Inhibition of a protein-protein interaction ... - The FASEB Journal

The FASEB Journal • Research Communication

Inhibition of a protein-protein interaction between INI1 and c-Myc by small peptidomimetic molecules inspired by Helix-1 of c-Myc: identification of a new target of potential antineoplastic interest L. Bagnasco,*,†,‡,1 L. Tortolina,*,† B. Biasotti,*,†,‡ N. Castagnino,*,† R. Ponassi,*,† V. Tomati,* E. Nieddu,§ G. Stier,储 D. Malacarne,† and S. Parodi*,†,‡ *Department of Oncology, Biology and Genetics, University of Genoa, Genoa, Italy; †Laboratory of Experimental Oncology, National Cancer Institute (IST), Genoa, Italy; ‡I.S.O. Istituto Superiore di Oncologia, Genoa, Italy; §Department of Pharmaceutical Sciences, University of Genoa, Genoa, Italy; and 储Structural and Computational Biology Unit, EMBL Heidelberg, Heidelberg, Germany c-Myc is a transcription modulator protooncogene. When overexpressed, it becomes an important contributor to the multi-hit process of malignant transformation. In two earlier papers in this journal (see refs. 19, 20) we reported that retro-inverso peptidomimetic molecules inspired by the Helix-1 of c-Myc motif could be sequence-specific antiproliferative agents active in the low micromolar range. We also found that our peptides were not opening the fouralpha-helix Myc:Max bundle. Their antiproliferative activity in cancer cell lines needs the presence of side chains projecting outside of the bundle in the corresponding native H1 motif. This observation suggested interference with an external partner. In this study we investigated the INI1:Myc interaction. INI1 is a subunit of the SWI/SNF complex (component of the enhanceosome surrounding Myc:Max heterodimer). The INI1: Myc interaction was confirmed via pull down, ELISA, and fluorescence anisotropy assays. According to the length of INI1 fragments used, we calculated Kds ranging between 1.3ⴛ10ⴚ6 and 4.8ⴛ10ⴚ7 M. The three different techniques applied showed that the INI1:Myc interaction was also the target of our retro-inverso peptidomimetic molecules, which seem to bind specifically at INI1. A Myc binding, 21aa INI1 fragment (minimum interacting sequence), could inspire the synthesis of a new class of more selective c-Myc inhibitors.—Bagnasco, L., Tortolina, L., Biasotti, B., Castagnino, N., Ponassi, R., Tomati, V., Nieddu, E., Stier, G., Malacarne, D., and Parodi, S. Inhibition of a protein-protein interaction between INI1 and c-Myc by small peptidomimetic molecules inspired by Helix-1 of c-Myc: identification of a new target of potential antineoplastic interest. FASEB J. 21, 1256 –1263 (2007) ABSTRACT

blasts after receiving fresh serum—respond with an early wave of neosynthesis of c-Myc mRNA and c-Myc protein. For many years, c-Myc has been thought to act as an early response transcription factor dimerizing with its obligatory partner Max and binding DNA as a hetero-dimer at a specific consensus sequence (the E-box [CACGTG]) (1, 2). The c-Myc N-terminal domain, called trans-activation domain (TAD), would bind bridge proteins with one of the three RNA polymerases, thereby playing a key role in regulating ribosome biogenesis and cell growth. Several genes downstream of c-Myc would be transcribed and finally translated into new proteins (3–5). This action could be compared with other similar proteins (Mad family proteins) belonging, as Myc itself does, to the helix-loop-helix superfamily but missing the TAD that could interact and dimerize with Max, inhibiting Myc effects of cell growth and proliferation. Papers describing genes downstream of c-Myc have appeared (6) and continue to appear at high frequency in the literature (see on-line database on web site: http://www.myc-cancer-gene.org/site/mycTargetDB. asp). A reasonable possibility is that c-Myc is necessary but not sufficient for specific gene transcription. If other transcription factors concur to transcription specificity, then the thousands of genes potentially downstream of c-Myc could undergo a more finely regulated expression. Concerning the initial processes of gene transcription, a complex cell machinery, called enhanceosome, has been described by some investigators for different transcription factors; c-Myc could take part in similar complexes (7). c-Myc is one of the most studied oncogenes. Its relevance in malignant transformation comes from the fact that its homeostasis has been directly or indirectly

Key Words: inhibitors 1

Many cells, after receiving a growth stimulation signal—for instance, cultured nontransformed fibro1256

