Expanding the Genetic Code: Incorporation of Functional Secondary Amines via Stop Codon Suppression

Enzymes are attractive catalysts for chemical industries, and their use has become a mature alternative to conventional chemical methods. However, biocatalytic approaches are often restricted to metabolic and less complex reactivities, given the limited amount of functional groups present. This drawback can be addressed by incorporating non-canonical amino acids (ncAAs) harboring new-to-nature chemical groups. Inspired by organocatalysis, we report the design, synthesis and characterization of a panel of ncAAs harboring functional secondary amines and their cellular incorporation into different protein scaffolds. D / L -pyrrolidine- and D / L -piperidine-based ncAAs were successfully site-specifically incorporated into proteins via stop codon suppression methodology. To demonstrate the utility of these ncAAs, the catalytic performance of the obtained artificial enzymes was investigated in a model Michael addition reaction. The incorporation of pyrrolidine-and piperidine-based ncAAs significantly expands the available toolbox for protein engineering and chemical biology applications.


Introduction
To counteract the harmful effects of global warming and pollutants due to industrialization in our society, efficient and sustainable industrial processes are urgently required. [1,2]In the transition toward less-hazardous strategies, several approaches have emerged, including the use of organocatalysis. [3,4]L-proline is one of the first organocatalysts successfully employed in asymmetric synthesis, due to its amine moiety this amino acid can form an iminium/enamine intermediate with a carbonyl moiety. [5]This property has been exploited in the field of organocatalysis, particularly in the alternative asymmetric synthesis of valuable building blocks. [5,6]Following up, other secondary amines such as Hayashi-Jorgensen and MacMillan catalysts have shown remarkable versatility in the synthesis of several drug precursors. [5]7] However, organocatalysts often display low turnover numbers (TON), require further optimization to access high chemo-and stereoselectivity and require the use of organic solvents. [5,6]hus, organocatalysts, while attractive to produce valuable, enantiomerically pure chemicals, do not yet satisfy all requirements of green chemistry. [8]Biocatalysis is a multidisciplinary field that emerged as an academic curiosity and has become a mature alternative to classical and energy-consuming chemical methods. [1,9]Enzymes are attractive catalysts because they perform their reactivities under aqueous, mild conditions with high efficiency.Moreover, enzymes show exquisite control over the chemo-, regio-, and enantioselectivity of their reactions. [10]evertheless, enzymes are generally limited to metabolicderived reactivities, and their reaction scope is constrained compared with the versatility of transformations accessible via more traditional chemical synthesis.To overcome this limitation, protein engineering campaigns toward new-to-nature reactions have recently emerged.For example, iminium biocatalysis has been performed through catalytic proteinogenic N-terminal secondary amine (proline), [11] primary amine (lysine), [12] artificial cofactor (biotin/pyrrolidine conjugate) [13] and non-canonical amino acids (ncAAs) (e. g. paminophenylalanine). [14,15] The latter strategy has been addressed through genetic code expansion.[18][19] Thus, this methodology is of great interest both for academia and industrial biotechnology, due to the possibility of introducing new catalytic residues in proteins.[19] The incorporation of ncAAs enables the manipulation of proteins with exquisite precision at the singleatom level. [16]This feature has been exploited by chemists and biologists in numerous studies of protein structure and function. [17,19]These ncAAs are used for mechanistic studies, isotopic labels for infrared and NMR studies, photocrosslinkers, heavy atoms for X-ray crystallography, spin labels and fluorescent labelling.However, the design and use of catalytic ncAAs with functional side chains is scarce.To date, few examples have been described, p-aminophenylalanine being one of the most exploited, due to its primary amine moiety. [15,16,20]Another relevant example is Nδ-methylhistidine, where an alkyl group fixes the tautomeric form of the imidazole ring. [21](2,2'bipyridin-5-yl)alanine is a dissimilar example, exploiting its metal-binding affinity can be used for the design of artificial metalloenzymes. [16,19]Although extremely exciting (bio)catalysis can be performed using functional ncAAs, the field is constrained by the limited availability of functional ncAAs.Inspired by the restricted accessibility of catalytic ncAAs and the benefits of organocatalysis, we aimed to incorporate new-to-nature functional secondary amines into biomolecular scaffolds to expand the toolbox of protein building blocks for their use in biocatalysis.[24] To date, attempts have been made to tackle this limitation, however, an efficient biological system for incorporating secondary amines into proteins has not been fully investigated and characterized.The first known study was aimed at obtaining structural analogs of pyrrolysine, [25] and in a more recent study using a similar chemical pathway, ncAAs having secondary amines and thiazolidine functional groups were introduced in a biomolecular scaffold. [26]Herein, we expand the panel of ncAAs harboring new-to-nature functional secondary amines, along with designing a new and more versatile synthetic route toward their production (Figure 1).Finally, the catalytic potential of prepared ncAAs was demonstrated for abiological iminium reactivity using the lactococcal multidrug resistance regulator (LmrR) as a biomolecular scaffold. [27]

