Cloning, protein production and purification
RAD and MBH genes were obtained from IDT and GenScript, respectively (sequences can be found in Supplementary Table 1). All genes were cloned into a vector containing a hexa-histidine-tag and TEV-cleavage site using Golden Gate assembly with a BsaI restriction enzyme54. The plasmid sequence is available at https://doi.org/10.5281/zenodo.15494922. Enzymes for amplification, mutagenesis and cloning were sourced from New England Biolabs.
For initial screening, the plasmids were transformed into Escherichia coli BL21 (DE3) STAR cells using standard procedure (30 min incubation on ice, 30 s heat shock at 42 °C, and regeneration in SOC medium at 37 °C, 1,000 rpm). A 96-well plate containing 500 µl of LB medium with 50 µg ml−1 kanamycin in each well was inoculated with a single colony for each design, including a positive control, a negative control and a sterile control (sequences in Supplementary Table 1). Cultures were grown overnight at 37 °C, 250 rpm. One hundred microlitres of these cultures was used to inoculate 24-well plates containing 3 ml of ZY autoinduction medium. For each sample, two replicates were performed. Cultures were grown at 37 °C, 200 rpm for 6 h, then at 18 °C, 200 rpm for 18 h.
Samples were collected by centrifugation (20 min, 4,800g) and resuspended in lysis buffer (20 mM sodium phosphate (sodium phosphate), 500 mM NaCl, 1% n-octyl-β-D-glucopyranoside, 20 µg ml−1 DNase I, 250 µg ml−1 lysozyme, 1× cOmplete protease inhibitor pill per 100 ml, pH 7.4). Lysis was performed at room temperature for 1 h on a vibrating plate at 1,000 rpm. The lysate was clarified (20 min at 4,800g), and the supernatant was purified in a 96-well plate using magnetic Ni-NTA beads and an Opentrons OT-2 pipetting robot. The lysate was incubated with the magnetic beads for 20 min at room temperature, and the supernatant was discarded. Beads were washed with 200 µl wash buffer (20 mM sodium phosphate, 500 mM NaCl, 2 mM tris(2-carboxyethyl)phosphine (TCEP), pH 7.4) once. The supernatant was discarded, and bead-bound proteins were eluted by adding 100 µl elution buffer (20 mM sodium phosphate, 500 mM NaCl, 250 mM imidazole, 2 mM TCEP, pH 7.4) twice per well, yielding 200 µl eluate for each replicate. Protein concentration was determined via Bradford assay (bovine serum albumin standard curve). Protein purity was confirmed using SDS–PAGE.
Active site lysine-to-alanine variants were generated by site-directed mutagenesis using primers designed with the NEBaseChanger tool. PCR reactions contained 10 µl of 5× NEB Q5 buffer, 32.5 µl ddH2O, 1 µl of dNTP mix (10 mM in ddH2O), 2.5 µl of forward and reverse primers (10 µM in ddH2O), 1 µl of template DNA (2 ng µl−1) and 0.5 µl Q5 DNA polymerase. Reactions were cycled for 25 rounds (98 °C denaturation, primer-specific annealing temperature, 72 °C extension), with initial denaturation (30 s) and final extension (2 min) steps. A complete list of primers and annealing temperatures used can be found in Supplementary Table 9. To 1 µl of the PCR reaction, 1 µl each of T4 DNA ligase buffer, T4 DNA ligase, T4 polynucleotide kinase, and DpnI as well as 5 µl of ddH2O were added and incubated at 22 °C for 1 h. Plasmids were transformed into E. coli TOP10 cells and plated on agar plates (50 µg kanamycin per ml agar). Plasmids from single colonies were isolated using a Monarch Spin Plasmid Miniprep kit and sequence-verified. Screening of lysine-to-alanine variants was performed in the same way as the original designs.
