Article Information

• PDF (1147 kb)
• Full Text with Large Figures
• Supporting Material
• Cited by in Scopus (11)

PubMed

• Articles by Huaiyu Yang
• Articles by Hualiang Jiang

Related Articles

• On the Equivalence Point for Ammonium (De)protonation during...

  On the Equivalence Point for Ammonium (De)protonation during Its Transport through the AmtB Channel Biophysical Journal, Volume 92, Issue 12, 15 June 2007, Pages L103-L105 David L. Bostick and Charles L. Brooks Abstract Structural characterization of the bacterial channel, AmtB, provides a glimpse of how members of its family might control the protonated state of permeant ammonium to allow for its selective passage across the membrane. In a recent study, we employed a combination of simulation techniques that suggested ammonium is deprotonated and reprotonated near dehydrative phenylalanine landmarks (F107 and F31, respectively) during its passage from the periplasm to the cytoplasm. At these landmarks, ammonium is forced to maintain a critical number (∼3) of hydrogen bonds, suggesting that the channel controls ammonium (de)protonation by controlling its coordination/hydration. In the work presented here, a free energy-based analysis of ammonium hydration in dilute aqueous solution indicates, explicitly, that at biological pH, the transition from ammonium () to ammonia (NH) occurs when these species are constrained to donate three hydrogen bonds or less. This result demonstrates the viability of the proposal that AmtB indirectly controls ammonium (de)protonation by directly controlling its hydration. Abstract | Full Text | PDF (217 kb)

• Ammonium Recruitment and Ammonia Transport by E. coli Ammoni...

  Ammonium Recruitment and Ammonia Transport by E. coli Ammonia Channel AmtB Biophysical Journal, Volume 91, Issue 12, 15 December 2006, Pages 4401-4412 Thomas P. Nygaard, Carme Rovira, Günther H. Peters and Morten Ø. Jensen Abstract To investigate substrate recruitment and transport across the Ammonia transporter B (AmtB) protein, we performed molecular dynamics simulations of the AmtB trimer. We have identified residues important in recruitment of ammonium and intraluminal binding sites selective of ammonium, which provide a means of cation selectivity. Our results indicate that A162 guides translocation of an extraluminal ammonium into the pore lumen. We propose a mechanism for transporting the intraluminally recruited proton back to periplasm. Our mechanism conforms to net transport of ammonia and can explain why ammonia conduction is lost upon mutation of the conserved residue D160. We unify previous suggestions of D160 having either a structural or an ammonium binding function. Finally, our simulations show that the channel lumen is hydrated from the cytoplasmic side via the formation of single file water, while the F107/F215 stack at the inner-most part of the periplasmic vestibule constitutes a hydrophobic filter preventing AmtB from conducting water. Abstract | Full Text | PDF (1357 kb)

• A Stable Water Chain in the Hydrophobic Pore of the AmtB Amm...

  A Stable Water Chain in the Hydrophobic Pore of the AmtB Ammonium Transporter Biophysical Journal, Volume 92, Issue 9, 1 May 2007, Pages L82-L84 Guillaume Lamoureux, Michael L. Klein and Simon Bernèche Abstract The accessibility of water molecules to the pore of the AmtB ammonium transporter is studied using molecular dynamics simulations. Free energy calculations show that the so-called hydrophobic pore can stabilize a chain of water molecules in a well of a few kcal/mol, using a favorable electrostatic binding pocket as an anchoring point. Moreover, the structure of the water chain matches precisely the electronic density maxima observed in x-ray diffraction experiments. This result questions the general assumption that the AmtB pore only contains ammonia (NH) molecules diffusing in a single file fashion. The probable presence of water molecules in the pore would influence the relative stability of NH and , and thus calls for a reassessment of the overall permeation mechanism in ammonium transporters. Abstract | Full Text | PDF (246 kb)

