Portal de Periódicos da CAPES

The allosteric inhibition of glycine transporter 2 by bioactive lipid analgesics is controlled by penetration into a deep lipid cavity

2021; Elsevier BV; Volume: 296; Linguagem: Inglês

10.1016/j.jbc.2021.100282

1083-351X

Katie A. Wilson, Shannon N. Mostyn, Zachary J. Frangos, Susan Shimmon, Tristan Rawling, Robert J. Vandenberg, Megan L. O’Mara,

Amino Acid Enzymes and Metabolism

The role of lipids in modulating membrane protein function is an emerging and rapidly growing area of research. The rational design of lipids that target membrane proteins for the treatment of pathological conditions is a novel extension in this field and provides a step forward in our understanding of membrane transporters. Bioactive lipids show considerable promise as analgesics for the treatment of chronic pain and bind to a high-affinity allosteric-binding site on the human glycine transporter 2 (GlyT2 or SLC6A5). Here, we use a combination of medicinal chemistry, electrophysiology, and computational modeling to develop a rational structure–activity relationship for lipid inhibitors and demonstrate the key role of the lipid tail interactions for GlyT2 inhibition. Specifically, we examine how lipid inhibitor head group stereochemistry, tail length, and double-bond position promote enhanced inhibition. Overall, the l-stereoisomer is generally a better inhibitor than the d-stereoisomer, longer tail length correlates with greater potency, and the position of the double bond influences the activity of the inhibitor. We propose that the binding of the lipid inhibitor deep into the allosteric-binding pocket is critical for inhibition. Furthermore, this provides insight into the mechanism of inhibition of GlyT2 and highlights how lipids can modulate the activity of membrane proteins by binding to cavities between helices. The principles identified in this work have broader implications for the development of a larger class of compounds that could target SLC6 transporters for disease treatment. The role of lipids in modulating membrane protein function is an emerging and rapidly growing area of research. The rational design of lipids that target membrane proteins for the treatment of pathological conditions is a novel extension in this field and provides a step forward in our understanding of membrane transporters. Bioactive lipids show considerable promise as analgesics for the treatment of chronic pain and bind to a high-affinity allosteric-binding site on the human glycine transporter 2 (GlyT2 or SLC6A5). Here, we use a combination of medicinal chemistry, electrophysiology, and computational modeling to develop a rational structure–activity relationship for lipid inhibitors and demonstrate the key role of the lipid tail interactions for GlyT2 inhibition. Specifically, we examine how lipid inhibitor head group stereochemistry, tail length, and double-bond position promote enhanced inhibition. Overall, the l-stereoisomer is generally a better inhibitor than the d-stereoisomer, longer tail length correlates with greater potency, and the position of the double bond influences the activity of the inhibitor. We propose that the binding of the lipid inhibitor deep into the allosteric-binding pocket is critical for inhibition. Furthermore, this provides insight into the mechanism of inhibition of GlyT2 and highlights how lipids can modulate the activity of membrane proteins by binding to cavities between helices. The principles identified in this work have broader implications for the development of a larger class of compounds that could target SLC6 transporters for disease treatment. One in 10 adults worldwide is diagnosed with chronic pain each year (1Goldberg D.S. McGee S.J. Pain as a global public health priority.BMC Public Health. 