Correspondence: Department of Oncology, Biology and Genetics, University of Genoa, L. go R. Benzi 10, Genoa 16132, Italy. E-mail: [email protected] doi: 10.1096/fj.06-7082com 0892-6638/07/0021-1256 © FASEB

altered in the majority of all solid and hematopoietic tumors (8). During the last decade, many attempts using small drug-like molecules to inhibit c-Myc have been reported, and papers concerning the development of c-Myc inhibitors have been published (9 –14). A great number of other proteins besides Max have been found to interact directly with c-Myc, modulating its action both positively and negatively (15, 16). For several years, our work has focused around c-Myc and the search of peptidomimetic c-Myc inhibitors (17, 18). In our 2005 paper, we demonstrated sequence-specific activity (complete growth inhibition at 10 ␮M) for some peptidomimetic molecules, made of D-amino acids and inspired from the first helix of c-Myc, in c-Myc overexpressing sensitive cell lines (18). These RI-peptides (19) were fused to a retro-inverso internalization sequence discovered by our group, and could easily and efficiently cross through cell membranes and reach the nucleus. Moreover, being constituted by D-amino acids, they were insensitive to the activity of cell proteases, with long half-lives both in vitro and in vivo (18, 20). Based on circular dichroism (CD) and antiproliferative activity data (18), we suggested that our retroinverso peptides could act externally of the “four alpha helix bundle” formed by the C-terminal domains of c-Myc and Max (21), acting as a wedge between helix 1 of c-Myc and an external protein belonging to the enhanceosome (or higher order structure surrounding c-Myc:Max). Based on literature data and considerations of molecular modeling, we proposed the protein INI1 as a possible candidate (22). In this work we demonstrate that our hypotheses were correct, with two independent molecular/biochemical methods showing that our retro-inverso peptides are specific in inhibiting the interaction between c-Myc and INI1. We had previously shown that some side chains of our retro-inverso peptidomimetic molecules were required for a partial interference at the level of the c-Myc:Max heterodimer, whereas different side chains were essential for a growth inhibitory activity. These last molecules are also capable of interfer-

ence at the level of the c-Myc:INI1 protein-protein interaction. By means of a third biophysical method based on fluorescence polarization, we have defined a minimum interacting sequence of INI1 with c-Myc; this sequence could be a starting point at the level of molecular modeling for future designs of c-Myc inhibitor peptides inspired by a INI1 motif and looking toward c-Myc.

MATERIALS AND METHODS Protein production and purification Recombinant 6-His-tagged proteins were produced in transformed Escherichia coli as described previously (18). Briefly, cDNA sequences encoding either the whole or part of the proteins were subcloned in pET-modified vectors and transformed in BL21 strain of E. coli. Bacterial pellets were harvested and lysed, and the proteins were purified on a Ni-NTA-column (Qiagen, Valencia, CA, USA). The C-terminal domain of c-Myc (b-HLH-zip), the RPT1–2 (aa 179 –326) region of INI1, and the entire Max protein were produced in pETM11 vector. This vector contains a TEV (Tobacco Etch Virus) protease recognition site between the His tag and the recombinant protein start codon. After purification, the tag could be cut out using a recombinant His-tagged TEV protease; when needed, a second Ni-NTA column was used to further purify the protein from both cleaved His tag and the protease itself. The whole INI1 protein was also produced as a nontagged protein, both in pETM13 and in pETM11. Peptide synthesis Peptides indicated in Table 1 were synthesized either with an Applied Biosystem 433A Peptide Synthesizer using a solidphase technique or manually by using the standard method of solid-phase peptide synthesis, which follows the 9-fluorenylmethoxycarbonyl (Fmoc) strategy with minor modifications. Additional details are given in ref. 18. The HLH-zip portion of c-Myc (QRR-H1-loop-H2-zip, aa 365– 439) was obtained by chemical solid-phase synthesis. This L-peptide was fluoresceinated at the amino-terminal for fluorescence anisotropy studies (Myc-F). The sequence of Myc-F is: Fl•QRRNELKRSFFALRDQIPELENNEKAPKVVILKKATAYILSVQAEEQKLISEEDLLRKRREQLKHKLEQLRNSCA.

TABLE 1. List of peptides studied Peptide

Sequence

L-Int-H1-S6A,F8A RI-Int-VV-H1-S6A,F8A RI-Int-VV-H1-S6A,F8A,Q13A

RQIKIWFQNRRMKWKKNELKRAFAALRDQI IQDRLAAFARKLENKKWKMRRNQFWVKVQR IADRLAAFARKLENKKWKMRRNQFWVKVQR

Fluoresceinated peptides INI_193-213 INI_203-223 INI_213-233 INI_223-243 INI_233-253 INI_243-263 INI_253-273

MEIDGQKLRDAFTWNMNEKLM AFTWNMNEKLMTPEMFSEILC MTPEMFSEILCDDLDLNPLTF CDDLDLNPLTFVPAIASAIRQ FVPAIASAIRQQIESYPTDSI QQIESYPTDSILEDQSDQRVI ILEDQSDQRVIIKLNIHVGNI

