Non-Covalent Binding of Tripeptides-Containing Tryptophan to Polynucleotides and Photochemical Deamination of Modified Tyrosine to Quinone Methide Leading to Covalent Attachment

A series of tripeptides TrpTrpPhe (1), TrpTrpTyr (2), and TrpTrpTyr[CH2N(CH3)2] (3) were synthesized, and their photophysical properties and non-covalent binding to polynucleotides were investigated. Fluorescent Trp residues (quantum yield in aqueous solvent ΦF = 0.03–0.06), allowed for the fluorometric study of non-covalent binding to DNA and RNA. Moreover, high and similar affinities of 2×HCl and 3×HCl to all studied double stranded (ds)-polynucleotides were found (logKa = 6.0–6.8). However, the fluorescence spectral responses were strongly dependent on base pair composition: the GC-containing polynucleotides efficiently quenched Trp emission, at variance to AT- or AU-polynucleotides, which induced bisignate response. Namely, addition of AT(U) polynucleotides at excess over studied peptide induced the quenching (attributed to aggregation in the grooves of polynucleotides), whereas at excess of DNA/RNA over peptide the fluorescence increase of Trp was observed. The thermal denaturation and circular dichroism (CD) experiments supported peptides binding within the grooves of polynucleotides. The photogenerated quinone methide (QM) reacts with nucleophiles giving adducts, as demonstrated by the photomethanolysis (quantum yield ΦR = 0.11–0.13). Furthermore, we have demonstrated photoalkylation of AT oligonucleotides by QM, at variance to previous reports describing the highest reactivity of QMs with the GC reach regions of polynucleotides. Our investigations show a proof of principle that QM precursor can be imbedded into a peptide and used as a photochemical switch to enable alkylation of polynucleotides, enabling further applications in chemistry and biology.


Introduction
Non-covalent binding of peptides/proteins and DNA or RNA is an important biological process enabling transcription of DNA and cell replication [1,2]. Understanding the recognition process between polynucleotides and peptides and being able to affect these events [3] provide immense scientific advantage, enabling numerous applications in biochemistry, biology, and medicine. Therefore, binding of peptides to polynucleotides has been intensively investigated [4,5]. Examples of DNA-binding peptides peptides/proteins are numerous, for instance zinc finger [6], or leucine zipper motives [7], as well as many other, even small peptides [4]. Furthermore, a special attention was devoted to synthetic peptides that selectively recognize DNA [8,9] or DNA-binding oligopeptides which can be modulated by light [10]. Peptide-binding units can easily be functionalized by fluorophores [11,12], providing facile methods for DNA fluorescent labeling and visualization [13].

Synthesis
The synthetic protocol for the preparation of tripeptides 1-3 was based on the stand ard Boc-chemistry in solution where the activation of carboxylic acids was achieved b the use of N-hydroxysuccinimide (NHS) and 1-ethyl-3-(3-dimethylaminopropyl)car bodiimide (EDC) [55]. In the first step, N-Boc-L-Trp-OH was activated by EDC and trans formed to the succinimide ester N-Boc-L-Trp-OSu, which was isolated (Scheme 1). Th following coupling with H-L-Trp-OH afforded dipeptide N-Boc-L-Trp-L-Trp-OH in exce lent yield. The free carboxylic functional group of the dipeptide was activated again b transforming it to a succinimide ester, which was coupled with Phe, Tyr, or modified Ty prepared according to the literature precedent [52]. The coupling afforded desired pep tides 1-3 in moderate yields. For the study with polynucleotides removal of the Boc wa facilitated by the treatment with HCl dissolved in dry EtOAc, whereupon HCl salts wer obtained.

Synthesis
The synthetic protocol for the preparation of tripeptides 1-3 was based on the standard Boc-chemistry in solution where the activation of carboxylic acids was achieved by the use of N-hydroxysuccinimide (NHS) and 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDC) [55]. In the first step, N-Boc-L-Trp-OH was activated by EDC and transformed to the succinimide ester N-Boc-L-Trp-OSu, which was isolated (Scheme 1). The following coupling with H-L-Trp-OH afforded dipeptide N-Boc-L-Trp-L-Trp-OH in excellent yield. The free carboxylic functional group of the dipeptide was activated again by transforming it to a succinimide ester, which was coupled with Phe, Tyr, or modified Tyr, prepared according to the literature precedent [52]. The coupling afforded desired peptides 1-3 in moderate yields. For the study with polynucleotides removal of the Boc was facilitated by the treatment with HCl dissolved in dry EtOAc, whereupon HCl salts were obtained.