Results and Discussion
For this work, we designed and synthesized ncAAs with a secondary-amine core inspired by well-known organocatalysts (Figure 1a).Based on L-proline and Hayashi-Jørgensen catalysts, 1 containing a proline core was designed (Figure 1b).Additionally, to investigate the role of the secondary amine ring size on the catalytic performances, piperidine-based ncAA was also studied (2).To evaluate the influence of the secondary amine contiguous stereocenter, for both proline-and piperidine-based ncAAs, the respective stereoisomers were synthesized (D and L).
To increase the chances of in vivo incorporation, the cyclic secondary amines were connected to the ɛ-nitrogen of the lysine, mimicking a pyrrolysine structural analog (Figure 1a). [17,18]urthermore, two MacMillan-like ncAAs were designed with an imidazolidinone core possessing a 2-carbon longer side chain.The linker length was shortened, originating ncAAs 3 and 4, characterized by C1 and C3 linkers, respectively.
Remarkably, all the ncAAs could be obtained by exploiting a similar retrosynthetic strategy (Scheme 1).The synthetic route involved the coupling of the carboxylic acid on the cyclic secondary amine-side chain with an L-amino acid with the appropriate side chain length (L-lysine for ncAA 1 and 2, Ldiaminopropionic acid for 3 and L-ornithine for 4), followed by the cleavage of the protecting groups.In detail, the synthesis of 1 and 2 started with the HATU-mediated coupling of Z-L-LysÀ OMe • HCl with Z-L-proline, Z-D-proline, Z-L-piperidine carboxylic acid and Z-D-piperidine carboxylic acid (Scheme 1a).The desired intermediates L-5, D-5, L-6 and D-6 were accessed with good yields (> 90 %).The synthetic route continued with the basic hydrolysis of the methyl ester, in the presence of LiOH.This step worked well for the two proline derivatives, producing L-7 and D-7 with decent yields after purification via flash chromatography (82-84 %).Under the same reaction conditions, the hydrolysis of L-6 and D-6 generated the desired products and an unknown impurity in an almost 1 : 1 ratio.As a consequence, L-8 and D-8 were isolated by reverse-phase chromatography with a lower yield (30-62 %).Then, Pdmediated hydrogenation allowed the cleavage of the Cbzprotecting groups to produce the ncAAs L-1, D-1, L-2 and D-2.For L-2 and D-2, the protocol was optimized by exchanging the order of the cleavage of the protecting groups (Scheme S3).In this way, L-6 and D-6 underwent the Pd-mediated hydrogenation and then the basic hydrolysis, which in both cases afforded the desired piperidine-based ncAAs with quantitative yield, without observing the formation of the previous impurities.
Compounds 3 and 4 were synthesized following a similar strategy, involving the coupling of the key carboxylic acid 13 with Z-L-DapÀ OMe • HCl (11) or Z-L-OrnÀ OMe • HCl ( 12), followed by hydrolysis of the methyl ester and hydrogenation of the Cbz units (Scheme 1b).Compound 13 was not commercially available and was synthesized starting from FmocÀ Glu(tBu)À OH (Scheme S3).Compound 13 was coupled with either Z-L-DapÀ OMe • HCl (11) or Z-L-OrnÀ OMe • HCl (12), previously produced from the corresponding carboxylic acids upon treatment with thionyl chloride in methanol.The hydrolysis of the accessed intermediates 14 and 15 with LiOH generated compounds 16 and 17 with moderate yields (30-60 % over the two steps).The ncAAs 3 and 4 were finally obtained upon hydrogenation.Upon full characterization, the obtained panel of ncAAs was used for further biochemical experiments.
For correct incorporation of newly synthesized ncAAs into proteins, a suitable aminoacyl tRNA synthetase-tRNA pair must be identified. [17,18]Due to the similarity of the synthesized ncAAs to pyrrolysine, we screened a panel of > 50 pyrrolysine tRNAsynthetase (PylRS) variants from different origins (Methanosarcina barkeri, Methanosarcina mazei and Methanomethylophilus alvus), containing between 0-7 mutations in their active site.The analysis aimed to site-specifically incorporate ncAAs 1-4 into C-terminally His-tagged superfolder green fluorescent protein (sfGFP) containing a premature amber codon (sfGFPÀ N150TAGÀ His x6 (Figure 2a).Gratifyingly, wild-type MbPylRS (Methanosarcina barkeri) directed incorporation of four of the synthesized ncAAs (L-1, D-1, L-2 and D-2).No incorporation was observed for 3 and 4 with any of the studied pairs (Figure 2b).To observe the incorporation, ncAA supplementation was performed at 5 and 15 mM.Wild-type MbPylRS showed for L-1, D-1, L-2 and D-2, at 15 mM ncAA a higher fluorescence intensity than at 5 mM.Additionally, fluorometric features remained unchanged.When supplemented with BocÀ lysine (BocK, positive control for MbPylRS system), the analysis suggests that there is a saturation of the expression from at least 1 mM.Final confirmation was performed with 50 mL cultures supplemented with 10 mM ncAA, and using wild-type sfGFP and BocK-incorporating variants as controls.The variants were purified via immobilized metal affinity chromatography and analyzed by SDS-PAGE and LC-MS.For the electrophoretic assay, the expected masses (ca.27.8 