For batch production, 10 ml of TB medium containing 100 mg l−1 kanamycin was inoculated with a single colony of BL21 (DE3) STAR cells containing the respective plasmid. After overnight growth at 37 °C, 140 rpm, 10 ml of the culture was used to inoculate 1 l TB medium (same antibiotic). Cultures were grown to an OD600 of 0.6–0.8 at 37 °C and 140 rpm, and induction was initiated by adding isopropyl β-d-thiogalactopyranoside (IPTG) to a final concentration of 0.1 mM. Cells were collected 4–5 h after induction via centrifugation (20 min, 4,000g). For MBH19-63, cultures were incubated at 20 °C, 140 rpm overnight following induction and collection the next morning. Pellets were washed once with 30 ml of 0.9% NaCl solution at room temperature and stored at −20 °C.
Pellets were thawed and resuspended in lysis buffer (20 mM sodium phosphate, 500 mM NaCl, and a spatula of DNase I and lysozyme per 200 ml, pH 7.4). The suspensions were sonicated for 15 min on ice and the lysate was centrifuged at 43,000g for 40 min. The supernatant was loaded onto gravity columns containing 1–2 ml nickel immobilized metal affinity chromatography (Ni-IMAC) resin equilibrated with lysis buffer and washed with wash buffer (see above). The purified proteins were eluted using an elution buffer (see above). Buffer was exchanged to storage buffer (20 mM sodium phosphate, 300 mM NaCl, 2 mM TCEP, pH 7.4 for RAD designs; 20 mM sodium phosphate, 150 mM NaCl pH 7.4 for MBH designs) using centrifugal filters. For RAD designs, His-tag cleavage was performed by adding 0.062 mg of TEV protease (produced in-house) per mg of protein and incubation at 4 °C overnight. The cleaved tag was removed using reverse Ni-IMAC. MBH His-tags were not removed. The final purification step consisted of gel filtration on a S75 Increase 10/300 GL or S75 10/300 column equilibrated with the respective storage buffer. Protein concentrations were determined by specific absorbance at 280 nm and Bio-Rad assay. Samples were flash-frozen in liquid nitrogen and stored at −80 °C.
Intact mass spectrometry
Five microlitres of protein samples (10 μM) in RAD storage buffer were desalted on a Shim-pack Scepter C4-300 (G) column (3 μm) by washing with 1% methanol in the presence of 0.1% formic acid. Increasing concentrations of acetonitrile (MeCN, 1–95%) with 0.6% formic acid eluted the proteins into an Impact II ESI-Q-TOF (Bruker) mass spectrometer. Protein signatures were integrated and deconvoluted using the DataAnalysis maximum entropy function.
Circular dichroism
CD and thermal denaturation experiments for retro-aldolase designs were performed on a JASCO-1500 CD-spectrophotometer in 10 mM sodium phosphate buffer, pH 7.4, containing 150 mM NaF and 0.5 mM TCEP at approximately 0.25 mg ml−1 protein concentration. Spectra were recorded in 1 mm quartz cuvettes with a cap. Thermal denaturation was performed at 3 °C min−1 while monitoring CD signal intensity at 220 nm. Spectra (190–260 nm) were recorded at 20 °C, 45 °C, 70 °C and 95 °C. Additional spectra were recorded after cooling down to 20 °C. Each spectrum consisted of three accumulations.
Chemical denaturation experiments were performed in 100 mM sodium phosphate buffer pH 7.4 containing 300 mM NaF and 1 mM TCEP at a final protein concentration of 0.6 mg ml−1. GdnHCl from a 7.4 M stock solution (concentration confirmed via refractometry) was added to final concentrations between 0 M and 7.1 M. Protein samples were incubated at room temperature overnight, and the CD signal at 220 nm was recorded in quartz capillaries. Denaturation midpoints were calculated from a sigmoidal fit with the Python SciPy library.
