Efficient Extracellular Expression of Metalloprotease for Z‑AspartameSynthesisFucheng Zhu,† Feng Liu,† Bin Wu,† and Bingfang He*,†,‡

†College of Biotechnology and Pharmaceutical Engineering and ‡School of Pharmaceutical Sciences, Nanjing Tech University, No. 30Puzhu South Road, Nanjing 211816, China

*S Supporting Information

ABSTRACT: Metalloprotease PT121 and its mutant Y114S (Tyr114 was substituted to Ser) are effective catalysts for thesynthesis of Z-aspartame (Z-APM). This study presents the selection of a suitable signal peptide for improving expression andextracellular secretion of proteases PT121 and Y114S by Escherichia coli. Co-inducers containing IPTG and arabinose were usedto promote protease production and cell growth. Under optimal conditions, the expression levels of PT121 and Y114S reached>500 mg/L, and the extracellular activity of PT121/Y114S accounted for 87/82% of the total activity of proteases. Surprisingly,purer protein was obtained in the supernatant, because arabinose reduced cell membrane permeability, avoiding cell lysis.Comparison of Z-APM synthesis and caseinolysis between proteases PT121 and Y114S showed that mutant Y114S presentedremarkably higher activity of Z-APM synthesis and considerably lower activity of caseinolysis. The significant difference insubstrate specificity renders these enzymes promising biocatalysts.

KEYWORDS: extracellular expression, metalloprotease, Escherichia coli, arabinose, Z-aspartame

■ INTRODUCTION

Aspartame (L-aspartyl-L-phenylalanine methyl ester, APM) is aprotected dipeptide sweetener, which is 200 times sweeter thansucrose.1 Since its approval, APM has been widely used as alow-calorie sweetener in soft drinks and food products.2 Morethan 19,000 t of APM are produced every year, making it themost highly synthesized peptide in the world.3 APM can beeasily obtained by deprotection of carboxybenzyl from Z-aspartame (N-carbobenzoxy-L-aspartyl-L-phenylalanine methylester, Z-APM).4 Members of metalloprotease family M4 (alsoknown as the thermolysin family) such as thermolysin andpseudolysin, as efficient biocatalysts, were utilized to catalyzethe synthesis of Z-APM from N-carbobenzoxy-L-aspartic acid(Cbz-Asp) and L-phenylalanine methyl ester (Phe-OMe).5 Aprevious study developed mutant Y114S, which can efficientlycatalyze the synthesis of Z-APM, by site-directed mutagenesisof protease PT121 from Pseudomonas aeruginosa.6 However,due to the limited expression level, these metalloproteases areone of the limiting factors in Z-APM enzymatic preparation.Various alkaline proteases have been produced at industrial

scale.7 However, to the best of our knowledge, the expressionlevels of many thermolysin family metalloproteases, such asthermolysin-like protease and pseudolysin, are still limited. InBacillus subtilis and Pichia pastoris hosts, extracellular expressionof thermolysin family proteases can reach 10−50 mg/L.8−10 Asthe most commonly used host, Escherichia coli is also utilized toexpress thermolysin family metalloproteases, but most of theseenzymes are expressed either in intracellular form, ranging from6 to 20 mg/L,11−13 or in inclusion body form.14 Because of itshigh proteolysis, mature proteases located inside cells couldcause recombinant cell growth inhibition, or even cell lysis.Targeting the protein to the culture medium may overcomethis obstacle.15 Additionally, extracellular expression mayfacilitate downstream processing and achieve higher produc-

tion.16 Previously, we have demonstrated the expression ofmetalloprotease PT121 and its mutant Y114S with improvedactivity of Z-APM synthesis.6 However, expressed PT121proteases were mainly obtained as intracellular form and theexpression level remained low. Therefore, exploring an effectiveapproach for improving extracellular expression is necessary.Many approaches have been reported to facilitate extrac-

ellular secretion of proteins in E. coli. These studies haveincluded almost every aspect of the secretion process, such asutilization of different secretion pathways,17 construction ofleaky strains,18 and coexpression of the key secretioncomponents,19 as well as optimization of culture conditions.20

Nevertheless, extracellular expression of the recombinantproteins in E. coli is still a daunting and challenging task.Signal peptide plays an important role in the expression andsecretion of proteins.21 The nucleotide sequence encodingsignal peptide affects the 5′mRNA secondary structure of thetranslation initiation region (TIR), which was confirmed to playa crucial role in the efficiency of gene expression.22 The stabilityof mRNA secondary structure is quantified as the value ofminimal fold free energy (ΔG). By increasing the ΔG of TIRfrom position −35 to +36, the human tumor necrosis factor αand extracellular domain of Her2/neu protein were overex-pressed in E. coli.23 Moreover, because the expressed proteasehas a toxic effect on cells due to high proteolytic activity,promoting cell growth is another key factor for proteaseexpression. Increasing evidence indicates that arabinosepromotes cell growth for overexpression of penicillin acylase.24

Xu et al. reported that arabinose efficiently reduced the amount

Received: September 19, 2016Revised: December 2, 2016Accepted: December 2, 2016Published: December 2, 2016

Article

pubs.acs.org/JAFC

© 2016 American Chemical Society 9631 DOI: 10.1021/acs.jafc.6b04164J. Agric. Food Chem. 2016, 64, 9631−9638

of inclusion bodies and significantly improved the activity ofpenicillin acylase.25 To our best knowledge, no study hasreported the arabinose-induced expression of protease.In the present study, replacement of signal peptide was

applied to promote the extracellular expression of metal-loprotease PT121 and its mutant Y114S. Arabinose supple-mentation significantly increased the expression level ofproteases and improved the purity of secreted proteases.Possible mechanisms are also discussed. Additionally, thesubstrate specificity of PT121 and Y114S in the catalyzedsynthesis of Z-APM and caseinolytic activity was investigatedand rationalized by assessing kinetic parameters and conductingmolecular dynamic simulation.

