Active membrane transporters are dynamic molecular machines that catalyse transport across a membrane by coupling solute movement to a source of energy such as ATP or a secondary ion gradient. A central question for many active transporters concerns the mechanism by which transport is coupled to a source of energy. The transport process and associated energetic coupling involve conformational changes in the transporter. For efficient transport, the conformational changes must be tightly regulated and they must link energy use to movement of the substrate across the membrane. The present review discusses active transport using the well-established energetic framework for enzyme-mediated catalysis. In particular, membrane transport systems can be viewed as ensembles consisting of low-energy and high-energy conformations. The transport process involves binding interactions that selectively stabilize the higher energy conformations, and in this way promote conformational changes in the system that are coupled to decreases in free energy and substrate translocation. The major facilitator superfamily of secondary active transporters is used to illustrate these ideas, which are then be expanded to primary active transport mediated by ABC (ATP-binding cassette) import systems, with a focus on the well-studied maltose transporter.

ENZYMES, REACTION CO-ORDINATES AND TRANSITION STATES

Active transport systems undergo cycles of conformational changes that are associated with the formation and disruption of domain and subunit interactions. What is the energetic nature of these conformational changes and subunit interactions that allows the system to remain dynamic, and how do these changes link energy consumption to substrate movement?

These questions can be addressed by consideration of enzyme-catalysed reactions (Figure 1). The fundamental key to catalysis is transition state stabilization by the enzyme [1,2]. Thus enzymes demonstrate the greatest affinity for the transition state of the reaction, and relatively low affinity for the substrate and product, explaining why transition state analogues are such effective enzyme inhibitors [3]. In the absence of enzyme, the high energy of the transition state dictates that relatively few substrate molecules will reach an energy level high enough to undergo a chemical change to product. However, enzymes have evolved to bind tightly to transition states; in this manner, the binding energy of the enzyme lowers the energy of the transition state and the reaction rate is increased in proportion to the decrease in activation energy. Since the reaction rate will be limited by the step with the highest energy, the enzyme works by ‘flattening’ the energetic landscape between substrate and product and thereby increasing the rate of interconversion between the two.

Enzyme catalysis

Figure 1
Enzyme catalysis

(A) An enzyme-catalysed isomerization reaction. The enzyme (E; yellow) has evolved to bind the transition state (TS; purple) with higher affinity than either the substrate (S; blue) or product (P; red). (B) Energetics of an enzyme-catalysed reaction compared with the non-catalysed reaction. Relatively tight binding of the transition state by the enzyme, compared with its binding of the substrate or product, lowers the activation energy (∆GActivation) of the catalysed reaction and accelerates the interconversion between substrate and product. The overall rate enhancement is related to the decrease in the activation energy by exp(∆∆GActivation/RT); for example, a reduction in activation energy of 25 kJ/mol would produce a 24000-fold rate enhancement at 25°C.

Figure 1
Enzyme catalysis

(A) An enzyme-catalysed isomerization reaction. The enzyme (E; yellow) has evolved to bind the transition state (TS; purple) with higher affinity than either the substrate (S; blue) or product (P; red). (B) Energetics of an enzyme-catalysed reaction compared with the non-catalysed reaction. Relatively tight binding of the transition state by the enzyme, compared with its binding of the substrate or product, lowers the activation energy (∆GActivation) of the catalysed reaction and accelerates the interconversion between substrate and product. The overall rate enhancement is related to the decrease in the activation energy by exp(∆∆GActivation/RT); for example, a reduction in activation energy of 25 kJ/mol would produce a 24000-fold rate enhancement at 25°C.

In a similar manner, membrane transporters can be thought of as systems that flatten the energetic landscape the solute has to navigate to cross the membrane. In contrast with enzymes, transporters do not bind preferentially to a transition state of the substrate, because, of course, the substrate is not changed chemically during the transport process. Instead, the substrate binds preferentially to a high-energy conformation of the transporter that is analogous to the transition state of an enzyme-catalysed reaction. To illustrate this, consider the case of a passive transport system that mediates diffusion of substrate across a membrane using an alternating-access mechanism (Figure 2).

Energetics of facilitated diffusion by an alternating-access transporter

Figure 2
Energetics of facilitated diffusion by an alternating-access transporter

(A) Transmembrane transport by a passive alternating-access system. The transporter facilitates diffusion of a solute (yellow) across the membrane by binding initially in an open-out (left) or open-in (right) conformation. The transporter is relatively flexible, and has evolved to bind the solute most tightly when it adopts an occluded (middle) conformation. (B) An energetic model for the alternating-access passive transporter in which ligand binds most tightly to the higher-energy occluded conformation. The ligand-free transporter was assigned relative energies of 20 kJ/mol, 45 kJ/mol and 23 kJ/mol for the open-out, occluded and open-in conformations respectively (see the Supplementary Online Data for a rationale and example calculations). These energies would lead to a conformational distribution for open-out/occluded/open-in equal to 77:0.003:23 at equilibrium in the absence of ligand. The ligand (yellow square) binds with dissociation constants of 500 μM to the open-out, 50 μM to occluded and 1 mM to the open-in conformation, leading to decreases in their energies of −18.8 kJ/mol, −24.5 kJ/mol and −17.1 kJ/mol upon ligand binding, assuming standard state conditions (25°C and ligand activity of 1 on both sides of the membrane). In the presence of saturating concentrations of ligand, there will be three bound species with relative energies of 1.2 kJ/mol, 20.5 kJ/mol and 5.9 kJ/mol for open-out, occluded and open-in, yielding a conformational distribution for open-out/occluded/open-in equal to 87:0.036:13. Relatively tight binding of ligand to the occluded conformation leads to a lower energetic barrier between open-out and occluded conformations (19.3 kJ/mol going from ligand-bound open-out to occluded, compared with 25 kJ/mol for the unliganded transporter) and therefore faster conversion (approximately 10-fold) between open-out and open-in conformations in the presence of ligand. Tight binding of the substrate by the occluded state of the transporter therefore facilitates interconversion between open-in and open-out, in the same way that tight binding of the transition state by an enzyme facilitates interconversion between substrate and product.

