adenoviral p53 effects and cell-specific e7 protein–protein interactions of human cervical cancer...

Biosensors and Bioelectronics 20 (2005) 2236–2243

Adenoviral p53 effects and cell-specific E7 protein–proteininteractions of human cervical cancer cells�

Jeong-Woo Choia, Woong Shick Ahnb, Su Mi Baec, Doo-Bong Leea, Yong-Wan Kima,∗a Department of Chemical and Biomolecular Engineering, and Interdisciplinary Program of Integrated Biotechnology Sogang University,

1 Sinsu-Dong, Mapo-Gu, Seoul 121-742, Republic of Koreab Department of Obstetrics and Gynecology, College of Medicine, The Catholic University of Korea, Seoul, Republic of Korea

c Catholic Research Institutes of Medical Science, College of Medicine, The Catholic University of Korea, Seoul, Republic of Korea

Received 11 August 2004; received in revised form 30 October 2004; accepted 12 November 2004Available online 2 Febuary 2005

Abstract

We investigated the time-course tumor growth suppression effects of recombinant adenovirus expressing p53 on human cervical cancer cellsand cell-specific E7 protein–protein interactions in cell lysates using surface plasmon resonance (SPR) biosensor. Six HPV-infected humanc negativeC lso, wep te. For eachc uppressionl ll growthi n SiHa andH p53 intoC ansfection.T between E7a ell death,w effects ofr©

K

1

tGpt

m-, in-orya-im-b is

,tedpethepro-

0d

ervical cancer cell lines (HPV 16-positive cells, CaSki and SiHa cells; HPV 18-positive cells, HeLa and HeLaS3 cells; and HPV33A and HT3 cells) were used. After infection with AdCMVp53, the cell-specific growth inhibition was studied in vitro and in vivo. Aroduced the recombinant E7 oncoprotein of HPV 16 type and tested chip-based protein–protein interactions with each cell lysaervical cancer cell, differential cell growth inhibitions were shown via cell count assay and MTT assay. Note that the same trend in sevels was shown in CaSki, HeLa and in SiHa, HeLaS3, respectively. In contrast, infection with AdCMVLacZ showed increased cen a manner similar to the negative control group. The levels of p53 protein were notably expressed in CaSki and HeLa more than ieLaS3 for 4 days. In contrast, p53 expression was continually maintained in C33A and HT3 for 6 days. After transfection AdCMVaSki- and SiHa-xenografted nude mice, the size of tumor was remarkably decreased in SiHa cells as compared to AdCMVLacZ trhe SPR sensor surface was successfully modified with the recombinant E7 oncoprotein and showed cell-specific interactionsnd its target proteins from cell lysates. The anti-tumor effects were accomplished via differential role of p53-specific apoptotic chich is dependent upon the cervical cancer cell line. Also, a molecular level understanding of cell-dependent protein interaction

ecombinant E7 was shown.2004 Elsevier B.V. All rights reserved.

eywords:Cervical neoplasia; Gene therapy; p53 Gene; Surface plasmon resonance

. Introduction

Human papillomavirus (HPV) 16 and 18 have been iden-ified in above 90% of cervical cancer (Shillitoe et al., 1994).enerally, after high-risk HPV infection, E6 and E7 onco-roteins are consistently expressed, and essential for immor-

alization and transformation of human squamous epithelial

� The paper was presented at 8th World Congress on Biosensors 2004.∗ Corresponding author. Tel.: +82 2 3273 0357; fax: +82 2 3273 0331.E-mail address:[email protected] (Y.-W. Kim).

cells (Kaur et al., 1989). The E6 and E7 proteins form coplexes with p53 and retinoblastoma (Rb), respectivelyhibiting the activities of the proteins in cell cycle regulatsystems (McMurray et al., 2001). These selective degradtions of tumor suppressor proteins are strategically veryportant in gene therapy, as the inactivation of p53 and Ressential for the induction of cervical carcinoma (Lee et al.2002). It is well known that cervical cancer cell lines infecwith HPV expressed normal pRB but low-levels of wild typ53. The trials for suppression of cancer cell growth viarestoration of the p53 mutation have been considerably

956-5663/$ – see front matter © 2004 Elsevier B.V. All rights reserved.oi:10.1016/j.bios.2004.11.022

