Sequence specific peptidomimetic molecules inhibitors of a protein-protein interaction at the helix...

©2005 FASEB

The FASEB Journal express article 10.1096/fj.04-2369fje. Published online January 25, 2005.

Sequence specific peptidomimetic molecules inhibitors of a protein–protein interaction at the helix 1 level of c-Myc E. Nieddu,* A. Melchiori,† M. P. Pescarolo,† L. Bagnasco,† B. Biasotti,‡ B. Licheri,‡ D. Malacarne,† L. Tortolina,‡ N. Castagnino,‡ S. Pasa,†,‡ G. Cimoli,‡ C. Avignolo,‡ R. Ponassi,‡ C. Balbi,† E. Patrone,§ C. D’Arrigo,§ P. Barboro,† F. Vasile,║ P. Orecchia,¶,# B. Carnemolla,¶ G. Damonte,** E. Millo,# D. Palomba,†† G. Fassina,†† M. Mazzei,* and S. Parodi†,‡

*Department of Pharmaceutical Sciences, University of Genoa, Genoa, Italy; †Laboratory of Experimental Oncology, National Cancer Institute, Genoa, Italy; ‡Department of Oncology, Biology and Genetics, University of Genoa, Genoa, Italy; §Institute for the Study of Macromolecules, Genoa section, Genoa, Italy; ║Dept. of Biophysical Sciences and Technology, University of Genoa, Genoa, Italy; ¶Laboratory of Cell Biology, National Cancer Institute, Genoa, Italy; #Institute Giannina Gaslini, Genova, Italy; **Dept. of Experimental Medicine, Biochemistry section c/o Center of Excellence for Biomedical Research, Genoa, Italy; ††Xeptagen S.p.A., Pozzuoli, Napoli, Italy

Corresponding author: E. Nieddu, Dept. of Pharmaceutical Sciences, University of Genoa, 16132 Genoa, Italy. E-mail: [email protected]

ABSTRACT

Our work is focused in the broad area of strategies and efforts to inhibit protein–protein interactions. The possible strategies in this field are definitely much more varied than in the case of ATP-pocket inhibitors. In our previous work (10), we reported that a retro-inverso (RI) form of Helix1 (H1) of c-Myc, linked to an RI-internalization sequence arising from the third α-helix of Antennapedia (Int) was endowed with an antiproliferative and proapoptotic activity toward the cancer cell lines MCF-7 and HCT-116. The activity apparently was dependent upon the presence of the Myc motif. In this work, by ala-scan mapping of the H1 portion of our molecules with D-aa, we found two amino acids necessary for antiproliferative activity: D-Lys in 4 and D-Arg in 5 (numbers refer to L-forms). In the natural hetero-dimer, these two side chains project to the outside of the four α-helix bundle. Moreover, we were able to obtain three peptides more active than the original lead. They strongly reduced cell proliferation and survival (RI-Int-VV-H1-E2A,S6A,F8A; RI-Int-VV-H1-S6A,F8A,R11A; RI-Int-VV-H1-S6A,F8A,Q13A): after 8 days at 10 µM total cell number was ~1% of the number of cells initially seeded. In these more potent molecules, the ablated side chains project to the inside in the corresponding natural four α-helix bundle. In the present work, we also investigated the behavior of our molecules at the biochemical level. Using both a circular dichroism (CD) and a fluorescence anisotropy approach, we noted that side chains projecting at the interior of the four α-helix bundle are needed for inducing the partial unfolding of Myc-H2, without an opening of the leucine zipper. Side chains projecting at the outside are not required for this biochemical effect. However, antiproliferative activity had the opposite requirements: side chains projecting at the outside of the bundle were

of 26(page number not for citation purposes)

essential, and, on the contrary, ablation of one side chain at a time projecting at the inside increased rather than decreased biological activity. We conclude that our active molecules probably interfere at the level of a protein–protein interaction between Myc-Max and a third protein of the transcription complex. Finally, CD and nuclear magnetic resonance (NMR) data, plus dynamic simulations, suggest a prevalent random coil conformation of the H1 portion of our molecules, at least in diluted solutions. The introduction of a kink (substitution with proline in positions 5 or 7) led to an important reduction of biological activity. We have also synthesized a longer peptido-mimetic molecule (RI-Int-H1-S6A,F8A-loop-H2) with the intent of obtaining a wider zone of interaction and a stronger interference at the level of the higher-order structure (enhanceosome). RI-Int-H1-S6A,F8A-loop-H2 was less active rather than more active in respect to RI-Int-VV-H1-S6A,F8A, apparently because it has a clear bent to form a β-sheet (CD and NMR data).

Key words: D-peptides • protein–protein contacts • growth inhibition • structural studies

n our search of lead molecules of potential antineoplastic interest, we have chosen c-Myc protein as a target. Its proto-oncogene can be converted into an oncogenic form through proviral insertion, chromosomal translocation, gene amplification, or mutations that increase

the rate of transcription and elongation or alter mRNA half-life (1). The gene is often amplified in nonsmall cell lung, breast, colon, prostate, and ovarian carcinoma, and elevated expression is found in colon and breast carcinomas (2). In colon cancer, its expression is increased and deregulated after APC inactivation (a “hard” DNA alteration present in more than 70% of colon cancers) (3). In conclusion, c-Myc deregulation has emerged as a central oncogenic switch in many human cancers (4). Which high incidence cancers will be especially “addicted”/hypersensitive to c-Myc inhibitors (perhaps in association with inhibitors of other signaling-proteins) is still unknown.

c-Myc interacts in vitro and in vivo with another cellular BR-HLH-Zip (BasicRegion-Helix1-Loop-Helix2-LeucineZipper) protein, Max, forming hetero-dimers which recognize and bind to a six-base palindromic DNA sequence known as E-box. Other direct or indirect DNA bindings are probably required to confer sufficient binding selectivity (5). The hetero-dimer interacts with other proteins within a higher-order structure, which has also been named enhanceosome (6).

We tried to interfere with c-Myc activity at this level through the synthesis of peptido-mimetic molecules that could work as a wedge between the Myc/Max hetero-dimer and a cross-talking protein of the enhanceosome (7). We took inspiration from Helix 1 (H1) of Myc, and we synthesized several peptido-mimetic analogs of this motif. We made them capable of internalization through a linkage to artificial retro-inverso (RI) basic sequences arising from the third α-helix of Antennapedia (Int) (8). These RI forms translocate very well across biological membranes both at 4°C and 37°C (9). To drastically reduce intracellular and extracellular peptidase activity we used D-aa, and we synthesized several RI-peptides [RI-Int, RI-Int-VV, RI-Int-VV-H1, and many other RI-Int-VV-H1 derivatives] (10).

We tested our RI-peptides on an HCT-116 cell line from a human colon carcinoma and on a MCF-7 cell line from a human breast cancer, both overexpressing c-Myc.

I


In this work, we confined our interest in H1 analogs, RI-peptides made of D-amino acids (D-aa), studying the effects of systematic mutations (one side-chain change at a time) in our amino acids. To understand which side chains are favorable or unfavorable to biological activity, a complete Ala-scan mapping of the RI H1 motif was carried out. We also investigated the effects of the substitution of two key residues (Arg5 or Phe7) with a proline on folding and the biological activity. In addition, we have also synthesized and studied a longer peptido-mimetic molecule similar in sequence to the Helix1-loop-Helix2 of Myc [RI-Int-H1-S6A,F8A-loop-H2].

MATERIALS AND METHODS

Peptides synthesis, purification, and characterization

A list of studied RI peptido-mimetic molecules can be found in Table 1. References to data in tables are indicated by “short-cut abbreviations” contained in Table 1.

