Novel enzyme-polymer conjugates for biotechnological applications

Materials

Sepharose^® 4BCL activated with cyanogen bromide (CNBr) were from GE Healthcare (Uppsala, Sweden). Dextran (Mr 1500, 6000), polyethyleneimine (PEI) (Mr 1500 kDa), 2-dipyridyldisulfide (2-PDS), diethyl-p-nitrophenylphosphate (D-pNP), dithiothreitol (DTT), L-aspartic acid, sodium borohydride, Triton^® X-100, dimethylsulfoxide (DMSO), 1-ethyl-3-(3-dimethylaminopropyl)-carbodiimide (EDC), hydroxylamine hydrochloride, n-hydroxysuccinimide (NHS), cysteine hydrochloride, ethylenediamine (EDA), thymidine, Dimethyl 3- phenylglutarate (2), p-nitrophenylbutyrate (pNPB) (3) and N-Succinimidyl-S-acetylthioacetate were from Sigma. Dextran-aldehyde, aspartic-dextran and amines-dextran prepared as previously described (Fuentes et al., 2004; Guisán et al., 1997). 3’,5’-di-O-acetylthymidine (1) was synthetized as previously described (Bavaro et al., 2009). Other reagents and solvents used were of analytical or HPLC grade.

Site-directed mutagenesis of BTL

B. thermocatenulatus lipase (Genkbank number X95309) mutants were prepared as previously described (Romero et al., 2012). Briefly, all site-directed mutagenesis experiments were carried out by PCR using mutagenic primers (See ESI, Table S1). To introduce the amino acid change, the corresponding pair of primers was used as homologous primer pair in a PCR reaction using a specific plasmid as a template and Prime Start HS Takara DNA polymerase. The product of the PCR was digested with endonuclease DpnI that exclusively restricts methylated DNA (Schmidt-Dannert et al., 1996) E. coli DH10B cells were transformed directly with the digested product. The plasmid with mutated BTL were identified by sequencing and then transformed into E. coli BL21 (DE3) cells to express the corresponding proteins. Firstly, C65S was created, and the resulting plasmid was used as a template to create the double mutant C65S/C296S-BTL. This plasmid was used as a template to construct additional mutations (A193C, L230C) using different mutagenic primer (Romero et al., 2012).

Cloning, expression, purification and immobilization of BTL variants

The gene corresponding to the mature lipase from G. thermocatenulatus BTL was cloned into pT1 expression vector, as previously described (Schmidt-Dannert et al., 1996). Cells carrying the recombinant plasmid pT1BTL2 were grown at 30 °C and over expression were induced by raising the temperature to 42 °C for 20 h. The enzyme was purified from E. coli crude extract by interfacial adsorption on Octyl-Sepharose as previously described (Fernandez-Lorente et al., 2008). The lipase was desorbed from the support adding 20 mL of 25 mM phosphate buffer pH 7 with 0.5% Triton X-100 (v/v) per gram of support. After that, the lipase was immobilized on CNBr-activated Sepharose at pH 7 in 25 mM sodium phosphate buffer for 1 h at 25 °C ( > 95% immobilization yields) with a final loading of 5 mg_lip/g_cat .

Enzymatic hydrolysis of pNPB

The activities of the soluble lipases (for the immobilization process) and their immobilized preparations were analyzed spectrophotometrically measuring the increment in absorbance at 348 nm produced by the release of p-nitrophenol (pNPOH) (∈ = 5.150 M⁻¹ cm⁻¹) in the hydrolysis of 0.4 mM pNPB in 25 mM sodium phosphate at pH 7 and 25 °C. To initialize the reaction, 0.05–0.2 mL of lipase solution or suspension was added to 2.5 mL of substrate solution. Enzymatic activity is given as micromole of hydrolyzed pNPB per minute per milligram of enzyme (IU) under the conditions described above.

