Oncogene (2006) 25, 7740–7746 & 2006 Nature Publishing Group All rights reserved 0950-9232/06 $30.00 www.nature.com/onc ORIGINAL ARTICLE AHNP-streptavidin: a tetrameric bacterially produced surrogate fusion against p185her2/neu

K Masuda, M Richter, X Song, A Berezov, K Masuda, R Murali, MI Greene and H Zhang

Department of Pathology and Laboratory Medicine, University of Pennsylvania School of Medicine, Philadelphia, PA, USA

The anti-p185her2/neu peptidomimetic (AHNP) is a small of an antibody is sufficient to mimic its parental exo-cyclic peptide derived from the anti-p185her2/neu monoclonal antibody function (Murali and Greene, rhumAb 4D5 (h4D5). AHNP mimics many but not all 1998). The CDR peptide is much smaller than the of the antitumor characteristics exhibited by h4D5. antibody but retains binding affinity to the antigen as However, the pharmacokinetic profiles of AHNP are less well as the biological activity of the antibody. Our than optimal for therapeutic or diagnostic purposes. To studies with CDR mimetics demonstrated that the improve the bindingaffinity to p185 her2/neu and the constrained CDRs could mediate binding in a context- antitumor efficacy, we have engineered a fusion protein independent manner and formed the theoretical basis containingAHNP and a nonimmunoglobulin protein for the technique of CDR grafting (Bruck et al., 1986; scaffold, (SA). The recombinant protein, Williams et al., 1989; Park et al., 2000). AHNP-SA (ASA) bound to p185her2/neu with high affinity, p185her2/neu (Her2, Neu, c-ErbB2) is overexpressed in inhibited the proliferation of p185her2/neu-overexpressing many breast and ovarian cancers (Slamon et al., 1989), cells, and reduced tumor growth induced by p185her2/neu- lung tumors (Kern et al., 1990) as well as tumors of the transformed cells. These data suggest that the bacterially pancreas (Williams et al., 1991). Extensive studies have produced tetrameric ASA can be used as an antibody- shown that p185her2/neu plays a dominant role in mediat- surrogate molecule. This class of molecule will play a role ing the malignant phenotype (Kokai et al., 1988; Cohen in the diagnosis and treatment of p185her2/neu-related et al., 1989; Lodato et al., 1990). The evolution of tumors. Our studies establish a general principle by which targeted therapy began with the development of mono- a small biologically active synthetic exo-cyclic peptide can clonal to p185her2/neu (Drebin et al., 1984, 1985, be engineered to enhance functional aspects by structured 1986, 1988) that directly lead to the reversion of the oligomerization and can be produced recombinantly using malignant phenotype. The anti-p185her2/neu antibody bacterial expression. ‘trastuzumab’ (h4D5 or Herceptin, Genentech, San Oncogene (2006) 25, 7740–7746. doi:10.1038/sj.onc.1209745; Francisco, CA, USA), which represented a humanized published online 19 June 2006 antibody of similar specificity, has been approved by FDA to treat breast cancers (Sliwkowski et al., 1999). Keywords: AHNP; P185her2/neu; streptavidin; tetramer Earlier, we reported the design of a p185her2/neu binding peptidomimetic (BiP) anti-p185her2/neu peptidomimetic (AHNP) (FCDGFYACYMDV) (Park et al., 2000; Berezov et al., 2001). AHNP was rationally derived from the structure of the CDR-H3 loop of the anti- Introduction p185her2/neu antibody h4D5, and possessed the ability to disable p185her2/neu tyrosine kinase activity in vitro and Monoclonal antibodies represent mainstream agents for in vivo similar to, but with far less affinity, than the molecular diagnosis and have recently become de facto monoclonal antibody. Unlike the intact phenotype- cancer therapeutics. There is a great interest in devel- reverting monoclonals, AHNP was unable to down- oping smaller versions of therapeutic antibodies for regulate p185her2/neu from the cell surface, indicating that several reasons: namely to improve tumor penetration, the small molecule functioned to directly disable the eliminate the humanization process, and reduce the cost p185her2/neu complex. These characteristics suggested a of production (Maynard and Georgiou, 2000; Hudson potential for antibody-mimicking peptidomimetics to and Souriau, 2003). Previously, we established a general serve as reagents for tumor therapy. principle that a single CDR3 loop from the heavy chain Streptavidin (SA), a produced by Streptomyces avidinii that binds to with an À15 Correspondence: Dr MI Greene, Department