Chemistry

Most of peptides were synthesized with an Applied Biosystems (Foster City, CA) 433A peptide synthesizer using a solid-phase technique. Rink amide-AM resin (Advanced Biotech) (substitution 0.6 mequiv/g) was used to obtain an amide at the C-terminal. Swelling was induced by washing the resin several times with N-methylpyrrolidone (NMP) and dichloromethane (DCM) and leaving it overnight in NMP. 10 equivalents of amino acid per 1 equivalent of resin were used. N-Fmoc-aa (Advanced Biotech) protected with Boc (Trp, Lys), tBu (Ser, Thr, Tyr, Asp, Glu), Trt (Cys, Asn, Gln, His), and Pbf (Arg) were used. The removal of FMOC was accomplished using piperidine (20% in NMP) for 25’. The conductivity of the deprotection solution (because of the conductive carbamate salt generated) was monitored. If necessary, a further deprotection of 15’ was made. The carboxyl group of the next amino acid (10eq) was activated by 9.5eq of HBTU with HOBt in DIEA/DMF solution. The coupling was 2 h long for the first amino acid and 1 h and 30 min for the others. If an additional deprotection occurred, an additional coupling (30 min) was made. The capping was made by 0.5 M acetic anhydride, 0.125 M DIEA and 0.015 M HOBt in NMP. The free NH2 of the last synthesized amino acid (usually I) was acetylated. The cleavage was made using 94.5 TFA, 2.5 Water, 2.5 EDT, and 1 TIS for 2-3 h at room temperature. Then, precipitation in ether occurred.

The analysis and purification were made by reversed-phase high-performance liquid chromatography (RPHPLC) (BioCAD 700E Perfusion Chromatography Workstation).

The analytical column was a RP-18 (5 µm) LiChrospher 100. The preparative column was a µBondapak C18, 125Å, 10 µm, 19 × 300 mm. Solvents are Water (0.1% TFA) and acetonitrile (0.1 %TFA).

Some of the peptides were analyzed for molecular mass by mass spectrometry using a single quadrupole HP Engine 5989-A equipped with an electrospray ion source (ES-MS) and set in the positive-ion mode and confirmed with a Kompact MALDI 4 TOF spectrometer (Kratos Analytical, Manchester, UK). Other peptides were analyzed for molecular mass by electrospray ionization (ESI)-quadrupole ion-trap mass spectrometry on a LCQDECA instrument (ThermoFinnigan, San Jose, CA).


HLH and peptides modified with proline were manually synthesized by using the standard method of solid-phase peptide synthesis, which follows the 9-fluorenylmethoxycarbonyl (Fmoc) strategy with minor modifications. Briefly, 65 µmol of deprotected Rink AM resin (Novabiochem, Lucerne, Switzerland) were treated for 40 min at 40°C with a coupling reaction mixture containing five equivalents (equiv) of the appropriate Fmoc-aa (Novabiochem), 4.5 equiv of o-(7-azabenzotriazol-1-yl)-1,1,3,3-tetramethyluroniumhexafluorophosphate (PRIMM), 5 equiv of N,N-diisopropilethylamine (Fluka Chemie AG, Buchs, Switzerland), 7.5 equiv of sim-collidina (Fluka) at 0.2 M amino acid final concentration in anhydrous N-methylpyrrolidone (Fluka).

The peptides were then purified and characterized as described previously.

The peptides Int, H1, Int-H1, HLH, R5A, Q13A were rhodaminated (conjugation of a rhodamine-B group at the NH2-terminal region), purified, and characterized as described previously.

Preparation of the Myc/Max biochemical target for studying the interference of our peptido-mimetic molecules

A QRR-H1-loop-H2-zip portion of c-Myc, 75 aa long, was obtained by chemical solid-phase synthesis. This L-peptide was prepared both cold (for CD studies) and fluoresceinated at the amino-terminal (for fluorescence anisotropy studies). A complete Max protein was instead obtained via molecular biology procedures. Both preparation methodologies, related to each of the two monomers, are described below.

Myc peptide

Sequence of fluoresceinated (or nonfluoresceinated) QRR-H1-loop-H2-zip of c-Myc (L-peptide used for dimerization with Max) is

Fl·QRRNELKRSFFALRDQIPELENNEKAPKVVILKKATAYILSVQAEEQKLISEEDLLRKRREQLKHKLEQLRNSCA

Chemistry

The peptide was synthesized with an Applied Biosystems 431A peptide synthesizer using a solid-phase technique. 9-fluorenylmethoxycarbonyl (Fmoc) aa and HMP resin (substitution 0.99 mmol/g) were manufactured by Novabiochem. The swelling was made washing the resin (160 mg) several times with N-methylpyrrolidone (NMP, Merck) and dichloromethane (DCM, Merck). One mmole of Fmoc-Ala-OH reacted with 0.5 mmoles of DCC (Applied Biosystems), giving 0.5 mmoles of the corresponding symmetric anhydride. The coupling with the resin was 34’ long, then the resin was washed several times with MeOH (Merck), DCM, and NMP. Then, the synthesis proceeded as described previously for peptide chemistry. A cutoff at aa 20, 40, and 60 occurred to test the synthesis progression.

For fluorescence polarization experiments a fluorescent QRR-H1-loop-H2-zip of c-Myc was made (Myc-F): the N-terminal amine of the protected peptide resin was treated with 1.5 equiv of


5,6carboxy fluorescein N-hydroxysuccinimide ester (Fluka) and 3eq of DIEA in NMP for 16 h at 25°C.

Cleavage, purification, and characterization were performed as described previously for RI peptido-mimetic molecules chemistry.

Max protein

The sequence encoding the whole human Max coding region was generated by RT-PCR. Total RNA was extracted from PBL. Peripheral blood lymphocytes were isolated from heparinized blood of healthy volunteers by Ficoll-Hypaque density gradient centrifugation, and mRNA was reverted using oligo-dT as primer. cDNA encoding for the full length of human Max was amplified using specific primers containing suitable restriction enzyme sites for further in-frame cloning (Forward: TAT CCA TGG GCG ATA ACG ATG ACA TCG A and Reverse: ATA GGT ACC TTA GCT GGC CTC CAT CCG G).

The PCR product was cleaved with NcoI and Acc65I restriction enzymes and cloned into different vectors (kindly donated by Gunter Stier, EMBL Heidelberg, Germany). These vectors are based on pET series backbone (Novagen), and expression is determined by isopropyl 1-thio-β-D-galactopyranoside (IPTG) induction, whereas purification is based on the Histidine Tag method.

Max expression vectors were transformed in Escherichia coli strain BL21 [DE3], and an expression run was performed to choose the best expressing construct.

Briefly, bacterial cultures were grown to an OD600 of 0.8–1.0, and protein expression was induced by adding IPTG to a final concentration of 0.5 mM. Cultures were incubated for overnight at 25°C before cell harvesting.

Cells were resuspended in buffer A (200 mM Tris pH 8, 10 mM imidazole, 150 mM NaCl, 2 mM β-mercaptoethanol) containing 0.2% NP-40, shortly sonicated and centrifuged at 10,000 g for 15 min: the cleared supernatant was directly loaded onto a Ni2+-NTA-Agarose column (Quiagen) previously equilibrate with buffer A. After successive washes in buffer A, containing 1 M NaCl and 30 mM imidazole, bound proteins were eluted with buffer A containing 300 mM imidazole.

