Modeling and Simulation of hGAT1: A Mechanistic Investigation of the GABA Transport Process

Human γ-Aminobutyric acid transporter 1 (hGAT1) is a Na+/Cl− dependent co-transporter that plays a key role in the inhibitory neurotransmission of GABA in the brain. Due to the lack of structural data, the exact co-transport mechanism of GABA reuptake by hGAT1 remains unclear. To examine the roles of the co-transport ions and the nature of their interactions with GABA, homology modeling and molecular dynamics simulations of the hGAT1 in the open-to-out conformation were carried out. Our study focused on the sequential preloading of Na+ and Cl− ions, followed by GABA binding. Our simulations showed pre-loading of ions maintains the transport ready state of hGAT1 in the open-to-out conformation essential for GABA binding. Of the four putative preloaded states, GABA binding to the fully loaded state is most favored. Binding of Na+ ion to the Na1 site helps to maintain the open-to-out conformation for GABA binding as compared to the Na2 site. GABA binding to the mono-sodium or the di-sodium loaded states leads to destabilization of Na+ ions within their binding sites. The two most prominent interactions required for GABA binding include interaction between carboxylate group of GABA with the bound Na+ ion in Na1 binding site and the hydroxyl group of Y140. Overall our results support the fully loaded state as the predominate state for GABA binding. Our study further illustrates that Na+ ion within the Na1 site is crucial for GABA recognition. Therefore, a revised mechanism is proposed for the initial step of hGAT1 translocation cycle.


Introduction
γ-Amino butyric acid (GABA) is a major endogenous inhibitory neurotransmitter in the central nervous system (CNS) [1]. Under normal physiological condition, GABA is released from vesicles into the synaptic cleft to restore the action potential of neurons. GABA reduces neuronal excitation by binding to the GABA receptors located on the surface of the post-synaptic neuron, resulting in a concentration gradient exchange of ions and hyperpolarization of the membrane's action potential [2]. After inhibitory neurotransmission, the extracellular concentration of GABA in the synaptic cleft is restored and maintained at low level by the feedback transport mechanism of the human GABA transporter subtype 1 (hGAT1) and GABA transporter subtype 3 (hGAT3) located on the presynaptic neuron surface and nearby astrocytes [3].
hGAT1 functions as the primary GABA transporter in the CNS requiring two sodium and one chloride co-transport ions to facilitate the GABA translocation process. At the basal state, hGAT1 interconverts between "open-to-out" and "open-to-in" (hGAT1 open-to-out and hGAT1 open-to-in ) conformations. During the reuptake process, commonly referred as the forward mode, GABA is loaded into the hGAT1 open-to-out conformation along with the sodium and chloride ions and is co-transported from the synaptic cleft into the cytoplasm of the presynaptic neuron. Buildup of intracellular concentration of GABA within the presynaptic neuron can also lead to GABA release. In this reverse mode GABA and co-transport ions bind to the hGAT1 open-to-in conformation and are co-transported into the synapse (Fig. 1) [4,5].
Abnormal neurotransmission that prevents the buildup of synaptic GABA concentration underlies the onset of various neurological disorders such as Alzheimer's disease [6], Parkinson's disease [7], schizophrenia [8,9] and, most notably, epilepsy [10]. Inhibitory neurotransmission can be restored by inhibiting the GABA reuptake process via hGAT1. This prolongs the availability of GABA for binding to the GABA receptors on the postsynaptic neurons [11]. To-date, the hGAT1 open-to-out  conformation involved in the GABA reuptake process has been established as a validated drug target [12] with Tiagabine as the only FDA approved drug for the treatment of epilepsy [13]. However, Tiagabine therapy has been associated with certain side effects including tremor, ataxia and sleep disorder, therefore, the quest for alternative hGAT1 inhibitors as potential antiepileptic agents has remained an active area of research for the past four decades [14][15][16][17]. Structural modeling of hGAT1 in the open-to-out conformation and its cotransport mechanism of GABA, facilitated by Na + and Cl − ions, should provide the atomistic insight essential for the design of novel GABA reuptake inhibitors.
