Alternative proton-binding site and long-distance coupling in Escherichia coli sodium–proton antiporter NhaA

Edited by Ernest M. Wright, David Geffen School of Medicine at UCLA, Los Angeles, CA, and approved August 11, 2020 (received for review March 23, 2020)
September 24, 2020
117 (41) 25517-25522


NhaA is the Escherichia coli sodium–proton antiporter responsible for maintaining cellular ion and volume homeostasis; however, despite two decades of research, the electrogenic transport mechanism remains poorly understood. To address an ongoing debate centering on the identity of the second proton-binding site, we carried out state-of-the-art molecular dynamics simulations of several NhaA mutants in which the proton-binding residue in question is mutated. Our simulations suggested that the mutants maintain the electrogenic transport of NhaA by utilizing an alternative proton-binding aspartate and in some cases two distant aspartates that can jointly bind one or two protons. Alternative proton-binding site and proton sharing between distant aspartates may represent important general mechanisms of proton-coupled transport in secondary active transporters.


Escherichia coli NhaA is a prototypical sodium–proton antiporter responsible for maintaining cellular ion and volume homeostasis by exchanging two protons for one sodium ion; despite two decades of research, the transport mechanism of NhaA remains poorly understood. Recent crystal structure and computational studies suggested Lys300 as a second proton-binding site; however, functional measurements of several K300 mutants demonstrated electrogenic transport, thereby casting doubt on the role of Lys300. To address the controversy, we carried out state-of-the-art continuous constant pH molecular dynamics simulations of NhaA mutants K300A, K300R, K300Q/D163N, and K300Q/D163N/D133A. Simulations suggested that K300 mutants maintain the electrogenic transport by utilizing an alternative proton-binding residue Asp133. Surprisingly, while Asp133 is solely responsible for binding the second proton in K300R, Asp133 and Asp163 jointly bind the second proton in K300A, and Asp133 and Asp164 jointly bind two protons in K300Q/D163N. Intriguingly, the coupling between Asp133 and Asp163 or Asp164 is enabled through the proton-coupled hydrogen-bonding network at the flexible intersection of two disrupted helices. These data resolve the controversy and highlight the intricacy of the compensatory transport mechanism of NhaA mutants. Alternative proton-binding site and proton sharing between distant aspartates may represent important general mechanisms of proton-coupled transport in secondary active transporters.
Cation–proton antiporters (CPAs) play important roles in pH, ion, and volume homeostasis by exchanging one (CPA1) or two (CPA2) protons for typically one sodium or potassium ion (14). One of the best-studied CPAs is Escherichia coli NhaA, which serves as a prototype for understanding the function of CPA2 family members. Over the past two decades, the ion transport mechanism of NhaA has been probed by a large number of biochemical and physiology experiments (1, 2), and two atomic-resolution crystal structures in the inward-facing conformation at low pH have been resolved (5, 6). The crystal structures show that the NhaA monomer is composed of a core and a dimer domain totaling 12 transmembrane (TM) helices. The ion-binding site, consisting of Asp163, Asp164, Lys300, and Thr132, is located in the middle of the core domain and near the intersection of two antiparallel discontinuous helices, TM4 and TM11 (Fig. 1).
Fig. 1.
Structure of NhaA (PDB 4au5, ref. 6) and a zoomed-in view of the ion-binding site. Important transmembrane helices and residues discussed in this work are labeled.
An ongoing debate about the NhaA antiport mechanism centers on the question of which two residues are involved in proton binding. General consensus is that 1) a competitive binding mechanism is at work whereby two protons and one sodium ion cannot be bound at the same time, 2) the sodium ion binds to Asp164, and 3) Asp164 is one of the proton-binding residues. However, the identity of the second proton-binding residue remains controversial. The “two-aspartate” model (4, 79), which was proposed after the first crystal structure of inactive NhaA was obtained (5), assumes that the second proton is carried by Asp163. In this model, Lys300 primarily fulfills a role of stabilizing the protein. The more recent “salt-bridge model” (6, 10), which was proposed after the second crystal structure of inactive NhaA was resolved (6), suggests that the second proton is carried by Lys300, which is stabilized by a salt bridge with the deprotonated Asp163. Binding of sodium to Asp164 and Asp163 destabilizes the salt bridge and lowers the pKa of Lys300 so that it releases the second proton.
Perhaps the best evidence for the two-aspartate model comes from the recent solid supported membrane (SSM)-based electrophysiology measurements of Lys300 mutants. In particular, the K300Q and K300A mutants showed electrogenic transport, albeit with a maximum current decreased to 2.5 and 7.5%, respectively (8, 9). Substitution with a charged residue, such as the K300R and K300H mutants, also produced electrogenic transport, albeit with a maximum current decreased to 13.3 and 18.3%, respectively (8). Interestingly, the double-mutant K300Q/D163N also showed electrogenic activity, although the maximum current was further decreased to 0.8% (9). To explain the latter, Fendler and coworkers (9) hypothesized that Asp133 may play the role of an auxiliary proton-binding site. Circular dichroism (CD) measurements demonstrated reduced thermal stability for all of the Lys300 mutants, and thus Lys300 was assigned a primarily structural role (8, 9). Most recently, a broad evolutionary analysis of CPAs revealed eight highly conserved residues, among which positions 6 and 7 correspond to Asp163 and Asp164 in NhaA and position 8 corresponds to Lys300 in NhaA (4). A subsequent experiment of the triple-mutant P108E/A106S/D163N showed electroneutral transport, thus offering a renewed support to the two-aspartate model.
Evidence for the salt-bridge model comes from the Asp163-Lys300 salt bridge in the second crystal structure of NhaA (6) as well as the analogous salt bridge (Asp156-Lys305) in the crystal structures of the homologous NapA from Thermus thermophilus in the inward- (11) and outward-facing states (12). The salt-bridge model is also supported by the recent fixed-charge (6) and continuous constant pH molecular dynamics (CpHMD) studies (10). The latter work (10) showed that the pKa of Asp164 is the highest among the active-site aspartates, confirming its role as a proton-binding residue; in contrast, the pKa of Asp163 is always lower than the solution value of 4, due to either the electrostatic interaction with Lys300 or sodium binding, indicating that it cannot accept a proton. Interestingly, the pKa of Lys300 depends on sodium binding. In the absence of sodium binding, Lys300 forms a salt bridge with Asp163, which stabilizes the charged form of Lys, leading to an upshift of the pKa to around 11.6, compared to the solution pKa of 10.4. However, when a sodium ion binds to Asp163 and Asp164, the salt bridge is disrupted, resulting in a downshift of the pKa by about 2.7 units to 8.9. In the presence of sodium binding, the simulated titration curve shows that lysine starts to deprotonate at a pH around 7 and the occupancy of the deprotonated state increases with pH. Taking into account the slight overestimation of the lysine pKa by simulation (10), the titration behavior of lysine is well correlated with the measured pH profile of the sodium current: no current at pH below 6.5 and a dramatic increase between pH 7 and 8.5 (7, 13, 14). Therefore, together with the calculated pKas of the aspartates, the CpHMD simulation suggested that Asp164 and Lys300 are the two proton-binding sites (10).
Given the strong evidence for Lys300 being the second proton-binding residue, why do K300 mutants retain electrogenic transport? Intrigued by the apparent contradiction, we set out to investigate the protonation states of key aspartic residues in the K300A, K300R, and K300Q/D163N mutants, using the membrane-enabled hybrid-solvent CpHMD method with pH replica exchange (10, 15), which has been previously applied to investigate the activation mechanism of NhaA and elucidate the role of Lys300 (10). Our simulations suggest that an alternative proton-binding residue, Asp133, can accept a proton when Lys300 is mutated. Surprisingly, titration of Asp133 is coupled to Asp163 in K300A or Asp164 in K300Q/D163N. To provide a testable hypothesis, simulations were also conducted for a triple-mutant K300Q/D163N/D133A.