Our best result was given by a plasmid carrying a TEV (Tobacco Etch Virus protease) cleavage site between the histidine tag and the multicloning site that allowed us to cut out the six histidines at the N-ter of recombinant Max. This plasmid became our standard for Max preparation. Because we used a recombinant His-Tag-TEV protease, a second Ni2+-NTA column was used to further purify the cleaved Max protein from its His-Tag and the protease.

Investigation of rhodaminated peptido-mimetic molecules internalization

MCF-7 and HCT-116 cells were cultured in RPMI 1640 supplemented with 5% FCS. Cells were grown on an appropriate cover slide (area of ~5 cm2). 50 µl of complete medium containing the fluorescent peptide (10 µM) were added over the intact cells. This cover slide was reverted and placed over a slide. The margins of the cover slides were sealed with a tiny amount of nail


polish. The chamber formed is ~0.1 mm deep. Four independent slides were used; ~100 HCT-116 cells were examined in each slide containing ~1000 cells. After 2 h of incubation at 37°C, the cells were examined with a Zeiss fluorescence microscope by using a 100-W (HBO 100W/2 OSRAM) lamp, an excitation cut-off at 546 nm, an emission cut-on at 580 nm, and a ×40 Zeiss Neofluar objective. Fluorescence intensity was measured as already reported (10).

Fluorescence anisotropy

The formation of the hetero-dimers was measured in terms of increased anisotropy (Polarion, Tecan Co.). All the experiments were performed in Polar Buffer (HEPES pH 7, 100 mM; KCl, 250 mM; MgCl2, 15 mM; DTT, 5 mM; EDTA 5 mM); the experimental volume was always 200 µl per well, and the plates used were 96-wells black nonbinding plates (Corning). 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 done at different times of incubation to assess the stability of the reactions. Results refer to a measurement after 75 min of incubation.

Peptides secondary structure analysis by CD

CD spectra were recorded on a Jasco J-700 spectropolarimeter. All spectra were recorded in nitrogen atmosphere at room temperature using a 0.05 cm path-length quartz cell. Each spectrum was an average of 10 scans over 260–180 nm.

The experiments on Int-H1 (0.26 mM) were recorded under the following conditions: 1) water solution 150 mM NaCl and 2) water/TFE mixture (50:50 and 10:90, respectively).

The experiments on HLH (0.26 mM) were recorded under the following conditions: 1) water solution 150 mM NaCl; 2) phosphate buffer pH = 5.5; 3) water/HFA mixture (75:25, 50:50, and 33:67, respectively); 4) water solution 10 mM of sodium dodecyl sulfate (SDS).

The experiments on R5P and F7P (0.29 mM) were recorded in the following conditions: 1) water solution 150 mM NaCl; 2) water/TFE mixture 50:50.

From the CD spectra, the peptide secondary structure elements were obtained through the comparison with CD spectra of proteins and peptides with known secondary structure using CONTIN or K2D programs (12–13).

Peptido-mimetics / hetero-dimer interaction studies by CD

Thermal denaturation experiments of Myc, Max, and of Myc-Max in the presence or absence of three different peptido-mimetic molecules (Int-H1, R5A, Q13A) were carried out in 12.5 mM phosphate (pH 7.4), 150 mM KCl, 0.25 mM EDTA, and 1 mM DTT in the temperature range from 6 to 60°C. CD spectra were obtained at each temperature on a Jasco J-500A spectropolarimeter equipped with a Jasco IF-500-2 data processor by averaging five scans recorded from 200 to 245 nm at a rate of 20 nm/min with a step resolution of 0.2 nm, a time constant of 4 s and a bandwidth of 2.0 nm. A thermo-stated 0.1-cm path-length cell was used. The data were expressed in terms of molar ellipticity [θ], the mean residue ellipticity in units of


degrees centimeter squared for each decimole of residue or as ellipticity at a fixed concentration. Contents of α-helix and “unordered” forms were obtained from the value of [θ] at 222 nm according to the method of Greenfield et al. (14). For comparison, some calculations were carried out by taking [θ]222=1580 deg cm2 dmol-1 for the unordered form as suggested by Chen et al. (15). This change introduced only negligible difference in the estimate of the α-helix content.

Nuclear magnetic resonance and molecular dynamic studies

1H-nuclear magnetic resonance (NMR) experiments were recorded on Bruker AMX 500 MHz. The probe temperature was maintained at 300 K, and the water suppression was carried out using the watergate scheme for the COSY and the DPFGSE scheme in the case of TOCSY and NOESY spectra.

In TOCSY experiments, a mixing time of 80 ms was applied to obtain remote scalar connectivities. NOESY spectra were recorded with mixing time of 150 and 300 ms for spin system and sequential assignments.

The analyzed samples of Int-H1 contained 1 mM peptide and 50 mM NaCl concentration, respectively, in H2O/D2O 90:10 and in H2O/D2O/TFE-d3 40:10:50.

After the sequence-specific assignment, 194 nonredundant restraints were obtained by NMR spectra and used for the determination of the Int-H1 structure.

Calculated structures were obtained from restrained molecular dynamics simulations using DYANA (ETH, Zurich) followed by energy minimization using Amber force field as implemented by the Discover program (Molecular Simulations Inc., San Diego, CA) on a Silicon Graphics Mod. O2 (O.S. IRIX 6.3) workstation.

D-aa were introduced in the DYANA library, changing signs of the z coordinates for all Hα and the side chain atoms. One hundred DYANA calculations were started with random polypeptide conformations, and the 20 resulting conformers with the lowest target function values were analyzed: these 20 structures were minimized performing 100 steps of steepest descent and 300 steps of conjugate gradient. The 10 structures with the lowest potential energy were drawn.

Assessment of c-Myc protein levels in our cancer lines

Our lines were characterized in terms of c-Myc gene expression: we used a Western blot method to determine quantitatively and specifically the number of c-Myc protein molecules per cell. (16). The c-Myc standard used was an amino terminal sequence (aa 1-262 of c-Myc) produced as a 65 kDa GST-tagged fusion protein) of known concentration (Santa Cruz Biotechnology, Santa Cruz, CA). The standard segment was used as recommended by the supplier and 0.5 ng or 1 ng was loaded on 12% SDS-PAGE next to the lane containing a cell protein extract.

All cell lines were plated and cultured in condition of subconfluence and exponential growth until recovery. Before cell lysis and protein extraction, cell density number (cell/cm2) was determined.


The cells were washed, scraped off with 10 ml cold PBS and pelleted by centrifugation at 3,000 g for 5 min. The supernatant was discarded, the pellet resuspended in 100 µL of cold lysis buffer (50 mM Tris-HCl, 250 mM NaCl, 0.1% Nodinet P40, 5 mM EDTA, 50 mM NaF, 2 mM PMSF, 20 µg/ml aprotinin and 20 µg/ml leupeptin), transferred into an Eppendorf tube and incubated for 30 min on ice. The extract was centrifuged for 5 min in a microcentrifuge at 4°C, and the supernatant stored at –20°C. 200,000 cells total protein lysates per lane were electrophoretically separated using 12% acrylamide gels and then transferred to a nitrocellulose membrane (Schleicher & Schuell, Keene, NH). After blocking with 5% milk in TBS-t buffer, the membrane was incubated overnight with a rabbit polyclonal anti-human c-Myc IgG antibody at concentration 1:5000 (N-262, Santa Cruz Biotechnology).

The nitrocellulose membrane was washed and exposed to an alkaline phosphatase-conjugate goat anti-rabbit IgG (Sigma, St. Louis, MO) at room temperature for 2 h, followed by exposure to the BCIP/NBT (5-bromo-4-chloro-3-indolylphosphate/Nitro Blue Tetrazolium substrate developer, Sigma). The standard band intensities (65 kDa standard segment) were compared with the cellular c-Myc ones (64 kDa) using a specific computer software (Scion Image). Myc concentration was reported as a ratio to the standard segment of 65 kDa. (16).

