Cloning, expression, and one-step purification/immobilization of two carbohydrate-binding module-tagged alcohol dehydrogenases

Histidine- and CBM-tagged enzyme cloning and overexpression

Molecular biology experiments led to the obtaining of E. coli M15ΔglyA cells capable to overexpress satisfactorily the histidine-tagged alcohol dehydrogenase from S. cerevisiae as well as CBM3-ScADH and CBM9-ScADH fusion proteins, not only in complex LB medium but also in minimum defined medium.

SLIC technique turned out to be an effective and cost-saving method, since DH5α cell colonies were grown in LB-agar plates in all cases after cell transformation process. DNA sequencing confirmed the correct cloning with no mutations detected in any case.

E. coli M15ΔglyA overexpression screening (Fig. 1) was useful to determine that all clones were able to produce the target proteins. Nevertheless, the percentage of recombinant protein with respect to total protein content (determined by SDS-PAGE) varied quite significantly between the different constructs, from 26% for His-ScADH (Fig. 1A) to 34 and 38% of relative band intensity for CBM3-ScADH (Fig. 1B) and CBM9-ScADH (Fig. 1C), respectively. This increased overexpression levels for CBM-fused variants could somehow reveal the beneficial role of these kind of protein domains, as previously reported [19].

Fig. 1figure 1

SDS-PAGE of E. coli M15ΔglyA final samples from well plate cultures in LB medium. A: His-ScADH. Lane M: molecular weight standard (kDa); lanes 1 to 10: induced cultures. B: CBM3-ScADH. Lanes 1 to 10: induced cultures; lane 11: negative control. C: CBM9-ScADH. Lanes 1 to 4: induced cultures; lane 5: negative control. Culture conditions: 2 mL LB, 24 °C, 24 h, 140 rpm, 0.4 mM IPTG. His-ScADH (37 kDa), CBM3-ScADH (55 kDa) and CBM9-ScADH (64 kDa) corresponding bands indicated with arrows

Minor differences were observed between clones of the same variant, and the colony of each construct which showed the greater overexpression level - according to SDS-PAGE -was picked to generate the cell bank for the subsequent production processes, being these colonies those corresponding to Fig. 1A lane 4, Fig. 1B lane 6 and Fig. 1C lane 3, respectively.

Enzyme production at bioreactor scale

Aiming not only to compare production yields between the three constructs but also to determine if the N-terminal-fused CBMs domains affected negatively to ScADH catalytic activity, the three recombinant proteins were produced in a 2 L bench-scale reactor, using minimum and free-antibiotic media with glucose as carbon source as described.

Bioprocess parameters such biomass, substrate, enzyme specific activity and specific mass production were analyzed along the three processes (Fig. 2).

Fig. 2figure 2

E. coli M15ΔglyA fed-batch cultures. Batch, fed-batch and induction ([IPTG] 0.25 mM) phases indicated. Culture conditions: 37 °C, pH 7.0, 450-1200 rpm, 60% PO2, 0.25 mM IPTG. A: His-ScADH, B: CBM3-ScADH and C: CBM9-ScADH. Arrows indicate the stop of the feeding (↓) and the resume of the feeding (↑)

Regarding the development of the three processes, batch phase lasted for about 16 to 18 hours, with an approximate biomass/substrate yield close to the predicted value of 0.3 g of biomass per gram of glucose. Fed-batch phase duration was also similar in all cases, reaching the desired biomass concentration for culture induction in about 8 to 10 hours of exponential growth. Specific growth rate (μ) of 0.18 h− 1 was calculated during exponential feeding addition, except in CBM3-ScADH production case (Fig. 2B), in which a μ of 0.15 h− 1 was determined. The differences between the pre-set μ in feeding addition and the experimentally measured μ could be due to the coefficient of cell maintenance [35], which was not considered. Besides, CBM9-ScADH production process (Fig. 2C) experienced an accumulation of glucose at the early stage of fed-batch phase, probably caused by an acetate accumulation at the end of batch phase. In that case, feeding was stopped until glucose concentration decreased below 5 g·L− 1 and acetate was completely exhausted.

In all cases, target protein overexpression mechanism was strongly repressed, since first cell lysate samples – corresponding to the moment prior to induction - showed negligible enzyme activity, which increased quite notably in later samples. Protein overexpression induction caused in all cases a metabolic imbalance that led to the accumulation of acetate, glucose, and the subsequent decrease in cell growth, as expected. Production parameters, including the activity units per mol of enzyme (specific activity) were determined for the three cases and were listed in Table 2.

