Synthesis and biological evaluation of α- and β-hydroxy substituted amino acid derivatives as potential mGAT1–4 inhibitors

In this study, we report the synthesis and biological evaluation of a variety of α- and β-hydroxy substituted amino acid derivatives as potential amino acid subunits in inhibitors of GABA uptake transporters (GATs). In order to ensure that the test compounds adopt a binding pose similar to that presumed for related larger GAT inhibitors, lipophilic residues were introduced either at the amino nitrogen atom or at the alcohol function. Several of the synthesized compounds were found to exhibit similar inhibitory activity at the GAT subtypes mGAT2, mGAT3, and mGAT4, respectively, as compared with the reference N-butylnipecotic acid. Hence, these compounds might serve as starting point for future developments of more complex GAT inhibitors.


Introduction
Gamma-aminobutyric acid (GABA) is the most abundant inhibitory neurotransmitter in the mammalian central nervous system (CNS) (Bowery and Smart 2006), with up to 40% of synapses estimated to be GABAergic (Meldrum and Chapman 1999). Deficient GABAergic neurotransmission is assumed to play a decisive role in the pathogenesis of several severe neurological diseases, including Alzheimer's disease (Lanctôt et al. 2004), depression (Kalueff and Nutt 2007), epilepsy (Treimann 2001), and neuropathic pain (Daemen et al. 2008). A promising therapeutical approach for the treatment of these diseases exists in the inhibition of the transport molecules responsible for the removal of GABA from the synaptic cleft, resulting in the prolongation of the effect exerted by the available GABA (Krogsgaard-Larsen et al. 1991). Belonging to the solute carrier 6 (SLC6) transporter gene family, these membrane-bound proteins termed GABA transporters (GATs) use the co-transport of sodium ions for the translocation of the substrate against the chemical gradient (Kristensen et al. 2011). The nomenclature of the four GAT subtypes depends on the species the transporters are cloned from. When originating from mouse brain cells, the GAT subtypes are termed mGAT1-4. For all other species, including humans, an alternate nomenclature is used which has also been adopted by the Human Genome Organisation, denoting the transporters GAT1 (=mGAT1), BGT1 (=mGAT2), GAT2 (=mGAT3), and GAT3 (=mGAT4) (Madsen et al. 2009). Hereafter the nomenclature referring to the murine transporters will be applied as the biological test system used in our group is based on these. mGAT1 and mGAT4 are the most abundant GATs in the mammalian CNS, with the former being predominantly located on presynaptic neuronal membranes mediating the neuronal GABA uptake, whereas mGAT4, which is mainly expressed on glia cells, is responsible for the glial GABA uptake (Minelli et al. 1996;Jin et al. 2011). The other two GAT subtypes, mGAT2 and mGAT3, are primarily located in the periphery, with the highest densities being found in liver and kidneys, and hence are thought to not play any significant role in the termination of GABAergic signalling in the CNS (Zhou et al. 2012).
When bulky, lipophilic residues were introduced at the nitrogen atom, this led to compounds with not only increased lipophilicity and hence improved BBB penetration, but also with significantly higher inhibitory potency. In this context, compounds possessing a diaryl methyl or a biaryl unit as lipophilic domain which is connected to the nipecotic acid partial structure via a flexible linker of 3-5 atoms proved to be highly potent and selective mGAT1 inhibitors, with the respective pIC 50 values rising from~5 to almost 7 as compared with unsubstituted (R)-nipecotic acid. Among these compounds, Tiagabine (Gabatril®, 4, Table 1, entry 5) stands out for being the only GAT inhibitor that has been approved for clinical use (Nielsen et al. 1991).
In a similar way the introduction of a lipophilic residue consisting of a triarylmethyl group, which is linked to the nipecotic acid subunit via a spacer of three atoms, furnishes mGAT4 selective inhibitors, with (S)-SNAP-5114 constituting the prototypic representative of this group (5, Table  1, entry 6) (Dhar et al. 1994). Remarkably, the (S)-isomer 5 exhibits higher inhibitory potency than the respective (R)isomer, running contrary to what is observed for the unsubstituted nipecotic acid [(R)-3, (S)-3] as well as for mGAT1 inhibitors such as Tiagabine (4).
Interestingly, small amino acids including β-alanine (6, Table 2, entry 1), isoserine (7a, Table 2, entry 2), 2,3-diaminopropionic acid (8, Table 2, entry 3) and (Z)-4-aminobut-2enoic acid (9, Table 2, entry 4) also show moderate biological activity at mGAT3 and mGAT4 that is comparable to that exerted by (R)-nipecotic acid [(R)-3, Table 1, entry 3], while being distinctly less potent inhibitors of mGAT1 (Kragler et al. 2005). Thus, substitution of the nipecotic acid partial structure in the lead compound (S)-SNAP-5114 (5) with a (S)-2hydroxy-2-[(R)-pyrrolidin-2-yl]acetic acid unit, which can be understood as rigidized derivative of rac-isoserine (7a), led to  (Steffan et al. 2015). For the development of mGAT4 inhibitors with increased subtype selectivity, this study hence aims to identify further cyclic and acyclic 2-and 3-hydroxyamino acids as possible alternatives to the nipecotic acid partial structure present in the scaffold of important mGAT4 inhibitors such as 5. In this context, the hydroxyl function plays a crucial role as it may enable additional interactions with the target, e.g., by establishing hydrogen bridges, hence increasing binding enthalpy. Proceeding from the basic structure of isoserine 7a, it was intended to implement several structural modifications, including elongation of the carbon chain by insertion of methylene groups at various positions, and the rigidization of the molecule by integrating the amino acid backbone into larger heterocycles. Accordingly, a set of compounds featuring both variations and thus deviating from the original isoserine (7a) structure should be synthesized and biologically evaluated (Scheme 1).
Despite the stereochemistry of amino acids such as nipecotic acid [(R)-3, (S)-3] being known to represent an important factor when it comes to the biological activity of mGAT4 inhibitors, we opted for the synthesis of the target compounds in racemic form as this provides information about the biological activity of both enantiomers.
According to the results of molecular modelling experiments previously performed by us for mGAT1 (Wein et al. 2016), small inhibitors such as nipecotic acid adapt a binding pose at which the amino nitrogen atom is facing towards the intracellular space. However, if bulky, lipophilic moieties are introduced at the nitrogen atom, as it is the case with Tiagabine (4, Table 1, entry 5), the binding pose is altered in a way that the nitrogen atom is orientated towards the extracellular site, with the lipophilic side chain looming into the extracellular vestibule, although the position of the carboxylic acid function remains largely unchanged (Scheme 2). Using the example of a N-butyl residue, Wein et al. demonstrated that even the presence of small Nsubstituents is sufficient to cause this change of the binding mode. Although no reliable in silico models exists so far for the other GAT subtypes, it seems highly likely that these findings also apply to mGAT2-4. Thus, following this assumption, all parent compounds included in this study were also evaluated for their inhibitory potential as N-butyl derivatives in order to ensure that the orientation of the parent compounds in the binding pocket corresponds to the orientation they would have as part of larger molecules comprising a lipophilic domain. In addition, unsubstituted αand β-hydroxyamino acids exhibiting higher biological activity than their corresponding N-butyl derivatives should in addition be substituted at the alcohol function. In detail, the alcohol function should be provided with a 4,4′,4″-trimethoxytrityloxyethyl residue   is a characteristic structural motif of mGAT4 inhibitors such as 5 and 10. According to the model pointed out above, this might allow the amino acid substructure to stay in the position with the nitrogen atom facing towards the intracellular space, which would evidently be more favourable than the antagonal orientation found in the Nsubstituted amino acids. At the same time, the newly introduced lipophilic domain might be accommodated in the vestibule, which is known to contribute in general significantly to inhibitory potency and selectivity of GAT inhibitors (compare Table 1, entries 3 and 4 with 6).