• …more

Copyright © 2007 The Biophysical Society. All rights reserved.
Biophysical Journal, Volume 92, Issue 3, 877-885, 1 February 2007

doi:10.1529/biophysj.106.090191

Channels, Receptors, and Electrical Signaling

Detailed Mechanism for AmtB Conducting NH₄⁺/NH₃: Molecular Dynamics Simulations

Huaiyu Yang, Yechun Xu, Weiliang Zhu^,  , Kaixian Chen and Hualiang Jiang^,  

Drug Discovery and Design Center, State Key Laboratory of Drug Research, Shanghai Institute of Materia Medica, Shanghai Institutes for Biological Science and Graduate School, Chinese Academy of Sciences, Shanghai 201203, China

Address reprint requests to Weiliang Zhu or Hualiang Jiang, Drug Discovery and Design Center, Shanghai Institute of Materia Medica, 555 Zuchongzhi Road, Shanghai 201203, China. Tel.: 86-21-5080-5020; Fax: 86-21-5080-7088.

Abstract

The mechanism by which the ammonium transporter, AmtB, conducts into the cytoplasm was investigated using conventional molecular dynamics (MD) simulations. These simulations revealed that the neutral molecule, NH₃, passes automatically through the channel upon its arrival at the Am2 site and that the function of the channel as a one-way valve for passage of NH₃ is not determined by the cytoplasmic exit gate but, rather, by the periplasmic entrance gate of the channel. The NH₃, produced by deprotonation of at the entrance gate, is spontaneously conveyed to the central region of the channel via a hydrogenbond network comprising His-168, His-318, Tyr-32, and the NH₃ molecule. Finally, the NH₃ molecule exits the channel through the exit gate formed by Phe-31, Ile-266, Val-314, and His-318. In addition, Ser-263 is shown to play a critical role in the final stages, acting as a pivoting arm to shunt the NH₃ molecule from the cytoplasmic exit gate of the channel out into the cytoplasm. This finding is further supported by another simulation which shows that NH₃ fails to be translocated through the channel formed by the Ser-263–Ala mutation. Thus, this study casts new insights on the mechanism of AmtB-mediated passage of NH₃ across cellular membranes.

The transport of ammonium across cellular membranes is critical for the acquisition and metabolism of nitrogen in a diverse range of organisms from bacteria to man 1,2. The ammonia/ammonium transporter protein family includes the ammonium transporters (Amts) and methylamine permeases (MEPs), and these proteins are typically 400–500 amino acids in length 3,4,5.

Ammonia/ammonium conductance facilitated by the Amt/MEP family of proteins is concentration dependent. At high extracellular ammonium concentrations, AmtB transporter activity is inhibited through complex formation with the regulatory GlnK protein, a member of the P_II protein family 2,6,7. Under these conditions, passive membrane permeation of NH₃ may be sufficient to promote cell growth 2. However, at low ammonium concentrations, GlnK does not interact with AmtB, and the function of AmtB for ammonia/ammonium translocation is activated. At the outset of our studies, the widely accepted view was that the target molecule in ammonia/ammonium translocation was . However, other experiments suggested that NH₃ could be the entity that is translocated 8,9. A series of studies investigating substrate specificity has been performed 2,10 and discordant conclusions have been drawn. Some studies indicate that the Amt/MEP family of proteins mediate the uptake of the analog, (methyl)ammonium, into the cell depending on the membrane potential and pH values, prompting the notion that these proteins are involved in the movement of from the periplasm into the cytoplasm 2,10. In contrast, other studies show that AmtB conduct NH₃ bidirectionally 8,9. Cellular studies conducted by Soupene et al. reveal that AmtB can facilitate the passage of NH₃ molecules into the cytoplasm.

