A Metal-Free DNA Nuclease Based on a Cyclic Peptide Scaffold

Bioconjugate Chem. 2010, 21, 1425–1431

1425

A Metal-Free DNA Nuclease Based on a Cyclic Peptide Scaffold Shadad Alkhader,† Aviva Ezra,† Jana Kasparkova,‡ Viktor Brabec,‡ and Eylon Yavin*,† The Institute for Drug Research, The School of Pharmacy, The Hebrew University of Jerusalem, Hadassah Ein-Karem, Jerusalem 91120, Israel, and Institute of Biophysics, Academy of Sciences of the Czech Republic, v.v.i., 61265 Brno, Czech Republic. Received December 9, 2009; Revised Manuscript Received May 5, 2010

The ability to cleave DNA with the aid of chemical nucleases has been a challenge in the scientific community, particularly in the absence of a redox active metal ion. Inspired by structural characterization of the active site found in Staphylococcal nuclease, we have designed a series of organic molecule comprising cyclic pentapeptides conjugated to a DNA intercalator (e.g., anthraquinone). The cyclic peptide is designed to cleave the phosphodiester backbone, whereas the intercalator is expected to improve binding affinity to the substrate (DNA). Our lead compound (1-AQ), composed of the cyclic peptide cyc-D-Lys-Gly-Arg-Ser-Arg conjugated to anthraquinone, degrades DNA into small fragments at physiologically relevant conditions (i.e., 37 °C, pH ) 7.4). We find that 1-AQ is highly effective in degrading duplex DNA at micromolar concentrations as corroborated by agarose and polyacrylamide gel electrophoresis. Changing the DNA intercalator to acridine (1-Ac) renders the compound comparable in nuclease activity to 1-AQ. In comparison to control compounds (Lin-1 and 1) that lack either the cyclic scaffold or the DNA intercalator, our lead compound (1-AQ) is found to be significantly more active as a DNA chemical nuclease. We have studied the importance of the triad (Arg-Ser-Arg) as the designed module for DNA cleavage. Changing L-Ser to L-Glu (cyc-D-Lys-Gly-Arg-Glu-Arg, Glu-AQ) results in an inactive compound, whereas the cyclic peptide Gly-AQ (cyc-D-Lys-Gly-Arg-Gly-Arg, where glycine replaces L-serine) has similar DNA nuclease activity to 1-AQ. In addition, changing the stereochemistry from D-lysine to L-lysine results in a cyclic peptide (1-L-AQ) exerting weak DNA nuclease activity, highlighting the importance of the cyclic backbone conformation for efficient DNA nuclease activity. The addition of ROS scavengers does not reduce DNA nuclease activity; an observation that supports a hydrolytic cleavage mechanism. Finally, we have estimated the kinetics of DNA cleavage of a 15-mer duplex DNA substrate by compound 1-AQ. By monitoring DNA duplex degradation by following the change in absorbance (hyperchromicity) at various 1-AQ concentrations, we report a maximal kobs value (as an underestimation of kmax) of 1.62 h-1 at a 7.5-fold of 1-AQ. We have also compared the other two active peptide conjugates, namely, 1-Ac and Gly-AQ to that of 1-AQ. Both compounds exert similar nuclease activity to that of 1-AQ. To the best of our knowledge, this is the most active metal-free DNA nuclease reported to date that exerts its DNA nuclease activity at biologically relevant conditions.

INTRODUCTION In recent years, there has been considerable interest in designing synthetic small organic molecules as catalysts that hydrolyze the phosphodiester bond (1, 2). Such compounds, when targeted to certain DNA/RNA sequences, might lead to the development of highly specific artificial restriction enzymes, as well as potential therapeutic molecules for specific gene silencing. Several research groups have been working on the development of DNA cleavage agents that are metal-free and hydrolytic in nature (1-10). The absence of a redox active metal ion has the advantage of avoiding complications such as metal dissociation and uncontrolled redox chemistry. Nonetheless, ligand-based hydrolytic DNA cleavage remains a challenge, as it is well-known that the DNA phosphodiester bond is very stable under physiological conditions with a half-life hydrolysis rate of about 200 million years (3). Various peptides have been used as metal-free DNA nucleases (1, 3-6); however, DNA cleavage was typically observed at high peptide concentration (e.g., 10 mM) (5) or at non-physiological conditions (e.g., 50 °C) (4). Another synthetic * Corresponding author. The Institute for Drug Research, The School of Pharmacy, The Hebrew University of Jerusalem, Hadassah EinKarem, Jerusalem 91120, Israel. E-mail: [email protected], Tel: 972-2-6758692, Fax: 972-2-6757076. † The Hebrew University of Jerusalem. ‡ Academy of Sciences of the Czech Republic.

approach is based on utilizing macrocyclic compounds as scaffolds to which various ligands (e.g., guanidine) are appended (7-10). In several cases, hydrolytic DNA cleavage was achieved at physiological conditions (pH 7.5, 37 °C) though ligand concentration was relatively high (100-200 µM) (7, 8). In a recent report, DNA cleavage was shown to be promoted by 2,6-pyridinecarboxamide substituted with two guanidine groups (11). As in the previous cases, a relatively high concentration of ligand was required for metal-free DNA hydrolytic cleavage. Staphylococcal nuclease from Staphylococcus aureus is a Ca2+ dependent extracellular enzyme that is secreted by these bacteria as a defense mechanism against viral DNA and RNA (12). This enzyme has been exploited for achieving site-specific DNA and RNA cleavage by it is conjugation to triplex-forming oligonucleotides (13, 14). Inspired by the detailed mechanistic and structural characterization of this enzyme (12, 15-24) indicating several amino acid residues (e.g., L-Arg and L-Glu) as well as Ca2+ as key players in the hydrolysis of the phosphodiester bond, we set out to prepare a cyclic pentapeptide scaffold presenting such amino acids. On the basis of extensive studies by the Kessler group (25, 26) on cyclic RGD peptides arranged in a βII′/γ-turn, we synthesized several cyclic pentapeptides (cRSRGK, cRSRGK, cRERGK, and cRGRGK where K ) D-lysine) to which DNA intercalators (anthraquinone ) AQ or acridine ) Ac) were conjugated via the ε-amine of (D or L) lysine (Scheme 1).