Chemistry
Moisture-sensitive reactions were carried out in oven-dried glassware under inert gas atmosphere. Commercially available starting materials were used without further purification. Tetrahydrofuran (THF) was freshly distilled from sodium benzophenone ketyl. All other solvents were distilled prior to use. Microwave reactions were carried out with Biotage Initiator™. Flash column chromatography was performed on Merck silica gel 60 (mesh 0.040-0.063 mm) as stationary phase; thin-layer chromatography (TLC) was carried out on Merck silica gel 60 F 254 sheets. Preparative MPLC was performed using a Buechi instrument (C-605 binary pump system, C-630 UV detector at 254 nm and C-660 fraction collector) and a Sepacore B-685 (26 V, 230 mm) glass column equipped with YMC Gel Triart Prep C18-S (12 nm, 5-20 µm). 1 H and 13 C NMR spectra were, unless stated otherwise, recorded at rt with JNMR-GX (JEOL 400 or 500 MHz) or Bruker BioSpin Avance III HD (400 or 500 MHz) and integrated with the NMR software MestRe-Nova. Deuterated solvents for NMR, which included CD 2 Cl 2 , CDCl 3 , MeOD, D 2 O, and 1,1,2,2-tetrachloroethane-d2, were purchased from Eurisotop, Saint-Aubin, France. NaOD was purchased from Sigma-Aldrich Chemie GmbH, Munich, Germany. IR samples were measured as KBr pellets or film with Perkin-Elmer FT-IR 1600. HRMS data were obtained with JMS-GCmate II (EI, Jeol) or Thermo Finnigan LTQ FT Ultra (ESI, Thermo Finnigan).
General procedure for the synthesis of the β-hydroxyamino acid esters (GP1) A solution of lithium-HMDS in MTBE (0.97 M) was cooled to −78°C and anhydrous ethyl acetate (EtOAc, 1.0 eq) was added dropwise. After complete addition the mixture was stirred for 25 min and a solution of ketone (1.0 eq) in dry THF was added. The reaction mixture was allowed to slowly warm up to −10°C, quenched with brine, diluted with H 2 O and extracted thrice with CH 2 Cl 2 . The combined CH 2 Cl 2 phases were dried over sodium sulfate (Na 2 SO 4 ) and concentrated in vacuum to yield the crude product.
General procedure for the deprotection and N-butylation of the amino acid amides and amino acid esters (GP2) A mixture of the benzyl or benzhydryl-protected amino acid derivative, palladium on charcoal (10% Pd, 0.1 eq) and butyraldehyde (2.5 eq) in EtOH was stirred vigorously under 15 bar hydrogen pressure for 16 h, filtered over Cealite® and reduced in vacuum.
General procedure for the hydrolysis of the amides and esters (GP3) A mixture of carboxamide or ester and barium hydroxide octahydrate (2.0-2.4 eq) was stirred under reflux conditions (carboxamides) or at rt (esters) in EtOH/H 2 O 1:1 for the appropriate time. Carbon dioxide was passed through until no further precipitate formed. The suspension was filtered over a cotton wool pad and a syringe filter (Perfect-Flow®, WICOM Germany GmbH, PTFE, 0.2 µM) and reduced in vacuum. The residue was solved in distilled water (2.0 ml) and lyophilized.
General procedure for the N-butylation of the acyclic amino acids (GP4) The amino acid (1.0 eq) and KOH (2.0 eq) were suspended in EtOH and H 2 O was added until the reaction mixture became homogeneous. 1-Bromobutane (0.9 eq) was added dropwise.
After stirring at rt for 16 h the reaction mixture was reduced in vacuum. The crude compound was purified by MPLC (eluent: MeOH/H 2 O 1:9).
General procedure for the protection of the acyclic amino acids I (GP5a) A solid, well-grounded mixture of the amino acid (1.0 eq) and phtalic anhydride (1.0 eq) was heated to 140°C, resulting in a colourless melting. After 30 min, the reaction mixture was cooled to rt and re-dissolved in EtOAc (300 ml). The solution was washed with 1 M sodium hydrogen sulfate solution (100 ml), water (3 × 100 ml), and brine (100 ml). The organic layer was dried over Na 2 SO 4 , filtered, and concentrated in vacuum. The resulting residue was solved in anhydrous MeOH (250 ml) and 2 M HCl in Et 2 O (250 ml) was added. The mixture was stirred at rt until TLC indicated complete consumption of the educt and reduced in vacuum. The crude compound was purified by flash column chromatography on silica (eluent: CH 2 Cl 2 / EtOAc 9:1).
General procedure for the protection of the acyclic amino acids II (GP5b) A solid, well-grounded mixture of the amino acid (1.0 eq) and phtalic anhydride (1.0 eq) was heated to 140°C, resulting in a colourless melting. After 30 min, the reaction mixture was cooled to rt, solved in anhydrous MeOH (250 ml) and 2M HCl in Et 2 O (250 ml) was added. The mixture was stirred at rt until TLC indicated complete consumption of the educt and reduced in vacuum. The crude compound was purified by flash column chromatography on silica (eluent: CH 2 Cl 2 /EtOAc 9:1).
General procedure for the formation of the ether function (GP6) tert-Butyl(2-iodoethoxy) diphenylsilane (1.4 eq) and Ag 2 CO 3 (4.0 eq) were added to a suspension of the amino acid derivative (1.0 eq) in toluene (10.0 ml). The reaction mixture was stirred in a pressure tube at 120°C until TLC indicated complete consumption of the amino acid derivative, cooled to rt, filtered through a paper filter and reduced in vacuum. The residue was purified by flash column chromatography on silica (eluent: pentane/Et 2 O 7:3).
General procedure for cleavage of the TBDPS protecting group (GP7) The TBDPS-protected compound (1.0 eq) was solved in THF/pyridine 9:1 (v/v) in a polypropylene tube. A 70% solution of HF-pyridine (5.0 eq) was added dropwise at 0°C. The suspension was stirred at rt and the progress of the reaction was monitored by TLC. After complete consumption of the educt phosphate buffer (pH = 6.0, 1.0 M, 100 ml) was added and the mixture was extracted with ethyl acetate (100 ml). The organic phase was washed with water (100 ml) and brine (100 ml), dried over MgSO 4 and reduced in vacuum. The crude product was purified by flash column chromatography on silica (eluent: CH 2 Cl 2 / EtOAc 65:35).
General procedure for the deprotection of the acyclic amino acid derivatives I (GP9a) 12 M NaOH (2.0 eq) was added to a solution of the protected compound (1.0 eq) in MeOH (15.0 ml). After stirring for 16 h at rt, 1,2-diaminoethane (7.0 eq) was added and the mixture was heated in a microwave for 16 h at 140°C. Finally, the reaction mixture was reduced in vacuum and purified by MPLC (eluent: MeOH/ H 2 O 7:3).
General procedure for the deprotection of the acyclic amino acid derivatives II (GP9b) 12 M NaOH (2.0 eq) was added to a solution of the protected compound (1.0 eq) in MeOH (15.0 ml). After stirring for 16 h at rt, the mixture was freeze-dried. 1,2-Diaminoethane (7.0 eq) was added and the mixture was heated in a microwave for 16 h at 140°C. Finally, the reaction mixture was reduced in vacuum and purified by MPLC (eluent: MeOH/H 2 O 7:3).