of Pathology and extremely high affinity (KDB10 M), represents an Laboratory Medicine, University of Pennsylvania School of Medicine, ideal scaffold to enhance the functional potential of 252 John Morgan Building, 36th Hamilton Walk, Philadelphia, PA AHNP. With a high affinity for biotin, the fusion 19104-6082, USA. E-mail: [email protected] protein will be able to carry biotinylated therapeutics Received 14 March 2006; revised 26 April 2006; accepted 26 April 2006; (such as cytotoxins) or labeling reagents to tumor cells published online 19 June 2006 with p185her2/neu overexpression. This strategy has already Tetrameric anti-p185her2/neu protein ASA K Masuda et al 7741 been tested on antibody-SA conjugates but the results hydrochloride under reducing conditions, refolded and are varied due to the large size of the antibody purified by Co2 þ resin under oxidative conditions. (Kalofonos et al., 1990). In addition, the new class of On Sodium dodecyl sulfate–polyacrylamide gel elec- CDR-carrying tetramers can also be used to replace trophoresis (SDS–PAGE), both SA and ASA revealed a antibodies in assays utilizing the biotin-SA system. In single band of B16kDa, with ASA at a slightly higher fact, the ASA can be also used as a diagnostic tool to position, after the were heated at 1001C (lanes detect p185her2/neu in early breast cancer developments 1, 3; Figure 1b). The observed sizes correlate with the using FACTT, an ultra-sensitive detection technology expected molecular weight of monomers. Under milder (Zhang et al., 2006). Third, the in vivo pharmacokinetics conditions (heated at 551C, in the absence of 2-ME), SA of SA have been well studied (Rosebrough, 1993) and and ASA appeared as a single band of B64 kDa (lanes SA constructs have been used in patients (Kalofonos 2, 4; Figure 1b), indicating that these proteins associate et al., 1990). However, the therapeutic characteristics as tetramers. of functional peptidomimetics fused to nonimmuno- A disulfide bond was designed to form between the globulin protein scaffolds are not known. two cysteines in the AHNP sequence. In antibody- In the present study, we have recombinantly fused the derived peptides or proteins, disulfide bonds are AHNP peptide to SA. Here, we show that ASA essential to constrain the mimetic peptides (Park et al., possesses the abilities to bind the p185her2/neu ectodomain 2000) and also play a critical role in maintaining the and to suppress the proliferation of p185her2/neu-trans- structural framework required for binding to antigens formed cells in vitro as well as in vivo. These studies serve (Proba et al., 1997; Jurado et al., 2002). We chose the as a principle to create new forms of antibody-surrogate sulfhydryl-alkylating reagent 4-acetamide-40-maleimidyl- 0 molecules in a modular fashion. Biologically active atylbene-2,2 -disulfonic acid (AMS, Mr B500 Da) to peptides can be linked to a suitable scaffold upon which determine if the fusion proteins form the required multiple copies of relevant CDRs are disposed. We have disulfide bonds. Reduced proteins bind to AMS termed the new species, which contains a functional covalently with free thiol groups and migrated as CDR peptide loop on a nonimmunoglobulin scaffold, molecules of higher molecular weight in nonreducing a ‘Multimeric Binding peptide (mBip)’. SDS–PAGE. As shown in Figure 2, ASA expressed in bacteria was reduced and bound to AMS independent of dithiothreitol (DTT) treatment. This is due to the reducing environment in the E. coli cytoplasm. After Results purification, ASA was oxidized and only bound to AMS if pretreated with DTT. Since SA has no cysteines, AMS Expression and purification of the ASA in did not affect the mobility of SA on the SDS–PAGE (Figure 2c). Our data suggest that the purification In ASA, a flexible 11 amino-acid (a.a.) linker procedure facilitated proper disulfide bond formation (GGGGSRSNSSS) connects the AHNP unit to the in ASA. N-terminus of the SA core. ASA also contains a 6 Â His To verify whether ASA is refolded properly, tag for the purification purpose. While SA was easily we studied its binding to biotin and p185her2/neu by expressed as a soluble protein in Escherichia coli -linked immunosorbent assays (ELISA). Using (E. coli), ASA was insoluble and formed inclusion bodies biotin-coated microplates, we found that both SA and (Figure 1a). Insoluble ASA was dissolved in guanidine ASA bound to biotin in a dose-dependent manner

