Expression and purification of soluble and active human enterokinase light chain in Escherichia coli

Highlights • Recombinant production of soluble, active enterokinase (EK) is challenging.• Maltose binding protein-fusion improves EK solubility but reduces activity.• GroEL/ES and Erv2/PDI induces correct refolding and improves EK activity.• Replacing free cysteine with serine dramatically improves EK activity.

Recombinant fusion technology has been used to enhance the expression level and solubility of target proteins, and to facilitate their purification [1,2]. Proteases including Factor Xa, thrombin, tobacco etch virus (TEV) protease, and enterokinase (EK) are used for the site-specific cleavage of recombinant tags from fusion proteins [3][4][5][6]. While Factor Xa, thrombin, and TEV protease cleave inside the recognition site, EK cleaves outside the site, thus it has a proteolytic activity regardless of the P1 0 position sequence.
In this study, we present strategies to produce active hEK L in E. coli cytoplasm. We report production of soluble, active hEK L with improved folding efficiency that can be used in-house. To produce active, cytoplasmic hEK L with the correct disulphide bonds, we constructed hEK L fused with MBP through the D 4 K cleavage site and expressed this in E. coli cells expressing chaperone proteins (Fig. 1a). A previous report demonstrated expression of soluble and active MBP-tagged hEK L [26]. However, we found that MBP-hEK L was unable to self-cleave, indicating an absence of the enzymatic activity (Figs. S1 and 1b). To test whether removal of MBP could restore the hEK L activity, an hEK L variant was constructed by replacing the D 4 K with the TEV protease recognition site (ENLYFQ). However, hEK L obtained by TEV cleavage of MBP-hEK L was still inactive (data not shown). To investigate whether the loss of activity resulted from a limited reduction of disulphide bonds or misfolding, we conducted a refolding process to rearrange disulphide bonds. Detection of self-cleaved forms of refolded hEK L indicated that the refolded enzyme was partially active (Fig. S2).  These results demonstrated that MBP fusion enhances the solubility of hEK L but does not allow its correct folding. We speculated that hEK L misfolding might result from incorrect disulphide bonds formed during expression in E. coli. Therefore, to promote the formation of the correct disulphide bonds in E.coli-expressed hEK L , we employed three strategies: (i) use of a trxB À , gor À , ahpC* + mutant expressing cytoplasmic DsbC (SHuffle strain) for oxidative folding, (ii) replacement of the free cysteine with serine (C112S), which bound to heavy chain, to reduce misfolding, and (iii) co-expression of molecular chaperones that isomerize disulphide bonds. First, when the SHuffle strain was used, self-cleaved hEK L was successfully detected, although at a low level (7.9 % of total MBP-D 4 K-hEK L ), in cells grown at 20 C (Fig. 1c). Use of the C112S mutated hEK L dramatically improved the ratio of self-cleaved hEK L to up to $49.5 % in cells grown at 20 C, which may be caused by the reduced mispairing of multiple disulphide bonds [12,27]. Remarkably, fully self-cleaved hEK L was detected from cell coexpressing GroEL/ES and Erv2/PDI grown at 20 C. In particular, the activity was slightly higher upon GroEL/ES co-expression. Notably, hEK L was not visible in the SDS-PAGE gel even when hEK L activity was observed. However, as shown in Fig. S3, when inactivated hEK L was produced by TEVp, hEK L was visible in the SDS-PAGE gel. Therefore, we assumed that the visibility of hEK L in the SDS-PAGE gel was influenced by its folding.
We further monitored the time profiles for cell growth and enzymatic activity of hEK L C112S ( Fig. 2a and b). After 27.5 h of culture, the cell growth reached the maximum (2.87 OD 600 ) and then sharply decreased. At that time, the hEK L activity in the soluble fraction reached the maximum value (372 U/mL) and then decreased to $22 U/mL. In contrast, hEK L in culture supernatants reached the maximum value (303 U/mL) after 75.5 h of culture. These results indicated that hEKL may be released into the extracellular fraction by autolysis of cell.
We attempted to obtain highly pure hEK L C112S from culture supernatants. The culture supernatant of E. coli SHuffle expressing pET-30a-MBP-D 4 K-hEK L C112S and pACYC-GroEL/ES was loaded on the affinity chromatography (HisTrap TM ) along with 1 mM DTT to improve the binding efficacy (Fig. 2c). The enzymatic activity was 306 AE 0 U/mL and 3085 AE 43 U/mL before and after purification, respectively ( Fig. 2d-g). A previous report [11] showed that a lowyield hEK L (10 %) can be purified from the culture media of P. pastoris using a two-step purification with several pre-treatment steps [11]. However, we could purify hEK L at high purity (>99 %) and yield (>99 %) using a simplified one-step method. Purified hEK L C112S had affinity to GD 4 K-na with K M = 0.287 AE 0.079 mM, turnover number K cat = 6.725 Â 10 4 AE 1.230 Â 10 4 s À1 , and catalytic efficiency K M /K cat = 2.385 Â 10 5 mM À1 s À1 .
In conclusion, we could purify soluble and active hEK L at a high yield using an MBP tag, replacing the free cysteine with serine, using E. coli strain promoting oxidative folding, co-expressing molecular chaperone that isomerise disulphide bonds, and culturing at low temperature. These findings provide strategies for purification of the complex, multiple disulphide-bonded hEK L from E. coli.

Data statement
All data reported in the paper are available from the corresponding author upon reasonable request. Materials and Methods in this study are described in the Supplementary information.

Declaration of Competing Interest
The authors have no competing interests to declare