A strategy for L-isoleucine dioxygenase screening and 4-hydroxyisoleucine production by resting cells

ABSTRACT L-Isoleucine dioxygenase (IDO) specifically converts L-isoleucine(L-Ile) to 4-hydroxyisoleucine(4-HIL). To obtain IDO with improved activity, a strategy was developed that is dependent on the restoration of succinate-minus E. coli cell growth by the coupling of L-Ile hydroxylation and the oxidation of α-ketoglutarate(α-KGA) to succinate. Five mutants were obtained with this strategy, and the characteristics of IDOM3, which exhibited the highest activity, were studied. The catalytic efficiency, thermal stability and catalytic rate of IDOM3 were significantly improved compared with those of wild-type IDO. Moreover, an efficient method for the biotransformation of 4-HIL by resting cells expressing IDOM3 was developed, with which 151.9 mmol of 4-HIL/L (22.4 g/L) was synthesized in 12 h while the substrates seldom exhibited additional consumption.


Introduction
4-Hydroxyisoleucine(4-HIL) is a natural nonproteinogenic amino acid that was first isolated from the seeds of Trigonella foenum-graecum and has been proven to exhibit glucose-dependent insulinotropic activity in rat models of type 2 diabetes mellitus(T2DM). Moreover, 4-HIL does not induce side effects such as hypoglycemia, which occurs during T2DM therapy. 1-4 4-HIL has also been found to be an effective drug to control body weight gain, glycemia, insulinemia and decreased plasma triglyceride levels in rodents. 5 Furthermore, recent studies have shown that 4-HIL exhibits effective antidiabetic activity in a model of type1 diabetes mellitus without insulin. 6 The primary method for 4-HIL production is extraction from fenugreek seeds; however, the yield by this method is rather low (approximately only150 mg of 4-HIL can be extracted from 1 kg of fenugreek seeds). 2,7 In addition to seed extraction, chemical and enzymatic synthesis methods have also been developed but were considered to result in lowefficiency, highcost and heavy pollution [8][9][10] 4-HIL has at least 8 stereo configurations, but only (2S, 3R, 4S)-4-HIL exhibits biologic activity. 11 L-Isoleucine dioxygenase (IDO) from Bacillus thuringiensis 2e2 has been found to specifically convert L-isoleucine(L-Ile) to (2S, 3R, 4S)-4-HIL, which is a member of the a-ketoglutarate(a-KGA)-dependent hydroxylase family, and (2S, 3R, 4S)-4-HIL has been successfully synthesized by IDO. 11 In a previous study, we cloned ido(Accession KC884243) from isolated B. thuringiensis TCCC 11826, and IDO was used to synthesize 4-HIL via biotransformation. 12 However, the 4-HIL production was rather low(44.64 mM in 36 h), and the substrates(L-Ile and a-KGA) were also found to exhibit additional consumption during the process of biotransformation.
To further gain IDO with improved activity, a strategy dependent on the coupling of L-Ile hydroxylation, the oxidation (decarboxylation) of a-KGA to succinate and cell growth was developed. Five mutants were obtained via this strategy, and the characteristics of the mutant exhibiting the highest activity were studied. Moreover, a method for 4-HIL production by resting Escherichia coli cells overexpressing activity-improved IDO was developed.