Figure 1 Expression and purification of the fusion protein ASA and SA. (a) Expression of ASA as insoluble proteins. Lane 1, total protein of uninduced cells (U); Lane 2, total protein of induced cells (I); lane 3, soluble fraction of induced cells (S); Lane 4, insoluble fraction of induced cells (IS). (b) Formation of tetramers observed in purified ASA and SA. Lanes 1 and 3, purified proteins were heated for 5 min in boiling water in the presence of SDS, DTT and 2-ME; Lanes 2 and 4, samples were heated for 5 min at 551C in the presence of SDS and DTT without 2-ME. Lanes 1 and 2, SA: Lanes 3 and 4, ASA.

Oncogene Tetrameric anti-p185her2/neu protein ASA K Masuda et al 7742

Figure 2 Disulfide bond formation in ASA and SA. AMS-alkylation was used to determine if disulfide bonds were formed in ASA produced in the cytoplasm of E. coli BL21(LysE) and after refolding. The alkylated and nonalkylated products, representing reduced and oxidized forms, respectively, were separated by their different mobility in nonreducing SDS–PAGE and visualized by Western blot probed with anti-SA antibody.

Figure 3 ELISA analysis of ASA for biotin and Her2-Fc binding ability. SA and ASA were 10-fold serially diluted from 1.1 mg/well as indicated. (a) SA or ASA was immobilized on a biotin-coated plate. Bound protein was detected by an anti-SA antibody-HRP. (b) After immobilization of SA or ASA on biotin-coated plate, Her2-human Fc fusion was added to each plate and bound fusion protein was detected using anti-Her2 antibody and anti-mouse IgG-HRP. Error bars represent the standard deviation from four replicates.

(Figure 3a). In a separate experiment, ASA was able to ASA (about 2.5 times) was required to achieve 50% cell bind to the Her2-Fc fusion in a dose-dependent manner, growth inhibition in vitro. Both ASA and AHNP were while SA demonstrated no binding even at the highest less potent than h4D5 in the inhibition of the anchorage- concentration (Figure 3b). These data indicate that the independent growth of T6–17 in the polyHEMA plate ASA retains a native structure, which permits an assay. integrated biophysical activity of AHNP and SA.

In vivo activity Kinetic analysis of ASA binding to HER2 We have established a mouse tumor model using the AHNP-SA (ASA) was also tested for binding to the transformed T6–17 cells. After T6–17 cells are inocu- HER2 receptor ectodomain by Biacore 3000 (Biacore lated into nude mice subcutaneously (s.c.), palpable International AB, Sweden). Kinetic constants were tumors appear after about 48–72 hours. To assess ASA estimated by global fitting analysis of the sensor gram as a protein therapeutic, we treated animals with either curves to the 1:1 Langmuirian interaction model, which 5 À1 À1 SA alone or ASA (3.75 mg/kg/dose, three times per gave a kon of 2.09 Â 10 M s , and a koff of À3 À1 week). Treatment started from day 0 and was provided 1.83 Â 10 s . The koff/kon ratio gave a value of to animals by three s.c. injections each week. Treatment 8.8 nM for the equilibrium dissociation constant (KD). with ASA prevented the appearance of palpable tumors The apparent affinity of ASA is about 40-fold higher until day 7 and reduced the tumor size by 62% at day 14 than AHNP (350 nM), suggesting that ASA might be (Figure 5a). useful for diagnostic and therapeutic purposes. We also studied in vivo ASA activity by intraperito- neal (i.v.) administration. Treatment, started day 3 after Biological activity of the ASA palpable tumors appeared (B2cm3) was also provided Next, we determined if ASA could inhibit proliferation to animals by three injections per week (Figure 5b–c). of p185her2/neu-transformed T6–17 cells (Di Fiore et al., Consistently, ASA demonstrated inhibition of tumor 1987). Treatment with ASA fusion proteins at two growth in a dose-dependent manner (Figure 5b). different concentrations resulted in 20–50% inhibition ASA (5 mg/kg, or 80 nmol/kg/dose) demonstrated 40 of cell proliferation in a dose-dependent manner times greater activity than the peptidomimetic AHNP (Figure 4a). Inhibitory effects were not observed in the (4 mg/kg, or 3 mmol/kg/dose) (Figure 5c), suggesting control NR6cell line (Figure 4b). However, compared that the pharmacokinetics of ASA is superior to AHNP with the AHNP peptide in molar concentration, more and is critical for in vivo activity. The result with ASA