1-Benzyl-3-hydroxypiperidine-3-carboxamide (13b)
1-Benzylpiperidin-3-one hydrochloride 13a (571 mg, 2.50 mmol) was solved in dry CH 2 Cl 2 (4.0 ml). Freshly distilled triethylamine (0.85 ml, 6.1 mmol) was added and the brownish suspension was placed in an ultrasound bath for 15 min. After addition of trimethylsilyl cyanide (0.80 ml, 6.3 mmol) the reaction mixture was stirred for 48 h, diluted with CH 2 Cl 2 (10.0 ml), filtered through a paper filter and reduced in vacuum. The oily residue was solved in CH 2 Cl 2 (7.0 ml) and cooled to 0°C. Concentrated sulfuric acid (0.70 ml, 13 mmol) was added and the biphasic mixture was stirred for 2 h at rt, after which the mixture was cooled to 0°C, diluted with H 2 O (5.0 ml) and alkalized with 25% ammonium hydroxide solution. Potassium sodium tartrate (0.50 g) was added and the mixture was extracted five times with CH 2 Cl 2 (20.0 ml). The combined organic phases were dried over MgSO 4 and reduced in vacuum. The crude product was purified by flash column chromatography on silica (eluent: ethyl acetate + 3% triethylamine) to afford the desired compound as amorphous off-white solid (567 mg, 96%) .