Growth curves of our lines in the presence of different concentrations of our molecules

To evaluate the capability of our RI-peptides to inhibit cell proliferation of different cancer cell lines, we followed the same schedule of treatment, already applied in our previous work (10). In synthesis, cells were plated at opportunity density at day 0 (for MCF-7: 2×104 cells/well; for HCT-116: 2.7×104 cells/well), and they were exposed to different concentrations of peptides for a total of three treatments at days 1, 4, and 7; for MCF-7 and HCT-116, the used concentration of all RI-peptidomimetic molecules was 10 µM.

Cells were detached from independent dishes and counted daily. Growth curves of treated and control cells were obtained by means of a hemocytometer count and the trypan blue dye exclusion test, in experiments performed in triplicate.

The statistical analysis of data was performed taking into account the following considerations. We know that exponential increase in cell number cannot continue ad infinitum: as the population size increases, doubling time gradually decreases and cell population reaches a peak followed by a gradual decline. This type of relationship is a curvilinear relationship: the quadratic polynomial growth curve fits reasonably well the log-transformed population data relative to the growth processes, allowing for some “curvature” in the growth of the variable under analysis over time (i.e., cell number).

ln(N) = at2+bt+ln(N0) [1]

The linear term represents the natural (exponential) growth of the population; the quadratic term represents a reduction of this natural growth caused, for example, by overcrowding or by toxic/biologic effects of the treatment.

To analyze the data we used SAS PROC MIXED, a statistical procedure for fitting multilevel mixed linear models (a quadratic polynomial equation is a function nonlinear in the variables but


linear in the parameters). From the analysis, a clear influence of the treatment seems to emerge on the behavior explained by the linear term b of the model, whereas the quadratic term of the equation appeared not significantly influenced by the imposed treatment. For this reason we expressed the activity as inhibition of proliferation comparing the linear terms (b) relative to each RI peptido-mimetic molecule:

The indexes were calculated as

[(b0-b)/b0] · 100 [2]

where b is the linear term for that peptide and b0 is the reference linear term specific for the respective control. b has a negative sign when, after treatment, the number of cells decreases. In this case the value of Eq. [2] can become higher than 100 (see Fig. 3).

RESULTS

Structural studies

Stability of Myc/Max hetero-dimer

CD study: changes in the secondary structure of Myc-Max induced by three different peptido-mimetic molecules Figure 1 shows representative CD spectra in the wavelength range 200–245 nm of Myc, Max-Max homo-dimer, and of Myc-Max hetero-dimer, in the absence or presence of three different peptido-mimetic molecules as reported in Materials and Methods.

For immediate appreciation of the spectral differences, the data are expressed as ellipticity in mdeg of solutions at a fixed (5.14 µM) concentration of each protein. The CD profiles clearly show that both Max and Myc-Max assume a coiled-coil conformation at 6°C, as judged from the [θ]222/[θ]208 ratio, which is higher than unity (17). On the contrary, the ratio is much less then unity (0.85) for Myc, which is poorly structured and does not undergo dimerization. Overall, the amount of secondary (α-helical) structure determined here for human Myc-Max, Max-Max, and Myc is in fairly good agreement with recent literature data. The number of residues in the α-helix conformation, NRα, is 122 for Max-Max at 6°C, which corresponds well to the total number of residues (116), which compose the H1, H2, and leucine zipper regions (30, 30, and 56, respectively). From the date by Fieber et al. (18) for Max-Max a value of NRα equal to 125 is obtained; likewise, NRα=110 has been calculated from the experiments by Krylov et al. (19). As concerns the hetero-dimer the works by Fieber et al. (18) and Krylov et al. (19) yield 136 and 107, respectively; the value determined in this work is 84. This discrepancy can be explained assuming that the deletion of 9 out of 12 aa of the basic region, corresponding to amino acid positions 355–364, destabilizes the α-helical H1 region. It has been reported that the α-helix content of Myc-Max shows a modest (3%) increase with respect to that of the homo-dimer (18). This result is not obtained in the present case, as the profile corresponding to the sum of the signals of Myc and Max (dotted line in Fig. 1) falls slightly below the spectrum of the hetero-dimer in the range 200–220 nm. It is clear that this apparent discrepancy is again well explained by assuming that the deletion of the basic region disfavors the nucleation of the α-helix in the H1 region.


Finally, the addition to the Myc-Max solution of an equimolecular amount of the peptide Int-H1 or R5A leads to an appreciable decrease in the α-helical content of the hetero-dimer. After subtraction of the spectrum of the peptide from that of the solution containing Myc-Max + Int-H1 or Myc-Max + R5A, we find that for either peptide, the α-helix content of the hetero-dimer undergoes a decrease by 31%, which corresponds approximately to 30 residues; this value is equal to the number of α-helical residues comprised in the H2 region of the hetero-dimer.

These results directly show that these molecules, mimics of the Myc H1 region, bind to their partner Max, introducing an additional disturbance in the formation of the four α-helix bundle (10), which results in the unfolding of the H2 region. Therefore, under these experimental conditions, the leucine zipper represents the only source of stabilization of the dimer. This extensively unfolded form occurs also as an intermediate state in the thermal denaturation of Max-Max at low pH (20). The peptide Q13A displays a different behavior: the α-helix content of the hetero-dimer undergoes a decrease by only 21%, which corresponds approximately to 18 residues showing less interaction with the hetero-dimer.

Additional structural information has been obtained by thermal denaturation experiments. Figure 2 shows the dependence of [θ]222 on the temperature for Max-Max, Myc-Max, and Myc, in the temperature range 6–60°C; the values of the dissociation constant Kd, and of the temperature and enthalpy of unfolding Tu and ∆Hu, obtained by applying van’t Hoff equation, are listed in Table 2.

Max-Max and Myc-Max show appreciable differences in the structural stability as judged from the values of Tu and ∆Hu (35 and 40°C; 56 and 40 kcal/mol, respectively). For Max-Max the values of the thermodynamic parameters are in agreement with previous data (18, 19). We note that the enthalpy of denaturation of Myc-Max is lower than that of the Max-Max, reflecting the lack of ordered structure in the H1 region; of course, this circumstance also accounts for the higher transition entropy of the homo-dimer compared with that of the hetero-dimer. [θ]222 of Myc increases almost linearly as a function of temperature, showing that the process conforms to the steady, noncooperative denaturation of short helical segments, as already reported by Fieber et al. (18). Finally, the denaturation of the hetero-dimer in the presence of an equimolecular amount of Int-H1 or R5A shows a decrease in Tu and an increase in Kd, which again reflects the unfolding of the H2 segment due to the action of the two molecules. On the contrary, the thermal denaturation in the presence of Q13A does not conform to the simple scheme proposed above. In fact, we find a lower value of the dissociation constant and a higher value of the melting temperature. Because the number of unfolding amino acids at 6°C is lower in the presence of Q13A, passing from 30 to 18, we can conclude (also in the light of the consideration that domains constituted of short peptides unfold in an extremely cooperative manner) that the mode of interaction of Int-H1 or R5A with the hetero-dimer is very different from the mode displayed by Q13A.