Oncogene Tetrameric anti-p185her2/neu protein ASA K Masuda et al 7743

Figure 4 Inhibition of cell proliferation by ASA. Effects of ASA treatment, or treatments with SA and the peptide AHNP, on the proliferation of T6–17 (p185her2/neu transformed cells, C) or NR6 (control cells, D) were shown. Cell proliferation was measured by the standard MTT assay performed on poly-HEMA plates (for T6–17) or regular tissue culture plates (for NR6). Error bars represent the standard deviation from six replicates.

(5 mg/kg) was similar to that with h4D5 (0.5 mg/kg) at day 14. Figure 5 Inhibition of tumor growth by ASA. One million T6–17 cells were injected s.c. into nude mice to induce tumor growth. (a) s.c. treatment with ASA or SA (3.75 mg/kg/dose, three times per week) started from day 0. (b–c) i.v. treatment started on day 3, Discussion three times per week, after tumors were palpable. (b) ASA was administered at 50 or 5 mg/kg/dose. (c) ASA, AHNP and h4D5 After we rationally designed the CDR exocyclic peptide, were administered at 5 mg/kg/dose (80 nmol/kg/dose), 4 mg/kg/ dose (3 mmol/kg/dose), and 0.5 mg/kg/dose, respectively. Error bars which mimics antibody function (Park et al., 2000), we represent the standard deviation from 2 to 3 animals. explored ways to both enhance the binding profile and extend therapeutic potentials of the peptide AHNP. We have developed the notion that a multivalent CDR We anticipated that four copies of AHNP in ASA exocyclic peptide might enhance its biological potential. might result in a much slower off-rate than the To verify this idea, we have engineered a novel fusion monomeric peptide. However, the disassociation rate protein, ASA. As we expected, ASA forms a tetramer (1.83 Â 10À3 MÀ1 sÀ1) is actually slightly faster than after being recombinantly expressed and refolded. The the AHNP peptide (4.74 Â 10À4 MÀ1 sÀ1) and h4D5 high affinity of ASA for p185her2/neu and biotin suggests (1.23 Â 10À4 MÀ1 sÀ1) (Park et al., 2000). Kinetic binding that this fusion protein is properly folded to retain its studies suggest that AHNP units in ASA may not be biological functions. proximal to each other and once the binding AHNP unit ASA exhibits 40-fold higher affinity to p185her2/neu disassociates from the target there are no other free units (KD ¼ 8.8 nM) as compared to the isolated AHNP nearby to replace it. molecule (KD ¼ ASA 350 nM). A faster on-rate of To determine the disposition of AHNP in ASA, we ASA is accounted for by the avidity of the tetramer. have partially resolved the three-dimensional structure