Synthesis
The synthetic protocol for the preparation of tripeptides 1-3 was based on the standard Boc-chemistry in solution where the activation of carboxylic acids was achieved by the use of N-hydroxysuccinimide (NHS) and 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDC) [55]. In the first step, N-Boc-L-Trp-OH was activated by EDC and transformed to the succinimide ester N-Boc-L-Trp-OSu, which was isolated (Scheme 1). The following coupling with H-L-Trp-OH afforded dipeptide N-Boc-L-Trp-L-Trp-OH in excellent yield. The free carboxylic functional group of the dipeptide was activated again by transforming it to a succinimide ester, which was coupled with Phe, Tyr, or modified Tyr, prepared according to the literature precedent [52]. The coupling afforded desired peptides 1-3 in moderate yields. For the study with polynucleotides removal of the Boc was facilitated by the treatment with HCl dissolved in dry EtOAc, whereupon HCl salts were obtained. Scheme 1. Synthesis of 1-3.

Photophysical Properties
Photophysical properties of 1-3, and the corresponding HCl salts, are important for the study of their binding to polynucleotides, as well as for deeper understanding of the photochemical reactivity of 3. Absorption and fluorescence spectra for 1-3 were measured in CH 3 CN (Figures S1-S8 in the SI), whereas assays with biomacromolecules for the corresponding salts were conducted in aqueous cacodylate buffer (pH = 7, 50 mM), to which some DMSO was added to assure solubility. Absorption spectra of 1-3 are shown in Figure 2. In the low energy region, the absorption band with a maximum at 280 nm and a shoulder at 290 nm dominate for all tripeptides, the typical for tryptophan. The third amino acid, Phe, Tyr, or Tyr[CH 2 N(CH 3 ) 2 ], has only a minor influence on the absorption spectra, with 1 and 2 showing higher absorptivity than H-Trp-Trp-OH, and 3 lower (Figures S2, S5 and S7 in the SI). The finding suggests that there is some electronic interaction between the chromophores in the ground state. Due to large aromatic residues, it is plausible that tripeptides 1-3 form α-helix where the first and the third amino acid are positioned in the proximity to interact, leading to the hyperchromic effect with 1 and 2, and hypochromic effect with 3.