MBH48 CD experiments were performed on an Applied Photophysics Chirascan V100 CD-spectrophotometer in 20 mM sodium phosphate buffer, pH 7.4, containing 150 mM NaF at a protein concentration of 0.1 mg ml−1. Spectra were recorded in a 1 mm quartz cuvette with a cap. Thermal denaturation was performed at 1.5 °C min−1, with full spectra recorded every 1 °C. An additional spectrum was recorded after returning to the starting temperature.
SAXS
SAXS profiles were recorded at the ESRF BM29 BioSAXS beamline55. Samples were prepared in 20 mM sodium phosphate buffer, pH 7.4, containing 150 mM NaCl, 1 mM TCEP and 3% glycerol, at a protein concentration of 2–4 mg ml−1. Blank measurements were performed using buffer from the flowthrough of centrifugal filters for each sample.
Raw data were processed using the ATSAS software package56. Frames were manually inspected, averaged and background subtracted. All parameters listed in Supplementary Table 2 were calculated using autorg/autognom functions in ATSAS. Low-q trimming was guided by autorg recommendations. Scattering profiles were fit with design models or available crystal structures using FoXS57, with offset correction and explicit hydrogens enabled, using q values up to 0.5. C1 and C2 parameters were set to be flexible between 0.99 and 1.05 and −2 and 4, respectively.
X-ray crystallography
Crystallization drops were set up with commercial crystallization screens using the vapour-diffusion method, using a mosquito Xtal3 crystallization robot (SPT Labtech) and incubated at 20 °C. The protein concentration varied between 10–30 mg ml−1 in 20 mM sodium phosphate, 150 mM NaCl, 1 mM TCEP, pH 7.4 (RADs) or 20 mM HEPES, 150 mM NaCl, pH 7.4 (MBHase). The drop volume was 400 nl, with a 1:1 protein to precipitant solution ratio. Crystallization drops were equilibrated against 40 µl of precipitant solution. Crystals of RAD32 and MBH2 were obtained from manually set-up drops of 2 µl with 80 µl of precipitant solution in the reservoir. Depending on the design, crystals appeared after 1–14 days. Crystallization conditions and data collection and refinement statistics are provided in Extended Data Table 1. The obtained crystals were collected with CryoLoops (Hampton Research) and cryo-protected in mother liquor containing 25% glycerol or 25% PEG400, followed by flash freezing in liquid nitrogen. Diffraction data were collected at 100 K on ESRF beamlines (Grenoble, France). Complete datasets (360°) were collected to 2.43 Å (RAD13), 2.9 Å (RAD17), 2.0 Å (RAD32), 1.73 Å (RAD36),1.13 Å (MBH2) and 1.93 Å (MBH48) resolution.
The collected data were processed using XDS58 with the provided input file from the beamline. Data resolution cutoffs were determined by pairef59. Structure determination was performed by molecular replacement using PHASER60 with the design models as search templates. The best solution was refined in reciprocal space with PHENIX61 with 5% of the data used for Rfree and by real-space fitting steps against σA-weighted 2Fo – Fc and Fo – Fc electron density maps using COOT62. For RAD17, feature-enhanced maps63 were generated to facilitate the modelling of sidechains of the active site residues (Supplementary Fig. 16).
Activity measurements
Catalytic activity in the retro-aldol reaction was determined by following the formation of the aldehyde product 2 via measuring fluorescence emission at 452 nm at an excitation wavelength of 330 nm. Reactions were carried out in reaction buffer 1 containing 20 mM sodium phosphate, 300 mM NaCl, and 1 mM TCEP in 5 vol% dimethyl sulfoxide (DMSO), pH 7.4 at a 200 µl reaction volume in a 96-well-plate format at a temperature of 29 °C. Michaelis–Menten kinetics were measured with eight 1:1 serial dilutions starting at a concentration of 1 mM to a final concentration of 0.0078 mM rac-methodol. Parameters kcat and KM were determined by fitting reaction velocity and substrate concentration in a Michaelis–Menten model using the Python library SciPy, v.1.13.1.