Table 2 Production parameters of ScADH using E. coli M15ΔglyA; comparison between the three different N-terminal-fused tags

Even if similar total amounts of target protein were produced, obtaining 3.58 g of His-ScADH, 3.13 g of CBM3-ScADH and 3.48 g of CBM9-ScADH, specific mass production (mg·g− 1DCW) of CBM3-fused enzyme was slightly higher than histidine- and CBM9-fused ADH (Table 2). This increase was mainly caused by the difference of total biomass obtained in each case, being similar for histidine- and CBM9-fused variants (93 and 90 g DCW, respectively) but much lower for CBM3-tagged enzyme (60 g DCW).

Minor differences were observed among the constructs for target protein overexpression levels, oscillating from 10 to 12%. However, higher values were observed in LB screening experiments, mainly due to temperature shift from well-plate cultures (24 °C) to bench-scale reactor (37 °C); temperature is a well-known and well-described key parameter in recombinant protein production, where greater fraction of synthetized protein tends to fold correctly rather than generate insoluble inclusion bodies at lower culture temperatures because metabolic imbalance produced by a strong protein overexpression induction is tightly affected by temperature [36]. This hypothesis reveals that it still exists room for process optimization, although that was not the point in that case.

Overall, considering that one of the main objectives was the assessment of the possible affectation of CBMs domains to alcohol dehydrogenase’s functionality, specific activity (AU·μmol− 1) of the three variants were determined (Table 2), for which histidine-fused version was 1.5-fold higher than CBM-fused ones. Despite the significant activity loss, the fused CBM domains do not seem to negatively affect the catalytic capability of ScADH enzyme, since the resulting polypeptides are functional and total produced activity values still fluctuate inside the same magnitude order.

Immobilization of CBM-fused proteins

Aiming to characterize the affinity of CBM domains towards cellulose, CBM-fused ScADH enzymes were immobilized to a RAC cellulose support. The characterization was carried out by loading approximately 30 AU·mL− 1 support, where no diffusional limitations were observed.

First batch experiments (Fig. 3A and B) showed the high affinity of both carbohydrate-binding modules towards cellulose, achieving an almost total binding of target protein after 5 minutes of incubation (Table 3, Fig. 3).

Fig. 3figure 3

Alcohol dehydrogenase activity along the immobilization processes of CBM3-ScADH (A) and CBM9-ScADH (B) fusion proteins with 1 mL RAC support

Table 3 Immobilization parameters of CBM-fused ScADH proteins onto RAC support

Legend: Experiment conditions: 24 °C, pH 7.5, 1 mL RAC support, roller agitation. RA, retained activity and IY, immobilization yield.

Results also showed a slight deactivation of CBM3-fused ScADH due to the immobilization process as can be observed in suspension and blank activity profiles (Fig. 3A). This fact led to higher retained activities for CBM9-fused enzyme (97.7%) compared to the CBM3-fused ScADH (86.1%). In addition, recovered activity obtained once the immobilized derivatives were washed was almost 20% higher when CBM9-tag was used.

Regarding the mass balances, in both cases was demonstrated the high specificity of the binding between CBM domains and the cellulosic support, since total protein content difference between initial and final supernatant samples were close to the amount of target protein bound to support. On the one hand, CBM3-ScADH content in supernatant decreased from 14.6 to 0.7%, while final supernatant quantity decreased to 2.12 ± 0.03 mg (83.5% of initial). On the other hand, CBM9-ScADH presence in lysate decreased from 12.2 to 1.1%, recovering an 89.1% of total protein content in final supernatant samples (2.11 ± 0.14 mg).

These results validate the CBM-tagged enzymes as a promising system for one step purification/immobilization process thanks to i) the high specificity of CBM domains towards RAC compared to the other proteins present in E. coli lysates and ii) the high retained activities obtained in the final immobilized derivatives. In order to compare the enzyme storage stability between soluble and immobilized derivatives, samples were kept under refrigeration (4 °C) and suspension activity was measured along time (Fig. 4), revealing that the immobilization process allowed a 2.9-fold increase of half-life - from 13.5 to 38.7 hours - for CBM9-ScADH (Fig. 4B) and 5.5-fold increase for CBM3-ScADH – from 31.6 to 173.6 hours – (Fig. 4A). Despite CBM9-fused protein showed a faster loose of activity than CBM3-ScADH (both soluble and immobilized), the two immobilized derivatives kept a final relative activity of almost a 40% a fortnight after immobilization experiments were carried out.

