Analytical Biochemistry 418 (2011) 155–157

Contents lists available at ScienceDirect

Analytical Biochemistry

journal homepage: www.elsevier .com/locate /yabio

Notes & Tips

Development of an enzyme-linked immunosorbent assay analytical platformfor refolding yield determination of recombinant hepatitis B virus X (HBx) protein

Anindya Basu, Susanna Su Jan Leong ⇑School of Chemical and Biomedical Engineering, Nanyang Technological University, 62 Nanyang Drive, Singapore

a r t i c l e i n f o a b s t r a c t

Article history:Received 15 April 2011Received in revised form 11 July 2011Accepted 13 July 2011Available online 21 July 2011

Keywords:Hepatitis B virus X (HBx) proteinHepatitis BELISARefolding yieldSoluble yield

0003-2697/$ - see front matter � 2011 Elsevier Inc. Adoi:10.1016/j.ab.2011.07.014

⇑ Corresponding author. Fax: +65 6794 7553.E-mail address: Sleong@ntu.edu.sg (S.S.J. Leong).

1 Abbreviations used: DTT, dithiothreitol; GST, gluhepatitis B virus; HBx, hepatitis B virus X; HCC, he6-histidine; nRTIs, nucleos(t)idic reverse transcriptase

We report the development of a novel ELISA platform to quantitate hepatitis B virus X (HBx) proteinrefolding yields, which is critical for rational design and scaleup of aHBx bioprocess. HBx refolding yieldswere measured by determining the amount of HBx bound to immobilized GST–p53 on a ‘‘reduced gluta-thione’’-functionalized maleimide surface. Refolding yields were distinguished from soluble yields, whichwere determined by measuring total HBx protein bound to a maleimide surface under reducing condi-tions. This platform is amenable to scaleup, and will expedite HBx production for structural and clinicalstudies, leading to the development of HBx-based therapy for liver cancer.

� 2011 Elsevier Inc. All rights reserved.

Hepatitis B virus (HBV)1 infection has been found to be a majorcausative factor for the initiation and development of hepatocellularcarcinoma (HCC), a killer disease affecting over 350 million livesevery year [1,2]. Currently the seven medications approved in theUnited States for treatment of this condition include (i) two type 1interferons, and (ii) five nucleos(t)idic reverse transcriptase inhibi-tors (nRTIs), none of which have been found to be able to completelyeradicate the infection [3]. Furthermore, long-term usage of thesemedications, particularly the nRTIs, may lead to the developmentof virus resistance and cross-tolerance, thereby rendering thesestrategies ineffective. Therefore, new drug therapeutic strategiesare urgently needed to reduce patient suffering and mortality ratesassociated with HCC [4]. Over the years, different studies have indi-cated that the hepatitis B virus X (HBx) protein expressed by the HBVis strongly associated with HCC development. Various studies havesuggested the involvement of HBx in HBV transcription, replication,and interaction with host cellular machineries to support its prolif-eration [5–7], leading to strong interest in developing HBx-targetedtherapeutic strategies for HCC treatment. However, the real bottle-neck in identifying a structure-based lead against HBx is the lackof knowledge on the HBx structure [5], thereby making its structuralcharacterization through homology modeling and subsequent drugdesigning studies by docking a difficult task [8]. Furthermore, the

ll rights reserved.

tathione S transferase; HBV,patocellular carcinoma; 6His,inhibitors.

low inherent expression of the protein within the infected host cellsmakes its structural characterization, from natively isolated proteins,an impossible endeavor. Biosynthesis of the HBx protein in high-yielding bacteria platforms thus seems to be an obvious solution.However, previous attempts to recombinantly produce the proteinin different expression systems, such as Escherichia coli, have consis-tently resulted in the production of insoluble HBx [9–12], therebynecessitating a refolding-based bioprocess. Although the in vitroactivity of recombinantly produced HBx protein has been reportedin previous studies [13,14], no quantitative method for HBx refoldingyield determination has been reported, which is a large hindrance toHBx bioprocess design and scaleup efforts. Confronted by this road-block, this study aims to develop an ELISA platform for rapid andquantitative refolding yield determination of a 6-histidine (His)-tagged HBx (6His–HBx). In this study, the ELISA analytical platformwas developed based on the well-studied mechanism of HBx cellularactivities which are mediated through various protein–protein inter-actions [15]. Among the different cellular tumor suppressor factorsmodulated by HBx, its interaction with the tumor suppressor proteinp53 is the most well established [7,15–19]. Therefore the HBx–p53interaction which correlates with the biological activity of HBx formsthe basis of our ELISA platform design.