The coupling stoichiometry of co-transport ions in the GABA translocation process remains unclear. Intense efforts have been made using the voltage clamp technique to shed light into the exact number of the sodium and chloride ions involved. Previous studies by Skovstrup et al., [18], Claxton et al., [19], and Singh et al., [20] have shown the fully loaded hGAT1 with two sodium and one chloride ions is required for the successful transport of GABA across the membrane (Fig. 2). However, Bicho et al., have suggested an alternative mechanism in which the Cl − ion may not be required [10]. Most recently, Willford et al., have provided new evidence that the transport of GABA may involve an additional cation, most likely a sodium ion, which would result in a three Na + and one Cl − ions coupling stoichiometry for the GABA transport in both GAT1 and GAT3 transporters [21].
The exact mechanism of how co-transport ions facilitate the GABA transport process is not well understood. Various experimental efforts have been made to ratify the exact sequential order of ion binding required for the Na + /Cl − -dependent GABA reuptake transport process [10,22,23]. Mager et al., have hypothesized that pre-loading of Na + and Cl − ions is required for GABA binding [22]. Bicho et al., have reported that in the forward mode of the transport process, the binding of a single sodium ion, followed by the second sodium ion and GABA, facilitates the initialization of the GAT1 mediated GABA reuptake process [10]. Most recently, Rosenberg et al., have proposed a two steps mechanism in which the binding of two Na + ions occurred, followed by the simultaneous binding of GABA and Cl − ion in the subsequent step [24].
To better understand the exact mechanism of the co-transport process for GABA, homology modeling and molecular dynamics (MD) simulation of hGAT1 open-to-out conformation was carried out using the Drosophila dopamine transporter (dDAT) as the X-ray crystallographic structural template [25]. We hypothesized the maintenance of the pre-loaded state of hGAT1 in the open-to-out conformation is essential for GABA binding prior to GABA translocation. To examine the role of sodium and chloride ion binding during the forward mode of the GABA reuptake process, we investigated each preloaded state of the hGAT1. This includes the apo, mono-sodium, di-sodium and the fully loaded (two Na + and one Cl − ) states of hGAT1 in the presence and absence of GABA. In the absence of a high resolution X-ray crystallographic structure of a homologous transporter consisting of a third sodium ion binding site, the proposed mechanism by Willford et al., [21] was not examined in the present study. While other studies have hypothesized that Cl − ion may not function as a co-transport ion [10], and is needed only to maintain the concentration gradient across the membrane [26], our study supports the widely accepted view of the fully loaded state of hGAT1 as the primary transport state for GABA [27].

Homology Modeling
Structural modeling was carried out using the Schrodinger modeling package [12]. Homology modeling of hGAT1 (UniProt: P30531) was based on the X-ray crystallographic structure of the open-to-out conformation of the dDAT in complex with cocaine and 2 Na + /1 Cl − co-transport ions (PDB ID: 4XP4). The sequence alignment between hGAT1 and dDAT was based on earlier studies by Yamashita et al., [28] and Beuming et al., [29] to identify the structurally conserved regions of hGAT1. The final model of hGAT1 consists of the crystallographic bound co-transport ions taken from dDAT as their binding sites are highly conserved between the two homologous transporters [25]. Using the standard protein preparation protocols, the ionization states of final hGAT1 model was prepared at physiological pH, followed by energy minimization using OPLS 2005 force field [30] with the implicit solvent generalized Born model [31] to optimize all hydrogen-bonding networks. The final model was evaluated by PROCHECK [32] (Supplementary Fig. S1) and ERRAT [33].

Modeling hGAT1 Complexes
The fully loaded homology model of hGAT1 consists of GABA along with its two sodium and one chloride ions. In order to explore the exact mechanism of the GABA transport process, each of the nine putative ion-preloaded states of the hGAT1 in the open-to-out conformation were examined. These include the (i) hGAT1 out , and (ix) hGAT1 out G, N1+2, C states where G, N and C represent the bound GABA, Na + and Cl − ions, respectively and the N 1 , N 2 and N 1+2 denote the mono or the di-sodium ions bound at the Na1 and Na2 sites. The coordinates of the bound ions were taken from the structural template of dDAT. GABA was subsequently modeled, energy minimized in implicit solvent, and docked into the putative hGAT1 binding site defined by the bound cocaine substrate within the dDAT template using Glide [34][35][36].

Molecular Dynamics Simulation
Each of the hGAT1 complexes was embedded within 15 Å buffer region inside a box of POPC lipid membrane bilayer and explicit TIP3P water using Desmond [37]. The system was electroneutralized by 0.15 M NaCl salt concentration as counter ions. Each of the complexes was initiated by Desmond's default initialization protocol, followed by 50 ns production simulation using the OPLS 2005 force field under constant area isothermal isobaric (NPAT) conditions at 300 K and 1 atm.