Photophysical Properties
Photophysical properties of 1-3, and the corresponding HCl salts, are important for the study of their binding to polynucleotides, as well as for deeper understanding of the photochemical reactivity of 3. Absorption and fluorescence spectra for 1-3 were measured in CH3CN (Figures S1-S8 in the ESI), whereas assays with biomacromolecules for the corresponding salts were conducted in aqueous cacodylate buffer (pH = 7, 50 mM), to which some DMSO was added to assure solubility. Absorption spectra of 1-3 are shown in Figure 2. In the low energy region, the absorption band with a maximum at 280 nm and a shoulder at 290 nm dominate for all tripeptides, the typical for tryptophan. The third amino acid, Phe, Tyr, or Tyr[CH2N(CH3)2], has only a minor influence on the absorption spectra, with 1 and 2 showing higher absorptivity than H-Trp-Trp-OH, and 3 lower (Figures S2, S5 and S7 in the ESI). The finding suggests that there is some electronic interaction between the chromophores in the ground state. Due to large aromatic residues, it is plausible that tripeptides 1-3 form α-helix where the first and the third amino acid are positioned in the proximity to interact, leading to the hyperchromic effect with 1 and 2, and hypochromic effect with 3.  Fluorescence spectra of 1-3 in CH3CN are characterized by one band with a maximum at 358 nm for 2 and 3, and 366 nm for 1, the typical for Trp. The position of the maxima and the shape of the spectra do not depend on the excitation wavelength, indicating that the emission originates from Trp only. However, fluorescence spectra of 1 in cacodylate buffer at higher concentration (2 mM) exhibit bathochromic shifts of 10 nm ( Figure S3 in the ESI), suggesting aggregation of the molecules. Such an aggregation was not observed for 2 and 3. Fluorescence spectra of 3 exhibit solvatochromic properties (Figure S8, left, in the ESI), which are in accordance with reports on Trp photophysics [56]. Thus, 3 in CH3CN and 3×HCl in CH3CN-H2O (1:9), show difference of 23 nm between the maxima of emission spectra, which is due to different spectral properties of Trp derivatives in solvents of different polarity and proticity [57]. Understanding spectral and photophysical properties, which are dependent on solvent polarity and proticity, is important for the study of non-covalent binding of these peptides to macrobiomolecules including polynucleotides.
Quantum yields of fluorescence (Φf) were measured for 1-3 in CH3CN, and for the corresponding salts Φf was measured in aqueous solution CH3CN-H2O (1:9), containing sodium cacodylate buffer (pH = 7.0, 50 mM), (Table 1). N-acetyltryptophanamide (NATA) in H2O was used as a reference (Φf = 0.14) [58]. Note that Φf for multichromophoric systems 1-3 depends on the excitation wavelength, and we report the average value. However, for Fluorescence spectra of 1-3 in CH 3 CN are characterized by one band with a maximum at 358 nm for 2 and 3, and 366 nm for 1, the typical for Trp. The position of the maxima and the shape of the spectra do not depend on the excitation wavelength, indicating that the emission originates from Trp only. However, fluorescence spectra of 1 in cacodylate buffer at higher concentration (2 mM) exhibit bathochromic shifts of 10 nm ( Figure S3 in the SI), suggesting aggregation of the molecules. Such an aggregation was not observed for 2 and 3. Fluorescence spectra of 3 exhibit solvatochromic properties ( Figure S8, left, in the SI), which are in accordance with reports on Trp photophysics [56]. Thus, 3 in CH 3 CN and 3×HCl in CH 3 CN-H 2 O (1:9), show difference of 23 nm between the maxima of emission spectra, which is due to different spectral properties of Trp derivatives in solvents of different polarity and proticity [57]. Understanding spectral and photophysical properties, which are dependent on solvent polarity and proticity, is important for the study of non-covalent binding of these peptides to macrobiomolecules including polynucleotides.
Quantum yields of fluorescence (Φ f ) were measured for 1-3 in CH 3 CN, and for the corresponding salts Φ f was measured in aqueous solution CH 3 CN-H 2 O (1:9), containing sodium cacodylate buffer (pH = 7.0, 50 mM), (Table 1). N-acetyltryptophanamide (NATA) in H 2 O was used as a reference (Φ f = 0.14) [58]. Note that Φ f for multichromophoric systems 1-3 depends on the excitation wavelength, and we report the average value. However, for aqueous and non-aqueous solution, Φ f is similar for 1 and 2, and about double than Φ f for 3, presumably due to photochemical deactivation delivering QMs from 3. Decays of fluorescence for 1-3 in CH 3 CN, as well as for 2×HCl and 3×HCl in CH 3 CN-H 2 O (1:9) were measured by time-correlated single-photon counting (TC-SPC) upon excitation at 280 nm (Table 1 and Figures S9 and S10 in the SI). None of the fluorescence decays could be fit to single exponential function, which is anticipated for multichromophoric systems. Furthermore, it is known that Trp derivatives often do not exhibit single exponential decay due to the presence of two excited states, L a and L b [59].

