Eia

SP

a

ARRA

KRMMETS

1

aflea(aapps

(

f

0d

Journal of Biotechnology 147 (2010) 64–72

Contents lists available at ScienceDirect

Journal of Biotechnology

journa l homepage: www.e lsev ier .com/ locate / jb io tec

xpression and purification of recombinant human �1-proteinase inhibitor andts single amino acid substituted variants in Escherichia coli for enhanced stabilitynd biological activity�

aurabh Agarwal, Shweta Jha, Indraneel Sanyal ∗, D.V. Amlalant Transgenic Lab, National Botanical Research Institute (CSIR), P.O. Box 436, Rana Pratap Marg, Lucknow, U.P. 226 001, India

r t i c l e i n f o

rticle history:eceived 26 May 2009eceived in revised form 12 March 2010ccepted 17 March 2010

eywords:ecombinant �1-proteinase inhibitorBP fusion proteinodified synthetic gene

. coli expression

a b s t r a c t

Human �1-proteinase inhibitor (�1-PI) is the most abundant protease inhibitor found in the blood andexpression of biologically active recombinant �1-PI has great potential in therapeutic applications. Wereport here the expression of a synthetic �1-PI gene and its variants in Escherichia coli. Modified �1-PIgene and its single amino acid variants were cloned in pMAL-c2X vector, which allowed expression ofrecombinant protein(s) as a fusion of maltose-binding protein (MBP) with factor Xa protease recognitionsite between the fusion partners. The synthetic gene(s) were expressed in different E. coli strains andmaximum expression of recombinant �1-PI and variants up to 24% of total soluble protein (TSP) wasachieved with engineered strain carrying extra copies of tRNAs for rare codons. Recombinant �1-PI pro-tein(s) were purified by amylose affinity chromatography with high homogeneity and overall yield ofabout 7–9 mg l−1 of bacterial culture (∼5.2 g wet cell mass). E. coli expressed recombinant �1-PI showed

hermostabilityite-specific mutation

specific anti-elastase activity and appeared as a single band of ∼45.0 kDa on SDS-PAGE. Primary structureof purified protein and integrity of N-terminus has been verified by mass spectrometric analysis. Recom-binant �1-PI expressed in E. coli was fully intact having molecular mass similar to native unglycosylatedprotein purified from human plasma. Increased thermostability and specific activities of purified �1-PIvariant proteins confirmed the stabilizing effect of incorporated mutations. Our results demonstrate effi-cient expression and purification of stable and biologically active �1-PI and its variants in E. coli for further

therapeutic applications.

. Introduction

Human �1-proteinase inhibitor (�1-PI), also known as alpha-1-ntitrypsin (AAT) is an archetype of the serine protease inhibitoramily and a major constituent in human plasma. Its key physio-ogical function in human and animals is inhibition of neutrophillastase in lungs, thus protecting pulmonary tissues from dam-ge (Blank and Brantley, 1994). A single amino acid mutationGlu342Lys) in the mobile domain results in �1-PI deficiencynd potentially lethal hereditary disease causing lung emphysema

nd liver disorders (Lomas, 2005). Intravenous augmentation ofurified �1-PI from human serum is the only clinical treatmentresently available, which is in great demand, always in limitedupply and associated with possibility of pathogen contamination

� This research was supported by Council of Scientific and Industrial ResearchCSIR), India under the Network Project CMM 0004.∗ Corresponding author. Tel.: +91 522 2297954/2297955;

ax: +91 522 2205836/2205839.E-mail address: i sanyal@rediffmail.com (I. Sanyal).

168-1656/$ – see front matter © 2010 Elsevier B.V. All rights reserved.oi:10.1016/j.jbiotec.2010.03.008

© 2010 Elsevier B.V. All rights reserved.

(Heresi and Stoller, 2008). Over expression of recombinant human�1-PI in diverse alternative host systems has been considered overthe period, however, none of them could fulfil the demand of clin-ically safe and biologically active form of protein for therapeuticapplications (Karnaukhova et al., 2006).

The tertiary structure of �1-PI share a common structure withother serpins and composed of three �-sheets (A, B, C) and nine�-helices (hA–hI) as shown in Fig. 1A. In the native strained (S)active conformation, the molecule is intact and the reactive centreloop (RCL) is exposed to proteolytic cleavage. The cleavage accom-panies an irreversible transition to a very stable relaxed (R) formwhere the newly created N-terminal portion of the cleaved loop iscompletely inserted as central strand of sheet-A, with concomitantloss of inhibitory activity (Whisstock et al., 2000). The shutter andthe breach are two important areas for this conformational change(S → R transition). The breach, located at the top of sheet-A, is the

region where the RCL first inserts. The shutter region is located inthe middle of the serpin and controls the opening of the sheet-A.Both regions contain a number of highly conserved residues andseveral positions at which mutations result in hyperstability, with-out affecting inhibitory activity (Lee et al., 1996).

S. Agarwal et al. / Journal of Biotechnology 147 (2010) 64–72 65

Fig. 1. Tertiary structure of human �1-PI protein, site-specific mutations in modified gene and chimeric gene construct. (A) Ribbon model of human �1-PI protein molecule(PDB ID 1hp7) highlighting RCL, other critical domains such as breach and shutter domain and positions of incorporated single amino acid substitutions (encircled) to generatefive variants for increased stability and biological activity in recombinant �1-PI. (B) Table showing five incorporated point mutations and corresponding codon replacementsi PI hard inducf ssion vw MMI

potHgceb1tcc(sciW

tvtpgtsgcf(

n the modified synthetic �1-PI gene and plasmid vectors. (C) Expression vector pMownstream to the malE gene encoding for MBP fusion tag under the control of IPTGactor Xa between MBP and �1-PI fusion partners. Similarly five independent expreith arrows and designated as (I) pMFC, (II) pMFL, (III) pMAG, (IV) pMMV and (V) p