2011; 11: 770Crossref PubMed Scopus (562) Google Scholar). Despite the high rate of chronic pain, there is a lack of safe and effective treatment options, which in turn has significant social and economic consequences (2Volkow N. Benveniste H. McLellan A.T. Use and misuse of opioids in chronic pain.Annu. Rev. Med. 2018; 69: 451-465Crossref PubMed Scopus (108) Google Scholar). In the mammalian central nervous system, the neurotransmitter glycine inhibits the pain-signaling pathway (3Todd A.J. Neuronal circuitry for pain processing in the dorsal horn.Nat. Rev. Neurosci. 2010; 11: 823-836Crossref PubMed Scopus (776) Google Scholar). Synaptic concentrations of glycine are controlled by the two glycine transporters, GlyT1 and GlyT2, which are responsible for clearing glycine from synapses (4Eulenburg V. Armsen W. Betz H. Gomeza J. Glycine transporters: Essential regulators of neurotransmission.Trends Biochem. Sci. 2005; 30: 325-333Abstract Full Text Full Text PDF PubMed Scopus (259) Google Scholar). GlyT2 is expressed by presynaptic neurons and is also responsible for replenishing presynaptic glycine concentrations to maintain glycinergic neurotransmission. Inhibitors of GlyT2 slow the reuptake of glycine to prolong glycine neurotransmission (5Cioffi C.L. Modulation of glycine-mediated spinal neurotransmission for the treatment of chronic pain.J. Med. Chem. 2018; 61: 2652-2679Crossref PubMed Scopus (15) Google Scholar) and have been developed as potential therapeutics in the treatment of chronic pain (6Mostyn S.N. Wilson K.A. Schumann-Gillett A. Frangos Z.J. Shimmon S. Rawling T. Ryan R.M. O'Mara M.L. Vandenberg R.J. Identification of an allosteric binding site on the human glycine transporter, GlyT2, for bioactive lipid analgesics.Elife. 2019; 8e47150Crossref PubMed Scopus (16) Google Scholar, 7Caulfield W.L. Collie I.T. Dickins R.S. Epemolu O. McGuire R. Hill D.R. McVey G. Morphy J.R. Rankovic Z. Sundaram H. The first potent and selective inhibitors of the glycine transporter type 2.J. Med. Chem. 2001; 44: 2679-2682Crossref PubMed Scopus (76) Google Scholar, 8Takahashi E. Arai T. Akahira M. Nakajima M. Nishimura K. Omori Y. Kumagai H. Suzuki T. Hayashi R. The discovery of potent glycine transporter type-2 inhibitors: Design and synthesis of phenoxymethylbenzamide derivatives.Bioorg. Med. Chem. Lett. 2014; 24: 4603-4606Crossref PubMed Scopus (7) Google Scholar, 9Vandenberg R.J. Ryan R.M. Carland J.E. Imlach W.L. Christie M.J. Glycine transport inhibitors for the treatment of pain.Trends Pharmacol. Sci. 2014; 35: 423-430Abstract Full Text Full Text PDF PubMed Scopus (50) Google Scholar, 10Xu T.-X. Gong N. Xu T.-L. Inhibitors of GlyT1 and GlyT2 differentially modulate inhibitory transmission.Neuroreport. 2005; 16: 1227-1231Crossref PubMed Scopus (29) Google Scholar, 11Mostyn S.N. Sarker S. Muthuraman P. Raja A. Shimmon S. Rawling T. Cioffi C.L. Vandenberg R.J. Photoswitchable ORG25543 congener enables optical control of glycine transporter 2.ACS Chem. Neurosci. 2020; 11: 1250-1258Crossref PubMed Scopus (3) Google Scholar). N-acyl amino acids that comprise an amino acid head group conjugated via an amide bond to a monounsaturated lipid tail represent a novel class of GlyT2 inhibitors (6Mostyn S.N. Wilson K.A. Schumann-Gillett A. Frangos Z.J. Shimmon S. Rawling T. Ryan R.M. O'Mara M.L. Vandenberg R.J. Identification of an allosteric binding site on the human glycine transporter, GlyT2, for bioactive lipid analgesics.Elife. 