J.-W. Choi et al. / Biosensors and Bioelectronics 20 (2005) 2236–2243 2237

gressed. The p53-dependent apoptosis have been regardedas a useful strategy for treating human cancer (Hamada etal., 1996a,b). Adenoviral vector has been used to transfer aspecific gene into cells (Graham and Prevec, 1995), gener-ating an overwhelming amount of data and mechanism ofp53-mediated inhibition (Huang et al., 2000), but yet remainfar from ideal in clinical cervical cancer therapy as the ef-fects of adenoviral vectors may occur differently in humansthan in animal models. Recently, it is becoming increasinglyuncertain whether there are wide variations in tumorigenicityinhibitory effect among different cell types (Xu et al., 1996).Also, p53 can induce multiple independently regulated apop-totic pathways depending on the physiological circumstancesand cell types, however, the exact molecular mechanism lead-ing to p53-mediated apoptosis are still unclear.

Biosensor technology based on the surface plasmon reso-nance (SPR) technique has been developed for the measure-ment of binding kinetics between a biomolecule in complexbiological media and its binding partner immobilized on asensor surface with high specificity and sensitivity, with ashort detection time, and with simplicity (Oh et al., 2003,2004; Choi et al., 2004). Applications of the SPR biosensorinclude the study of receptor–ligand interactions (Logsdon etal., 2004), the formation of signaling complexes (Hu et al.,2004), epitope mapping (Schlattner et al., 2002) and protein-DNA binding (Huber et al., 2004). However, the enhance-m odyb ns ofb

p53a ervi-c cell-d lso,w ypea t them singt dif-f ellsi tedb ch isd mor-s ntialr

2

2

PV1 itivec neg-a p53g s alsou anka . The

cells were routinely propagated in monolayer culture in Dul-beco’s Modified Eagle’s Medium (DMEM) (GIBCO-BRL,Paisley, Scotland) supplemented with 5% heat-inactivated fe-tal bovine serum (Gibco-BRL), 0.37% sodium bicarbonate,30 mM 2-(4-(2-hydroxyethyl)-1-piperazinyl)ethanesulfonicacid (HEPES), and penicillin/streptomycin (Gibco-BRL).The cells were cultured in a 5% CO2 incubator (Sanyo, Tokyo,Japan) at 37◦C. Unless otherwise specified, all chemicalswere purchased from Aldrich and Sigma Chemical.

2.2. Recombinant adenoviral vector and titration ofAdCMVp53 viral stock

Recombinant adenoviral vector (AdCMVp53) containinghuman wild-type P53 gene was constructed as follows. A1.7 kb wild-type p53 gene with the HCMV promoter se-quence was cloned into theEcoRV andXba1 sites of p�

E1sp1B, and then co-transfected with the adenovirus plas-mid, pBHGE3, into 293 cell, and cultured in complete mediafor about 12 days. The p53 recombinant adenoviruses wereisolated from a single plaque and expanded in a 293 cell,and purified by double CsCl gradient at ultracentrifugation(25,000 rpm, 4◦C, and 2 h). The final infectious titers weredetermined to be 108 plaque-forming units (pfu) per ml. Ad-CMVLacZ was used as previously described (Hamada et al.,1996a,b).

2

asr ine.T RI,a E7D oRIs las-ma puri-fi in-c nmw . Pro-t io-� erec d andp lex( ro-t fol-l tionw

2

12-w nD mlo ered of

ent of sensitivity is required to detect the peptide–antibinding and protein–protein binding, as the concentratioiomolecules in a biological system are extremely low.

In this study, we constructed a recombinant AdCMVnd evaluated the effect of p53 adenoviral vector on cal cancer cells to investigate the molecular pathway ofependent tumor suppression effect in vitro and in vivo. Ae produced the recombinant E7 protein of HPV 16 tnd analyzed its binding partners in cell lysates to geolecular basis of tumor-specific effects in E7-expres

umors. We observe that p53 delivery results in a moreerential growth inhibition pattern in cervical cancer cn vitro and in vivo. This inhibition appears to be mediay p53-dependent apoptosis and cell cycle arrest, whiependent on the cervical cancer cell lines. Also, the tupecific growth suppression effects showed the differeegulations of E7 oncoprotein and p53 expression.