kDa) were found for all variants, (Figure 2c).In contrast, cultures induced but not supplemented did not show noticeable protein expression, suggesting that there was no misincorporation using this system (Figure S1  Once confirmed that the wild-type MbPylRS pair enables successful cellular incorporation of ncAAs 1 and 2, we decided to showcase their functional potential in preparation of an artificial enzyme.LmrR was chosen as a model biomolecular scaffold for the design.LmrR has been previously reported as an attractive scaffold for the incorporation of a range of ncAAs, primarily due to its large hydrophobic pocket at the homodimer interface. [27]The ncAA incorporation into position V15 has previously exhibited relatively high catalytic performance, due to the orientation of the residue towards the pocket and its close proximity to W96 and W96' residues. [20,28]Therefore, we evaluated the ncAA incorporation at this position (Figure 3a).Similar to sfGFP N150TAG, the cellular incorporation of the panel was evaluated and confirmed via SDS-PAGE (Figure 3b and S5a) and LC-MS (Figure 3c-d, S6 and S7).Furthermore, after purification, the thermostability of all the variants was assessed by measuring thermal unfolding towards a temperature ramp.All samples maintained similar apparent melting temperatures compared with the wild-type (ca.Δ 2 °C) (Figure S5b).The narrow differences in the melting temperature profiles suggest that the incorporation of the ncAAs does not have a detrimental effect on the global thermostability of the variants or on their binding network.
The catalytic properties of the obtained LmrR variants were evaluated on a model iminium reaction: Michael addition of nitromethane to cinnamaldehyde (Table 1, Figure S8-S9).The small-scale reactions were performed using 25 μM of dimeric LmrR variants as catalysts in the presence of 50 mM nitromethane and 1 mM cinnamaldehyde.After 25 h, the uncatalyzed reaction did not form a noticeable product (< 5 %), while in the presence of LmrR wild-type (V15), a 7 � 2 % conversion was observed.The reaction formulated with LmrR and an equimolar amount of pyrrolidine or piperidine exhibited similar conversion levels.This suggested that there was no entropic effect.In case of LmrR variant incorporating D-1 at position 15, the conversion of 31 � 2 % was observed, while for L-1-variant the conversion was lower (15 � 2 %).Furthermore, D-2 and L-2incorporating variants exhibited similar conversion levels to those of the wild-type (8-11 %).For protein scaffolds, a constrained binding pocket can hinder the activity.Given the slight size difference of piperidine in comparison to the pyrrolidine ring, it is possible that this may have a negative impact on catalysis, either due to the size of the side chain or its spatial orientation.Regarding enantiospecificity, for the respective controls and scaffolds incorporating 2, a racemic mixture of 4-nitro-3-phenylbutanal was obtained.Remarkably, both scaffolds incorporating D-1 and L-1, resulted in enantioselective catalysis (ee = 23 and 38 % for (S)-product, respectively).The enantioselectivity remained constant over time (Figure S10).
To determine the nature of the catalytic activity, the D-1 variant was incubated in the presence of 2-hydroxycinnamaldehyde to spectrophotometrically detect the iminium-activated species (Figure S11).Not surprisingly, the variant harboring the pyrrolidine moiety showed a higher absorbance than the wildtype variant.Furthermore, after chemically reducing the variant incorporating D-1 in the presence of cinnamaldehyde with NaBH 3 CN, time-of-flight mass spectrometry analysis showed the corresponding covalently bound reduced Schiff-base adduct (Figure S12).These results suggest an iminium ion-based biocatalysis performed by D-1.
Albeit modest, this observed catalytic selectivity of the 1variants is of great relevance because can allow further optimization for the synthesis of valuable enantiopure γ-nitroaldehydes for pharmaceutical precursors. [29]In comparison, the evolution of a class I aldolase for the same Michael asymmetric addition reaction has been described. [12]An engineered enzyme (DERA-MA, 12-fold mutant) capable of catalyzing the same model reaction with high efficiency (TON = 190 in 1 h) was obtained. [12]DERA-MA exhibits an ee = 99 % for (R)-product, and variants able to access (S)-product were not achieved.Furthermore, 4-oxalocrotonate tautomerase (4-OT) has been extensively investigated to perform a similar reaction exploiting an Nterminal proline as a nucleophile to promote enamine catalysis. [30]Additionally, 4-OT has also been evolved, and an 11fold mutant was achieved to promote Michael addition of nitromethane to cinnamaldehyde. [31]The obtained enzyme also only produced (R)-γ-nitroaldehyde product with an ee = 98 %.This encourages further work on the 1-variants for the synthesis of enantiopure (S)-4-nitro-3-phenylbutanal and derivatives.This can be achieved by using powerful screening methods in  campaigns of directed evolution of artificial enzymes, [20] facilitating the production of valuable pharmaceutical building blocks.