INHIBITION OF INTERACTION BETWEEN INI1 AND C-MYC

1257

The peptide was then purified and characterized as described previously (18). Pull down assay The histidine-tagged b-HLH-zip domain of c-Myc was produced as described above. To proceed for the pull down assay, this target was loaded onto a Ni-NTA-agarose column and the resin was washed extensively with urea 8 M at a gradient of pH ranging between 8 and 5.9, and to renaturate the recombinant protein, with 50 mM Tris/HCl pH 7. A small amount of resin was then used to assay the purity (by SDS-PAGE) and the concentration (by Bradford assay; Bio-Rad, Hercules, CA, USA) of the attached protein. Resin-linked His-tagged b-HLH-zip domain of c-Myc at a final concentration of 1 ␮M was incubated with a bacterial lysate containing the whole untagged INI1 in blocking buffer (200 mM KCl, 20 mM HEPES pH 7, 5 mM DTT, 20 mM imidazole, 5 mg/ml BSA, 0.1% IGEPAL, Pefablock) for 60 min at 4°C with gentle rocking. The resin was washed several times with Wash buffer (20 mM HEPES pH 7, 0.1 mM EDTA, 0.5 mM DTT, 4 mM MgCl2, 300 mM KCl, 20 mM imidazole, 0.05% IGEPAL, 10% glycerol), boiled for 5 min, run on a SDS-PAGE, transferred on a nitrocellulose membrane, and blotted with an anti-INI1 antibody (sc-13055 Santa Cruz; Santa Cruz Biotechnology Inc., Santa Cruz, CA, USA). A secondary HRP anti-rabbit antibody was used for detection. ELISA Two different enzyme-linked immunosorbent assays were performed to study binding between INI1 and c-Myc. The first assay involves interaction of INI1 with immobilized c-Myc; the second involves interaction of c-Myc with immobilized INI1. In the first case, 96-microtiter plate wells were coated with purified b-HLH-zip domain of c-Myc in 50 ␮l of PBS (NaCl 137 mM, KCl 2.7 mM, Na2HPO4 4.3 mM, KH2PO4 1.4 mM) overnight at 4°C. Wells were then rinsed with wash buffer (PBS, 0.05% Tween) and blocked with 200 ␮l of blocking buffer (PBS, 1% BSA) for 3 h at 37°C. After rinsing three times with wash buffer, 100 ␮l of binding buffer (PBS, 0.05% Tween, 1% BSA) containing 2 ␮M of purified INI1 whole protein was added to the wells and let to react for 1 h at 37°C. Wells were rinsed again three times and 100 ␮l of binding buffer containing 200 ng/ml of a primary antibody anti-INI1 (sc-13055 Santa Cruz) was added and incubated for 1 h at 37°C. After three additional rinsings, 100 ␮l of the same buffer containing a secondary HRP anti-rabbit antibody was added and incubated for another hour at 37°C. Finally, 100 ␮l of ABTS substrate (2,2⬘-azinobis [3-ethylbenzothiazoline-6-sulfonic acid]-diammonium salt) was applied and the plate was kept in the dark until the color intensity of positive controls was maximum and the negative controls did not develop nonspecific reactions (⬃6 –10 min). The ELISA plate was scanned in a Biotech ELISA reader at 450 nm. For the second assay, the same method was applied, but INI1 was immobilized in the wells and incubated with c-Myc. Primary antibody anti-c-Myc (sc-40 Santa Cruz) was used to reveal the binding reaction. Fluorescence anisotropy The formation of the hetero-dimers was measured in terms of increased anisotropy (Polarion; Tecan Trading AG, Switzer1258

Vol. 21

April 2007

land). All experiments were performed in polar buffer (20 mM Tris-HCl, 150 mM NaCl, 1 mM EDTA); the experimental volume was always 200 ␮l per well, and the plates used were 96-well black nonbinding plates (Corning Glassworks, Corning, NY, USA). Proteins were mixed at room temperature and maintained at target temperature (30°C) in the dark for the whole duration of the experiments. Anisotropy measurements were performed at different incubation times to assess the stability of the reactions. The reported results refer to measurements performed after 60 min of incubation.

RESULTS Pull down assays were always carried out in the presence of at least two negative controls: we used clean untreated resin saturated with blocking buffer to assess a nonspecific binding of untagged INI1 directly to the agarose (Fig. 1, lane 5). To exclude a nonspecific antibody reaction, we applied to the resin bound to the tagged “bait” protein (b-HLH-zip of c-Myc), a binding buffer solution not containing the “prey” protein (INI1) (Fig. 1, lane 4). As shown in Fig. 1, only in lane 2, corresponding to the pull down of the prey protein INI1, it was possible to detect a signal similar in intensity to those of lanes 7 and 8 (positive controls). Thus, we can confirm that there is a direct, specific interaction between c-Myc and INI1. In our previous work (18) we speculated that our peptides, active in inhibiting c-Myc at a cellular level, could interact externally to the c-Myc:Max four alpha helix bundle, probably acting between the helix 1 of c-Myc and an external protein such as INI1. Biological activity of our retro-inverso peptidomimetic molecules, in fact, required Helix-1 amino acid side chains looking outside of the four alpha helix bundle in the corresponding H1 motif of c-Myc. To confirm this hypothesis, we used some of our active peptides (see Table 1 in Materials and Methods) to dismantle the pull down reaction at concentrations

Figure 1. Pull down assay. Immunoblot of proteins bound to His-b-HLH-zip of c-Myc (1.3 ␮M) in a pull down assay, using anti-INI1 antibody. Lanes 1, 3, 6: molecular weight markers. Lane 2: INI1 (1.0 ␮M). Lane 4: blank. Lane 5: Ni-NTA resin incubated directly with INI1 alone in binding buffer. Lane 7: input (bacterial lysate directly loaded onto the gel at a similar protein concentration (⬃1.0 ␮M) as used for binding. Lane 8: purified His-INI1 (higher concentration: 2.5 ␮M). Typical results of experiments performed in triplicate.