Figure 2
Energetics of facilitated diffusion by an alternating-access transporter

(A) Transmembrane transport by a passive alternating-access system. The transporter facilitates diffusion of a solute (yellow) across the membrane by binding initially in an open-out (left) or open-in (right) conformation. The transporter is relatively flexible, and has evolved to bind the solute most tightly when it adopts an occluded (middle) conformation. (B) An energetic model for the alternating-access passive transporter in which ligand binds most tightly to the higher-energy occluded conformation. The ligand-free transporter was assigned relative energies of 20 kJ/mol, 45 kJ/mol and 23 kJ/mol for the open-out, occluded and open-in conformations respectively (see the Supplementary Online Data for a rationale and example calculations). These energies would lead to a conformational distribution for open-out/occluded/open-in equal to 77:0.003:23 at equilibrium in the absence of ligand. The ligand (yellow square) binds with dissociation constants of 500 μM to the open-out, 50 μM to occluded and 1 mM to the open-in conformation, leading to decreases in their energies of −18.8 kJ/mol, −24.5 kJ/mol and −17.1 kJ/mol upon ligand binding, assuming standard state conditions (25°C and ligand activity of 1 on both sides of the membrane). In the presence of saturating concentrations of ligand, there will be three bound species with relative energies of 1.2 kJ/mol, 20.5 kJ/mol and 5.9 kJ/mol for open-out, occluded and open-in, yielding a conformational distribution for open-out/occluded/open-in equal to 87:0.036:13. Relatively tight binding of ligand to the occluded conformation leads to a lower energetic barrier between open-out and occluded conformations (19.3 kJ/mol going from ligand-bound open-out to occluded, compared with 25 kJ/mol for the unliganded transporter) and therefore faster conversion (approximately 10-fold) between open-out and open-in conformations in the presence of ligand. Tight binding of the substrate by the occluded state of the transporter therefore facilitates interconversion between open-in and open-out, in the same way that tight binding of the transition state by an enzyme facilitates interconversion between substrate and product.

The alternating-access model was originally proposed in 1966 by Jardetzky [4], and applies to a number of different transport systems on the basis of biochemical and structural studies [59]. In the original treatment, the transporter exists in two conformations, ‘open-in’ and ‘open-out’, with the substrate-binding site accessible from the inside or outside surface of the membrane. However, it is clear that there will also be a point at which the transporter assumes an ‘occluded’ conformation that is intermediate between open-in and open-out, and in which the substrate-binding site is inaccessible from both sides of the membrane. Proteins are dynamic entities, and additional conformations will exist; in fact, the three conformations, i.e. open-in, open-out and occluded, may be regarded as ensembles of conformations, each populated with a range of energies [10]. For simplicity, and to maintain the analogy with enzyme catalysis, the discussion will be limited to the three conformations in the alternating-access mechanism that exhibit obvious functional differences: the open-in and open-out conformations, to which substrate can associate or dissociate, and the occluded conformation, in which the substrate binding site is inaccessible.

The three conformations will have different energies that will dictate their relative abundance at equilibrium [11]. On this basis, an alternating-access transporter in the absence of substrate should exhibit a relatively high energy for the occluded conformation compared with the open-in and open-out conformations. If this were not the case, then the unliganded transporter would be substantially populated in the occluded conformation and unable to bind substrate. The difference in energy between the occluded conformation and the open-in and open-out conformations will determine its abundance at equilibrium; for example, if the occluded conformation were 20 kJ/mol greater in energy (roughly one hydrogen bond) than both the open-in and open-out conformations, then it would be present at approximately 0.016% and the open-in and open-out forms would predominate, allowing substrate to access the binding site. The energy of the occluded conformation will also determine the rate of interconversion between the open-in and open-out conformations. A 20 kJ/mol difference (8 kBT) is compatible with relatively large domain movements and slow interconversion on a micro- to milli-second timescale [10].

The energetics of the transport system will be altered in the presence of substrate. In particular, the binding site will have a slightly different structure in each conformation, and, on this basis, each conformation will exhibit a unique binding mode and affinity for substrate. In this regard, it is worth asking which conformation is expected to have the highest affinity for the substrate. Enzymes have evolved to bind the transition state more tightly than either substrate or product because relatively tight binding of substrate or product would slow the reaction by effectively increasing the activation energy. Similarly, an alternating-access transporter is expected to exhibit the highest affinity for substrate when it is in an occluded conformation. In this way, the presence of substrate will lower the energy of the occluded conformation relative to the open-in and open-out conformations, and increase the rate of exchange between the two. In such a system, the substrate would bind initially with relatively low affinity to either the open-in or open-out conformation, and then promote a conformational change towards the occluded conformation by binding progressively more tightly to transporter conformations that approach the occluded state. That is, high-affinity binding of substrate to the relatively unstable occluded conformation would promote conformational change in the system, just as enzymes promote chemical change by binding tightly to an unstable transition state.

‘TRANSITION STATE’ STABILIZATION IN THE MAJOR FACILITATOR SUPERFAMILY

The major facilitator superfamily (MSF) of secondary active transporters provides examples of active alternating-access transporters. The energy for active transport by these systems comes from a secondary electrochemical gradient. These systems facilitate transport of substrate molecules across the membrane that is coupled to dissipation of the secondary electrochemical gradient. The decrease in free energy of the secondary gradient can therefore be used by the system to move the substrate against an electrochemical gradient.

In the lactose permease symporter, uptake of lactose into the cell is coupled to a proton electrochemical gradient: lactose and a proton are moved together across the membrane (Figure 3). To effect coupled movement of lactose and a proton, the system can alternate between open-in and open-out conformations only if is empty or if both lactose and a proton are bound. Key to the energetic coupling is the idea that the permease with either lactose or a proton bound is hindered from changing conformation. If the system could alternate between open-in and open-out with only one or the other of lactose or a proton bound, then it would simply facilitate diffusion of these molecules across the membrane.