The crystal structures of Escherichia coli AmtB show that it is a channel that spans the cellular membrane 11 times (Figure 1A) 11,12. On its periplasmic side (referred to as the top hereafter, Figure 1A), the channel is sheltered by Phe-107 and Phe-215, whereas the cytoplasmic side (referred to as the bottom) of the channel is blocked by Phe-31. The presence of an electron density peak outside the entrance gate in the crystal structure indicates that this could be the binding site (Am1 site in Figure 1A) 11. Cation-π interactions with Phe-103, Phe-107, and Trp-148, as well as hydrogenbond interactions with Ser-219, were postulated to be molecular determinants of binding in AmtB 13. Furthermore, there are three weak peaks (Am2–Am4 sites in Figure 1A) inside the channel which are presumed to be three NH₃ molecules 11. On the other hand, two significant studies challenged the view that NH₃ is observed in the center of the channel 12,14. Two histidine residues, His-168 and His-318, located in the middle of the channel, are linked via a hydrogen bond formed in two possible ways: between deprotonated His-168N^δ and the H of His-318N^δH (Figure 1B) or between deprotonated His-318N^δ and the H of His-168N^δH (Figure 1C). Hence, there are two plausible states of the His-168-His-318 system: 1), hydrogen atoms are added to N^ɛ of His-168 and N^δ of His-318 (referred to as the first protonation state); or 2), hydrogen atoms are added to N^δ of His-168 and N^ɛ of His-318 (referred to as the second protonation state). Khademi et al. proposed that unprotonated His-168N^δ and His-318N^δH are held rigidly by a hydrogen bond, and the CO of Leu-269 forms another hydrogen bond with the H of His-318N^δH 11.

Display large version of this figure

Figure 1

(A) Structure of AmtB (1U7G.pdb). Residues 214–272 are not shown for clarity. Some important residues are shown in stick representation. or NH₃ are variably put at the four sites (Am1–Am4, red spheres) in simulations. Entrance region, hydrophobic region, and exit region are simply indexed by red, black, and brown circles, respectively. (B) The first state of the neutral His-168-His-318 system where the hydrogen atoms are added to N^ɛ of His-168 and N^δ of His-318. (C) The second state of the neutral His-168-His-318 system where the hydrogen atoms are added to N^δ of His-168 and N^ɛ of His-318.

Although mechanistic information on the channel is available, further insights into the mode of ammonia/ammonium conductance at the atomic level has yet to be elucidated, especially the dynamic process of translocation. In addition, the exact site where releases H⁺ remains unknown. Deprotonation may take place either outside the channel or inside the channel, probably at the Am1 or Am2 site. Furthermore, it is also difficult to clarify the effects of the two protonation states of His-168 and His-318 on ammonia/ammonium conductance using crystallographic and biochemical methods. In addition, two different conformational states of the cytoplasmic exit of the channel have been observed in two different crystal forms 12. It is compelling to ask, then, what is the exit gating mechanism of the channel responsible for transport of small molecules? Does the conformational structure of the channel change from one form to another during NH₃ translocation?

Molecular dynamics (MD) simulations have been used to study membrane permeation of small molecules, a process that resembles that of ammonia translocation across membranes. For example, Groot et al. studied water permeation through aquaporin 15. Here, we report new insights into the mechanistic mode of ammonia/ammonium conductance through AmtB derived from a series of long time MD simulations on the systems where complexes were embedded in a water-solvated dipalmitoyl phosphatidylcholine (DPPC) bilayer. For the first time, the simulations show that NH₃ molecules can pass through AmtB spontaneously in both states of His-168-His-318 protonation via an exit gate facilitated by Ser-263.

The coordinates of AmtB were obtained from the crystal structure determined at pH 6.5 14 (Protein Data Bank (PDB) entry 1U7G). Mutations (F68S, S126P, and K255L) were modified by using the molecular modeling software Sybyl 6.8 (Tripos, St. Louis, MO). The protein was fitted into the DPPC bilayer with 309 lipids (15,450 atoms) to generate a suitable membrane system (Fig. 2). Coordinates for crystal water molecules near the entrance gate were kept and hydrogen atoms were added using the software Sybyl 6.8. Then the protein/DPPC systems were solvated in a bath of 16,661 SPC water molecules 16. Water molecules in the region which are occupied by a β-octyl glucoside molecule in the crystal structure were deleted. Thereafter, NH₃ molecules were variably added to three sites (Am2, Am3, and Am4 in Figure 1A) as shown in Table 1. To perform the modeling under simulated physiological conditions, Na⁺/Cl⁻ ions were then added to neutralize the modeled system.