Preparation of different thiol-polymers

Preparation of thiol-aspartic-dextran (DextCOOH)

To a 100 mL of dextran-aspartic polymer solution (10 mg/mL) Mw. 1.500 in 5 mM sodium acetate buffer at pH 5.0 were added 16 and 9.6 mg (0.125 eq) of EDC and NHS, respectively. Similar to before, to a 100 mL of dextran-aspartic polymer solution (10 mg/mL) Mw. 6000 in 5 mM sodium acetate buffer at pH 5.0 were added 1.0 and 0.6 mg (0.0313 eq) of EDC and NHS, respectively. For both polymers, the reaction was maintained for 2 h at 25 °C. After that the mixture was then dialyzed 3 times against distilled water (300 volumes). Then, the pH of polymer solution was set at 7.5 in 5 mM sodium phosphate buffer pH 7.5. Subsequently, 420 µL (0.0625 eq) or 26 µL (0.0156 eq) of a 100 mM cysteine solution were added to polymers of Mw 1.500 and 6.000, respectively. The reaction was maintained for 2 h at 25 °C and finally the mixture was extensively dialyzed (See Fig. 3).

Protection of thiol in BTL variants by 2-PDS

Different BTL variants (0.2 g of the immobilized form) were incubated in 2 mL of DTT solution (50 mM in 25 mM sodium phosphate at pH 8) for 30 min to avoid oxidation and permitting the posterior disulfide exchange. After, the reduced biocatalysts were washed with distilled water after the DTT smell disappeared.

Then 0.2 g of reduced BTL variants was added to 3 mL of 2-PDS solution (1.5 mM substrate in a mixture of DMSO (5%, v/v) in 25 mM phosphate buffer at pH 8.0) for 1 h. The cysteine PDS activation was followed spectrophotometrically by measuring the increase of the absorbance at 343 nm (by the release of 2-mercaptopyridine, which quickly tautomerizes into 2-thiopyridone) of the solution. A full modification was found in all BTL variants (Godoy et al., 2010).

Chemical incorporation of tailor-made polymers on different BTL variants

0.7 mL of respective polymer solution was dissolved in 2.3 mL of 500 mM sodium phosphate buffer with 0.5% Triton-X100 at pH 8. Then, 0.2 g of immobilized PDS-BTL variants was added. After 1 h, the modification was confirmed spectrophotometrically by measuring the increase of the absorbance at 343 nm. The biocatalyst was filtered.

Enzymatic hydrolysis of different substrates

Irreversible inactivation of BTL immobilized preparations by diethyl-p-nitrophenylphosphate (D-pNP)

0.2 g of different BTL immobilized preparations were suspended in 4 mL of 25 mM sodium phosphate buffer solution at pH 7 and 25 °C with or without the presence of 0.5% of Triton X-100. Then, 1.5 mM of inhibitor (D-pNP) was added to this solution. The reaction was maintained until the activity – measured using pNPB assay – of the immobilized enzyme was zero.

Mutagenesis, purification and immobilization of BTL variants

Lipase BTL has the particularity that its open configuration involves a large structural rearrangement, and the concerted movements of a tricky lid formed by two different loops. We therefore set out to create semisynthetic BTL variants with only one cysteine located at lid zone. Based on a bioinformatics study using the crystallographic open conformation of the enzyme (Carrasco-López et al., 2009) and the charged amino acid distribution in the lid (Baker et al., 2001), two different positions were selected to be mutated: Ala193 (in the internal loop) and Leu230 (in the external loop), see Fig. 4.

First, focused on the protein engineering of this lipase, the two cysteines (Cys65, Cys296) in the wild type enzyme (BTL-WT) were replaced by two serines. The new-engineered enzyme (BTL-C65S/C296S: BTL*) was expressed in E. coli without detriment to the enzyme activity. Therefore, using BTL C65S/C296S as template, residues Ala193 and Leu230 were replaced by cysteine residues by directed mutagenesis to create the two single cysteine variants: BTL*-A193C and BTL*-L230C. Both variants were expressed in E. coli with a good production similar than BTL-WT and BTL*.

