation in gluconate vesicles nor on ascorbate-phenazine methosulfate-dependent transport in either preparation. It is also noteworthy that essentially the same results were obtained when alanine, leucine, proline, lysine, and aspartic acid were used as transport substrates in both vesicle preparations.
Anaerobiosis, amytal, 2-heptyl-4-hydroxyquinoline-N-oxide, and cyanide produce rapid efflux of serine from glucose vesicles preloaded in the presence of L-lactate (Fig. 5) Addition of sodium dithionite produced the same difference spectrum as that shown in ZZ.
FIQ. 5 (right). Effect of the inhibition of electron transfer on the steady state level of serine accumulation.
Reaction mixtures containing glucose vesicles were prepared and assayed as described in Fig. 1 from the intravesicular pool. Oxalate does not inhibit the exchange of serine across the vesicle membrane ( Fig. 6), indicating that it does not have a direct effect on the carriers per se. Since vesicles incubated in the presence of oxalate do not catalyze net uptake or efflux of serine, it seems likely that the exchange reaction observed in the presence of oxalate is limited to a one-for-one interchange between internal and external serine.* Qualitatively similar results were obtained when these experiments were carried out with alanine, leucine, proline, lysine, or aspartic acid, and with the /3-galactoside (18), amino acid (19) Reaction mixtures (50 ~1 total volume) were prepared as described in Fig. 1 As shown in Fig. 7, this prediction is borne out.
Addition of oxalate prior to cyanide, amytal, or gassing with argon virtually abolishes serine efflux induced under the latter conditions (Fig. 7A). With 2-hepty%hydroxyquinoline-N-oxide (Fig. 7B), oxalate markedly inhibits the initial rate of efflux, but the effect is not so great as that observed with cyanide, amytal, or anaerobiosis. Finally, as shown in Fig. 7C, oxalate also markedly inhibits efflux in the presence of the uncoupling agent 2,4-dinitrophenol. These experiments provide further support for the hypothesis that carrier activity refiects primarily the oxidation-reduction potential of the energycoupling site for transport.
Kinetics of Amino Acid Uptake and Z@lux-In the mechanism proposed for respiration-dependent transport by B. coli membrane vesicles (3,4,18), the affinity of the carriers is determined by the oxidation-reduction state of the respiratory chain at the site of energy coupling. cr-Glycerol-P-or L-lactate-dependent amino acid upt,ake by S. aureus membrane vesicles is a high affinity process, exhibiting apparent Michaelis constants in the micromolar range.
With serine, specifically, the apparent K, is 10 to 12 PM and the VI,,,, 5 to 6 nmoles per min per mg of membrane protein (2). Under anaerobic conditions, however, where active serine accumulation and L-lactate (or cu-glycerol-P) oxidation are markedly inhibited, the kinetics of serine uptake are strikingly altered (Figs. 8 and 9). As shown in Fig. 8, when the rate and extent of serine uptake are measured under argon, approximately 5 min is required for external serine to equilibrate with the intramembranal pool at serine concentrations ranging from approximately 0.26 to 6.9 mM. A reciprocal plot of the initial rates at each serine concentration is given in Fig. 9. Clearly, the data yield a linear function intersecting the z and y axes at the origin, suggesting that the ratelimiting step for serine uptake under these conditions is a nonsaturable process with no K,. The vesicle suspensions were gassed with argon for 3 min, n-lactate was added to a final concentration of 20 mM, and the incubation was continued under argon for an additional 10 min at 25".
[UJ4C]Serine (25.7 mCi per mmole) was then added to the reaction mixtures to give final concentrations ranging from 0.26 to 6.91 mM as shown. At the times indicated, the reactions were terminated as described previously (1,2) with the exception that 5 ml of 0.1 M LiCl wash and 47-mm cellulose-nitrate filters were used. The filters were dried, dissolved in 10 ml of Instabray scintillator (Yorktown Research, New York, New York), and counted in a Beckman liquid scintillation counter. The broken line in each panel represents the serine concentration of the intravesicular pool at equilibration. FIG. 9. Kinetics of serine uptake under anaerobic conditions. Reciprocal initial rates of serine uptake at each serine concentration were calculated from the data given in Fig. 8. The results were then plotted by the method of Lineweaver and Burk (21).
When cyanide is added to vesicles which have been loaded to various intravesicular serine concentrations by incubation with either cu-glycerol-I' (gluconate membranes) or L-lactate (glucose membranes), the rates of efflux observed exhibit saturation with respect to the internal serine concentration (Fig. 10). The apparent K, values for serine efflux calculated from Hofstee plots (insets in Fig. 10)  After 10 min, one set of control samples was assayed.
Potassium cyanide (20 mM final concentration) was then added to another set of identical samples and the incubations were continued at 25" for 30 s. The reactions were then terminated and the samples were assayed as described previously (1,2,8). The differences between the control samples and the samples incubated for another 30 s in the presence of cyanide (expressed as nanomoles per mg of membrane protein) are presented as a function of the intramembranal serine concentration prior to the addition of cyanide.