Results and Discussion

Calculated pKas Confirm Asp164 as the First Proton Carrier and Suggest Asp133 Is the Second Proton Carrier in the K300 Mutants.

The membrane-enabled hybrid-solvent CpHMD simulations with pH replica exchange (10, 15) were initiated from the crystal structure of NhaA (Protein Data Bank [PDB] 4au5, ref. 6). The production runs contained 16 replicas in the pH range 2.5 to 8 and simulations continued for 28 to 45 ns per replica until the protonation state fractions at all pH values (and correspondingly the pKas) of the relevant residues were converged. The aggregate sampling time was 432 to 688 ns for each mutant. Convergence analysis and replica-exchange statistics are provided in SI Appendix, Table S1 and Figs. S1–S6. Consistent with our previous CpHMD simulations of membrane transporters (16, 17), the pH replica-exchange protocol significantly accelerated the pKa convergence, as the direct coupling between conformation and titration coordinates allows random walks in the pH space to enhance the sampling of both conformational and protonation states (15).
Table 1 summarizes the calculated pKa values of the active-site aspartates in the wild-type (WT) and K300 mutants of NhaA. Asp164 is a consensus proton-binding residue and, in agreement with this role, shows a pKa of 5.1 in the WT and 5.0 to 5.5 in the mutants, higher than the model (solution) pKa value of 4. Asp163 has a depressed pKa of 2.4 in WT (6, 10), making it unlikely to be protonated in the periplasmic pH of about 6.5. In K300A, the pKa of Asp163 is higher than in WT due to the lack of salt-bridge interaction with Lys300; however, it remains lower than the model value of 4, due to hydrogen bonding with Thr132 and Ala131 (see later discussion). In K300R, the pKa of Asp163 is even lower than in WT due to the salt-bridge interactions with Arg300. Thus, the low pKa of Asp163 makes it implausible to accept a proton at relevant pH. Interestingly, the pKa of Asp133 remains at 4.8, similar to the pKa of Asp164. These data support the hypothesis (9) that Asp133 acts as an alternative proton acceptor in the absence of Lys300; i.e., in the absence of the salt-bridge mechanism, Asp133 assumes the role of the second proton carrier. Fendler and coworkers (9) invoked this hypothesis to explain the electrogenic transport of the double-mutant K300Q/D163N, but our data suggest that it is also operative for the single mutants of K300.
Table 1.
Calculated pKas of the active-site aspartic acid residues in the WT and K300 mutant of NhaA
n/a, not applicable.
Estimated pKa may not be very accurate, as it is at or below the lowesttitration pH.
Macroscopic pKas of the coupled titration of Asp133/Asp163.
Macroscopic pKas of the coupled titration of Asp133/Asp164.
Asp133 has been repeatedly implicated as an essential residue for the stability and function of NhaA. Various D133 mutants were highly unstable (18); D133N resulted in the loss of antiporter activity (19) and D133C, D133A, and D133K mutations significantly impeded binding of sodium or lithium ions (18, 20, 21). Asp133 is a well-conserved residue next to the sodium-binding residue Thr132 (10, 21) and is likely accessible from either periplasm or the cytoplasmic side when the protein undergoes the alternating-access conformational change. If Asp133 indeed plays the role of an alternative proton acceptor, then mutating it in addition to removing K300 should lead to an electroneutral transporter. The CpHMD titration of a triple-mutant K300Q/D163N/D133A shows that Asp164, the only remaining ionizable residue in the active site, has a pKa of 5.3, which is 0.3 units higher than that of the WT. We hypothesize that although the simultaneous mutation of Asp133 and Asp163 would significantly impede ion binding, this triple mutant may still be able to transport sodium in exchange for one proton.