Results and discussion
Restoration of E. coli DsucADaceA-ido growth by IDO via the coupling of L-Ile hydroxylation and the oxidation of a-KGA to succinate One unique property of the a-KGA-dependent hydroxylase reaction is the coupling of substrate hydroxylation and the oxidation (decarboxylation) of a-KGA to succinate. 13 Thus, IDO may shunt the TCA cycle when succinate synthesis is blocked, thereby coupling L-Ile hydroxylation and cell growth.
In E. coli, there are 4 pathways for succinate synthesis: (1) from a-KGA by a-ketoglutarate dehydrogenase via the TCA cycle under aerobic conditions, (2) from isocitrate by isocitrate lyase via the glyoxylate pathway, (3) from oxaloacetate through the reductive branch of the TCA cycle under anaerobic conditions, and (4) from L-glutamate and L-arginine via the g-aminobutyrate catabolic pathway. 14 To block succinate synthesis in E. coli K-12 MG1655, both the sucA-encoding submit of a-ketoglutarate dehydrogenase and aceA-encoding isocitrate lyase were deleted, resulting in E. coli DsucADaceA. The plasmid pWSK-ido was subsequently introduced into E. coli DsucADaceA, and E.coliDsucADaceA-ido was obtained. E. coli DsucADaceA could not grow in M9 medium without the addition of succinate during aerobic cultivation although the g-aminobutyrate catabolic pathway was still retained (Fig. 1A). The effect of succinate addition on the growth of E.col-iDsucADaceA was detected, and the biomass of E.col-iDsucADaceA was enhanced with the increased addition of succinate. Notably, E.coliDsucADaceAido induced by IPTG restored growth in M9 medium supplemented with L-Ile and a-KGA, indicating that IDO activity shunts the blocked TCA cycle and restores cell growth coupling with L-Ile hydroxylation (Fig. 1B).
4-HIL accumulation was detected in culture broth during cell growth. The biomass (Fig. 1B) and 4-HIL accumulation(data not shown) were enhanced with the increase in IPTG addition. Because 4-HIL accumulation is coupled with succinate synthesis, it was deduced that the higher biomass was due to increased succinate synthesis catalyzed by the enhanced expression of ido induced by an increased IPTG concentration.

Screening the IDO variants with improved activities
Because IDO can couple L-Ile hydroxylation and E.coliDsucADaceA cell growth and because the biomass was dependent on the level of succinate synthesis, one would expect that a higher IDO activity would correspond to a greater supply of succinate and larger colonies of strains expressing ido on the plates. Error-prone PCR was performed with ido as a template, and the mutated gene products were subsequently ligated into the vector pWSK29. The recombinant plasmids were transformed into E. coliDsucADaceA to obtain a mutant library. After the screening of approximately 9,000 clones, 5 were found to show considerably larger colonies than the wild type. The 5 clones were collected, and 4 clones repeatedly showed increased 4-HIL production compared with the strains with wild-type IDO, indicating improved IDO activity (Fig. 2).
Sequence analysis of IDO M3 , which exhibited the highest IDO activity, showed that the residues Leu27, Glu80, Gly169 and Ser182 were substituted for Ile27, Asp80, His169 and Asp182, respectively. The active site structures of a-KGA-dependent hydroxylase exhibit nearly identical arrangements of the 3 amino acid side chains, a His 1 -X-Asp/Glu-Xn-His 2 motif. 13 As a member of the superfamily, IDO also has the only His 159 -X-Asp 161 -X 50 -His 212 motif, which might be the active site. Thus, one could reason that the 2 mutated amino acid residues(Gly169 and Ser182) located in the motif resulted in the improved activity of IDO M3 .

Expression and characterization of IDO M3
Ido M3 was expressed by E. coli BL-ido M3 under the induction of IPTG, and recombinant IDO M3 was purified by a Ni 2C -NTA affinity column. The concentrations of the purified recombinant IDO and IDO M3 were 5.1 and 4.2 mg/mL, respectively. To calculate the K m , k cat and V max values of IDO and IDO M3 , the activities of the 2 enzymes toward different concentrations of L-Ile were measured, and the data were plotted according to the Michaelis-Menten equation. As shown in Table 1, K m was lower, but V max and K cat were higher for IDO M3 than IDO. The catalytic efficiency k cat /K m was approximately 1.5-fold higher for IDO M3 than for IDO under the measured conditions.
The optimum temperature for the activity of IDO M3 and IDO was determined within a range of 10 C to 60 C. As shown in Fig. 3A, the optimum temperature for both enzymes was approximately 35 C. Thermal stability assays showed that both IDO and IDO M3 were stable below 40 C, while the activities decreased dramatically when the temperature was above 40 C (Fig. 3B). More than 60% of the maximum activity was observed after incubation at 60 C for 30 min (Fig. 3C). Notably, IDO M3 exhibited a slightly higher stability than IDO. After incubation at 60 C for 40 min, 53% of IDO M3 activity was retained, compared with 40% of IDO activity (Fig. 3C). The mutated amino acid residues at sites 27 and 80 may be responsible for the improved thermal stability. The optimum pH for the activity of IDO M3 and IDO was determined over a pH range of 3.0 to 11.0. The activity of the 2 enzymes rose with the increased pH until 7.0 and then decreased, with no difference in the optimum pH for the 2 enzymes (Fig. 3D).