Display large version of this figure

Figure 2

Side view of the simulation system. The AmtB protein is shown as a green ribbon, phosphate atoms are drawn as cyan spheres, the other atoms of the lipid are represented as magenta lines, and water molecules are displayed as red and white sticks. The front half of the bilayer is not shown for the sake of clarity.

Table 1 Detailed information for MD simulations

	Trajectory	Am2	Am3	Am4	His-168	His-318	Time (ns)
	A1	NH3	NH3	NH3	NɛH	NδH	8
	A2	NH3	NH3	NH3	NδH	NɛH	10
	B*	NH3	NH3	NH3	NɛH	NδH	20
	C1		–	–	NδH	NɛH	2
	C2		–	–	NδH	NɛH	2

* Ser-263 was mutated to alanine.

Details of these MD simulations are listed in Table 1. In simulations A1, A2, and B, three NH₃ molecules were placed at the Am2–Am4 sites, respectively, whereas in simulations C1 and C2, only one was placed at the Am2 site. Simulations A1 and B adopted the first protonation state for His-168-His-318 involving addition of hydrogen atoms to the N^ɛ and N^δ atoms of His-168 and His-318, respectively (Figure 1B), whereas the other simulations used the second state of protonation for His-168-His-318 whereby the N^δ atom of His-168 and the N^ɛ atom of His-318 were protonated (Figure 1C). Simulation B on the mutant Ser-263-Ala was performed to investigate the role of Ser-263 in NH₃ translocation.

MD simulations were carried out with the GROMACS package 17,18, using NPT and periodic boundary conditions. The GROMOS87 force field 19 was applied to the protein, and the parameters for the lipid were those used in previous MD studies of lipid bilayers 20,21,22,23. The charges of NH₃ and were obtained by a restrained ESP-fit method using the ChelpG approach 24. The charge on the nitrogen of NH₃ is −1.077, and the charges on the hydrogen atoms are +0.359. In , the nitrogen charge is −0.824, and the hydrogen charges are +0.456. The linear constraint solver (LINCS) method 25 was used to constrain bond lengths, allowing an integration step of 2 fs. Electrostatic interactions were calculated with the particle mesh Ewald algorithm 26,27. A constant pressure of 1 bar was applied independently in X, Y, and Z directions of the whole system with a coupling constant of 1.0ps 28. The systems were subjected to energy minimizations to remove unfavorable contacts. Water, lipids, , NH₃, and protein were coupled separately to a temperature bath at 323K using a coupling time of 0.1ps 28.

To study NH₃ translocation under the first protonation state of His-168-His-318, simulation A1 was carried out. All the NH₃ molecules that were initially located at Am2–Am4 sites (referred to as nha1, nha2, and nha3 hereafter) moved to the cytoplasmic pore exit (bottom) region of the channel within the simulation time of 1310ps (Fig. 3). Four snapshots (Figure 3C) demonstrated clearly how the NH₃ molecules moved toward the cytoplasmic exit gate to the bottom of the channel. It was observed that hydrogen bonding between the NH₃ molecules and His-168 ceased after ∼1310ps (Figure 3A), with a concomitant increase in the number of hydrogen bonds formed between the NH₃ and Tyr-32 or His-318, especially after 1310ps (Figure 3B). The N^ɛ of His-318 and hydroxyl of Tyr-32 could form hydrogen bonds with the NH₃ molecules, pulling the molecules to the bottom area of the channel. Thus, hydrogen bonding between N^ɛ of His-318 and hydroxyl of Tyr-32 provides the driving force for transport of NH₃ molecules through the hydrophobic central core to the bottom region of the channel.