Titration of Asp133 Is Coupled to Asp163 in the Single-Mutant K300A NhaA.

Coupled proton titration is a phenomenon, whereby protonation/deprotonation of two residues affects each other, resulting in irregular titration curves and the necessity to invoke a two-proton model for calculating macroscopic pKas (22). Typically, the coupled residues are hydrogen bonded to each other, such as the aspartyl dyad in enzymes (23, 24), or spatially proximal to each other, such as the adjacent Asp407/Asp408 in the E. coli multidrug efflux transporter AcrB, which pumps out one drug molecule for two protons (17). Interestingly, we found that the titration curve of Asp163 in K300A NhaA is irregular (SI Appendix, Fig. S7), prompting us to test coupled titration between all pairs of titratable residues in the ion-binding site of K300A, K300R, and WT NhaA. While no coupling between any residue pairs was found for K300R and WT NhaA, Asp133/Asp163 in K300A NhaA was found to influence each other’s titration, even though the Cγ distance is about 13 Å in the crystal structure (Fig. 1). The titration curve of Asp163 shifts its position dependent on the protonation state of Asp133. When Asp133 is deprotonated, the pKa of Asp163 is shifted higher, indicating that it is more likely to be protonated; conversely, when Asp133 is protonated, the pKa of Asp163 is shifted lower, indicating that it is more likely to be deprotonated (SI Appendix, Fig. S7). A similar behavior is seen for the titration of Asp133, which is dependent on the protonation state of Asp163 (SI Appendix, Fig. S7).
In light of the coupling between Asp133 and Asp163 in the K300A mutant, we calculated the macroscopic pKas using the coupled two-proton model, which resulted in pK1/pK2 of 5.5/3.5 (Table 1). The splitting between pK1 and pK2 is 2, which is 1 unit larger than the splitting between the site-specific pKas of 4.8 and 3.8, reflecting the aforementioned anticooperative behavior of Asp133 and Asp163 titrations. Below we focus on pK1, as it represents the pKa of the relevant protonation event where pH is decreased from high (cytoplasmic) to low (periplasmic).
We examined the details of proton coupling by calculating the probabilities of having zero proton (D133/D163), one proton (D133H/D163 or D133/D163H), and two protons (D133H/D163H). Interestingly, as pH is lowered from 8 to about 5.5 (pK1), either Asp133 or Asp163 accepts a proton, as demonstrated by the equal probabilities of the two singly protonated states (Fig. 2A). These data indicate that both Asp133 and Asp163 are responsible for binding the second proton in K300A NhaA (with Asp164 being the first proton-binding residue). As a comparison, Fig. 2B shows a similar analysis for K300R NhaA. As pH is lowered from 8, the singly protonated state is exclusively D133H/D163, which indicates that Asp133 is solely responsible for binding the second proton as pH is lowered from cytoplasmic to periplasmic pH in the K300R mutant. This is expected, as Asp163 is always deprotonated due to the salt-bridge interaction with Arg300 until low pH conditions.
Fig. 2.
Titration of Asp133 and Asp163 is coupled in K300A but not in K300R or WT NhaA. (A and B) Probabilities of the doubly deprotonated (P0, gray), (combined) singly protonated (P1, purple), and doubly (P2, red) protonated states of Asp133/Asp163 in NhaA mutants K300A (A) and K300R (B). Similar data for WT NhaA are shown in SI Appendix, Fig. S8. Probabilities of the two separate singly protonated states, D133H/D163 and D133/D163H, are shown in blue and green, respectively. pK1 corresponds to the crossing between P0 and P1 curves. (C and D) Snapshots showing the hydrogen bonds around Asp133 and Asp163 in K300A NhaA, when Asp163 is deprotonated and Asp133 is protonated (C) and when Asp163 is protonated and Asp133 is deprotonated (D).
Given the long distance between Asp133 and Asp163, how is the coupling realized? Our analysis revealed that the coupling is communicated through the flexible intersection of TM4 and TM11, whereby changes in the protonation states of Asp133 and Asp163 affect their hydrogen-bonding patterns. Consistent with our previous findings that deprotonated aspartates tend to be stabilized by accepting hydrogen bonds (23, 24), the deprotonated Asp163 is a hydrogen bond acceptor to the sidechain hydroxyl of Thr132 and the backbone amides of Ala131 as well as Thr132 (Fig. 2C). However, when Asp133 becomes deprotonated, it forms hydrogen bonds with the sidechain hydroxyl and backbone amide of Thr340, which pulls on the linker region of TM4, preventing TM4 residues Thr132 and Ala131 from forming hydrogen bonds with Asp163 (Fig. 2D and SI Appendix, Fig. S9).