4-HIL synthesis by resting cells expressing IDO M3
Under industrial application conditions, purified enzymes are not a suitable enzyme source. Biotransformation by E. coli harboring ido seems to be an appropriate method for 4-HIL synthesis. However, 50 nmol of glucose (150 nmol in total) was additionally consumed for cell growth rather than for 4-HIL synthesis during process biotransformation. 15 Moreover, additional consumption of L-Ile and a-KGA was also detected in our previous study(data not shown). One attractive solution would be to use resting cells, a strategy that has been successfully used in nitrile hydrolase  biocatalysis. 16 However, the cell membrane seems to be the main limitation for the uptake of L-Ile and a-KGA and the export of 4-HIL. Thus, to improve the cell permeability of the substrate and products, harvested E. coli BL21(DE3) cells expressing IDO M3 were frozen at ¡80 C to obtain resting cells, and the strategy for 4-HIL synthesis was evaluated. As shown in Fig. 4, the resting cells could successfully transform L-Ile to 4-HIL, while intact cells that had not been frozen were not successful.   least 4 times while exhibiting almost the same yield (data not shown).

Strains and plasmids
The strains and plasmids used in this work are listed in Table 2.

Primers
The primers used in this work are listed in Table 3.

Construction of E. coli DsucADaceA
Knockout of sucA and aceA was performed by PCR-based λ-red recombination as described previously. 17 A 1114-bp DNA fragment containing the chloramphenicol resistance gene cassette from pKD3 was amplified using primers sucA-1 and sucA-2.The PCR products were electroporated into E. coli K-12 MG1655 carrying the λ-red recombinase expression plasmid pKD46. Cells in which homologous recombination occurred were selected on an agar plate containing chloramphenicol and were identified by direct colony PCR using primers sucA-3 and sucA-4. The antibiotic marker was eliminated using the helper plasmid pCP20. The resulting strain was denoted E. coli DsucA. aceA in E. coli DsucA was deleted with the same method, resulting in E. coli DsucADaceA.

Construction of an ido mutant library and ido overexpression strain
A mutant library of the ido gene was constructed by error-prone PCR from pET-ido using an instant error-prone PCR kit(Tiandz, Inc.; 101005) and the  primers ido-1 and ido-2. The mutated gene products were consequently ligated into the vector pWSK29, which had been digested by Xba I via the ClonEx-pressTM II One Step Cloning Kit(Vazyme Biotech Co., Ltd; C11201), using a homologous recombinase. The recombinant plasmids were transformed into E. coli DsucADaceA, and the cells were grown on plates containing M9 medium(glycerol as a carbon source) supplemented with L-Ile and a-KGA(1 g/L) as well as ampicillin(100 mg/L) and IPTG(0.05 mM). The ido gene was amplified from pWSK-ido or pWSK-ido M3 by PCR using primers ido-3 and ido-4. The obtained PCR products were digested by BamH I and Hind III and were cloned into the expression vector pET-His. The recombinant plasmids(pET-IDO and pET-IDO M3 ) were further transformed into competent E. coli BL21(DE3) cells, and the cells were selected on LB plates containing 100 mg/L ampicillin, resulting in BL-ido and BL-ido M3 .