Escherichia coli is often the host of choice for its ability to multi-ly rapidly to high cell densities on inexpensive media, availabilityf versatile vectors and achieving expression levels exceeding morehan 30% of total cellular protein (Makrides, 1996; Baneyx, 1999).owever, problems arise with expression of several heterologousenes for yield, stability and solubility of foreign proteins in E.oli due to factors including, presence of rare codons, translationalfficiency, stability of mRNA, insolubility, formation of inclusionodies and complex post-translational modifications (Makrides,996; Kane, 1995). Several strategies have been used to circumventhe codon bias, increase yields and stability of foreign proteins in E.oli such as exchange of rare codons with E. coli preferred codons,o-expression of desired tRNA genes and use of fusion proteinsBaneyx, 1999). Different fusion tags seem to improve solubility,tability, translational efficiency and folding by acting as molecularhaperones in context to the partner proteins besides facilitat-ng rapid purification (Esposito and Chatterjee, 2006; Kapust and

augh, 1999).We have designed and developed a highly modified syn-

hetic �1-PI gene and its five single amino acid substitutionariants by extensive codon-optimization for dicot plant geneso achieve safe, stable and biologically active recombinant �1-PIrotein in transgenic plants (Agarwal et al., 2008). The designedene contains 69 codons out of 394 (17%), which are reportedo be rarely used in E. coli and their presence inhibits expres-

ion (Makrides, 1996). In the present study, the modified �1-PIene and its single amino acid substituted variants were effi-iently expressed under optimized culture conditions as MBPusion protein, driven by isopropyl-�-d-thiogalactopyranosideIPTG) inducible ptac promoter by overcoming the limitations

bouring the modified synthetic �1-PI coding sequence cloned in vector pMAL-c2X,ible ptac promoter and lacZ� terminator with an in-frame cleavage site of proteaseectors for �1-PI variants were also generated with specific point mutations shown

respectively.

of codon bias, insolubility, misfolding, degradation or aggrega-tion. The expressed protein(s) were purified by amylose affinitychromatography and characterized for its integrity, stability andbiological activity.

2. Materials and methods

2.1. Construction of E. coli expression vectors with modified ˛1-PIgene

The coding sequence of human �1-PI gene was codon-optimizedaccording to the codon usage frequencies of highly expressed dicotplant genes and synthesized by PCR-based gene synthesis usingoverlapping oligonucleotides (Agarwal et al., 2008; Stemmer etal., 1995). Similarly, five variants of modified �1-PI gene for sin-gle amino acid substitutions at Phe51 to Cys (FC), Phe51 to Leu(FL), Ala70 to Gly (AG), Met358 to Val (MV) and Met374 to Ile (MI)have been developed by site-directed mutagenesis of respectivecodons as shown in Fig. 1B (An et al., 2005). These positions wereselected for substitution as they were reported to be present at crit-ical sites of the �1-PI protein molecule (Im et al., 2004). Synthetic�1-PI gene and its variants were amplified using forward primer 5′-CGGAATTCggatccATGGAAGATCCTCAAGGAGATGCTGC-3′ and rev-erse primer 5′-GGTACCTCTAGaagcttTTACTACTTCTGAGTAGGGTT-AACC-3′ to incorporate initiation and stop codons (underlined) and

BamHI and HindIII restriction sites (lower case) at 5′ and 3′ endof the gene respectively. The amplified PCR products were clonedinto E. coli expression vector pMAL-c2X (NEB, USA), downstreamto the malE gene of MBP to develop vectors pMPI with modifiednative �1-PI gene and pMFC, pMFL, pMAG, pMMV and pMMI with

6 Biotec

g(

2

((ca0AhbT1sf

obptf(Maparp(

2

fsRw(pwiawbs

2

mcbwabso(

2

e

6 S. Agarwal et al. / Journal of

ene variants, using the standard molecular cloning techniquesSambrook and Russel, 2001).

.2. Expression and purification of recombinant ˛1-PI from E. coli

E. coli strains DH5�, TB1 (NEB, USA) and BL21-CodonPlus-RILBL21CP) carrying tRNA genes for codon AGG/AGA, ATA and CTAStratagene, USA) were used as expression hosts. Recombinant E.oli strains were grown in LB medium with 100 �g ml−1 ampicillint 37 ◦C up to OD600 0.5, 1.0 or 2.0. The cultures were induced with.3 mM IPTG and further grown at either 37 ◦C or 30 ◦C, for 1–8 h.liquots of induced cultures were withdrawn at periodic intervals,arvested by centrifugation (4000 × g for 10 min at 4 ◦C) followedy washing and resuspension in extraction buffer (EB; 20 mMris–HCl, pH 7.4, 200 mM NaCl, 1 mM EDTA, 1 mM sodium azide,0 mM �-mercaptoethanol). Cells were lysed by sonication withhort pulses and debris was removed by centrifugation (9000 × gor 20 min at 4 ◦C).

The crude extract from 1 l E. coli culture was diluted and appliedn a pre-equilibrated amylose column (5 ml, NEB, USA) followedy washing with 12 column volumes of EB and elution of boundroteins with EB supplemented with 10 mM maltose. The frac-ions containing fusion protein were pooled and concentratedollowed by cleavage with protease factor Xa in cleavage buffer20 mM Tris–HCl, pH 8.0, 100 mM NaCl, 2 mM CaCl2) to separate

BP and �1-PI fusion partners. The cleaved reaction products weregain applied on second amylose column followed by collection ofure �1-PI protein in flow-through fractions. Purified protein wasnalyzed by SDS-PAGE, Western immunoblotting, DAC-ELISA andesidual PPE activity assay. TSP was determined by dye bindingrocedure taking bovine serum albumin as a reference standardBradford, 1976).

.3. SDS-PAGE and Western immunoblotting

Protein samples combined with Laemmli buffer were boiledor 10 min and fractionated on 10% SDS-PAGE gel, followed bytaining with 0.1% Coomassie brilliant blue R-250 (Sambrook andussel, 2001; Laemmli, 1970). Electrophoresed protein samplesere transferred onto polyvinylidene difluoride (PVDF) membrane

Bio-Rad, USA) in transfer buffer (25 mM Tris base, 192 mM glycine,H 8.3, 0.1% SDS) for Western immunoblotting. The membranesere blocked in 5% non-fat dry milk (in PBST) at 25 ◦C followed by

ncubation with either rabbit anti-human �1-PI antibody or rabbitnti-MBP antibody. The membranes were washed and incubatedith goat anti-rabbit IgG-alkaline phosphatase conjugated anti-

ody followed by colour development with BCIP-NBT substrateolution.