■ MATERIALS AND METHODSChemicals, Reagents, and Materials. N-Carbobenzoxy-L-

aspartic acid (Cbz-Asp) and L-phenylalanine methyl ester (Phe-OMe) were obtained from GL Biochem Co., Ltd. (GLS, Shanghai,China). o-Nitrophenyl-β-D-galactopyranoside (ONPG) and N-phenyl-α-naphthylamine (NPN) were purchased from Aladdin Co., Ltd.(Aladdin, Shanghai, China). All primers were synthesized at Invitrogen(Invitrogen, Shanghai, China). DNA polymerase, restriction enzymes,and T4 DNA ligase were purchased from Takara (Takara, Dalian,China).Molecular Techniques. All primers used in this study are shown

in Table S1. Genomic DNA of P. aeruginosa PT121 was used astemplate for PCR.26 The 1497 bp fragment of gene lasB was amplifiedwith cds-F and cds-R as the nucleotide primers. The amplified DNAproduct was inserted into the pMD-18T vector (Takara). Therecombinant plasmid was used as a template for PCR. The 1425 bpfragment containing the gene encoding the pro-peptide and maturepeptide was amplified using the primers lasB-F1 and lasB-R1. Thus,the original pelB signal sequence of pET-22b (Takara) was retained.The product of PCR was digested with restriction enzymes NcoI andBamHI and then ligated into pET-22b. Primers lasB-F2 and lasB-R2were used to obtain the DNA product with its native signal peptide byPCR. The product was digested with the restriction enzymes NdeI andBamHI and then ligated into pET-22b. As a result, the native signalpeptide of metalloprotease PT121 was retained. The nucleotidesequence of signal peptide ompA was fused to the lasB gene withoutits native signal peptide by overlap PCR using primer-rF1, primer-rR1,primer-rF2, and primer-rR2. Then the constructed fragment wascloned into vector pET-22b. The recombinant pET-22b harboring theinserted gene sequence was transformed into E. coli BL21 (DE3).Mutant Y114S was generated using a QuikChange site-directedmutagenesis kit (Stratagene, La Jolla, CA, USA), and the recombinantplasmid was transformed into E. coli BL21. The correct clones wereselected for DNA sequencing (GenScript, Nanjing, China).Medium and Culture Conditions. The transformed BL21 was

inoculated in 5 mL of LB medium and grown with shaking at 37 °Covernight to serve as the seed. A 2% (v/v) concentration of theinoculum was inoculated in 50 mL of LB medium containing 100 μg/mL ampicillin and then incubated at 37 °C until an optical density at600 nm (OD600) reached 1.0. Then, 10 μM isopropyl thio-β-D-galactopyranoside (IPTG) was added to induce protein expression at30 °C for 24 h. The E. coli transformed with pET-22b was taken as thecontrol group for analysis.The pooled supernatant was dialyzed against 50 mM Tris-HCl (pH

8.0) buffer to remove salt and other small molecules, and then thesample was submitted for assessment of its casein hydrolysis activity.The purity and molecular weight of the target protein were measuredusing SDS-PAGE with 12.5% separating gel as described byLaemmli.27

Purification of Protease. The protease purification wasperformed as described by Tang et al.28 with slight modification.The culture supernatant was harvested by centrifugation at 12000g and4 °C for 10 min and loaded onto a Phenyl Sepharose column (GEHealthcare, Uppsala, Sweden), which was equilibrated with 50 mM

Tris-HCl (pH 8.0) containing 1.0 M NaCl. After binding, the columnwas eluted with 50 mM Tris-HCl (pH 8.0) at a flow rate of 1.0 mL/min. The absorbance of eluent at 280 nm was measured, and the activefractions were collected and dialyzed for further research. Bradford’smethod was used for assaying the protein concentration.

Assay of Caseinolytic Activity. The activity of the supernatanttoward casein was measured as previously described.26 Two millilitersof the diluted protease was added to 2 mL of Tris-HCl buffer (50 mM,pH 8.0) containing 2% (w/v) casein. The reaction mixture wasincubated at 40 °C for 10 min and terminated by adding 4.0 mL ofTCA mixture containing trichloroacetic acid (0.11 M), sodium acetate(0.22 M), and acetic acid (0.33 M). The mixture was then centrifugedat 12000g for 10 min. The absorbance at 280 nm of the reactionsupernatant was measured against a blank control. The amount ofenzyme that hydrolyzes casein and produces 1 μg of tyrosine perminute is defined as 1 unit (U) of protease activity. All above assayswere performed in triplicates for calculating the mean and standarddeviation.

Effect of IPTG and Arabinose on Expression of Proteases. Toexamine the effect of IPTG and arabinose on the expression ofproteases PT121 and Y114S, IPTG (0−100 μM) and arabinose (0−5.0 mg/mL) were applied to induce the expression of proteases,respectively. The flask cultures were further shaken at 30 °C foranother 24 h. Then, the cultures were centrifuged at 4 °C and 12000gfor 10 min. The supernatant of culture was assayed for extracellularcaseinolytic activity. The cell pellet was resuspended in Tris-HCl buffer(50 mM, pH 8.0). The cell suspension was sonicated for 5 min byusing an ultrasonic processor and then centrifuged at 12000g and 4 °Cfor 5 min. The supernatant of cell extracts was assayed for intracellularprotease activity. All above assays were performed in triplicates forcalculating the mean and standard deviation.

