Chemoenzymatic Synthesis of ortho-, meta-, and para-Substituted Derivatives of l-threo-3-Benzyloxyaspartate, An Important Glutamate Transporter Blocker

A simple, three-step chemoenzymatic synthesis of l-threo-3-benzyloxyaspartate (l-TBOA), as well as l-TBOA derivatives with F, CF3, and CH3 substituents at the aromatic ring, starting from dimethyl acetylenedicarboxylate was investigated. These chiral amino acids, which are extremely difficult to prepare by chemical synthesis, form an important class of inhibitors of excitatory amino acid transporters involved in the regulation of glutamatergic neurotransmission. In addition, a new chemical procedure for the synthesis of racemic mixtures of TBOA and its derivatives was explored. These chemically prepared racemates are valuable reference compounds in chiral-phase HPLC to establish the enantiopurities of the corresponding chemoenzymatically prepared amino acids.

l-Glutamate is the major excitatory neurotransmitter in the mammalian central nervouss ystem and, as such, contributes to neuronal signaling throughi ts activation of av ariety of glutamate-gated ion channels. [1] Excitatory amino acid transporters (EAATs)a re responsible for the uptake of glutamate from the synaptic cleft, which therebyt erminates the glutamatergic neurotransmitter signal. [1,2] Hence, EAATs present on neurons and surrounding glia cells play ac ritical role in regulatings ynaptic glutamate concentrations.A ccumulationo fe xcitotoxic levels of extracellularg lutamate may lead to overactivation of glutamate-gated ion channels and, consequently,n euronal injury.G lutamate-mediated neuronal injury has been linked to severaln eurologicald isorders such as amyotrophic lateral sclerosis, Alzheimer's disease, epilepsy,a nd Huntington's disease. [3] Studies on EAATs ,o fw hich five subtypes (EAAT1-5) have been identified, have been largely dependentu pon the development of selective and potent inhibitors that can be used to probe the physiological roles of these transporters in the regulation of glutamatergicneurotransmission or in the pathogenesis of neurologicald iseases. [1,3] Aspartate derivatives with large aryl or aryloxy substituents at the C3 position form an important class of inhibitors of EAATs . [3,4] This is exemplified by one of the first EAATi nhibitors to be reported, l-threo-3-benzyloxyaspartate (l-TBOA), which is aw idely used nontransportable blocker for all five EAATs ubtypes. [4a-c] However,t he chemical synthesis of l-TBOA,t he enantiomer of TBOA with the most potent inhibitory properties,i sahighly challenging 11-step procedure startingf rom (R)-Garner aldehyde. [4c,d] Therefore, there is great interest in the development of alternative procedures that provide simple and environmentally friendly access to l-TBOA and derivatives of l-TBOA with variouss ubstituents on the aromatic ring.
Several research groups have explored ammonia lyases as biocatalysts in the asymmetric amination of unsaturated acids to yield chiral a-amino acids. [5] This is av ery attractive strategy for amino acid synthesis, because it makes use of readily available starting compounds without the requirement for recycling of cofactors, implementation of dynamic kinetic resolution strategies, or additional enzymes. The academic and industrial interesti na spartate derivatives, combined with the potential advantages of replacing chemical processes with biocatalysis, has promptedu st of ocus our attention on methylaspartate ammonia lyase (MAL). [6] This enzyme catalyzes the reversible addition of ammonia to mesaconate to give l-threo-3-methylaspartate and l-erythro-3-methylaspartate as products. [7] However,t he wild-type enzymeh as an arrow electrophile scope and only displays amination activity towards fumaratea nd its derivatives with as mall substituent at the C2 position. [8][9][10] Interestingly,r ecent protein engineering work on MAL has shown that ag lycineo ra na lanine mutation at position Leu384 resulted in mutant enzymes with the ability to aminate fumarate derivatives with large substituents at the C2 position. [10] The synthetic potentialo ft he L384A mutant has been demonstrated by its use as ab iocatalyst in the asymmetric synthesis of various3 -substituteda spartate derivatives, including the important EAATinhibitor TBOA. [10,11] In this report, we describe the synthetic potentialo ft he L384A and L384G mutants of MAL in the selectivep reparation of both enantiomerso fT BOA (1a= l-TBOA), as well