The FASEB Journal

BAGNASCO ET AL.

of 1 ␮M, 10 ␮M, and 100 ␮M. An unrelated peptide (BH3 of Bak) was used as a negative control. In Fig. 2 we show that only our active retro-inverso peptide RI-Int-H1-S6A,F8A was able to inhibit the binding between c-Myc and INI1 in a pull down assay, whereas L-Int-H1-S6A,F8A seemed inactive. This last outcome did not surprise us since it could be explained considering that the experiment was run in the presence of a raw bacterial lysate, probably rich in peptidases. The results of the subsequent ELISA assay were also compatible with an intrinsic weaker activity of the L-peptide with respect to the corresponding retro-inverso peptide. To assess the activity of our L- and RI-peptides with a different technique and in a “peptidase-free” environment, we implemented an ELISA assay. As for the pull down, we used similar negative

Figure 2. Pull down assay. Immunoblot of proteins bound to His-b-HLH-zip of c-Myc (1.3 ␮M) in a pull down assay, using anti-INI1 antibody. Upper panel: lane 1: INI1 bound to His-b-HLH-zip of c-Myc Ni-NTA resin complex. Lanes 2, 7: molecular weight markers. Lanes 3–5: as lane 1, but complex incubated with L-Int-H1-S6A,F8A (100 ␮M, 10 ␮M, and 1 ␮M, respectively). Lane 6: as in lane 1, but complex incubated with unrelated L-peptide BH3 of Bak (100 ␮M). Lane 8: blank. Lane 9: Ni-NTA resin incubated directly with INI1 alone in binding buffer. Lane 10: input (bacterial lysate directly loaded onto gel, containing a similar INI1 amount as used for binding). Lower panel: lane 1: INI1 bound to His-b-HLH-zip of c-Myc Ni-NTA resin complex. Lanes 2, 7: molecular weight markers. Lanes 3–5: as in lane 1, but complex incubated with RI-Int-H1-S6A,F8A (100 ␮M, 10 ␮M, and 1 ␮M, respectively). Lane 6: as in lane 1, but complex incubated with unrelated D-peptide RI-BH3 of Bak (100 ␮M). Lane 8: blank. Lane 9: Ni-NTA resin incubated directly with INI1 alone in binding buffer. Lane 10: input (bacterial lysate directly loaded onto gel containing an amount of INI1 similar to that used for binding). Typical results of experiments performed in triplicate.


Figure 3. ELISA: binding of INI1 with b-HLH-zip domain of c-Myc. Plates were coated with INI1 (1 ␮M) and incubated with increasing concentrations of c-Myc. Lane A: blank (uncoated wells). Lane B: c-Myc (1.4 ␮M). Lane C: c-Myc (3.5 ␮M). Lane D: c-Myc (7.1 ␮M). Average of triplicate experiments (vertical bars: se).

controls to avoid unspecific reactions. Also in this case, results fitted with our hypothesis. Figure 3 shows a binding interaction between b-HLH-zip of c-Myc and INI1. Addition of our active L-peptide to the reaction buffer (Fig. 4, upper panel) at concentrations ranging between 1 and 100 ␮M leads to a clear inhibition of signal intensity in a dose-dependent manner, indicating a binding inhibition between our two protein targets. Moreover, in this case we tested the sequence specificity of our peptide: in the lower panel of Fig. 4 the effects (at high concentration) of our RI active peptide and an RI unrelated peptide are shown, whereas 100 ␮M RI-Int-H1-S6A-F8A induces a close to 90% binding inhibition and 100 ␮M RI-Bak (BH3 of Bak) is completely inactive. In agreement with the pull down assay, our RI-peptide is much more potent than its corresponding L-form. This is also true at the cellular level (20). Finally, we run another set of ELISA assays in which other RI-peptides, found to be active at the cellular level (18), were tested. Figure 5A shows a representative ELISA assay where the RI-Int-H1-S6A,F8A,Q13A peptide was used; results observed at the cellular level were confirmed. Another set of experiments aimed at finding the minimum interacting sequence of INI1 with c-Myc was performed. The final goal of these studies is to conceive and synthesize new c-Myc-inhibiting molecules inspired by a specific INI1 sequence. Data in the literature describe the fragment of INI1 corresponding to the region around the RPT1 (imperfect repeat motif conserved region) aa 183–243 as interacting with c-Myc (22). To confirm these data, a fluorescence polarization test of interaction between Myc-F (see Materials and Methods) and a recombinant fragment of INI1, referred as RPT1–2 (aa 179 –326) (22), was performed; Fig. 6 shows a clear interaction between these two protein fragments. A Kd was calculated using the equation proposed by Lundblad for simple binding interaction assays (23). 1259

Figure 5. ELISA: binding inhibition of b-HLH-zip domain of c-Myc with INI1. Plates were coated with b-HLH-zip (1 ␮M) and incubated with purified INI1 at the fixed concentration of 2.1 ␮M alone (lane A) or in the presence of RI-Int-H1S6A,F8A,Q13A 1 ␮M (lane B), 10 ␮M (lane C), 100 ␮M (lane D). Values are blanked and expressed as percentage of noninhibited binding (column A: 100%). Blanks are on the order of 12–15% of full binding. Results are the average of triplicate experiments (vertical bars: se).