The lactose permease symporter

Figure 3
The lactose permease symporter

The lactose permease symporter catalyses transmembrane movement of a lactose molecule (blue) and proton (red) together. To efficiently couple lactose movement to the proton electrochemical gradient, only the empty transporter (left side) or the ternary complex (right side) can exchange between the open-out and open-in conformations; the transporter with either a proton or lactose molecule alone (middle species) is not able to exchange between the open-out and open-in conformations.

Figure 3
The lactose permease symporter

The lactose permease symporter catalyses transmembrane movement of a lactose molecule (blue) and proton (red) together. To efficiently couple lactose movement to the proton electrochemical gradient, only the empty transporter (left side) or the ternary complex (right side) can exchange between the open-out and open-in conformations; the transporter with either a proton or lactose molecule alone (middle species) is not able to exchange between the open-out and open-in conformations.

Lactose permease uses an alternating-access mechanism [5], and the enzymatic framework used previously for passive transport can be used to explain energetic coupling in lactose permease. For the permease in the absence of substrate, the energy of the occluded conformation should be somewhat greater than the open-in and open-out conformations (Figure 4A). An energetic difference of approximately 35 kJ/mol corresponds roughly to values obtained through Arrhenius analysis of the lactose permease [12], is roughly compatible with conformational transitions between open-in and open-out on the millisecond timescale [13], and is sufficiently high so that there will be relatively little empty occluded transporter present at equilibrium (see Supplementary Table S1 for additional information).

Energetic coupling in lactose permease

Figure 4
Energetic coupling in lactose permease

(A) Structures and energy levels for unliganded (empty) lactose permease and the ternary complex. In an enzymatic model of function, lactose permease works analogously to the system for facilitated diffusion, with relatively small energetic barriers for conformational change between open-in and open-out when the permease is unliganded (green bars) or in the ternary complex (purple bars), bound to both a proton (red diamond) and lactose (blue polygon). In the example shown, there is a 35 kJ/mol difference in energy between the empty open-out and occluded transporters. Binding of lactose and a proton lowers the energy of the open-out form by 24 kJ/mol and the occluded form by 30 kJ/mol; because of tighter binding to the occluded form, binding of lactose and a proton lowers the energetic difference between the open-out and occluded forms to 29 kJ/mol, and so the conformational change to the open-in form will be faster than for the unliganded permease. (B) Structures and energy levels for binary complexes of the permease. Efficient energetic coupling dictates that the binary complexes are hindered from alternating between open-in and open-out. On this basis, the activation energy for the conformational transition must be greatest for the two binary complexes. This could come about by unfavourable interactions of a single ligand with the occluded conformation. For example, the presence of lactose alone in the occluded conformation may raise its energy if a protonated amino acid is required for high-affinity binding, a situation that may not apply for the open-in and open-out conformations. Alternatively, the activation energy of the binary complex could be relatively high because of tight binding of a single ligand to the open-in and open-out conformations. In the case of the proton (red diamond and bars), relatively tight binding to the open-in and open-out conformations with little effect on the energy of the occluded conformation would increase the energetic barrier to conformational change. This could occur if the pKa of an ionizable group were lower in the occluded form than in the open-in and open-out conformations. (C) As in (A) and (B), except that the energies of the open-in and open-out conformations are put at the same relative position to illustrate changes in the activation energy that result from the differences in ligand-binding affinities to either the occluded or open-in/open-out states. On their own, the proton (red) and lactose (blue) bind the occluded conformation relatively weakly, leading to its high energy relative to the open-in and open-out conformations, to which they bind with greater affinity. On the other hand, the empty transporter (green) is able to alternate relatively freely (the empty occluded conformation is only modestly higher in energy), and a proton and lactose together in the ternary complex (magenta) exhibit a relatively high affinity for the occluded conformation, lowering its energy with respect to the binary complexes. The coupling efficiency will ultimately be determined by the difference in energy between the occluded conformations: a 25 kJ/mol difference in energy would be expected to produce a >24000-fold difference in the rate of conversion between open-in and open-out. With such an energetic profile, the empty transporter and the ternary complex will undergo hundreds or thousands of open-in/open-out conversions for every such conversion of a binary complex.

Figure 4
Energetic coupling in lactose permease

(A) Structures and energy levels for unliganded (empty) lactose permease and the ternary complex. In an enzymatic model of function, lactose permease works analogously to the system for facilitated diffusion, with relatively small energetic barriers for conformational change between open-in and open-out when the permease is unliganded (green bars) or in the ternary complex (purple bars), bound to both a proton (red diamond) and lactose (blue polygon). In the example shown, there is a 35 kJ/mol difference in energy between the empty open-out and occluded transporters. Binding of lactose and a proton lowers the energy of the open-out form by 24 kJ/mol and the occluded form by 30 kJ/mol; because of tighter binding to the occluded form, binding of lactose and a proton lowers the energetic difference between the open-out and occluded forms to 29 kJ/mol, and so the conformational change to the open-in form will be faster than for the unliganded permease. (B) Structures and energy levels for binary complexes of the permease. Efficient energetic coupling dictates that the binary complexes are hindered from alternating between open-in and open-out. On this basis, the activation energy for the conformational transition must be greatest for the two binary complexes. This could come about by unfavourable interactions of a single ligand with the occluded conformation. For example, the presence of lactose alone in the occluded conformation may raise its energy if a protonated amino acid is required for high-affinity binding, a situation that may not apply for the open-in and open-out conformations. Alternatively, the activation energy of the binary complex could be relatively high because of tight binding of a single ligand to the open-in and open-out conformations. In the case of the proton (red diamond and bars), relatively tight binding to the open-in and open-out conformations with little effect on the energy of the occluded conformation would increase the energetic barrier to conformational change. This could occur if the pKa of an ionizable group were lower in the occluded form than in the open-in and open-out conformations. (C) As in (A) and (B), except that the energies of the open-in and open-out conformations are put at the same relative position to illustrate changes in the activation energy that result from the differences in ligand-binding affinities to either the occluded or open-in/open-out states. On their own, the proton (red) and lactose (blue) bind the occluded conformation relatively weakly, leading to its high energy relative to the open-in and open-out conformations, to which they bind with greater affinity. On the other hand, the empty transporter (green) is able to alternate relatively freely (the empty occluded conformation is only modestly higher in energy), and a proton and lactose together in the ternary complex (magenta) exhibit a relatively high affinity for the occluded conformation, lowering its energy with respect to the binary complexes. The coupling efficiency will ultimately be determined by the difference in energy between the occluded conformations: a 25 kJ/mol difference in energy would be expected to produce a >24000-fold difference in the rate of conversion between open-in and open-out. With such an energetic profile, the empty transporter and the ternary complex will undergo hundreds or thousands of open-in/open-out conversions for every such conversion of a binary complex.