as as eries of TBOA derivatives (i.e., 1b-f, 1h,a nd 1i)w ith different substituents at the ortho, meta,a nd para positions of the aromatic ring (Scheme 1). This convenient three-step chemoenzymatic methodfor the stereoselective synthesis of TBOA and ring-substitutedd erivatives of TBOA (starting from dimethyl acetylenedicarboxylate 4)a ppears to be an elegant alternative for existing, highlyc hallengingm ultistep-synthesis procedures. [4c,d] To investigate the efficiency of the L384G and L384A mutants to aminate 2-benzyloxyfumarate (2a), kinetic parameters were determined ( Table 1). The data indicatet hat the two MAL mutantsa minate 2a with similar kinetic parameters and catalytic efficiency,h aving respectable k cat values of 6-7 s À1 .W e also compared the ability of these mutant enzymes to aminate 2a under optimized reaction conditions ( Figure 1). With 0.01 mol %biocatalysta nd a100-fold molar excessofa mmonia (5 m)r elative to the amount of 2a (50 mm), the reactions were complete within just af ew hours at pH 9.0 and room temperature, and excellent conversionso f> 80 %w ere achieved.T hese results clearly demonstrate the potential of these mutantsf or application in the synthesis of 3-benzyloxyaspartate.
To further demonstratet he synthetic usefulness of the MAL mutants, ap reparative-scale reaction using 0.01 mol %b iocatalyst (the L384Gm utant), ammonia (5 m), and 2a (0.75 mmol, 50 mm)w as performed at pH 9.0 and room temperature. The reactionw as stopped after 24 ha tafinal conversion of 90 % ( Table 2, entry 1). The amino acid product was purified (78 % yield) and identified ast he desired threo isomer of 3-benzyloxyaspartate (de > 95 %) by comparison of its 1 HNMR signals to those of chemically synthesized authentic standards with known threo or erythro configuration (see the Supporting Information). No other regio-or diastereoisomer was observed. The absolutec onfigurationo ft he product was determined by high-performancel iquid chromatography (HPLC)o nachiral stationaryphase by using chemically prepared authentic standards with known dl-threo or l-threo configuration (FigureS3, Supporting Information). This analysisr evealed that the product had the l-threo configuration and was presenta sasingle enantiomer (1a, l-TBOA; ee > 99 %).
The preparative-scale experiment wasrepeated under identical conditions but by using the L384A mutant as ab iocatalyst insteado ft he L384Gm utant.A nalysis of the isolateda mino acid product revealed that the L384A-catalyzed amination of 2a was also highly regio-and stereoselective and that l-TBOA was exclusively produced (de > 95 %, ee > 99 %). We also explored the L384A mutant for the kinetic resolution of aracemic mixture of dl-TBOA (5 mg, 0.021 mmol) in an analytical-scale experiment. The reaction was followed by monitorings ubstrate depletion and product formation by HPLC on ac hiral stationaryp hase. With 0.1 mol %b iocatalyst, the reaction was complete within 24 ha tp H8.0 and room temperature, and   Next, as eries of 2-benzyloxyfumarate derivatives 2b-j (Scheme 1) was preparedf rom dimethyl acetylenedicarboxylate (4)a nd evaluated as substrates for the L384G and L384A mutants. The ability of these MAL mutants to catalyze the amination of substrates with as mall (e.g.,F )o rm ore bulky (e.g., CF 3 and CH 3 )s ubstituent at the ortho, meta,o rpara positiono f the aromatic ring (i.e., 2b-j)w as tested by using 1 HNMR spectroscopy( Table 2). The L384Am utant displayed activity for 2benzyloxyfumarate derivatives 2b-e, 2h,a nd 2i.R eaction mixtures were incubated at pH 9.0 and room temperature, and conversionso f> 50 %a fter 3h and > 90 %a fter 24 hw ere achieved. Strikingly,m utant L384G not only showed activity for compounds 2b-e, 2h,a nd 2i,b ut it also processed 2-benzyloxyfumarate derivative 2f,w hichw as not converted at all by the L384A mutant. Again, excellent conversions were achieved after incubation for 24 ha tp H9.0 and room temperature. The aryl binding pocket of theL 384Gm utant is most likely slightly larger than that of the L384A mutant, [10] and this rationalizes its ability to convert compound 2f,w hichc arries al arge trifluoromethyl group at the meta position. Compounds with at rifluoromethyl( 2g)o rm ethyl (2j)s ubstituent at the para positionw ere not accepted as substrates by the L384A mutant nor the L384Gm utant. Because the L384G mutant showed al arger substrate scope, it was selected as the biocatalyst for applicationi np reparative-scale reactions.
