3.3 TIPO SUBJETIVO

2.3.1 TbINO1 Protein Overexpression and Purification

TbINO1 was recombinantly expressed and purified on a large scale using the construct pET15b-TbINO1 in BL21 Rosetta (DE3) cells.102 The protein was purified by Ni affinity chromatography, eluted with an increasing imidazole gradient (10 mM to 400 mM) in a 20 mM Tris acetate (pH 7.5) buffer containing 300 mM NaCl. Removal of imidazole was achieved by dialysis against 20 mM Tris acetate (pH 7.5), 50 mM NaCl and 5 mM dithiothreitol (DTT). The protein was then stored containing 20 % glycerol at -80 °C.102 When conducted, 8 L of E. coli cultures were grown and, typically, 600 mg of TbINO1 was collected after purification. This gives 75 mg L-1 of TbINO1, a very high yielding expression. Electrophoresis gels provided evidence of a good level or purity from the procedure (Figure 2.3.1).

Figure 2.3.1 – Protein electrophoresis gel from a standard protein purification. A – Whole cell, B – supernatant, C – Flow through, D – 10 mM imidazole wash, E – 20 mM imidazole wash, F – 50 mM imidazole wash, G – 100 mM imidazole wash, H – 250 mM imidazole wash, I – 400 mM imidazole wash.

2.3.2 Initial Optimisation

Before any work could be done on the scale up of the biotransformation, it was necessary to optimise the process on a small scale. As high conversion had been highlighted as an important target, work on the scale up of the process would not be started until a conversion of >95 % had been achieved. The original system22 had achieved such conversions, but switching to an ammonium bicarbonate buffer could have a big effect on the proficiency of the process. The first experiments used the same amount of protein (87 µg), D-glucose 6-phosphate (18, 10 mM) and DTT (1 mM) throughout, and the total volume of the reaction mixture was 150 µL. The concentration of ammonium bicarbonate in the buffer was kept at 50 mM and the temperature of the reaction was always 37 °C. The pH and concentrations of NAD+ and ammonium acetate were investigated.

Initial results were poor, with the IMPase assay demonstrating conversions as low as 8 %. The first variable that was investigated was the pH of the ammonium bicarbonate buffer.

for these reactions were again very low, with only 16 %, 23 % and 20 % being observed respectively (Table 2.3.1). With conversions this low, it was hypothesised that the

TbINO1 used had lost some activity after a long period of storage. Therefore, a new batch of TbINO1 was expressed and purified. Utilising the new TbINO1, the reaction gave a 53 % conversion with the buffer at pH 8.5.

TbINO1 pH [NAD+_{] Conversion}

Old 8.0 1 mM 16 % Old 8.5 1 mM 23 % Old 9.0 1 mM 20 % New 8.5 1 mM 53 % New 8.5 2 mM 56 % New 8.5 4 mM 50 %

Table 2.3.1 – Calculated conversions showing differences in pH, old and new TbINO1 batches and NAD+ concentration. Other variables were set to 87 µg of TbINO1, 10 mM

D-glucose 6-phosphate (18), 2 mM ammonium acetate and 50 mM ammonium bicarbonate as buffer. Reactions incubated at 37 °C overnight.

As the cofactor, NAD+ plays an essential role in the reaction. Potential optimisation of the NAD+ concentration was conducted by running three overnight experiments in parallel, with varying concentrations of NAD+ (Table 2.3.1). With only a small variation in conversion observed between these reactions, it is difficult to draw any conclusions as to the optimal concentration of NAD+. A concentration of 2 mM was chosen for remaining experiments. The concentration of ammonium acetate within the reaction mixture, seemed to have little or no effect on conversion. The previous work22 had shown that ammonium ions were needed for the biotransformation to progress. The selected ammonium bicarbonate buffer now contained this ion in high abundance. Doubling the ammonium acetate concentration to 4 mM had little to no effect, and indeed, in later experiments, completely omitting the addition of ammonium acetate had no effect on conversion. The DTT is included in the reaction mixture to prolong the life of the protein by relieving oxidative stress. It reduces and prevents enzyme S-S bonds

forming intra- or intermolecularly with cysteine residues. Reaction conversion was therefore not affected by an increase in DTT concentration.

Having investigated each of the components of the reaction, it was hypothesised that there could be an issue with the IMPase assay. If the IMPase was not efficiently cleaving the phosphate from L-myo-inositol 1-phosphate (19) then a true representation of the conversion would not be observed. The reaction conditions used for the IMPase reactions were initially 50 mM ammonium bicarbonate (pH 8.5), 5 mM magnesium chloride and 1.25 mU IMPase (bovine brain, Sigma) for 3 hours at 37 °C. Increasing the reaction time had little to no effect on the conversions observed. Leaving the reaction overnight led to the reaction standard (containing no TbINO1) being observably very green in colour after addition of Malachite Green, suggesting that IMPase is able to cleave the phosphate from

D-glucose 6-phosphate (18). Although this suggests that IMPase is not selective for the reaction product, D-glucose 6-phosphate (18) is hydrolysed at a much slower rate than L-

myo-inositol 1-phosphate (19), and within the preferred 3 hour reaction time for the IMPase reaction, the amount of free phosphate present from the starting material would be negligible. Using an increased amount of IMPase gave no discernible increase in conversion, and changes to magnesium chloride concentration also had little to no effect.

Finally, the Tris acetate buffer was used for this reaction. This was the original buffer used by the previous work,22 but had been changed to ammonium bicarbonate in error causing the IMPase to act less efficiently. Using the original buffer for this reaction, pushed the conversion of 56 % that had been observed previously, up to 80 %.

On this initial 150 µL scale, the desired conversion of >95 % had still not been observed up to this point. It was hypothesised that switching the buffer of the stored enzyme prior to addition to the reaction mixture, may have a desirable effect. The enzyme had been stored in the previously used Tris acetate buffer, and it was discovered that by switching this to the ammonium bicarbonate buffer the reaction conversion was pushed to 92 % on the first occasion and since then, conversions >98 % have been observed. The buffer was switched using centrifugal filters with 30 kDa cut off, washing the protein with ammonium bicarbonate buffer three times before adding to the reaction.

Variable Optimum D-Glucose 6-phosphate (18) 10 mM TbINO1 87 µg NAD+ 2 mM DTT 1 mM Ammonium bicarbonate 50 mM (pH 8.5)

Table 2.3.2 – Table showing the required conditions to achieve full conversion of D- glucose 6-phosphate (18) to L-myo-inositol 1-phosphate (19) at 37 °C overnight. Total reaction volume of 150 µL.

The final system used to obtain >95 % conversion on this 0.4 mg scale (10 mM D- glucose 6-phosphate (18) in 150 µL reaction) was therefore 87 µg TbINO1, 50 mM ammonium bicarbonate (pH 8.5), 2 mM NAD+ and 1 mM DTT conducted at 37 °C overnight (Table 2.3.2).

2.4 Scale-Up “In-Batch”