Display large version of this figure

Figure 3

NH₃ molecules move to the bottom of the channel in trajectory A1. Numbers of the hydrogen bond between NH₃ molecules and His-168 (A) and between NH₃ molecules and His-318, Tyr-32 (B) versus time. Curves are obtained by 10ps averaged. (C) Side views of the starting structure and three snapshot structures (1000, 1300, and 1310ps). In side views, NH₃ molecules are drawn as white and blue sticks and water molecules are displayed as red and white sticks. F31, F107, H168, F215, S263, and H318 are displayed as colored sticks. The position of Y32 is indicated by an arrow.

Fig. 4 shows nha2 going into aqueous solution at ∼1550ps and is clearly the first NH₃ molecule leaving the channel. After passage of the NH₃ molecules to the bottom region of the channel at ∼1300ps, nha2 formed a hydrogen bond with the hydroxyl group of Ser-263 (Fig. 4). This hydrogen bond drew the ammonia close to the exit gate, thus allowing the formation of new hydrogen bonds between the NH₃ and water molecules in aqueous solution. The time-dependent hydrogen bonds formed between nha2 and the atoms inside the channel (Figure 4A) and those between nha2 and the atoms outside the channel (Figure 4B) show that the number of internal hydrogen bonds decreased from ∼1545ps onward, whereas the external hydrogen bonds increased from ∼1542ps onward. Therefore, Ser-263 is deemed to be the key residue in NH₃ translocation through the channel. This residue acts as a pivoting arm, drawing the NH₃ molecule out of the channel. This result is consistent with a biological role for the conserved Ser-263 residue present in the transporter protein family. We postulated that Ser-263 plays an essential role in proteins belonging to the ammonium transporter family.

Display large version of this figure

Figure 4

Process of the first NH₃ (nha2) leaving the channel in trajectory A1. Time-dependent hydrogen bonds between nha2 and the atoms inside the channel (including other NH₃ molecules, His-318 and Tyr-32) (A) and between nha2 and the atoms outside the channel (Ser-263 and water molecules) (B). Curves are obtained by 5ps averaged. (C) Side views of six snapshot structures (1321, 1336, 1549, 1550, 1553, and 1554ps, respectively). Hydroxyl of Ser-263 interacting with NH₃ through direct hydrogen bonding (1321, 1336, and 1549ps) or water bridge (1550ps).

The exit processes of the other two NH₃ molecules in the channel are very similar to that of ammonia nha2 (Fig. S1, Supplementary Material ). The hydroxyl of Ser-263 interacts with nha3 after the departure of nha2, pulling the NH₃ molecule out of the cytoplasmic exit pore and initiating the departure of the molecule nha3 from the channel at 1845ps (Fig. S1A). The last NH₃ molecule (nha1) remained at the bottom of the channel by interactions primarily with Tyr-32 and His-318. In slight contrast with nha2 and nha3, nha1 interacts with Ser-263 via a water bridge (Fig. S1B). At ∼5000ps, nha1 exits the channel into the cytoplasm along with the water molecule.

Molecules nha1, nha2, and nha3 are initially located at Am2–Am4 sites, but they do not sequentially leave the channel in numerical order. Rather, nha2 is the first molecule that leaves the channel, followed by nha3, and finally by nha1. This is attributed to the random movement of these molecules while in the chamber, shifting forward and backward to varying degrees. They frequently change the relative order of their positions in the chamber of the channel, especially from 1000ps onward.

In contrast to simulation A1, hydrogen atoms were added to the system His-168-His-318 in simulation A2 in accordance with the second protonation state explained earlier (Table 1 and Figure 1C). It took ∼8ns for the three NH₃ molecules to exit the channel. The NH₃ molecule nha2 exited the channel at ∼500ps (Fig. 5), followed by nha1 at ∼850ps (Fig. S2, Supplementary Material ). Similar to trajectory A1, the hydroxyl groups of Ser-263 and water molecules were also found to play important roles in the translocation of the ammonia molecules through the channel in trajectory A2. But, in comparison with trajectory A1, the two ammonia molecules passed the channel much faster in trajectory A2. This could be attributed to the strong hydrogen-bond network formed among the three NH₃ molecules and the N^ɛH of His-318 (Fig. 5). In other words, His-318 functions as a ‘springboard’ to facilitate the passage of the three ammonia molecules through the central core to the bottom area of the channel, thus allowing the ammonia molecules to interact with Ser-263. Hence, the consequent location of the NH₃ molecule near the cytoplasmic exit gate favors its departure from the channel into aqueous solution.