Anisotropy studies Fluoresceinated H1-Loop-H2-zip aa 365–439 c-Myc domain (Myc-F) and full-length Max were first mixed at different concentrations to determine an optimal experimental condition, using the minimum well-detectable amount of fluoresceinated target. These preliminary experiments indicated that Myc-F can be used at 1–5×10−8 M; and that a concentration of Max 5×10−7 M is sufficient for almost complete binding of Myc-F to Max (unpublished observation).

The anisotropy value of Myc-F alone is ~60 mA at 30°C in Polar Buffer, whereas a plateau value ranging between 110 and 120 mA is reached when the protein is all bound to its partner Max (unpublished observation).

Subsequent competition experiments were performed using the above Myc-F/Max concentrations, as a hetero-dimer target, while a range of concentrations between 1×10−7 and 5×10−6 of cold Myc was used for competition.

A competing concentration of synthetic Myc (Myc cold) 10 times higher than Max was sufficient to make ~90% of Myc-F free (range of mA values obtained in multiple experiments: 73–77).

We then studied different synthetic L- and D-peptides for their ability to dissociate our Myc-F/Max dimer. As shown in Table 3, the unrelated L-peptide Bak (BH3 domain of Bak proapoptotic factor), the L-H1-S6A-F8A, the RI-H1-S6A-F8A (H1), and the RI Int-H1-S6A,F8A (Int-H1) were not able to affect the Myc-F/Max dimer. Even an Int-H1 made of L-aa and tested at 10−4 M could not separate the two monomers (data not reported). Our 75 aa synthetic cold Myc was used again as a positive control.

Thus, we can conclude that our molecules are not able to completely separate the two monomers of the Myc/Max hetero-dimer, even if CD data suggest a partial opening of the structures.

RI-peptides structures

The secondary structures of R5P, F7P, HLH, and Int-H1 were analyzed. In Table 4 the conformational features of our molecules, obtained by CD at room temperature in physiological solution and in water/TFE mixture are reported (note that TFE is the solvent most widely used to favor helical structure formation in polypeptides; its effect occurs through either a direct or an indirect strengthening of the intra-α-helical hydrogen bonds).

As expected, the introduction of a proline residue in both cases causes a decrease of helical content. HLH has a low propensity to form a helix structure in the analyzed conditions, while CD data indicate the presence of an important percentage of β-sheet conformation.

The Int-H1 peptide was also investigated by NMR in water and in a water/TFE 50:50 mixture. From the analysis of the NMR spectra, the complete assignment of the amino acids was obtained, and a set of long range correlations was collected.

Using these data, we carried out a DYANA simulation with Int-H1: a left-handed helix structure was obtained. Structure calculations suggest that the second half of the RI-peptide (corresponding to the Int region) shows a well-defined α-helical structure: the root mean square


deviation (RMSD) obtained from the superposition of the backbone atoms of the 10 best structures is RMSDbb(Int) = 0.81 ± 0.27.

A large conformational flexibility exists in the middle of the molecule. The N-terminal half of the molecule (the area of the H1 motif) appears not ordered. A much less defined left-handed helix can be obtained only between residues 6–12 of H1 (RMSDbb(6–12)=0.91±0.30).

Internalization

RI-rhodaminated peptides internalization

Quantitative penetration into cancer cell lines was assayed labeling the peptides with the chromophore rhodamine at the N-terminus. After an incubation of 2 h at 37°C (peptides concentration 10 µM), we investigated the penetration and the subcellular localization of these RI peptides (Table 5) compared with Bak (unrelated L-peptide that is not able to enter the cells: data not reported).

In conclusion, a moderately higher degree of internalization was shown for Int, whereas no significant differences in internalization properties were found among the 30-mer peptides analyzed (Q13A, K4A, and Int-H1). We have confirmed (10) that our molecules not only are capable of entering our cells, but they do concentrate significantly at the intracellular level, probably as a consequence of an enormous intracellular concentration of low-affinity binding sites.

c-Myc protein levels in our lines

The values of total c-Myc protein molecules per cell in our cancer lines in exponential growth was on average 61,300 ± 13,200 in MCF-7 cells and 142,900 ± 13,800 in human colon HCT-116 cells, whereas in rat embryo fibroblasts Rat-1a cells (spontaneously immortalized) was 13,900 ± 10,000.

Because of the function of c-Myc as a transcription factor, the protein localized to the nucleus might be considered as “active Myc” in comparison to the “inactive” cytoplasmic form (only 13% in MCF7 cells and only 4% in HCT-116 cells, respectively).

No proteins with lower molecular weights were detected, indicating that neither degradation nor nonspecific antibody binding had occurred.

Inhibition of growth by our peptido-mimetic molecules in MCF-7 and HCT-116 lines

Peptides coming from Ala-scan were tested only on the HCT-116 cell line, the most rapidly dividing line (from day one to day eight: T ≈ 17 h), whereas other peptides were tested on both cell lines.

In Figure 3, a histogram representing the percentages of inhibition of HCT-116 cell growth is shown. Similar results (not all molecules tested) were obtained with MCF-7 cells (unpublished observation). Looking at Fig. 3, we can subdivide our molecules into six categories: 1) internalization fragment Int; 2) H1: a RI peptide without the RI-Int internalization sequence; 3)


lead (10) Int-H1; 4) molecules less active than the lead; 5) molecules substantially equiactive with the lead; and 6) molecules more active than the lead.

Internalization fragment The retro-inverso form of the IIIrd α-helix of Antennapedia (Int), with the two Ile substituted with Val, is largely less active than the lead, even if it is able to internalize quite well (Table 5). The percentage of inhibition is less than 30%. RI-Int (without Val substitutions) was comparable to Int and is not reported.

H1 H1 is inactive and this molecule is unable to enter intact cells.

Lead Int-H1 The lead Int-H1 is internalized into the cell and active in inhibiting proliferation: it is considered our reference lead molecule (10). The percentage of inhibition was 58.6% in HCT-116 and 83.7% in MCF-7 cells, respectively.

Less active molecules in respect to the lead Lys in 4 and Arg in 5 turn out to be necessary for antiproliferative activity, because the activity of related peptides (K4A and R5A) is strongly reduced; in fact, the percentage of inhibition in HCT-116 cells was about 3 times lower than the lead. For R5P and F7P, an even greater decrease in activity was noted: the percentage of inhibition in HCT-116 cells was about 4 times lower; in MCF-7, cells were about 5–6 times lower, in respect to the lead. For HLH (our longest peptido-mimetic molecule), the percentage of inhibition was 2–3 times lower in respect to the lead.

Molecules substantially equiactive with the lead Substitution (one at a time) with an alanine of most of the H1 amino acids (I14, D12, L10, F7, L3, and N1) leaves the antiproliferative activity substantially unchanged. The functional groups present in these positions seemed unnecessary for inhibitors’ effectiveness, and elimination of these side chains did not improve activity. Differences in percent inhibition, with respect to the lead, were not biologically relevant for I14A, D12A, L10A, F7A, L3A, and N1A.

Molecules more active than the lead By substitution (one at a time) of Gln in 13, Arg in 11 and Glu in 2 with Ala (Q13A, R11A, and E2A), a complete inhibition of cell proliferation was observed, even accompanied with a decrease of the initial seeded cell number. These data suggest that ablation of a single side chain does not always decrease, but may sometimes increase biological activity.

DISCUSSION

In our previous work (10), we reported the following essential properties of our lead molecule Int-H1: 1) The presence of the RI-Int sequence was essential for internalization and antiproliferative activity of our RI peptido-mimetic molecules; 2) an RI-isoamphipathic variant in the Myc-H1 region of our lead was 3–10 times less active; 3) modulation of transcription levels of ornithine decarboxylase (inhibited), p53 (inhibited), and glyceraldehyde-3-phosphate dehydrogenase (not inhibited) was well compatible with an interference by our lead molecule at the level of Myc transcriptional activity; and 4) on a molar basis, RI peptides were ~5–10 times more potent and 30–35 times more stable (a half-life of about 10 days in complete medium) than their corresponding L-forms.