2019; 8e47150Crossref PubMed Scopus (16) Google Scholar). Our previous studies have shown that despite the high level of sequence conservation between GlyT2 and GlyT1, these compounds do not inhibit GlyT1 (6Mostyn S.N. Wilson K.A. Schumann-Gillett A. Frangos Z.J. Shimmon S. Rawling T. Ryan R.M. O'Mara M.L. Vandenberg R.J. Identification of an allosteric binding site on the human glycine transporter, GlyT2, for bioactive lipid analgesics.Elife. 2019; 8e47150Crossref PubMed Scopus (16) Google Scholar, 12Mostyn S.N. Carland J.E. Shimmon S. Ryan R.M. Rawling T. Vandenberg R.J. Synthesis and characterization of novel acyl-glycine inhibitors of GlyT2.ACS Chem. Neurosci. 2017; 8: 1949-1959Crossref PubMed Scopus (20) Google Scholar, 13Mostyn S.N. Rawling T. Mohammadi S. Shimmon S. Frangos Z.J. Sarker S. Yousuf A. Vetter I. Ryan R.M. Christie M.J. Vandenberg R.J. Development of an N-acyl amino acid that selectively inhibits the glycine transporter 2 to produce analgesia in a rat model of chronic pain.J. Med. Chem. 2019; 62: 2466-2484Crossref PubMed Scopus (15) Google Scholar). One of the most promising bioactive lipids from this series, oleoyl d-lysine (C18ω9-d-Lys), is a selective and potent GlyT2 inhibitor that is metabolically stable, blood–brain barrier permeable, and produces analgesia in a rat model of neuropathic pain with minimal side effects (13Mostyn S.N. Rawling T. Mohammadi S. Shimmon S. Frangos Z.J. Sarker S. Yousuf A. Vetter I. Ryan R.M. Christie M.J. Vandenberg R.J. Development of an N-acyl amino acid that selectively inhibits the glycine transporter 2 to produce analgesia in a rat model of chronic pain.J. Med. Chem. 2019; 62: 2466-2484Crossref PubMed Scopus (15) Google Scholar). The reduced toxicity of these compounds is attributed to partial inhibition of GlyT2. They reduce but do not completely block glycine transport, which allows presynaptic glycine concentrations to be maintained for subsequent repackaging into synaptic vesicles to maintain glycinergic neurotransmission (14Vandenberg R.J. Mostyn S.N. Carland J.E. Ryan R.M. Glycine transporter2 inhibitors: Getting the balance right.Neurochem. Int. 2016; 98: 89-93Crossref PubMed Scopus (14) Google Scholar). The most potent bioactive lipids bear positively charged (Lys) or aromatic (Trp) amino acid head groups and inhibit GlyT2 with IC50 concentrations of less than 50 nM (13Mostyn S.N. Rawling T. Mohammadi S. Shimmon S. Frangos Z.J. Sarker S. Yousuf A. Vetter I. Ryan R.M. Christie M.J. Vandenberg R.J. Development of an N-acyl amino acid that selectively inhibits the glycine transporter 2 to produce analgesia in a rat model of chronic pain.J. Med. Chem. 2019; 62: 2466-2484Crossref PubMed Scopus (15) Google Scholar). Within the allosteric-binding site, oleoyl-l-lysine (C18ω9-l-Lys) orients tail down so that the tail intercalates between aliphatic-rich regions of TM5 and TM8 and the oleoyl double bond is in close proximity to TM5 (6Mostyn S.N. Wilson K.A. Schumann-Gillett A. Frangos Z.J. Shimmon S. Rawling T. Ryan R.M. O'Mara M.L. Vandenberg R.J. Identification of an allosteric binding site on the human glycine transporter, GlyT2, for bioactive lipid analgesics.Elife. 2019; 8e47150Crossref PubMed Scopus (16) Google Scholar). The amino acid head group is accessible to the extracellular solution and stabilized by aromatic residues in TM7, TM8, and EL4 (6Mostyn S.N. Wilson K.A. Schumann-Gillett A. Frangos Z.J. Shimmon S. Rawling T. Ryan R.M. O'Mara M.L. Vandenberg R.J. Identification of an allosteric binding site on the human glycine transporter, GlyT2, for bioactive lipid analgesics.Elife. 