All BTL variants were efficiently purified by hydrophobic chromatography (Fernandez-Lorente et al., 2008) and immobilized on CNBr-activated Sepharose (>95% immobilization yields) with a final loading of 5 mg_lip/g_cat .

Site-directed polymers incorporations

For the strategic site-directed chemical incorporation of polymers on lid zone of lipase, the single cysteine variants, already immobilized on CNBr-agarose support, were protected using 2,2-dithiodipyridine disulfide (2-PDS) to allow the disulphide exchange modification and avoiding undesired oxidation (Fig. 5). First, the immobilized variants of BTL, previously reduced with DTT, were incubated in a solution of 1.5 mM 2-PDS at pH 8.0. Then, the modification was followed spectrophotometrically at 343 nm and obtained a modification yield over 95%.

Finally, the different protected BTL variants were specifically modified with a set of tailor-made thiolated-ionic-polymers. The incorporation of the thiolated-ionic-polymer was followed spectrophotometrically by the release of 2-thiopyridone (as in the previous step) (Fig. 5). After 1 h of incubation, a quantitative yield of modification was achieved in all cases.

The BTL* (variant without cysteines) was subjected to a similar treatment in order to confirm that the chemical modification only involved the thiol group. This mutated BTL could not be modified at all by any of these polymers (data not shown).

Effect of the chemical modification on activity and hyperactivation

The effect of the modification of immobilized BTL* variants using different polymers (different sort and sizes) on the specific activity towards p-nitrophenyl butyrate (3) is shown in Table 1.

The BTL*-A193C variant maintained the same specific activity than BTL-WT and BTL* variants. The modification of BTL*-A193C with Dext-COOH 1500 caused a slight increase on the specific activity (up to 118%). The enzyme modification with others polymers decreased considerably the activity (from 78% to 40%).

In the case of BLT*-L230C variant, the mutation produced a slight decrease in the activity, but the polymer incorporation improved the activity value in all cases, reaching 162% with Dext-COOH 6000.

The modification of the BTL variants with Dext-COOH 20000 greatly decreased the specific activity of the lipase in all cases (data not shown), this might be caused by steric hindrances with the incorporation of a “big size” polymer.

In order to investigate the effect of the polymer incorporation on the interfacial activation of the lipase, the specific activity in the hydrolysis of 3 was measured in the presence of 0.5% triton-X100 (v/v) (Table 1). As previously reported (Godoy et al., 2011), BTL-WT and BTL* are hyperactivated around 2.5 fold in presence of triton-X100, due to the stabilization of the open form of lipase molecule. Both mono-cysteine variants presented a similar hyperactivation (2 fold).

The hyper-activation showed by the polymer-modified BTL*-A193C variant was similar or even lower than non-modified enzyme. However, this site-directed chemical modification of the BTL*-L230C lipase variant caused a higher hyperactivation of the enzyme in the presence of detergent. In particular, BTL*-L230C-DextNH₂6000 showed the highest hyperactivation (approx. 4-fold) (Table 1).

These results show, activity and hyperactivation, that the specific incorporation of ionic polymers on the lid area alter the natural closed-open lid movement and maybe this effect is caused by steric hindrances, or the generation of a different lid conformational restructuration.

Effect of the chemical modification on the irreversible inhibition with D-pNP

In order to evaluate the effect of the polymer incorporation of the lipase surface on the enzyme catalytic mechanism, the irreversible inhibition of the enzyme by using diethyl-p-nitrophenylphosphate (D-pNP) was studied. A faster decrease on the specific activity of the enzyme variants was observed after the chemical incorporation of the polymers (Fig. 6).

In the case of the BTL*-A193C variant, after 10 min, this mutated enzyme preserved 57% of the initial activity while BTL*-A193C-DextNH₂6000 maintained only 29% of activity. When the inhibition was performance in the presence of 0.5% Triton (condition where the conformational equilibrium of the lipase is displaced to the open form), BTL*-A193C conserved the 22% of activity at the same time (Fig. 6A). For the variant BTL*-L230C similar results were observed (Fig. 6B). This result might involve that a more open conformation structure of the lipase was favored in both cases after the polymer modification. Moreover, the size of the polymer also had influence on the final protein structure. The lipase modification with higher molecular size (6000 kDa) polymers caused a faster inhibition than the smaller ones (1500 kDa).