In this paper, we have extended our investigation in two main directions: 1) the role of the different side chains of D-aa forming the H1 of Myc homologous sequence in our lead, in terms of antiproliferative activity, as well as investigation of three additional variants of our lead molecule: R5P; F7P; HLH; and 2) the biochemical interaction of our peptido-mimetic molecules with the hetero-dimer [HLH-zip of c-Myc | entire Max].

We showed that our molecules, which arise from Helix-1 of c-Myc, are able to inhibit cancer cell growth at low micromolar concentrations. c-Myc was overexpressed in our target cancer lines. Our purpose was to work at a protein–protein interaction level: c-Myc is not an isolated protein, but it acts within a higher-order structure named enhanceosome (6); in fact, it must bind with other entities of that complex structure to become active (21). We wanted to interfere at the level of c-Myc higher-order structure disrupting one of its connections. One fundamental step is the hetero-dimerization with Max.

Our RI peptides mimic H1 of Myc. Our molecules were therefore tested for their ability of interfering at the biochemical level with the hetero-dimer (CD data). The lead molecule opens the four α-helix bundle at the H2 level, without opening the leucine zipper. Anisotropy data confirm that we never achieve a complete separation of the two monomers at the biochemical level. For this interference at the level of H2, side chains projecting at the inside of the bundle are essential. Q13A misses an important Gln side chain projecting at the inside and is very poor at interfering with the four α-helix bundle (CD data). On the contrary, R5A, which has lost an important side chain projecting at the outside, is as good as the lead in its interference with the H2 chain of the bundle. For antiproliferative activity the opposite is required: side chains projecting at the outside are essential, whereas the loss of side chains projecting on the inside confers an advantage in terms of antiproliferative potency.

Our peptido-mimetic molecules displayed different potencies for the same degree of internalization. Our biologically most active molecules seem to act as a wedge between the hetero-dimer itself and a third external protein, part of the Myc/Max enhanceosome.

In what way can Int-H1, our most studied molecule, resemble a Helix1 of Myc? We know through CD data that our lead RI-peptide has ~17%–35% (depending on different buffers) propensity to form a helix, but in solution a D-aa, L-helical conformation is present mostly in the Antennapedia part, whereas H1 of Myc seems to be mostly a random coil (NMR data). Unusual L-helices can sometimes be found in proteins made of L-aa (22), driven by the structural context. Something similar (in a specular form) could perhaps happen to the H1 portion of our RI-peptides, when positioned in a convenient macromolecular surrounding. It is well known on the other hand that, as random coils, L-peptides and corresponding RI-peptides maintain the same spatial position of their respective side chains (23). We learned by Ala-scan mapping that the presence of two adjacent external basic side chains is required for antiproliferative activity: Lys in 4 and Arg in 5. On the contrary, our lead molecule becomes more potent if the following side chains are taken out: Glu in 2, Arg in 11, Gln in 13.

We looked at 1NKP (Protein Data Bank) Myc/Max/DNA X-ray structure (11). Our molecules could mimic a Myc-H1 interaction with a third protein external to the hetero-dimer. Myc binds Max forming the four-α-helix bundle. We referred to residues involved in this interaction as “inner” with respect to the hetero-dimer. Helix 1 of each monomer binds in part to Helix 1 of the


other monomer and in part to Helix 2 of the same peptide. Furthermore, amino acids at the C-terminal of each Helix 1 are nearer to and more internal than amino acids of their own loop; therefore, they are hidden by loop residues. In other words, amino acids of the C-terminal of Myc H1 from Ser 6 to Ile 14 (from aa 920 to aa 928 of chain A of 1NKP) can make contacts within the helices or with their own loop, but not with the outside of the hetero-dimer. Also Glu 2 (Glu 916 of chain A of 1NKP) is directed toward the inside of the hetero-dimer, keeping contact with Max residues (11).

It is essential that our peptides mimic H1 residues freely exposed at the outside of the hetero-dimer, that is, aa nearer to the basic helix, as Lys in 4 (918 of chain A of 1NKP) and Arg in 5 (919 of chain A of 1NKP).

Lys in 4 and Arg in 5 turned out to be indeed necessary, because the activity of related peptides without these side chains was strongly reduced compared with the lead Int-H1 activity. These basic amino acids in Myc Helix 1 structure (11) project toward the outside of the hetero-dimer and are presumably involved, maybe together with some loop residues and BH basic side chains, in interactions with a local third protein. Arg 904 of 1NKP is, in fact, freely projecting its side chains to the outside (in the corresponding position of our lead a Lys is instead present), whereas Lys 902, Arg 903, His 906, Arg 911, Arg 913, Arg 914, and Lys 939, together with Asn 907, Glu 910, and Pro 938, are involved in interactions with phosphate groups or bases of the DNA chain.

Our D-peptide amino acidic residues, corresponding to those of H1 of Myc involved in four α-helix bundle interactions (inside), were useless for “talking” with the third local protein of the enhanceosome, and could even be a steric or electrostatic hindrance for receptor surface approaching by our peptide-mimetic molecules’ [Gln 13, Arg 11, and Glu 2 of Myc-H1 (927, 925 and 916 of chain A of 1NKP)] and seem indeed to be involved in these internal interactions, projecting their side chain toward the inside of the hetero-dimer. They cannot interact with a third protein, obviously present at the outside (e.g., Gln 927 seems to interact with Gln 251 of Max, chain B of 1NKP). Gln 13 (927) and Arg 11 (925) side chains could represent only a hindrance (a potential for low-affinity interactions with nonspecific secondary sites) from the point of view of a selective interaction with an external protein. This could explain why a reduction of these active side chains to the methyl group of alanine could improve the biological activity of our peptidomimetic molecules. Similar considerations can be made for Glu in position 2 (916).

Substitutions with Pro favor the general framework of our reasoning; in fact, in R5P (R in 919 of chain A of 1NKP), we change a useful residue (Arg) and introduce a disturbing kink, drastically reducing antiproliferative activity. In F7P (F in 921 of chain A of 1NKP), we change a nonessential amino acid, but we still introduce a disturbing kink in the structure, decreasing the possibility of a correct molecular docking toward the target protein.

The HLH molecule maintains the RI-sequence of Helix1-loop-Helix2, but it cannot be made of two D-helices separated by an intervening loop, also because it has a clear tendency to assume a β-sheet conformation (CD data and not reported NMR data). This predominant and incorrect secondary structure, very different from the c-Myc portion of Myc/Max hetero-dimer, is probably responsible for the low activity of this larger peptido-mimetic molecule.


In conclusion, the activity of our molecules seems to be side chain sequence-specific. In addition, kinks and β-sheets are forbidden secondary structures. An interaction with an external third protein of the Myc/Max enhanceosome seems to be the favored hypothesis.

It has been reported (24) that the domain BR-H1-loop-H2 of c-Myc binds the ATP-ase protein INI-1, between aa 181 and 240 (of INI-1).