2019; 8e47150Crossref PubMed Scopus (16) Google Scholar). In the present work, we use a combination of medicinal chemistry, electrophysiology, and computational modeling to explore the structure–activity relationship for lipid inhibitors of GlyT2 to understand how head group stereochemistry and chemical features of the acyl tail affect inhibitor activity and interaction with GlyT2 at a molecular level. The effects of head group stereochemistry are investigated for the most potent previously identified lipid inhibitors, C18ω9-Lys and C18ω9-Trp. To further understand the effect of structure on inhibitor potency, a series of acyl-lysine analogs were synthesized with variations in the tail length (C18, C16, and C14) and position of the double bond within the lipid tail (ω9, ω7, ω5, and ω3; Fig. 1). Molecular dynamics (MD) simulations were used to provide a structural basis for the effect of chemical changes on lipid inhibitor properties. We demonstrate that the potency of the lipid inhibitors was greatly influenced by a combination of both the stereochemistry of the head group and the length and saturation of the lipid tail. Furthermore, inhibitor potency is shown to depend on deep penetration of the lipid tail into a stabilized location in the allosteric-binding site. This study develops a comprehensive structure–activity relationship for lipid inhibitors of GlyT2. This is critical for future rational design of more effective GlyT2 inhibitors for the treatment of chronic pain and may have broader implications for modulation of other SLC6 transporters. Bioactive lipids bearing Lys or Trp head groups in the l-configurations are among the most potent GlyT2 inhibitors in the series, inhibiting GlyT2 with IC50 concentrations of 25.5 and 54.6 nM, respectively (Table 1). Interestingly, when the head group is converted to the d-configuration, the Lys analog (C18ω9-d-Lys) retains potency, whereas the Trp analog (C18ω9-d-Trp) is inactive (Table 1) (6Mostyn S.N. Wilson K.A. Schumann-Gillett A. Frangos Z.J. Shimmon S. Rawling T. Ryan R.M. O'Mara M.L. Vandenberg R.J. Identification of an allosteric binding site on the human glycine transporter, GlyT2, for bioactive lipid analgesics.Elife. 2019; 8e47150Crossref PubMed Scopus (16) Google Scholar). To investigate the molecular basis of this activity, MD simulations of 500 ns were performed in triplicate, in which one molecule of either the l- or d-stereoisomers of C18ω9-Lys or C18ω9-Trp was docked into the extracellular allosteric-binding site of our GlyT2 homology model (15Subramanian N. Scopelitti A.J. Carland J.E. Ryan R.M. O'Mara M.L. Vandenberg R.J. Identification of a 3rd Na+ binding site of the Glycine transporter, GlyT2.PLoS One. 2016; 11e0157583PubMed Google Scholar). Only binding poses in which the lipid tail was inserted into the extracellular allosteric-binding site and the double bond was in close proximity to TM5 were considered (6Mostyn S.N. Wilson K.A. Schumann-Gillett A. Frangos Z.J. Shimmon S. Rawling T. Ryan R.M. O'Mara M.L. Vandenberg R.J. Identification of an allosteric binding site on the human glycine transporter, GlyT2, for bioactive lipid analgesics.Elife. 2019; 8e47150Crossref PubMed Scopus (16) Google Scholar). Regardless of head group amino acid type or stereochemistry, the transporter remains in an outward-occluded conformation (where the intracellular gate distance is >4.1 Å, and the extracellular gate distance is 10 μMaPreviously published data from Mostyn et al., 2019 (6).>10 μMaPreviously published data from Mostyn et al., 2019 (6).14.2aPreviously published data from Mostyn et al., 2019 (6).>10 μMaPreviously published data from Mostyn et al., 2019 (6).C18ω9 Lys25.5aPreviously published data from Mostyn et al., 2019 (6).86.8aPreviously published data from Mostyn et al., 2019 (6).>3 μMaPreviously published data from Mostyn et al., 2019 (6).48.3aPreviously published data from Mostyn et al., 2019 (6).91.0aPreviously published data from Mostyn et al., 2019 (6).