MS Binding Assays
The MS Binding Assays were performed as reported (Zepperitz et al. 2006) with mGAT1 membrane preparations obtained from a stable HEK293 cell line and NO711 as unlabelled marker in competitive binding experiments.

Synthesis of the cyclic N-butylamino acid derivatives
For the synthesis of the cyclic N-butylhydroxyamino acids 11d, 11g, 12g, and 13d we intended to start from the benzyl-or benzhydryl-protected cyclic aminoketones 11a, 12a and 13a, respectively. Reaction of 11a and 13a with tetramethylsilyl cyanide followed by acidic hydrolysis of the thus to be formed TMS-protected cyanohydrines should furnish the corresponding α-hydroxycarboxamides 11b and 13b. Likewise, reaction of the cyclic aminoketones 11a and 12a with lithium 1ethoxyethen-1-olate should lead to the β-hydroxyesters 11e and 12e. Deprotection of the amino nitrogen atom of 11b, 11e, 12e, and 13b, followed by introduction of a n-butyl rest via reductive amination and hydrolysis of the carboxamide and ester function, respectively, should finally furnish the desired free amino acids 11d, 11g, 12g, and 13d.
Synthesis of the cyclic α-hydroxycarboxamide 11b was performed according to literature (Lamb 2008). The same reaction sequence was applied for the synthesis of its ring-expanded analogue 13b. Hence, the Nbenzylpyrrolid-3-one 13a was reacted with trimethylsilyl cyanide to give the respective TMS-protected cyanohydrine. Due to the lability of the TMS ether function, the TMS-protected cyanohydrine was not purified and characterized, but used for the next reaction step, the hydrolysis with concentrated sulfuric acid. This gave the corresponding α-hydroxycarboxmide 13b in an excellent yield of 96% over both reaction steps. The β-hydroxyesters 11e and 12e were obtained in good yields of 80-93% from the cyclic aminoketones 11a and 12a by reaction with lithium 1-ethoxyethen-1-olate, which was generated from ethyl acetate and LiHMDS at low temperature (Scheme 3).