.4. Quantitative estimation of recombinant ˛1-PI by DAC-ELISA

Recombinant �1-PI protein was quantified by DAC-ELISAethod (Agarwal et al., 2008). Aliquots of protein samples were

oated in wells of microtiter plate followed by blocking and incu-ation with 1:5000 dilution of rabbit anti-human �1-PI antibody,ashing with PBST and incubation with 1:8000 dilution of goat

nti-rabbit IgG-alkaline phosphatase conjugated antibody followedy colour development with p-nitrophenyl phosphate substrateolution. Expression levels of recombinant �1-PI were quantifiedn a linear standard curve plotted with pure human �1-PI proteinSigma, USA).

.5. Residual PPE activity assay and thermostability analysis

The biological activity of recombinant �1-PI in cell freextracts and purified samples was determined as residual porcine

hnology 147 (2010) 64–72

pancreatic elastase (PPE) activity using N-succinyl-Ala-Ala-Ala-p-nitroanilide as chromogenic substrate (Agarwal et al., 2008).Protein samples (50 �l) were added into 100 �l of assay buffer(20 mM Tris–HCl, pH 8.0, 150 mM NaCl, 0.01% Tween-80) followedby addition of 50 �l (3.0 �g ml−1) PPE, incubation at 37 ◦C for 15 minand addition of 50 �l of 2 mM chromogenic substrate and fur-ther incubation for 2 h at 25 ◦C. The activity of residual PPE wasdetermined by measuring the release of p-nitroaniline from chro-mogenic substrate at 405 nm. Pure human �1-PI protein was usedas reference standard for corresponding recombinant �1-PI con-centration and residual PPE activity. Specific activity was based onamount of biologically active �1-PI per mg of total soluble protein.The stability of wild type and mutant recombinant �1-PI proteinsexpressed in E. coli was determined as percentage of remainingbiological activity at 37 ◦C and 54 ◦C and compared with nativeglycosylated human serum �1-PI. The aliquots (5 �g) of purifiedprotein samples (50 �g ml−1) were incubated at 37 ◦C and 54 ◦C for2.5 h and 1 h, respectively. Aliquots were withdrawn after intervalof every 30 min and 10 min at 37 ◦C and 54 ◦C respectively and ana-lyzed for the remaining �1-PI activity by residual PPE enzymaticactivity assay. All the assays were performed in triplicates alongwith internal control.

2.6. Mass spectrometric analysis

Coomassie brilliant blue stained protein bands of interest werecut out from the SDS-PAGE gel and ‘in-gel’ digested as described byShevchenko et al. (2006). Proteins were reduced with dithiothreitoland alkylated with iodoacetamide before overnight digestion with0.02 �g �l−1 trypsin (Proteomics grade, Sigma, USA) at 37 ◦C. Therecovered peptides and intact protein samples were desalted withC18 and C4 Zip-Tip, respectively according to the manufacturer’sinstructions (Millipore, USA). Samples were prepared by the ‘drieddroplet method’ using �-cyano-4-hydroxycinnamic acid (CHCA) asmatrix dissolved in 50% (v/v) acetonitrile and 0.1% (v/v) trifluo-roacetic acid. The MS and MS/MS spectra were acquired using 4800MALDI–TOF/TOF mass spectrometer (Applied Biosystems, USA)equipped with a Nd:YAG (355 nm, 200 Hz) laser. The instrumentwas operated with delayed extraction (1100 ns) at acceleratingvoltage of 20 kV. The 4700 cal mix (having mixture of six stan-dard peptides) and bovine serum albumin (Applied Biosystems,USA) were used as external calibrants for reflector and linear mode,respectively. The MS and MS/MS spectra were typically acquired byaveraging 30 sub-spectra from a total of 900 shots of the laser withthe intensity set at 4200 and 4900, respectively.

Protein identification from the generated data was performedby searching against the MSDB database using online Mascotsearch engine (http://www.matrixscience.com/). Peptide mass tol-erance of ±100 ppm and fragment ion mass tolerance of ±0.2 Dawas set and maximum missed cleavages allowed was one. Car-bamidomethylation (C) as fixed modification, deamidation (NQ)and oxidation (M) as variable modifications were considered. Theprobability score calculated at p < 0.05 was used as the criteria forcorrect identification of proteins.

3. Results

3.1. Synthesis of modified ˛1-PI gene and recombinant vectors

The coding sequence of human �1-PI gene was designed

to display codon usage pattern of abundantly expressed dicotplant genes. In the modified �1-PI gene (GenBank accession no.EF638826), 205 codons out of 394 (52%) were replaced with sub-stitution of 281 favoured nucleotides. Codon frequency analysisrevealed that modified �1-PI gene contains 69 codons (17%), which

S. Agarwal et al. / Journal of Biotechnology 147 (2010) 64–72 67

Fig. 2. Expression of MBP–�1-PI fusion protein in E. coli under different culture conditions. (A) Growth pattern of recombinant bacterial strains DH5� (�), TB1 (�) andBL21CP (�) with pMPI expression vector at 37 ◦C after induction with IPTG at OD600 0.5 (arrow). (B) Comparative yield of expressed fusion protein in E. coli BL21CP withI fferen( inducer 3–10,r row.

wmefimtSab

3

iBigIib(sDi3esloBoifrp8

PTG induction at 25 ◦C (�), 30 ◦C (�) and 37 ◦C (�). (C) Yield of fusion protein in diD) SDS-PAGE analysis of cell-free extracts of recombinant BL21CP–pMPI cultureecombinant BL21CP having pMAL-c2X vector without �1-PI modified gene; laneespectively at 30 ◦C with expression of ∼87.5 kDa fusion protein as indicated by ar

ere reported to be rare for E. coli and among them 19 (5%) areost rarely used codons (AGA/AGG, ATA, CTA, CCA and CCC). To

xpress recombinant �1-PI and its variants in E. coli, the modi-ed �1-PI coding sequence(s) were inserted downstream of thealE gene encoding MBP fusion tag in the bacterial expression vec-

or pMAL-c2X under the control of IPTG inducible ptac promoter.ix expression vectors such as pMPI, pMFC, pMFL, pMAG, pMMVnd pMMI containing in-frame cleavage site of factor Xa proteaseetween MBP–�1-PI fusion partners were developed (Fig. 1C).