>3 μMaPreviously published data from Mostyn et al., 2019 (6).C16ω7 Lys66.6 (49.9–88.7)90.2 ± 2.1>3 μM602 (373–856)96.2 ± 2.1>3 μMC14ω5 Lys703 (414–1250)96.2 ± 7.3>3 μM1380b95% Confidence interval was not able to be calculated. n = 3 to 7 for all measurements.79.7 ± 3.6>3 μMC18ω5 Lys67.5 (31.7–143)81.3 ± 5.1>3 μM64.9 (36.3–123)87.6 ± 3.7>3 μMC16ω3 Lys10.8 (8.37–13.8)94.9 ± 1.6>3 μM699 (343–1480)91.8 ± 8.6>3 μMa Previously published data from Mostyn et al., 2019 (6Mostyn S.N. Wilson K.A. Schumann-Gillett A. Frangos Z.J. Shimmon S. Rawling T. Ryan R.M. O'Mara M.L. Vandenberg R.J. Identification of an allosteric binding site on the human glycine transporter, GlyT2, for bioactive lipid analgesics.Elife. 2019; 8e47150Crossref PubMed Scopus (16) Google Scholar).b 95% Confidence interval was not able to be calculated. n = 3 to 7 for all measurements. Open table in a new tab Both C18ω9-l-Lys and C18ω9-d-Lys remain bound in the allosteric site throughout the combined 1500 ns of MD simulation with the lipid tail in an extended conformation (∼17–20 Å measured from the end of the tail to the stereocenter; Fig. S1), intercalated between TM5, TM7, and TM8 (Figs. 2A and S2). The amino acid head groups remain in close proximity to the protein/bilayer–water interface and interact with the extracellular regions of TM5, TM7, TM8, and EL4, including hydrogen bonding with cationic arginine residues (i.e., R439, R531, and R556) and stacking with nonpolar residues (i.e., F526 and W563). The tail is positioned in a hydrophobic pocket lined by L436, V523, Y550, A553, and F567 (Fig. 2B and C), in agreement with previous studies (6Mostyn S.N. Wilson K.A. Schumann-Gillett A. Frangos Z.J. Shimmon S. Rawling T. Ryan R.M. O'Mara M.L. Vandenberg R.J. Identification of an allosteric binding site on the human glycine transporter, GlyT2, for bioactive lipid analgesics.Elife. 2019; 8e47150Crossref PubMed Scopus (16) Google Scholar). Interactions that occur with both the head group and tail are sustained for >75% of the total simulation time (Table S3). Notably, the more potent inhibitor, C18ω9-l-Lys, has deeper penetration into the binding pocket than C18ω9-d-Lys. Y550 is a key residue in the interaction of both isomers. Y550 interacts with C18ω9-l-Lys above the oleoyl double bond. The change of stereochemistry to C18ω9-d-Lys decreases the depth of penetration of the oleoyl tail and shifts the interaction with Y550 below the double bond of C18ω9-d-Lys. In this position, the Y550 hydroxyl forms hydrogen bonds with the backbone of W563, so that the lipid tail is sandwiched between Y550 and W563, with W563 interacting directly with the double bond with a CH–π interaction. The interactions of Y550 and W563 with both isomers correlate with mutagenesis data reporting that the GlyT2 Y550L and W563L mutants are not inhibited by C18ω9-l-Lys or C18ω9-d-Lys (6Mostyn S.N. Wilson K.A. Schumann-Gillett A. Frangos Z.J. Shimmon S. Rawling T. Ryan R.M. O'Mara M.L. Vandenberg R.J. Identification of an allosteric binding site on the human glycine transporter, GlyT2, for bioactive lipid analgesics.Elife. 2019; 8e47150Crossref PubMed Scopus (16) Google Scholar). In contrast to the lysine-based analogs, the potency of the bioactive lipids bearing Trp head groups is significantly impacted by stereochemistry such that the d-isomer is inactive against GlyT2 (Table 1) (6Mostyn S.N. Wilson K.A. Schumann-Gillett A. Frangos Z.J. Shimmon S. Rawling T. Ryan R.M. O'Mara M.L. Vandenberg R.J. Identification of an allosteric binding site on the human glycine transporter, GlyT2, for bioactive lipid analgesics.Elife. 