Photochemistry
Photodeamination of aminomethylphenoles in aqueous CH 3 OH leads to photomethan olysis products via QM intermediates [52,53]. To demonstrate that photometahnolysis of tripeptide 3 also takes place, we performed preparative irradiation of the fully protected compound (bearing Boc and Bn) to make the photoproduct isolation easier. The irradiations were conducted in CH 3 OH solutions at 300 nm and the composition of the solutions was analyzed by HPLC. Methanolysis of 3 run to the conversion of 65% gave methyl ether 3-OMe, which was isolated in 22% yield (Scheme 2). In addition, photomethanolysis of salt 3×HCl was conducted, but it was analyzed by HPLC only. It gave cleanly only one photoproduct, presumably the photomethanolysis ether. aqueous and non-aqueous solution, Φf is similar for 1 and 2, and about double than Φf for 3, presumably due to photochemical deactivation delivering QMs from 3. 2.9 ± 0.1 (29%) 1 Quantum yield of fluorescence was measured using NATA in mQ-H2O (Φf = 0.14) [58]. The fluorescence spectra were measured by exciting at 280, 290, and 300 nm, and the average values of the quantum yields (Equation (S1) in the ESI) were calculated. The associated errors correspond to the maximum absolute deviation. 2 Decay times were measured by TC-SPC. Fluorescence decays were obtained by exciting samples at 280 nm and detecting fluorescence at 370 nm. Relative contribution of decay components is given in parenthesis. The associated errors correspond to maximal standard deviations obtained from the fitting. 3 CH3CN-H2O (1:9), containing sodium cacodylate buffer (pH = 7.0, 50 mM).
Decays of fluorescence for 1-3 in CH3CN, as well as for 2×HCl and 3×HCl in CH3CN-H2O (1:9) were measured by time-correlated single-photon counting (TC-SPC) upon excitation at 280 nm (Table 1 and Figures S9 and S10 in the ESI). None of the fluorescence decays could be fit to single exponential function, which is anticipated for multichromophoric systems. Furthermore, it is known that Trp derivatives often do not exhibit single exponential decay due to the presence of two excited states, La and Lb [59].

Photochemistry
Photodeamination of aminomethylphenoles in aqueous CH3OH leads to photomethanolysis products via QM intermediates [52,53]. To demonstrate that photometahnolysis of tripeptide 3 also takes place, we performed preparative irradiation of the fully protected compound (bearing Boc and Bn) to make the photoproduct isolation easier. The irradiations were conducted in CH3OH solutions at 300 nm and the composition of the solutions was analyzed by HPLC. Methanolysis of 3 run to the conversion of 65% gave methyl ether 3-OMe, which was isolated in 22% yield (Scheme 2). In addition, photomethanolysis of salt 3×HCl was conducted, but it was analyzed by HPLC only. It gave cleanly only one photoproduct, presumably the photomethanolysis ether.

Scheme 2. Photomethanolysis of 3.
Efficiency of the photomethanolysis reaction (ΦR) was measured by the use of a primary actinometer, KI/KIO3 (Φ254 = 0.74) [58,60], and the samples were excited at 254 nm. The efficiency was measured for 3 and 3×HCl, providing similar values of ΦR = 0.13 ± 0.01 and ΦR = 0.11 ± 0.02, respectively. Relative efficient formation of the photomethanolysis product is highly indicative that deamination takes place via a QM intermediate, as reported in the literature precedent [52,53].

Scheme 2. Photomethanolysis of 3.
Efficiency of the photomethanolysis reaction (Φ R ) was measured by the use of a primary actinometer, KI/KIO 3 (Φ 254 = 0.74) [58,60], and the samples were excited at 254 nm. The efficiency was measured for 3 and 3×HCl, providing similar values of Φ R = 0.13 ± 0.01 and Φ R = 0.11 ± 0.02, respectively. Relative efficient formation of the photomethanolysis product is highly indicative that deamination takes place via a QM intermediate, as reported in the literature precedent [52,53].