. Materials and methods

.1. Cell lines and culture conditions

Four HPV-infected human cervical cancer cell lines (H6-positive cells, CaSki and SiHa cells; and HPV 18-posells, HeLa and HeLaS3 cells) were used. Also, HPVtive C33A and HT3 cell lines that have a mutation onene were used. Human hepatoma cell line Hep-G2 wased. The cell lines were obtained from the cell line bt Seoul National University’s Cancer Research Center

.3. Production of recombinant E7 protein

The full protein coding region of the HPV 16 E7 gene weverse transcription-PCR amplified from a CaSki cell lhe amplified DNA was digested with BamHI and Econd the resulting DNA fragment was gel purified. TheNA fragments were then cloned into the BamHI and Ecite of the pET vector (Novagen, Madison, WI). The pid construct was transformed intoEscherichia coliDH5�nd selected against kanamycin. pET-E7 vector wased and again transformed into BL21(DE3) cells, andubated in Luria–Bertani broth until absorbance at 600as between 0.6 and 0.8 absorbance units (log phase)

eins were induced by addition of 1 mM isopropyl-1-th-d-galactopyranoside (IPTG) for 3 h. The cell pellets wollected and lysed. The cell supernatants were collecteassed through for the Ni(II)–nitrilotriacetic acid compNi–NTA) resin column according to the manufacturer’s pocol (Qiagen, Valencia, CA). The protein solution wasowed by overnight dialysis in PBS. The final protein soluas stored at−70◦C.

.4. Cell growth suppression and MTT assay

Cervical cancer cell lines were first inoculated fromell plates into 5× 105 to 5× 106 cells/wells, cultured iMEM containing 10% FBS overnight, washed with 1f DPBS, infected with recombinant adenoviruses that wiluted at 31.25, 62.5, 125, 250, and 500 multiplicity

2238 J.-W. Choi et al. / Biosensors and Bioelectronics 20 (2005) 2236–2243

infection (MOI), and cultured for 1 h. Then, the numberof cells was measured under microscopy everyday for thegrowth suppression analysis. The measurement of cell num-ber was repeated three times, and the average was used. For 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide(MTT) assay, after the incubation period, 20�l of 5× MTT(final concentration 0.5 mg/ml to each well) was added tocell cultures in each well and cultured for 4 h in a humidifiedatmosphere. Two hundred microliters of DMSO was addedand shaken for 10 min in a shaker. After complete solubi-lization of the purple formazan crystals, the absorbance wasmeasured with ELISA-reader at 570 nm.

2.5. Western blot

The cell extracts were electrophoresed for 2 h withSDS–PAGE at 10 mA, and blotting was performed for 1 hand 30 min with a Hybond-ECL membrane (Amersham, Up-psala, Sweden) at 100 V. The blotted membrane was blockedwith 5% skim milk and reacted with primary antibody,mouse anti-human p53 DO7 (Novocastra, Newcastle, UK)and secondary antibody, horseradish peroxidase-conjugatedgoat anti-mouse IgG (Novocastra, Newcastle, UK). For eachstep, the membrane was washed with TBST (Tris bufferedsaline with 0.1% Tween-20) three times. Protein bands werevisualized using an ECL kit according to the manufacturer’sp

2

tedi d foreo tot s, tu-m . Thet ulard eraged , andf inw ure’sp

2

atesa llsw tube.T illedp (1%( 4b antso .

SPRs t aw 8

50◦. The instrumental configuration of the laser light source,polarizer, photo multiplier tube (PMT), and attenuated to-tal reflection (ATR) coupler were performed according tothe manufacturer’s protocol. The polarizer permits measure-ments both with p-polarized and s-polarized light, and thep-polarized light is used as a reference. The resolution of theangle reading of the goniometer was 0.001. A BK 7 typecover glass plate (18 mm× 18 mm, Superior, Germany) wasused as the solid support. The self-assembled monolayer of11-mercaptoundecanoic acid (MUA) on the Au surface wasfabricated by submerging the prepared Au substrate into aglycerol/ethanol (1:1, v/v) solution containing 150 mM of11-MUA for at least 12 h. For chemical binding between the11-MUA adsorbed on the Au substrate and E7 protein, thecarboxyl group in 11-MUA was activated by submerging theAu substrate modified with 11-MUA into a solution of 10%1-ethyl-3-(3-dimethylaminopropyl)carbodiimide hydrochlo-ride (EDAC) in water/ethanol (10:1, v/v) for 2 h at room tem-perature. The self-assembled E7 protein layer was fabricatedby the incubation of the activated Au substrate in a solutionof E7 protein in 10 mM phosphate buffer (PBS, pH 7.4) con-taining 0.14 mol/l NaCl and 0.02% (w/v) thimerosal (PBS)at room temperature for 2 h. To associate with the lysates, theE7 protein layer by self-assembly technique was immersedin a solution containing lysates in a PBS buffer. After 4 h ofincubation at 4◦C, the surface was rinsed with a PBS buffer,f eE on,t ntil as e car-rb

2

ent’st erec

3

3

d-C ibi-t e ofc cellc Ski( aS3( wni iHa( andH up-p iHa,H in-

rotocol (Amersham, Arlington Heights, IL).