The root mean square deviation of the C α atoms (C α RMSD) from its initial coordinates and the root mean square fluctuation of the C α atoms from its mean position (C α RMSF) serve as the reference measure for the overall stability and conformational movements of the hGAT1 upon co-transport ions and GABA binding. Since all the nine modeled preloaded states were based on the hGAT1 homology model, the evaluation of C α RMSD for all simulations utilized the same reference structure.

Homology Modeling
Numerous factors such as sequence identity, functional similarity, structural resolution, and the sequence alignment are essential in determining the structural template used in homology modeling [38,39]. The latter is particularly critical for identifying structurally conserved regions, binding sites, and transmembrane regions. Here the homology modeling of hGAT1 (UniProt ID: P30531) in the open-to-out conformation was carried out based on the X-ray crystallographic structure of dDAT (PDB ID: 4XP4) in complex with its cocaine substrate and cotransport di-sodium and chloride ions. The X-ray crystallographic structural resolution between Aquifex aeolicus leucine transporter (A a LeuT) and dDAT was 2.0 Å and 2.8 Å, respectively. The overall sequence homology between hGAT1 and dDAT was 66% (Fig. 3), significantly higher than the 36% sequence homology of the A a LeuT used previously [18,28,29]. All three homologs (hGAT1, dDAT and A a LeuT) possess functional similarity as small molecules co-transporters, with nearly~75% shared sequence homology defining their structurally conserved ion and substrate binding sites. This includes the highly conserved Y139, Y140, I143, W146, F294, N327 and S396 binding site residues of hGAT1, whose importance has been well articulated through previous mutagenesis studies [18,24,27,40]. Finally, A a LeuT is a Cl − independent co-transporter which lacks the structurally conserved chloride ion binding site likely due to the S331E mutation [41][42][43]. As such, dDAT was selected as the preferred structural surrogate for the development of the revised hGAT1 homology model [18,28,29]. Subsequent interconversion to the hGAT1 in G, N1+2, C conformation enables the release of the GABA and co-transport ions into the intracellular membrane as proposed by Skovstrup et al., [18], Claxton et al., [19], and Singh et al., [20].
The homology model of hGAT1 showed the characteristic topology of 12 transmembrane (TM) helices along with the intracellular (IL) and extracellular loops (EL) (Supplementary Fig. S1A). The quality of the hGAT1 model was evaluated using Ramachandran plot [32] and ERRAT [33]. The Ramachandran plot displayed 93.5% of the residues of hGAT1 in most favored regions, 5.2% residues in additionally allowed regions, 0.4% residues in generously allowed regions, and 0.9% residues in disallowed regions (Supplementary Fig. S1B). The residues in the disallowed region include N184, M200, T201, and D202 located within the EL2. This was not surprising as the EL2 in the homologs possess less number of amino acid residues and hence was not modeled properly in hGAT1 [44]. Moreover, ERRAT [33] scored 82.7% hereby supporting the overall reliability of the hGAT1 model.

Molecular Dynamics Simulations
The present study explores the co-transport mechanism of hGAT1 by examining each of the four putative preloaded states of hGAT1 and its interaction with GABA. The preloaded states include the mono-sodium ion loaded state in the Na1 or Na2 site (hGAT1 out N1 and hGAT1 out N2 ), the di-sodium loaded state in both the Na1 and Na2 sites (hGAT1 out N1+2 ), and the fully loaded state with two sodium and one chloride ions in the Na1, Na2 and Cl sites (hGAT1 out N1+2, C ). The four putative transport ready states involve each of the explored preloaded states with bound GABA (hGAT1 out G, N1 , hGAT1 out G, N2 , hGAT1 out G, N1+2 and hGAT1 out G, N1+2, C ).
The apo state (hGAT1 out free ) was included as the reference state for the study.