Laser Flash Photolysis (LFP)
To detect QM intermediates and other plausible intermediates in the photochemistry of tripeptide 3, LFP was used. The samples were excited with a Nd:YAG laser at 266 nm. The measurements were performed in N 2 and O 2 -purged CH 3 CN solution, where O 2 was expected to quench triplets and radicals, but not QMs. Moreover, the spectra and decay kinetics were measured in CH 3 CN and CH 3 CN-H 2 O (1:1) where the difference was expected due to excited state proton transfer (ESPT) pathways, possible in aqueous solvent only [61][62][63][64] (for all transient absorption spectra and the associated decay kinetics see Figures S11-S37).
In Ar-purged CH 3 CN solution of 3 we detected two transients absorbing over the whole spectrum with a maximum at ≈350 nm. The lifetimes of the transients were 200 ± 50 ns, probably corresponding to the triplet excited state or some radical (assignment based on the quenching by O 2 ), and 450 ± 100 µs tentatively assigned to QM-3 since in the O 2 -purged solution it has a similar lifetime, ≈400 µs. The spectra show additional transient absorbing at ≈350 nm and ≈550 nm that is not quenched by O 2 . It may correspond to species generated from Trp, such as indolyl radical-cation and N-radical [65][66][67].
In the aqueous Ar-purged solution of 3 two transients were detected with the lifetimes of 10-30 ns, probably corresponding to the triplet excited state or some radical and the one with the lifetime 190 ± 10 µs, tentatively assigned to the QM since it is not quenched by O 2 .
In the LFP experiments for the salt 3×HCl, in Ar-purged solution we detected three transients with the lifetimes of ≈10 ns, probably corresponding to the triplet excited state or some radical, 10.6 ± 0.6 µs, not assigned and 1.3 ± 0.6 ms, tentatively assigned to the QM. In O 2 -purged solution we detected a similar long-lived transient with the lifetime, ≈0.5-10 ms whose precise determination of decay kinetics was difficult due to very low intensity and long decay time.
Since the very low intensity of the transient (∆A < 10 −3 ) that was tentatively assigned to QM-precluded quenching experiments, we performed control LFP experiments for 2, where the formation of QM is not possible. Figure 3 shows transient absorption spectra of 2 and 3 recorded after a delay of 27 µs after the laser pulse, revealing some difference at 300-400 nm, which may be assigned to the presence of QM based on the position of the anticipated maximum in the absorption spectra [52,53]. This long-lived transient with the lifetime of ≈400 µs was seen from 3 (but not from 2), we therefore assigned to QM-3. Interestingly, in the aqueous solution, a difference between the transient absorption spectra between 2 and 3 was not observed, presumably due to pronounced formation of N-Trp radicals from both peptides, in the process that involves photoionization to radical cation and its deprotonation [65][66][67].
To detect QM intermediates and other plausible intermediates in the photochemistry of tripeptide 3, LFP was used. The samples were excited with a Nd:YAG laser at 266 nm The measurements were performed in N2 and O2-purged CH3CN solution, where O2 was expected to quench triplets and radicals, but not QMs. Moreover, the spectra and decay kinetics were measured in CH3CN and CH3CN-H2O (1:1) where the difference was ex pected due to excited state proton transfer (ESPT) pathways, possible in aqueous solven only [61][62][63][64] (for all transient absorption spectra and the associated decay kinetics see Fig  ures S11-S37).
In Ar-purged CH3CN solution of 3 we detected two transients absorbing over the whole spectrum with a maximum at ≈350 nm. The lifetimes of the transients were 200 ± 50 ns, probably corresponding to the triplet excited state or some radical (assignmen based on the quenching by O2), and 450 ± 100 μs tentatively assigned to QM-3 since in the O2-purged solution it has a similar lifetime, ≈400 μs. The spectra show additional transien absorbing at ≈350 nm and ≈550 nm that is not quenched by O2. It may correspond to spe cies generated from Trp, such as indolyl radical-cation and N-radical [65][66][67]. In the aque ous Ar-purged solution of 3 two transients were detected with the lifetimes of 10-30 ns probably corresponding to the triplet excited state or some radical and the one with the lifetime 190 ± 10 μs, tentatively assigned to the QM since it is not quenched by O2. In O2 purged solution the longest-lived transient has a similar lifetime, ≈150-200 μs.
In the LFP experiments for the salt 3×HCl, in Ar-purged solution we detected three transients with the lifetimes of ≈10 ns, probably corresponding to the triplet excited state or some radical, 10.6 ± 0.6 μs, not assigned and 1.3 ± 0.6 ms, tentatively assigned to the QM. In O2-purged solution we detected a similar long-lived transient with the lifetime ≈0.5-10 ms whose precise determination of decay kinetics was difficult due to very low intensity and long decay time.