Titration of Asp133 Is Coupled to Asp164 in the Double-Mutant K300Q/D163N NhaA.

Surprisingly, we found that the titration of Asp133 is coupled to that of Asp164 in the double-mutant K300Q/D163N, with the two macroscopic pKas of 5.9 and 4.4 (Table 1). Compared to the site-specific pKas of 5.5 for Asp164 and 4.8 for Asp133, the splitting between the macroscopic pKas of 1.5 units is much larger, suggesting a high degree of anticooperativity. We performed the analysis for the coupled protonation events of Asp133/Asp164 (Fig. 3). As pH is lowered from 8 to 5.9 (pK1), the first proton is accepted mainly by Asp164, as demonstrated by the much higher probability of D133/D164H compared to D133H/D164 (Fig. 3A). As pH is further lowered to 4.4 (pK2), the second proton is accepted, as demonstrated by the sizable probability of the doubly protonated state. This analysis suggests that Asp164 remains the first proton-binding residue and Asp133 participates in binding the second proton in K300Q/D163N NhaA. However, the second proton-binding event is shifted to a lower pKa, compared to K300A NhaA, which binds the second proton with a pK of 5.5 (attributable to Asp133/Asp163), and K300R NhaA, which binds the second proton with a pK of 4.8 (attributable to Asp133). We suggest this may be a contributor to the extremely low peak current of the double mutant, in addition to the reduced sodium-binding affinity due to the mutation of Asp163.
Fig. 3.
Coupled titration of Asp133 and Asp164 in K300Q/D163N NhaA. (A) Probabilities of the doubly deprotonated (P0, gray), (combined) singly protonated (P1, purple), and doubly protonated (P2, red) states of Asp133/Asp164 in the double mutant. Probabilities of the two separate singly protonated states, D133H/D164 and D133/D164H, are shown in blue and green, respectively. pK1 and pK2 correspond to the crossings between P0 and P1 and between P1 and P2 curves, respectively. (B) Probability of the protonated Asp164 acting as a hydrogen bond donor and acceptor to Thr132, when Asp133 is deprotonated and accepting at least one hydrogen bond from Thr340. (C and D) Snapshots showing hydrogen bonds around Asp133 and Asp164 when Asp133 is deprotonated and Asp164 is protonated (C) and when Asp133 is protonated and Asp164 is deprotonated (D).
The mechanism for the coupling between Asp133 and Asp164 in K300Q/D163N is similar to that of the coupling between Asp133 and Asp163 in K300A. Analysis showed that when Asp164 is protonated, it can both accept and donate a hydrogen bond with the backbone amide and carbonyl of Thr132, and this configuration occurs when Asp133 is deprotonated and accepts a hydrogen bond from Thr340 (Fig. 3 B and C). However, when Asp133 becomes protonated, it no longer forms a hydrogen bond with Thr340, which affects the linker region of TM4, changing the position of Thr132 and ending the bidentate hydrogen bond with Asp164 (Fig. 3D).
Finally, we examined the conformational dynamics of the mutants during the simulations. While the overall structures of K300R and K300Q/D163N stayed close to that of the WT, K300A showed significant conformational change during the simulation (SI Appendix, Fig. S10). TM5 and TM10 stayed in proximity during the simulations of K300R and K300Q/D163N, due to the R300-D163 salt bridge or the water-mediated hydrogen bonds between Gln300 and Asn163; however, TM5 became bent and displaced TM10 in the simulation of K300A NhaA (Fig. 4). This conformational change is consistent with our previous simulations of WT NhaA, which showed bending of TM5 following deprotonation of Lys300 (10).
Fig. 4.
Simulation of K300A NhaA showed major conformational rearrangements. (A) Overlay of the K300A (red) and WT (blue) NhaA structures. For clarity, loop regions are hidden. (B) Zoomed-in view of the active sites of the K300A (red) and WT (blue) NhaA.