Fig. 4figure 4

Alcohol dehydrogenase stability for CBM3-ScADH (A) and CBM9-ScADH (B) under refrigeration storage conditions (4 °C). Experiments conditions: 50 mM Tris-HCl buffer pH 7.50, 24 °C, roller agitation. Error bars correspond to standard error of three replicates

Once again, the obtained results consolidate CBM domains as feasible one-step purification/immobilization tags due to the improvement of enzyme stability once the fusion peptides are bound to the support.

The results obtained are in accordance with other immobilisation techniques found in literature; a 90% of RA was achieved by using carboxymethyl dextran (CMD) coated magnetic nanoparticles (CMD-MNPs) activated with epoxy groups, using epichlorohydrin (EClH) [37], by which a 75% of immobilised activity was maintained after 21 days of storage at 4 °C. Besides, other recent studies performed with ADH enzymes reported lower RA values; a 58 and a 62% of RA were reached for ADH variants from Artemisia annua and Streptococcus mutans, respectively, that were immobilised onto agarose resins functionalised with epoxy groups [3839].

Therefore, CBM-based immobilisation method stands as a time- and cost-saving immobilisation alternative, which also enables to reach one of the highest immobilisation yields reported so far.

The maximum enzyme loading capacity of RAC was analyzed for both CBM-fused enzymes by increasing the offered enzyme quantities (Fig. 5). Both fused enzymes could be successfully immobilized under high loads. However, due to mass transfer limitations, steric hindrances and other possible phenomena commonly associated to highly loaded immobilization supports, retained activity values were underestimated [40]. Thus, RA coefficient previously assessed with no-limiting conditions was used to calculate the theoretical final activity of the high-loaded derivatives by assuming to be equal in all cases, since it is not dependent on enzyme amount [14].

Fig. 5figure 5

Correlation between alcohol dehydrogenase activity offered and the theoretical retained activity in the support. Experiments conditions: 50 mM Tris-HCL pH 7.50 buffer, 24 °C, roller agitation

1 mL of RAC support was able to bind up to 7500 ± 275 AU of CBM9-ScADH enzyme from cell lysate, which would correspond to approximately to 115 ± 4 mg of target protein. For CBM3-ScADH protein, RA was significantly lower, binding up to 4300 ± 287 AU per RAC mL, corresponding to 66 ± 4 mg of protein.

Nevertheless, considering that diverse CBM families can be established according to the type of compounds by which these domains present a greater binding affinity (see introduction), it was assumable that a higher amount of CBM9-ScADH fusion protein would be attached to the support, rather than CBM3-ScADH, since the first family modules are characterized to bind amorphous cellulose while the second ones are not [18].

For that reason, an immobilization experiment with increasing amounts of CBM3-ScADH was carried out concurrently, but cell lysate was mixed with Avicel® microcrystalline cellulose instead of amorphous cellulose, aiming to corroborate that CBM3 bounds with higher affinity towards non-treated cellulose (Fig. 5). In that sense, unequivocal results were obtained, since 99.8% of IY was achieved when 5000·AU were offered to 1 mL support and 94.5% of IY was measured when 10,000 UA were offered, demonstrating that the most suitable strategy with CBM3 would be using microcrystalline cellulose instead of RAC.

Summarizing, both CBM domains have proved to be useful tags for ScADH one-step immobilization with cellulosic supports. Even if maximum load capacity of RAC support varied notably depending on the CBM, one positive aspect noticed for both fusion proteins is that enzyme activity remains almost unaltered, making this immobilization method a promising strategy, which also increases storage stability compared with soluble enzyme.

Use of CBM domains as purification tags

Another possible application that CBMs can provide is their use just as purification tags, based on the reversibility of the bound between cellulose and the protein. This way, several purification methods have already been established, including fast protein liquid chromatography (FPLC) processes [25]. Aiming to compare the efficiency of purification process depending on which CBM is fused to target enzyme, FLCP experiments have been performed, in which enzymes were firstly immobilized to the cellulosic support and were then desorbed with glucose.