2019; 8e47150Crossref PubMed Scopus (16) Google Scholar). In MD simulations, C18ω9-l-Trp binds to the GlyT2 allosteric site in a similar conformation as C18ω9-l-Lys. C18ω9-l-Trp is again intercalated between TM5, TM7, and TM8 with the double bond in close proximity to TM5 (Figs. 3A and S3). The d-isomer, C18ω9-d-Trp, alters its orientation around the stereocenter so the C18ω9 tail points toward EL4, rather than facing the pocket formed by TM5, TM7, and TM8. While C18ω9-d-Trp still binds to the GlyT2 extracellular allosteric site, it adopts a much more curled conformation than C18ω9-l-Trp (∼13 Å versus ∼18 Å measured from the end of the tail to the stereocenter; Fig. S1). This curled conformation results in a shallower binding interaction and reduced depth of penetration of C18ω9-d-Trp into the extracellular allosteric site, which alters the coordination of the lipid tail within the binding pocket. Key head group interactions with R439 and F526 were maintained. Residues L436, V523, Y550, A553, and L557 that interact with the lipid tail of C18ω9-d-Lys, C18ω9-l-Lys, and C18ω9-l-Trp instead form interactions with the head group of C18ω9-d-Trp (Fig. 3B and C and Table S3). Only one interaction forms with the C18ω9-d-Trp lipid tail for >75% of the simulation time. Without these key residues stabilizing the position of the acyl tail of C18ω9-d-Trp within the binding site, the tail does not remain bound between TM5, TM7, and TM8 over the course of the MD simulation as observed for the active lipid inhibitors. Instead the lipid tail leaves the allosteric-binding site and reorients in the solution toward EL4, where it adopts a variety of conformations. Despite C18ω9-d-Trp remaining bound in the extracellular allosteric-binding site throughout the total simulation time and the seemingly favorable interaction of the C18ω9-d-Trp head group, no inhibition of GlyT2 is achieved. This indicates that in order for the bioactive lipids to inhibit GlyT2, the tail must be stabilized within the extracellular allosteric-binding site, positioned between TM5, TM7, and TM8. This is in agreement with the inability of free amino acids to cause inhibition (16Carland J.E. Thomas M. Mostyn S.N. Subramanian N. O'Mara M.L. Ryan R.M. Vandenberg R.J. Molecular determinants for substrate interactions with the glycine transporter GlyT2.ACS Chem. Neurosci. 2017; 9: 603-614Crossref PubMed Scopus (15) Google Scholar). The lipid inhibitors described previously contain 18 carbon acyl tails with a cis-double bond in the Δ9 position (i.e., nine bonds from the amide linkage to the amino acid head group), and penetration of the lipid tails into the allosteric site appears to be a critical determinant of inhibitory activity. To investigate the effect of tail length on GlyT2 inhibitory activity, we synthesized d- and l-lysine–based inhibitors with truncated tails. Double bonds were maintained in the Δ9 position, and overall chain lengths were reduced to C16 or C14 (C16ω7-Lys and C14ω5-Lys, respectively). The chemical structures of these new lipids are shown in Figure 1, and their synthesis and characterization are described in the Supporting information section. Newly synthesized acyl lysine analogs were then tested against GlyT2 and also tested for selectivity by testing against the closely related GlyT1 transporter using two-electrode voltage clamp electrophysiology (see Table 1 for inhibitory data). While C18ω9-l-Lys and C18ω9-d-Lys had similar levels of activity, head group conformation greatly affected the potencies of the chain-shortened analogs. C16ω7-l-Lys inhibited GlyT2 with an IC50 of 66.6 nM, but the corresponding d-isomer was ninefold less potent (IC50 of 602 nm; Fig. 4B). Further shortening of the tails to Lys C14ω5 produced a similar preference for l- versus d-, albeit with marked decrease in the potency compared with the C18 analogs (IC50 concentrations of 770 and 1380 nM for Lys-l-C14ω5 and