Conclusions
In this work, we present a new and versatile route for the synthesis of ncAAs harboring secondary amines, [25] extended to piperidine-and McMillan-based ncAAs.Interestingly, although it had already been suggested that only the D-stereoisomer of 1 is a substrate for PylRS and could be incorporated into proteins, [25] we showed that incorporation of the L-isomer is also possible (Figure 2 and 3).Nevertheless, expression yields were higher when supplementing with D-than with L-ncAAs.
We propose this may be due to their structural (stereo)similarity with pyrrolysine (Figure 1a).][34][35] Results with 3 and 4 could be caused by too great of a structural change to the natural pyrrolysine (linker length, size of the imidazolidinone core).Further evaluation of cellular incorporation of these ncAAs is ongoing.Regarding the catalytic performance of the obtained artificial enzymes, the best LmrR variant (incorporating D-1) exhibited a TON = 12.3 after 25 h (Table 1).Although modest, similar values have been obtained using LmrR as an artificial scaffold harboring p-aminophenylalanine as catalytic residue (ca.2-50). [14,28,36]However, leveraging this robust platform enables to perform protein engineering campaigns targeting improvements in yield, enantioselectivity, and other iminium-based reactivities.
In conclusion, herein we have explored a versatile and modular synthetic route for the design, synthesis and characterization of novel ncAAs harboring secondary amines; and their incorporation into biomolecular scaffolds.The in vivo incorporation of the panel shows the extremely biotechnologically attractive potential of incorporating a range of new-to-nature functional secondary amines into proteins, expanding the available toolbox for protein engineering and sustainable, abiological biocatalysis.

