Enantioselective Synthesis of α‐Thiocarboxylic Acids by Nitrilase Biocatalysed Dynamic Kinetic Resolution of α‐Thionitriles

Abstract The enantioselective synthesis of α‐thiocarboxylic acids by biocatalytic dynamic kinetic resolution (DKR) of nitrile precursors exploiting nitrilase enzymes is described. A panel of 35 nitrilase biocatalysts were screened and enzymes Nit27 and Nit34 were found to catalyse the DKR of racemic α‐thionitriles under mild conditions, affording the corresponding carboxylic acids with high conversions and good‐to‐excellent ee. The ammonia produced in situ during the biocatalytic transformation favours the racemization of the nitrile enantiomers and, in turn, the DKR without the need of any external additive base.

Abstract: The enantioselective synthesis of a-thiocarboxylic acids by biocatalytic dynamick inetic resolution (DKR) of nitrile precursors exploitingn itrilase enzymes is described.Ap anel of 35 nitrilase biocatalysts were screened and enzymes Nit27 and Nit34 were found to catalyse the DKR of racemic a-thionitriles under mild conditions, affording the corresponding carboxylic acids with high conversionsa nd good-to-excellent ee. The ammonia produced in situ during the biocatalytic transformationf avours the racemizationo ft he nitrile enantiomersa nd, in turn, the DKR withoutt he need of any external additive base.
Enantiomerically pure a-mercapto(thio)carboxylic acids are ubiquitous structural motifs found in pharmaceutical ingredients, [1] peptides, [2] metal chelators [3] and S,O-ligands [4] or used as precursors in the asymmetrics ynthesis of drugs like tiopronin 2 and the IMP1 inhibitor 3 ( Figure 1). [5] Mercapto/thio-carboxylic acids also belongt ot he class of volatile sulphur compounds (VSC) andf ound application as flavouring and aroma agents in food and fragrance industries. [6] Compounds such as 4 and 5 have ac haracteristicm eaty-cheese aroma or fruity/ tropical flavour, respectively and these organoleptic properties are related to the configuration of their CÀSs tereocentre. Despite the key importanceo fc hiral a-thiocarboxylic acids, the availablea nd efficient protocols for their asymmetric synthesis are limited. In fact, the construction of molecules bearing a stereodefined CÀSc entre poses ad istinct challenge in organic chemistry mainly because sulphur compounds have different reactivity which can be incompatible with the type of asymmetric chemistry exhibited by their O-or N-derivativea na-logues. The most straightforwardm ethods to access enantiomericallyp ure a-thiocarboxylic acids rely on stereospecific substitutionr eactions of enantioenriched halo-, [7] hydroxy-or aminoa cid [8] derivatives with sulphur nucleophiles. Enzymatic and non-enzymatic dynamic kinetic resolution (DKR)a lso represents an efficient approach to access enantioenriched a-thiocarboxylic acids from esters, due to the ability of aC ÀSs tereocentrei na lpha to an electronw ithdrawing group to racemise unders pecific conditions. Examples of DKR of thioesters using hydrolase enzymes [9a-c] or thiocarboxylic acids using Nocardia diaphanozonaria cells [9d] have been reported by Drueckhammer and Otha, respectively,w hile Birman described an on-enzymat-   ic DKR of a-thiocarboxylic acids using homo-benzotetramisole (S)-HBTM. [10] However,d espite the excellent conversion and selectivity,a ll these methods show some limitations in terms of green chemistry,s uch as the need of additives and bases (iPr 2 NEt or trioctylamine) to promote the dynamic racemization of the substrates, the use of ester protecting groups with consequent production of waste in the chemical process and the need of oxygen-free conditions to avoid the formation of side products.
Due to our interest in the development of green and sustainable biocatalytic methodologiest oa ccess VSC and druglike compounds, [11] herein we describe aD KR approacht os ynthesise enantiomerically pure a-thiocarboxylic acids from racemic a-thionitriles exploiting nitrilase enzymes. Nitrilases [12] catalyse the hydrolysis of nitriles into the corresponding carboxylic acids with the formation of ammonia as side product. We hypothesised that the NH 3 produced in situ during the course of the biotransformation would be sufficient to promote the racemizationo ft he thionitrile substrates and their DKR, as shown in Figure 1, without the need of additional externalb ases to the reaction mixture or the use of protecting groups.
In addition, the enzymatich ydrolysis would allow the efficient synthesis of a-thiocarboxylic acids under mild conditions. In fact, even if it is widely documented that nitrilesc an be hydrolysed into carboxylic acids under acidic or basic conditions, [13] we found that a-thionitriles decompose into volatile by-products in the presence of strong acids/bases, [14] thus supporting the need for am ild methodology to access a-thiocarboxylic acids.