A major hurdle for the development of a quantitative analyticalplatform for a protein like HBx is the absence of native standardsneeded for generating the calibration curves for quantifying theamount of HBx bound on the ELISA plates. In this study, we showhow this drawback can be overcome by using the 6His tag as acommon platform for detecting the amount of 6His–HBx boundbased on their interaction with 6His antibodies. In the following

Fig.1. Schematic of the developed ELISA platform to quantify HBx refolding and soluble yields. The marked 6His tags were used as common platform for detection of boundproteins. (a) Strategy to construct the calibration curve: 6His–GST protein at different known concentrations were covalently attached to the maleimide surface andincubated with the HRP-conjugated 6His primary antibody. On addition of the TMB substrate, the intensity of the color development was measured by recording absorbanceat 650 nm, where increasing absorbance correlates with increasing 6His primary antibody concentration and 6His proteins bound on the plate. (b) Strategy to determine HBxrefolding yields: GST–p53 protein was first bound to a GSH-functionalized maleimide surface, to which the refolded HBx sample was added. The amount of HBx proteinbound to GST–p53 was determined by incubation with HRP-conjugated 6His primary antibody. (c) Soluble HBx yield was determined by covalently binding the HBx proteinsto the maleimide surface (under mild reducing conditions), followed by incubation with the HRP-conjugated 6His primary antibody for subsequent detection. N.B. Theschematic diagram may not reflect the true binding stoichiometry between the different proteins.

Fig.2. ELISA results: (a) Calibration curve obtained by covalent immobilization of 6His–GST on the maleimide surface; (b) HBx-soluble and refolding yields obtained from theELISA platform; (c) comparison of the specificity of HBx binding to GST–p53 and BSA. All data represented within the figure show the mean ± standard deviation of theobserved values.

156 Notes & Tips / Anal. Biochem. 418 (2011) 155–157

paragraphs, we discuss the rationale and strategies adopted forconstructing the calibration curve for our ELISA, followed by deter-mination of HBx-soluble and refolding yields.

Construction of a calibration curve to quantify bound 6His–HBx: Acalibration curve to quantify the amount of bound 6His–HBx onthe ELISA plate was constructed using commercially available6His–GST (glutathione S transferase) proteins (Abcam, USA). First,6His–GST proteins at varying known concentrations were cova-lently bound to the maleimide-functionalized plates (SulfhydrylBind, Corning, USA), shown schematically in Fig. 1a. As per thesupplier’s instructions, the 6His–GST protein solutions were pre-pared in Tris–EDTA (TE) buffer (pH 6.5) containing 1 � 10�3 mMDTT, before being added to the plates. After suitable washing andblocking steps, a horseradish peroxidase (HRP)-tagged 6His pri-mary antibody (Abcam, USA) was used to quantify the amount of6His–GST proteins bound on the plate surface. After 1.5 h of incu-bation with the antibody, equal volumes of TMB substrate develop-ing solution (Millipore, Singapore) were added to the wells, and thesample absorbance of each well was measured using a microplatereader. Reverse-phase HPLC analysis confirmed that all the addedGST proteins bound to the maleimide surface (data shown in thesupplementary material). The calibration curve which correlatesabsorbance units to the amount of 6His–GST proteins (nM) bound

(Fig. 2a) can thus be used to quantitate the amount of 6His–HBximmobilized on the ELISA plate.

Determination of HBx refolding yield: HBx refolding yield is deter-mined based on the interaction of correctly refolded HBx with p53.To enable binding of the GST-tagged p53 to the plate surface, reducedglutathione (GSH) was first immobilized on a maleimide-functional-ized microplate via covalent binding of the free cysteine residues ofGSH and the maleimide surface (Fig. 1b). Incubation of GST–p53 pro-tein allows binding of the GST protein to GSH, and hence avails thep53 molecule to bind with biologically active HBx proteins whichare present within the refolding samples. Using this strategy, HBxrefolding yields were determined from the equation

Refolding Yield ð%Þ

¼ ½Mass of refolded HBx bound to GST� p53� � 100%

½Mass of denatured reduced HBx� ð1Þ

The GSH-coated ELISA plate surface was first incubated with50 ll of 10 lg/ml GST–p53 protein solution (Millipore, Singapore);binding of the GST–p53 protein to this surface was confirmed withreverse-phase HPLC analysis (data provided in the supplementarymaterial). After blocking this surface with 5% nonfat milk, refolded6His–HBx samples, purified and refolded using the same methodsreported in our earlier work [20], were introduced to the ELISA