.6. Inhibition of tumor growth

Cancer cells (107 cells/mouse) in 0.1 ml PBS were injecnto the 6-week-old female mice. Seven mice were useach group. Six days later, a solution containing 5× 109 pfuf AdCMVp53, AdCMVLacZ, or PBS were injected in

he area where tumor cells were generated. For 4 weekor formation and size were evaluated once in 2 days

umors were measured with calipers for two perpendiciameters, and tumor size was calculated based on avimensions. The tumors were resected at indicated day

rozen in−70◦C for analysis. Total RNA and total proteere extracted with Trizol as described in the manufactrotocol for Western analysis.

.7. SPR spectroscopy

Each cell line was grown to confluence in 150-mm plnd incubated with at 37◦C as described above. The ceere scraped by a cell scraper and collected in a microhe cells were then washed rapidly three times with chhosphate-buffered saline and then lysed with sonicationw/v) lysozyme in PBS buffer). The lysate was cleared at◦Cy centrifugation at 13,000 rpm for 30 min. The supernatf the cell lysate were used in subsequent experiments

The bimolecular interactions were monitored using apectroscopy (MultiskopTM, Optrel GbR, Germany) aavelength of 632.8 nm and at incidence angles from 3◦ to

ollowed by the injection of 40�l of the lysate to captur7-binding protein from cell lysates. After lysate injecti

he sensor chip surface was washed with PBS buffer utable baseline was reached. All SPR experiments weried out at 25◦C with a constant flow rate of 8�l/min in PBSuffer.

.8. Statistical analysis

Statistical analysis was done using the paired Stud-test and ANOVA. Values between different groups wompared.P-values <0.05 were considered significant.

. Results

.1. Cell growth inhibition by AdCMVp53

When cervical cancer cells were infected with AMVp53 for increasing incubation times, complete inh

ion of cell growth was observed over time. The degreell growth inhibition was measured via MTT assay, andount. Inhibition of cell growth by MTT assay showed Ca81% suppression), SiHa (47%), HeLa (78.5%), HeL51%), C33A (94%), and HT3 (76%) on day 6, as shon Fig. 1. Cell counts assay showed CaSki (97.4%), S81.6%), HeLa (95%), HeLaS3 (85%), C33A (94%),T3 (74%) on day 6. Note that the same trend in sression levels was shown in CaSki, HeLa and in SeLaS3. In contrast, infection with AdCMVLacZ showed


Fig. 1. Growth-inhibitory effects of adenovirus p53 delivery in various cervical cancer cell lines. Cells (105 cells/well) were cultured in 12-well plates in triplicateovernight and infected with adenoviruses expressing p53 and LacZ at 50 MOI. After infection, cells were cultured for various periods and then trypsinized forcounting under a microscope. The mean cell counts in triplicate were plotted. (�) Mock-infected; (�) AdCMVLacZ-infected; (�) AdCMVp53-infected.

creased cell growth in a manner similar to the negative controlgroup.

To determine whether the AdCMV53 vector overex-presses p53 protein in each cell, Western blot analysis wasperformed after transfection with AdCMVp53 into each cellline as shown inFig. 2. The results showed that exogenousp53 protein showed a high level of expression in HPV posi-tive cells on day 2 and 4, but not on day 6. Of the HPV positivecells, the p53 was only detected in HeLaS3 on day 6. In con-trast, p53 expression was continually maintained in C33A andHT3 during the same periods. Note that the same trends in p53expression levels on day 4 were shown in CaSki, HeLa and inSiHa, HeLaS3, which were consistent with the trends in cellgrowth inhibition assay. There was no expression of p53 inmock-infected cells or cells infected with AdCMVLacZ. Theresults suggest the view that p53 expression is responsiblefor the differential cell growth inhibition in cervical cancercells.