10.1021/bc900543b  2010 American Chemical Society Published on Web 07/19/2010

1426 Bioconjugate Chem., Vol. 21, No. 8, 2010 Scheme 1. Synthesis of Cyclic Pentapeptides Conjugated to DNA Intercalators (Anthraquinone or Acridine)

These cyclic peptides were designed to bring about the following properties: (1) Three amino acid side chains (ArgSer-Arg, Arg-Gly-Arg, or Arg-Glu-Arg) as a triad (γ-turn) are designed to bind to the phosphodiester backbone and potentially cleave DNA; (2) the D-Lysine is then oriented to allow conjugation of a DNA binding moiety (intercalators AQ or Ac) affording a βII′/γ-turn arrangement with D-lysine at the i + 1 position (25, 26). Herein, we report on DNA nuclease activity of these peptide-conjugates highlighting the various chemical features that are required for efficient DNA degradation by such metal-free DNA cutters.

MATERIALS AND METHODS All commercially available chemical reagents were used without further purification. The solvents DMF, DCM, and MeOH were purchased from Acros (dry under molecular sieves) and used for peptide synthesis without further purification. Phosphoramidites and reagents for DNA synthesis were purchased from Biolab (Jerusalem, Israel) and used as received. Fmoc-6-aminocaproic acid (Fmoc-Ahx) (27), anthraquinone2-cabonyl chloride (AQ-Cl), and acridine-9-carbonyl chloride (28) were synthesized according to the literature. Thin-layer chromatography (TLC) was performed on aluminum sheets Merck silica gel 60 F254. Compounds were visualized by UV absorption at 254 nm and stained with 3% ninhydrin solution in 97% EtOH to monitor the presence of free amines. Semipreparative RP-HPLC separations were performed on a Shimadzu instrument (LC-2010C) by using a Phenomenex C18 column (300A0, 5 µm, 250 mm × 20 mm) with a flow rate of 4 mL/min-1. The eluent was a linear gradient from 95% water (with 0.1% TFA) to 30% acetonitrile over 5 min followed by a

Alkhader et al.

gradient of 30% to 80% acetonitrile over 25 min. The detection was performed at 255 nm. ESI-MS spectra were carried out either on an Orbi-trap mass spectrometer (Thermo Finnigan) using a nanospray attachment or on a ThermoQuest Finnigan LCQ Duo MS. Preparative normal-phase CombiFlash separations were performed on a Mercury instrument. The eluent was a linear gradient from 0% MeOH to 20% MeOH (in DCM). General Procedures for Solid-Phase Peptide Synthesis. Ia. Coupling with HBTU/HOBt (used for Ser/Gly/Glu/Lys Amino Acids and Ahx). Fmoc-amino acid (4 equiv), 2-(1Hbenzotriazole-1-yl)-1,1,3,3-tetramethyluronium hexafluorophosphate (HBTU, 4 equiv), 1-hydroxybenzotriazole (HOBt, 4 equiv), and N,N-diisopropylethylamine (DIEA, 7 equiv) were dissolved in DMF (5 mL) to give a reaction mixture which was added to the 2-chlorotrityl-Gly resin (0.5 g, theoretical loading 0.55 mmol g-1, 0.275 mmol). The reaction mixture was shaken at room temperature (2 × 90 min) and washed with DMF (3 × 5 mL) and DCM (3 × 5 mL). 90% average yield/amino acid based on monitoring Fmoc removal by UV-vis spectroscopy (ε300 ) 7800 cm-1 M-1). Ib. Coupling with DIC/HOBt (used for Arg Amino Acid). Fmoc-amino acid (4 equiv), N,N-diisopropylcarbodiimide (DIC, 4 equiv), HOBt (4 equiv), DIEA (7 equiv), and 4-(dimethylamino)-pyridine (DMAP, 0.1 equiv) were dissolved in DMF to give a clear solution which was added to the resin. The reaction mixture was shaken at room temperature for 90 min, and then a fresh solution (of the above mixture) was added to the resin and shaken overnight. The resin was washed with DMF (3 × 5 mL), DCM (3 × 5 mL). Typically 85% yield. II. Fmoc Deprotection. The Fmoc-protected resin was suspended in a solution of 20% piperidine in DMF (5 mL) for 2 × 10 min. The resin was washed with DMF (3 × 5 mL) and DCM (3 × 5 mL). III. Peptide CleaVage from Resin by TFA Treatment. The resin was washed with DCM (3 × 5 mL) and treated with a solution of 3% TFA/DCM (5 mL) for 2 × 10 min. After removal of the resin by filtration, the filtrates were combined and the product was precipitated by the addition of cold Et2O (25 mL). After filtration, the product was obtained as a white powder. IV. Cyclization of the Linear Pentapeptide. The linear pentapeptide was dissolved in DMF (high dilution, C ) 2 × 10-3 M). After addition of 2-(1H-7-azabenzotriazol-1-yl)-1,3,3tetramethyluranium hexafluorophosphate methanaminium (HATU, 2 equiv) and DIEA (7 equiv), the suspension was stirred for 12 h at room temperature. After the solution was concentrated under vacuum to a minimum volume of DMF, the cyclic peptide was precipitated by the addition of water. V. Pbf and OtBu Deprotection with TFA. The protected peptide was dissolved in TFA/H2O/Et3SiH/phenol (88:5:2:5). After 3 h, Et2O was added to the reaction mixture and a white precipitate was collected by filtration. The final product was HPLC purified using conditions stated above. Synthesis of Compound 1-AQ. Linear-Gly-Arg(Pbf)-Ser(tBu)Arg(Pbf)-D-Lys(Boc). The linear peptide GRSRK was initially synthesized on the solid support as described in the general procedures (sections I-III). 235 mg of crude linear peptide was subsequently cyclized without purification. Cyclic-Gly-Arg(Pbf)-Ser(tBu)-Arg(Pbf)-D-Lys(Boc). The linear peptide was cyclized according to general procedure (section IV). The cyclic peptide was purified by column chromatography (CombiFlash) using conditions described in materials and equipment. Yield: 164 mg (0.13 mmol, 70%) of cyclic peptide. Cyclic-Gly-Arg(Pbf)-Ser(tBu)-Arg(Pbf)-D-Lys. Boc Deprotection with TFA: The protected cyclic peptide was dissolved in 25% TFA/DCM for 30 min at room temperature. After addition of cold Et2O, the white precipitate was collected by filtration.

