Amination of ω-Functionalized Aliphatic Primary Alcohols by a Biocatalytic Oxidation–Transamination Cascade

Amination of non-activated aliphatic fatty alcohols to the corresponding primary amines was achieved through a five-enzyme cascade reaction by coupling a long-chain alcohol oxidase from Aspergillus fumigatus (LCAO_Af) with a ω-transaminase from Chromobacterium violaceum (ω-TA_Cv). The alcohol was oxidized at the expense of molecular oxygen to yield the corresponding aldehyde, which was subsequently aminated by the PLP-dependent ω-TA to yield the final primary amine product. The overall cascade was optimized with respect to pH, O2 pressure, substrate concentration, decomposition of H2O2 (derived from alcohol oxidation), NADH regeneration, and biocatalyst ratio. The substrate scope of this concept was investigated under optimized conditions by using terminally functionalized C4–C11 fatty primary alcohols bearing halogen, alkyne, amino, hydroxy, thiol, and nitrile groups.

Biocatalytic cascades have emerged as at ime-, resource-, and cost-saving strategy in bioorganic synthesis. [1] The use of several enzymes in ao ne-pot fashion avoids purification/isolationo f (unstable)i ntermediates and the associated unavoidable loss of material. Numerous examples of multienzymatic processes of ever increasing complexity for the production of valuable compounds [2] indicate that the areaso f" systemsb iocatalysis" [3] and pathway engineering [4] are beginning to merge. The synthesis of amines dominates current cascade design because the occurrence of amines is underrepresented in the pool of renewable carbon sources, in contrast to their frequentn eed in chemicalsynthesis. [5] For instance, terminal alkylamino functionalization of alkanes and fatty acid methyl esters was achieved by combining an alkane monooxygenase (AlkBGT) and a w-transaminase in as ingle designed whole-cells ystem. [6] The coupling of a w-transaminases with other enzymes, such as acetohydroxyacid synthase,t ransketolase, varioush ydrolases, and alcohol dehydrogenases enabled the synthesis of (chiral) amine derivatives. [7] The direct transformation of alcohols to amines is only feasible by metal catalysts, [8] no enzymei sk nown for this reaction. However,b iocatalytic two-step oxidation-reductivea mination sequences are known. Oxidationo fa na lcohol by an alcohol dehydrogenase yields the corresponding aldehyde/ketone, which can be reductively aminated by an w-transaminase. The eleganceo ft his redox-neutral process is the internal cofactor recycling, in which NADH generated during alcoholo xidation is employed in the reductivea mination step. This concept has been successfully appliedt oab road range of linear and cyclic aliphatic primaryand secondary alcohols, aryl-alkanols, benzylic alcohols, and a,w-diols for the synthesis of (di)amines. [7] Alcohol oxidases represent an attractive, but underrepresented, alternative to (thermodynamically disfavoured) nicotinamide-dependent alcohol oxidationc atalyzed by alcohold ehydrogenases. These enzymes are commonly flavin-or Cu-dependenta nd use O 2 as an electrona cceptor. [9] Two-electron transfer yields H 2 O 2 as abyproduct, which is destroyedb ycatalase or by the horseradish peroxidase (HRP)/2,2'-azino-bis(3ethylbenzothiazoline-6-sulfonic acid (ABTS) system.T his method is a" green" alternative to traditional protocols, which requiret ransition metals, dimethylsulfoxide (e.g. Swern, Pfitzner-Moffato xidation), [10,11] or nitroxyl radicals (e.g. TEMPO). [12] Recently,wee stablished atwo-step one-pot oxidation-transamination cascade based on Cu I -dependent galactose oxidase (GOase) in combination with a w-transaminase. [13] Dictated by the substrate characteristics of GOasef rom Fusarium NRRL 2903, only electronically activated benzylic and cinnamic alcohols werea ccepted, and this method was not applicable to nonactivated aliphatic (fatty) alcohols. To broaden the substrate scope of this protocol, as earch for as uitable alcohol oxidase revealed ap utative flavin-dependent long-chain alcohol oxidasef rom Aspergillus fumigatus (LCAO_Af) as ap romising candidate. [14] The enzyme shows homology and sequence identity( 30-40 %) with other flavoprotein alcohol oxidases and containst he conserved flavin-bindingd omain (pfam00732) of the glucose-methanol-choline (GMC) oxidasef amily.L CAO showsa ctivity in the H 2 O 2 /HRP/ABTS and supplementary flavin adenine dinucleotide (FAD) enhances the activity of the enzyme( data not shown). Preliminary resultsr eportedt he oxidation of C 6 -C 8 fattya lcohols yielded the corresponding aldehydes. Reductivea mination of aliphatic aldehydes through wtransaminases is well known. [15] To shift the unfavorable equi-librium towards the amine, the well-established l-alanine donor system, employing an alanine dehydrogenase( Ala-DH) combinedw ith an adequate NADH-recycling system, completed the overall cascade (Scheme1). [16] In contrast to Cu I -dependent GOase, where alcohol oxidation selectively stops at the aldehydes tage, undesired over oxidation of aldehydes to carboxylic acids (a common phenomenon for flavin-dependent alcohol oxidases) by LCAO_Af had to be taken into account. [17] The overall performance of the cascade was tuned with respect to the following parameters by using 1-hexanol as substrate (detailed data are given in the Supporting Information): 1) For the removal of H 2 O 2 ,d isproportionation catalyzed by catalase from Micrococcus lysodeiktikus or two-electron transfer mediated by HRP employing ABTS as an electron sink workede quallyw ell. The latter (chromophoric) approachh as advantages in screeningc onditions, whereas the former is more suitable for preparative-scale transformations (Supporting Information, Ta ble S1). 2) For the reductive amination step, variousn icotinamide cofactor recycling systems based on glucose/glucose dehydrogenase (GDH), formate/formate dehydrogenase (FDH), and phosphite/phosphite dehydrogenase (PtDH) werecompared: The phosphite/PtDH systemw as the least efficient, whereas the formate/FDH system led to significant over oxidation of the aldehyde intermediate to yield the undesired carboxylic acid. The bestr esults were obtained with glucose/GDHb yu sing standard conditions, but this result was not further optimized ( Table S2). 3) As ignificant advantage of separateo verexpression of the oxidase and the w-transaminase is the possibilityt oa djust the ratio of both enzymes.T he best results were obtained when LCAO Af and w-TAC v( employed as whole lyophilized cells) were used in ar atio of 2:1( Ta ble S3). 4) The pH profile was investigated within ar ange of pH [5][6][7][8][9][10][11][12], which revealed that the efficiency of the cascade continuouslyi ncreased from pH 6( conversion < 2%) to am aximum at pH 10, followed by as harp drop at pH 12. 5) The efficiency of the cascade was very sensitive towards elevated product concentrations, deduced from the fact that increasing amounts of substrate (10-75 mm)l ed to as teady decrease in conversion and gave comparable levelsofp roduct ( % 18-20mm,F igure S3). 6) Sincet he solubility of the oxidantO 2 in aqueous systemsi s limited, experimentsw ere conducted under atmospheric conditions and at elevated pressure, which revealed an optimal performance at 2-4bar (TableS4). 7) To evaluate the overall performance of the optimized cascade reaction compared to the initial conditions, the aminef ormation was followed over time. Engineering of the reactionp arameters led to ac onsiderable improvement of the overall efficiency with full conversion of 1-hexanol to the corresponding amine within 10 h( Figure S4).
To evaluate the substratet olerance (see Table 1) of the biocatalytic oxidation-transamination cascader eaction, primary alcohols with ac hain length ranging from C 4 -C 11 and derivatives bearing ah alogen,a lkyne,a mino, hydroxy,t hiol, and nitrile group were subjected to the enzymatic amination cascade.
The limitations of the cascade are set by the incompatibility of amino, hydroxy (diol), and thiol substituents (14 a-16 a,e ntries [14][15][16]. Steric factors can be neglected, therefore, the nonacceptance of substrates bearing polar end groups is most likely because of their heavy hydration in aqueous medium. The same is true for secondary aliphatic and activated benzylic or allylic alcohols, which were not converted independent of their E/Z configuration ( Figure S5).
In view of the preparative-scale applicationo ft he oxidation-aminationc ascade, as even-fold upscaleo ft he optimized screening conditions was performed with hydroxy-nitrile 12 a and hydroxy-alkyne 13 a as substrates. Both products were isolated after derivatization to the corresponding ethyl N-carbamates of 12 b and 13 b in 42 %a nd with complete conversion in case of the w-nitrile analogue.
In conclusion, af ive-enzyme cascade for the aminationo f primary alcohols by coupling al ong-chain alcohol oxidase with a w-transaminaseh as been successfully extended to also encompass nonactivated fatty alcohols as substrates. Long-chain alcohol oxidase from A. fumigatus was exploited as an excellent oxidation catalyst for ab road range of w-functionalized aliphatic C 4 -C 11 alcohols, which gave the corresponding primary amines with good-to-excellentc onversionb yu sing w-TA from C. violaceum.