Hit evaluation of the PylRS screen
Plasmids harboring the wild-type PylRS, its dedicated PylT and the His x6 -tagged sfGFP150TAG reporter were cotransformed into chemocompetent E. coli K-12 cells via heat shock.Transformants were grown in 2xYT containing appropriate antibiotics at 37 °C for 8 hours.The culture was diluted to an OD 600 nm of 0.01 into autoinduction medium containing antibiotics and incubated overnight in the absence or presence of 2 mM ncAA.Approximately 5x10 8 cells were harvested by centrifugation (7,000 g, 5 min) and lysed via boiling in 1x SDS loading buffer for 5 min at 95 °C.SDS-PAGE and coomassie staining revealed approximate sfGFP production in comparison to negative control. [37]Leftover cells were pelleted (4,000 g, 10 min), resuspended in lysis buffer (20 mM Tris pH 8.0, 300 mM NaCl, 30 mM imidazole, 1 mM PMSF) and lysed via sonication in an ice-water bath.The crude lysate was centrifuged (14,000 g, 20 min, 4 °C) and the cleared lysate was incubated with Ni Sepharose 6 Fast Flow resin (Cytiva) for 1 hour at 4 °C while rotating.The mixture was collected in a plastic column and washed with 10 column volumes of wash buffer (20 mM Tris pH 8.0, 300 mM NaCl, 30 mM imidazole) and subsequently eluted with wash buffer supplemented with 300 mM imidazole.Correct incorporation of ncAA was confirmed via LC-MS.