Metal-Free DNA Nuclease

Cyclic- Gly-Arg(Pbf)-Ser(tBu)-Arg(Pbf)-D-Lys(AQ). Coupling of cyclic peptide with AQ-Cl: Cyclic peptide cGRSRK (0.13 mmol), DIEA (4 eq (0.52 mmol)), and AQ-Cl (0.14 mmol) were dissolved in 10 mL dry DCM under an argon atmosphere. The reaction mixture was stirred at room temperature for 2 h. The reaction progress was monitored by TLC and was terminated by disappearance of the starting material (cyclic peptide). Deprotection of Cyclic-Gly-Arg-Ser-Arg-D-Lys(AQ). Final deprotection of Pbf and tBu protecting groups was carried out as described in general procedures (section V). Total yield of coupling to AQ-Cl and deprotection was 5.4 µmol (4%). MS: [M+1] (Observed) ) 820.01; [M+1] (expected) ) 819.88. Synthesis of Compound Glu-AQ. Linear-Gly-Arg(Pbf)Glu(tBu)-Arg(Pbf)-D-Lys(Boc). The cyclic peptide cRERGK was initially synthesized on the solid support as described in the general procedures (sections I-III). 181 mg of crude linear peptide were subsequently cyclized without purification. Cyclic-Gly-Arg(Pbf)-Glu(tBu)-Arg(Pbf)-D-Lys(Boc). Peptide cyclization: The linear peptide was cyclized according to general procedure (section IV). The cyclic peptide was purified by column chromatography (CombiFlash) using conditions described in materials and equipment. Yield: 126 mg (0.1 mmol, 70%) of cyclic peptide. Cyclic-Gly-Arg(Pbf)-Glu(tBu)-Arg(Pbf)-D-Lys. Boc deprotection with TFA: The protected cyclic peptide was dissolved in 25% TFA/DCM for 30 min at room temperature. After addition of cold Et2O, the white precipitate was collected by filtration. Cyclic-Gly-Arg(Pbf)-Glu(tBu)-Arg(Pbf)-D-Lys(AQ). Coupling of cyclic peptide with AQ-Cl: Cyclic peptide cGRSRK (0.1 mmol), DIEA (4 equiv (0.4 mmol)) and AQ-Cl (0.11 mmol) were dissolved in 10 mL dry DCM under an argon atmosphere. The reaction mixture was stirred at room temperature for 2 h. The reaction progress was monitored by TLC and was terminated by disappearance of the starting material (cyclic peptide). Deprotection of Cyclic-Gly-Arg-Glu-Arg-D-Lys(AQ). Final deprotection of Pbf and tBu protecting groups was carried out as described in general procedures (section V). Total yield of coupling to AQ-Cl and deprotection was 4.5 µmol (5%). MS: [M+1] (observed) ) 861.75; [M+1] (expected) ) 861.92. Synthesis of Compound 1-L-AQ. Cyclic-Gly-Arg-Ser-ArgLys(AQ) was synthesized as described for compound 1-AQ. 290 mg of crude linear peptide was subsequently cyclized without purification. Yield: 223 mg (0.179 mmol, 78%) of cyclic peptide. Total yield of coupling to AQ-Cl and final deprotection was 10 µmol (6%). MS: [M/2] (observed) ) 410.3; [M+1] (expected) ) 819.88. Synthesis of Compound 1-Ac. Cyclic-Gly-Arg-Ser-ArgLys(Ac) was synthesized as described for compound 1-AQ. 210 mg of crude linear peptide were subsequently cyclized without purification. Yield: 170 mg (0.136 mmol, 82%) of cyclic peptide. Total yield of coupling to Ac-Cl and final deprotection was 13.9 µmol (10%). MS: [M/2] (observed) ) 395.93; [M+1] (expected) ) 789.88. Synthesis of compound Gly-AQ. Linear-Gly-Arg(Pbf)-GlyArg(Pbf)-D-Lys(Boc). The linear peptide GRGRK was initially synthesized on the solid support as described in the general procedures (sections I-III). 230 mg of crude linear peptide was subsequently cyclized without purification. Cyclic-Gly-Arg(Pbf)-Gly-Arg(Pbf)-D-Lys(Boc). The linear peptide was cyclized according to general procedure (section IV). The cyclic peptide was purified by column chromatography (CombiFlash) using conditions described in materials and equipment. Yield: 190 mg (0.164 mmol, 84%) of cyclic peptide. Cyclic-Gly-Arg(Pbf)-Gly-Arg(Pbf)-D-Lys. Boc deprotection with TFA: The protected cyclic peptide was dissolved in 25%