Determination of Cell Lysis. Cell lysis was determined as thepercentage of extracellular activity versus the total activity containingextracellular and intracellular portion by using β-galactosidase as thereporter protein.18,29 Activity of β-galactosidase was assayed aspreviously reported.30

Assay of Membrane Permeability. Permeability of the innermembrane was determined by measuring the influx of ONPG, asubstrate of cytosolic β-galactosidase. β-Galactosidase localized withinthe cytoplasm can hydrolyze ONPG that passes the inner membrane,resulting in absorbance at 420 nm. Thus, an increase in the rate ofabsorbance indicates the permeability of the inner membrane. In thisstudy, the permeability of the inner membrane was assayed bymeasuring the access of ONPG to the cytoplasm in accordance with apreviously reported method with minor modifications.31 Therecombinant E. coli BL21 containing the constructed plasmid wasobtained by centrifugation (12000g), rinsed once, and then suspendedin 10 mM sodium phosphate buffer (pH 7.4) to an OD600 of 0.15.ONPG with a final concentration of 100 μg/mL was added to anELISA plate containing 200 μL of the cell suspension. Substratecleavage by β-galactosidase was measured by light absorption every 3min at 420 nm in a microplate reader (BioTek, Winooski, VT, USA).

The hydrophobic fluorescent probe NPN was used as an indicatorof outer membrane integrity. NPN has a low fluorescence quantumyield in aqueous solution, but its fluorescence is strong in thehydrophobic environment of cell membrane. Normally, NPN isexcluded from the lipopolysaccharide layer of the outer membrane butcan enter at points where membrane integrity is compromised. Thus,the permeability of the outer membrane can be indicated byfluorescence value and its increased rate. Outer membranepermeability was measured according to a previous method withminor modifications.31 NPN with a final concentration of 10 μM wasadded to a quartz cuvette containing 3 mL of cell suspension.Fluorescence absorption was assayed using a spectrofluorometerShimadzu RF-1501 (Shimadzu, Kyoto, Japan) with 5 nm slit widths.The emission and excitation wavelengths were set to 420 and 350 nm,respectively.

Kinetic Constants of Recombinant PT121 and Y114S. Theinitial rate of tyrosine formation, presented as the initial rate ofcaseinolysis, was determined at a fixed concentration of purified

Journal of Agricultural and Food Chemistry Article

DOI: 10.1021/acs.jafc.6b04164J. Agric. Food Chem. 2016, 64, 9631−9638

9632

enzyme (0.1 mg/mL) and at different concentrations of casein (0.5−10.0 mg/mL) for 5 min. All above assays were performed in triplicatesfor calculating the mean and standard deviation. The determined valuewas plotted by nonlinear fitting using Origin 9.0.To measure the kinetic parameters of the synthetic reaction for Z-

APM, various Cbz-Asp and Phe-OMe concentrations of 10−100 mMat fixed concentrations of 500 mM Phe-OMe and 100 mM Cbz-Aspwere used as substrates, respectively. Purified enzyme (0.9 μM) wasadded into the reaction mixture, and the reaction system was shaken at37 °C and 180 rpm for 1 h. All samples were analyzed by high-performance liquid chromatography (HPLC) as previously reported.6

The kinetic constants KM were calculated from Lineweaver−Burkplots.We calculated the KM1 of Cbz-Asp (ZD) and the KM2 of Phe-OMe

(FM), respectively. Random rapid equilibration mechanism wasapplied to calculate the kinetic constants of Z-APM synthesisaccording to a previous method.32 The initial reaction rate v0 of Z-APM synthesis was obtained at 100 mM Cbz-Asp and 500 mM Phe-OMe. The kcat would be obtained by the following equation:

=+ + +

vk

K K K K[E] [ZD] [FM]

[FM] [ZD] [ZD] [FM]0cat 0 0 0

M1 M2 M1 0 M2 0 0 0

RNA Secondary Structure Prediction. The secondary structureand Gibbs free energy (ΔG) of the mRNA TIR were predicted andcalculated using an online server (http://mfold.rna.albany.edu/?q=mfold/RNA-Folding-Form).Molecular Simulation. The crystal structure (PDB code: 1EZM)

determined by Thayer et al.33 was used for the molecular modeling ofPT121 and Y114S. The 3D structure of Phe-OMe was stripped from

the crystal structure of thermolysin (PDB code: 3QGO). Moleculardocking of the substrate was performed using LeDock (http://lephar.com). The software in optimizations of the ligand pose (position andorientation) and its rotatable bonds is based on simulated annealing aswell as genetic algorithm.34 All crystal waters were removed from thecrystal structure, and polar hydrogen atoms were added to thestructure of protein by LePro module in the docking suite LeDock. Allcatalytic residues were included in the active pocket. Moleculardynamic simulation was performed using Amber 12 with the Amber99SB force field.35 Ligand topology was generated using the LeapProgram. Sodium ions were added to maintain system neutrality. Theneutral system was subjected to energy minimization with the steepestdescent of 1000 steps and conjugate gradient of 2000 steps. Then, the50,000 steps position-restrained dynamic simulation was applied to thesystem at 300 K. Finally, MD was performed for 40 ns at 300 K.Postprocessing and analysis were performed using standard Ambertools and customized scripts.

■ RESULTS AND DISCUSSION

Effect of Signal Peptides on Expression of PT121 andY114S. The effect of signal peptides (pelB, ompA, and itsnative signal peptide) on the expression of PT121 and Y114Swas studied. The recombinant E. coli cells harboring Sp-PT121(BL21/Sp-PT121), ompA-PT121 (BL21/ompA-PT121), pelB-PT121 (BL21/pelB-PT121), and pelB-Y114S (BL21/pelB-Y114S) were induced for 24 h. In Figure 1, SDS-PAGE showsthat most of the proteases PT121 and Y114S wereextracellularly expressed by using pelB or ompA as signal

Figure 1. (A) SDS-PAGE analysis of proteases in the culture supernatant and cell-extracted supernatant. Lanes: M, protein marker; 1−10,extracellular fraction and intracellular fraction of BL21/pET22b (lanes 1 and 2), BL21/Sp-PT121 (lanes 3 and 4), BL21/pelB-PT121 (lanes 5 and6), BL21/pelB-Y114S (lanes 7 and 8), and BL21/ompA-PT121 (lanes 9 and 10). (B) Extracellular and intracellular activities of BL21/pET22b,BL21/Sp-PT121, BL21/pelB-PT121, BL21/pelB-Y114S, and BL21/ompA-PT121. Values are the mean ± standard deviation from threeindependent experiments.