The open-in, open-out and occluded conformations will have particular affinities for the substrate (in this case lactose and a proton together), so the reaction profile will change when substrates are present. Assuming the occluded conformation has the highest energy, efficient transport requires that the substrate bind with greatest affinity to the occluded conformation. Binding preferentially to either the open-in conformation or the open-out conformation, which already have a lower energy than the occluded conformation, will lead to slower exchange between open-in and open-out owing to an increase in the relative energy of the occluded substrate-bound state. On the other hand, preferential binding to the occluded conformation will lower its energy and enhance the rate of exchange between open-in and open-out. Thus the presence of the substrate can flatten the energetic landscape of the transporter, and increase conformational exchange between open-in and open-out (Figure 4A). For example, if lactose and a proton together bind with a KD of 1 mM to the open-in and open-out states, but with a 10-fold higher affinity to the occluded state, then the relative energy of the occluded conformation will decrease from 35 kJ/mol to 29.3 kJ/mol due to the 5.7 kJ/mol in greater binding energy. This change will produce a 10-fold increase in the relative concentration of the occluded conformation, and a corresponding 10-fold increase in the rate of exchange between open-in and open-out.

The key to efficient energetic coupling is that the binding of a proton or lactose alone must raise the energy of the occluded state relative to open-in and open-out states (Figure 4B). For example, if the presence of a proton on its own raised the energy of the occluded state by 25 kJ/mol (roughly the energy of a hydrogen bond), then its relative concentration would decrease over 4 orders of magnitude, and the rate of open-in/open-out exchange would be roughly 10000-fold lower than for the unoccupied transporter. Energetic coupling results from conformational changes between open-in and open-out that are hundreds or thousands of times faster when the permease is empty or bound to both lactose and a proton, compared with when it is bound to either a proton or lactose alone (Figure 4C).

The energetics of the transport reaction co-ordinate appear similar to those of an enzyme-catalysed reaction. In the case of an enzyme-catalysed reaction, the substrate exists in a low-energy ground state that binds relatively weakly to the enzyme. Once in the active site, the propensity of the enzyme to bind with high affinity to the transition state will effectively lower the energy of the transition state and increase its relative population, thereby increasing the rate of interconversion between substrate and product. For transmembrane transport, the occluded conformation of the transporter is analogous to the transition state of the substrate in an enzyme-catalysed reaction: the occluded conformation has a relatively high energy that can be lowered by the presence of the transport substrate. Thus the presence or absence of substrate will modulate the energy of the occluded conformation relative to the open-in and open-out states, and thereby affect the rate of exchange between open-in and open-out. For secondary active transporters, energy coupling results from particular ligand-bound (or free) occluded states having much higher energies.

The same energetic framework can be used for MFS antiporters, such as the glycerol 3-phosphate transporter that moves organic phosphates into the cell and couples this to movement of inorganic phosphate out of the cell. In this case, efficient coupling dictates that the empty occluded conformation has a relatively high energy compared with the phosphate-bound forms. The phosphate-bound forms exchange rapidly, while uncoupled dissipation of the inorganic phosphate gradient is prevented by slow exchange of the empty transporter due to the relatively high energy of the empty occluded conformation. These types of energetic considerations could also be applied to the neurotransmitter/sodium symporter family, represented by the leucine transporter, for which there are structures in all three conformations, i.e. open-in and open-out and the occluded state, bound to leucine and two sodium ions [8,9]. These crystal structures have been complemented by a comprehensive EPR study that relates binding of sodium and leucine to conformational changes in the system [14].

ENERGETICS AND DYNAMICS IN THE MALTOSE TRANSPORTER

ABC (ATP-binding cassette) transporters work using an alternating-access model with the energy required for mechanical work supplied by ATP binding and hydrolysis. ABC import systems are complicated by a second substrate-binding site on a peripheral binding protein. Nevertheless, the energetic framework based on the idea of transition state stabilization in enzymes can be applied to ABC importers to provide insight into the dynamics and energetics of their transport cycle.

The maltose transporter consists of two membrane subunits, MalF and MalG, and two ATP-binding cassettes, MalK (Figure 5A). There are crystal structures of MalFGK2 in three conformations: open-in [6], open-out [15] and occluded [7]; importantly, only the open-out conformation can hydrolyse ATP because this is the only conformation in which the MalK cassettes are able to tightly dimerize to assemble the ATPase catalytic machinery. A mechanism for maltose transport is illustrated in Figure 5A. EPR measurements indicate MalFGK2 exists in the membrane in an open-in conformation, consistent with a relatively low energy for this conformation [16]. EPR measurements in detergent, proteoliposomes and membrane nanodiscs have also shown that the presence of both ATP and MBP-maltose (maltose binding protein in complex with maltose), but not ATP alone, will facilitate tight dimerization of the cassettes, indicating that MBP–maltose helps to provoke a change towards the open-out conformation [16,17]. How does MBP–maltose do this?