Intact mass determination of purified sfGFP from screening
Intact mass analysis was carried out on an Agilent Technologies 1260 Infinity LC-MS system with a 6310 Quadrupole spectrometer using a solvent system that employed 0.1 % formic acid in water as solvent A and 0.1 % formic acid in ACN as solvent B. Proteins were separated on a Phenomenex Jupiter C4 300A LC Column (150×2 mm, 5 μm) using a gradient from 10 % to 55 % B in 1.65 min followed by a gradient from 55 % to 90 % in 0.85 min at a flow rate of 0.9 mL min À 1 .The protein samples were analyzed in positive mode as well as by UV absorbance at 193, 254 and 280 nm .Deconvolution was performed with the Agilent OpenLAB CDS ChemStation LC/MS software using standard parameters.

Molecular biology
Vector D4 harbors the MbPylRS gene and OA12 harbors the gene with amber codon and pylT, both under araBAD promoter.Sitedirected mutagenesis was performed via QuikChange methodology.Primers were designed following the methodology described elsewhere. [38]Primers for LmrR V15TAG site-directed mutagenesis as follows (5'), Fw: GCT CAA ACC AAT GTG ATC CTG CTG AAT GTC and Rv: CAT TGG TTT GAG CAC GCA GCA TTT CTT TC.Primers for sfGFPTAG150N site-directed mutagenesis as follows (5'), Fw: CAG CCA TAA CGT GTA TAT TAC CGC CGA TAA AC and Rv: CAC GTT ATG GCT GTT GAA ATT ATA TTC CAG TTT ATG ACC.Reactions were performed in 25 μL following the recommendation of the manufacturer (Pfu Turbo DNA polymerase, Agilent).The following PCR cycles were used: 95 °C for 2 min; 95 °C for 30 s, 55 °C for 30 s and 72 °C for 7 min, (the thermal cycle was repeated 26 times); final extension at 72 °C for 10 min and hold at 4 °C.Reaction products were incubated with restriction endonuclease DpnI at 37 °C for 16 h.Subsequently, 5 μL were transformed into chemocompetent E.coli NEB10β cells via heat shock.Plasmids were isolated and confirmed by sequencing service (Eurofins Genomics).Cloning of LmrR gene into OA12 vector was performed by Golden Gate assembly by inserting flanking BsaI sites via PCR and continuing with the cloning methodology. [39]

Protein purification for upscaled expression cultures
After determining the suitable orthogonal pair through the library screening, the constructs (OA12 vector harboring sfGFP150TAG, sfGFP wild-type, LmrR V15TAG or LmrR wild-type) were cotransformed into chemocompetent E.coli NEB10β cells via heat shock with the corresponding vector harboring the PylRS gene (D4 for MbPylRS).Single colonies were selected for precultures and then inoculated 1 : 100 for 10 or 50 mL (in lysogeny broth culture, supplemented with kanamycin 50 μg mL À 1 and tetracycline 17.5 μg mL À 1 ) at 37 °C with constant shaking at 135 r.p.m.When the culture reached an optical density at 600 nm of 0.6-0.7, the expression was supplemented with the corresponding ncAAs (at 0, 1, 5, 10 or 15 mM) and induced with L-arabinose (at 0 or 0.05 % w/ w) at 37 °C for 16 h, 24 h and 37 h for sfGFP variants and 28 °C for 16 h for LmrR variants.Controls were performed in biological triplicate.Preliminary screening was performed with instrumental duplicates.Additionally, cultures were supplemented with 5 mM nicotinamide (CobB inhibitor).Cultures were harvested by centrifugation and resuspended in BugBuster Master Mix (EMD Millipore Corp) or 50 mM TrisHCl, 500 mM NaCl, pH 8.0 buffer with 1 mM PMSF.Cell-free extracts (CFE) were obtained after lysing with BugBuster or sonicating (5' on and 5' off, for 5 min at 35 % amplitude, Branson 550; microtip model 102 C (CE)) and then centrifuging the lysate at 14,000 g, 10 min at 4 °C.Purified variants were obtained by immobilized metal affinity chromatography using Ni Sepharose High-Performance resin (Cytiva).The resin was equilibrated with 50 mM Tris HCl pH 8.0 and 500 mM NaCl (buffer C), then CFEs were loaded and the mixture was incubated at 4°for 15 min.The resin was washed with 10 CV of buffer C with 40 mM imidazole and eluted with buffer C with 400 mM imidazole.Elution fractions were loaded on pre-equilibrated Econo-Pac 10DG desalting columns (Bio-Rad).For LmrR variants, the concentration of the proteins was determined by using the calculated extinction coefficient ɛ 280 nm = 19,940 M À 1 cm À 1 .The final sample was flashfrozen with liquid nitrogen and stored at À 70 °C.The purity of each purified batch was confirmed by SDS-PAGE analysis.

LC-MS analysis for upscaled culture expressions
After purification, analysis was performed on an InfinityLab II LC coupled to a single quadrupole MS (LC/MSD XT, Agilent Technologies) using an AdvanceBio RP-mAb Wide Pore Reversed-Phase column (2.1×50 mm, particle size 3.5 μm, Agilent Technologies) and a linear gradient of 0-95 % B in 15 min (solvent A: H 2 O + 0.1 % TFA; solvent B: 80 % isopropanol, 10 % acetonitrile, 10 % H 2 O + 0.1 % TFA; flow rate of 0.3 mL min À 1 ).The protein samples were analyzed in positive mode as well as by UV absorbance at 210 and 280 nm .Deconvolution was performed with the Agilent OpenLAB CDS ChemStation LC/MS software using standard parameters.