Bioconjugate Chem., Vol. 21, No. 8, 2010 1427

TFA/DCM for 30 min at room temperature. After addition of cold Et2O, the white precipitate was collected by filtration. Cyclic-Gly-Arg(Pbf)-Gly-Arg(Pbf)-D-Lys(AQ). Coupling of cyclic peptide with AQ-Cl: Cyclic peptide cGRGRK (0.13 mmol), DIEA (4 eq (0.52 mmol)), and AQ-Cl (0.14 mmol) were dissolved in 10 mL dry DCM under an argon atmosphere. The reaction mixture was stirred at room temperature for 2 h. The reaction progress was monitored by TLC and was terminated by disappearance of the starting material (cyclic peptide). Deprotection of Cyclic-Gly-Arg-Gly-Arg-D-Lys(AQ). Final deprotection of Pbf and tBu protecting groups was carried out as described in general procedures (section V). Total yield of coupling to AQ-Cl and deprotection was 7.8 µmol (5%). MS: [M/2] (observed) ) 395.3; [M+1] (expected) ) 788.85. Synthesis of Compound 1 (Without Intercalator). CyclicGly-Arg-Ser-Arg-Gly. The cyclic control peptide GRSRG was synthesized on the solid support as described in the general procedures (sections I-IV). 205 mg of linear peptide was subsequently cyclized without purification. Cyclic peptide was purified by column chromatography (CombiFlash) using the conditions described in materials and equipment. Total yield of cyclization was 170 mg (0.158 mmol, 84%). Deprotection of Cyclic Peptide. Final deprotection of Pbf and tBu protecting groups was carried out as described in general procedures (section V). MS: [M/2] (observed) ) 257.64; [M+1] (expected) ) 513.55. Synthesis of Compound Lin-1. Linear-Gly-Arg-Ser-ArgAhx-(AQ). The linear peptide GRSR-Ahx was initially synthesized on the solid support (150 mg Rink Amide resin, L ) 0.5 mmol/g) as described in general procedures (sections I-II). Coupling of AQ-Cl to Resin-Bound Linear Peptide. AQ-Cl (4 equiv) and DIEA (7 equiv) dissolved in dry DCM (3 mL) were added to the linear peptide GRSRAhx (on the solid support) and shaken for 90 min. Cleavage from the resin and removal of the protecting groups were carried out as described in general procedures (section V). 6.5 mg of compound Lin-1 were obtained (7.9 µmol, 10% yield). MS: [M/2] (observed) ) 411.2; [M] (expected) ) 820.89. DNA Cleavage Monitored by Agarose Gel Electrophoresis. DNA cleavage experiments were performed using 95 ng DNA (PCR product or plasmid) per reaction. DNA (dissolved in 50 mM Tris-HCl/10 mM CaCl2 buffer, pH ) 7.4) was incubated (in the dark) at 37 °C with the various peptides. After incubation, 1.5 µL of the loading buffer (30 mM EDTA, 0.05% (w/v) glycerol, 36% (v/v) bromophenol blue) was added to each sample followed by loading samples (total volume ) 15 µL) on a1% agarose gel containing 1.0 µg/mL ethidium bromide. Electrophoresis was carried out at 90 V for 1.5 h in 0.5 M TAE buffer. Bands were visualized by UV light and photographed. DNA Cleavage Monitored by Polyacrylamide Gel Electrophoresis. Double-stranded oligonucleotide (50 bp), 5′-end labeled at the top strand, was incubated at 37 °C with 1-AQ in a 50 mM Tris, in 10 mM CaCl2 buffer (pH 7.4), and in the dark. The aliquots were withdrawn at various time intervals, and the reaction was stopped by placing samples on dry ice. DNA cleavage products were resolved by polyacrylamide gel electrophoresis under denaturing conditions (20% PAA/8 M urea). The autoradiograms were visualized by using the bioimaging analyzer BAS-2500 (Fuji Photo Film Co. Ltd., Tokyo, Japan) and Aida image analyzer software (Raytest GmbH, Strauben, Germany). DNA Synthesis. Oligonucleotides (15-mer 5′-CGCGATGACTGTACT and its complementary sequence) were synthesized on an Applied Biosystems 3400 DNA/RNA synthesizer, purified by reverse-phase HPLC (Phenomenex, Clarity 5 µ oligo RP), and characterized by MALDI-TOF MS.

1428 Bioconjugate Chem., Vol. 21, No. 8, 2010

Alkhader et al.

DNA CleaVage as Monitored by Hyperchromicity. A 60 µL quartz cuvette was loaded with 1.8 µM dsDNA (15 mer-5′CGCGATGACTGTACT and its complementary sequence) dissolved in Tris buffer (50 mM Tris, 10 mM CaCl2, pH 7.4). The peptide conjugate (at a given concentration between 1.8 µM to 18 µM) was added, and the adsorption of the sample at 260 nm was monitored by UV-vis spectroscopy every 5 min. Binding Affinity of Inactive Peptide Conjugate (GluAQ) to CT-DNA as Determined by EtBr Displacement. A 1 mL four-sided quartz cuvette was loaded with ethidium bromide (1.3 µM) dissolved in Tris buffer (50 mM Tris, 10 mM NaCl, pH 7.4). The fluorescence was measured (ex. 545 nm, em. 595 nm, RT). Next, CT-DNA was added (254 µM) in aliquots (2 µL), and the increase in fluorescence was measured after the continuous addition of CT-DNA. This procedure was continued until no change in fluorescence was noticed (saturation). Finally, a solution of Glu-AQ (4 µM) was added in aliquots of 2.6 µL, and fluorescence quenching was measured. The binding affinity of Glu-AQ was estimated according to Kapp ) (KEBX[EB])/[Glu-AQ]50 where KEB ) 1 × 107 M-1, [EB] ) 1.3 × 10-6 M, and [GluAQ]50 ) concentration of Glu-AQ where 50% reduction in fluorescence is achieved (29).

RESULTS AND DISCUSSION Structural Considerations. On the basis of the active site found in Staphylococcal nuclease which consists of two L-Arg (Arg35 and Arg87) that electrostatically bind and activate the phosphodiester bonds toward hydrolysis by an activated water as nucleophile, several cyclic pentapeptides were synthesized (Scheme 1) with the following variations: Triads of the type Arg-X-Arg (where X ) L-Ser, Gly, or L-Glu) positioned on the γ-turn (25, 26) as mediators of DNA cleavage were studied. The choice of X ) L-serine (compounds 1-AQ and 1-Ac) stems from the potential of the hydroxyl group to act as a nucleophile; attacking the phosphodiester bond that is more susceptible to nucleophilic attack due to it is electrostatic interactions with two L-Arg. In the case of L-Glu (compound Glu-AQ), the carboxyl side chain of this amino acid is designed to activate a water molecule (as elucidated in the enzyme’s active site) that, in turn, would then attack the susceptible phosphodiester bond. Positioning Gly (compound Gly-AQ) between two L-Arg should determine whether it is necessary to have a nucleophile at position X, as H2O may provide this role. In addition, we have synthesized a cyclic pentapeptide conjugated to AQ (1-L-AQ, Vide post) where L-lysine replaces D-Lys. This change in stereochemistry should dramatically change the cyclic backbone conformation, thus allowing assessment of the importance of the designed cyclic peptides (with the βII′/γ-turn arrangement 25, 26) to DNA nuclease activity. Chemical Synthesis. The synthesis of the cyclic pentapeptides (Scheme 1) was carried out by synthesizing the linear peptide (GRSRK, GRSRK, GRGRK, and GRERK, where K ) D-lysine) on the solid support (2-chlorotrityl Gly resin) followed by C- to N-terminus peptide cyclization in a diluted DMF solution using HATU as a coupling reagent. After cyclization, peptides were purified by column chromatography and the BOC protected ε-amine was removed by treating the cyclic peptides with 25% TFA (in DCM) for 30 min at room temperature. Anthraquinone-2-carbonyl chloride or acridine-9- carbonyl chloride (obtained by reflux of the acid in neat SOCl2) was then added to the peptides and stirred for 2 h in dry DCM in the presence of excess base (DIEA). The crude products were dried in vacuo and redissolved in the deprotection solution (88:2:5:5 TFA/triethylsilane/water/phenol)