Display large version of this figure

Figure 5

Process of the ammonia nha2 leaving the channel in trajectory A2. The starting structure and five snapshots (463, 464, 466, 471, and 477ps) are shown.

Interestingly, recurrence of the translocation process was also discovered during the simulation. For example, the molecule nha1, which had exited the channel at ∼850ps, was found in the channel again at ∼1000ps (Fig. S2). In addition, a water molecule was also found to have completely entered the channel at ∼1310ps. Thus, the simulation indicated that both water and NH₃ molecules could return to the bottom area of the channel, implying that the characteristic of the channel as a one-way valve to conduct NH₃ is not determined by the bottom exit gate but rather by the top entrance gate of the channel. Indeed, upward diffusion of NH₃ from the bottom region of the channel to the area above the His-168 residue was not observed during the whole simulation.

The ammonia molecule, nha3, also formed hydrogen bonds with Ser-263 and finally exited the channel at ∼8100ps (Fig. S3, Supplementary Material ). Thereafter, the water molecule was found to leave the channel into the cytoplasm. The process by which the water molecule exited the hydrophobic channel was similar to the translocation process of NH₃. The conserved Ser-263 residue plays critical roles in these processes via hydrogen bonding with ammonia or water molecules.

At the cytoplasmic exit of the channel, two different conformational states, P63 (PDB: 1XQF) and R3 (PDB: 1XQE), have been observed in two different crystal forms 12. The distance between Phe-31 and Val-314 is ∼8.0Å in the P63 form and 10.5Å in the R3 form. On the basis of these experimental results, it was postulated that the exit gating mechanism requires remarkable structural changes to the channel during NH₃ translocation 12,14 with a fluctuation distance between Phe-31 and Val-314 from 8.0Å to 10.5Å. To determine if this proposed exit gating mechanism is reasonable, the distance of Phe-31 to Val-314 was monitored along the simulation time in trajectory A1 (Fig. 6). We did not observe the expected large fluctuation in distance during the translocation of NH₃ through the channel. In fact, limited distance fluctuations (from 8.0 to 8.8Å) were observed throughout the three periods coinciding with the exit of the three NH₃ molecules from the channel (Fig. 6). Therefore, the proposed exit gating mechanism requiring significant structural contortion of the channel might not exist. The same observations are also made in the simulation for trajectory 2. It is reasonable to conclude, therefore, that marked conformational change is not essential for the passage of NH₃ through the channel.

Display large version of this figure

Figure 6

Distance between Phe-31 and Val-314 versus time in the processes of NH₃ molecules leaving in trajectory A1. Curve is obtained by 10ps averaged.

Based on the above analyses of simulations A1 and A2, a new exit gating mechanism is proposed: NH₃ can move through the channel toward the cytoplasmic pore region and exit the channel into the cytoplasm via the exit gate formed by Phe-31, Ile-266, Val-314, and His-318. Hydrogen bonding is the most important driving force for the translocation of NH₃ molecules through the channel. His-318 serves as the springboard to facilitate the passage of the ammonia molecule toward the cytoplasmic end of the channel. In this process, significant conformational change of the channel is not essential. Instead, Ser-263 plays an essential role, acting as a pivoting arm to finally mediate the departure of NH₃ from the channel.