To explore the overall stability and dynamic behavior of hGAT1, molecular dynamics (MD) simulations were carried out for each of its putative state within a POPC phospholipid membrane bilayer with TIP3P explicit water model at 0.15 ionic salt concentration. The C α RMSD (A-C) and the C α RMSF (D-F) for each of the MD simulations are shown in Fig. 4. The reference simulation of the apo state (hGAT1 out free ) showed convergence after the first 10 ns of the simulation with an average C α RMSD over 2.9 Å, indicative of a stable hGAT1 homology model undergoing conformational optimization within the phospholipid bilayer membrane-solvent environment. Both the mono-sodium preloaded states, hGAT1 out N1 and hGAT1 out N2 with sodium ion bound to either the N1 or N2 site in hGAT1 exhibited an average C α RMSD of 2.4 Å and 2.9 Å, respectively (Fig. 4A). Subsequent binding of GABA to these two states resulted in an average C α RMSD of 2.4 Å and 3.3 Å for hGAT1 out G, N1 and hGAT1 out G, N2 complexes, respectively. The marked increase in the average C α RMSD for hGAT1 out G, N2 over hGAT1 out G, N1 suggests preloading of sodium ion to the Na1 site is crucial, and may be preferred over the Na2 site, for stabilizing the hGAT1 in the open-to-out conformation for both GABA and co-transport ion binding prior to translocation. For the hGAT1 out N1+2 complex, binding of di-sodium ions resulted in an average C α RMSD of 2.9 Å similar to both the reference hGAT1 out free and the hGAT1 out N2 states. Subsequent addition of GABA to the di-sodium complex (hGAT1 out G, N1+2 ) increases the average C α RMSD to 3.5 Å (Fig. 4B), comparable to the observed changes for GABA binding to the hGAT1 out N2 state. The overall increase in the average C α RMSD for both the hGAT1 out G, N1+2 and the hGAT1 out G, N2 complexes, as compared to their unbound GABA states (hGAT1 out N1+2 and hGAT1 out N2 ), suggests neither preloaded states afford sufficient stabilization environment for GABA binding. For the hGAT1 out N1+2, C complex, addition of a single Cl − ion to the di-sodium complex resulted in an average C α RMSD value of 3.0 Å (Fig. 4C). In this preloaded state, binding of GABA (hGAT1 out G, N1+2, C ) also resulted in an average C α RMSD of 3.0 Å. In contrast to the hGAT1 out G, N1+2 and the hGAT1 out G, N2 states, the presence of the Cl − ion with the di-sodium ions helped to stabilize the overall hGAT1 structure both prior to and during GABA binding. Therefore, we propose the fully loaded state with the di-sodium and chloride ions as the preferred preloaded state for GABA binding required for the reuptake transport process. To validate the overall stability of the fully loaded complex (hGAT1 out N1+2, C ), its simulation was further extended to 100 ns with a sustained average C α RMSD of 3.0 Å (Supplementary Fig. S2).
To examine the effect of ion binding on conformational changes, we further evaluated the root mean square fluctuations of the backbone C α atoms (C α RMSF) for all nine hGAT1 complexes (Fig. 4 D-F). The highest C α RMSF values correspond to the residues located in the hydrophilic ELs, the ILs and terminal loops. The residues in the TM segments showed the lowest values of C α RMSF. The highest C α RMSF values observed corresponded to the EL2 and EL3 at 3.4 Å and 3.7 Å, respectively. EL2 located between the TM3 and TM4 underwent major conformational shift during translocation of GABA in forward mode. It was also observed that cytoplasmic region of the TM3 become more flexible due to the unwinding over the course of simulation period. This might be a result of poor solvent-lipid interactions as it was observed in all of the hGAT1 proposed model states. Overall, the C α RMSF have shown that the ligand-protein complex of fully loaded hGAT1 is reasonably stable in the lipid bilayer-aqueous environment.
Binding of the di-sodium ions is required in all the three homologs of A a LeuT, dDAT and hGAT1, but the exact order of the sodium ions preloading process remains unclear. Bicho et al., has proposed that a single sodium ion binds first into the hGAT1 binding pocket, followed by other ions and substrate [10]. In order to explore Bicho's ion transport mechanism, we examine the interatomic distances between each of the sodium ions and GABA with its nearest neighboring residues in hGAT1 (Table 1). For the hGAT1 out N1 and hGAT1 out N2 complexes with the mono-sodium ion located at either the Na1 and Na2 sites in the absence of GABA, the average distances for the A61, N66, S295 and N327 residues surrounding the sodium ion at the Na1 site were 3.2 Å, 2.5 Å, 2.5 Å and 2.5 Å, respectively during the 50 ns MD simulation. For the sodium ion located at the Na2 binding site, the average distances for G59, I62, L392 and D395 were 2.7 Å, 2.4 Å, 3.1 Å and 2.3 Å, respectively. Examining the binding of GABA to the mono-sodium complex (hGAT1 out G, N1 or hGAT1 out G, N2 ) helped to identify its effect on sodium ions binding. In the hGAT1 out G, N1 complex, binding of GABA alters the Na1 interactions with surrounding A61 and S295, resulting in an average distance of 4.4 Å and 4.9 Å while forming a direct charge-charge interaction between its negatively charged carboxylate group with the positively charged Na1 at 3.0 Å. For the hGAT1 out G, N2 complex the average distance between Na2 with residues G59 and I62 was dramatically increased to 3.9 Å and 6.7 Å respectively. Since the GABA binding site was far away from the Na2 site, the average distance between GABA's carboxylate group to the Na2 sodium ion was 8.9 Å. This is consistent with current view that the hGAT1 out G, N2 complex is considered as a "non-GABA reuptake/transport state" because of the absolute requirement for sodium ion interaction with GABA's carboxylate group [45].