Lys-d-C14ω5, respectively; Fig. 4D). To provide a structural explanation of this difference in activity with changing tail length, C16ω7 and C14ω5 acyl lysines in the l- and d-configurations were docked to the extracellular allosteric-binding site, and 500 ns of unrestrained MD simulations were performed in triplicate. Throughout all simulations, GlyT2 remains in an outward-occluded conformation, regardless of the tail length of the acyl lysines (Table S1). The membrane properties are not affected (Table S2). MD simulations showed that both C16ω7-l-Lys and C14ω5-l-Lys remained bound in the allosteric-binding site throughout the simulation, with the tail positioned between TM5, TM7, and TM8 (Figs. 5A and S4). The Lys head groups of both C16ω7-l-Lys and C14ω5-l-Lys remain in close proximity to the protein/bilayer–water interface, interacting with the extracellular regions of TM5, TM7, TM8, and EL4. C16ω7-l-Lys adopts a similar orientation to that observed for C18ω9-l-Lys, in close proximity to the key binding pocket residues (Fig. 5B and Table S4). As was observed for C18ω9-l-Lys, the C16ω7-l-Lys head group interacts with R436 and F526. Similarly, the l-Lys C16ω7 tail is located between TM5, TM7, and TM8 in the extracellular allosteric-binding site where it interacts with L436, V523, Y550, L557, and F567, and the bottom of the pocket is flanked by V214. As was the case for C18ω9-l-Lys, Y550 interacts with the lipid tail just above the double bond of C16ω7-l-Lys. The similarities between the overall orientation and interactions of C18ω9-l-Lys and C16ω7-l-Lys with key residues in the binding pocket provide a structural basis for the potent inhibition of GlyT2 by both molecules. Further shortening of the lipid tail to give C14ω5-l-Lys significantly reduces the depth of the tail penetration into the binding pocket (Fig. 5A). Furthermore, shortening of the lipid tail dramatically alters C14ω5-l-Lys head group interactions. The C14ω5-l-Lys head group interacts with the membrane, forming hydrogen bonds with 1-palmitoyl-2-oleoyl phosphatidylcholine (POPC) head groups (Fig. S5). There are no interactions with the C14ω5-l-Lys head group that persisted for >75% of the total simulation time (Fig. 5C and Table S4). In contrast, the tail maintains interactions with key residues (i.e., L436, V523, Y550, and F567) for >75% of the total simulation time. In both C14ω5-l-Lys and C16ω7-l-Lys, the tail only penetrates as deep as F567 (Fig. 5, B and C), which interacts with C6 of C18ω9-l-Lys and C5 of C16ω7-l-Lys. This change in orientation of C14ω5-l-Lys (IC50 of 703 nm) relative to C16ω7-l-Lys (IC50 of 66.6 nm) and C18ω9-l-Lys (IC50 of 25.5 nm) correlates with the dramatic decrease in GlyT2 inhibition observed for C14ω5-l-Lys. On change of stereochemistry to the d-configuration, C16ω7-d-Lys and C14ω5-d-Lys only remain bound in the extracellular allosteric site for approximately one third (∼500 ns) of the total simulation time. This is in contrast to C18ω9-d-Lys, which remains bound throughout the entire simulation. Furthermore, C16ω7-d-Lys and C14ω5-d-Lys adopt a different conformation in the extracellular allosteric-binding pocket to C18ω9-d-Lys (Figs. 6A and S6). Specifically, the C16ω7-d-Lys head group interacts with the head groups of membrane POPC lipids (Fig. S4B) while maintaining interactions with key GlyT2 residues (F526, R439, and R556; Fig. 6B and Table S4). While the C16ω7-d-Lys tail group is positioned in the extracellular allosteric-binding pocket in a similar orientation to C18ω9-d-Lys and interacts with L436, V523, Y550, and F567, none of these interactions persist for >75% of the total simulation time (Fig. 6B and Table S4). The reduced occupancy and different binding orientation of