To illustrate the importance of Ser-263 in NH₃ translocation, simulation B was performed on a mutant of AmtB, in which Ser-263 was mutated to an alanine residue. Hydrogen atoms were added to His-168-His-318 in the first protonation state described earlier (Figure 1B and Table 1). In contrast with trajectory A1 whereby the three NH₃ molecules are shown to exit the channel in 5ns, none of the NH₃ molecules was observed to leave the hydrophobic channel in the duration of the 20-ns simulation time in trajectory B. Fig. 7 depicts the snapshot at 20ns from trajectory B. It is clear from the persistent presence of all three NH₃ molecules within the channel that the S263A mutation abrogated the function of Ser-263 in mediating the final departure of NH₃ into the cytoplasm. Taken together, our data confirm that Ser-263 plays a key role in the translocation of NH₃ molecules.

Display large version of this figure

Figure 7

Snapshot at 20ns from trajectory B. Ammonia molecules are shown as white and blue sticks.

Further analysis revealed that there are fewer water molecules around the exit gate in trajectory B than in trajectory A1. Because NH₃ is hydrophilic by nature, the predominantly more hydrophilic feature of trajectory A1 compared with trajectory B should provide a more conducive environment for the passage of NH₃ through the channel to the deep cytoplasmic bottom region and finally its departure through the exit pore into the essentially hydrophilic cytoplasm. On the other hand, the Ser-263–Ala mutation created a hydrophobic block by virtue of the aliphatic nature of alanine thereby effectively hindering the passage of hydrophilic NH₃ through the channel.

It has been suggested that the perisplasmic entrance gate is formed by Phe-107 and Phe-215 11,30. Analysis of the motions of these two residues in MD simulation, especially in conventional MD simulation, could provide useful information about the entry mechanism. Therefore, the distance between Phe-107 and Phe-215 against time in trajectory A1 was monitored (Figure 8A), and large fluctuations were observed ∼2000–4000ps. Upon further analysis, it was found that Phe-107 fluctuated on a large scale in the simulation (Figure 8B). Although Phe-215 does not show the large fluctuations seen in Phe-107, the aromatic ring of Phe-215 was observed to rotate ∼360° during the simulation from 3100 to 3800ps. Thus, the entrance gate could open spontaneously, and this mechanistic model is in contrast to the suggestion by Mo's group that was responsible for opening the entrance gate via interaction with Phe-107 30.

Display large version of this figure

Figure 8

(A) Distance between two aromatic rings of Phe-107 and Phe-215 versus time. (B) The superposition of the starting structure (black) and the snapshot structure of 2500ps (cyan) for Phe-107 and Phe-215.

It is noteworthy that although the simulations described so far have successfully elucidated the dynamic mechanism by which NH₃ molecules move through the hydrophobic channel into the aqueous cytoplasmic environment, the exact juncture where releases its proton to form NH₃ remains a mystery. Does the release its proton inside the channel? Our previous study using quantum chemistry methods reveals that proton transfer may take place between the imidazole ring of His-168 and 29. To explore whether the could enter freely into the channel to reach the His-168 residue, simulations were carried out to investigate the motions of the molecule in the channel with the hydrogen atom added in the second protonation state described earlier. It was found that the molecule did not enter the channel but exited the channel into the periplasm very rapidly (trajectory C1, Figure 9A), reaching the external Am1 site at ∼1100ps. Figure 9A shows that the distance of from the starting position is time dependent, indicating that resides over a longer period at the 3.3Å and 7Å positions compared with other positions before its final escape from the channel. To validate this observation, one more simulation (trajectory C2 in Table 1) was performed on the same system but with different initial velocities assigned to each atom in the simulation (Figure 9B). Remarkably, trajectory C2 is very similar to trajectory C1, indicating that rapid entry of into the periplasm through the entrance pore was not an incidental event. Therefore, deprotonation of the at the periplasm is a critical prerequisite for its entry into the channel to reach the Am2 site. Recent work that supports this viewpoint was reported by Mo's group using steered molecular dynamics (SMD) simulations 30. They speculated that the deprotonation process occurs at the Am1 site with Asp-160 as the proton acceptor.

Display large version of this figure

Figure 9

Time dependence of distance of from the starting position in trajectory C1 (A) and C2 (B).