Figure 4. ELISA: binding inhibition of b-HLH-zip domain of c-Myc with INI1. Plates were coated with b-HLH-zip of c-Myc (1 ␮M), then INI1 was added to the wells and incubated for 1 h at 37°C. Values are blanked and expressed as percentage of noninhibited binding (columns A: 100%). Blanks are on the order of 12–15% of full binding. Results are the average of triplicate experiments (vertical bars: se). Upper panel: column A: INI1 full binding to c-Myc, no inhibition. Columns B–D: inhibitor L-peptide Int-H1-S6A,F8A was added at growing concentrations (1, 10, and 100 ␮M, respectively) along with INI1. Lower panel, column A: INI1 alone, no inhibition. Column B: inhibitor RI-peptide RI-Int-S6A,F8A was added along with INI1 at a concentration of 100 ␮M. Column C: inhibition with unrelated peptide RI-BH3 of Bak (100 ␮M).

For our graph, the best-fitting Kd was 1.3⫻10⫺6 M (se⫽1.7⫻10⫺7 M). All the best-fitting computations and statistical analysis were performed using the SAS/ STAT software, Version 8, of the SAS System for [Unix], copyright, SAS Institute Inc. (Cary, NC, USA). Encouraged by these results, a set of seven consecutive fluoresceinated peptides, each 21 aa long and overlapping each other by 10 amino acids, encompassing the putative interacting fragment of INI1 (Table 2), were synthesized as described in Materials and Methods. Each peptide was tested for interaction with c-Myc b-HLH-zip domain using the fluorescence anisotropy technique; only the fluoresceinated peptide INI_213– 233 demonstrates a binding activity with b-HLH-zip, as shown in Fig. 7. We calculated the Kd as described 1260

Vol. 21

April 2007

above and computed a Kd of 4.8⫻10⫺7 M (se⫽9.8⫻10⫺8 M). We also demonstrated that this interaction is reversible by inhibiting the binding with the same nonfluoresceinated peptide. Figure 8 shows an inhibition curve obtained incubating a fluoresceinated INI1_213–233 and b-HLH-zip of c-Myc at the fixed concentrations of 7.5⫻10⫺7 M and 5.0⫻10⫺7 M, respectively, in the presence of an increasing concentration of “cold” INI1_213–233. Finally we demonstrated that our peptidomimetic molecules RI-Int-H1-S6A,F8A; RI-Int-H1-S6A,F8A,Q13A

Figure 6. Fluorescence anisotropy. Anisotropy values of binding interaction at a fixed concentration of fluoresceinated c-Myc HLH-zip (Myc-F) (5⫻10⫺9 M) and increasing concentrations of RPT1–2. These results are referred to 60 min of incubation and represent the mean of three different experiments, each performed in triplicate. Where not indicated, ses are not larger than the black squares. Kd ⫽ 1.3⫻10⫺6 M (se⫽1.7⫻10⫺7 M).

The FASEB Journal

BAGNASCO ET AL.

TABLE 2. Schematic representation of a set of seven consecutive fluoresceinated peptides, each 21aa long, and overlapping each other by 10 amino acids, encompassing the putative interacting fragment of INI1(RPT1-RPT2): Only the peptide INI_213–233 (in boldface) demonstrates a binding activity with b-HLH-zip. 189 RPT1 245 259 RPT2 319

193 MEIDGQKLRDAFTWNMNEKLM 213 203 AFTWNMNEKLMTPEMFSEILC 223 213 MTPEMFSEILCDDLDLNPLTF 233 223 CDDLDLNPLTFVPAIASAIRQ 243 233 FVPAIASAIRQQIESYPTDSI 253 243 QQIESYPTDSILEDQSDQRVI 263 253 ILEDQSDQRVIIKLNIHVGNI 273

and L-Int-H1-S6A,F8A, inspired by H1 of c-Myc, are able to interact with the INI1_213–233 fragment (Fig. 9). At high concentrations, L-Int-H1-S6A,F8A (black squares) reaches a higher plateau level of anisotropy with respect to both RI-Int-H1-S6A,F8A (black triangles) and RI-Int-H1-S6A,F8A,Q13A (black circles). However, the inflection point for the three molecules binding with fluoresceinated INI1_213–233 is always ⬃3⫻10⫺7 M (3.4⫻10⫺7 M for L-Int-H1-S6A,F8A, 2.9⫻10⫺7 M for RI-Int-H1-S6A,F8A, and 3.2⫻10⫺7 M for RI-Int-H1-S6A,F8A,Q13A, respectively). The difference among the three values is not statistically significant. An intrinsic higher fluorescence anisotropy of the completely bound complex [fluoresceinated INI1_213– 233]:[L-Int-H1-S6A,F8A] suggests a stiffer hetero-dimer structure further away from a spherical/ellipsoidal conformation with respect to the other two hetero-dimers, a complex endowed with a slower tumbling time (23). When using the ELISA technique (Figs. 4 and 5) with