Energetic coupling in the maltose transporter

Figure 5
Energetic coupling in the maltose transporter

(A) The maltose transporter consists of MBP (blue), integral membrane subunits MalF and MalG (green), and the MalK ATPase subunits (salmon). Maltose transport is coupled to ATP hydrolysis in several distinct steps. First, the presence of maltose is signalled through the interaction between MBP–maltose and MalFGK2 [19]. Rather than binding to, and stabilizing, the ground state, MBP–maltose binds to the occluded state [7]. Secondly, as the system achieves an open-out conformation, the P3 loop (‘scoop loop’) is inserted into the maltose-binding site of MBP, ensuring that maltose has moved into the MalFG-binding site when ATP is hydrolysed [15,22,23]. After ATP hydrolysis, MBP is expected to stay bound to MalFGK2 until the occluded state is reached, so that maltose cannot diffuse back into the periplasm. MBP dissociates as MalFGK2 goes from the occluded to open-in conformation, when maltose diffuses into the cytoplasm. (B) The reaction co-ordinate for the open-in to open-out conformational change is shown for MalFGK2 in the presence of ATP (black lines) and in the presence of both ATP and MBP–maltose (blue lines). In the absence of MBP–maltose, MalFGK2 exists predominantly in an open-in conformation, which is unable to hydrolyse ATP due to the separation in the ATP-binding cassettes. In the presence of MBP–maltose, a very stable complex is formed, provided ATP hydrolysis is prevented. On this basis, the open-out MBP–MalFGK2–ATP complex is given the lowest energy. Open MBP binds with submicromolar affinity to the open-out conformation [18,25,26], indicating that the free energy of binding is of the order of 40 kJ/mol. Maltose–MBP stimulates the ATPase activity with a Km of approximately 15 μM [20], corresponding to an interaction energy of approximately 30 kJ/mol. On this basis, binding of MBP–maltose to the occluded conformation would lead to a >100000-fold rate enhancement for the conformational change from open-in to open-out and thereby contribute to the effect of MBP–maltose on the ATPase activity of MalFGK2.

Figure 5
Energetic coupling in the maltose transporter

(A) The maltose transporter consists of MBP (blue), integral membrane subunits MalF and MalG (green), and the MalK ATPase subunits (salmon). Maltose transport is coupled to ATP hydrolysis in several distinct steps. First, the presence of maltose is signalled through the interaction between MBP–maltose and MalFGK2 [19]. Rather than binding to, and stabilizing, the ground state, MBP–maltose binds to the occluded state [7]. Secondly, as the system achieves an open-out conformation, the P3 loop (‘scoop loop’) is inserted into the maltose-binding site of MBP, ensuring that maltose has moved into the MalFG-binding site when ATP is hydrolysed [15,22,23]. After ATP hydrolysis, MBP is expected to stay bound to MalFGK2 until the occluded state is reached, so that maltose cannot diffuse back into the periplasm. MBP dissociates as MalFGK2 goes from the occluded to open-in conformation, when maltose diffuses into the cytoplasm. (B) The reaction co-ordinate for the open-in to open-out conformational change is shown for MalFGK2 in the presence of ATP (black lines) and in the presence of both ATP and MBP–maltose (blue lines). In the absence of MBP–maltose, MalFGK2 exists predominantly in an open-in conformation, which is unable to hydrolyse ATP due to the separation in the ATP-binding cassettes. In the presence of MBP–maltose, a very stable complex is formed, provided ATP hydrolysis is prevented. On this basis, the open-out MBP–MalFGK2–ATP complex is given the lowest energy. Open MBP binds with submicromolar affinity to the open-out conformation [18,25,26], indicating that the free energy of binding is of the order of 40 kJ/mol. Maltose–MBP stimulates the ATPase activity with a Km of approximately 15 μM [20], corresponding to an interaction energy of approximately 30 kJ/mol. On this basis, binding of MBP–maltose to the occluded conformation would lead to a >100000-fold rate enhancement for the conformational change from open-in to open-out and thereby contribute to the effect of MBP–maltose on the ATPase activity of MalFGK2.

The central idea behind the proposed energetic framework is that the occluded conformation of MalFGK2 has a relatively high energy compared with the open-in and open-out conformations (Figure 5B). In this way, the occluded conformation represents an energetic barrier to free interconversion between the open-in and open-out conformations, and its presence limits the potential for uncoupled ATP hydrolysis. With this energetic framework in mind, a role for MBP–maltose becomes clear: it binds and stabilizes the occluded conformation of MalFGK2. MBP–maltose drives conformational change in the system by binding to a higher-energy state of MalFGK2, similar to the way an enzyme will drive chemical change by binding to a high-energy transition state of the substrate.

In many schematics of ABC import systems, the liganded binding protein is shown to interact with the open-in conformation, and this interaction leads to a change in conformation towards open-out. However, if MBP–maltose were to bind preferentially to open-in MalFGK2, it would decrease ATPase activity by stabilizing a low-energy conformation that is not able to hydrolyse ATP. Binding to a higher-energy occluded conformation, intermediate between open-in and open-out, offers an energetically reasonable pathway for MBP–maltose to effect conformational change in the system. The high energy of the occluded conformation will make its interaction with MBP inherently unstable, and the system will either dissociate and return to the open-in conformation or it will proceed in the other direction towards the open-out conformation of MalFGK2. ATP participates in this process as well, since it is required for the system to attain the tightly closed and catalytically competent conformation where ATP can be hydrolysed [16,18]. Thus the population of each MalK-binding site with ATP is expected to lower the energy of the occluded state, along with binding of MalFGK2 on the outer surface.

The separation of the open-in and open-out conformations by a high-energy occluded conformation explains the longstanding observation that unliganded MBP stimulates ATPase activity by MalFGK2 [19,20]. Addition of unliganded MBP to membrane-embedded MalFGK2 brings about a small, but significant, increase in ATPase activity, with a Km below 1 μM, which makes this interaction completely different from the interaction between maltose-bound MBP and MalFGK2. A similar stimulation of ATPase activity by the unliganded binding protein was also observed for the histidine transporter [21]. This kinetic effect is most likely to be due to high-affinity binding of unliganded MBP to the open-out conformation of MalFGK2; there is evidence that this interaction contributes to efficient and transport-coupled ATP hydrolysis [22,23]. This effect of unliganded MBP on MalFGK2 ATPase, characterized by a submicromolar Km and low Vmax, is therefore consistent with a minor population of the open-out conformation being present in membranes, and a kinetic barrier, i.e. the occluded conformation, that prevents high levels of ATP hydrolysis in the absence of maltose-loaded MBP.