Iminium ion formation determination
Detection of iminium ion within the LmrR variant was followed by spectrophotometry incubating 55 μM of protein (dimer) with 2 mM 2-hydroxycinnamaldehyde in 50 mM HEPES 150 mM NaCl buffer, pH 6.5. [40]Subsequently, absorbance was followed at 516 nm , 30 μM DERA-MA was used as a positive control.To covalently label LmrRV15D-1 with cinnamaldehyde, a previously reported method was modified. [12]48 μM of protein (dimer) was incubated with 0.6 mM of substrate in 50 mM HEPES, 150 mM NaCl, pH 6.5 and 5 % v/v ethanol at room temperature overnight (12 μL).Then, NaBH 3 CN was added to reduce the iminium ion (150 mM final concentration).The reaction was incubated for 2 h at room temperature, and the buffer was exchanged using a centrifugal filter (Amicon Ultra, Ultracel 10 KDa cut-off).As a control, a sample with LmrRV15D-1 was identically treated, without adding substrate.The samples were analyzed through time-of-flight mass spectrometry (Agilent 6230 LC/TOF), using as eluents the same solvents as LC-MS analysis for upscaled culture expressions.The protein samples were analyzed in positive mode.Deconvolution was performed using Agilent MassHunter BioConfirm Software.

Thermal stability assay
The apparent melting temperatures were determined by measuring fluorescence response to a temperature ramp.Protein samples were set at 30 μM (dimer) following manufacturer recommendations (Tycho, Nanotemper).The capillary was heated from 35 to 95 °C, increasing the temperature by 3 °C every min, using a Tycho NT.6 instrument (Nanotemper).By measuring total fluorescence intensity (sum brightness at 350 and 330 nm ).The unfolding temperature was determined as the maximum of the derivative of the sigmoidal curve.

Small-scale conversions
For the determination of the catalytic properties of LmrR variants, small-scale conversions of Michael addition of nitromethane into cinnamaldehyde were evaluated.Small-scale conversions were formulated at 200 μL in 2 mL tubes in 50 mM HEPES, 150 mM NaCl and 5 % v/v ethanol, pH 6.5.Reactions were prepared with 25 μM dimer LmrR, 1 mM cinnamaldehyde and 50 mM nitromethane, using ethanol as cosolvent (n = 3).Controls without protein, wildtype LmrR with 25 μM pyrrolidine, piperidine or none and without protein and 25 μM pyrrolidine or piperidine were carried out.Reactions were incubated at 25 °C for 24 h with constant shaking.Then, samples were extracted with one volume of ethyl acetate containing 0.02 % v/v mesitylene as an external standard for 45 s.Then, to remove residual water, anhydrous sulfate magnesium was added to the organic solution.Separation was carried out using a GCMS-QP2010 Ultra instrument (Shimadzu) with a Chiraldex GTÀ A column (0.12 μm×0.25 mm×30 m) and helium as mobile phase.The temperature method was a gradient from 70 to 170 °C at 10 °C min À 1 .Calibration curves of substrate and product were analyzed to measure conversion levels.The racemic product was synthesized according to a (slightly modified) reported literature procedure by using pyrrolidine instead of a chiral organocatalyst. [41]dentification of enantiomers was performed by comparing the reaction catalyzed with DERA-MA, described to form (R)-product with ee = 99 %. [12]

Figure 1 .
Figure 1.a) Structures of well-known secondary amine-based organocatalysts and pyrrolysine.b) Workflow of this work, including the structures of the ncAAs panel synthesized in this work.The scheme was partially created with BioRender.com.