Figure 7. Fluorescence anisotropy. Anisotropy values of binding interaction at a fixed concentration of fluoresceinated INI1_213–233 (7.5⫻10⫺8 M) and increasing molar concentrations of b-HLH-zip of c-Myc. These results are referred to 60 min of incubation and represent the mean of three different experiments, each performed in triplicate. Where not indicated, the ses are not larger than the black squares. Kd ⫽ 4.8⫻10⫺7 M (se⫽9.8⫻10⫺8 M). INHIBITION OF INTERACTION BETWEEN INI1 AND C-MYC

our three peptidomimetic molecules mentioned above, we were inhibiting the formation of a relatively large hetero-dimer (probably endowed with a significant degree of quaternary structure) [b-HLH-zip (88 aa)]: [INI1 (385 aa)]. This situation is quite different from that shown in Fig. 9, where a direct binding study between short 21- and 30-aminoacid-long peptides is reported. Perhaps a slightly different ranking of the potency of our three molecules with the two methods is justified by the decidedly different experimental conditions.

DISCUSSION The results reported in our previous paper (18) suggested that our retro-inverso peptides, inspired by Helix-1 (H1) of c-Myc, in order to display biological antiproliferative activity required the presence of

Figure 8. Fluorescence anisotropy. Anisotropy values of an inhibition assay (black squares) of binding interaction between fluoresceinated INI1_213–233 and b-HLH-zip of c-Myc at the fixed concentrations of 7.5⫻10⫺8 M and 5.0⫻10⫺7 M, respectively, in the presence of an increasing molar concentration of “cold” INI1_213–233. The black triangle indicates the anisotropy value of fluoresceinated INI1_213–233 alone. These results are referred to 60 min of incubation and represent the mean of three different experiments, each performed in triplicate. Where not indicated, the ses are not larger than the black symbols. 1261

Figure 9. Fluorescence anisotropy. Anisotropy values of binding interaction at a fixed concentration of fluoresceinated INI1_213–233 (7.5⫻10⫺8 M) and increasing concentrations of L-Int-H1-S6A,F8A (black squares), RI-Int-H1-S6A,F8A (black triangles), and RI-Int-H1-S6A,F8A,Q13A (black circles). Best-fitting sigmoid curves are also indicated as follows: L-Int-H1-S6A,F8A (dashed curve), RI-Int-H1-S6A,F8A (dashed/ dotted curve), and RI-Int-H1-S6A,F8A,Q13A (dotted curve). Estimated Kd values are 3.4⫻10⫺7 M for L-Int-H1-S6A,F8A, 2.9⫻10⫺7 M for RI-Int-H1-S6A,F8A, and 3.2⫻10⫺7 M for RI-Int-H1-S6A,F8A,Q13A, respectively. Differences among the three Kds are not statistically significant. These results are referred to 60 min of incubation and represent the mean of three different experiments, each performed in triplicate. Where not indicated, the ses are not larger than the black symbols.

amino acid side chains projecting outward from the c-Myc:Max four alpha-helix bundle (making reference to the corresponding H1 of c-Myc motif). In addition, CD data had clearly shown that the ability of some of our molecules to interfere with the c-Myc:Max four alpha-helix bundle never induced a complete separation of the two monomers and was not a requirement for biological activity. Literature reports (22) referred to INI1, a component of the DNA remodeling complex SWI/SNF, as a protein binding to c-Myc perhaps at the basic helix-H1 region, external to the c-Myc:Max four alpha helix bundle. We therefore explored whether our peptidomimetic molecules could be active at the level of this protein-protein interaction. We first wanted to confirm whether recombinant c-Myc and INI1 (whole proteins or fragments) were able to establish a protein-protein interaction. As shown in Figs. 1 and 3, we have confirmed such interaction. At this point we wanted to explore whether our peptidomimetic molecules would interfere with this c-Myc:INI1 interaction. Since we could inhibit the binding between b-HLH-zip of c-Myc and the entire INI1 protein by means of our lead retro-inverso molecule RI-Int-F6A,S8A both in a pull down assay (Fig. 2, lower panel) and by ELISA (Fig. 4, lower panel), we could say that our hypothesis was correct. L-Int-F6A,S8A was apparently unable to inhibit that interaction at the level of a pull down assay (Fig. 2, upper panel). This 1262