In di-sodiumhGAT1 complex (hGAT1 out N1+2 ), both Na1 and Na2 ions were observed at a interatomic distance of 7.6 Å. Both sodium binding sites are separated and screened from one another by residues A61 and I62, each of which forms a direct interaction with one of the sodium ions [46,47]. The addition of GABA to the di-sodium complex (hGAT1 out G,

N1+2
) resulted in an increase in the interatomic distances between Na1 with S295 to 5.6 Å and N327 to 5.1 Å while the interactions between Na1 and A61, N66, S295 were relatively preserved. The average interatomic distance between GABA's carboxylate group to Na1 in the di-sodium complex was slight lower at 2.5 Å as compared to the 3.0 Å in the hGAT1 out G, N1 complex.
Addition of Cl − ion to the di-sodium complex in the absence of GABA (hGAT1 out N1+2, C ) showed the lowest overall interatomic distances for Na1 and Na2 with the nearby residues all between the ranges of 2.5 to 3.0 Å. The interatomic distances between Cl − ion with Y86, S295, N327 and S331 were all between 3.3 and 4.1 Å. In the fully loaded hGAT1 out G, N1+2, C complex with bound GABA, the overall average interatomic distances for the ion binding site residues surrounding the Na2 and Cl − ions diminishes, indicative of enhance stability and binding affinity. The interatomic distances between the ions and their respective binding site residues for the fully loaded hGAT1 complex with and without GABA over the course of MD simulation are shown in Supplementary Fig. S3. For the fully loaded GABA bound complex, we were able to identify six stable interactions surrounding Na1 ion, as compared to the typical four found in the hGAT1 out G, N1 and hGAT1 out G, N1+2 states ( Table 1). The interatomic distance between GABA and Na1 was 2.3 Å, the shortest of all four simulated GABA bound complexes. Overall, our interatomic distance analysis showed similar interaction profile for mono-sodium and di-sodium binding to the Na1 and Na2 sites. The presence of a sodium ion in the Na1 site is crucial for GABA binding. The binding of the Cl − ion is essential for stabilizing the di-sodium ions both in the GABA unbound and bound complex. Its presence further improves the interatomic distance between GABA and the sodium ion in the Na1 site. Based on MD simulation and detailed analysis of the protein-ligand interactions, the fully loaded hGAT1 out G, N1+2, C state is the most promising of the four examined GABA bound hGAT1 complexes for GABA transport. In addition to its interaction to the Na1 ion, the GABA's carboxylate group also formed a hydrogen bonding interaction with the hydroxyl group of Y140 ( Fig. 5A and Fig. 6). The importance of Y140 in hGAT1 has been demonstrated experimentally as a known binding determinant of GABA [48,49]. Mutation of Y140 has been shown to abolish its transport ability of GABA in the forward mode [50]. Its importance has also been explored by Skovstrup et al., with constrained MD simulation [18]. Furthermore, as a zwitterion, GABA's positively charged ammonium group also forms hydrogen bond with the hydroxyl groups of  Y60 and S396. Interestingly, S396 has been shown to play a significant role in forming the lid type structure of EL4 that helps in the establishment of a hydrophobic cavity around hGAT1 binding site [51,52]. It has also been established that binding of the GABA requires a major conformational change in EL4 [53]. The interactions of sodium and chloride ions with the surrounding binding site residues in fully loaded hGAT1 were also examined. Na1 occupied its respective position in the binding pocket by interacting with the side chain atoms of N66, S295 and N327 and carbonyl oxygen of the A61 and S295 (Fig. 5B) along with electrostatic interaction with carboxylate group of GABA (Fig. 6). Yamashita et al., also observed direct interaction between the Na1 and leucine in A a LeuT previously [28] while it was water-mediated between the amino group of dopamine and Na1 in dDAT [25]. On the other hand, the surrounding residues of Na2 showed interactions with backbone atoms of G59, I62, L392 and side chain atoms of D395 (Fig. 5C). In addition, the side chain of G63 formed a bridge between Na1 and Na2 in fully loaded hGAT1 (Fig. 5C). The importance of G63 is well studied by Kanner in GABA permeation pathway through mutagenesis study [1]. It has been established that the mutation of G63 to either cysteine or serine abolishes the Na + and Cl − dependent GABA transport both in forward and backward mode [1].
