2.2. Halomonas elongata Aminotransferase
HEWT gene (spuC
) was trialled first in both constructs (T4L-HEWT_L1 and T4L-HEWT_L2, nine and three amino acid long spacers, respectively). The expression level of both chimeras was lower than what obtained for HEWT, resulting in 12 and 20 mg of protein per litre of media, after IMAC purification, for T4L-HEWT_L1 and L2, respectively (Table 1
). The proteins analysed by SDS-page electrophoresis did not show formation of any inclusion bodies caused by incorrect folding, and the decrease expression might simply be attributed to physiological limitation of the E. coli
host in translating a longer gene. Although fusion tags have been reported to have a positive influence on protein solubility and expression, they do not function equally well with all target proteins [12
]. Furthermore, analysis of the purified fractions confirmed the integrity of the constructs showing a molecular weight in the correct range of expected mass (Figure S1
). HEWT specific activity was also affected by the presence of the lysozyme, with values of 1.37 and 0.50 U/mg for the two T4L-HEWT (Table 1
). Such an effect could be explained with the fact that the N-term is a delicate structural domain which forms part of the active site; an exogenous module linked to it has a direct influence on its position and consequently on its catalytic efficiency. Indeed, with the longer linker the effect on activity was less severe, while with only three amino acids separating the lysozyme from HEWT the activity dropped 10 times with respect to the control. The proximity of the T4L can also hinder the substrate accessibility to the active site further influencing the reactivity.
The immobilisation of the chimeric constructs was carried out on two supports: Sepabeads EC-EP/S are described to have an average pore diameter of 10–20 nm, and ReliZyme EP403/S with a larger pore size (20–40 nm) which should favour better protein distribution and prevent ‘caging’ of the enzymes into narrow pores. In fact, the additional bulk of the T4L makes the chimeric construct significantly larger and this could potentially lead to immobilisation difficulties when the pore size of the resin is small. Immobilisation of the constructs followed the same procedure described previously [18
], maintaining cobalt(II) as the coordinating metal. Initially, 5 mg of enzyme per gram of resin were used, allowing the protein to react with the epoxy groups on the carrier for 24 h (Table 1
On Sepabeads EC-EP/S, the control HEWT sample yielded 22% of recovered activity, while both T4L-HEWTs showed a marked improvement in the retention of activity with 80.4% and 59.8%, for the nine- and three-amino acid linker, respectively. These results confirm that the TL4 prevents direct attachment of the enzyme to the resin and this effect is enhanced by a longer spacer. On ReliZyme EP403/S the smaller wild type HEWT behaved virtually in the same manner, and likewise T4L-HEWT_L1 for which the rescued activity was already very high and the benefits of larger pores were not particularly significant. However, for the T4L-HEWT_L2 construct, the change observed in recovered activity was important, with almost a 30% improvement. Larger pores in this case may impact the number or covalent bonding with the surrounding surface leading to a reduced rigidity and distortion of the protein as well as generally increasing the substrate diffusion rate.
A range of concentrations for HEWT and T4L-HEWTs were then screened to assess the loading capacity of the resin, and whether the specific activity as well as the rescued activity could be further improved. The behaviour of HEWT and T4L + HEWT_L2 did not change significantly between 1 and 5 mg. At 10 mg of enzyme, the rescued activity decreased presumably due to surface overloading. On the other hand, the T4L + HEWT_L1 presents a bell-shape behaviour with a peak at 3 mgenzyme
); it is plausible that the T4L coats the surface of the resin further protecting the enzyme from any binding to the resin yielding to almost equal activity as recorded for the free enzyme and a specific activity of the imm-T4L-HEWT_L1 of 1.28 U/mg versus 1.38 U/mg of the free T4L-HEWT_L1.
The incubation time to facilitate enzyme binding was investigated. It was postulated that the high distribution of lysines on the T4L surface could promote a faster immobilisation, and shorter incubation time may result in higher rescued activity. The immobilisation was thus carried out by leaving the enzymes to react with the resin between 1 and 24 h. Wild-type HEWT rescued activity was unaffected by the incubation time and consistent around 25–30%, while T4L + HEWT_L1 showed an unexpected behaviour with optimal value of rescued activity only after 24 h (Figure S2
). 24 h was therefore selected as the ideal incubation time for further assays for all the HEWT variants.