.2. Expression of recombinant ˛1-PI in E. coli

In order to optimize the host strain and conditions for max-mum induction, the recombinant E. coli strains DH5�, TB1 andL21CP transformed with pMPI expression vector were grown and

nduced separately by addition of 0.3 mM IPTG after the bacterialrowth has attained the optical density of 0.5, 1.0 or 2.0 at 600 nm.t was observed that growth of all the three E. coli cultures werencreased exponentially up to 5 h post-induction. The induction ofacterial cultures with IPTG supplementation at optical density 0.5OD600) showed maximum increase in growth as well as expres-ion of recombinant protein in BL21CP strain followed by TB1 andH5� strains (Fig. 2A). Recombinant BL21CP–pMPI culture was

nduced with IPTG at OD600 0.5 and cultivated at 25 ◦C, 30 ◦C and7 ◦C for 8 h to determine the optimum growth temperature forxpression of fusion protein. The results analysed by DAC-ELISAhowed significant increase in fusion protein content at 30 ◦C fol-owed by 25 ◦C while minimum recovery at 37 ◦C (Fig. 2B). The yieldf specific fusion protein was significantly higher in recombinantL21CP strain with 24% of TSP as compared to 18% and 15% TSPbtained in TB1 and DH5� strains respectively at 30 ◦C post IPTG

nduction (Fig. 2C). Recombinant BL21CP culture was grown at dif-erent pH of LB medium ranging from pH 6.0 to 9.0. The results onecovery of fusion protein showed maximum yield of 24% TSP atH 7.0 which decreased significantly to 12–14% TSP at pH above.0 (data not shown). Maximum induction of fusion protein of

t strains of E. coli BL21CP (�), TB1 (�) and DH5� (�) at 30 ◦C after IPTG induction.d for different time periods. Lane 1, prestained molecular mass standard; lane 2,recombinant BL21CP transformed with pMPI expression vector induced for 1–8 h

87.5 kDa in recombinant BL21CP–pMPI culture was induced withIPTG at 30 ◦C after 6 h incubation as revealed by SDS-PAGE analyses(Fig. 2D). The yield of fusion protein did not increase significantlybeyond 6–8 h of incubation. The size of fusion protein on SDS-PAGEwas approximately 87.5 kDa, which matched well with sum of thetheoretical mass of non-glycosylated �1-PI (45.0 kDa) and MBP(42.48 kDa).

3.3. Purification of recombinant ˛1-PI protein

Recombinant �1-PI protein and its variants were purified up tohomogeneity as evident from SDS-PAGE and Western immunoblotwith maximum yield of about ∼7–9 mg from 1 l of recombinantBL21CP culture, grown and induced under optimized conditions.The MBP–�1-PI fusion protein was purified from cell-free extractsof bacteria by affinity chromatography on amylose resin. Major-ity of the fusion protein was eluted in first few fractions after theaddition of maltose in extraction buffer. SDS-PAGE of eluted pro-tein fraction showed one major band of ∼87.5 kDa and anotherminor band of ∼70.0 kDa, while the immunoblot analysis con-firmed the ∼87.5 kDa band as MBP–�1-PI fusion protein (Fig. 3A,B). The minor band of ∼70 kDa may be an intermediate or degra-dation product of fusion protein. The purified fusion protein wasdigested with factor Xa protease that showed the appearance ofonly two protein bands which migrated with apparent molecularmasses of ∼45.0 and ∼42.5 kDa on SDS-PAGE corresponding to �1-PI and MBP respectively (Fig. 3A). Immunoblotting of above gel withanti-�1-PI and anti-MBP antibodies further confirmed the cleavageof fusion protein into �1-PI and MBP (Fig. 3B, C). Fusion proteinof MBP–Paramyocin �Sal of 70.2 kDa as internal control was alsodigested with factor Xa, which produced two proteins of molecu-

lar mass 42.5 and 27.7 kDa corresponding to MBP and Paramyocin�Sal respectively that reflected the efficacy and specificity of fac-tor Xa cleavage (Fig. 3A). The factor Xa digested fusion protein wasagain chromatographed on second amylose column, which in turnbinds with cleaved MBP fusion tag while unbound pure �1-PI was

68 S. Agarwal et al. / Journal of Biotechnology 147 (2010) 64–72

Fig. 3. Characterization of MBP–�1-PI fusion protein expressed in BL21CP at 30 ◦C. (A) Coomassie brilliant blue stained 10% SDS-PAGE gel. Lane 1, prestained molecular massstandard; lane 2, amylose affinity chromatography purified fusion protein of ∼87.5 kDa; lane 3, purified MBP–�1-PI fusion protein digested with factor Xa protease showingcleaved protein of ∼45.0 kDa �1-PI and ∼42.5 kDa MBP respectively; lane 4, fusion protein MBP–paramyocin �Sal (∼70.2 kDa) digested with factor Xa as control showingtwo cleaved protein bands of ∼42.5 kDa MBP and ∼27.7 kDa paramyocin �Sal respectively; lane 5, pure human �1-PI of ∼52.0 kDa and MBP of ∼42.5 kDa. (B) Westernimmunoblot of the gel A with anti-human �1-PI antibody exhibiting cross-reaction with; lane 1, fusion protein; lane 2, �1-PI fragment of factor Xa digested fusion protein;lane 3, no cross-reaction with MBP and paramyosin proteins; lane 4, pure human �1-PI protein. (C) Western immunoblot of the gel A with anti-MBP antibody exhibitingcross-reaction with; lane 1, ∼87.5 kDa fusion protein; lane 2, MBP fragment of factor Xa digested fusion protein; lane 3, MBP fragment in factor Xa digested control substrateMBP–paramyocin �Sal; lane 4, pure MBP protein.