Therefore, it seems that the chemical modification is favouring a higher exposition of the catalytic Ser to the medium, possibly caused by different structural changes in the lid region. It is very likely that it favours an open form of the enzyme with changes in the catalytic properties (Godoy et al., 2010; Romero et al., 2012).

Biotechnological applications of BTL modified variants

The biotechnological potential of these semisynthetic enzymes has been analyzed in two interesting biotransformations: the regioselective deprotection of per-O-acetylated thymidine (1) and the desymmetrization of dimethylphenylglutaric acid diester (2).

First, the BTL variants were tested in the hydrolysis of 1 at pH 7.0 (Table 2). Both mutated variants (BTL*-A193C and BTL*-L230C) showed a slight increase on activity and regioslectivity than the BTL-WT and BTL* toward the production of 3’-O-acetylthymidine (4). The incorporation of DextCOOH-6000 and DextNH₂-6000 in the BTL*-A193C increased the lipase activity in 2.5 and 3 fold, respectively. Furthermore, the BTL*-A193C modified with DextNH₂6000 exhibited a significant increase in the regioselectivity, reaching a yield of 89% of (4). However, the modification at the mutated Cys230 did not produce better results, in both activity and regioselectivity. The BTL*-L230C-DextCOOH-6000 variant showed a decrease in activity in more than 3.5 fold and BTL*-L230C-DextNH₂-1500 exhibited the lowest regioselectivity value (37% of 4).

In order to investigate the importance of the ionization state of the new biocatalysts, the pH of the reaction was decrease (from 7 to 5), showing significant differences (Table 3). The BTL-WT and BTL* showed around 30% less activity at pH 5.0 than at pH 7.0, but a slightly higher regioselectivity. The most interesting changes were obtained in the BTL*-L230C variant. The specific incorporation of DextCOOH-1500 and DextNH₂-1500 produced an increase on the activity in 4.5 and 6.0 fold, more than BTL-WT. These new biocatalysts showed better regioselectivity, 84 and 80% yield of (4), respectively.

The BTL*-A193C variant exhibited the higher regioselectivity value at pH 5.0, with 91% yield of 4. The modification with DextCOOH-6000 increased the activity in nearly 35% without decreasing in the regioselectivity (for time reaction course see Fig. S1).

The desymmetrization of dimethyl phenylglutaric acid diester (2) at pH 7.0 and 5.0 was studied after. The two monocysteine-BTL* variants showed higher enantiomeric excess than BTL-WT and BTL* (Tables 4 and 5).

In the hydrolysis of 2 at pH 7, the activity of BTL*-A193C decreased considerably compared with BTL-WT (more than 17 fold) (Table 4). The incorporation of tailor-made polymers improved the activity and the enantioselectivity. When the BTL*-A193C variant was modified with DextCOOH₂-6000 the activity increased 3.5 fold. The modification with DextNH₂-1500 gave a higher enantiomer excess (87%). Moreover, the best result was found after the chemical modification with DextNH₂-6000. The initial enzymatic activity increased more than 5 fold after this modification and this variant catalyzed the hydrolytic reaction with excellent enantiomeric excess (>99%) (for time reaction course see Fig. S2).

The incorporation of the cysteine in 230 generated a higher active and selective enzyme variant (BTL*-L230C) in the hydrolysis of (2) (93% e.e.) (Table 4). However, any modification improved this result, the modified BTL*-230 catalyst with DextCOOH-1500, DextNH₂-1500 and DextNH₂-6.000 has a similar enantioselectivity but lower activity. With the performance of BTL*-230, the reaction at pH 5.0, involved a slight increase of both activity and enantioselectivity around 20% and reaching 96% e.e., respectively (Table 5).