Vol. 21

April 2007

result could be explained at least in part by considering that a raw bacterial lysate enriched in recombinant INI1 was run on the column containing a histidinetagged b-HLH-zip of c-Myc linked to the resin. This microenvironment could contain bacterial peptidases that might have partially but sufficiently digested the 30 amino acid-long L-Int-F6A,S8A. In fact, transferring this test to the “cleaner” ELISA environment showed a partial inhibition of the c-Myc:INI1 interaction by the L-peptide, but less efficient than the inhibitory activity of the corresponding RI-peptide (Fig. 4, upper panel and lower panel, respectively). Other RI-peptides found to be active at the cellular level in our earlier work (18) have been ELISA tested and found to agree reasonably with previously observed biological results. For instance, at the biochemical level of the ELISA, RI-Int-H1-S6A,F8A,Q13A (Fig. 5) is slightly less potent than RI-Int-H1-S6A,F8A (Fig. 4, lower panel) but is slightly more active at the cellular level (18). Obviously the intracellular multiprotein complexity can finely modulate the biological effect beyond the simplified situation of a biochemical test. During the last decade, much effort has been exerted in an attempt to find c-Myc inhibitors. These investigators (9 –14) tried to find small drug-like molecules capable of interfering at the level of the c-Myc:Max interaction. In this paper we reported inhibitors of a c-Myc interaction with a different partner. It could be important to define a new protein:protein interaction target useful for testing different families of potential c-Myc inhibitors. For this reason we planned a series of experiments aimed at finding a minimum interacting sequence of INI1 with the carboxyl-terminal region of c-Myc. The assessment via fluorescence anisotropy of the binding between the RPT1–2 motif of INI1 and a fluoresceinated HLH-zip fragment of c-Myc (Fig. 6) suggested that a possible smaller fragment of INI1, still binding with c-Myc, could be searched within the RPT1–2 INI1 region. To pursue this hypothesis, an array of seven fluoresceinated peptides, 21 amino acid long and overlapping by 10 aa, encompassing the major interacting region of INI1 with c-Myc, were synthesized. By means of the fluorescence anisotropy technique, they were tested to assess their binding with b-HLH-zip of c-Myc. Only the fluoresceinated peptide INI1_213–233 was able to interact (Fig. 7) reversibly (Fig. 8) with the above c-Myc domain, with a calculated Kd of 4.8 ⫻ 10⫺7 M. The calculated Kd (1.3⫻10⫺6 M) for the entire RPT1–2 region was slightly higher (Fig. 6). The c-Myc fragments used to test RPT1–2 and INI1_213–233 fragments were not identical, however. Only the c-Myc fragment binding the fluoresceinated INI1_213–233 fragment possessed a complete basic helix region. This could be a possible explanation for the better affinity observed for INI1_213–233 with respect to the larger INI1 fragment tested with a fluoresceinated QRR-helix-loop-helix-zip of Myc. The results reported in Fig. 9 suggest that our

The FASEB Journal

BAGNASCO ET AL.

peptides (both L- and RI-) interact directly with the INI1_213–233 fragment. The results reported in Fig. 9 have prompted us to think that the RI-peptidomimetic molecules reported earlier (18) may not have been optimally conceived considering they look at a modular protein, INI1, used by the cell for many different chromatin remodeling and accessibility functions rather than directly at a region of the c-Myc onco-protein. At the same time, the 21 aa INI_213–233 fragment could offer an inspiration at the level of modeling for conceiving and synthesizing molecules looking toward c-Myc, rather than toward INI1. This INI1:c-Myc interaction region is rich in acid aa on the INI1 side, basic aa on the c-Myc side, and is relatively poor in hydrophobic aa. It could turn out to be a favorable hot spot for the search of drug-like or more extended in space peptidomimetic inhibitors. We set up a new well-defined and apparently robust biochemical target for screening (potentially including HTS of combinatorial libraries); this new target could be the starting point for searching a new class of c-Myc inhibitors: inhibitors of the INI1:c-Myc protein-protein interaction. Looking at the problem of finding c-Myc inhibitors from a more general point of view, several proteins interact, for instance, with the important TAD domain of c-Myc. We think that the number of c-Myc proteinprotein interactions of potential interest as targets for inhibitors is much larger than the Myc:Max heterodimerization region (probably the most evolutionary conserved region). In this work we suggest a new possibility and invite a look at a higher order structure surrounding c-Myc. This work was supported by Ministero della Salute Finalizzato 2003 conv. No. 137, MIUR PRIN 2003 prot. 2003061530, MIUR PRIN 2005-prot. 2005061408_005.

6. 7. 8. 9.

10. 11. 12.

13. 14. 15. 16.

17.

18.

19.

20.

REFERENCES 1. 2. 3. 4.

5.