Concluding Discussion

To resolve the controversy between the two-aspartate and salt-bridge models for the antiport mechanism of NhaA, we investigated the protonation states of the active-site aspartic residues in the K300A, K300R, K300Q/D163N, and K300Q/D163N/D133A mutants using the state-of-the-art CpHMD simulations in the explicit lipid bilayer. Our data confirmed that Asp164 is the first proton-binding site and offered direct evidence to support the hypothesis (9) that Asp133 can serve as an alternative second proton site in place of Lys300 in K300 NhaA mutants. While Asp133 is solely responsible for binding the second proton in K300R NhaA, surprisingly, our simulations revealed coupled proton transfer mechanisms in two mutants, K300A and K300Q/D163N. In K300A NhaA, the titration of Asp133 is coupled to Asp163 such that they both participate in binding the second proton. The involvement of Asp163 lends support to the two-aspartate model, implicating that Asp163 can play the role of the second proton carrier in some circumstances. In K300Q/D163N NhaA, Asp164 and Asp133 jointly bind two protons, with Asp164 being primarily responsible for binding a proton at higher pH.
We note that the current study and our previous work (10) are based on the inward-facing crystal structures of NhaA. A question arises as to whether the present interpretations would be supported by the calculated pKas of the active-site residues in the outward-facing state. Although such calculations are not conducted due to the lack of an outward-facing crystal structure of NhaA, comparison of the crystal structures of the analogous NapA in the inward- and outward-facing states suggests that the order of pKas in the two states is likely the same; i.e., Asp164 (Asp157 in NapA) > Asp163 (Asp156 in NapA), and the pKa of Lys300 (Lys305 in NapA) remains above the model value of 10.4. Thus, the same two residues that are suggested as proton donors for the cytoplasmic inward-facing state are proton acceptors for the periplasmic outward-facing state. In the latter scenario, Asp164 (Asp157 in NapA) can bind a proton, whereas Asp163 (Asp156 in NapA) cannot because its pKa is too low. In contrast, Lys300 (Lys305 in NapA) can bind a proton to form a salt bridge with Asp163 (Asp156 in NapA).
Our data are consistent with the functional studies showing that the K300 mutants maintain the electrogenic transport function of NhaA (8, 9) and suggest that while the salt-bridge model involving Lys300 as the second proton acceptor (6, 10, 11) is operative for the WT NhaA, K300 mutants may invoke Asp133 alone (e.g., in K300R) or Asp133 and Asp163 together (e.g., in K300A) as the second proton acceptor. The K300 double mutant, e.g., K300Q/D163N, can also utilize Asp164 and Asp133 together for electrogenic transport. Thus, our work suggests a revision of the two-aspartate model (4, 79), in that Asp164 and Asp133, and sometimes jointly with Asp163, are responsible for the electrogenic transport of K300 mutants. This revised model may also explain why the K305Q mutant of the analogous NapA supports electrogenic transport (9), as Glu333, which is located on TM10 and occupies a similar region of the flexible linker as Asp133 in NhaA, may serve the role of a substitute proton carrier. Finally, our simulations also produced an experimentally testable hypothesis that the triple-mutant K300Q/D163N/D133A may conduct electroneutral transport of sodium albeit low activity. Note that the conclusion of this work is based on the simulations and experimental data of WT, single, and double mutants of NhaA; experimental verification of the triple-mutant hypothesis would provide additional data but is not in the scope of our work.
Our data are compatible with the recent broad evolutionary analysis of CPAs (4), according to which both Asp163 and Lys300 are among the eight specificity-determining residues in CPAs. However, why does a triple-mutant P108E/A106S/D163N NhaA show electroneutral transport (4)? As an alternative explanation to the two-aspartate model, we suggest that, in this triple mutant, Lys300 is unable to lose a proton because its salt-bridge partner Glu108 does not bind a sodium (4), which is a requirement for the salt-bridge disruption and subsequent deprotonation of Lys300, as demonstrated in our previous simulation work (10). Another seeming contradiction with the evolutionary analysis concerns the observation that 90% of electroneutral CPA1 electroneutral antiporters have a residue equivalent to Asp133 in NhaA (4). In the absence of CpHMD simulations of a CPA1 antiporter and considering our data of K300Q/D163N NhaA, we speculate that the second proton-binding event, in which Asp133 is coupled to Asp164, may occur at a very low pH condition such that it is not relevant to experimental measurements of electrogenicity.
Point mutations are frequently used to probe or infer proton-binding residue(s) in mechanistic studies of secondary active transporters. Our work demonstrated that the WT NhaA function can be restored by a mutant transporter through a compensatory mechanism, in which an alternative residue or a coupled residue pair takes on the role of proton binding/release. A related observation was made recently for a CPA1 antiporter MjNhaP1, whereby the ion selectivity was found to be robustly encoded in the amino acids of the first and the second shell (25). Alternative proton-binding site and coupling between key aspartates highlight the intricacy of the compensatory transport mechanism of NhaA. Coupled protonation by two adjacent aspartates is common (26) and has been demonstrated in the one-proton exchange mechanism of a multidrug transporter AcrB by a simulation study (17); however, the coupling between two distant aspartates enabled through the proton-coupled hydrogen-bonding network at the flexible crossing of two disrupted helices (as in K300A and K300Q/D163N NhaA) has not been demonstrated before. We speculate that alternative proton-binding site and long-distance coupling may represent important general mechanisms of proton-coupled transport in secondary active transporters.
A recent MD study using a transition path shooting method revealed a putative inward-to-outward conformational transition of the CPA1 antiporter PaNhaP (27). CpHMD simulations of a kinase protein demonstrated that the process of large conformational changes can be facilitated by a shift in the protonation-state populations (28). Thus, it is conceivable that the transition and/or intermediate states of NhaA may have different protonation states. Detailed proton-coupled antiport mechanisms of CPAs await future exploration by combining CpHMD and advanced sampling protocols.