Fig. 4. Purification of recombinant �1-PI protein expressed in BL21CP strain. (A) Lane 1, prestained molecular mass standard; lane 2, IPTG induced crude cell-free extract ofrecombinant BL21CP harbouring pMAL-c2X vector without �1-PI gene; lane 3, induced recombinant BL21CP–pMPI extract showing expression of fusion protein of ∼87.5 kDa.(B) Coomassie brilliant blue stained 10% SDS-PAGE gel. Lane 1, prestained molecular mass standard; lane 2, amylose affinity purified fusion protein of ∼87.5 kDa; lane 3, factorXa digested fusion protein showing ∼45.0 kDa �1-PI and ∼42.5 kDa MBP; lane 4, purified �1-PI protein of ∼45.0 kDa eluted from second amylose column; lane 5, purifiedh showna i-huml inantm ∼45.0

cotprw

TP

uman �1-PI protein of ∼52.0 kDa. A minor band of ∼70.0 kDa was also observed (fter second affinity chromatography. (C) Western immunoblot of the gel B with antane 2, �1-PI fragment in factor Xa digested fusion protein; lane 3, purified recomb

ass of purified fusion, human and recombinant �1-PI protein of ∼87.5, ∼52.0 and

ollected in the flow-through (Fig. 4A–C). The stepwise recoveryf recombinant � -PI protein from 1 l of recombinant BL21CP cul-

1ure is summarized in Table 1. Western immunoblot analysis ofurified protein samples with anti-�1-PI antibody showed cross-eaction with single discrete band of ∼45.0 kDa, which matchedell with the theoretical molecular mass of non-glycosylated �1-

able 1urification steps of recombinant �1-PI from E. coli BL21CP straina.

Purification step Protein content (mg)

Total soluble proteinb Recombin

Cell-free extract 144.0 33.50First amylose affinity chromatographyf 38.30 29.13Second amylose affinity chromatographyg 8.38 7.25

a Starting with 1 l of bacterial culture grown at 30 ◦C for 6 h after IPTG induction.b Total soluble protein determined by Bradford assay.c Recombinant protein estimated by DAC-ELISA using anti �1-PI antibody.d Biological activity (elastase inhibitory capacity) estimated by residual PPE activity asse Specific activity is based on amount of biologically active �1-PI per mg of total solublf Purified MBP–�1-PI fusion protein.g Factor Xa cleaved purified recombinant �1-PI protein.

by arrow), which was co-purified with fusion protein and subsequently removedan �1-PI antibody. Lane 1, exhibiting cross-reaction with ∼87.5 kDa fusion protein;�1-PI protein; lane 4, pure human �1-PI protein. Arrow corresponds to molecularkDa respectively.

PI and reflected significant homogeneity of the purified protein(Fig. 4C). The final yield of pure recombinant � -PI was about 7 mg

1from 1 l of bacterial culture (∼5.2 g wet cell weight). Similar purifi-cation profile and yields ranging between 7 and 9 mg 1−1 bacterialcultures were obtained for different variants of recombinant �1-PIprotein.

Specific activitye Yield (%)

ant �1-PIc Biologically active �1-PId

8.80 0.06 100.007.21 0.19 87.003.77 0.45 21.6

ay.e protein.

S. Agarwal et al. / Journal of Biotechnology 147 (2010) 64–72 69

Fig. 5. Identification and characterization of �1-PI protein by MALDI–TOF/TOF. (A) The observed MS spectrum (peptide mass fingerprint) of the tryptic digest of recombinant�1-PI purified from E. coli. Further MS/MS analysis and peptide sequencing was performed by selecting at least ten precursor ions (data not shown). Inset shows the observedp n, whs e obset ass mc

3

nlsoeTtobap

gtrutoe

eak of 44328.35 Da for mass of singly charged ion species (M+H)+ of intact proteipectrum of the tryptic digest of pure human serum �1-PI protein. Inset shows thheoretical mass (52 kDa) of glycosylated human �1-PI. The decrease in observed maused by matrix properties.

.4. Recombinant ˛1-PI protein analysis by MALDI–TOF/TOF

Accurate masses of recombinant �1-PI expressed in E. coli andative human serum �1-PI protein were determined by MS in

inear high-mass positive ion mode with low mass gate (LMG)et at 100 Da. The observed peaks showed molecular weightsf 44328.36 Da and 49824.49 Da for purified recombinant E. colixpressed and native human �1-PI, respectively (Fig. 5A, B-inset).he observed mass for recombinant protein was matched well withheoretical mass of unglycosylated �1-PI, while a 2 kDa decrease inbserved mass of human �1-PI (theoretical mass 52.0 kDa) maye due to fragmentation of glycan chains and removal of sialiccid residues from the glycosylated human �1-PI caused by ‘hot’roperties of CHCA matrix.

For unambiguous identification of protein, peptide fragmentsenerated by ‘in-gel’ trypsin digestion were analyzed by MS (pep-ide mass fingerprinting) and MS/MS (peptide sequencing) in

eflector positive ion mode. The resulting spectra were searchedsing Mascot search engine and both human and recombinant pro-eins were significantly identified as �1-PI with a MOWSE scoref 416 and 590, respectively at p < 0.05 with high sequence cov-rage. The MS spectra of tryptic peptides of both the proteins are

ich corresponds to theoretical mass of unglycosylated �1-PI. (B) The observed MSrved molecular mass of 49824.49 Da for intact protein, which is ∼2 kDa less thanay be due to fragmentation of glycan chains and desialylation of the glycoprotein

shown in Fig. 5A and B. These results suggested that recombinant�1-PI expressed in E. coli is fully intact with complete N-terminalsequence and molecular mass is similar to the native unglycosy-lated human protein purified from serum.

3.5. Biological activity and thermostability of recombinant ˛1-PIprotein

Biological activity of E. coli expressed recombinant �1-PI pro-tein(s) was monitored by residual PPE activity assay that showedefficient inhibition of elastase activity with specific activity of about0.45 ± 0.04 for wild type �1-PI after two cycles of amylose affinitychromatography. Biological activity of factor Xa cleaved recombi-nant �1-PI protein was found relatively enhanced in comparison tofusion complex (Table 1). The specific activities of the cleaved �1-PIvariant proteins were significantly enhanced up to 0.68 ± 0.07 for�1-PIFC, 0.61 ± 0.05 for �1-PIFL, 0.52 ± 0.06 for �1-PIAG, 0.45 ± 0.03

for �1-PIMV and 0.47 ± 0.04 for �1-PIMI, which suggest the stabi-lizing effects of incorporated mutations in the expressed protein(Table 2).

Thermal stability analysis of the purified wild type recombi-nant �1-PI and variant �1-PI proteins was performed at 37 ◦C and

70 S. Agarwal et al. / Journal of Biotechnology 147 (2010) 64–72

Table 2Yield, recovery and specific activities of purified recombinant �1-PI and variants from E. colia.