By methods described previously (19), it was determined that each milligram of membrane protein contained approximately 1.7 ~1 of intramembranal fluid.
It was assumed that all of the intramembranal serine was in free solution. Inset, data plotted by the method of Hofstee (15). case membranes, respectively, while the V,,, is 5 to 6 nmoles per min per mg of membrane protein in both preparations. Thus, compared to the influx process under optimal conditions for transport (i.e. in the presence of ar-glycerol-P or L-lactate under aerobic conditions), the apparent K, for serine efflux is approximately 400 to 500 times higher, but the V,,, is almost identical.
In addition to demonstrating that the S. aureus system exhibits similar properties to those observed previously in E. coli (3, 4, B-20), these results suggest that the carriers can catalyze facilitated diffusion in the direction of efflux only.

DISCUSSION
Data presented in this paper demonstrate that active transport of amino acids by S. aureus membrane vesicles requires the oxidation of either cY-glycerol-P or L-lactate depending on the growth conditions of the parent cells, Thus, vesicles prepared from S. aureus U-71 grown on gluconate as a primary carbon source exhibit an absolute specificity for ar-glycerol-P as a physiological electron donor for transport, whereas vesicles prepared from the same cells grown on glucose exhibit an absolute specificity for L-lactate as an electron donor for transport. It is apparently the coupling between these dehydrogenases and transport which is altered rather than the presence or absence of the particular dehydrogenase, as both preparations exhibit by guest on March 24, 2020 http://www.jbc.org/ Downloaded from similar dehydrogenase activities qualitatively. Moreover, NADH is oxidized by both vesicle preparations at rates which exceed those of either a-glycerol-P or L-lactate. The absolute dependence of these respiration-linked transport systems on a specific physiological electron donor is unique to the Sfaphylococcal membrane vesicle system. In E. coli membrane vesicles, for instance, although nlactate is by far the most effective physiological electron donor for transport, succinate and NADH, as well as other electron donors, will support transport to some extent when the appropriate dehydrogenases are induced (3, 4, 13, 14, U-20).
A somewhat analogous situation has been reported in n-lactate dehydrogenase mutants of E. coli, however, where it has been shown that the coupling between succinic dehydrogenase and transport is markedly enhanced (22).
Active transport in S. uureus membrane vesicles and in vesicles prepared from a number of other bacterial species is directly dependent on electron transfer (l-4, 9). Inhibition of a-glycerol-P or L-lactate-dependent respiration in S. uureus vesicles by a variety of electron transfer inhibitors results in inhibition of amino acid uptake, and spectrophotometric and other evidence presented in this and other communications (1) demonstrates that the site of energy coupling between active transport and the respiratory chain occurs between the primary dehydrogenase(s) and the cytochrome chain.
Although the precise mechanism by which electron transfer is coupled to active transport is unknown, experiments presented here provide a strong indication that carrier activity is related to the oxidation-reduction potential of the respiratory chain at the site of energy coupling. The evidence rests primarily on the observations that inhibition of electron transfer after the energy-coupling site (i.e. with anaerobiosis, cyanide, 2-heptyl-4-hydroxyquinoline-N-oxide, or amytal) induces rapid efflux of solutes accumulated in the intravesicular pool, while inhibition before the site of energy coupling (i.e. with oxalate) does not induce efflux despite almost complete inhibition of oxidation and the initial rate of uptake. Similar arguments have been presented for E. coli membrane vesicles (4,3,11,18). In the Sfaphylococcal system, moreover, this argument is strengthened by the observation that oxalate not only fails to induce efflux, but also blocks efflux produced by electron transfer inhibitors which inhibit after the energycoupling site.
In addition, the finding that oxalate inhibits the rate of efflux induced by 2,4-dinitrophenol is striking, and suggests that the proton-conducting properties of this uncoupling agent cannot fully explain its inhibitory activity in this system. In the case of the lactose and proline transport systems in E. coli vesicles, where kinetics of influx and efflux have been studied in detail (18,19)) the concentrating ability of the vesicles is directly related to the ratio of the K, values for influx and efflux (4). A similar situation exists for the serine transport system in S. uureus vesicles.
The K, for active serine uptake with either a-glycerol-P or L-lactate is 10 to 12 PM, while the K, for efflux is approximately 5 mbr, yielding a ratio of 500. It can be calculated from the data given in Fig. 10 that the vesicles accumulate serine to an intravesicular concentration of approximately 5 mM at an external serine concentration of 10 PM, giving a distribution ratio of 500. These results provide further evidence that one of the primary effects of energy coupling is to change the affinity of the carriers for ligand.
Finally, the possibility that facilitated diffusion is the ratelimiting step for active serine accumulation is inconsistent with at least two observations: (a) the initial rate of serine uptake under anoxic conditions exhibits linear kinetics and it takes approximately 5 min for external serine to equilibrate with the intramembranal pool; (5) oxalate does not induce efflux of serine from the intravesicular pool despite almost complete inhibition of L-lactate oxidation and active serine uptake.