Figure 1. (A) Agarose gel (1%) electrophoresis of supercoiled DNA (pSP73 plasmid, superhelical density ) -0.063) incubated for 4 and 16 h with 1-AQ at 37 °C. Lanes: 0, 4, 16 h ) DNA incubated with 1-AQ for 0, 4, and 16 h, respectively; M ) 1 kb DNA ladder. (B) DNA nuclease activity of 1-AQ on a 32P-labeled 50-mer DNA duplex as corroborated on a 20% polyacrylamide/8 M urea gel. Lanes: C+T and G+A ) Maxam-Gilbert sequencing reactions; no peptide ) control, no peptide added; 0, 4, and 16 ) DNA incubated with 1-AQ for 0, 4, and 16 h, respectively. For both gels: Incubation temperature ) 37 °C, [peptide]:[DNA in bp] ) 1:1, final peptide concentration ) 20 µM. All samples were kept in the dark. Incubation buffer: [Tris] ) 50 mM, [CaCl2] ) 10 mM, pH ) 7.4.

for 3 h. All final cyclic compounds were precipitated in cold diethyl ether and HPLC purified on a C18 column. DNA Nuclease Activity of Cyclic Peptides. Chemical nuclease activity was initially evaluated in a phosphate buffer after incubating both 1-AQ and Glu-AQ (20 µM) with plasmid DNA (pSP73) in the dark to avoid UV-induced DNA photocleavage by AQ (27, 28). At 37 °C, no activity of either cyclic pentapeptide was seen. However, at 50 °C, 1-AQ was extremely active (Figures S15-S16, Supporting Information). Taking into account that a high concentration of phosphate ions (phosphate buffer) might interfere with the nuclease activity of the cyclic pentapeptides, we decided instead to incubate the cyclic pentapeptides in a TRIS buffer. In this buffer (50 mM TRIS, 10 mM CaCl2, pH ) 7.4), incubation of 1-AQ with plasmid DNA at 37 °C leads to complete disappearance of the band after overnight incubation (Figure 1A). To verify that this observation is due to DNA cleavage, a 50-mer 32P-radiolabeled duplex was incubated with 1-AQ. As shown in Figure 1B, small DNA fragments are observed already after 4 h of incubation. On the contrary, Glu-AQ is completely inactive (data not shown). The inactivity of Glu-AQ might be a consequence of an electrostatic interaction between the guanidine group(s) of L-Arg (positively charged at pH ) 7.4) and the carboxylate group of L-Glu. Thus, the designed activity for both L-Arg and L-Glu might be hampered by such an electrostatic interaction. As Glu-AQ showed no DNA nuclease activity, it was exploited in order to estimate the apparent binding affinity of this class of peptide conjugates to dsDNA. An EtBr displacement experiment was conducted (see experimental section for details). By following the reduction in EtBr fluorescence (when excluded from CT-DNA by Glu-AQ), we have determined a

Metal-Free DNA Nuclease

Figure 2. Fluorescence decay of EtBr bound to CT-DNA after the addition of Glu-AQ. The estimated binding affinity (Kapp ) 5.4 × 105 M-1) of Glu-AQ to dsDNA was determined according to the amount of Glu-AQ required for a 50% decrease in fluorescence (see text for formula).

Bioconjugate Chem., Vol. 21, No. 8, 2010 1429

Figure 4. DNA nuclease activity of compounds 1-AQ, Gly-AQ, 1-Ac, and 1-L-AQ in [Tris] ) 50 mM, [CaCl2] ) 10 mM, pH ) 7.4 buffer at 37 °C as corroborated on an agarose gel. In lanes 2-12: 95 ng PCR product (591 b.p.). Lane 1, DNA ladder; lane 2, DNA control (without peptides); lanes 3-5, 1-AQ at 4, 16, and 64 µM, respectively; lanes 6-8, Gly-AQ at 4, 16, and 64 µM, respectively; lanes 9-11, 1-Ac at 4, 16, and 64 µM, respectively; lane 12, 16 µM of 1-L-AQ. DNA incubated with all cyclic peptides for 4 h. All samples were kept in the dark.

Scheme 2. Chemical Structures of Control Peptides Lin-1 and 1

Figure 5. DNA nuclease activity of 1-AQ in the presence of ROS scavengers. 95 ng PCR product (591 bp) was incubated with 1-AQ (20 µM) for 4 h at 37 °C. Lane 1, DNA ladder; lane 2, only DNA; lane 3, no inhibitor added; lane 4, 10 mM NaN3; lanes 5-6, 1 mM of DMSO and tBuOH, respectively; and lane 7, 10 mM KI. All samples were kept in the dark during incubation. Incubation buffer: [Tris] ) 50 mM, [CaCl2] ) 10 mM, pH ) 7.4.