Blackwell, T. K., Kretzner, L., Blackwood, E. M., Eisenman, R. N., and Weintraub, H. (1990) Sequence-specific DNA binding by the c-Myc protein. Science 250, 1149 –1151 Blackwood, E. M., Lucher, B., and Eisenman, R. N. (1992) Myc and Max associate in vivo. Genes Dev. 6, 71– 80 Oster, S. K., Ho, C. S., Soucie, E. L., and Penn, L. Z. (2002) The myc oncogene: marvelously complex. Adv. Cancer Res. 84, 81– 154 Arabi, A., Wu, S., Ridderstrale, K., Bierhoff, H., Shiue, C., Fatyol, K., Fahlen, S., Hydbring, P., Soderberg, O., Grummt, I., et al. (2005) c-Myc associates with ribosomal DNA and activates RNA polymerase I transcription. Nat. Cell Biol. 7, 303–310 Oskarsson, T., and Trumpp, A. (2005) The Myc trilogy: lord of RNA polymerases. Nat. Cell Biol. 7, 215–217


21. 22.

23.

Dang, C. V. (1999) Minireview: c-Myc target genes involved in cell growth, apoptosis, and metabolism. Mol. Cell. Biol. 19, 1–11 Merika, M., and Thanos, D. (2001) Enhanceosomes. Curr. Opin. Genet. Dev. 11, 205–208 Henriksson, M., and Lucher, B. (1996) Proteins of the Myc network: essential regulators of cell growth and differentiation. Cancer Res. 68, 109 –182 Berg, T., Cohen, S. B., Desharnais, J., Sonderegger, C., Maslyar, D. J., Goldberg, J., Boger, D. L., and Vogt, P. K. (2002) Small-molecule antagonists of Myc/Max ibroblasts. Proc. Natl. Acad. Sci. U. S. A. 99, 3830 –3835 Xu, Y., Shi, J., Yamamoto, N., Moss, J. A., Vogt, P. K., and Janda, K. D. (2006) A credit-card library approach for disrupting protein-protein interactions. Bioorg. Med. Chem. 14, 2660 –2673 Yin, X., Giap, C., Lazo, J. S., and Prochownik, E. V. (2003) Low molecular weight inhibitors of Myc-Max interaction and function. Oncogene 22, 6151– 6159 Mo, H., and Henriksson, M. (2006) Identification of small molecules that induce apoptosis in a Myc-dependent manner and inhibit Myc-driven transformation. Proc. Natl. Acad. Sci. U. S. A. 103, 6344 – 6349 Kiessling, A., Sperl, B., Hollis, A., Eick, D., and Berg, T. (2006) Selective inhibition of c-Myc/Max dimerization and DNA binding by small molecules. Chem. Biol. 13, 745–751 Ponzielli, R., Katz, S., Barsyte-Lovejoy, D., and Penn, L. Z. (2005) Cancer therapeutics: targeting the dark side of Myc. Eur. J. Cancer 41, 2485–2501 Sakamuro, D., and Prendergast, G. C. (1999) New Myc-interacting proteins: a second Myc network emerges. Oncogene 18, 2942–2954 Kohn, K. W., Aladjem, M. I., Pasa, S., Parodi, S., and Pommier, Y. (2003) Molecular interaction map of mammalian cell cycle control. In Encyclopedia of the Human Genome, pp. 457– 474, Nature Publishing Group, London Giorello, L., Clerico, L., Pescarolo, M. P., Vikhanskaya, F., Salmona, M., Colella, G., Bruno, S., Mancuso, T., Bagnasco, L., Russo, P, and Parodi, S. (1998) Inhibition of cancer cell growth and c-Myc transcriptional activity by a c-Myc helix 1-type peptide fused to an internalization sequence. Cancer Res. 58, 3654 –3659 Nieddu, E., Melchiori, A., Pescarolo, M. P., Bagnasco, L., Biasotti, B., Licheri, B., Malacarne, D., Tortolina, L., Castagnino, N., Pasa, S., et al. (2005) Sequence specific peptidomimetic molecules inhibitors of a protein-protein interaction at the helix 1 level of c-Myc. FASEB J. 19, 632– 634 Guichard, G., Benkirane, N., Zeder-Lutz, G., van Regenmortel, M. H., Briand, J. P., and Muller, S. (1994) Antigenic mimicry of natural L-peptides with retro-inverso-peptidomimetics. Proc. Natl. Acad. Sci. U. S. A. 91, 9765–9769 Pescarolo, M. P., Bagnasco, L., Malacarne, D., Melchiori, A., Valente, P., Millo, E., Bruno, S., Basso, S., and Parodi, S. (2001) A retro-inverso peptide homologous to Helix 1 of c-Myc is a potent and specific inhibitor of proliferation in different cellular system. FASEB J. 15, 31–33 Nair, S. K., and Burley, S. K. (2003) X-ray structure of Myc-Max and Mad-Max recognizing DNA. Molecular bases of regulation by proto-oncogenic transcription factors. Cell 112, 193–205 Cheng, S. W., Davies, K. P., Yung, E., Beltran, R. J., Yu, J., and Kalpana, G. V. (1999) c-MYC interacts with INI1/hSNF5 and requires the SWI/SNF complex for transactivation function. Nat. Genet. 22, 102–105 Lundblad, J. R., Laurance, M., and Goodman, R. H. (1996) Fluorescence polarization analysis of protein-DNA and proteinprotein interactions. Mol. Endocrinol. 10, 607– 612 Received for publication August 18, 2006. Accepted for publication November 9, 2006.

1263