Materials and Methods

System Preparation.

The structures of the mutants, K300A, K300R and K300Q/D163N, K300Q/D163N/D133A were built from the final snapshot of the equilibration simulation of WT NhaA (10), which started from the monomer B of the inward-facing crystal structure of NhaA (PDB 4au5) (6). The protein was embedded in a bilayer composed of 135 (63 upper and 72 lower leaflets) 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC) lipid molecules. Both sides of the lipid bilayer were hydrated with a 15-Å slab of explicit water. Sodium and chloride ions were added to neutralize the system at physiological pH and reach the ionic strength of 0.15 M. The K300A mutation was made by manually truncating the Lys sidechain, while the other mutations were made using the mutation command in the MMTSB toolset (29). The number of ions in the solution was adjusted accordingly. Each mutant system underwent energy minimization using 50 steps of steepest descent followed by 10 steps of adopted basis Newton–Raphson algorithms. Other details of system preparation are given in our previous work (10).

Continuous CpHMD Simulations.

Membrane-enabled hybrid-solvent continuous CpHMD (10, 15) simulations were conducted using the CHARMM program version c42a2 (30). Following the energy minimization, the mutant systems were briefly equilibrated using CpHMD at pH 4 (crystallization pH) for 2.4 ns, whereby the harmonic restraints on the protein heavy atoms were reduced from 1 kcalmol−1Å−1 to zero. The systems were then subjected to pH replica-exchange CpHMD simulations. Sixteen replicas in the pH range 2.5 to 8 were used for the mutant simulations. Each replica was simulated under constant NPT conditions at 310 K and 1 atm. In the replica-exchange protocol, an attempt to swap between adjacent pH replicas was made every 500 MD steps. The aggregate sampling time was 423 ns for the WT, 688 ns for K300A, 448 ns for K300R, 432 ns for K300Q/D163N, and 392 ns for K300Q/D163N/D133A. We note that within these simulation times, the heavy atom root-mean-square deviations (rmsd) and pKas of relevant residues were converged (SI Appendix, Figs. S1 and S2). All analyses were based on the last 240 ns of mutant simulations.
In the membrane-enabled hybrid-solvent CpHMD method (10), conformational sampling is conducted in a fully explicit solvent and lipid bilayer, while the forces on the titration coordinates are calculated with the membrane GBSW implicit-solvent model (31). Mixing solvent models is an approximation, and one potential artifact is a spike in potential energy due to the lack of solvent relaxation after a λ update (15). To avoid this, a GBSW calculation and update of titration coordinates are executed every five MD steps, as in our previous NhaA simulations (10) and other work (16). Another potential artifact that is perhaps more subtle is related to the lack of a formal Hamiltonian in the mixed-solvent simulations. As a consequence, the use of a Hamiltonian (pH) replica-exchange protocol may result in a small bias in the conformational distributions. A related thermostat artifact has been previously demonstrated in temperature replica-exchange simulations of peptide folding (32). This is a topic that warrants a detailed investigation in the future.
For the membrane GBSW model (31), the thickness of the implicit bilayer was calculated as 30 Å, using the average distance between the C2 atoms of the lipids in the cytoplasmic- and periplasmic-facing leaflets. The implicit membrane was excluded from the interior of the protein using an exclusion cylinder with a radius of 15 Å placed in the center of the protein. This distance provided maximum coverage of the protein while overlapping with as few lipids as possible. A switching distance of 5 Å was used for the transition between the low dielectric slab and bulk solvent. All other settings used the default options, consistent with our previous work (10, 15).

Molecular Dynamics Protocol.

The protein was represented by the CHARMM22/CMAP all-atom force field (33, 34), the lipids were represented by the CHARMM36 lipid force field (35), and water was represented with the CHARMM modified TIP3P water model (30). Temperature was maintained at 310 K using a modified Hoover thermostat (36), and the pressure was maintained with the Langevin piston coupling method (37). The van der Waals interactions were smoothly switched to zero between 10 and 12 Å. The particle mesh Ewald method (38) was used to calculate long-range electrostatics, with a real space cutoff of 12 Å and a sixth-order interpolation with a 0.9-Å grid spacing. All bonds involving hydrogen were restrained using CHARMM’s SHAKE algorithm (39), allowing for a 2-fs time step.

pKa Analysis.

Following Ullmann (40) and our previous work (41), we write the probabilities of the doubly deprotonated (P0), singly protonated (P1), and doubly protonated (P2) states for a coupled titration pair (e.g., Asp163/Asp133) as
where pK1 refers to the macroscopic pKa for gaining the first proton, e.g., D133/D163D133/D163H or D133H/D163; and pK2 refers to the macroscopic pKa for gaining the second proton, e.g., D133/D163H or D133H/D163D133H/D163H; and Z is the partition function
The total number of bound protons is therefore (40, 41)
Fitting of P vs. pH to the above equation gives the two macroscopic pKas.

Data Availability

All study data are included in this article and SI Appendix, Figs. S1–S10 and Table S1.


We acknowledge financial support provided by the National Institutes of Health (R01GM118772).