Kapp value of 5.4 × 105 M-1 (Figure 2). This value is in the range of the intrinsic binding constants reported for peptidyl anthraquinones (30). Several control peptides were synthesized (Scheme 2). Compound Lin-1 is a linear version of the cyclic peptide where a 6-aminocaproic acid (Ahx) linker is used to separate AQ from the triad (Arg-Ser-Arg). Compound 1 (Scheme 2) is a cyclic pentapeptide that lacks AQ. Both compounds were incubated with a PCR product (591 bp’s) and compared to the DNA nuclease activity of the parent compound 1-AQ (Figure 3). Both compounds Lin-1 and 1 were considerably less active. Thus, it is apparent that both cyclic scaffold and DNA intercalator are required for the superior DNA nuclease activity exerted by compound 1-AQ. In order to gain some insight into the importance of the triad (Arg-Ser-Arg), a cyclic peptide was synthesized where glycine replaces L-Serine. The title compound (cRGRGK conjugated

Figure 3. DNA nuclease activity of compounds 1-AQ, 1, and Lin-1 in [Tris] ) 50 mM, [CaCl2] ) 10 mM, pH ) 7.4 buffer at 37 °C as corroborated on an agarose gel. In all lanes: 95 ng PCR product (591 bp). Lane 1, control (without peptide); lane 2, 4 µM of 1-AQ; lane 3, 16 µM of 1-AQ; lane 4, 16 µM of 1; lane 5, 48 µM of 1; lane 6, 144 µM of 1; lane 7, 16 µM of Lin-1; lane 8, 48 µM of Lin-1; lane 9, 144 µM of Lin-1. DNA incubated with all peptides for 4 h. All samples were kept in the dark.

to AQ, Gly-AQ) has two L-arginine amino acids for phosphodiester binding and activation; however, it lacks a nucleophile. DNA nuclease activity of such a conjugate would suggest that a water molecule acts as a nucleophile. Indeed, we observe that Gly-AQ exerts similar DNA nuclease activity to that of 1-AQ (Figure 4, lanes 6-8). The high DNA nuclease activity observed for these conjugates also coincides with water as a nucleophile. In such a scenario, phosphodiester cleavage would regenerate the chemical nuclease providing catalytic activity to such conjugates. DNA nuclease activity of both cyclic peptides with the ArgSer-Arg triad but with different DNA intercalators (AQ and Ac) were compared (Figure 4, compounds 1-AQ and 1-Ac, lanes 9-11). No significant difference in activity was observed, highlighting the versatility of such cyclic peptide-DNA intercalator conjugates as metal-free chemical nucleases. L-Lysine was introduced into the cyclic pentapeptide (cRSRGK, 1-L-AQ) replacing D-lysine. This is expected to bring about a dramatic change in the conformation of the cyclic peptide’s backbone. As shown in Figure 4 (lane 12), a concentration of 16 µM of 1-L-AQ results only in modest DNA nuclease activity (as corroborated by the weakened DNA band) in comparison to complete disappearance of the DNA band obtained with 1-AQ, 1-Ac, and Gly-AQ at the same concentration. Mechanism and Potency of DNA Cleavage. To gain evidence that supports a hydrolytic mechanism of cleavage by the active cyclic pentapeptides, ROS scavengers were added to incubated samples (1-AQ + DNA) in order to exclude an oxidative mechanism for DNA cleavage. As shown in Figure 5, the addition of a singlet oxygen quencher (NaN3), hydroxyl radical scavengers (DMSO or tBuOH), and a superoxide scavenger (KI) had no effect on 1-AQ’s DNA nuclease activity. In all cases, the plasmid is cleaved as observed for the control sample (lane 3) without inhibitor.

1430 Bioconjugate Chem., Vol. 21, No. 8, 2010

Alkhader et al.

Figure 6. (A) Hyperchromicity of 15-mer dsDNA substrate after the addition of peptide conjugates 1-AQ, Gly-AQ, and 1-Ac at 2.5 molar excess concentrations. (B) Estimated kobs values of 0.70 h-1, 0.85 h-1, and 0.65 h-1 for peptide conjugates 1-AQ, Gly-AQ, and 1-Ac, respectively. kobs values were extracted form the linear fit (slope ) kobs) for all three peptides conjugates.

The fact that DNA cleavage of plasmid DNA or duplex DNA (591 bp) by 1-AQ (and other active cyclic peptides) yielded either DNA smears or complete disappearance of the parent band precluded the straightforward analysis of cleavage kinetics (i.e., nicking (form II-coiled) and double strand breaks (form IIIlinear)). Thus, we decided to monitor DNA cleavage of the active peptide conjugates by following DNA degradation by monitoring the change in absorbance at 260 nm (hyperchromicity). A synthetic 15-mer DNA duplex was chosen as a DNA substrate. The various peptide conjugates were added to this 15-mer duplex DNA substrate, and the change in adsorption was followed at 260 nm. Degradation of dsDNA to ssDNA and to further single bases is expected to lead to an increase in absorption (hyperchromicity) (31, 32). As we have observed complete degradation of the radiolabeled 50-mer duplex (Figure 1B), we assume that the change in adsorption on a longer time scale in which the change in O.D. is negligible (e.g., 2 h) represents a good approximation for complete degradation of the DNA substrate. Accordingly, we have estimated the amount of intact DNA as follows: % intact DNA ) 100 - (O.D.t - O.D.0 /O.D.∞) where O.D.t ) measured adsorption, O.D.0 ) measured adsorption at t ) 0, and O.D.∞ ) measured adsorption at the end of the reaction (plateau region). By plotting the logarithm of this value as a function of time (within the first hour of kinetics), we obtain a linear fit where the slope is an estimation for kobs. Figure 6 presents the hyperchromicity data (change in O.D.) related to the initial DNA nuclease activity (within 2 h) and the linear plots of the three active peptide conjugates, namely, 1-AQ, 1-Ac, and Gly-AQ at 2.5 equiv peptide conjugate to dsDNA substrate. On the basis of the kobs values that range 0.65-0.85 h-1, all three peptide conjugates seem to be similarly active. The peptide conjugate (Gly-AQ) that has no bulky group situated between the two L-Arg in its triad (Gly instead of L-Ser) seems to be the most active of all three compounds. We speculate that the introduction of L-Ser at this position diminishes DNA nuclease activity perhaps due to steric hindrance. However, further peptide analogues are required in order to verify this argument. The small differences in kobs between 1-AQ and 1-Ac are attributed to the DNA intercalator (anthraquinone vs acridine), as this is the only difference in structure between these peptide conjugates. We have also repeated these experiments with one of the active compounds (1-AQ) at various peptide-conjugate concentrations. Table 1 lists the kobs values determined for this peptide conjugate (see Supporting Information Figure S17 for kinetic profiles of 1-AQ at various concentrations). Interestingly, we find that at a high peptide conjugate concentration (10 equiv) the kobs value is reduced. Thus, an optimal value is obtained at