3.2. Anti-tumor effects of AdCMVp53 in nude mice

We focused on CaSki and SiHa cells, as these two repre-sentative types of cells displayed a differential growth inhibi-tion pattern to AdCMVp53. To investigate whether overex-pression of p53 protein induces cell-dependent apoptosis invivo, we infected AdCMVp53 with cancer cells-xenograftednude mice and then measured each tumor at day 3, 5, 8, 11, 14,17, 20, and 23. As shown inFig. 3, in CaSki cell-xenograftedtumor tissue, the size was decreased to 23.2, 34.3, 33.9, 35.8,19.1, 20.8, 15.7, and 12.7%, as compared to AdCMVLacZand PBS injection, respectively. Also, in the case of SiHa cell-xenografted, the size reduction was 27.4, 26.6, 25.2, 28.0,28.8, 28.7, 28.9, and 27.1%, respectively, even though statis-tical significance in the early days was not meaningful. It is,however, notable that the levels of tumor growth inhibitionin the early stages (for about 11 days) are consistent with theresults of in vitro assay, in which the cell growth inhibition


Fig. 2. p53 Protein expression in cervical cancer cells after AdCMVp53infection. Each cell line infected with AdCMVp53 at 50 MOI was incubatedfor 6 days. Cells were harvested for specific days to obtain cell crude extracts.Ten microgram proteins were subjected to 10% SDS polyacrylamide gel, andanalyzed with ECL detection method. (A) CaSki; (B) SiHa; (C) HeLa; (D)HeLaS3; (E) C33A; (F) HT3.

Fig. 3. Tumor growth-inhibitory effects of adenovirus p53 in CaSki andSiHa cell-xenografted nude mice. Mice were injected subcutaneously with107 cells/mouse. Two weeks later, 100�l of AdCMVp53 (1010 pfu/ml), Ad-CMVLacZ (1010 pfu/ml) and PBS alone was injected into tumors of 7–8 mmin diameter. Seven mice were used for each treatment. Statistically signifi-cant atP< 0.05 using the paired Student’s test compared to PBS control wasindicated as * character. (*,P< 0.05).

Fig. 4. (A) SDS–PAGE of E7 proteins. The arrow points the 23-kDa proteindesignated E7 proteins. (B) Western blot analysis of E7 protein from differentcolonies (lanes 1 and 2). MW: molecular weight markers.

was remarkable at CaSki. No significant differences in anti-tumor effect of PBS and AdCMVLacZ were observed in thisassay.

3.3. Preparation of self-assembled E7 protein layer

The recombinant E7 protein encompassing 98 amino acidresidues of the full-lengthE7 gene was expressed and pu-rified in E. coli. The recombinant E7 protein migrated as a23 kDa protein in SDS–PAGE and was reactive to the HPV16 E7 monoclonal antibodies as shown inFig. 4. The changesof SPR curves by adsorbing 11-MUA, and by binding of E7protein in series on Au substrate are shown inFig. 5. As aresult, the SPR angle was shifted significantly from 43.3 to43.6 by the adsorption of 150 mM 11-MUA on Au surface.And, the SPR angle was shifted from 43.6 to 44.0 by chemicalbinding between E7 protein and the activated carboxyl groupof 11-MUA with EDAC. In principle, a surface plasmon res-

F mbled1 .

ig. 5. The changes of the SPR curve by E7 protein binding to self-asse1-MUA modified Au surface. (a) Bare Au; (b) 11-MUA; (c) E7 protein


Fig. 6. Sensorgrams for the capturing of E7-binding proteins from each celllysate by the immobilized E7 protein. (A) Injection of cell lysate from theSiHa cells; (B) injection of cell lysate from the CaSki cells; (C) injection ofcell lysate from the Hep-G2 cells over the sensor chip immobilized with E7protein.

onance is extremely sensitive to the interfacial architecture.An adsorption process leads to a shift in the plasmon reso-nance and allows monitoring the mass coverage at the surfacewith a high accuracy. Therefore, the shift in the SPR angleverified that thin layer of 11-MUA on Au surface was formedand E7 protein was well bound with 11-MUA adsorbed onAu substrate. From the above results, it can be concluded thatthe self-assembled E7 protein layer was fabricated on the Ausubstrate.

