One‐Step Biocatalytic Synthesis of Sustainable Surfactants by Selective Amide Bond Formation

Abstract N‐alkanoyl‐N‐methylglucamides (MEGAs) are non‐toxic surfactants widely used as commercial ingredients, but more sustainable syntheses towards these compounds are highly desirable. Here, we present a biocatalytic route towards MEGAs and analogues using a truncated carboxylic acid reductase construct tailored for amide bond formation (CARmm‐A). CARmm‐A is capable of selective amide bond formation without the competing esterification reaction observed in lipase catalysed reactions. A kinase was implemented to regenerate ATP from polyphosphate and by thorough reaction optimisation using design of experiments, the amine concentration needed for amidation was significantly reduced. The wide substrate scope of CARmm‐A was exemplified by the synthesis of 24 commercially relevant amides, including selected examples on a preparative scale. This work establishes acyl‐phosphate mediated chemistry as a highly selective strategy for biocatalytic amide bond formation in the presence of multiple competing alcohol functionalities.


Methods and materials Materials and instrumentation
All chemicals and buffers were bought from Sigma Aldrich, Fluorochem or Fischer Scientific. Medium for cell growth was bought from Formedium. All materials relating to molecular biology work were purchased from New England Biolabs (NEB). All NMR spectra were recorded using a Bruker Avance 400 instrument.
HPLC analyses were performed using an Agilent 1260 Infinity II system. LC/MS analyses were performed using an Agilent 1200 series LC system equipped with a G1379A degasser, a G1312A binary pump, a G1329 autosampler unit, a G1316A temperature-controlled column compartment and a G1315B diode array detector. Compounds were ionized using API-electrospray technique and detected in positive mode on the LCMS System. Drying gas temperature 250 °C at 12 L min-1, and nebulizer pressure at 25 psig.
On both LC/MS and HPLC systems an ACE5 C18 column was used (Dimensions: 250 x 4.6 mm).
For HRMS analyses an Agilent 1200 series LC system was used, coupled to an Agilent 6520 QTOF mass spectrometer, ESI positive mode. The data was analysed using Agilent MassHunter software.

Protein expression and purification
CARmm-A and CHU genes, plasmids and expression strains (E. coli BL21 (DE3)) were prepared using previously described methods. [1] For protein expression, autoclaved baffled flasks containing 700 ml auto-induction medium containing the appropriate antibiotic, were inoculated with E. coli BL21 (DE3) cells and were grown at 30°C for 72 hours. Cells were harvested by centrifugation and the cell pellet was stored in zip-lock bags at -80°C.
To lyse the cells for purification, the cell pellet was resuspended in Equilibration buffer (50 mM Tris.HCl pH 8, 200 mM NaCl). The cells were then sonicated 20s on/20s off 25 times. The lysis mixture was subsequently centrifuged, the supernatant collected and the pellet discarded. The supernatant was then mixed with Ni-NTA agarose and left shaking at 4°C for 30 minutes. This mixture was then poured into a gravity column and was washed with Wash Buffer (10 mM imidazole, 50 mM Tris.HCl pH 8, 200 mM NaCl). The protein was then eluted using Elution Buffer (200 mM imidazole, 50 mM Tris.HCl pH 8, 200 mM NaCl). The eluted protein was concentrated using Vivaspin centrifugal concentrators (30.000 MWCO, Sartorius) and then desalted using PD-10 columns (GE Healthcare) following the respective protocols. Purity was checked using SDS-PAGE (staining with Instant Blue (Expedeon)) and concentration was determined by measuring absorbance at 280 nm using Nanodrop (Thermo Fisher).
For the production of cell-free lysates, the frozen cell pellets were resuspended in reaction buffer (100 mM HEPBS pH 8.5), then sonicated and subsequently centrifuged as described above. The supernatant protein concentration was measured, aliquoted and stored at -20°C.