Table 1. Kinetics of DNA Cleavage by 1-AQ as Determined by Hyperchromicity at 260 nm fold of 1-AQ

kobs (h-1)

1 2.5 5 7.5 10

0.26 0.70 1.13 1.62 0.53

7.5 equiv of 1-AQ. It is possible that, at higher concentrations, compound 1-AQ forms aggregates in buffer by π-stacking of the appended anthraquinone moieties on each D-Lys of the cyclic peptide. At 7.5 equiv of 1-AQ, a kobs of 1.62 h-1 is found based on the hyperchromicity measurements. To the best of our knowledge, this value is by an order of magnitude higher than the most efficient metal-free DNA nucleases reported to date (at physiological conditions). In summary, we have synthesized a family of metal-free DNA nucleases that exert high DNA nuclease activity at physiological conditions and at micromolar concentrations. It was found that both amino acid composition as well as conjugation of the cyclic pentapeptide to DNA binders (intercalators) are essential features required for such activity. As DNA intercalators such as anthraquinone (33) have been shown to exert some sequence specificity when bound to short peptides, it is possible to exploit this discrimination for selective DNA cleavage. Alternatively, the conjugation of a triplex-forming oligonucleotides (TFO) in place of a DNA intercalator now has the potential to bring about sequence-specific DNA cleavage by such cyclic peptide-TFO conjugates; a goal that we are currently pursuing.

ACKNOWLEDGMENT We thank the authority for R&D of the Hebrew U. and the Alex Grass Center for Drug Design for financial support. This research was also supported by the Academy of Sciences of the CR (IAA400040803), the Ministry of Education of the CR (MSMT LC06030, ME08017, OC08003), the Academy of Sciences of the CR (Grants KAN200200651, AV0Z50040507). J.K. is the international research scholar of the Howard Hughes Medical Institute. Supporting Information Available: Gel electrophoresis of compounds 1-AQ and Glu-AQ in phosphate buffer, MS and HPLC chromatograms of compounds 1-AQ, 1-L-AQ, 1-Ac, Gly-AQ, Glu-AQ, 1, and Lin-1, and kinetics plots of 1-AQ at various concentrations. This material is available free of charge via the Internet at http://pubs.acs.org.

Metal-Free DNA Nuclease

LITERATURE CITED (1) Ma, Y., Chen, X., Sun, M., Wan, R., Zhu, C., Li, Y., and Zhao, Y. (2008) DNA cleavage function of seryl-histidine dipeptide and its application. Amino Acids 35, 251–256. (2) Razkin, J., Lindgren, J., Nilsson, H., and Baltzer, L. (2008) Enhanced complexity and catalytic efficiency in the hydrolysis of phosphate diesters by rationally designed helix-loop-helix motifs. ChemBioChem 9, 1975–1984. (3) Cheng, C. T., Lo, V., Chen, J., Chen, W. C., Lin, C. Y., Lin, H. C., Yang, C. H., and Sheh, L. (2001) Synthesis and DNA nicking studies of a novel cyclic peptide: Cyclo[Lys-Trp-LysAhx-]. Bioorg. Med. Chem. 9, 1493–1498. (4) Feng, Y. P., Cao, S. L., Xiao, A. S., Xie, W. J., Li, Y. M., and Zhao, Y. F. (2006) Studies on cleavage of DNA by N-phosphoryl branched peptides. Peptides 27, 1554–1560. (5) Li, Y. S., Zhao, Y. F., Hatfield, S., Wan, R., Zhu, Q., Li, X. H., McMills, M., Ma, Y., Li, J., Brown, K. L., He, C., Liu, F., and Chen, X. Z. (2000) Dipeptide seryl-histidine and related oligopeptides cleave DNA, protein, and a carboxyl ester. Bioorg. Med. Chem. 8, 2675–2680. (6) Schmuck, C., and Dudaczek, J. (2007) Screening of a combinatorial library reveals peptide-based catalysts for phosphorester cleavage in water. Org. Lett. 9, 5389–5392. (7) Sheng, X., Lu, X. M., Chen, Y. T., Lu, G. Y., Zhang, J. J., Shao, Y., Liu, F., and Xu, Q. (2007) Synthesis, DNA-Binding, cleavage, and cytotoxic activity of new 1,7-dioxa-4,10-diazacyclododecane artificial receptors containing bisguanidinoethyl or diaminoethyl double side arms. Chem.sEur. J. 13, 9703–9712. (8) Sheng, X., Lu, X. M., Zhang, J. J., Chen, Y. T., Lu, G. Y., Shao, Y., Liu, F., and Xu, Q. (2007) Synthesis and DNA cleavage activity of artificial receptor 1,4,7-triazacyclononane containing guanidinoethyl and hydroxyethyl side arms. J. Org. Chem. 72, 1799–1802. (9) Wan, S. H., Liang, F., Xiong, X. Q., Yang, L., Wu, X. J., Wang, P., Zhou, X., and Wu, C. T. (2006) DNA hydrolysis promoted by 1,7-dimethyl-1,4,7,10-tetraazacyclododecane. Bioorg. Med. Chem. Lett. 16, 2804–2806. (10) Zhao, Y. C., Zhang, J., Huang, Y., Wang, G. Q., and Yu, X. Q. (2007) DNA cleavage promoted by 2,9-dimethyl-4,7diazadecane-2,9-dithiol (DDD) derivatives. Bioorg. Med. Chem. Lett. 17, 2745–2748. (11) Shao, Y., Sheng, X., Li, Y., Jia, Z.-L., Zhang, J.-J., Liu, F., and Lu, G. Y. (2008) DNA binding and cleaving activity of the new cleft molecule N,N′-bis(guanidinoethyl)-2,6-pyridinedicarboxamide in the absence or in the presence of copper(II). Bioconjugate Chem. 19, 1840–1848. (12) Cotton, F. A., Hazen, E. E., and Legg, M. J. (1979) Staphylococcal nuclease - proposed mechanism of action based on structure of enzyme-thymidine 3′,5′-bisphosphate-calcium ion complex at 1.5-Å resolution. Proc. Natl. Acad. Sci. U.S.A. 76, 2551–2555. (13) Pei, D. H., Corey, D. R., and Schultz, P. G. (1990) Site-specific cleavage of duplex DNA by a semisynthetic nuclease via triplehelix formation. Proc. Natl. Acad. Sci. U.S.A. 87, 9858–9862. (14) Zuckermann, R. N., and Schultz, P. G. (1989) Site-selective cleavage of structured RNA by a staphylococcal nuclease DNA hybrid. Proc. Natl. Acad. Sci. U.S.A. 86, 1766–1770. (15) Cole, H. B. R., Sparks, S. W., and Torchia, D. A. (1988) Comparison of the solution and crystal-structures of staphylococcal nuclease with C-13 and N-15 chemical-shifts used as structural fingerprints. Proc. Natl. Acad. Sci. U.S.A. 85, 6362– 6365. (16) Deiters, J. A., Gallucci, J. C., and Holmes, R. R. (1982) Pentacoordinated molecules 0.41. computer-simulation of staphylococcal nuclease action on thymidine 3′,5′-bis(phosphate) (Pdtp). J. Am. Chem. Soc. 104, 5457–5465. (17) Leung, K. W., Liaw, Y. C., Chan, S. C., Lo, H. Y., Musayev, F. N., Chen, J. Z. W., Fang, H. J., and Chen, H. M. (2001) Significance of local electrostatic interactions in staphylococcal