E7-binding protein was captured from the cell lysate di-rectly without the need for purifying the receptor protein.A typical sensorgram for the immobilization of the specificproteins is shown inFig. 6. The differential changes in sen-sor signal correspond to approximately 0.8 different bindingpatterns of E7 proteins-immobilized, based on the estima-tion that cell-specific pathway is existed. The captured E7-

binding proteins could not be washed away by PBS buffer butcould be completely washed away by the solution contain-ing 10 mM NaOH and 0.5% SDS. The regenerated antibodysurface could be reused for subsequent E7-binding proteinscapturing. The cell lysate containing E7-binding proteins wasextracted from the HPV-infected human cervical cancer cellssuch as CaSki, and SiHa overexpressing E6 and E7 oncopro-teins. In contrast, the lysate from the human hepatoma Hep-G2 cell line has significant E7-binding proteins. It was foundthat the amount of the captured proteins was completely de-pendent on the type of the cell line. Differential amounts ofE7-binding proteins were captured on the sensor chips usinghuman cervical cancer cells. The amount of E7-binding pro-teins captured from HPV 16-positive CaSki cells was greaterthan that from the SiHa cells. Hep-G2 cells infected HBVcontained more E7-binding molecules per cell than humancervical cancer cells, as expected.

4. Discussion

The cervical cancer cells that are HPV 16- or 18-positiveare reported to express only low concentration of p53 tran-script, due to E6-activated-ubiquitin dependent protease di-gestion. Our results showed that after infection with aden-ovirus p53, the expression of p53 protein was highly main-t ex-p 3. Ith in thep fur-t andp iptw thata cells,t oteinr ted-u otedt s re-s latedt sus-c (a PVp ies),a cellg s inS iods.T pat-t rente in-c comem me-t Thus,a g onc

co-p cer-

ained for several days and p53 transcripts were highlyressed for a long time with no cleaved fragments of p5as been shown that p53 can be cleaved to produce p35resence of double-stranded DNA, and that p35 could

her cleave p53 at the N- and C-terminals to produce p5040 (Okorokov et al., 1997). The expression of E6 transcras gradually decreased (data not shown), indicatinglthough exogenous p53 has sufficiently expressed in

he expression difference between p53 transcript and presults from the state of how much the level of E6-regulabiquitin-mediated protein degradation is. It has been n

hat cell-dependent significant growth suppression waulted from exogenous p53 overexpression. It was specuhat cells with a smaller HPV copy number might be moreeptible to growth inhibition by adeno-associated virusSund Wu, 1996), such as SiHa cells (one to two copies of Her cell) rather than CaSki (600 copies), HeLa (50 copnd HeLaS3 (10 copies). In CaSki cells, the inhibition ofrowth was significant in the early days in vitro, whereaiHa, the inhibition was not detected for the same perhe in vitro results was due to the differential expression

erns in p53 and E6 oncoprotein, leading to actually diffeffects on tumor growth inhibition in vivo. It is becomingreasingly clear that even though, gene transfer has beore efficient, another problem is that there could be so

hing specific to the disease treated or to the gene used.n improved strategy for cancer gene therapy dependinancer cell-dependent pathway should be studied.

Differential protein–protein interactions of several onroteins are functionally related to cell cycle regulation in


vical cancer, in which the pathways leading to cell growth andapoptosis play a key role in cervical cancer growth and reg-ulation. Although a number of cervical cancer-related genesand proteins have been studied, many of the molecular eventsinvolved in the cervical cancer pathway are still unclear be-cause a protein can be involved in multiple independentlyregulated cancer-specific pathways. For the purpose of iden-tifying protein interactions, the SPR analysis can overcomethe complexity of the biomolecule interactions in the HPV16-associated pathway, and identify several cancer-specificcellular functions as well as proteins of unknown function(Nelson et al., 1999). Also, it can become a major com-peting platform for the proteome-wide characterization ofcarcinogenesis. In this study, we used the SPR-based pro-tein chip for the detection of cell-specific E7 protein–proteininteractions, suggesting the differential binding characteris-tics of E7 with E7-binding proteins and which biologicalfunctions primarily affected by human cervical cancer type.The expression of HPV oncogenic proteins, E6 and E7, isrequired for carcinogenesis and maintenance of the tumorstate (Scheffner et al., 1991). Furthermore, E7-specific im-mune responses are detected in cervical cancer patients (deGruijl et al., 1996). In this regard, E7-specific interactionpatterns should be evaluated in real model systems. The SPRsensorgrams exhibit behavior characteristic of the specificbinding of each cell lysate to the E7-immobilized surface ofb r pro-t directa can-d turea avec mayi p53-m thee thati backo s fors elyp di-r tedb siblep iatedc