Journal of Agricultural and Food Chemistry Article

DOI: 10.1021/acs.jafc.6b04164J. Agric. Food Chem. 2016, 64, 9631−9638

9633

peptide, whereas most of the target proteins were intracellularlyexpressed for Sp-PT121 constructs. The culture supernatantprotease activity of recombinant BL21/Sp-PT121 reached only61.6 ± 9.6 U/mL, which accounted for 10% of the total activity.By comparison, the signal peptides pelB/ompA effectivelypromoted the extracellular secretion of protease PT121, andthe highest expression of protease PT121 using pelB as signalpeptide was observed as shown in Figure 1. The proteaseactivities of pelB-PT121 and pelB-Y114S in the culturesupernatant reached 5200 ± 230 and 2700 ± 170 U/mL,respectively, which constituted 91.5 and 88.4% of the totalactivity. Even though the concentrations of PT121 and Y114Sin supernatant were almost the same, the caseinolytic activity ofY114S was only half that of the PT121. This might be due tothe difference in substrate specificity between proteases PT121and Y114S.Signal peptide not only promoted extracellular secretion but

also affected the 5′mRNA secondary structure. In the presentstudy, the secondary structures of the mRNA TIR (fromnucleotide points −35 to +36) of three constructs named TIR-pelB (pelB signal peptide), TIR-ompA (ompA signal peptide),and TIR-Sp (native signal peptide of protease PT121) werepredicted theoretically. The essential architecture of the TIR isshown in Figure 2. The ΔG values of mRNA TIR in TIR-pelB,TIR-ompA, and TIR-Sp were −9.5, − 9.7, and −10.5 kcal/mol,respectively. Previous studies demonstrated that increasing theΔG of TIR increases protein expression.36,37 In addition, thestart AUG codon was demonstrated as a key limiting factor forprotein expression in E. coli.38 The AUG start codons in TIR-pelB and TIR-ompA were fully exposed, compared with that in

TIR-Sp. The exposure of initiator AUG was speculated toreduce the blockade element for gene translation and toimprove protein expression. Therefore, the pelB signal peptideswere more suitable for secretory expression of proteases PT121and Y114S. To our knowledge, this study is the first report onthe extracellular expression of thermolysin-like protease fromP. aeruginosa in an E. coli system. The strategies describedherein may be feasible for rational selection of a suitable signalpeptide for efficient protein expression.

Effect of Inducers (IPTG and Arabinose) on Expressionof Protease PT121. Mutant Y114S exerts little effect onprotein expression. Thus, protease PT121 was selected as amodel to investigate. IPTG is commonly used as an inducer ofprotein expression regulated by T7 promoter, and arabinosewas also able to induce penicillin acylase expression regulatedby T7 promoter in E. coli.25 The effects of IPTG and arabinoseon the expression of protease PT121 were investigated.As the concentration of IPTG increased, the expression level

increased, and a maximum value was obtained at 10 μM (Table1). Further increase of IPTG concentration beyond 10 μMsignificantly decreased the expression level of protease. On theother hand, the cell density (OD600) significantly decreasedeven when 5 μM IPTG was added. High concentrations ofIPTG negatively influenced the expression of the target protein.This result may be attributed to the fact that overexpression oftarget protease inhibits cell growth or even causes cell lysis.39,40

Thus, lowering the concentration of IPTG could promoteprotease expression and cell growth.Arabinose has been applied to induce the expression of

penicillin acylase,24 but whether arabinose can induce protease

Figure 2. Predicted secondary structure and minimum free energy of the 5′mRNA translation initiation region (TIR), including 36 bases upstreamand 35 bases downstream of the start codon. The start codon AUG is labeled in red.

Table 1. Expression of Protease PT121 under Various Inducer Conditionsa

IPTG (μM) arabinose (g/L)

inducer 0 5 10 50 100 0.5 1 2 3 4

protease activity 210 ± 6 4900 ± 150 5600 ± 270 4800 ± 150 4100 ± 120 550 ± 17 660 ± 18 1500 ± 45 2200 ± 67 670 ± 17

OD600 5.4 ± 0.2 2.7 ± 0.2 2.6 ± 0.3 2.6 ± 0.2 2.5 ± 0.3 6.8 ± 0.3 7.2 ± 0.3 6.0 ± 0.3 5.3 ± 0.2 5.0 ± 0.1

proteaseactivity/OD600

39 ± 1.2 1800 ± 50 2100 ± 60 1800 ± 50 1600 ± 30 81 ± 2.5 91 ± 3.2 250 ± 6 420 ± 13 130 ± 3.9

aExperiments were performed in triplicate and are presented as the mean ± standard deviation.

Journal of Agricultural and Food Chemistry Article

DOI: 10.1021/acs.jafc.6b04164J. Agric. Food Chem. 2016, 64, 9631−9638

9634

expression is unclear. In this study, the effects of arabinoseconcentration on cell growth and protease production wereinvestigated (Table 1). The extracellular protease activity andcell density initially increased with increasing arabinoseconcentration and then decreased when this concentrationexceeded 3.0 g/L. The maximum protease activity achieved was2200 U/mL. Although the protease activity in the arabinose-induced culture was lower than that in the IPTG-inducedculture, the cell density was >2-fold that when IPTG was usedas an inducer. The result indicates that arabinose not onlyenhances cell growth but also induces protease PT121expression.Conventionally, cell growth is severely inhibited by protease

expression because of its high proteolytic activity.41 Toovercome these deficiencies, arabinose and low-concentrationIPTG (10 μM) were used as co-inducers for proteaseexpression. As shown in Figure 3, the supernatant activity