Construct �1-PI variants Protein content (mg) Specific activitye

Total soluble proteinb Recombinant �1-PIc Biologically active �1-PId

pMPI WT 8.38 7.25 ± 0.28 3.77 ± 0.34 0.45 ± 0.04pMFC F51C 9.48 8.06 ± 0.08 6.45 ± 0.66 0.68 ± 0.07pMFL F51L 10.28 8.74 ± 0.11 6.27 ± 0.51 0.61 ± 0.05pMAG A70G 8.71 7.41 ± 0.17 4.53 ± 0.52 0.52 ± 0.06pMMV M358V 7.94 6.75 ± 0.23 3.57 ± 0.24 0.45 ± 0.03pMMI M374I 10.64 8.12 ± 0.22 5.01 ± 0.43 0.47 ± 0.04

a Determined for factor Xa cleaved recombinant �1-PI protein(s) after final purification step.

ity assolubl

5etmar�twc�tTmotow(

Fiaeps

b Total soluble protein determined by Bradford assay.c Recombinant protein estimated by DAC-ELISA using anti �1-PI antibody.d Biological activity (elastase inhibitory capacity) estimated by residual PPE active Specific activity is based on amount of biologically active �1-PI per mg of total s

4 ◦C. Most mutant proteins except �1-PIMV showed significantlynhanced stability at 37 ◦C over time as compared to the wildype recombinant �1-PI. Only 20% loss in activity was observed for

utant variants after 150 min of incubation as compared to >35%nd 80% loss of protein activity in M358 V variant and �1-PIWTespectively, while activity of purified glycosylated human native1-PI remained unaffected (Fig. 6A). The results of heat inactiva-

ion of protein activity at 54 ◦C showed that activity of recombinantild type �1-PI was 50% inhibited after 7.5 min of incubation and

ompletely lost after 30 min of treatment, whereas recombinant1-PI variant proteins showed significantly enhanced resistance to

hermal inactivation in comparison to wild type �1-PI (Fig. 6B).he variant �1-PI protein having F51C substitution showed maxi-um protection to thermal inactivation exhibiting 50% inhibition

f activity after 24 min of incubation, followed by �1-PI substi-uted with F51L, A70G, M374I and M358 V, respectively. Retentionf maximum biological activity after 60 min of incubation at 54 ◦Cas obtained in variant protein �1-PIFC (11%), followed by �1-PIFL

10.2%), �1-PIAG (9.8%), �1-PIMI (7.1%) and �1-PIMV (1.4%), while

ig. 6. Thermostability analysis of E. coli expressed recombinant �1-PI protein andts variants at 37 ◦C and 54 ◦C. (A) Inactivation curves of native human serum �1-PInd cleaved �1-PI protein(s) purified from recombinant BL21CP having wild typexpression vector (pMPI) and vectors having single point mutations (pMFC, pMFL,MAG, pMMV and pMMI) at 37 ◦C. (B) Heat inactivation analysis at 54 ◦C for theame. Approximately 5 �g of purified �1-PI protein was used in each sample.

ay.e protein.

glycosylated native human serum �1-PI lost only 20% of its activityafter 60 min of treatment (Fig. 6B).

4. Discussion

We have demonstrated high-level expression of plant codon-optimized modified synthetic �1-PI gene and its variants consistingrare codons in E. coli under optimized culture conditions. The mod-ified �1-PI genes were expressed as MBP fusion protein, driven byptac promoter for high-level expression together with overcom-ing the limitations of insolubility, misfolding-mediated instability,degradation or aggregation of recombinant foreign protein (Kapustand Waugh, 1999). Comparative expression profile showed consis-tently maximum expression of recombinant �1-PI and its variantproteins (22–24% of TSP) in BL21CP strain of E. coli. This strain car-ries extra copies of tRNA genes for arginine, isoleucine and leucine,therefore, may augment for rare codons in the modified �1-PI geneand leads to higher level of expression. Similar pattern of expressionwas observed when a plant codon-optimized synthetic cry gene ofBacillus thuringiensis was expressed as NusA fusion protein in E. colicarrying copies for rare tRNA codons (Kumar et al., 2005). In addi-tion, we have also observed expression of fusion protein up to 18%and 15% of TSP in TB1 and DH5� strains respectively. This level ofrecombinant �1-PI expression in E. coli strains without any extracopies of tRNA genes could be attributed to the optimum cultureconditions and also to the presence of N-terminal MBP fusion tagthat increases stability of expressed mRNA, translational efficien-cies (Kapust and Waugh, 1999; Lian et al., 2009) and also protectingthe passenger protein from intracellular proteolysis (Bach et al.,2001). Similar kind of observations has been reported owing toN-terminus addition of GST fusion tag that significantly increasedthe expression of genes containing 33% of rare codons (Wu andOppermann, 2003; Wu et al., 2004). The N-terminal fusion of MBPwith �1-PI coding sequence leads to late emergence of rare codonthat mimics stabilizing effect for translational complex formationand allows formation of specific secondary structure for increasedgene expression (Goldman et al., 1995; Gursky and Beabealashvilli,1994). Over expression of ScFv fraction of antibodies and otherheterologous proteins in E. coli as MBP or GST fusion, however,support our results and hypothesis (Bird et al., 2004; Esposito andChatterjee, 2006; Bach et al., 2001).

Determination of exact molecular mass and primary structure ofthe expressed recombinant protein is essential for molecular iden-tification and characterization. Our results of mass spectrometricanalyses and peptide mass fingerprinting of the purified recom-

binant �1-PI expressed in E. coli and native human serum �1-PIshowed fully intact state of recombinant �1-PI protein with identi-cal N-terminal sequence and molecular mass. These results reflectthe structural integrity of the recombinant �1-PI expressed in E. coliand retention of corresponding specific biological activity, which

Biotec

an

ciiBPpi1ohfwlTbrfiaibip

oadia(cicTcsttltmoTtshrtclcMieioCtrca

sr

S. Agarwal et al. / Journal of

re the critical parameter for therapeutic application of recombi-ant proteins (Karnaukhova et al., 2006; Hepner et al., 2005).