5

er iso s. Yeta enet o, iti rans-f therec o theg ime-c ade-

novirus expressing p53 on human cervical cancer cells andcell-specific E7 protein–protein interactions in cell lysatesusing SPR biosensor. For each cervical cancer cell, differen-tial cell growth inhibitions were shown via cell count assayand MTT assay. The same trend in suppression levels wasshown in CaSki, HeLa and in SiHa, HeLaS3, respectively.The levels of p53 protein were notably expressed in CaSkiand HeLa more than in SiHa and HeLaS3. After transfectionAdCMVp53 into CaSki- and SiHa-xenografted nude mice,the size of tumor was remarkably decreased in SiHa cellsas compared to AdCMVLacZ transfection. The SPR sensorsurface was successfully modified with the recombinant E7oncoprotein and showed cell-specific interactions betweenE7 and its target proteins from cell lysates. The present ex-perimental system is not already optimized, but it providesthe possibility of investigating SPR to understanding of cell-dependent protein–protein interaction in cervical cancer. Inthe future, a large subset of cellular proteomic changes viaSPR-based protein chip will be described. This could be ap-plied to construct the tumor classification and determinationof tumor progression with high efficiency.

Acknowledgement

andE En-v tuteo

R

C ofBio-

d em-per,viruselial

. 77,

G virus

H R.,r ofncer.

H oth,an-On-

H talkyn-

acti-

H s inncerancer

iosensor chip. The use of the E7-capturing approach foein interaction study has certain advantages over thentigen–antibody sensor chip. That is, specific bindingidates can be recruited for cell lysate with in vivo strucnd without the need for purified proteins. Our results honfirmed that the cell-specific E7 protein interactionsnduce the differential anti-cancer effect via exogenous

ediated apoptosis. Note that the in vivo results providexistence of cell-dependent growth-inhibitory pathway

s consistent with the SPR biosensor results. One drawf operating on lysates is that the sensitivity of SPR is lesmall molecules that for large molecules. This study likrovides an additional option for analyzing biomoleculesectly from complex mixtures and identifying E7-targeiomolecules from nontargeted biomolecules as a posroteome-wide diagnostic modality against HPV-assocervical cancer.

. Conclusions

The transfer of genes from one organism to anothne of the basic essential techniques of modern geneticfully satisfactory technique for efficient and reliable g

ransfer to a range of cell types is yet to be found. Alss becoming increasingly clear that even though, gene ter has become more efficient, another problem is thatould be something specific to the disease treated or tene used. The aim of this work is to investigate the tourse tumor growth suppression effects of recombinant

This work was supported in part by the Korea Sciencengineering Foundation (KOSEF) through the Advancedironment Monitoring Research Center at Kwangju Instif Science and Technology.

eferences

hoi, J.W., Nam, Y.S., Fujihira, M., 2004. Nanoscale fabricationbiomolecular layer and its application to biodevices. Biotech.process Eng. 9, 76–85.

e Gruijl, T.D., Bontkes, H.J., Stukart, M.J., Walboomers, J.M., Rmink, A.J., Verheijen, R.H., Helmerhorst, T.J., Meijer, C.J., ScheR.J., 1996. T cell proliferative responses against human papillomatype 16 E7 oncoprotein are most prominent in cervical intraepithneoplasia patients with a persistent viral infection. J. Gen. Virol2183–2191.

raham, F.L., Prevec, L., 1995. Methods for construction of adenovectors. Mol. Biotechnol. 3, 207–230.

amada, K., Alemany, R., Zhang, W.W., Hittelman, W.N., Lotan,Roth, J.A., Mitchell, M.F., 1996a. Adenovirus-mediated transfea wild-type p53 gene and induction of apoptosis in cervical caCancer Res. 56, 3047–3054.

amada, K., Sakaue, M., Alemany, R., Zhang, W.W., Horio, Y., RJ.A., Mitchell, M.F., 1996b. Growth inhibition of human cervical ccer cells with the recombinant adenovirus p53 in vitro. Gynecol.col. 60, 373–379.

u, Y., Gonzalez-Martinez, D., Kim, S.H., Bouloux, P., 2004. Crossof anosmin-1, the protein implicated in X-linked Kallmann’s sdrome, with heparan sulphate and urokinase-type plasminogenvator. Biochem. J. 23, 114–120.

uang, T.G., Ip, S.M., Yeung, W.S., Ngan, H.Y., 2000. Changep21WAF1, pRb, Mdm-2, Bax and Bcl-2 expression in cervical cacell lines transfected with a p53 expressing adenovirus. Eur. J. C36, 249–256.