and total activity increased with increasing arabinose concen-tration. The optimum arabinose concentration was 3.0 mg/mL,and the total activity reached 11900 ± 500 U/mL, which was2.1-fold that of the arabinose-free culture. Significantly, theextracellular protease reached 10300 ± 460 U/mL (468 mg/L).Similarly, the total activity/OD600 increased when the arabinoseconcentration was increased from 0.0 to 3.0 mg/mL, butremarkably decreased when the concentration was furtherincreased to 4 mg/mL. The maximum total activity/OD600 wasachieved at 3.0 mg/mL arabinose and was 1.7-fold that withoutarabinose addition. The OD600 was 1.2-fold that withoutarabinose addition. The production of mutant Y114S was alsodetermined under the same conditions. The total andextracellular activities of Y114S reached 6600 ± 240 and5400 ± 200 U/mL, respectively, both of which were almost 2.0-fold that of the arabinose-free cultivation. Interestingly, SDS-PAGE showed that the proteases in the culture supernatantwith added 3.0 g/L arabinose were purer than that in thearabinose-free culture (Figure 4). In particular, the specificactivity of protease PT121 in culture supernatant with 3.0 g/Larabinose was 22,000 U/mg, which was almost equal to that ofpurified native protease PT121.28 These results suggest thatarabinose may prevent cell lysis by decreasing the leakage ofintracellular proteins, thus improving the purity of target

protease in the culture supernatant. Arabinose enhances theproduction of penicillin acylase by mediating the processingand correct folding of this enzyme.25,42 However, the decreaseof the leakage of intracellular proteins in E. coli by arabinose hasnot been reported.In early papers, the full length of the lasB gene with the

native signal sequence integrated into vector pUC19 wasintracellularly expressed in E. coli JM109 with 23.7 U/mL.13

Odunuga et al. cloned the mature pseudolysin coding sequenceof lasB gene into pET28a and achieved the yield of 40 mg/L inE. coli BL21 as inclusion bodies followed by refolding.43 InP. pastoris, P. aeruginosa elastase was successfully expressed andsecreted into cultures with 330 U/mL, and the expressedprotease was glycosylated.10 In this study, addition of arabinosesignificantly promoted expression of protease PT121 and keptcells from leakage. High-purity extracellular protein wouldconsiderably simplify the process of purification. This studymay provide a useful strategy in expressing proteases.

Determination of Cell Lysis and Membrane Perme-ability. Extracellular expression of protease may causehydrolysis of membrane proteins, causing cell lysis. To explorethis possibility, β-galactosidase was taken as the reporterprotein. β-Galactosidase is a cytoplasmic protein, and its level inthe culture medium is very low under normal conditions; thus,the presence of extracellular β-galactosidase activity would bean indication of cell lysis.18,29 In this study, 91.5% of the totalexpressed protease PT121 and 13.6% of the total expressed β-galactosidase were found in the arabinose-free culture medium.However, 86.9% of the total expressed protease PT121 and<1% of the total β-galactosidase were found in the arabinose-supplemented culture medium, and this amount of β-galactosidase was similar to that of the control E. coli cellsharboring pET-22b (BL21/pET22b) without the target geneinserted. The addition of arabinose reduced the percentage ofβ-galactosidase in the culture supernatant, but the percentage oftarget protease in the culture supernatant remained almost thesame. These results suggest that the presence of extracellularprotease is not due to cell lysis. Additionally, supplementationof arabinose could prevent cell lysis.We analyzed cell membrane permeability to further explore

the mechanism by which arabinose prevents cell lysis and

Figure 3. Synergistic effect of IPTG and arabinose on protease PT121production: extracellular activity (cross-hatched bars); total activity(black bars); total activity/OD600 (light gray bars); pH of supernatant(slashed bars). Results are expressed as the mean ± standard deviationfrom three independent experiments.

Figure 4. SDS-PAGE analysis of proteins in the culture supernatant.Lanes: M, protein marker; 1 and 3, culture supernatant of BL21/pelB-PT121 and BL21/pelB-Y114S induced by IPTG; 2 and 4, culturesupernatant of BL21/pelB-PT121 and BL21/pelB-Y114S induced byIPTG and arabinose.

Journal of Agricultural and Food Chemistry Article

DOI: 10.1021/acs.jafc.6b04164J. Agric. Food Chem. 2016, 64, 9631−9638

9635

increases the purity of the recombinant enzyme. As shown inFigure 5, under the arabinose-free supplemented culture, the

inner and outer membrane permeabilities of BL21/pelB-PT121were significantly higher than those of the control BL21/pET22b, which may be the result of hydrolysis by expressedprotease. However, when BL21/pelB-PT121 was cultured inarabinose-supplemented LB medium at the inductive phase,inner membrane permeability significantly decreased to a leveleven lower than that of BL21/pET22b in the culture withoutaddition of arabinose. The outer membrane permeability alsodecreased when arabinose was added. However, the outermembrane maintained its higher permeability even whenarabinose was added. Proteases of the thermolysin family arefolded and matured in periplasmic space, and then a correctlyfolded mature protease is secreted.44 The higher permeability ofthe outer membrane promoted the secretion of therecombinant protease located in periplasmic space into culture.The inner membrane maintained its lower permeability withthe addition of arabinose, and this phenomenon might havetwo benefits. One is prevention of cytoplasmic protein fromleaking out of the cytoplasm to express purer secretedproteases; the other is maintenance of bacterial growth.