The racemic thionitrile 8a,s ynthesised from the bromonitrile 7 under microwave irradiation on water (Scheme 1), was chosen as am odel substrate to identify the most appropriate hydrolyticb iocatalysts from ap ool of 35 nitrilases, identified and isolated through am etagenomic approachf rom Prozomix's library.Ac olourimetrica ssay, [12d, 15] able to detect the NH 3 formed during the nitrile hydrolysis reaction by exploiting ophthalaldehyde (OPA), was carried out, leading to the identification of eighth it nitrilases (blue/dark wells in Scheme 1). The conversion and enantiomeric excess( ee)f or each of the 8h it enzymesw ere then analysed by chiral HPLC and the results are reported in Table 1.
Surprisingly,s everalo ft he nitrilases which hadb een tipped for activity based on the assay plate demonstrated no conversion of 8a into the carboxylic acid 9a (Table 1, entries 2,3 and 7-8). Thism ay be due to the sensitivity of the assay which is able to detect ammonia andl ead to ac olour change even in the presence of little enzymatic activity (< 1% conversion). On the other hand, Nit27 afforded 9a with 36 %c onversion and an excellent 95 %ee( entry 6), while the startingn itrile was recovered with 22 % ee. Higher conversion (89 %) waso bserved with Nit20 (entry 5) but both the nitrile 8a and the acid 9a were recovered with negligible selectivity.
The absolute configurationo ft he acid was assigneda s(S)-9a by comparison of its alpha value with that described in the literature. [16] Duetothe promisingly high ee,Nit27 was selected for further optimization of the reaction conditions. Thee ffects of temperature and pH on the biocatalytic transformation were then investigated (Table 2). When the biotransformation Scheme1.Synthesis of a-thionitriles and colorimetric screening assay.
[ b] Intensity of the colour is described on the basiso fp ersonal perception as showni nS cheme 1.  was carried out for al ongert ime (24 h) an increase in the conversion and the ee was observed. Similard ata were obtained at 25 8C, 30 8Cor378C(entries 2-4), while adrop in the conversion was observed at 4 8C( entry 1), likely due to al oss in enzyme activity at this temperature. On the other hand, the pH of the buffer solution had am ore incisive effect on the biocatalytic hydrolysis.I nf act, at pH 8t he acid (S)-9 a was obtained with 68 %c onversion (entry 8), clearly indicatingt hat a DKR reaction was occurring. The rate of substrate racemization of 8a is faster at higherp Hdue to ah igher proportion of the generated ammonia being present in the non-protonated form rather than as an ammonium ion. The increased nucleophilicity of the non-protonated NH 3 may lead to ag reater proton exchange for the a-thionitrile substrate favouring, in turn, the formation of (S)-9 a with conv. > 50 %, as shown in Scheme 2a.
When the reaction was carried out at ah igher pH (pH 8.5), a decrease in the conversion was initially observed, likely due to ar educed activity of the enzyme at this pH. However,w hen the reaction was left stirring foralongert ime (7 days, Table 2, entry 9), excellent (75 %) conversion and 94 % ee of (S)-9 a were obtained. Interestingly,u nder these conditions, the unreacted nitrile 8a was recovered with 60 % ee,w hile, when the reaction was left for 13 days, 8a was obtainedw ith only 8% ee (Table 2, entry 10). It is plausible that the maximum conversion of the biotransformation at pH 8.5 is reached in 7days due to the loss of enzyme activity after this time in the presence of ammonia.D ue to the basic environment, the unreacted nitrile 8a racemises affording, after 13 days, 8a with 8% ee,t husi ndirectly confirmingt he proposed mechanism for the DKR reaction. Noteworthy,n or acemization of the acid 9a occursa fter 13 days. Finally, 8a was treated atp H7.2 for 7days to investigate if longer reactiontimes at lower pH could also lead to improvedc onversions and ee. Under these conditions, an excellent conversion of 90 %w as observed but the ee of 9a slightly dropped down to 81 %( entry 11). [17] It is likelyt hat after some hours at pH 7.2 the Nit27 beginst oa lso hydrolyse the (R)-8 a enantiomer,t hus resulting in al ower ee of 9a.I ncreasing the loading of the biocatalysts proved detrimental leading to 9a with poor (36 %) ee (Table 2, entry 13). On the other hand, the lower conversions observed at pH 8.5 after 7days must be ascribable to the reduced activity of the enzymeu nder these conditions, which counterbalances the racemization process. Even if at pH 7.2 the amount of non-protonated NH 3 is low,i t appearst ob es ufficient to catalyset he racemization of nitrile 8a,i na greement with previous literature data. [9] In order to confirm our hypothesis on the role of ammonia in the racemization of 8a,t wo additional experiments werec arried out (Scheme 2b). The enantioenriched nitrile 8a (ee = 72 %, Table 2, entry 12) was suspended in abuffer solution (pH 7.2) and treated with 1equiv of aqueous ammonia. After 48 h, the full racemizationo f8a was observed. On the contrary,w hen the enantioenriched 8a was left stirring for 48 hi nb uffer solution (pH 7.2) without anyb ase,n or acemization occurred and only as mall decrease in the ee was observed. [18] These experiments show that, even if ammonia at pH 7.2 mainly exists in its protonated form,t he amount of free base is sufficient to favour the racemization of the a-thionitrile substrates, thus supporting the hypothesised mechanism of the biotransformation.