Oncogene Tetrameric anti-p185her2/neu protein ASA K Masuda et al 7744 of ASA by X-ray crystallography at 2.8 A˚ . A detailed has better activity than ASA in vitro, the results report on the crystal structure of ASA will be published observed in vivo suggest that the reduced biological elsewhere. Briefly, AHNP and the 11 a.a. linker in ASA activity of AHNP is related to its pharmacokinetics and are structurally disordered and not visible in the electron not due its binding affinity. Although we have not density map. However, the approximate position of obtained data on the ASA pharmacokinetics yet, we AHNP can be located based on the crystal packing expect ASA to have a serum half-life close to that of SA restraints. Two SA monomers in the asymmetric unit (12 h) (Rosebrough, 1993) since they have similar core form a dimer (A–B) and are related by the noncrystallo- structure and molecular weight. In comparison, AHNP graphic twofold symmetry. This dimer then associates only has a serum half-life of about 2–3 h. with its symmetry (twofold axis)-related molecule, In separate studies we have found that several forming a tetramer (A-B-A0-B0) (Figure 6). The N- anti-p185her2/neu antibodies can promote tetrameric terminal AHNP units of the A–B dimer are directed to receptor complexes, which are less kinase active and two opposite directions. Hence, it is unlikely that the primed for the receptor downregulation (Furuuchi et al., two AHNP units in A–B dimer can interact with the manuscript submitted). Based on these observations, same p185her2/neu structure (either homo- or hetero-dimer) we hypothesize that ASA might promote signaling- simultaneously. However, the N-terminus of A and A0 defective receptor ensembles. These signaling-defective point in the same direction to position the two AHNP receptor ensembles (inactive tetrameric complexes) peptides adjacent to each other (35 A˚ ). The same is true might be functionally similar to defective dimers of the two AHNP units in B and B0 (25 A˚ ). Thus, it induced by AHNP. We can extend these scaffolds to appears that the tetrameric molecule contains two other polymerizing molecules. Thus, AHNP fusions separate p185her2/neu-binding surfaces, each consisting of with dimeric, trimeric or tetrameric scaffolds such as two copies of AHNP units. The tetramer really helical bundles may create species suitable for human functions as two dimers. therapeutic and diagnostic applications. ASA displays activity in vitro and in vivo. In summary, our study has demonstrated a Surprisingly, ASA demonstrated less antiproliferative rational protein engineering approach to produce in activity than the AHNP peptide in vitro in an bacteria therapeutically useful fusions carrying anchorage-independent assay using polyHEMA, but anti-p185her2/neu peptide, AHNP. The bacterial system is was about 40-fold more active than AHNP in reducing the most inexpensive expression system for recombinant tumor growth in vivo. These studies reaffirm the notion proteins. Bacterially expressed ASA has an economic that in vitro assays of malignant phenotype are never benefit compared to Herceptin. ASA also has the as revealing as in vivo tumor growth. Earlier while potential of replacing antibodies in the tumor-imaging developing the AHNP species, we found that the in vitro field. In the most successful two-step tumor imaging biological activity of peptidomimetic correlated with its procedure (Kalofonos et al., 1990), the antibody slow dissociation rate of binding (Berezov et al., 2001). has to be linked to SA first before it can be used in a It is possible that in the cell-based assay condition, ASA pretarget step. ASA, with full biotin-binding activity binding to the receptor is weak owing to the long flexible of SA, is completely compatible with such a two-step linkers between AHNP and SA. This is clearly reflected system. With engineered SA of lower immunogenicity in the SPR studies, in which the AHNP disassociation (Chinol et al., 1998; Meyer et al., 2001), ASA rate is comparable to antibody (in the order of 10À4), can be used to deliver biotinylated therapeutics, such while ASA disassociates is slightly faster than AHNP. as toxins or heavy metals, to enhance its in vivo activity. In vivo, ASA consistently exhibited potent antitumor Finally, the advantage of our approach is that the activity. Recombinantly produced ASA showed a construct is modular. Small peptide mimetics remain as comparable antitumor effect (50–60% inhibition) as functional units but linked to the scaffolds to render the antibody (h4D5) and had a much higher activity bioactive peptides into useful therapeutics that can be than AHNP. Considering the observation that AHNP expressed inexpensively in bacteria.

Figure 6 Modeling of ASA. The scaffold protein SA (rendered in ribbon model) formed a compact tetramer. The engineered AHNP loop (rendered in sphere model) extended far away (B25 A˚ ) from the molecular surface of the tetramer, through an elongated linker. The structure of AHNP loop was taken from the complex structure of HER2 and Herceptin Fab (PDB code: 1N8Z). The AHNP loop and the connecting linker were positioned according to the crystal packing restraints. The distances between the neighboring AHNP loops were different (B25 and 35 A˚ respectively). Image B is obtained via a 901 horizontal rotation of image A.