Supporting Information

Appendix (PDF)


E. Padan, Functional and structural dynamics of NhaA, a prototype for Na+ and H+ antiporters, which are responsible for Na+ and H+ homeostasis in cells. Biochim. Biophys. Acta 1837, 1047–1062 (2014).
E. Padan, M. Landau, Sodium-proton (Na+/H+) antiporters properties and roles in health and disease. Met. Ions Life Sci. 16, 391–458 (2016).
C. L. Brett, M. Donowitz, R. Rao, Evolutionary origins of eukaryotic sodium/proton exchangers. Am. J. Physiol. Cell Physiol. 288, C223–C239 (2005).
G. Masrati et al., Broad phylogenetic analysis of cation/proton antiporters reveals transport determinants. Nat. Commun. 9, 4205 (2018).
C. Hunte et al., Structure of a Na+/H+ antiporter and insights into mechanism of action and regulation by pH. Nature 435, 1197–1202 (2005).
C. Lee et al., Crystal structure of the sodium-proton antiporter NhaA dimer and new mechanistic insights. J. Gen. Physiol. 144, 529–544 (2014).
T. Mager, A. Rimon, E. Padan, K. Fendler, Transport mechanism and pH regulation of the Na+/H+ antiporter NhaA from Escherichia coli. J. Biol. Chem. 286, 23570–23581 (2011).
O. Călinescu, M. Dwivedi, M. Patiño-Ruiz, E. Padan, K. Fendler, Lysine 300 is essential for stability but not for electrogenic transport of the Escherichia coli NhaA Na+/H+ antiporter. J. Biol. Chem. 292, 7932–7941 (2017).
M. Patiño-Ruiz et al., Replacement of Lys-300 with a glutamine in the NhaA Na+/H+ antiporter of Escherichia coli yields a functional electrogenic transporter. J. Biol. Chem. 294, 246–256 (2019).
Y. Huang, W. Chen, D. L. Dotson, O. Beckstein, J. Shen, Mechanism of pH-dependent activation of the sodium-proton antiporter NhaA. Nat. Commun. 7, 12940 (2016).
C. Lee et al., A two-domain elevator mechanism for sodium/proton antiport. Nature 501, 573–577 (2013).
M. Coincon et al., Crystal structures reveal the molecular basis of ion translocation in sodium/proton antiporters. Nat. Struct. Mol. Biol. 23, 248–255 (2016).
D. Taglicht, E. Padan, S. Schuldiner, Overproduction and purification of a functional Na+/H+ antiporter coded by nhaA (ant) from Escherichia coli. J. Biol. Chem. 266, 11289–11294 (1991).
T. Mager et al., Differential effects of mutations on the transport properties of the Na+/H+ antiporter NhaA from Escherichia coli. J. Biol. Chem. 288, 24666–24675 (2013).
J. A. Wallace, J. K. Shen, Continuous constant pH molecular dynamics in explicit solvent with pH-based replica exchange. J. Chem. Theory Comput. 7, 2617–2629 (2011).
W. Chen, B. H. Morrow, C. Shi, J. K. Shen, Recent development and application of constant pH molecular dynamics. Mol. Simul. 40, 830–838 (2014).
Z. Yue, W. Chen, H. I. Zgurskaya, J. Shen, Constant pH molecular dynamics reveals how proton release drives the conformational transition of a transmembrane efflux pump. J. Chem. Theory Comput. 13, 6405–6414 (2017).
A. Rimon, M. Dwivedi, A. Friedler, E. Padan, Asp133 residue in NhaA Na+/H+ antiporter is required for stability cation binding and transport. J. Mol. Biol. 430, 867–880 (2018).
H. Inoue, T. Noumi, T. Tsuchiya, H. Kanazawa, Essential aspartic acid residues, Asp-133, Asp-163 and Asp-164, in the transmembrane helices of a Na+/H+ antiporter (NhaA) from Escherichia coli. FEBS Lett. 363, 264–268 (1995).
L. Galili, A. Rothman, L. Kozachkov, A. Rimon, E. Padan, Transmembrane domain IV is involved in ion transport activity and pH regulation of the NhaA-Na+/H+ antiporter of Escherichia coli. Biochemistry 41, 609–617 (2002).
M. Maes, A. Rimon, L. Kozachkov-Magrisso, A. Friedler, E. Padan, Revealing the ligand binding site of NhaA Na+/H+ antiporter and its pH dependence. J. Biol. Chem. 287, 38150–38157 (2012).
H. Webb et al., Remeasuring HEWL pKa values by NMR spectroscopy: Methods, analysis, accuracy, and implications for theoretical pKa calculations. Proteins 79, 685–702 (2011).
Y. Huang, Z. Yue, C. C. Tsai, J. A. Henderson, J. Shen, Predicting catalytic proton donors and nucleophiles in enzymes: How adding dynamics helps elucidate the structure-function relationships. J. Phys. Chem. Lett. 9, 1179–1184 (2018).
R. C. Harris, J. Shen, GPU-accelerated implementation of continuous constant pH molecular dynamics in amber: Predictions with single-pH simulations. J. Chem. Inf. Model. 59, 4821–4832 (2019).
J. Warnau et al., Ion binding and selectivity of the Na +/H + antiporter MjNhaP1 from experiment and simulation. J. Phys. Chem. B 124, 336–344 (2020).
J. Shen, Zooming in on a small multidrug transporter reveals details of asymmetric protonation. Proc. Natl. Acad. Sci. U.S.A. 115, 8060–8062 (2018).
K.-i. Okazaki et al., Mechanism of the electroneutral sodium/proton antiporter PaNhaP from transition-path shooting. Nat. Commun. 10, 1742 (2019).
C. C. Tsai, Z. Yue, J. Shen, How electrostatic coupling enables conformational plasticity in a tyrosine kinase. J. Am. Chem. Soc. 141, 15092–15101 (2019).
M. Feig, J. Karanicolas, C. L. Brooks III, MMTSB tool set: Enhanced sampling and multiscale modeling methods for applications in structural biology. J. Mol. Graph. Model. 22, 377–395 (2004).
B. R. Brooks et al., CHARMM: The biomolecular simulation program. J. Comput. Chem. 30, 1545–1614 (2009).
W. Im, M. Feig, C. L. Brooks III, An implicit membrane generalized Born theory for the study of structure, stability, and interactions of membrane proteins. Biophys. J. 85, 2900–2918 (2003).
E. Rosta, N. V. Buchete, G. Hummer, Thermostat artifacts in replica exchange molecular dynamics simulations. J. Chem. Theory Comput. 5, 1393–1399 (2009).
A. D. MacKerell Jr et al., All-atom empirical potential for molecular modeling and dynamics studies of proteins. J. Phys. Chem. B 102, 3586–3616 (1998).
A. D. MacKerell Jr, M. Feig, C. L. Brooks III, Extending the treatment of backbone energetics in protein force fields: Limitations of gas-phase quantum mechanics in reproducing protein conformational distributions in molecular dynamics simulations. J. Comput. Chem. 25, 1400–1415 (2004).
J. B. Klauda et al., Update of the CHARMM all-atom additive force field for lipids: Validation on six lipid types. J. Phys. Chem. B 114, 7830–7843 (2010).
W. G. Hoover, Canonical dynamics: Equilibration phase-space distributions. Phys. Rev. A 31, 1695–1697 (1985).
S. E. Feller, Y. Zhang, R. W. Pastor, B. R. Brooks, Constant pressure molecular dynamics simulation: The Langevin piston method. J. Chem. Phys. 103, 4613–4621 (1995).
T. Darden, D. York, L. Pedersen, Particle mesh Ewald: An N log(N) method for Ewald sums in large systems. J. Chem. Phys. 98, 10089–10092 (1993).
J. P. Ryckaert, G. Ciccotti, H. J. C. Berendsen, Numerical integration of the Cartesian equations of motion of a system with constraints: Molecular dynamics of n-alkanes. J. Comput. Phys. 23, 327–341 (1977).
G. M. Ullmann, Relations between protonation constants and titration curves in polyprotic acids: A critical view. J. Phys. Chem. B 107, 1263–1271 (2003).
J. A. Wallace, J. K. Shen, Charge-leveling and proper treatment of long-range electrostatics in all-atom molecular dynamics at constant pH. J. Chem. Phys. 137, 184105 (2012).