In simulation C2, moves outward into the periplasm from the ∼450-ps time point onward. Once the reaches the external region, the molecule circulates around the Am1 site because of its interactions with Phe-107 and Trp-148 and Ser-219, suggesting that cation-π interaction is important for the process (Fig. S4, Supplementary Material ). This postulation is in agreement with Liu's observations 13.

Based on the results of our simulations, it is postulated that the first protonation state of the AmtB His-168-His-318 system (Figure 1B) is representative of the physiological status of this transporter protein, and this view is consistent with the observations of Khademi et al. First, our data on the fluctuation patterns of His-168-His-318 in trajectory A1 and A2 support this view. No significant structural changes to His-168 and His-318 in trajectory A1 (Fig. S5, Supplementary Material ) were observed, and this is consistent with the known stability of His-168-His-318 in crystal structures. On the other hand, the His-168-His-318 in trajectory A2 showed very large-scale fluctuations that were incompatible with the B-factor in crystal structures (Fig. S5). Hence, the implication is that the structure of His-168-His-318 in the crystal form was abolished in trajectory A2. Second, the phenomena of water entering the pocket in trajectory A1 again suggest that the protein adopts the first protonation state of the His-168-His-318 system. Unlike the previously reported nonadiabatic molecular dynamics simulation 11, one water molecule was found to enter and remain inside the deep pocket of the channel via the exit pore after the departure of three ammonia molecules from the channel (Figure 1A, and Fig. S6 of Supplementary Material ). Another water molecule dynamically entered and exited this pocket region via the exit pore. The water molecule(s) interacted with the N^ɛ of His-318 and hydroxyl of Tyr-32 to form hydrogen bonds. Indeed, a crystal structure of AmtB (PDB: 1XQF) shows that there is a water molecule in the pocket, whereas another crystal (PDB: 1XQE) demonstrates that there are two water molecules in the pocket, suggesting that our simulation results are in good agreement with the experimental results. However, no similar phenomenon was observed through trajectory 2, again implying that the physiologically relevant protonation state of His-168-His-318 adopted by the protein involves the addition of hydrogen atoms to N^ɛ of His-168 and N^δ of His-318, and this view is well supported by the work of Khademi et al.

Simulation results indicate that NH₃ molecules leave the channel with the aid of Ser-263 through hydrogen-bond formation that drives the movement of NH₃ downward to the exit region. Subsequently, water molecules in the vicinity of the exit region form further hydrogen bonds to draw the NH₃ completely out of the channel. Further compelling evidence for this mechanism is that the substitution of Ser-263 by alanine abolishes the departure of NH₃ from the channel in trajectory B. The implication is that the residue Ser-263 functions like a ‘pivoting arm’ to shunt NH₃ into the cytoplasm.

On the basis of SMD 30 and MD simulations, it was deduced that releases a proton in the periplasm and reaches the Am2 site as NH₃. Mo et al. surmised that the deprotonation process occurs at the Am1 site with Asp-160 as the proton acceptor 30. We showed strong evidence that there is no obligatory requirement for for the entrance gate to attain the open state. In addition, we found that the ammonia molecule is conveyed to the central region of the channel via a hydrogenbond network involving His-168, His-318, Tyr-31, and the NH₃ molecule itself. Finally, the NH₃ exits the channel mediated by Ser-263 and water molecules present in the vicinity. Fig. 10 depicts the proposed translocation pathway of NH₃ through AmtB. However, more quantum mechanics-based simulations are required to resolve at the atomic level the deprotonation process by which an molecule converts into NH₃ during translocation across the channel.

Display large version of this figure

Figure 10

Deduced mechanism of the conductance of NH₃. Ammonium releases its proton at Am1 site. The ammonia moves to the bottom region of the channel with the help of the interactions from Tyr-31 and His-318. Finally, the hydroxyl of Ser-263 forms a hydrogen bond with NH₃, to pull the NH₃ from the bottom region of the channel into the cytoplasm.