Major limitations for expression of heterologous proteins in E.oli are misfolding-mediated aggregation, formation of insolublenclusion bodies, loss of specific biological activity and instabil-ty due to lack of post-translational modifications (Makrides, 1996;ird et al., 2004; Lian et al., 2009). Expression of human gene for �1-I in E. coli has been attempted earlier to achieve biologically activerotein. However, activity of recombinant �1-PI protein expressed

n E. coli was decreased due to protein oligomerization (Kwon et al.,995), N-terminal extension of the protein (Courtney et al., 1984),r very complex purification procedure (Bischoff et al., 1991). Weave expressed recombinant �1-PI in E. coli strain BL21CP as MBP

usion protein and achieved higher expression (upto 24% of TSP),ith enhanced solubility, retention of structural integrity, high bio-

ogical activity and rapid purification on amylose affinity column.he purified recombinant �1-PI was unglycosylated and less sta-le but showed high biological activity since glycosylation is notequired for its activity (Courtney et al., 1984). We have developedve variants of the protein with single amino acid substitutionst critical domains to increase the stability of the molecule. Then-vivo half-life of unglycosylated protein can also be increasedy PEGylation that prevents rapid clearance from the blood after

ntravenous infusion and decreases immunogenicity of therapeuticroteins (Cantin et al., 2002).

Enhanced stability, structural integrity and functional efficacyf recombinant �1-PI and other therapeutic proteins expressed inlternate host are of major concern and always in high demand toecipher the mechanism of proteinase inhibition, folding, misfold-

ng and aggregation for developing therapeutic strategies (Cabritand Bottomley, 2004). The native form of serine protease inhibitorsserpins) exist in metastable strained state that reflect on the effi-acy of their biological activity. However, this could be relieved byncorporation of different single amino acid substitutions at spe-ific sites for more stable conformation of serpins (Im et al., 2004).herefore five variants of �1-PI protein were developed via spe-ific point mutations in strategic domains to relieve the strainedtate of native �1-PI protein. Phe51 and Met374 residues lie inhe breach domain as strands of sheet-B at hydrophobic core ofhe protein molecule. Substitution of these positions with smallerinear aliphatic residues like Cys, Leu and Ileu would decreasehe size of chains around hydrophobic core of �1-PI and allow

ore freedom and improved tertiary packing preventing openingf sheet-A and RCL insertion (Kwon et al., 1994; Im et al., 2004).his would eventually increase the stability of native �1-PI pro-ein. Increased backbone freedom is another option for stabilizationtate of inhibitory serpins. Ala70 is located at the beginning of C-elix and its substitution with small flexible residue like Gly mayesult in better packing of proximal residues, thereby increasinghermostability of serpins by releasing the energy constraint asso-iated with Ala70 (Im et al., 2004). Oxidation of Met358 residueocated at P1 position of RCL domain of �1-PI results into signifi-ant loss of inhibitory activity for elastase, therefore, replacement ofet358 with Val, which is relatively refractory to oxidation should

mprove efficiency for elastase inhibition (Travis et al., 1985; Levinat al., 2009). Most of the substitutions engineered in serpins tomprove their stability and efficiency for proteolytic destructionf elastase were resulted into increased thermostability as well.onsidering this as an important parameter we have comparedhermostability of wild type and variant �1-PI molecules and ouresults are in agreement with earlier reports and suggest signifi-

ance of point mutations at critical sites to improve the biologicalctivity and stability of �1-PI (Kwon et al., 1994; Im et al., 2004).

To conclude, our results have demonstrated high-level expres-ion and simple purification of stable and biologically activeecombinant �1-PI protein and its variants from E. coli. This

hnology 147 (2010) 64–72 71

approach is quick, cost-effective and especially suitable for design-ing expression strategies in structural and functional genomicsfor expressing therapeutic proteins in alternative heterologoushosts. In addition, with the availability of extensive studies ondifferent mutated �1-PI, it will be possible to understand the molec-ular mechanisms of proteinase inhibition, folding, misfolding andaggregation towards developing therapeutic strategies. This typeof protein engineering opens a new dimension in the area of ther-apeutic protein production with increased stability and efficacy.

Acknowledgements

We are grateful to Director, NBRI, Lucknow for infrastructuralsupport, encouragement and valuable suggestions. We thankfullyacknowledge Council of Scientific and Industrial Research (CSIR),India for providing funds and senior research fellowships to SA andSJ. This work was carried out under the CSIR Network Project CMM0004.

References

Agarwal, S., Singh, R., Sanyal, I., Amla, D.V., 2008. Expression of modified gene encod-ing functional human alpha-1-antitrypsin protein in transgenic tomato plants.Transgenic Res. 17, 881–896.

An, Y., Ji, J., Wu, W., Lv, A., Huang, R., Wei, Y., 2005. A rapid and efficient methodfor multiple-site mutagenesis with a modified overlap extension PCR. Appl.Microbiol. Biotechnol. 68, 774–778.

Bach, H., Mazor, Y., Shaky, S., Shoham-Lev, A., Berdichevsky, Y., Gutnick, D.L., Benhar,I., 2001. Escherichia coli maltose-binding protein as a molecular chaperone forrecombinant intracellular cytoplasmic single-chain antibodies. J. Mol. Biol. 312,79–93.

Baneyx, F., 1999. Recombinant protein expression in Escherichia coli. Curr. Opin.Biotechnol. 10, 411–421.

Bird, P.I., Pak, S.C., Worral, D.M., Bottomley, S.P., 2004. Production of recombinantserpins in Escherichia coli. Methods 32, 169–176.

Bischoff, R., Speck, D., Lepage, P., Delatre, L., Ledoux, C., Brown, S.W., Roitsch, C., 1991.Purification and biochemical characterization of recombinant �1-antitrypsinvariants expressed in Escherichia coli. Biochemistry 30, 3464–3472.

Blank, C.A., Brantley, M., 1994. Clinical features and molecular characteristics of�-1-antitrypsin deficiency. Ann. Allergy 72, 105–121.

Bradford, M.M., 1976. A rapid and sensitive method for the quantitation of proteinutilizing the principle of protein–dye binding. Anal. Biochem. 72, 248–254.

Cabrita, L.D., Bottomley, S.P., 2004. How do proteins avoid being too stable? Bio-physical studies into metastable proteins. Eur. Biophys. J. 33, 83–88.