Bioconjugate Chem., Vol. 21, No. 8, 2010 1431 nuclease studied by site-directed mutagenesis. J. Biol. Chem. 276, 46039–46045. (18) Libson, A. M., Gittis, A. G., and Lattman, E. E. (1994) Crystalstructures of the binary Ca2+ and Pdtp complexes and the tertiary complex of the Asp(21) -] Glu mutant of staphylococcal nuclease - implications for catalysis and ligand-binding. Biochemistry 33, 8007–8016. (19) Serpersu, E. H., Shortle, D., and Mildvan, A. S. (1986) Kinetic and magnetic-resonance studies of effects of genetic substitution of a Ca-2+-liganding amino-acid in staphylococcal nuclease. Biochemistry 25, 68–77. (20) Shan, L., Tong, Y. F., Xie, T., Wang, M., and Wang, J. F. (2007) Restricted backbone conformational and motional flexibilities of loops containing peptidyl-proline bonds dominate the enzyme activity of staphylococcal nuclease. Biochemistry 46, 11504–11513. (21) Ting, A. Y., Shin, I., Lucero, C., and Schultz, P. G. (1998) Energetic analysis of an engineered cation-pi interaction in staphylococcal nuclease. J. Am. Chem. Soc. 120, 7135–7136. (22) Wall, M. E., Ealick, S. E., and Gruner, S. M. (1997) Threedimensional diffuse x-ray scattering from crystals of Staphylococcal nuclease. Proc. Natl. Acad. Sci. U.S.A. 94, 6180–6184. (23) Wang, J. F., Hinck, A. P., Loh, S. N., Lemaster, D. M., and Markley, J. L. (1992) Solution studies of staphylococcal nuclease H124l 0.2. H-1, C-13, and N-15 chemical-shift assignments for the unligated enzyme and analysis of chemical-shift changes that accompany formation of the nuclease thymidine 3′,5′-bisphosphate calcium ternary complex. Biochemistry 31, 921–936. (24) Weber, D. J., Libson, A. M., Gittis, A. G., Lebowitz, M. S., and Mildvan, A. S. (1994) NMR docking of a substrate into the X-Ray structure of the Asp-21 -] Glu mutant of staphylococcal nuclease. Biochemistry 33, 8017–8028. (25) Haubner, R., Gratias, R., Diefenbach, B., Goodman, S. L., Jonczyk, A., and Kessler, H. (1996) Structural and functional aspects of RGD-containing cyclic pentapeptides as highly potent and selective integrin alpha(v)beta(3) antagonists. J. Am. Chem. Soc. 118, 7461–7472. (26) Kessler, H., Gratias, R., Hessler, G., Gurrath, M., and Muller, G. (1996) Conformation of cyclic peptides. Principle concepts and the design of selectivity and superactivity in bioactive sequences by ‘spatial screening’. Pure Appl. Chem. 68, 1201– 1205. (27) Khandazhinskaya, A. L., Kukhanova, M. K., and Jasko, M. V. (2005) New nonnucleoside substrates for terminal deoxynucleotidyl transferase: Synthesis and dependence of substrate properties on structure. Russ. J. Bioorg. Chem. 31, 352–356. (28) Tierney, M. T., and Grinstaff, M. W. (2000) Synthesis and characterization of fluorenone-, anthraquinone-, and phenothiazine-labeled oligodeoxynucleotides: 5 ′-probes for DNA redox chemistry. J. Org. Chem. 65, 5355–5359. (29) Begum, M. S. A., Saha, S., Nethaji, M., and Chakravarty, A. R. (2010) Iron(III) schiff base complexes of arginine and lysine as netropsin mimics showing AT-selective DNA binding and photonuclease activity. J. Inorg. Biochem. 104, 477–484. (30) Gatto, B., Zagotto, G., Sissi, C., Cera, C., Uriarte, E., Palu`, G., Capranico, G., and Palumbo, M. (1996) Peptidyl anthraquinones as potential antineoplastic drugs: Synthesis, DNA binding, redox cycling and biological activity. J. Med. Chem. 39, 3114–3122. (31) Tataurov, A. V., You, Y., and Owczarzy, R. (2008) Predicting ultraviolet spectrum of single stranded and double stranded deoxyribonucleic acids. Biophys. Chem. 133, 66–70. (32) See website: http://www.owczarzy.net/extinct.htm. (33) Gatto, B., Zagotto, G., Sissi, C., and Pulambo, M. (1997) Preferred interaction of D peptidyl-anthraquinones with doublestranded B-DNA. Int. J. Biol. Macromol. 21, 319–326. BC900543B