Notes & Tips / Anal. Biochem. 418 (2011) 155–157 157

plate. Refolded HBx samples which were incubated over differenttime intervals in the refolding buffer (2 M urea, 0.25 M arginine,0.1 mM GSH, and 0.01 mM GSSG in 50 mM Tris at pH 7.4) were de-salted into 50 mM Tris buffer (pH 7.4) and diluted 1000 times be-fore being loaded onto the ELISA plate. A fixed refoldingconcentration of 0.1 mg/ml was used throughout this study. Todetermine the total denatured-reduced HBx concentration (Eq.(1)), solubilized HBx protein (in 8 M urea and 10 mM DTT) was firstdiluted in a denaturing buffer (containing 8 M urea) to refoldingconcentrations, and then diluted 1000-fold in TE buffer (pH 7.4)for loading onto the ELISA plate. Denatured-reduced HBx was thuscovalently bound on the maleimide surface under reducing condi-tions (Fig. 1c). Carryover of DTT did not interfere with binding ofdenatured-reduced HBx to the maleimide surface which wasdetermined by performing a protein mass balance (data notshown). Specificity of refolded HBx binding to GST–p53 was testedby measuring the ability of the refolded HBx to bind to BSA whichwas covalently immobilized on the maleimide surface at 1 mg/ml(detailed protocol presented in the supplementary material).

Determination of HBx soluble yields: The design of our ELISA plat-form also aimed to distinguish HBx refolding yields from HBx-sol-uble yields. Equation (2) was used to measure HBx soluble yields:

Soluble Yield ð%Þ

¼ ½Mass of HBx bound on the maleimide surface� � 100%

½Mass of total denatured reduced HBx�ð2Þ

To determine the mass of HBx bound on the maleimide surfacefor determination of HBx soluble yields, the refolded HBx samplewas incubated in 1 � 10�3 mM DTT to enable covalent bindingbetween the free cysteine residues of HBx and the maleimide sur-face. The amount of HBx bound was determined by using the 6Histag antibody detection strategy as detailed above.

The results of quantitating soluble and refolding yields ofHBx are summarized in Fig. 2b. Fig. 2c compares the amountof HBx protein bound to GST–p53 with nonspecific binding of theHBx protein to BSA. Nonspecific binding of HBx to BSA (surfaceimmobilized at 1 mg/ml) for each point was found to be signifi-cantly smaller (P < 0.01, data tested with Student’s t test) com-pared to HBx binding to GST–p53 (surface-immobilized at 10 lg/ml). As expected, HBx refolding yield increased with time (up to3 days), followed by a decrease in yield, presumably due to unpro-ductive interaction between misfolded proteins and the refoldedprotein and/or the oxidation of GSH [20].

By using a common platform that quantitates the interactionbetween 6His-tag and 6His-antibodies, this study reports thedevelopment of a novel ELISA methodology to determine HBxrefolding and soluble yields based on the specific interaction ofHBx with the tumor suppressor protein, p53 [17,18]. Since thereare no high-purity HBx standards that are commercially availableto date, this ELISA platform will be instrumental in facilitating fu-ture design, optimization, and scaleup studies of HBx bioprocesses.The maleimide attachment chemistry [21] allowed us to create aGSH surface that was needed to immobilize the p53 protein mole-cules so that it remains readily available for binding with the HBx.The success of our strategy is well reflected not only by the repro-ducibility of the values at each data point (coefficient of variation<6% for all the measured soluble and refolded HBx concentrationmeasurements), but also the data trends which agreed well with

our previous work on HBx solubility studies based on second virialcoefficient measurements [20]. To the best of our knowledge, thisis the first report on the development of an ELISA-based analyticalplatform for evaluation of recombinant HBx bioactivity, where thedistinction between the soluble and the refolding yields of the HBxprotein could be established.

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at doi:10.1016/j.ab.2011.07.014.

References

[1] D. Lavanchy, Hepatitis B virus epidemiology, disease burden, treatment, andcurrent and emerging prevention and control measures, J. Viral Hepatitis 11(2004) 97–107.

[2] D.M. Parkin, P. Pisani, J. Ferlay, Estimates of the worldwide incidence of 25major cancers in 1990, Int. J. Cancer 80 (1999) (1990) 827–841.