Screening IDO mutants with improved activity
The clones exhibiting larger colonies than the wild type were transferred to 24-well microplates with 800 mL of M9 medium supplemented with L-Ile (1 g/L), a-KGA(1 g/L) and IPTG(0.05 mM). After incubation at 37 C for 24 h, 40 mL of the culture in each well was then inoculated into the corresponding well of another 24-well microplate with the same medium and was then cultured for another 24 h. The biomass(OD 600 )was detected, and variants exhibiting a higher biomass were chosen for further analysis.
Biotransformation of 4-HIL by E. coli DsucADaceA harboring ido variants and resting cells of BL-ido M3 E. coli DsucADaceA cells harboring ido variants were cultured in 24-well microplates with M9 medium supplemented with L-Ile(1 g/L), a-KGA(1 g/L) and IPTG (0.05 mM) at 37 C for 24 h, and then 0.1 mL of the culture broth was transferred to another 24-well microplate with the same medium, was cultured for another 24 h and was centrifuged at 4 C for 5 min. Five hundred microliters of supernatant was collected for a quantitative analysis of 4-HIL.
BL-ido M3 cells were harvested after induction by IPTG for 6 h and were frozen at ¡80 C for 2 h to obtainresting cells. The cells (20 g) were suspended in a 1-L shake flask with 200 mL of Tris-HCl (100 mM, pH 7.0) containing 160 mM L-Ile and a-KGA, 5 mM FeSO 4 and 10 mM ascorbic acid; the cells were thenshaken at 37 C at 200 rpm for 16 h.

Protein expression and purification
IPTG (0.1 mM) was added when the BL-ido and BLido M3 cells were grown in LB medium(with 100 mg/L ampicillin) to the midexponential stage. The cells were harvested by centrifuging and broken with sonication after cultivation for another 4 h. The expression of recombinant IDOs was confirmed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). The 6£His-tagged IDOs were purified using a Ni 2C -NTA affinity column. The concentration of the purified proteins was determined by the Bradford method using bovine serum albumin as a standard.

Characterization of IDO and IDO M3
For a catalytic characterization of IDO and IDO M3 , the reaction mixture composed of 10 mM a-KGA and L-Ile, 5 mM FeSO 4 , 10 mM ascorbic acid and 100 mM Tris-HCl(pH 7.0) was combined with 0.5 mg/mL of purified recombinant IDO, and the reaction was performed at 30 C for 30 min. The enzymatic activity was determined by the production of 4-HIL, as measured by high-performance liquid chromatography (HPLC).
The optimal temperatures for IDO and IDO M3 were determined by evaluating the activities at temperatures ranging from 10 C to 70 C. The thermal stability of the enzymes was determined by assessing their residual activities after incubation at temperatures from 10 C to 60 C for 1 h. The time-course thermal stability of the enzymes at 60 C was determined by evaluating their residual activities after incubation for 0 min, 10 min, 20 min, 30 min, 40 min, 60 min, 90 min and 120 min.
To determine the K m values, L-Ile was used at concentrations of 0.02 to 5 mM. To examine the pH dependency of the reaction, citric acid sodium-citrate buffer (pH 3.0-5.0), phosphate-buffered saline (pH 6.0-8.0), Tris-HCl (pH 9.0) and sodium bicarbonatesodium carbonate buffer (pH 10.0-11.0) were used.

Analytical determination of 4-HIL concentration by HPLC
Four-HIL was analyzed by precolumn derivatization using 2, 4-fluoro-dinitrobenzene and was detected by HPLC using an Agilent C18 column (150 mm £ 4.6 mm, 3.5 mm). Elution was performed using a gradient of 50% acetonitrile (v/v)/50mM (CH 3 COONa), fed at a constant flow rate of 1.0 mL/min. UV absorption was measured at 360 nm, and the column temperature was maintained at 33 C.

Conclusions
A strategy for the selection of IDO with improved activity was developed based on the coupling of L-Ile hydroxylation, the oxidation of a-KGA to succinate and cell growth. Five mutants were obtained using this strategy. The catalytic efficiency, thermal stability, and catalytic rate of the IDO M3 variant selected by the strategy were significantly improved compared with those of wild-type IDO. A method for 4-HIL biotransformation by resting cells expressing IDO M3 was developedand was shown to be suitable for 4-HIL synthesis with substrates only seldom being subjected to additional consumption.

Disclosure of potential conflicts of interest
No potential conflicts of interest were disclosed.