Oncogene Tetrameric anti-p185her2/neu protein ASA K Masuda et al 7745 Materials and methods Determination of disulfide bond formation by protein alkylation with AMS Cell lines Disulfide bond formation of ASA was determined by the AMS T6–17 (gift from Dr JH Pierce) was derived from NIH3T3 by method described previously (Invitrogen) (Kobayashi and Ito, overexpressing the p185her2/neu receptor (Di Fiore et al., 1987). 1999). Briefly, bacterial lysates (100 ml) or purified protein NR6is a mouse fibroblast cell line that lacks endogenous EGF samples (2 mg) were treated with DTT (100 mM) for 10 min at receptors. T6–17 and NR6 cells were grown in Dulbecco’s 1001C. Proteins were then precipitated with trichloroacetic modified Eagle medium supplemented with 10% heat inacti- acid (TCA, 10% W/V) for 10 min at 41C. After centrifugation vated fetal calf serum, L-glutamine (2 mM), penicillin (100 U/ml) at 10 000 g for 15 min, the pellet was washed once with 250 ml and streptomycin (100 mg/ml) at 371C in a humidified 5% CO2 of ice-cold acetone and air-dried after centrifugation again. atmosphere. Alkylation was performed in 20 ml of the buffer containing 150 mM Tris-Cl (pH 7.5), 2% SDS (W/V) and 15 mM AMS for 1 h at room temperature. 20 ml of SDS–PAGE sample buffer Construction of plasmids (2 Â ) lacking 2-mercaptoethanol was added to all samples, Two plasmids, pKMSH and pKMASH, were constructed to which were then analysed by SDS–PAGE and immunoblotting express SA and ASA fusion protein, respectively, under with horseradish peroxidase (HRP)-conjugated anti-SA anti- the control of T7 promoter and lacIq. The following primers body (Invitrogen, Carlsbad, CA, USA). were used to clone SA or the fusion protein: For SA, primer19 0 (5 -AAAAAACATATGGAAGCAGGTATCACCGGCACA MTT assay CCTGGTACAACCAGCTCGGCTCGACCTTCATCGTGA To measure cell proliferation, we used the modified anchorage- 0 0 CCGCGGC-3 ) and primer3–2dch (5 -AAGCTTTTATTA independent 3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyl tetrazo- 0 ATGATGATGGTG-ATGATGG-3 ); for AHNP fusion pro- lium bromide (MTT) assay as previously described (Hansen 0 tein ASA, primer1 (5 -AAAAAACATATGGTCGACTAT et al., 1989). T6–17 cells (800 cells/well, poly 2-hydroxyethyl TGCGATGGCTTTTATG CGTGCTATATGGATGTGGG methacrylate (poly-HEMA) coated) or NR6cells (1200 cells/ 0 0 TGGTGGTGGTAG-3 ), primer2 (5 -GGATGTGGGTGGT well) were seeded in 96-well plates in the presence of SA, ASA GGTGGTAGCAGATCTAACAGCAGCAGCGAAGCAG or the AHNP peptide and incubated at 371C in a humidified GTATCACCGGACAATGGTACAC-30) and primer3–2dch. 5% CO2 atmosphere for 48 h. A total of 25 ml of MTT solution The amplified DNAs were ligated to pCR2.1 vector using (5 mg/ml in PBS) was added to each well, and after 2 h of the TA Cloning Kit (Invitrogen, Carlsbad, CA, USA). The incubation at 371C, 100 ml of the extraction buffer (20% w/v plasmid pKMSH was constructed by inserting the core SDS, 50% N,N-dimethyl formamide, pH 4.7) was added. After sequence of SA into the NdeI and HindIII restriction sites of an overnight incubation at 371C, the optical density at 570 nm pET21 (a) þ . pKMSH was then used as a platform for the was measured using an ELISA reader. construction of the ASA fusion protein expression vector pKMASH. Two restriction enzyme sites (SalI and BglII) were engineered at the N-terminus of AHNP and inside the linker, Enzyme-linked immunosorbent assays her2/neu respectively, for the convenience of replacing the AHNP The binding of ASA to biotin and p185 was determined sequence in the future. by an ELISA assay modified from the previously described method (Farlow et al., 2002). Purified SA or ASA was immobilized on a biotin-coated ELISA plate (Pierce, Rock- 1 Fusion protein production, cyclization and purification ford, IL, USA) for 1 h at 4 C in PBS. Excess protein was E. coli BL21 Origami (LysS) (Novagen) cells harboring washed out and then 50 ng/well of purified Her2-human Fc either pKMSH or pKMASH were grown with 50 mg/ml fusion (Park et al., 2000) protein in PBS was added to each 1 ampicillin at 251C. When the OD at 600 nm reached 0.4, SA well. After a 3 h incubation at 4 C, plates were washed 6times M with PBS containing 0.1% Tween-20 (PBS-T; 150 ml/well) and and ASA were induced to express with 0.5 m isopropyl her2/neu b-D-thiogalactopyranoside (IPTG) and cultured at 371C for anti-p185 antibody was added in PBS (1:500 dilution; 2.5 h. The recombinant proteins were insoluble and existed in 50 ml/well). After a 1 h further incubation, plates were washed the inclusion body. After cells were sonicated, the pellet six times with PBS (200 ml/well) and HRP-conjugated anti- (inclusion body fraction) was obtained by centrifugation at mouse IgG antibody was added (1:5000 dilution; 50 ml/well). 10 000 g for 30 min and then solublized at 10 mg/ml with Finally, plates were washed six times with PBS (200 ml/well) M M and bound anti-mouse IgG antibody-HRP was developed 100 m Tris-Cl (pH 7.5), containing 6 guanidine-HCl and 0 0 1mM DTT. The supernatant was obtained by centrifugation at using a substrate solution (100 ml/well) containing 3,3 ,5,5 - 10 000 g for 30 min and then diluted against 20 times volume of tetramethylbenzene (1 tablet/10 ml solution; Sigma, St Louis, M 100 mM Tris-Cl (pH 7.5), containing 1 mM reduced form MO, USA) in phosphate-citrate buffer (103 m dibasic M glutathione and 0.1 mM oxidized form glutathione. After sodium phosphate, 24 m citric acid, pH 5.0; Sigma). The dilution, the supernatant was obtained by centrifugation at reaction was allowed to proceed in the dark for 2 min at RT, 10 000 g for 30 min and then incubated for 45 min at 41C with and stopped with 2 M H2SO4 (50 ml/well). A450 readings of the 1 ml of Talon resin (Clontech, Mountain View, CA, USA), samples were determined using an ELISA reader. Bound SA which had been equilibrated with buffer A, composed of was detected by the same method using an anti-SA-HRP in 100 mM Tris-Cl (pH 7.5) and 300 mM NaCl. The resin was PBS (1:500 dilution; 50 ml/well). packed into a column, and washed with 20 ml of buffer A and then 10 ml of buffer A supplemented with 30 mM imidazole. Study of ASA and p185her2/neu interaction using the Biacore SA or ASA was eluted with buffer A containing 250 mM Binding experiments were performed with the surface plasmon imidazole and dialysed against phosphate-buffered saline resonance-based biosensor instrument Biacore 3000 (Biacore (PBS). Alternatively, insoluble fusion proteins were also AB, Uppsala, Sweden), at 251C, as described previously solublized by urea (8 M), purified through a Ni-NTA agarose (Berezov et al., 2002). Immobilization of HER2 on the sensor column (Qiagen), and refolded by stepwise reduction of urea in surface was performed following the standard amine coupling the dialysis buffer. procedure according to the manufacturer’s instructions.