The Cl − ion in fully loaded hGAT1 was found in near proximity of Na1. The residues surrounding the Cl − ion include Y86, S295, N327 and S331 as shown in Fig. 5D. Mari et al., [11], Skovstrup et al., [18] and Zomot et al., [26] demonstrated previously that the side chain of S295 forms a direct bridge between Na1 and Cl − ions, thereby keeping them intact in their respective positions in hGAT1 binding pocket. Although, the side chains of amino acid residues surrounding all of the three co-transport ions (2 Na + and 1 Cl − ) were flexible throughout unconstrained 50 ns MD simulation, all of the ions remained consistently bound from their respective first frame position to the last frame position, demonstrating the rigor of our homology model and the stability of the fully loaded hGAT1.
Based on the C α RMSD and inter-atomic distances between all the putative ions preloaded states of hGAT1, we proposed a more detailed GABA translocation cycle by hGAT1 (Fig. 7). According to our results, the addition of mono-sodium ion (hGAT1 out N1 or hGAT1 out N2 ) to the apo state (hGAT1 out free ) does not affect the overall stability of hGAT1. Formation of the di-sodium complex is also tolerable. Subsequent addition of GABA to both the mono-sodium and the di-sodium bound states resulted in the overall increase in the interatomic distances between the bound sodium ion and its neighboring residues, suggesting premature binding of GABA could lead to destabilization of the pre-loaded cotransport ions. The binding of Cl − ion to di-sodium complex showed marked decrease in the overall interatomic distances between the bound sodium ions with its neighboring residues. Its presence helped to stabilize the pre-bound sodium ions within their binding sites, making the hGAT1 out N1+2, C the most favored state of all four co-transport ions preloaded states. GABA binding in the fully loaded complex (hGAT1 out G, N1+2, C ) is the only GABA bound state with overall decrease in the interatomic distances for all pre-bound co-transport ions with their neighboring residues. It also exhibited the nearest distance between its carboxylate group with the Na1 ion relative to the GABA bound mono-sodium and di-sodium states. Simultaneous binding of GABA and Cl − to the disodium state is possible to form the fully loaded complex. Overall, our model is consistent with the previous findings that the binding of GABA to the sodium ion is crucial for its successful transport [47,52].

Conclusion
With the availability of a high resolution X-ray crystallographic structure of dopamine transporter, homology modeling and molecular dynamics simulation of hGAT1 were carried out to examine the exact mechanism of the GABA transport process. Previous homology models of hGAT1 were based on AaLeuT with only 36% sequence homology Fig. 6. Interatomic distances between GABA carboxylate oxygen atoms to the Na + ion in the Na1 site and the hydroxyl group of Y140. and lack a chloride ion binding site. In the current study, we have modeled and simulated the first full length homology model of hGAT1 based on highly homologous dDAT. Moreover, this is the first 50 ns molecular dynamics simulation of the full length hGAT1 to-date to explore multiple preloaded states (mono-sodium, di-sodium, and fully loaded) of hGAT1 involved in the translocation cycle of GABA. We hypothesized that the preloaded state of hGAT1 in open-to-out conformation is essential for GABA binding prior to the reuptake of GABA. Our results support the fully loaded hGAT1 as the most favored state for GABA translocation. It further establishes the importance of sodium ion binding within the Na1 binding site for GABA recognition. Although the flipping of hGAT1 from open-to-out to open-to-in conformation was not examined in the present study, our revised mechanistic model for the GABA translocation cycle may provide an improve framework for understanding the initial step in the GABA reuptake process.