Mutant Y114S with Efficient Synthesis of Z-APM andLower Caseinolysis. The activities of Z-APM synthesis andcaseinolysis were compared between protease PT121 and itsmutant Y114S with the same concentrations of proteins. Asshown in Figure 6, protease PT121 presented a higher initial

reaction rate of caseinolysis than that of Y114S. The specificactivity of PT121 on caseinolysis was 22,000 U/mg, which wasalmost 2-fold that of Y114S (12,000 U/mg). The results maydemonstrate that PT121 has higher hydrolysis activitycompared with Y114S.Interestingly, mutant Y114S facilitates efficient synthesis of

Z-APM. As shown in Table 2, v0 and kcat values of Y114S in Z-

APM synthesis were 8.2- and 8.5-fold those of PT121,respectively, whereas both KM1 and KM2 of mutant Y114Swere slightly declined compared with those of PT121. MutantY114S showed remarkably higher activity of Z-APM synthesisand considerably decreased the activity of caseinolysis. Thisdifference may be the result of alteration in substrate specificity.In the present study, the metalloprotease presented efficientpeptide synthesis through site-directed mutagenesis, which mayillustrate that the mutation site Tyr114 is a key residue formediating the substrate specificity of protease PT121.To rationalize this difference between PT121 and Y114S,

molecular docking was applied to obtain the enzyme−substratecomplex as shown in Figure S1. Molecular dynamic simulationwas used to calculate the binding free energies of two enzymestoward Phe-OMe, and the binding free energies were −20.23and −23.34 kcal/mol for PT121 and Y114S, respectively, bysampling the last 5 ns of trajectories. This result illustrates thatmutant Y114S has stronger binding ability to Phe-OMe than

Figure 5. Permeability of the inner (A) and outer (B) membranes ofE. coli BL21(DE3) containing target plasmid: (■) cultivation of BL21/pET22b added with arabinose; (□) cultivation of BL21/pET22b; (▲)cultivation of BL21/pelB-PT121 with added arabinose; (△)cultivation of BL21/pelB-PT121.

Figure 6. Caseinolytic initial rate at various casein concentrations forPT121 (■) and Y114S (●). Experiments were carried out in triplicate,and results are expressed as the mean ± standard deviation from threeindependent experiments.

Table 2. Kinetic Parameters of Z-APM Synthesis of ProteasePT121 and Mutant Y114Sa

enzyme v0 (μM·s−1) KM1 (mM) KM2 (mM) kcat (s−1)

mutant Y114S 7.4 ± 1.2 46 ± 3.2 32 ± 2.1 12.8 ± 0.8protease PT121 0.9 ± 0.1 48 ± 3.1 38 ± 2.8 1.5 ± 0.2

aZ-APM synthesis reaction was carried out in aqueous solvent at 37°C and pH 6.0 for 60 min using various concentrations (10−100 mM)of Cbz-Asp and Phe-OMe at fixed 500 mM Phe-OMe and 100 mMCbz-Asp, respectively. Values are expressed the mean ± standarddeviation.

Journal of Agricultural and Food Chemistry Article

DOI: 10.1021/acs.jafc.6b04164J. Agric. Food Chem. 2016, 64, 9631−9638

9636

that of PT121. Previous studies also reported only onemutation that considerably affects substrate specificity.45

Protease PT121 and its mutant Y114S might be useful in thedegradation and engineering of food proteins46 and in thesynthesis of bioactive peptides, such as aspartame, respectively.Further work will be conducted to engineer proteases PT121and Y114S for enhancing yield in the synthesis of other activepeptides.In summary, we demonstrated that replacement of the signal

peptide promotes the extracellular expression of metal-loprotease PT121 and its mutant Y114S. The addition ofarabinose significantly promoted the expression of proteasesand improved the purity of the proteases in E. coli. In addition,the residue 114 has been well characterized as a key residue toalter substrate specificity. Molecular simulation suggested thatresidue 114 plays a positive role in substrate binding.

■ ASSOCIATED CONTENT

*S Supporting InformationThe Supporting Information is available free of charge on theACS Publications website at DOI: 10.1021/acs.jafc.6b04164.

Table S1: Primers used in this study. Figure S1: (A) 3Dstructure of Y114S-Phe-OMe complex interaction; (B)structure of Phe-OMe in the active center; (C) RMSDsof the Cα atoms for PT121 and Y114S over alltrajectories at 300 K (40 ns) (PDF)

■ AUTHOR INFORMATION

Corresponding Author*(B.H.) E-mail: bingfanghe@njtech.edu.cn. Phone: 86-25-58139902. Fax: 86-25-58139902.

ORCIDBingfang He: 0000-0002-4502-9531FundingThis research was supported by the National Natural ScienceFoundation of China (21376119, 21506099, 81503012),Natural Science Foundation of Jiangsu Province of China(BK20150963), and the Specialized Research Fund for theDoctoral Program of Higher Education (20123221130001).We also acknowledge the fund sponsored by the ResearchInnovation Program for College Graduates of Jiangsu Province(KYZZ_0228).

NotesThe authors declare no competing financial interest.

■ REFERENCES(1) Mazur, R. H.; Schlatter, J. M.; Goldkamp, A. H. Structure-tasterelationships of some dipeptides. J. Am. Chem. Soc. 1969, 91, 2684−2691.(2) Khurana, H. K.; Cho, I. K.; Shim, J. Y.; Li, Q. X.; Jun, S. J.Application of multibounce attenuated total reflectance fouriertransform infrared spectroscopy and chemometrics for determinationof aspartame in soft drinks. J. Agric. Food Chem. 2008, 56, 778−783.(3) Yagasaki, M.; Hashimoto, S. I. Synthesis and application ofdipeptides; current status and perspectives. Appl. Microbiol. Biotechnol.2008, 81, 13−22.(4) Isowa, Y.; Ohmori, M.; Ichikawa, T.; Mori, K.; Nonaka, Y.;Kihara, K. I.; Oyama, K.; Satoh, H.; Nishimura, S.; Isowa, Y. Thethermolysin-catalyzed condensation reactions of N-substituted asparticand glutamic acids with phenylalanine alkyl esters. Tetrahedron Lett.1979, 20, 2611−2612.