Since the T4L chimeric proteins showed greater retained activities, they were further analysed to investigate the effect of the linker protein on the stability of the enzyme. Under standard storage conditions (4 °C, buffer phosphate pH 8, 0.1 mM PLP), the immobilised T4L-HEWTs showed unchanged specific activity for over a month. When the stability was evaluated by incubating the resin at 37 °C and 45 °C, no significant differences were observed with respect of the control HEWT immobilisation. On the other hand, the stability under working conditions (consecutive cycles of bioconversion), showed a slight negative trend in performance for the T4L-HEWT_L1 (Figure S3
—circle), while the T4L-HEWT_L2 (Figure S3
—square) was more stable with a virtually unchanged activity after 10 cycles, similarly to the imm-HEWT behaviour [18
A significant feature of the imm-HEWT was the improved ability to withstand the presence of organic co-solvent in the reaction media with respect to the free enzyme. The construct harbouring the lysozyme spaced by the longer linker showed a similar behaviour to the free enzyme (Figure S4
). This is not unexpected since the rationale of the protein linker was to prevent a direct bonding of the enzyme to the resin allowing a greater degree of freedom to the catalyst, which in this way no longer benefits from the imposed rigidity of the multi-point attachment to the carrier. However, when the shorter linker construct was assessed, the catalyst performed marginally better than the imm-HEWT (Figure S4
), which is indicative of an enhanced stability induced by the close proximity of T4L to protein, similarly to the effect that BSA may have on pure proteins [23
2.3. Bacillus subtilis Esterase
To validate the approach of the protein linker as a beneficial addition in enzymatic immobilisation, the strategy was extended to a second enzyme: Bacillus subtilis
esterase, specifically a mutant variant (BS2m) engineered to augment its amidase activity [21
]. Immobilisation of the purified enzyme via covalent attachment proved to be virtually impossible with almost complete loss of activity yielding negligible nominal activity of the support (1.3 U/g) when 5 mgenzyme
were applied, equal to a rescued activity of 0.4% (Table 2
). Such inability to retain activity upon covalent immobilisation is characteristic of many esterases [24
]. Gross and co-workers speculated that this class of proteins tend to create a superficial layer preventing an efficient substrate diffusion [26
]. As an example, the commercial catalyst NovoZyme 435 (Candida antarctica
lipase-B, CAL-B) is immobilised at 82 mgenzyme
but yields only 7.5% recovery activity, highlighting how this class of enzyme is particularly affected by immobilisation [27
]. In addition, in the case of the BS2m variant, the superficial lysines, are concentrated in proximity to the active site which could drive the orientation of the enzyme in such a way that the catalytic pocket faces the support limiting its accessibility.
The same strategy applied for the HEWT was thus evolved towards the BS2m creating two chimeric constructs harbouring the 9- and 3-amino acid linkers. In this particular case, both longer and shorter linkers (T4L-BS2m_L1 and T4L-BS2m_L2, respectively) showed a less severe reduction in expression levels when compared to the original BS2m (Table 2
, Figure S5
) and an activity of 31 and 35 U/mg, respectively (Table 2
). Covalent immobilisation onto the Sepabeads EC-EP/S epoxy-resin of the T4L-BS2m_L1 yielded an active resin with 5.5 U/g equal to 3.6% recovery activity, a notable 10-fold improvement on the original BS2m (Table 2
). T4L-BS2m_L2, as observed for HEWT, did not perform as well as the L1 linker, with lower, yet still improved, immobilisation qualities (2.9 U/g and a 1.7% recovery activity). The co-immobilisation of T4L and BS2m, separately expressed and purified, was also tested to exclude a possible non-specific effect of the T4L but no enhancement of the recovered activity was noted in this case (data not shown). This further confirmed that the fused T4L physically shields the enzyme from covalent attachment and does not directly affect enzyme activity.
Immobilisation of BS2m, T4L-BS2m_L1, and L2 was also tested on the ReliZyme (EP403/S). Interestingly, the new support offered better performance than the Sepabeads, allowing for a more efficient immobilisation for all variants (though still lower for the BS2m control). The recovered specific activity of T4L-BS2m_L1 reached 6.2 U/g (3.9%) and 3.7 U/g (2.1%) for the construct with the shorter link, confirming the benefits of a facilitated diffusion inside the solid support. This strategy therefore allowed to prepare an immobilised esterase with an activity considerably higher than the commercial NovoZyme 435 (4.5 U/g, tested under the same condition) using 16-fold less enzyme. Furthermore, the immobilised enzyme exhibited a great stability retaining 87% of its initial activity after five reaction cycles, comparable to the commercial CAL-B (83%) [27
2.4. Horse Liver Alcohol Dehydrogenase
Finally, horse liver alcohol dehydrogenase (HLADH) was selected as an additional test enzyme, since its immobilisation had already been carried out on the same epoxy resin, providing a good reference for the experiment [28
]. This enzyme presents additional challenges to general immobilisation strategies since its cofactor is not tightly bound and the cofactor binding domain needs to be accessible during the catalytic cycle [29
]. For these reasons, structural alterations may have a major impact on the reactivity, unlike in the case of the aminotransferase where the PLP cofactor is permanently buried inside the active site.
gene was subcloned into both vectors similarly to the other two genes. However, for both constructs, expression level and specific activity dropped dramatically with an activity 10 times lower, and yielding less than 1 mg of protein per litre of culture media after IMAC purification (Figure S6
). Nevertheless, immobilisation of both chimeras and wild type HLADH was performed. The specific activity of the resin following immobilisation of the T4L-HLADHs (1 mgenzyme
) was very low overall compared to the native imm-HLADH (Table 3
). However, similarly to what observed for the HEWT, the T4L-shielded enzymes showed an improvement in the recovered activity although the overall nominal activity remained rather low. Immobilisation of the enzymes on Relizyme EP403/S in this case did not offer any appreciable improvement.
The original HLADH lost about two-thirds of the activity, while the two chimeric enzymes responded better to the immobilisation, reaching around 50% of recovered activity. The stability of the immobilised enzymes was excellent for all the constructs under storage condition (4 °C in Tris-HCl buffer pH 8), with almost unchanged activity for over a month.