C16ω7-d-Lys (IC50 of 602 nm) in the extracellular allosteric-binding site in part explain the 12-fold reduced potency relative to C18ω9-d-Lys (IC50 of 48.3 nm). In the case of C14ω5-d-Lys, the d-Lys head group interacts with F526 and R439 in a similar manner to C18ω9-d-Lys. However, the C14ω5-d-Lys tail protrudes into the surrounding membrane where it interacts with POPC at the protein–lipid interface (Fig. S7). In this orientation, L443 is the only key residue interacting with the lipid tail for >75% of the total simulation time (Fig, 6C). The lack of interactions and reduced occupancy in the extracellular allosteric-binding site may in part explain why C14ω5-d-Lys is not an effective inhibitor of GlyT2 (IC50 of 1380 nm). To assess the effect of the double-bond position on activity, we synthesized C18ω5-Lys and C16ω3-Lys, which contain cis-double bonds in the Δ13 position (Fig. 1). Both l- and d-isomers of C18ω5-Lys inhibit GlyT2 with IC50 concentrations of 67.5 and 64.9 nM, respectively (Fig. 4A and Table 1). This trend is consistent with the results for C18ω9-l-Lys and C18ω9-d-Lys, where the configuration of the amino acid head group did not greatly alter the activity. However, when the length of the acyl chain was decreased to C16, the position of the double bond produced a marked difference in inhibitory activity between isomers. Thus, C16ω3-l-Lys is a potent inhibitor of GlyT2 (IC50 of 10.8 nM), whereas the corresponding d-isomer was 65-fold less potent, with an IC50 of 699 nm (Fig. 4C). To provide a structural explanation of this difference in activity with changing tail length, the l- and d-isomers of C18ω5 and C16ω3 Lys were docked to the extracellular allosteric-binding pocket and simulated for 500 ns in triplicate. Regardless of the presence of a bound inhibitor, GlyT2 again remains in an outward-occluded conformation (Table S1) throughout the simulations, and the membrane properties are not affected (Table S2). Both stereoisomers with C18 tails (C18ω5-l-Lys and C18ω5-d-Lys) remained bound in the allosteric pocket throughout all simulations. The overall binding conformation and potencies of C18ω5-l-Lys and C18ω5-d-Lys were similar to that of the C18ω9-l-Lys and C18ω9-d-Lys (Fig. 7A). The head group interactions with F526, L443, and R439 are maintained (Figs. 7B and C and S8 and Table S5), and the tails of C18ω5-l/d-Lys are located between TM5, TM7, and TM8, interacting with the nonpolar residues (e.g., V214, L436, Y550, L557, and F567) and Y550 interacts with the lipid tail above the double bond (Fig. 7), giving a structural basis for their activity. Major differences in the binding interactions of the C16 isomers were observed, consistent with their IC50 concentrations. C16ω3-l-Lys, the most potent inhibitor of GlyT2 in this lipid series, remains bound in the allosteric pocket throughout the combined MD simulations, whereas C16ω3-d-Lys only remains bound in the extracellular allosteric-binding pocket for approximately one-third of the total simulation time. In the case of C16ω3-l-Lys, the head group interacts with F526, L443, and R439, as observed for C18ω9-l-Lys (Figs. 8 and S9). Unlike other inhibitory lipids, the C16ω3-l-Lys tail does not adopt an extended conformation but instead has a curled conformation in the allosteric-binding site between TM5, TM7, and TM8. The curled C16ω3-l-Lys tail interacts with W215, Y550, L557, F567, and L436 (Fig. 8B and Table S5). The interaction between C16ω3-l-Lys and W215 is unique and notable because W215 is directly adjacent to the glycine-binding site and physically separates the extracellular allosteric site and the vestibular substrate-binding site. Interactions with W215 may reflect communication between the extracell