The idea that MBP–maltose binds preferentially to the occluded conformation of MalFGK2 is supported by the crystal structure of closed ligand-bound MBP in complex with the occluded conformation of MalFGK2 [7]. The presence of this crystal structure, which incorporates a high-energy conformation of MalFGK2, raises the issue of detergent effects on membrane proteins. MalFGK2 in phospholipid bilayers exhibits exceptionally low rates of basal ATPase activity that are increased roughly 10-fold when MalFGK2 is solubilized in detergent. This increase in ATPase activity is consistent with a loss of structural constraints in detergent micelles compared with the phospholipid bilayer. That is, the differences in energies for the various conformations of MalFGK2 are attenuated when it is solubilized in detergent. On this basis, the detergent solubilization required for crystallization probably helps to lower the energy of conformations that may only be represented to a minor degree in a phospholipid bilayer.

COUPLING OF ATP HYDROLYSIS TO MALTOSE TRANSPORT

With this framework in place, it is possible create a model for the catalytic cycle that can explain how the system has evolved to couple ATP hydrolysis to movement of maltose across the membrane. In fact, enough is known about the individual steps in the ATP hydrolytic and transport cycle that the ‘biological logic’ that drove the evolution of the system is coming into focus. The available data support a model where ATP binding and hydrolysis is coupled to the movement of maltose in several discrete steps.

  1. In the absence of maltose-loaded MBP, MalFGK2 adopts predominantly an open-in conformation in which the ligand-binding site is exposed to the cytoplasm, and the ABC subunits are separated and unable to hydrolyse ATP. Rapid exchange between open-in and open-out does not take place because of the barrier imposed by the high-energy occluded conformation.

  2. MBP–maltose and ATP together initiate the catalytic cycle by binding to, and lowering the energy of, the occluded conformation, facilitating the transition to the open-out conformation in which ATP can be hydrolysed. This is the first step in the mechanism that couples ATP hydrolysis to maltose transport. In particular, the requirement for maltose-bound MBP ensures that the ATPase-competent open-out conformation is only achieved when maltose is present outside of the cell.

  3. When the open-out conformation is attained, interaction between a loop in MalG and the empty maltose-binding site of MBP is required for efficient ATP hydrolysis [15,22,23]. This is the second step in the energetic coupling mechanism: it ensures that maltose is transferred from its binding site in MBP to the binding site in MalFGK2 before ATP is hydrolysed.

  4. The high-affinity association between unliganded MBP and open-out MalFGK2 prevents maltose diffusing out of its low-affinity binding site in MalFG and back into the external environment. On this basis, it is expected that MBP will stay associated with MalFGK2 until the occluded conformation is reached.

ATP performs two important functions in this cycle. First, the ATP molecules bound to the cassettes are required to drive the conformational change from open-in and through the occluded to the open-out conformation [1618]. This process represents a descent in free energy as interactions between the ATP molecules and the binding sites in the cassettes become stronger and more numerous as the cassettes adopt a closed tightly dimerized conformation. When ATP hydrolysis is prevented by mutation (E159Q) in the MalK active sites, the ATP-bound MBP–MalFGK2 becomes arrested in the stable low-energy conformation that was crystallized [15]. In the case of the isolated ATP-binding cassette from MJ0796 of Methanococcus jannaschii with an equivalent ATPase-inactivating mutation (E171Q), the presence of ATP decreases the dimer dissociation constant from 208 μM to 70 nM [24], representing a decrease in free energy of 20 kJ/mol. Thus tight ATP binding by the cassettes drives the system into a stable and relatively low-energy state.

The second function of ATP is to undergo hydrolysis and raise the energy of the system. This stage of the transport cycle depends only on relatively subtle structural changes. Thus a conformation similar to that obtained with the MalK(E159Q) mutation was produced by incubating MalFGK2 and MBP with ATP and vanadate [15,25,26]. More recently, crystal structures of MBP–MalFGK2 arrested with AMP-PNP (adenosine 5′-[β,γ-imido]triphosphate), ADP·BeF3, ADP·VO4, and ADP·AlF4 have been solved [27]. The overall structures of the complexes are essentially identical, but there are subtle changes in the active site. Most importantly, a comparison of the complexes representing ATP bound in the ground state (i.e. AMP-PNP and ADP·BeF4) and the two representing the catalytic transition state (ADP·VO4 and ADP·ALF4) indicated the position of the attacking water molecule. Because the overall structure of the MBP–MalFGK2 complex is virtually identical in all four cases, the arrested complex is often referred to as representing the ‘transition state for ATP hydrolysis’, although in a catalytically competent MBP–MalFGK2 system with Mg2+-ATP, the actual transition state will be much higher in energy due to the activation energy associated with hydrolysis of ATP.

EXPERIMENTAL SUPPORT FOR THE MALTOSE TRANSPORT MODEL AND UNANSWERED QUESTIONS

The first steps in the proposed coupling mechanism involve binding of MBP–maltose to an occluded conformation of MalFGK2 that promotes a conformational change from the resting state (open-in) to the open-out conformation where ATP is poised for hydrolysis. This is followed by movement of maltose from its binding site in MBP to its MalFGK2-binding site and hydrolysis of ATP. These steps have been supported experimentally by both biochemical and structural studies. Supporting evidence includes:

  1. Open-in resting-state conformation:

    • Crystal structure of unliganded MalFGK2 in open-in conformation [6].

    • EPR measurements consistent with open-in as the predominant conformation in detergent and membranes in the presence or absence of ATP [16,17].

  2. Interaction of maltose-bound MBP with the occluded state:

    • Robust stimulation of ATPase activity by maltose–MBP, with a Km of approximately 15 μM [20].