Synthesis of the acyclic N-butylamino acid derivatives
For the synthesis of the N-butyl derivatives 7b, 14b, and 15b from the cyclic hydroxyamino acids 7a, 14a, and 15a, a specialized procedure developed for the monobutylation of β-alanine (Santimukul and Perez 2011) was followed. When according to this procedure 7a, 14a, and 15a were treated with n-bromobutane in a mixture of methanol and water under reflux the desired test compounds 7b, 14b, and 15b were obtained in yields of 75-79% (Scheme 4).

Synthesis of O-alkylated hydroxyamino acid derivatives
Since the unsubstituted hydroxyamino acids 7a, 14a, and 15a had been found to exhibit higher inhibitory potencies at all GAT subtypes than the corresponding N-butyl derivatives 7b, 14b, and 15b (see chapter 3 "biological evaluation", Table 3), also the respective amino acid derivatives with an tris(4-methoxyphenyl)methyloxyethyl attached to the hydroxy function of the parent compounds should be included in this study.
As starting compounds for this synthesis of the target compounds 7g, 14g, and 15g, the derivatives 7c, 14c, and 15c seemed well suited, as they should allow a selective functionalization of the OH group as the carboxylic acid and the amino moieties in 7g, 14g, and 15g are protected in form of ester and phthalimidic moieties, respectively. The preparation of 14c has been described in literature (Farkas et al. 2009), the synthesis of 7c and 15c should be accomplished in an analogous manner. Based on these starting materials, 7c, 14c, and 15c, in the next steps first a 2-hydroxyethyl residue should be attached to the free OH function to serve as linker to the lipophilic domain of the target compound. The trityl based lipophilic moiety should be introduced only after that, as the reaction conditions required for the formation of the first ether function were thought to cause side reactions if the terminal trityl moiety were already present. Deprotection of the carboxylic acid and the amino group in 7f, 14f, and 15f should finally lead to the target compounds 7g, 14g, and 15g.
The final deprotection of the amino and the carboxylic acid function of 7f, 14f, and 15f in order to obtain the free amino acids 7g, 14g, and 15g was first attempted in a twostep reaction sequence. Hydrazinolysis of the phthalimide moiety should liberate the terminal amino group, and subsequently hydrolysis of the methyl ester function under alkaline conditions (NaOH) the carboxylic acid moiety. Unfortunately, when the primary amine was formed in the first reaction step, it immediately reacted with the methyl ester function leading to the formation of the corresponding lactame, which could not be cleaved again without destruction of the molecule. Hence, the sequence of the deprotection was altered applying first NaOH to hydrolyse the methyl ester function. Thereby, also the phthaloyl group protecting the amino function was partially cleaved leading to the corresponding phthalamide moieties. Still, the free amino acids 7g, 14g, and 15g could be obtained by subjecting the thus obtained crude reaction product without prior isolation to heating with 1,2-diaminoethane. This furnished the desired target compounds 7g, 14g, and 15g in yields of 69-83% over both reaction steps (Scheme 6).