(5) Birrane, G.; Bhyravbhatla, B.; Navia, M. A. Synthesis of aspartameby thermolysin: an X-ray structural study. ACS Med. Chem. Lett. 2014,5, 706−710.(6) Liu, K.; Zhu, F.; Zhu, L.; Chen, G.; He, B. Highly efficientenzymatic synthesis of Z-aspartame in aqueous medium via in situproduct removal. Biochem. Eng. J. 2015, 98, 63−67.(7) Vojcic, L.; Pitzler, C.; Koerfer, G.; Jakob, F.; Martinez, R.;Maurer, K. H.; Schwaneberg, U. Advances in protease engineering forlaundry detergents. New Biotechnol. 2015, 32, 629−634.(8) MarieClaire, C.; Ruffet, E.; Antonczak, S.; Beaumont, A.;O'Donohue, M.; Roques, B. P.; FournieZaluski, M. C. Evidence bysite-directed mutagenesis that arginine 203 of thermolysin and arginine717 of neprilysin (neutral endopeptidase) play equivalent critical rolesin substrate hydrolysis and inhibitor binding. Biochemistry 1997, 36,13938−13945.(9) O'Donohue, M. J.; Roques, B. P.; Beaumont, A. Cloning andexpression in Bcillus subtilis of the npr gene from Bacillusthermoproteolyticus rokko coding for the thermostable metalloproteasethermolysin. Biochem. J. 1994, 300, 599−603.(10) Han, M. H.; Ding, H. Y.; Wang, J. L.; Jin, M. Y.; Yu, X. B.Expression of the lasB gene encoding an organic solvent-stable elastasein Pichia pastoris and potential applications of the recombinantenzymes in peptide synthesis. Biochem. Eng. J. 2013, 77, 154−160.(11) Marie-Claire, C.; Roques, B. P.; Beaumont, A. Intramolecularprocessing of prothermolysin. J. Biol. Chem. 1998, 273, 5697−5701.(12) Marie-Claire, C.; Ruffet, E.; Beaumont, A.; Roques, B. P. Theprosequence of thermolysin acts as an intramolecular chaperone whenexpressed in trans with the mature sequence in Escherichia coli. J. Mol.Biol. 1999, 285, 1911−1915.(13) Ogino, H.; Yokoo, J.; Watanabe, F.; Ishikawa, H. Cloning andsequencing of a gene of organic solvent-stable protease secreted fromPseudomonas aeruginosa PST-01 and its expression in Escherichia coli.Biochem. Eng. J. 2000, 5, 191−200.(14) Mansfeld, J.; Petermann, E.; Durrschmidt, P.; Ulbrich-Hofmann,R. The propeptide is not required to produce catalytically activeneutral protease from Bacillus stearothermophilus. Protein ExpressionPurif. 2005, 39, 219−228.(15) Cheng, J.; Wu, D.; Chen, S.; Chen, J.; Wu, J. High-levelextracellular production of α-cyclodextrin glycosyltransferase withrecombinant Escherichia coli BL21 (DE3). J. Agric. Food Chem. 2011,59, 3797−3802.(16) Makrides, S. C. Strategies for achieving high-level expression ofgenes in Escherichia coli. Microbiol. Rev. 1996, 60, 512−538.(17) Su, L.; Chen, S.; Yi, L.; Woodard, R. W.; Chen, J.; Wu, J.Extracellular overexpression of recombinant Thermobif ida fuscacutinase by α-hemolysin secretion system in E. coli BL21(DE3).Microb. Cell Fact. 2012, 11, 8.(18) Shin, H. D.; Chen, R. R. Extracellular recombinant proteinproduction from an Escherichia coli lpp deletion mutant. Biotechnol.Bioeng. 2008, 101, 1288−1296.(19) Matos, C. F. R. O.; Branston, S. D.; Albiniak, A.; Dhanoya, A.;Freedman, R. B.; Keshavarz-Moore, E.; Robinson, C. High-yield exportof a native heterologous protein to the periplasm by the tattranslocation pathway in Escherichia coli. Biotechnol. Bioeng. 2012,109, 2533−2542.(20) Li, Z. F.; Li, B.; Liu, Z. G.; Wang, M.; Gu, Z.-B.; Du, G. C.; Wu,J.; Chen, J. Calcium leads to further increase in glycine-enhancedextracellular secretion of recombinant α-cyclodextrin glycosyltransfer-ase in Escherichia coli. J. Agric. Food Chem. 2009, 57, 6231−6237.(21) Haryadi, R.; Ho, S.; Kok, Y. J.; Pu, H. X.; Zheng, L.; Pereira, N.A.; Li, B.; Bi, X.; Goh, L. T.; Yang, Y.; Song, Z. Optimization of heavychain and light chain signal peptides for high level expression oftherapeutic antibodies in CHO cells. PLoS One 2015, 10, e0116878.(22) Wu, X. Q.; Jornvall, H.; Berndt, K. D.; Oppermann, U. Codonoptimization reveals critical factors for high level expression of two rarecodon genes in Escherichia coli: RNA stability and secondary structurebut not tRNA abundance. Biochem. Biophys. Res. Commun. 2004, 313,89−96.