From previous experiments,i tw as clear that while the conversions of the biocatalytic transformation increase at lower pH, the best conv./ee ratio were obtained at higherpH. The scope of the biocatalysed hydrolysis of a-thionitriles 8a-m with nitrilase Nit27 was thus investigated under different pH conditions. The thionitrile substrates 8a-m were synthesised from a-bromonitrile 7 or from the appropriate cyanohydrins 10 a-d as shown in Table 3. The a-thiopropanoic acids 9b-d (Table 3, entries 1-3) wereo btained from the corresponding nitriles with high to excellentc onversions( 50 to 93 %) and excellent ee (89-99 %) regardless of the pH at which the reaction was carriedo ut.T he unreacted nitriles are generally recovered with good ee when the reactionw as performed at pH 7.2. [19,17] The thiocarboxylic acids 9e,f were also obtained with excellent ee but lower conversions ( Table 3, entries 4a nd 5), most likely due to presence of bulkier substituents on the S-phenyl ring that maya ffect their interaction with the nitrilase catalytic site, while the allyl nitrile 8g was fully and enantioselectively converted into 9g (ee 99 %, Table 3, entry 6). All the athiopropanoic acids were obtained with absolute configuration S,t hus confirming the selectivity of the nitrilase Nit27 for the enantiomers (S)-8. [16] When the methyls ubstituent on the stereocentre was replaced with bulkier groups in derivatives 8hm,l ow or no conversion into the corresponding acids 9 was observed. The a-thiobutanoic acids 9h,i (Table 3, entries 7a nd 8) were obtained with excellent ee but poor conversions, while no conversion was observedf or derivatives 9j-m bearing bulkier groups. Since the reason for the poor activity of Nit27 on nitriles 8h-m was ascribable to their steric hindrance, a second screening of the nitrilase pool from Prozomix's library was carriedo ut. For the substrate 8h only the nitrilaseN it34 gave ap ositive response in the colorimetric assay,w hile 4 nitrilases( Nit2, Nit6, Nit20 and Nit34) were identified to be active for the phenylacetonitrile derivative 8j (Table 4). As expected, Nit27 was not highlighted in the colorimetric assay, which was consistent with the lack of conversion observed in the initial reactions cope experiments. Nit20 and Nit34 fully converted the phenylacetonitrile 8j into the acid 9j although with poor ee (22 %a nd 5% respectively,T able 4, entry 3), while irrelevant conversion was observed with Nit02 and Nit06. While both Nit20 and 34 are active upon the a-phenyl-substituted thionitrile 8j,t he lack of enantioselectivity indicates that these nitrilases are less able to differentiate the two enantiomers. Surprisingly,e ven if the nPr and the allyl groups are similar in terms of size, the derivative 8p was not convertedb y any of the four nitrilases (Table 4, entry 4). It is plausible that the shorter length of the double bond together with the different electronic properties of the allyl and nPr groups may affect the binding of 8j and 8p to the catalytic site of Nit20 and thus account for the different conversionso bserved. On the other  hand, nitrilase Nit34 proved to be effective on derivatives 8h and 8k leading to the corresponding acids 9 with good ee and conversions > 50 %, clearly indicating aD KR of the substrate. Compounds 9i and 9l-o were obtained with good-excellent ee (up to 93 %) but with poor conversionsa nd yields most likely due to steric factors which may preventt he substrates 8 to interact with the enzymec atalytic site. [20] In conclusion, nitrilase enzymes provedt ob eefficient biocatalysts for the enantioselective synthesis of a-thiocarboxylic acids 9 through base-free dynamic kinetic resolution of the correspondingr acemic nitriles. Within this work two biocatalysts, namely Nit27 and Nit34, have been identified from a panel of 35 nitrilases and have shown the ability to catalyse the enantioselective synthesis of a-thiocarboxylic acids 9 with good-to-excellent conversion and ee by DKR reactionu nder mild conditions.
The main advantageo ft he methodology relies in the exploitation of the ammonia formed during the biocatalytic hydrolysis to catalytically facilitatet he in situ racemizationo fn itrile 8 (R)-enantiomers, avoiding the need of external bases,w hile the nitrilases electively hydrolyses the (S)-enantiomers. To the best of our knowledge, this is the first example of base-free DKR of a-thionitriles by nitrilasebiocatalysed hydrolysis andrepresents am ild, greener and straightforward means to access enantiomericallyp ure a-thiocarboxylic acids, such as the flavouring compounds 4 and 5.