    • Crystal structure of MBP–maltose bound to an occluded conformation of MalFGK2 [7].

    • Ability of a cross-linked closed form of MBP to act as a dominant negative in maltose transport [28].

    • Little or no binding of maltose–MBP with cross-linked forms of either the open-in or open-out conformations [18].

    • EPR measurements that indicate maltose–MBP together with ATP is required for conformational change from the open-in to open-out forms [16,17].

  3. Hydrolysis of ATP coupled to movement of maltose by interaction of the MalG P3 loop (i.e. the ‘scoop loop’) with the maltose-binding site of MBP:

    • Crystal structure of MalFGK2–MBP complex showing insertion of the MalG scoop loop into the maltose-binding site of MBP [15].

    • Mutations in the maltose-binding site of MBP drastically lower the ability of the system to hydrolyse ATP [22].

    • Truncation of the MalG P3 loop leads to constitutive uncoupled ATPase activity [23].

The last step, in which the system moves from open-out to open-in after ATP hydrolysis, is difficult to interrogate. If the occluded conformation really is a higher energy state in going from open-in to open-out, then it is reasonable to assume that the same conformation will also have a relatively high energy in going from open-out to open-in. It is not clear how this energetic barrier is surmounted in the absence of a stabilizing interaction with MBP–maltose. An obvious difference in the system as it proceeds through the cycle is the identity of the nucleotides bound to the ATP-binding cassettes. In the forward reaction, both ATP and MBP–maltose are required for the conformational shift from open-in to open-out. With the post-hydrolysis state, the presence of ADP and inorganic phosphate in the cassettes, maltose in the MalFG-binding site and/or unliganded MBP bound to the exterior may contribute to lowering the energy of the occluded state, but detailed molecular mechanisms are difficult to envisage.

CONCLUSION

The fundamental thermodynamic concepts of microstates, the Boltzmann distribution and transition state stabilization have been used, and continue to be used, to understand mechanisms of enzyme catalysis. These same principles can be applied to membrane transport systems to interpret biochemical results and integrate them into models of conformational change and chemical coupling. The value of treating transport systems as catalytic systems or enzymes has been suggested previously [29,30], and the idea of an energetic barrier that is lowered by substrate binding has been advanced in the case of the MFS proteins, and specifically the glycerol 3-phosphate antiporter [31,32]. These ideas have been developed and extended to the maltose transporter, a primary active transport system. For enzymes, chemical change is promoted by tight binding and stabilization of a high-energy transition state by the enzyme. For transporters, the central mechanistic questions have to do with how conformational change is promoted. By analogy with enzymes, conformational change will be promoted by stabilization of a higher-energy intermediate (occluded) conformation. The stabilization can be accomplished by high-affinity binding of the transport substrate to the occluded conformation, or by binding of the occluded conformation to a low-energy form of a transporter subunit, such as maltose-bound MBP.

Abbreviations

     
  • ABC

    ATP-binding cassette

  •  
  • AMP-PNP

    adenosine 5′-[β,γ-imido]triphosphate

  •  
  • MBP

    maltose-binding protein

  •  
  • MSF

    major facilitator superfamily

Thanks to members of the Shilton laboratory and especially Patrick Telmer and Alister Gould for their work on the transport system, and to Janet Wood at the University of Guelph, Guelph, Canada, for comments and discussion about the review.

FUNDING

A Discovery Grant from the Natural Sciences and Engineering Research Council of Canada supports our research on transport systems.