Biological evaluation
The amino acids 7a, 7b, 7g, 11d, 11g, 12g, 13d, 14a, 14b, 14g, 15a, 15b, 15g, as well as the carboxamide and ester derivatives 11c, 11f, 12f, and 13c were tested for their inhibitory potencies on the four GABA transporter subtypes mGAT1-4 in a [ 3 H]GABA uptake assay previously developed by our group (Kragler et al. 2008) The tests were performed in a standardized manner in triplicates using HEK293 cell lines, each expressing one of the four GAT subtypes. In addition, binding affinities towards mGAT1 were examined employing a standardized MS Binding Assay with NO711 as native MS marker (Zepperitz et al. 2006). The results are summarized in Table 3. Inhibitory potencies and binding affinities of the tested compounds are represented as pIC 50 and pK i , respectively. Each test compound was characterized in three independent experiments performed in triplicates and the standard error of mean (SEM) is given. If the determination of the pIC 50 value proved not feasible due to low inhibitory potency, as percentage of the remaining [ 3 H] GABA uptake at 100 µM concentration of the test compound. Correspondingly, the percentage of remaining MS marker is Scheme 5 Synthesis of the protected amino acid derivatives 7c, 14c, and 15c. Reagents and conditions: a 1. Phthalic anhydride (1.0 eq), 140°C, 30 min; 2. MeOH, 2 M HCl in Et 2 O, rt given in cases when the tested compound caused only a minor reduction of the MS marker binding.
Unexpectedly, rigidization of the N-butylisoserine (7b) molecule by linking the C-2 atom with the nitrogen atom via a methylene bridge, resulting in azetidine heterocycle 11d, leads to a compound which exerts its highest inhibitory potency at mGAT2 (pIC 50 = 3.38 ± 0.08, Table 3, entry 6). At the same time the effects of 11d at the other GATs are minor, the [ 3 H]GABA uptake amounting to 72% (mGAT3), 85% (mGAT4), and 89% (mGAT1). Formal enlargement of the azetidine ring present in 11d to a piperidine ring causes the inhibitory potency at mGAT2 to completely vanish (13d, [ 3 H]GABA at 100 µM = 104%, Table 3, entry 12). In addition, the activity exerted at the other GAT subtypes is extremely low, the values for the remaining [ 3 H]GABA uptake at 100 µM being 96% (mGAT1), 85% (mGAT3), and 91% (mGAT4), respectively. As compared with its desoxy analogoue N-butylnipecotic acid (16, Table 3, entry 1), this constitutes a distinct decline of inhibitory potency at all GATs, hence demonstrating that the introduction of a hydroxy group into the 3-position of the nipecotic acid scaffold is associated with a strong reduction of biological activity.
Linking the C-3 atom and the amino nitrogen atom of 15b via a C 2 -bridge, resulting in a pyrrolidine substructure, is likewise accompanied by a complete loss of inhibitory potency at mGAT3-4, the values for the remaining [ 3 H] GABA uptake being nominally 103% and 105%, respectively (12g, Table 3, entry 10). However, the compound exerts some activity at mGAT2 (pIC 50 = 3.17). For mGAT1, a value of 94% remaining GABA uptake at 100 µM was determined, which is consistent with the parent compound 15b not showing any effect at this subtype (15b, remaining [ 3 H]GABA uptake = 100%, Table 3, entry 17).
For all test substances the binding affinities at mGAT1 were found to be very low as compared with the reference compound 16 (pK i = 3.36 ± 0.02, Table 3, entry 1), the only exception being 14a (pK i = 3.61 ± 0.05, Table 3, entry 11).

Conclusion
A series of cyclic and acyclic hydroxyamino acid derivatives was synthesized and biologically evaluated for their potential as amino acid subunits in GAT inhibitors. Among the compounds synthesized for this study, we identified 1butyl-3-hydroxyazetidine-3-carboxylic acid and, even more so, 2-(1-butyl-3-hydroxypyrrolidin-3-yl)acetic acid to be selective and moderately potent inhibitors of mGAT2, with the respective pIC 50 values amounting to 3.17 and 3.38 ± 0.08, respectively.
Interestingly, the pIC 50 values of the unsubstituted, openchained hydroxyamino acids are in most cases more than one log unit higher at all GAT subtypes than the values of the respective N-butyl analogues, indicating that Nderivatisation might carry an energy penalty by forcing the compound into a less favourable binding pose. Therefore, each of the hydroxyamino acids was derivatized by linking the alcohol function to a C 2 spacer bearing a 4,4′,4″trimethoxytrityloxy moiety, which is a common structural motif of mGAT4 inhibitors such as (S)-SNAP-5114. In theory, this would allow the amino acid subunit to adapt the more favourable binding pose, while at the same time enabling interactions between the target and the lipophilic domain. Unfortunately, the resulting compounds displayed poor inhibitory potency and selectivity as compared with the unsubstituted hydroxyamino acids and, even more so, as compared with (S)-SNAP-5114.
Acknowledgements Open Access funding provided by Projekt DEAL.

Compliance with ethical standards
Conflict of interest The authors declare that they have no conflict of interest.
Publisher's note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons. org/licenses/by/4.0/.