Huber, W., Perspicace, S., Kohler, J., Muller, F., Schlatter, D., 2004. SPR-based interaction studies with small molecular weight ligands usinghAGT fusion proteins. Anal. Biochem. 333, 280–288.

Kaur, P., McDougall, J.K., Cone, R., 1989. Immortalization of primaryhuman epithelial cells by cloned cervical carcinoma DNA containinghuman papillomavirus type 16 E6/E7 open reading frames. J. Gen.Virol. 70, 1261–1266.

Lee, S.J., Namkoong, S.E., Lee, W.C., Sul, J.W., Jee, S.H., You, Y.K.,Lee, J.E., Park, J.S., 2002. Polymorphisms of p53, p21 and IRF-1 andcervical cancer susceptibility in Korean women. Cancer Res. Treat.34, 357–364.

Logsdon, N.J., Jones, B.C., Allman, J.C., Izotova, L., Schwartz, B.,Pestka, S., Walter, M.R., 2004. The IL-10R2 binding hot spot on IL-22 is located on the N-terminal helix and is dependent on N-linkedglycosylation. J. Mol. Biol. Sep. 342, 503–514.

McMurray, H.R., Nguyen, D., Westbrook, T.F., McCance, D.J., 2001.Biology of human papillomaviruses. Int. J. Exp. Pathol. 82, 15–33.

Nelson, R.W., Jarvik, J.W., Taillon, B.E., Tubbs, K.A., 1999. BIA/MSof epitope-tagged peptides directly fromE. coli lysate: multiplex de-tection and protein identification at low-femtomole to subfemtomolelevels. Anal. Chem. 15, 2858–2865.

Oh, B.K., Lee, W.C., Lee, W.H., Choi, J.W., 2003. Nano-scale probefabrication using self-assembly technique and application to detectionof Escherichia coliO157:H7. Biotech. Bioprocess Eng. 8, 227–232.

Oh, B.K., Kim, Y.K., Park, K.W., Lee, W.H., Choi, J.W., 2004. Surfaceplasmon resonance immunosensor for the detection ofSalmonella ty-phimurium. Biosens. Bioelectron. 19, 1497–1504.

Okorokov, A.L., Ponchel, F., Milner, J., 1997. Induced N- and C-terminalcleavage of p53: a core fragment of p53, generated by interaction withdamaged DNA, promotes cleavage of the N-terminus of full-lengthp53, whereas ssDNA induces C-terminal cleavage of p53. EMBO J.16, 6008–6017.

Scheffner, M., Munger, K., Byrne, J.C., Howley, P.M., 1991. The state ofthe p53 and retinoblastoma genes in human cervical carcinoma celllines. Proc. Natl. Acad. Sci. U.S.A. 88, 5523–5527.

Schlattner, U., Reinhart, C., Hornemann, T., Tokarska-Schlattner, M., Wal-limann, T., 2002. Isoenzyme-directed selection and characterization ofanti-creatine kinase single chain Fv antibodies from a human phagedisplay library. Biochim. Biophys. Acta 1579, 124–132.

Shillitoe, E.J., Kamath, P., Chen, Z., 1994. Papillomaviruses as targetsfor cancer gene therapy. Cancer Gene Ther. 1, 193–204.

Su, P.F., Wu, F.Y., 1996. Differential suppression of the tumorigenicityof HeLa and SiHa cells by adeno-associated virus. Br. J. Cancer 73,1533–1537.

Xu, H.J., Zhou, Y., Seigne, J., Perng, G.S., Mixon, M., Zhang, C., Li,J., Benedict, W.F., Hu, S.X., 1996. Enhanced tumor suppressor genetherapy via replication-deficient adenovirus vectors expressing an N-terminal truncated retinoblastoma protein. Cancer Res. 56, 2245–2249.

adenoviral p53 effects and cell-specific e7 protein–protein interactions of human cervical cancer...

Documents