Experimental Section
General procedure for amination of alcohols Lyophilized whole E. coli BL21(DE3) cell preparations containing overexpressed genes of w-TA_Cv [18] (20 mg) and LCAO_Af (40 mg) were both resuspended in sodium phosphate buffer (500 mLe ach, 100 mm,p H10) supplemented with pyridoxal-5'-phosphate (PLP, 2mm), NAD + (2 mm)a nd FAD( 1mm). The samples were shaken at 30 8Ca nd 120 rpm for 30 min in Eppendorf vials (horizontal position) and were combined after rehydration. As olution of l-alanine (100 mm), ammonium chloride (67 mm), and d-glucose (80 mm)i n sodium phosphate buffer (500 mL, 100 mm,p H10.0) was added. Alanine dehydrogenase (Ala-DH) from B. subtilis (0.013 U), glucose dehydrogenase (GDH from DSM, 2U), and catalase (1700 U) from M. lysodeikticus were added. Finally,t he substrate (10 mm)w as added and the reaction mixture was placed into an oxygen pressure chamber. [19] The apparatus was primed with oxygen (technical grade) for about 1min and pressurized to 2bar.T he reaction mixture was shaken at RT and 170 rpm for 24 h. The conversion was determined by GC-MS analysis after derivatization of the corresponding primary amine with ethyl(succinimidooxy)formate (see the Supporting Information).