pH profiles for RAD designs were determined at 50 µM rac-methodol concentration in 384-well plates at a total sample volume of 100 µl and a temperature of 25 °C. Final enzyme concentrations were set to 2 µM (RAD13), 12.6 µM (RAD17), 0.03 µM (RAD29), 0.06 µM (RAD32), 0.03 µM (RAD35) and 3.6 µM (RAD36). To 10 µl of protein samples in 10 mM sodium phosphate, 300 mM NaCl, pH 7.4 90 µl of reaction buffer 2 at various pH values (100 mM phosphate, 100 mM borate, 100 mM acetate buffer containing 55.56 µM rac-methodol and 5.56 vol% DMSO) were added. The pH of reaction buffer 2 was adjusted to values between 4.6 and 10.6 in 0.4 steps. kcat/KM was determined via monitoring of fluorescence as described above. For each measurement, the background reaction at the respective pH was subtracted. pKa values were obtained according to a fit with a two-pKa model. Detailed information on software and equations can be found in the Supplementary Methods.
The aldol addition reaction was performed using a final enzyme concentration of 5 μM in reaction buffer 3 (20 mM sodium phosphate, 300 mM NaCl, pH 7.4). 5 mM of 2 was added from a stock solution in acetonitrile (MeCN, 15 vol% final concentration) followed by 5 vol% acetone. The reactions (final volume 500 μl) were incubated for 24 h at 30 °C and 120 rpm. Control reactions were run in the absence of enzymes. All reactions were performed in triplicate. The products were extracted using ethyl acetate (2× 250 μl) spiked with 0.5 vol% acetophenone as internal standard. The combined extracts were dried over anhydrous sodium sulfate and analysed by chiral-phase high-performance liquid chromatography (HPLC) as indicated below. The results were corrected for the background reaction under the same conditions. The identity of the compounds was further confirmed by gas chromatography–mass spectrometry and the use of authentic reference material. Substrate consumption was measured using a calibration curve.
For conversion and substrate enantiomeric excess determination in the retro-aldol reaction, a final enzyme concentration of 20 µM in reaction buffer 3 was used to catalyse the reaction with 2 mM rac–1 as the substrate, added from a stock solution in MeCN (15 vol% final concentration). The samples were incubated for 24 h at 30 °C and 120 rpm. All experiments were performed as technical triplicates (for biological replicate measurements see Supplementary Table 4). The reaction was analysed using the same method as the forward aldol reaction, monitoring the consumption of both enantiomers and the formation of the aldehyde product 2.
HPLC analyses were performed on a Shimadzu system (DGU-20A On-line Degasser, LC-20AD pump, SIL-20AC autosampler, CBM-20A system controller, SPD-M20A photodiode array detector, Shimadzu CTO-20AC column oven). The samples (5 µl) were analysed with an isocratic flow according to the following methods:
For elution order and absolute configuration: analysis on a Daicel Chiralpak IB (250 mm, ID 4.6 mm, particle size 5 µm) using n-heptane/2-propanol 90:10 (isocratic, flow rate of 1 ml min−1, 30 °C, wavelength 254 nm); retention times: (S)-methodol 12.6 min, (R)-methodol 13.2 min (ref. 64) (Supplementary Fig. 17a,c).
For enantiomeric excess determination, analysis was performed on a Daicel Chiralcel OD-H (250 mm, ID 4.6 mm, particle size 5 µm) using n-heptane/2-propanol 92:8 (isocratic, flow rate of 1 ml min−1, 30 °C, wavelength 254 nm), as complete baseline separation was achieved on an OD-H column. Retention times: (S)-methodol 18.5 min, (R)-methodol 20.0 min (see Supplementary Fig. 17b,d).