Preparative-scale experimentsw erep erformedt oa llow unambiguous product identification by HRMS and 1 HNMR, 13 CNMR, and/or 19 FNMR spectroscopya nd thus to ascertain that the L384G-catalyzed amination of 2b-f, 2h,a nd 2i yields the corresponding amino acid products 1b-f, 1h,a nd 1i (Scheme 1). Ammonia (5 m), 2-benzyloxyfumarate derivative (0.75 mmol, 50 mm), and biocatalyst (either 0.01 or 0.05 mol %) were incubated at pH 9.0 and room temperature,a nd reactions were stopped after 24 h, which resulted in excellent conversions of > 89 %( Ta ble 2). The products were purified (57-77 % yield) and identified as the corresponding 3-benzyloxyaspartate derivatives 1b-f, 1h,a nd 1i.A mino acid product 1i was furtheri dentified as the desired threo isomer of 3-(3-methyl)benzyloxyaspartate( de > 95 %) by comparison of its 1 HNMR signals to those of chemically synthesized authentic standards with known threo or erythro configuration.A lthough the relative configurationo fp roducts 1b-f and 1h was not determined by comparison to authentic standards, we assume the relative configuration to be threo for all products on the basis of analogy (see the Supporting Information). [10] Analysis of products 1c, 1f,a nd 1i by HPLC on ac hiral stationary phase by using chemically prepared authentic standards with known dl-threo configurationr evealedt hat the L384G-catalyzed amination reactions were highly enantioselective andt hat these amino acids were produced with > 99 % ee (Table 2; Figures S5-S7). Controlexperiments showedt hat the 2-benzyloxyfumarate derivatives did not undergo amination in the absence of the enzyme. These results show that the L384G mutant has striking potential for application in the regio-and stereoselective synthesis of derivatives of TBOA.
In conclusion, we demonstrated the chemoenzymatic synthesis of l-threo-3-benzyloxyaspartate (l-TBOA)a nd variousd erivatives of TBOA with substituents on the aromatic ring by asymmetrica mination of 2-benzyloxyfumarate derivatives by using engineered methylaspartate ammonia lyase variants. This chemoenzymatic approach appears to be an elegant alternative for existing chemical methods. [4a-d] This was convincingly demonstrated by the 3-step chemoenzymatic synthesis of l-TBOA,t he selective chemical preparation of which is ah ighly challenging 11-step procedure starting from (R)-Garner aldehyde. [4c] Appropriate experiments to investigate the inhibitory properties of the newly synthesized TBOA derivatives against aspartate/glutamate transporters are underway. 2j -0.05 0000--- [a] Yield of isolated product after ion-exchangechromatography.
[b] The diastereomeric excess (de)isdefined as the excess of threo isomer over the erythro isomer.
[ c] The purified aminoa cid product had the threo configuration, as determined by comparison of its 1 HNMR signalst ot hose of chemically synthesizeda uthentic standardsw ith known threo or erythro configuration.
[d] The purifieda mino acid product was tentatively assigned the threo configuration on the basis of analogy (see the SupportingI nformation).
[e] The enantiomeric excess of the isolated productw as determined by HPLC on ac hiral stationary phase by usingachemically synthesized authentic standard with known dl-threo configuration.