Stabilization of thymidine phosphorylase from Escherichia coli by immobilization and post immobilization techniques

Abstract

Homodimeric thymidine phosphorylase from Escherichia coli (TP, E.C. 2.4.2.4) was immobilized on solid support with the aim to have a stable and recyclable biocatalyst for nucleoside synthesis. Immobilization by ionic adsorption on amine-functionalized agarose and Sepabeads^® resulted in a very high activity recovery (>85%). To prevent undesirable leakage of immobilized enzyme away from the support, the ionic preparations were cross-linked with aldehyde dextran (MW 20 kDa) and the influence of the dextran oxidation degree on the resulting biocatalyst activity was evaluated. Although in all cases the percentage of expressed activity after immobilization drastically decreased (≤25%), this procedure allowed to obtain an active catalyst which resulted up to 6-fold and 3-fold more stable than the soluble (non immobilized) enzyme and the just adsorbed (non cross-linked) counterpart, respectively, at pH 10 and 37 °C. No release of the enzyme from the support could be observed. Covalent immobilization on aldehyde or epoxy supports was generally detrimental for enzyme activity. Optimal TP preparation, achieved by immobilization onto Sepabeads^® coated with polyethyleneimine and cross-linked, was successfully used for the one-pot synthesis of 5-fluoro-2′-deoxyuridine starting from 2′-deoxyuridine or thymidine (20 mM) and 5-fluorouracil (10 mM). In both cases, the reaction proceeded at the same rate (3 μmol min⁻¹) affording 62% conversion in 1 h.

Introduction

Nucleoside phosphorylases (NPs) participate in the salvage pathway of nucleoside synthesis [1], [2]. These enzymes naturally catalyze the cleavage of ribonucleosides and deoxyribonucleosides to the free base (B₁) plus ribose 1-phosphate or deoxyribose-1-phosphate, respectively; in the presence of an acceptor nucleobase (B₂) the one-pot formation of a new nucleoside may take place by an equilibrium controlled transglycosylation process [3] (Scheme 1).

NPs from different sources, mainly bacterial, have been widely investigated as catalysts for the synthesis of unnatural nucleosides with antiviral and anticancer activity [4], [5], [6], [7]. The enzymatic syntheses of both natural and unnatural nucleosides reported in the last two decades offer several advantages over the chemical methods since highly regio- and stereo controlled reactions occur [4], [5], [6]. Moreover, as a result of route redesign using NPs, straightforward, greener and cost efficient syntheses can be achieved.

Generally speaking, while the use of enzymes as biocatalysts to assist in the preparative/industrial manufacture of fine chemicals has tremendous potential, application is frequently limited [8]. Biocatalyst stability is a major concern because it affects catalytic efficiency, frequency of catalyst replacement and product yields. The causes of poor biocatalyst stability are closely associated with the process conditions, and may include extreme temperatures or pH, or the presence of organic solvents that are outside the operating “stability window” of the catalyst, but that are often necessary to solubilize poorly water-soluble substrates in high concentrations. A particularly complex issue is the stabilization of multimeric enzymes. These enzymes may be easily inactivated by dissociation of the protein subunits, or by their uncorrected assembly.

Increased biocatalyst stability has mostly come from immobilization and post immobilization cross-linking [8], [9], [10], [11]. However, stabilization of enzymes can be achieved also by genetic engineering [12], or by coupling mutagenesis and immobilization [13]. In the case of immobilization–stabilization strategy, the multimeric nature of the enzyme should be considered to pursue not only the multipoint attachment of the protein, but also the multisubunit attachment [11].

Thymidine phosphorylase from Escherichia coli (TP, E.C. 2.4.2.4) catalyzes the reversible phosphorolysis of thymidine and other pyrimidine 2′-deoxynucleosides with the exception of 2′-deoxycytidine [14], [15]. This enzyme also accepts as a substrate many 5-substituted uracils and 2-S-substituted uracils (i.e. 2-thiouracil and 2-thiothymine) [16]. Conversely, aza-derivatives of uracil and thymine are not accepted as substrates, as well as compounds where C-6 is substituted by methyl or keto groups (5,6-dimethyluracil, 5,6-dihydrouracil, and barbituric acid) [16]. Substrate specificity of E. coli TP to thymidine derivatives modified at 2′-, 3′-, and 5′-positions of the sugar moiety has been also assessed [15], [16], [17]. The 2′-OH group enables more effective binding, but ribonucleosides decrease the reaction rate. The presence of 3′-OH group is essential for the phosphorolysis reaction to occur, while the removal of 5′-OH group has no effect either on the reaction rate or binding.