Journal of Agricultural and Food Chemistry Article

DOI: 10.1021/acs.jafc.6b04164J. Agric. Food Chem. 2016, 64, 9631−9638

9637

(23) Yin, J.; Tian, H.; Bao, L.; Dai, X.; Gao, X.; Yao, W. Analternative method of enhancing the expression level of heterologousprotein in Escherichia coli. Biochem. Biophys. Res. Commun. 2014, 455,198−204.(24) Srirangan, K.; Orr, V.; Akawi, L.; Westbrook, A.; Moo-Young,M.; Chou, C. P. Biotechnological advances on penicillin G acylase:pharmaceutical implications, unique expression mechanism andproduction strategies. Biotechnol. Adv. 2013, 31, 1319−1332.(25) Xu, Y.; Rosenkranz, S.; Weng, C. L.; Scharer, J. M.; Moo-Young,M.; Chou, C. P. Characterization of the T7 promoter system forexpressing penicillin acylase in Escherichia coli. Appl. Microbiol.Biotechnol. 2006, 72, 529−536.(26) Tang, X. Y.; Pan, Y.; Li, S.; He, B. F. Screening and isolation ofan organic solvent-tolerant bacterium for high-yield production oforganic solvent-stable protease. Bioresour. Technol. 2008, 99, 7388−7392.(27) Laemmli, U. K. Cleavage of structural proteins during theassembly of the head of bacteriophage T4. Nature 1970, 227, 680−685.(28) Tang, X. Y.; Wu, B.; Ying, H. J.; He, B. F. Biochemicalproperties and potential applications of a solvent-stable protease fromthe high-yield protease producer Pseudomonas aeruginosa PT121. Appl.Biochem. Biotechnol. 2010, 160, 1017−1031.(29) Cai, Z.; Xu, W.; Xue, R.; Lin, Z. Facile, reagentless and in siturelease of Escherichia coli intracellular enzymes by heat-inducibleautolytic vector for high-throughput screening. Protein Eng., Des. Sel.2008, 21, 681−687.(30) Katrolia, P.; Yan, Q.; Jia, H.; Li, Y.; Jiang, Z.; Song, C. Molecularcloning and high-level expression of a β-galactosidase gene fromPaecilomyces aerugineus in Pichia pastoris. J. Mol. Catal. B: Enzym. 2011,69, 112−119.(31) Li, Z.; Gu, Z.; Wang, M.; Du, G.; Wu, J.; Chen, J. Delayedsupplementation of glycine enhances extracellular secretion of therecombinant α-cyclodextrin glycosyltransferase in Escherichia coli. Appl.Microbiol. Biotechnol. 2010, 85, 553−561.(32) Kusano, M.; Yasukawa, K.; Inouye, K. Synthesis of N-carbobenzoxy-L-aspartyl-L-phenylalanine methyl ester catalyzed bythermolysin variants with improved activity. Enzyme Microb. Technol.2010, 46, 320−325.(33) Thayer, M. M.; Flaherty, K. M.; Mckay, D. B. Three-dimensional structure of the elastase of Pseudomonas aeruginosa at1.5-A resolution. J. Biol. Chem. 1991, 266, 2864−2871.(34) Zhao, H.; Caflisch, A. Discovery of dual ZAP70 and Syk kinasesinhibitors by docking into a rare C-helix-out conformation of Syk.Bioorg. Med. Chem. Lett. 2014, 24, 1523−1527.(35) Lindorff-Larsen, K.; Piana, S.; Palmo, K.; Maragakis, P.; Klepeis,J. L.; Dror, R. O.; Shaw, D. E. Improved side-chain torsion potentialsfor the Amber ff99SB protein force field. Proteins: Struct., Funct., Genet.2010, 78, 1950−1958.(36) Yin, J.; Bao, L.; Tian, H.; Gao, X.; Yao, W. Quantitativerelationship between the mRNA secondary structure of translationalinitiation region and the expression level of heterologous protein inEscherichia coli. J. Ind. Microbiol. Biotechnol. 2016, 43, 97−102.(37) Bai, C.; Wang, X.; Zhang, J.; Sun, A.; Wei, D.; Yang, S.Optimisation of the mRNA secondary structure to improve theexpression of interleukin-24 (IL-24) in Escherichia coli. Biotechnol. Lett.2014, 36, 1711−1716.(38) Zhang, W.; Xiao, W.; Wei, H.; Zhang, J.; Tian, Z. mRNAsecondary structure at start AUG codon is a key limiting factor forhuman protein expression in Escherichia coli. Biochem. Biophys. Res.Commun. 2006, 349, 69−78.(39) Faust, G.; Stand, A.; Weuster-Botz, D. IPTG can replace lactosein auto-induction media to enhance protein expression in batch-cultured Escherichia coli. Eng. Life Sci. 2015, 15, 824−829.(40) Duan, X.; Chen, J.; Wu, J. Optimization of pullulanaseproduction in Escherichia coli by regulation of process conditionsand supplement with natural osmolytes. Bioresour. Technol. 2013, 146,379−385.

(41) Esakkiraj, P.; Meleppat, B.; Lakra, A. K.; Ayyanna, R.; Arul, V.Cloning, expression, characterization and application of proteaseproduced by Bacillus cereus PMW8. RSC Adv. 2016, 6, 38611−38616.(42) Narayanan, N.; Hsieh, M. Y.; Xu, Y.; Chou, C. P. Arabinose-induction of lac-derived promoter systems for penicillin acylaseproduction in Escherichia coli. Biotechnol. Prog. 2006, 22, 617−625.(43) Odunuga, O. O.; Adekoya, O. A.; Sylte, I. High-level expressionof pseudolysin, the extracellular elastase of Pseudomonas aeruginosa, inEscherichia coli and its purification. Protein Expression Purif. 2015, 113,79−84.(44) Gao, X.; Wang, J.; Yu, D. Q.; Bian, F.; Xie, B. B.; Chen, X. L.;Zhou, B. C.; Lai, L. H.; Wang, Z. X.; Wu, J. W.; Zhang, Y. Z. Structuralbasis for the autoprocessing of zinc metalloproteases in thethermolysin family. Proc. Natl. Acad. Sci. U. S. A. 2010, 107, 17569−17574.(45) Wang, L.; Duan, X.; Wu, J. Enhancing the α-cyclodextrinspecificity of cyclodextrin glycosyltransferase from Paenibacillusmacerans by mutagenesis masking subsite-7. Appl. Environ. Microbiol.2016, 82, 2247−2255.(46) Asaoka, K.; Yasukawa, K.; Inouye, K. Coagulation of soyproteins induced by thermolysin and comparison of the coagulationreaction with that induced by subtilisin Carlsberg. Enzyme Microb.Technol. 2009, 44, 229−234.

Journal of Agricultural and Food Chemistry Article

DOI: 10.1021/acs.jafc.6b04164J. Agric. Food Chem. 2016, 64, 9631−9638

9638