RAD turnover number experiments were performed at 0.1 µM final enzyme concentration and 2 mM final substrate concentration (rac–1) in reaction buffer 3 containing 15 vol% DMSO. Samples were incubated for 48 h at 29 °C. All experiments were performed in triplicate. Product formation was monitored using fluorescence and comparison to a calibration curve. The background reaction of the substrate without enzyme was subtracted.
Conversion screening of MBH designs was performed at an enzyme concentration of 100 µM. Reactions were performed in 96-microwell plates in 20 mM phosphate buffer pH 7.4 with 150 mM NaCl and 10 vol% DMSO at substrate concentrations of 25 mM 3 and 5 mM 4. Samples were taken after 8 h of incubation at 40 °C and 800 rpm by quenching 10 µl of the reaction mixture with 10 µl of MeCN and subsequently used for HPLC analysis.
Michaelis–Menten kinetics for the reaction of 3 with 4 were recorded at 60 µM (MBH48) and 90 µM (MBH18) concentration in 20 mM sodium phosphate, pH 7.4, containing 150 mM NaCl and 10 vol% DMSO at 100 µl reaction volume in polypropylene 96-microwell plates. For reactions versus 3, the concentration of 4 was fixed at 5 mM, whereas that of 3 ranged from 0.5 mM to 32 mM. For reactions versus 4, the concentration of 3 was fixed at 25 mM, whereas that of 4 ranged from 0.1 mM to 6.4 mM. Reaction progress was sampled every 50 min (MBH48) or 2 h (MBH18) by quenching 10 µl of the reaction mixture with 10 µl of MeCN and subsequently analysed by HPLC. All reactions were performed in triplicate. For MBH48, biological replicates were performed as well, with values for KM and kcat within the standard deviations of the first measurement. Initial reaction velocities (V0) at each substrate concentration were determined from linear fits of conversion versus time. Information on software and equations is provided in the Supplementary Methods. Fits for individual measurements versus 3 and 4 are shown in Supplementary Fig. 13.
Conversion and kinetic measurements for MBH designs were analysed using a ThermoScientific UltiMate 3000 HPLC equipped with a Gemini SecurityGuard 4×2.0 cartridge and a Kinetex 5 µm XB-C18 100 Å column (50 × 2.1 mm, Phenomenex) at a flow rate of 1 ml min−1. An isocratic method using 22 vol% MeCN in water at 20 °C was used for all measurements. The 2-(hydroxy(4-nitrophenyl)methyl)cyclohex-2-en-1-one (5) product retention time and concentration were determined via comparison to a calibration curve prepared from a chemically synthesized standard.
Molecular dynamics simulations
Molecular dynamics simulations were performed using GROMACS65 version 2023.4. The GRO coordinates of the design models and the topology files were generated using the pdb2gmx tool with the amber99sb-ildn force field. The molecular dynamics unit cell was defined as a dodecahedron with a minimum distance of 1.0 nm between the complex and the box edges using the gmx editconf tool. The system was solvated with the SPC/E water model using the gmx solvate tool and neutralized by adding Na+ ions using the gmx grompp and gmx genion tools with a salt concentration of 150 mM. The system was then subjected to standard energy minimization and equilibration using the gmx grompp and gmx mdrun tools. Energy minimization was performed using the steepest descent algorithm with a maximum force of 1,000 kJ mol−1 nm−1 and a maximum of 5,000 steps. The equilibration consisted of two phases: NVT (constant number of particles, volume and temperature) and NPT (constant number of particles, pressure and temperature). The NVT phase was run for 100 ps with a temperature coupling of 298 K using the v-rescale thermostat. The NPT phase was run for 200 ps with a pressure coupling of 1 bar using the Parrinello–Rahman barostat. Backbone Cα RMSDs were followed throughout the trajectories to evaluate simulation stability (Supplementary Fig. 18).