References

References
1
Wolfenden
R.
Transition state analog inhibitors and enzyme catalysis
Annu. Rev. Biophys. Bioeng.
1976
, vol. 
5
 (pg. 
271
-
306
)
[PubMed]
2
Garcia-Viloca
M.
Gao
J.
Karplus
M.
Truhlar
D.G.
How enzymes work: analysis by modern rate theory and computer simulations
Science
2004
, vol. 
303
 (pg. 
186
-
195
)
[PubMed]
3
Schramm
V.L.
Transition states, analogues, and drug development
ACS Chem. Biol.
2013
, vol. 
8
 (pg. 
71
-
81
)
[PubMed]
4
Jardetzky
O.
Simple allosteric model for membrane pumps
Nature
1966
, vol. 
211
 (pg. 
969
-
970
)
[PubMed]
5
Smirnova
I.
Kasho
V.
Kaback
H.R.
Lactose permease and the alternating access mechanism
Biochemistry
2011
, vol. 
50
 (pg. 
9684
-
9693
)
[PubMed]
6
Khare
D.
Oldham
M.L.
Orelle
C.
Davidson
A.L.
Chen
J.
Alternating access in maltose transporter mediated by rigid-body rotations
Mol. Cell
2009
, vol. 
33
 (pg. 
528
-
536
)
[PubMed]
7
Oldham
M.L.
Chen
J.
Crystal structure of the maltose transporter in a pretranslocation intermediate state
Science
2011
, vol. 
332
 (pg. 
1202
-
1205
)
[PubMed]
8
Krishnamurthy
H.
Gouaux
E.
X-ray structures of LeuT in substrate-free outward-open and apo inward-open states
Nature
2012
, vol. 
481
 (pg. 
469
-
474
)
[PubMed]
9
Yamashita
A.
Singh
S.K.
Kawate
T.
Jin
Y.
Gouaux
E.
Crystal structure of a bacterial homologue of Na+/Cl−-dependent neurotransmitter transporters
Nature
2005
, vol. 
437
 (pg. 
215
-
223
)
[PubMed]
10
Henzler-Wildman
K.
Kern
D.
Dynamic personalities of proteins
Nature
2007
, vol. 
450
 (pg. 
964
-
972
)
[PubMed]
11
Garcia
H.G.
Kondev
J.
Orme
N.
Theriot
J.A.
Phillips
R.
Thermodynamics of biological processes
Methods Enzymol.
2011
, vol. 
492
 (pg. 
27
-
59
)
[PubMed]
12
Zhang
W.
Kaback
H.R.
Effect of the lipid phase transition on the lactose permease from Escherichia coli
Biochemistry
2000
, vol. 
39
 (pg. 
14538
-
14542
)
[PubMed]
13
Henzler-Wildman
K.A.
Lei
M.
Thai
V.
Kerns
S.J.
Karplus
M.
Kern
D.
A hierarchy of timescales in protein dynamics is linked to enzyme catalysis
Nature
2007
, vol. 
450
 (pg. 
913
-
916
)
[PubMed]
14
Kazmier
K.
Sharma
S.
Quick
M.
Islam
S.M.
Roux
B.
Weinstein
H.
Javitch
J.A.
McHaourab
H.S.
Conformational dynamics of ligand-dependent alternating access in LeuT
Nat. Struct. Mol. Biol.
2014
, vol. 
21
 (pg. 
472
-
479
)
[PubMed]
15
Oldham
M.L.
Khare
D.
Quiocho
F.A.
Davidson
A.L.
Chen
J.
Crystal structure of a catalytic intermediate of the maltose transporter
Nature
2007
, vol. 
450
 (pg. 
515
-
521
)
[PubMed]
16
Orelle
C.
Ayvaz
T.
Everly
R.M.
Klug
C.S.
Davidson
A.L.
Both maltose-binding protein and ATP are required for nucleotide-binding domain closure in the intact maltose ABC transporter
Proc. Natl. Acad. Sci. U.S.A.
2008
, vol. 
105
 (pg. 
12837
-
12842
)
[PubMed]
17
Alvarez
F.J. D.
Orelle
C.
Davidson
A.L.
Functional reconstitution of an ABC transporter in nanodiscs for use in electron paramagnetic resonance spectroscopy
J. Am. Chem. Soc.
2010
, vol. 
132
 (pg. 
9513
-
9515
)
[PubMed]
18
Bao
H.
Duong
F.
ATP alone triggers the outward facing conformation of the maltose ATP-binding cassette transporter
J. Biol. Chem.
2013
, vol. 
288
 (pg. 
3439
-
3448
)
[PubMed]
19
Davidson
A.
Shuman
H.
Nikaido
H.
Mechanism of maltose transport in Escherichia coli: transmembrane signaling by periplasmic binding proteins
Proc. Natl. Acad. Sci. U.S.A.
1992
, vol. 
89
 (pg. 
2360
-
2364
)
[PubMed]
20
Gould
A.D.
Telmer
P.G.
Shilton
B.H.
Stimulation of the maltose transporter ATPase by unliganded maltose binding protein
Biochemistry
2009
, vol. 
48
 (pg. 
8051
-
8061
)
[PubMed]
21
Ames
G.F.
Liu
C.E.
Joshi
A.K.
Nikaido
K.
Liganded and unliganded receptors interact with equal affinity with the membrane complex of periplasmic permeases, a subfamily of traffic ATPases
J. Biol. Chem.
1996
, vol. 
271
 (pg. 
14264
-
70
)
[PubMed]
22
Gould
A.D.
Shilton
B.H.
Studies of the maltose transport system reveal a mechanism for coupling ATP hydrolysis to substrate translocation without direct recognition of substrate
J. Biol. Chem.
2010
, vol. 
285
 (pg. 
11290
-
11296
)
[PubMed]
23
Cui
J.
Qasim
S.
Davidson
A.L.
Uncoupling substrate transport from ATP hydrolysis in the Escherichia coli maltose transporter
J. Biol. Chem.
2010
, vol. 
285
 (pg. 
39986
-
39993
)
[PubMed]
24
Moody
J.E.
Millen
L.
Binns
D.
Hunt
J.F.
Thomas
P.J.
Cooperative, ATP-dependent association of the nucleotide binding cassettes during the catalytic cycle of ATP-binding cassette transporters
J. Biol. Chem.
2002
, vol. 
277
 (pg. 
21111
-
21114
)
[PubMed]
25
Chen
J.
Sharma
S.
Quiocho
F.A.
Davidson
A.L.
Trapping the transition state of an ATP-binding cassette transporter: evidence for a concerted mechanism of maltose transport
Proc. Natl. Acad. Sci. U.S.A.
2001
, vol. 
98
 (pg. 
1525
-
1530
)
[PubMed]
26
Austermuhle
M.I.
Hall
J.A.
Klug
C.S.
Davidson
A.L.
Maltose-binding protein is open in the catalytic transition state for ATP hydrolysis during maltose transport
J. Biol. Chem.
2004
, vol. 
279
 (pg. 
28243
-
28250
)
[PubMed]
27
Oldham
M.L.
Chen
J.
Snapshots of the maltose transporter during ATP hydrolysis
Proc. Natl. Acad. Sci. U.S.A.
2011
, vol. 
108
 (pg. 
15152
-
15156
)
[PubMed]
28
Zhang
Y.
Mannering
D.
Davidson
A.
Yao
N.
Manson
M.
Maltose-binding protein containing an interdomain disulfide bridge confers a dominant-negative phenotype for transport and chemotaxis
J. Biol. Chem.
1996
, vol. 
271
 (pg. 
17881
-
17889
)
[PubMed]
29
Jencks
W.P.
From chemistry to biochemistry to catalysis to movement
Annu. Rev. Biochem.
1997
, vol. 
66
 (pg. 
1
-
18
)
[PubMed]
30
Klingenberg
M.
Transport viewed as a catalytic process
Biochimie
2007
, vol. 
89
 (pg. 
1042
-
1048
)
[PubMed]
31
Law
C.J.
Yang
Q.
Soudant
C.
Maloney
P.C.
Wang
D.-N.
Kinetic evidence is consistent with the rocker-switch mechanism of membrane transport by GlpT
Biochemistry
2007
, vol. 
46
 (pg. 
12190
-
12197
)
[PubMed]
32
Law
C.J.
Maloney
P.C.
Wang
D.-N.
Ins and outs of major facilitator superfamily antiporters
Annu. Rev. Microbiol.
2008
, vol. 
62
 (pg. 
289
-
305
)
[PubMed]