Oncogene Tetrameric anti-p185her2/neu protein ASA K Masuda et al 7746 Briefly, 35 ml of a solution containing 0.2 M N-ethyl-N0- Mice (dimethylaminopropyl) carbodiimide and 0.05 M N-hydroxy- NCR homozygous athymic (nude) mice (6to 8-week-old) were succinimide, were injected at a flow rate of 5 ml/min to activate purchased from the National Cancer Institute. To induce carboxyl groups on the sensor chip surface. HER2 (40 ng/ml in tumor formation, 106 transformed T6–17 cells were suspended 10 mM sodium acetate buffer, pH 5.0) was flowed over the chip in 100 ml of PBS and injected s.c. into the flank of each animal. surface at a flow rate of 20 ml/min until the desired level of Animals were maintained in accordance with guidelines of the bound protein (2000 RU) was reached. Unreacted protein was Institutional Animal Care and Use Committee (IACUC) of the washed out and activated groups that had not reacted were University of Pennsylvania. Tumor volume was calculated by blocked by the injection of 35 ml of 1 M ethanolamine at 5 ml/ the formula: p  length  width  Height/6. min. A reference surface was generated simultaneously under the same conditions but without receptor injection and used as Acknowledgements a blank to correct for instrument and buffer artefacts. ASA was injected at variable concentrations at a 20 ml/min flow rate This work was partially funded by grant from NCI (5P01 CA and binding to the HER2 receptor immobilized on the chip 89480). We thank Mayosha H Mendis for her technical was monitored in real time as a series of sensor grams. support in the purification of ASA.