Information & Authors


Published in

Go to Proceedings of the National Academy of Sciences
Go to Proceedings of the National Academy of Sciences
Proceedings of the National Academy of Sciences
Vol. 117 | No. 41
October 13, 2020
PubMed: 32973095


Data Availability

All study data are included in this article and SI Appendix, Figs. S1–S10 and Table S1.

Submission history

Published online: September 24, 2020
Published in issue: October 13, 2020


  1. secondary active transporters
  2. proton transport
  3. protein electrostatics
  4. cation–proton antiporters
  5. molecular dynamics


We acknowledge financial support provided by the National Institutes of Health (R01GM118772).


This article is a PNAS Direct Submission.



Department of Pharmaceutical Sciences, University of Maryland School of Pharmacy, Baltimore, MD 21201;
College of Computer Engineering, Jimei University, Xiamen 361021, Fujian Province, China;
Department of Physics, Arizona State University, Tempe, AZ 85287
Department of Pharmaceutical Sciences, University of Maryland School of Pharmacy, Baltimore, MD 21201;


To whom correspondence may be addressed. Email: [email protected].
Author contributions: O.B. and J.S. designed research; J.A.H. performed research; J.A.H., Y.H., O.B., and J.S. analyzed data; and J.A.H., O.B., and J.S. wrote the paper.

Competing Interests

The authors declare no competing interest.

Metrics & Citations


Note: The article usage is presented with a three- to four-day delay and will update daily once available. Due to ths delay, usage data will not appear immediately following publication. Citation information is sourced from Crossref Cited-by service.

Citation statements



If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download.

Cited by


    View Options

    View options

    PDF format

    Download this article as a PDF file


    Get Access

    Login options

    Check if you have access through your login credentials or your institution to get full access on this article.

    Personal login Institutional Login

    Recommend to a librarian

    Recommend PNAS to a Librarian

    Purchase options

    Purchase this article to get full access to it.

    Single Article Purchase

    Alternative proton-binding site and long-distance coupling in Escherichia coli sodium–proton antiporter NhaA
    Proceedings of the National Academy of Sciences
    • Vol. 117
    • No. 41
    • pp. 25183-25947







    Share article link

    Share on social media