[3] A.S. Lok, Navigating the maze of hepatitis B treatments, Gastroenterology 132(2007) 1586–1594.

[4] M. Nassal, New insights into HBV replication: new opportunities for improvedtherapies, Future Virol. 4 (2009) 55–70.

[5] S. Benhenda, D. Cougot, M.A. Buendia, C. Neuveut, Chapter 4 hepatitis B virus Xprotein. Molecular functions and its role in virus life cycle and pathogenesis,Adv. Cancer Res. 103 (2009) 75–109.

[6] L.L. David, H. Nguyen, Jianming Hu, Hepatitis B virus–cell interactions andpathogenesis, J. Cell. Physiol. 216 (2008) 289–294.

[7] H. Tang, N. Oishi, S. Kaneko, S. Murakami, Molecular functions and biologicalroles of hepatitis B virus x protein, Cancer Sci. 97 (2006) 977–983.

[8] S. Mandal, M. Moudgil, S.K. Mandal, Rational drug design, Eur. J. Pharmacol.625 (2009) 90–100.

[9] A. Gupta, T.K. Mal, N. Jayasuryan, V.S. Chauhan, Assignment of disulphidebonds in the X protein (HBx) of hepatitis B virus, Biochem. Biophys. Res.Commun. 212 (1995) 919–924.

[10] I. Marczinovits, C. Somogyi, A. Patthy, P. Németh, J. Molnár, An alternativepurification protocol for producing hepatitis B virus X antigen on a preparativescale in Escherichia coli, J. Biotechnol. 56 (1997) 81–88.

[11] G.Y. Hwang, C.Y. Lin, L.M. Huang, Y.H. Wang, J.C. Wang, C.T. Hsu, S.S. Yang, C.C.Wu, Detection of the hepatitis B virus X protein (HBx) antigen and anti-HBxantibodies in cases of human hepatocellular carcinoma, J. Clin. Microbiol. 41(2003) 5598–5603.

[12] S.M. Zhang, D.C. Sun, S. Lou, X.C. Bo, Z. Lu, X.H. Qian, S.Q. Wang, HBx protein ofhepatitis B virus (HBV) can form complex with mitochondrial HSP60 andHSP70, Arch. Virol. 150 (2005) 1579–1590.

[13] D. Liu, L. Zou, W. Li, L. Wang, Y. Wu, High-level expression and large-scalepreparation of soluble HBx antigen from Escherichia coli, Biotechnol. Appl.Biochem. 54 (2009) 141–147.

[14] E. Rui, P.R. De Moura, K. de Almeida Gonçalves, J. Kobarg, Expression andspectroscopic analysis of a mutant hepatitis B virus onco-protein HBx withoutcysteine residues, J. Virol. Methods 126 (2005) 65–74.

[15] M.C. Kew, Hepatitis B virus x protein in the pathogenesis of hepatitis B virus-induced hepatocellular carcinoma, J. Gastroenterol. Hepatol. 26 (2011) 144–152.

[16] M.A. Feitelson, M. Zhu, L.X. Duan, W.T. London, Hepatitis B x antigen and p53are associated in vitro and in liver tissues from patients with primaryhepatocellular carcinoma, Oncogene 8 (1993) 1109–1117.

[17] P.R. de Moura, E. Rui, K. de Almeida Gonçalves, J. Kobarg, The cysteine residuesof the hepatitis B virus onco-protein HBx are not required for its interactionwith RNA or with human p53, Virus Res. 108 (2005) 121–131.

[18] Y. Lin, H. Tang, T. Nomura, D. Dorjsuren, N. Hayashi, W. Wei, T. Ohta, R. Roeder,S. Murakami, The hepatitis B virus X protein is a co-activator of activatedtranscription that modulates the transcription machinery and distal bindingactivators, J. Biol. Chem. 273 (1998) 27097–27103.

[19] R. Truant, J. Antunovic, J. Greenblatt, C. Prives, J.A. Cromlish, Direct interactionof the hepatitis B virus HBx protein with p53 leads to inhibition by HBx of p53response element-directed transactivation, J. Virol. 69 (1995) 1851–1859.

[20] A. Basu, W. Chen, S. Leong, A rational design for hepatitis B virus X proteinrefolding and bioprocess development guided by second virial coefficientstudies, Appl. Microbiol. Biotechnol. 90 (2011) 181–191.

[21] G.T. Hermanson, Bioconjugate Techniques, Elsevier Academic Press,Amsterdam, 2008.