References

Berezov A, Chen J, Liu Q, Zhang HT, Greene MI, Murali R. Kern JA, Schwartz DA, Nordberg JE, Weiner DB, Greene (2002). J Biol Chem 277l: 28330–28339. MI, Torney L et al. (1990). Cancer Res 50l: 5184–5187. Berezov A, Zhang HT, Greene MI, Murali R. (2001). J Med Kobayashi T, Ito K. (1999). EMBO J 18l: 1192–1198. Chem 44l: 2565–2574. Kokai Y, Wada T, Myers JN, Brown VI, Dobashi K, Cohen J Bruck C, Co MS, Slaoui M, Gaulton GN, Smith T, Fields BN et al. (1988). Princess Takamatsu Symp 19l: 45–57. et al. (1986). Proc Natl Acad Sci USA 83l: 6578–6582. Lodato RF, Maguire HC, Jr, Greene MI, Weiner DB, LiVolsi Chinol M, Casalini P, Maggiolo M, Canevari S, Omodeo ES, VA. (1990). Mod Pathol 3l: 449–454. Caliceti P et al. (1998). Br J Cancer 78l: 189–197. Maynard J, Georgiou G. (2000). Ann Rev Biomed Eng 2l: Cohen JA, Weiner DB, More KF, Kokai Y, Williams WV, 339–376. Maguire Jr HC et al. (1989). Oncogene 4l: 81–88. Meyer DL, Schultz J, Lin Y, Henry A, Sanderson J, Jackson Di Fiore PP, Pierce JH, Fleming TP, Hazan R, Ullrich A, King JM et al. (2001). Protein Sci 10l: 491–503. CR et al. (1987). Cell 51l: 1063–1070. Murali R, Greene MI. (1998). Immunol Res 17l: 163–169. Drebin JA, Link VC, Greene MI. (1988). Oncogene 2l: 387–394. Park BW, Zhang HT, Wu C, Berezov A, Zhang X, Dua R Drebin JA, Link VC, Stern DF, Weinberg RA, Greene MI. et al. (2000). Nat Biotechnol 18l: 194–198. (1985). Cell 41l: 697–706. Proba K, Honegger A, Pluckthun A. (1997). J Mol Biol 265l: Drebin JA, Link VC, Stern DF, Weinberg RA, Greene MI. 161–172. (1986). Sympn Fund Cancer Res 38l: 277–289. Rosebrough SF. (1993). Nucl Med Biol 20l: 663–668. Drebin JA, Stern DF, Link VC, Weinberg RA, Greene MI. Slamon DJ, Godolphin W, Jones LA, Holt JA, Wong SG, (1984). Nature 312l: 545–548. Keith DE et al. (1989). Science 244l: 707–712. Farlow SJ, Wang RJ, Pandori MW, Sano T. (2002) 516l: Sliwkowski MX, Lofgren JA, Lewis GD, Hotaling TE, Fendly 197–200. BM, Fox JA. (1999). Sem Oncol 26l: 60–70. Hansen MB, Nielsen SE, Berg K. (1989) 119l: 203–210. Williams TM, Weiner DB, Greene MI, Maguire Jr HC. (1991). Hudson PJ, Souriau C. (2003). Nat Med 9l: 129–134. Pathobiology 59l: 46–52. Jurado P, Ritz D, Beckwith J, de Lorenzo V, Fernandez LA. Williams WV, Moss DA, Kieber-Emmons T, Cohen JA, (2002). J Mol Biol 320l: 1–10. Myers JN, Weiner DB et al. (1989). Proc Natl Acad Sci USA Kalofonos HP, Rusckowski M, Siebecker DA, Sivolapenko 86l: 5537–5541. GB, Snook D, Lavender JP et al. (1990). J Nucl Med 31l: Zhang H, Cheng X, Richter M, Greene MI. (2006). Nat Med 1791–1796. 12: 473–477.

Oncogene