Cantin, A.M., Woods, D.E., Cloutier, D., Dufour, E.K., Leduc, R., 2002. Polyethyleneglycol conjugation at Cys 232 prolongs the half life of �1-proteinase inhibitor.Am. J. Respir. Cell. Mol. Biol. 27, 659–665.

Courtney, M., Buchwalder, A., Tessier, L.H., Jaye, M., Benavente, A., Balland, A., Kohli,V., Lathe, R., Tolstoshev, P., Lecocq, J.P., 1984. High-level production of biologi-cally active human �1-antitrypsin in Escherichia coli. Proc. Natl. Acad. Sci. U.S.A.81, 669–673.

Esposito, D., Chatterjee, D.K., 2006. Enhancement of soluble protein expressionthrough the use of fusion tags. Curr. Opin. Biotechnol. 17, 353–358.

Goldman, E., Rosenberg, A.H., Zubay, G., Studier, F.W., 1995. Consecutive low-usageleucine codons block translation only when near the 5′ end of a message inEscherichia coli. J. Mol. Biol. 245, 467–473.

Gursky, Y.G., Beabealashvilli, R., 1994. The increase in gene expression induced byintroduction of rare codons into the C terminus of the template. Gene 148, 15–21.

Hepner, F., Cszasar, E., Roitinger, E., Lubec, G., 2005. Mass spectrometrical analy-sis of recombinant human growth hormone (Genotropin) reveals amino acidsubstitutions in 2% of the expressed protein. Proteome Sci. 3 (1), 1–12.

Heresi, G.A., Stoller, J.K., 2008. Augmentation therapy in alpha-1 antitrypsin defi-ciency. Expert Opin. Biol. Ther. 8, 515–526.

Im, H., Ryu, M.J., Yu, M.-H., 2004. Engineering thermostability in serine proteaseinhibitors. Protein Eng. Des. Sel. 17, 325–331.

Kane, J.F., 1995. Effects of rare codon clusters on high-level expression of heterolo-gous proteins in Escherichia coli. Curr. Opin. Biotechnol. 6, 494–500.

Kapust, R.B., Waugh, D.S., 1999. Escherichia coli maltose-binding protein is uncom-monly effective at promoting the solubility of polypeptides to which it is fused.Protein Sci. 8, 1668–1674.

Karnaukhova, E., Ophir, Y., Golding, B., 2006. Recombinant human alpha-1 pro-teinase inhibitor: towards therapeutic use. Amino Acids 30, 317–332.

Kumar, S., Birah, A., Chaudhary, B., Burma, P.K., Gupta, G.P., Pental, D., 2005. Plant

codon-optimized cry genes of Bacillus thuringiensis can be expressed as solubleproteins in Escherichia coli BL21 Codon Plus strain as NusA-Cry protein fusions.J. Invertebr. Pathol. 88, 83–86.

Kwon, K.-S., Kim, J., Shin, H.S., Yu, M.-H., 1994. Single amino acid substitutions of�1-antitrypsin that confer enhancement in thermal stability. J. Biol. Chem. 269,9627–9631.

7 Biotec

K

L

L

L

L

L

M

S

codon genes from hyperthermophilic archaea by the GST gene fusion system. J.

2 S. Agarwal et al. / Journal of

won, S.-K., Lee, S., Yu, M.-H., 1995. Refolding of �1-antitrypsin expressed as inclu-sion bodies in Escherichia coli: characterization of aggregation. Biochim. Biophys.Acta 1247, 179–184.

aemmli, U.K., 1970. Cleavage of structural protein during the assembly of the headof bacteriophage T4. Nature 227, 680–685.

ee, K.N., Park, S.D., Yu, M.-H., 1996. Probing the native strain in �1-antitrypsin. Nat.Struct. Biol. 3, 497–500.

evina, V., Dai, W., Knaupp, A.S., Kaiserman, D., Pearce, M.C., Cabrita, L.D., Bird, P.I.,Bottomley, S.P., 2009. Expression, purification and characterization of recom-binant Z �1-antitrypsin—the most common cause of �1-antitrypsin deficiency.Protein Exp. Purif. 68, 226–232.

ian, J., Ding, S., Cai, J., Zhang, D., Xu, Z., Wang, X., 2009. Improving aquaporin Zexpression in Escherichia coli by fusion partners and subsequent condition opti-mization. Appl. Microbiol. Biotechnol. 82 (3), 463–470.

omas, D.A., 2005. Molecular mousetraps, alpha-1-antitrypsin deficiency and theserpinopathies. Clin. Med. 5, 249–257.

akrides, S.C., 1996. Strategies for achieving high-level expression of genes inEscherichia coli. Microbiol. Rev. 60, 512–538.

ambrook, J., Russel, D.W., 2001. Molecular Cloning: A Laboratory Manual. ColdSpring Harbor Laboratory Press, New York.

hnology 147 (2010) 64–72

Shevchenko, A., Tomas, H., Havlis, J., Olsen, J.V., Mann, M., 2006. In-gel digestionfor mass spectrometric characterization of proteins and proteomes. Nat. Prot. 1,2856–2860.

Stemmer, W.P.C., Crameri, A., Ha, K.D., Brennan, T.M., Heyneker, H.L., 1995.Single-step assembly of a gene and entire plasmid from large numbers ofoligodeoxyribonucleotides. Gene 164, 49–53.

Travis, J., Owen, M., George, P., Carrel, R., Rosenberg, S., Hallewell, R.A., Barr, P.J.,1985. Isolation and properties of recombinant DNA produced variants of human�1-protinase inhibitor. J. Biol. Chem. 260 (7), 4384–4389.

Whisstock, J.C., Skinner, R., Carrell, R.W., Lesk, A.M., 2000. Conformational changesin serpins. I. The native and cleaved conformations of �1-antitrypsin. J. Mol. Biol.295, 651–665.

Wu, X., Oppermann, U., 2003. High-level expression and rapid purification of rare-

Chromatogr. B 786, 177–185.Wu, X., Jornvall, H., Berndt, K.D., Oppermann, U., 2004. Codon optimization reveals

critical factors for high level expression of two rare codon genes in Escherichiacoli: RNA stability and secondary structure but not tRNA abundance. Biochem.Biophys. Res. Commun. 313, 89–96.