E. coli TP has been well characterized also for its three-dimensional structure [18], [19], catalytic mechanism [2] and the range of possible applications [20], [21], [22]. Moreover, this enzyme has been used as whole cells [23], crude extract [24] and purified enzyme [16], [25], [26], [27] for the synthesis of a wide variety of nucleosides. However, to our best knowledge, differently from other pyrimidine nucleoside phosphorylases [28], [29], [30], TP has never been conveniently immobilized as a purified enzyme. TP can be considered, therefore, an interesting candidate for an immobilization study aimed at preparing a stable, active and reusable biocatalyst for transglycosylation reactions.

TP from all the reported sources acts as a dimer [31], [32], [33], [34] of identical subunits with a global mass ranging from 110 kDa in mammals to 90 kDa in E. coli. In TP from E. coli [19], each subunit contains an active site composed of two distinct domains: a small helical domain (α domain) and a mixed α–β domain separated by a large cleft. The binding site for the phosphate ion is located in the α–β domain while the binding site for the thymine/thymidine is located in the cavity between the two domains. The distance between the two sites is such that domains’ movement is necessary to bring the substrates close one to each other. This movement also shields the substrates from the solvent, preventing the escape of the intermediates before catalysis is complete [34]. Mostly important, due to the role played by domains’ movement in catalysis, this should be preserved also after immobilization.

In this work, several biocatalysts have been prepared by immobilizing TP from E. coli on supports characterized by different physico-chemical properties (hydrophilicity degree) and by different functional groups and, therefore, following different immobilization chemistries.

Taking into account the peculiar dynamic catalytic mechanism of TP, immobilization procedures resulting in a mild interaction between the enzyme and the support have been firstly considered. To this aim, adsorption on different ionic supports was studied and postimmobilization cross-linking by aldehyde dextran was performed to further stabilize the enzyme, according to the strategy successfully adopted for other multimeric enzymes [35] as well as for uridine phosphorylase (UP) from B. subtilis [28]. Polyethyleneimine (PEI) was indeed used as coating agent for ionic support preparation besides low molecular weight molecules such as N,N-diethylethylenediamine (DEAE) and ethylendiamine (EDA). PEI can be more advantageous over DEAE or EDA due to its polymeric nature which can be quite likely to interact with areas on the protein surface located in different enzyme subunits [36], and therefore desirable in case of multimeric enzymes such as TP.

Covalent immobilization was also performed, as previously reported for different enzymes. Both aldehyde [37], [38] and epoxide [39], [40] activated supports were used. Agarose activated with aldehyde groups (glyoxyl agarose) is routinely used as a hydrophilic carrier for immobilization and stabilization of enzymes [37], [41], [42]. This immobilization procedure relies on the formation of Schiff bases between the enzyme amino groups (mostly surface Lys) and the support aldehydes, followed by chemical reduction of imino bonds to stable covalent C–N.

Epoxy-activated acrylic resins are reported as ideal carriers for the industrial development of stable biocatalysts [39], [40], [43], [44]. A preliminary hydrophobic adsorption to the support, enhanced at high ionic strength, is followed by a covalent attachment which involves the nucleophilic groups of the enzyme (Lys, Cys, Tyr) and the epoxy groups. Furthermore, epoxy-activated carriers can be easily functionalized with aldehyde [45] or ionic groups [34]. Comparison of hydrophilic (agarose) and less hydrophilic supports (acrylic resins) with the same activation (aldehyde) was investigated in this work to highlight the effect of the nature of the matrix on TP activity and stability.

To test the activity of the better performing immobilized enzymes, the synthesis of 5-fluoro-2′-deoxyuridine (FdUrd, 3) via transglycosylation was carried out using both 2′-deoxyuridine (dUrd, 1a) and thymidine (Thd, 1b) as the donor nucleoside (Scheme 2). Differently from other 2′-deoxynucleosides, dUrd and Thd can be readily synthesized at a moderate price by chemical methods resulting, therefore, in quite convenient raw materials [46], [47].