A His-tagged carboxy-terminal part of c-Myc (encompassing the BR-HLH-Zip dimerization domain) was expressed in E. coli, and the bacterial lysate was purified on an Ni-NTA-agarose resin, leaving the c-Myc peptide bound to the resin. In a similar way, we produced the whole INI-1 protein (without His-tag), and this second bacterial lysate was applied to the resin bound to c-Myc; the mixture was allowed to interact for 60 min at 4°C in binding buffer according to Cheng et al. (24). After washing, an appropriate amount of the preboiled denatured resin-protein complex was run on SDS-PAGE. The gels were then transferred on a nitrocellulose membrane and blotted against either mAb anti-c-Myc (9E10 Santa Cruz) or polyclonal Ab anti-INI-1 (H-300 Santa Cruz). Unfortunately, for the moment, we have been unable to confirm the results of Cheng et al. (24): only c-Myc could be detected, but no bound INI-1. Small variations in the experimental conditions could be important. We continue to work in this direction.

We have also started to study some basic pharmacokinetic properties of our peptido-mimetic molecules. These 30 D-aa long molecules, capable of efficient cellular internalization, seem quite interesting in this respect. To our Int-H1 lead, we added a C14 labeled Gly. Nine mice (Balb-c female mice 1.5 month old) were injected intravenously in the tail vein, 5 mg/kg of RI-Int-H1-14CGly, 10,000 cpm/g of mice. Peak levels in tissues were observed around 6 hours: ~2,000 cpm/g for liver and lung, 4,000 cpm/g for spleen, and 12,000 cpm/g for kidney. Negligible levels were found in the brain (no crossing of the hemato-liquoral barrier). Our type of molecules seems capable of reaching good concentration levels in the examined tissues, with the exception of the brain, and these molecules have a rather long half-life in vivo, roughly 24 h (unpublished observation). They could represent a favorable new class of potential drugs. We intend to expand further these pharmacokinetic studies.

Finally, we would like to prepare molecules looking directly toward the external part of c-Myc. They could be potentially more selective than inhibitors of a third protein. This one could easily be a modular protein of several enhanceosomes.

ACKNOWLEDGMENTS

We thank Dr. Gennaro Citro of Institute Regina Elena, Rome, Italy, for his collaboration with pharmacokinetic studies, Dr. Menotti Ruvo of the University of Naples for his valuable teachings on peptide synthesis, Dr. Luciano Zardi of Institute Giannina Gaslini of Genoa for fruitful discussions, and Prof. Claudio Nicolini of the University of Genoa for the systematic usage of instruments from his laboratory. We thank Prof. Martino Bolognesi, Dr. Domenico Bordo, and Dr. Camillo Rosano for useful suggestions about molecular design. This work was supported by Ministero della Salute Finalizzato 2001 conv. N°133, Ministero della Salute Finalizzato 2002 conv. N°180, Ministero della Salute Finalizzato 2003 conv. N°137, MIUR 25 COFIN 2003, CNR CU02.00265.ST/971997, FIRB RBNEO1X3NB, PNR Oncologia 2000, FIRB RBAU01Y3SN, Associazione Italiana per la Ricerca sul Cancro. We are grateful to the Istituto


Superiore di Oncologia for the systematic usage of the Applied Biosystems 433A peptide synthesizer, the BioCAD 700E Perfusion Chromatography Workstation, and the LCQDECA mass spectrometer.

REFERENCES

1. Bhatia, K., Spangler, G., Gaidano, G., Hamdy, N., Dalla-Favera, R., and Magrath, I. (1994) Mutations in the coding region of c-myc occur frequently in acquired immunodeficiency syndrome-associated lymphomas. Blood 84, 883–888

2. Nesbit, C. E., Tersak, J. M., and Prochownik, E. V. (1999) Myc oncogene and human neoplastic disease. Oncogene 18, 3004–3016

3. He, T. C., Sparks, A. B., Rago, C., Hermeking, H., Zawel, L., da Costa, L. T., Morin, P. J., Volgestein, B., and Kinzler, K. W. (1998) Identification of c-myc as a target of the APC pathway. Science 281, 1509–1512

4. Cimoli, G., Bagnasco, L., Pescarolo, M. P., Avignolo, C., Melchiori, A., Pasa, S., Biasotti, B., Taningher, M., and Parodi, S. (2001) Signaling proteins as innovative targets for antineoplastic therapy: our experience with the signaling protein c-Myc. Tumori 87, S20–S23

5. Ma, P. C. M., Rould, M. A., Weintraub, H., and Pabo, C. O. (1994) Crystal structure of MyoD bHLH domain-DNA complex: perspectives on DNA recognition and implications for transcriptional activation. Cell 77, 451–459

6. Merika, M., and Thanos, D. (2001) Enhanceosomes. Curr. Opin. Genet. Dev. 11, 205–208

7. Henriksson, M., and Luscher, B. (1996) Protein of the Myc network, essential regulators of cell growth and differentiation. Adv. Cancer Res. 68, 109–182

8. Giorello, L., Clerico, L., Pescarolo, M. P., Vikhanskaya, F., Salmona, 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

9. Derossi, D., Calvet, S., Trembleau, A., Brunissen, A., Chassaing, G., and Prochiantz, A. (1996) Cell internalization of the third helix of the Antennapedia Homeodomain is receptor-independent. J. Biol. Chem. 271, 18188–18193

10. 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

11. 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


12. Provencher, S. W., and Glockner, J. (1981) Estimation of globular protein secondary structure from circular dichroism. Biochemistry 20, 33–37

13. Andrade, M. A., Chacon, P., Merelo, J. J., and Moran, F. (1993) Evaluation of secondary structure of proteins from UV circular dichroism using an unsupervised learning neural network. Prot. Eng. 6, 383–390

14. Greenfield, N., Davidson, B., and Fasman, G. D. (1967) The use of computed optical rotatory dispersion curves for the evaluation of protein conformation. Biochemistry 6, 1630–1637

15. Chen, Y. H., Yang J. T., and Chau, K. H. (1974) Determination of the helix and beta form of proteins in aqueous solution by circular dichroism. Biochemistry, 13, 3350-3359.

16. Rudolph, C., Adam, G., and Simm, A. (1999) Determination of copy number of c-Myc protein per cell by quantitative western blotting. Anal. Biochem. 269, 66–71

17. Cooper, T. M., and Woody, R. W. (1990) The effect of conformation on the CD of interacting helices: a theoretical study of tropomyosin. Biopolymers 30, 657–676

18. Fieber, W., Schneider, M. L., Matt, T., Kräutler, B., Konrat, R., and Bister, K. (2001) Structure, function, and dynamics of the dimerization and DNA-binding domain of oncogenic transcription factor v-Myc. J. Mol. Biol. 307, 1395–1410

19. Krylov, D., Kasai, K., Echlin, D. R., Taparowsky, E. J., Arnheiter, H., and Vinson, C. (1997) A general method to design dominant negatives to B-HLHZip proteins that abolish DNA binding. Proc. Natl. Acad. Sci. USA 94, 12274–12279

20. Naud, J. F., Frédéric, G., Raymund, W., Benoit, C., and Lavigne, P. (2003) Improving the thermodynamic stability of the leucine zipper of Max increases the stability of its b-HLH-LZ:E-box complex. J. Mol. Biol. 326, 1577–1595

21. Kurt, W. Kohn, Mirit I. Aladjem, Stefania Pasa, Silvio Parodi, and Yves Pommier, "Molecular Interaction Map of Mammalian Cell Cycle Control." In: Encyclopedia of the Human Genome, Nature Publishing Group, London, 2003, pp. 457–474.

22. Kleywegt, G. J. (1999) Recognition of spatial motifs in protein structure. J. Mol. Biol. 285, 1887–1897

23. Guichard, G., Benkirane, N., Zeder-Lutz, G., Van Regenmortel, M. H. V., Briand, J. P., and Muller, S. (1994) Antigenic mimicry of natural L-peptides with retro-inverso-peptidomimetics. Proc. Natl. Acad. Sci. USA 91, 9765–9769

24. Cheng, S. W. G., 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

Received June 10, 2004; accepted December 6, 2004.