However, RAC obtained from Avicel® PH-101 – which has a particle size of approximately 50 μm – was unviable for FPLC performance, because the cellulose bread ended up compacting and column flowthrough collapsed. For that reason, same cellulose support was used but with higher particle size (Avicel® PH-200, ~ 180 μm).

For CBM9-ScADH protein, three affinity purification processes performed consecutively resulted in a 94.7 ± 2.3% recovery of activity in average. A trivial ScADH fraction within the clarified lysate load was lost in the column flow-through (2.3 ± 0.05%) and no activity was measured at any of the column washes. The resulting chromatogram and the corresponding SDS–PAGE gel documentation of the purification processes is shown in Figs. 6 and 7, respectively, whereas the purification metrics are provided in Table 4.

Fig. 6figure 6

Chromatogram of three consecutive CBM9–ScADH purifications on pre-treated Avicel® PH-200 support. Experiment conditions: 15 mL column volume, 24 °C, pH 7.50

Fig. 7figure 7

SDS–PAGE of FPLC affinity purifications of CBM9–ScADH on a 10 mL RAC column. Lanes M: molecular weight standard (kDa); lanes 1, 5 and 9: clarified cell extract prior to column loading; lanes 2, 6 and 10: column flow through; lanes 3, 7 and 11: column wash; lanes 4, 8 and 12: purified CBM9–ScADH eluted with 50 mM Tris-HCl buffer containing 2 M glucose. CBM9-ScADH (64 kDa) corresponding band indicated with arrow

Table 4 FPLC-based ScADH purification results for CBM9-ScADH (top) and CBM3-ScADH (bottom)

As expected, a 2 M glucose solution was effective in desorbing all specifically bound target enzyme (Fig. 6), which elutes from the column in a single and clear peak. In addition, it has been proven that RAC support can be reused in consecutive purification batches, since process efficiency and product recovery did not vary significantly among the three experiments performed.

Nevertheless, SDS-PAGE reveals the presence of other proteins in elution fraction (Fig. 7, lanes 4, 8 and 12) – in fact, CBM9-ScADH purity is about 80% -. Although it has not been determined to what corresponds the band that weights around 30 KDa, it could be a broken fraction of the fusion protein which contains the CBM9 domain, that remains attached to the support until the elution step, but these bands cannot correspond to the CBM fragment (26 kDa) nor the ScADH enzyme (37 kDa). Besides, purification processes have been performed by adding protease inhibitor (PMSF) to cell extract to precisely prevent the breaking of the fusion protein.

Legend: Avicel® PH-200 RAC used as immobilization support and 2 M glucose (in 50 Tris-HCl buffer) used for protein elution. Experiment conditions: 15 mL column volume, 24 °C, pH 7.50.

On the other hand, CBM3-fused protein affinity purification resulted in a 6% recovery of initial activity (Table 4). Neither the flow-through nor the wash fractions presented any enzyme activity, disregarding then a loss of target protein in previous fractions. Must be stated that elution step resulted in the appearance of a single and clear peak in the chromatogram, but notably smaller than the observed for CBM9-fused variant, meaning that 2 M glucose solution was unable to unbind CBM3 domain from cellulose.

CBM3 domain has been successfully eluted with other compounds such ethyl glycol [2223], EDTA [41] or trimethylamine [21]. However, these compounds could not be used in our case of study since they would significantly compromise alcohol dehydrogenase activity.

Protein analysis confirmed what was observed by chromatography, given that elution fraction only contained the 2% of total protein, as opposed to CBM9 case, where 20% of total protein content was recovered in elution fraction.

Moreover, when microcrystalline cellulose was used instead of RAC, only was recovered a 1% of initial activity, which is in accordance with results observed for immobilization process. In other words, the more affinity towards substrate, the stronger bound is established, and the harder to desorb CBM3-fused proteins.

In consequence, these results revealed that CBM3-fused enzymes are suitable for a one-step purification/immobilization process but not applicable, under the tested conditions, for purification process based on the affinity interaction. For that purpose, a CBM9-fused strategy would be a better option since it allows both a single purification process and a one-step purification/immobilization process, and the recovery of a highly active enzyme, which has not been reported in most of the bibliography about CBM domains.

Moreover, it can be also concluded that both CBM3 and CBM9 tags are suitable and cheaper alternative to traditional polyhistidine tag used to purify proteins by IMAC chromatography.

留言 (0)

沒有登入
gif