Biotransformation procedure
In an example CARmm-A biotransformation, 5 mM of the carboxylic acid substrate (from a 0.5M stock in DMSO), 50 mM of amine (from a 0.25 M stock in buffer, adjusted to pH 8.5), 17.1 mM ATP (from a 0.1 M stock in buffer, adjusted to pH 8.5), 66.5 mM MgCl2 and CARmm-A (1 mg/mL) were added to HEPBS buffer (100 mM, pH 8.5) to a total volume of 0.5 mL in a 1.5 mL Eppendorf tube. The reaction was placed in a 37°C incubator for 16 hours shaking at 250 rpm.
The reaction was stopped by adding an equal volume of MeOH and shaking the mixture. This mixture was centrifuged and the supernatant was filtered and added to an HPLC vial for analysis.

Acylation of 13 C-decanoic acid
To investigate the reaction selectivity of amidation versus esterification, 13 C-labelled decanoic acid was reacted with amino sugar 1 ( Figure S1). The crude biotransformation was mixed 1:1 with MeOD and analysed using 13 C-NMR. For a negative control, a reaction without adding the enzyme catalyst was included, as well as a commercial standard of MEGA-10. The carbonyl regions of the 13 C-NMR spectra of these samples are shown in Figure S2. As the carbonyl region of the 13 C-NMR spectrum does not show any peaks other than the acid and amide peaks, it suggests the reaction is selective towards a single product without any unwanted ester byproducts.

F-NMR
To investigate the activity of amino sugars, initial CARmm-A catalysed biotransformations using 3fluoro cinnamic acid and amino sugar 1 were performed and analysed by 19 F-NMR (using previously described methods [1] , figure S3).  Figure S4 shows the crude biotransformation (top) and the same biotransformation when spiked with the 3-fluoro cinnamic acid substrate (bottom). This indicated that the substrate in the biotransformation has been completely converted to product. To identify the reaction product in this reaction, it was additionally analysed by LC-MS ( Figure S5  We also investigated wether sorbitol (a poly-alchohol derivative of glucose) would lead to ester formation when used as a nucleophile using the optimized reaction conditions and 3-fluoro cinnamic acid ( Figure S6). We observed that the 19 F-NMR spectra for the sorbitol experiment was identical to the experiment that contained no nucleophile. A very small new peak appeared in these experiments which is the small amount of acyl adenylate that is present in solution. This peak did not show in the no enzyme control experiment as expected. Therefore we concluded that no ester formation occurs when using sorbitol as a nucleophile. Furthermore, a positive control reaction between 3-fluoro cinnamic acid and 1 was performed showing full conversion to the amide product.

Optimisation
Initial reaction test for reacting 7 with 1 resulting in MEGA-8 (11) was performed using previously reported conditions using an excess of amine and ATP. [1] The conversion was determined by RP-HPLC at a wavelength of 210 nm, using a commercial standard as a reference for a calibration curve ( Figure S7 and S8). The calculated conversion of this reaction was >99%.  For reaction optimisation we used the reaction between 9 and 1 resulting in MEGA-10 (13), using the CHU enzyme to regenerate ATP from AMP and polyphosphate as a model reaction (figure S9). Figure S9: Reaction scheme for the CARmm-A catalysed reaction between 9 and 1 to synthesise 13, using the CHU enzyme to regenerate ATP from AMP and polyphosphate.
Using previously reported reaction conditions we investigated the effect of amine concentration on the conversion of substrates to 13 ( Figure S10). We performed design of experiments using the software JMP®, (Version 16 Pro. SAS Institute Inc., Cary, NC, 1989-2022 to optimize and better understand the CHU system. We constructed an empirical model for the effect of polyphosphate, AMP and Mg 2+ concentrations on conversion using data from a set of biotransformation conditions generated by the software ( Figure S11). Using the maximize desirability option in the prediction profiler, it was found that the optimum reaction conditions were 17.1 mM AMP, 66.5 mM MgCl2, and 14.9 mg/ml polyphosphate.

Analysis of analytical scale biotransformations
Reactions shown in Table 1 were stopped after 16 hours by adding methanol in a 1:1 ratio, the mixture was centrifuged and the supernatant was used for reversed-phase HPLC and LC/MS analysis.
To calculate conversions, a calibration curve was made using a dilution series of a commercially bought standard of MEGA-10 (13) using the concentration range 0.625 mM, 1.25 mM, 2.5 mM, 5 mM and 10 mM. These samples were run using the HPLC conditions described above, detecting the amide at 210 nm, taking the mAU value of the peak of the standard at a retention time of approximately 4.8 min.