Table 1 List of studied peptides with shortcut abbreviations and sequences Peptides Short-cut

abbreviation Sequence

RI-Int KKWKMRRNQFWIKIQR *RI-Int-VV <Int> KKWKMRRNQFWVKVQR *RI-H1-S6A,F8A <H1> IQDRLAAFARKLEN *RI-Int-VV-H1-S6A,F8A <Int-H1> IQDRLAAFARKLENKKWKMRRNQFWVKVQR *RI-Int-H1-S6A,F8A-Loop-H2 <HLH> VSLIYATAKKLIVVKPAKENNELEP

IQDRLAAFARKLENKKWKMRRNQFWIKIQR RI-Int-H1-R5P,S6A,F8A <R5P> IQDRLAAFAPKLENKKWKMRRNQFWIKIQR RI-Int-H1-S6A,F7P,F8A <F7P> IQDRLAAPARKLENKKWKMRRNQFWIKIQR RI-Int-VV-H1-S6A,F8A,I14A <I14A> AQDRLAAFARKLENKKWKMRRNQFWVKVQR *RI-Int-VV-H1-S6A,F8A,Q13A <Q13A> IADRLAAFARKLENKKWKMRRNQFWVKVQR RI-Int-VV-H1-S6A,F8A,D12A <D12A> IQARLAAFARKLENKKWKMRRNQFWVKVQR RI-Int-VV-H1-S6A,F8A,R11A <R11A> IQDALAAFARKLENKKWKMRRNQFWVKVQR RI-Int-VV-H1-S6A,F8A,L10A <L10A> IQDRAAAFARKLENKKWKMRRNQFWVKVQR RI-Int-VV-H1-S6A,F7A,F8A <F7A> IQDRLAAAARKLENKKWKMRRNQFWVKVQR RI-Int-VV-H1-R5A,S6A,F8A <R5A> IQDRLAAFAAKLENKKWKMRRNQFWVKVQR *RI-Int-VV-H1-K4A,S6A,F8A <K4A> IQDRLAAFARALENKKWKMRRNQFWVKVQR RI-Int-VV-H1-L3A,S6A,F8A <L3A> IQDRLAAFARKAENKKWKMRRNQFWVKVQR RI-Int-VV-H1-E2A,S6A,F8A <E2A> IQDRLAAFARKLANKKWKMRRNQFWVKVQR RI-Int-VV-H1-N1A,S6A,F8A <N1A> IQDRLAAFARKLEAKKWKMRRNQFWVKVQR Bak (L-peptide) QVGRQLAIIGDDINR Natural H1 of c-Myc is taken as a reference for our conventional numeration: N1 E2 L3 K4 R5 S6 F7 F8 A9 L10 R11 D12 Q13 I14. The six molecules preceded by a green asterisk have also been synthesized with a rhodamine flag at their amino-terminal, for internalization studies.


Table 2 Thermodynamic parameters of the unfolding of Max-Max and of Myc-Max in the presence or absence of different peptido-mimetic molecules

Protein Kd(37), M ∆Hu, kcal/mol

∆Su, e.u.

Tu, °C

Max-Max Myc-Max

Myc-Max-<Int-H1> Myc-Max-<R5A>

Myc-Max-<Q13A>

1.3 10–5

1.6 10–6

3.0 10–6

2.0 10–6

1.4 10–6

56 40 39 41 42

182 128 126 132 135

35 40 38 38 39


Table 3 Competition study: anisotropy values

Myc L-H1-S6A,F8A Bak <H1> <Int-H1>

Competitor 0 M (all bound) 106.67 ± 1.33 112.00 ± 1.00 111.33 ± 0.67 112.00 ± 1.53 109.00 ± 1.15

Competitor 1×10-6 M 106.00 ± 1.53 112.33 ± 0.33 111.67 ± 0.67 111.67 ± 0.33 110.00 ± 3.21

Competitor 5×10-6 M 87.33 ± 1.76 113.33 ± 0.33 113.67 ± 1.86 113.67 ± 0.67 107.33 ± 0.67

Competitor 1×10-5 M 80.33 ± 2.96 113.00 ± 0.58 112.67 ± 0.67 113.33 ± 1.33 111.67 ± 2.67

Free 59.33 ± 1.76 59.67 ± 0.88 57.33 ± 0.88 58.00 ± 1.53 59.00 ± 1.00

Data are given in triplicate ± SE. Max is always 5×10–7 M Myc-F is always 5×10–8 M. Competitors are used at different concentrations.


Table 4

Estimation of helix percentage obtained from the CONTIN or K2D programs

Peptide Buffer Alpha Beta Random <Int-H1> H2O 150 mM NaCl 17% 25% 58% <Int-H1> H2O-TFE 50-50 35% 20% 45%

<HLH> H2O 150 mM NaCl 5% 47% 48% <HLH> H2O-TFE 50-50 26% 20% 54%

<R5P> H2O 150 mM NaCl 8% 18% 74% <R5P> H2O-TFE 50-50 20% 15% 65%

<R7P> H2O 150 mM NaCl 8% 19% 73% <R7P> H2O-TFE 50-50 20% 12% 68%

Peptides are analyzed in H2O 150 mM NaCl and H2O-TFE 50-50. Measurements are also made in a DTT intracellular buffer (19); the percentages are similar to the others. id est alpha-helix value is low; a lot of noise is present in spectra (data not shown).


Table 5 Fluorescence intensity (arbitrary units) ± SE in cytoplasm (C), nucleus (N), nucleolus (Nl), and extracellular (E) registered with different peptides

Peptide C

N

Nl

E

<Q13A> 4.75 ± 0.095 4.5 ± 0.101 5.75 ± 0.139 1.00 ± 0.015

<K4A> 4.80 ± 0.104 5.25 ± 0.116 6.10 ± 0.154 1.00 ± 0.016

<Int-H1> 4.70 ± 0.127 5.65 ± 0.149 7.15 ± 0.202 1.00 ± 0.020

<Int> 7.10 ± 0.125 7.60 ± 0.138 12.10 ± 0.195 1.00 ± 0.016

Values are given as means ± SE. Values were normalized for extracellular fluorescence intensity (concentration 10 µM), assuming an average cell depth of 20 µm and considering that the depth of the chamber is 100 µm.


Fig. 1

Figure 1. CD spectra of Int-H1 (a), R5A (b), Q13A (c), Myc (d), Max-Max (g) and of Myc-Max in absence (i) or presence of Int-H1, R5A, Q13A (e, f and h, respectively). The concentration of all the components was 5.14 µM. The dotted line corresponds to the sum of the signals of Myc and Max-Max.


Fig. 2

Figure 2. The dependence of [θ]222 on the temperature of Myc (�), Max-Max (�) and of Myc-Max in the absence (�) or presence (�) of Int-H1, (�) R5A, and (�) Q13A, respectively.


Fig. 3

Figure 3. Percent inhibition of growth, induced in HCT-116 cells by our peptido-mimetic molecules at 10 µM, defined as [(b0-b)/b0]·100, where [ln(N)=at2+bt+ln(N0)], according to Equation 2 and Equation 1 of text, respectively. The SE

is added in dark gray to the right of each horizontal light gray bar.


Sequence specific peptidomimetic molecules inhibitors of a protein-protein interaction at the helix...

Documents

Transcript of Sequence specific peptidomimetic molecules inhibitors of a protein-protein interaction at the helix...