Prior to the 20 ns replicate simulations, individual 50 ns simulations were run to confirm proper equilibration. The parameters for both runs were as follows: the temperature and pressure were set at 298 K and 1 bar by the v-rescale thermostat and Parrinello–Rahman barostat, respectively; hydrogen bonds were constrained using the LINCS algorithm; the Verlet cutoff scheme was used to process intra-atomic interactions; the PME method was implemented to account for Coulombic and Lennard–Jones interactions; and a van der Waals cut-off radius of 1.0 was applied. The simulations were analysed using the MDAnalysis Python package version 2.8.0 (ref. 66).
Synthesis
All reagents used for synthesis were obtained from Sigma-Aldrich (except for pyridine, obtained from ThermoFisher Scientific) with min. 98% purity and used without further purification. Solvents were of HPLC grade. NMR spectra were measured on a Bruker Avance III 300 MHz NMR spectrometer. Chemical shifts are reported in ppm relative to TMS (δ = 0.00 ppm) and the coupling constants (J) in Hertz (Hz).
rac-Methodol (rac-1) synthesis
6-Methoxy-2-naphthaldehyde 2 (500 mg, 2.69 mmol, 1 equivalent) was added to a 1:4 mixture of acetone (5.4 ml) and aqueous phosphate solution (22 ml, 10 mM NaH2PO4, 111 mM NaCl, 2.7 mM KCl in water, pH 7.4) (1:4). L-proline (62.2 mg, 0.2 equivalent) was added to the solution, and the reaction was stirred at room temperature for 48 h. Since the reaction had not sufficiently proceeded based on thin-layer chromatography monitoring, more acetone (5.4 ml) and L-proline (62.2 mg) were added after 48 h, and the reaction was kept stirring for another 48 h, after which conversion appeared complete on thin-layer chromatography. Purification by flash chromatography (cyclohexane/ethyl acetate, 2:1) and evaporation of the solvent under reduced pressure yielded the final racemic product methodol (4-hydroxy-4-(6-methoxy-2-naphthalenyl)-2-butanone 1) as a white solid (400 mg, 1.64 mmol) in 61% yield. The spectral data were in accordance with the literature67 (Supplementary Figs. 19 and 20). 1H NMR (300 MHz, CDCl3) δ 7.76–7.68 (m, 3H), 7.42 (dd, J = 8.5, 1.9 Hz, 1H), 7.19– .08 (m, 2H), 5.29 (dd, J = 8.7, 3.7 Hz, 1H), 3.92 (s, 3H), 3.28 (br s, 1H), 3.03–2.80 (m, 2H), 2.21 (s, 3H). 13C NMR (75 MHz, CDCl3) δ 209.28 (s), 157.87 (s), 137.94 (s), 134.24 (s), 129.58 (s), 128.86 (s), 127.33 (s), 124.42 (s), 124.41 (s), 119.18 (s), 105.78 (s), 70.11 (s), 55.44 (s), 52.07 (s), 30.96 (s).
Synthesis of MBH product 2-(hydroxy(4-nitrophenyl)methyl)cyclohex-2-en-1-one (5)
The product standard for the MBH reaction was synthesized as described previously4. The product was purified by flash column chromatography (9:1 pentane:ethyl acetate), yielding white crystals (190 mg, 7.7% yield). The spectral data were in accordance with the literature68 (Supplementary Figs. 21 and 22). 1H NMR (300 MHz, CDCl3) δ 8.18 (d, J = 8.8 Hz, 2H), 7.54 (d, J = 8.2 Hz, 2H), 6.82 (t, J = 4.2 Hz, 1H), 5.60 (d, J = 6.0 Hz, 1H), 3.57 (d, J = 6.0 Hz, 1H), 2.49 – 2.39 (m, 4H), 2.00 (p, J = 6.2 Hz, 2H). 13C NMR (75 MHz, CDCl3) δ 200.26, 149.42, 148.33, 147.35, 140.30, 127.26, 123.66, 72.17, 38.55, 25.91, 22.50.
Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.