Since the very low intensity of the transient (ΔA < 10 −3 ) that was tentatively assigned to QM-precluded quenching experiments, we performed control LFP experiments for 2 where the formation of QM is not possible. Figure 3 shows transient absorption spectra o 2 and 3 recorded after a delay of 27 μs after the laser pulse, revealing some difference a 300-400 nm, which may be assigned to the presence of QM based on the position of the anticipated maximum in the absorption spectra [52,53]. This long-lived transient with the lifetime of ≈400 μs was seen from 3 (but not from 2), we therefore assigned to QM-3. In terestingly, in the aqueous solution, a difference between the transient absorption spectra between 2 and 3 was not observed, presumably due to pronounced formation of N-Trp radicals from both peptides, in the process that involves photoionization to radical cation and its deprotonation [65][66][67].    For 1×HCl, 20% DMSO had to be added to the solution to increase the solubility of the peptide in H 2 O and for that reason only preliminary titration with calf thymus-DNA (ct-DNA) was performed ( Figure S38). For better solubility 2×HCl and 3×HCl titrations were performed with all DNA/RNA, resulting in quite different emission response. Namely, addition of guanine-containing DNAs (ct-DNA and GC-DNA) exclusively quenched emission of both, 2×HCl ( Figure S39) and 3×HCl (Figure 4). At variance, addition of AT-DNA caused either very small changes (2×HCl, Figure S40) or bisignate changes: (a) quenching of emission of 3×HCl ( Figure 4) at excess of dye (r[dye]/[DNA] > 0.5) followed by (b) emission increase at an excess of DNA (r < 0.5). The addition of AU-RNA yielded also biphasic changes with much more pronounced emission increase (3×HCl Figure 3 and 2×HCl Figure S42). The moderate fluorescence of the studied 1×HCl-3×HCl allowed fluorescence titrations with various double stranded (ds)-DNA/RNA, whereby we used for excitation λexc = 295 nm (selective for tryptophan) to avoid absorbing the light by increasing the amounts of DNA/RNA (for all data see Figures S38-S54 in the ESI).
For 1×HCl, 20% DMSO had to be added to the solution to increase the solubility of the peptide in H2O and for that reason only preliminary titration with calf thymus-DNA (ct-DNA) was performed ( Figure S38). For better solubility 2×HCl and 3×HCl titrations were performed with all DNA/RNA, resulting in quite different emission response. Namely, addition of guanine-containing DNAs (ct-DNA and GC-DNA) exclusively quenched emission of both, 2×HCl ( Figure S39) and 3×HCl (Figure 4). At variance, addition of AT-DNA caused either very small changes (2×HCl, Figure S40 [68], to calculate the binding constants ( Table 2). The affinities of 2×HCl, 3×HCl to all studied ds-polynucleotides were similar (logKa = 6.0-6.8), thus suggesting that the observed difference in emission response (Figure 4b, Table 2) is not a result of the binding strength, but more likely the result of fine differences in positioning of the fluorophore within the DNA/RNA-binding site, affecting the Trp photophysical properties.  [68], to calculate the binding constants ( Table 2). The affinities of 2×HCl, 3×HCl to all studied ds-polynucleotides were similar (logK a = 6.0-6.8), thus suggesting that the observed difference in emission response (Figure 4b, Table 2) is not a result of the binding strength, but more likely the result of fine differences in positioning of the fluorophore within the DNA/RNA-binding site, affecting the Trp photophysical properties.
Comparison of fluorescence responses of synthetic DNA/RNAs can be correlated to the corresponding secondary structures. Namely, AT-DNA minor groove size and shape is excellently suited for small molecule binding [69,70], allowing also the insertion of dimeric species at excess of dye over DNA-binding sites (r[dye]/[DNA] > 0.5) [71], which can at excess of DNA (r < 0.25) de-aggregate and each molecule binds to individual binding site. At contrast to AT-DNA, GC-DNA is characterized by guanine amino groups protruding into minor groove, which sterically allows insertion of only single molecules. Moreover, guanine is the nucleobase with the lowest oxidation potential [72], allowing for photoinduced electron transfer (PET) between the peptide and guanine and leading exclusively to the fluorescence quenching. PET has been also documented for some 4,9-pyrene and acridine derivatives, which showed guanine-induced quenching of emission [73]. Finally, AU-RNA is well-known for narrow and deep, exceedingly hydrophobic major groove, which can bind small molecules, and due to efficient H 2 O exclusion yield strong emission increase of hydration-sensitive fluorophores (like Trp in this case).  3 Performed in 20% DMSO/buffer, and therefore, not comparable to other values obtained in buffer. 4 Too small emission changes for accurate processing of data. 5 Almost linear increase of emission allowed only approximation of logK a and Int.

Thermal Denaturation
The thermal denaturation experiments provide information about the ds-polynucleotide helix thermal stability as a function of interaction with added small molecules [74]. The difference between the T m value of free ds-polynucleotide and a complex with a small molecule (∆T m value) is an important factor in the characterization of small molecule/dspolynucleotide interactions. For instance, moderate to strong stabilization (∆T m > 5 • C) supports pronounced intercalative or minor groove-binding interaction [70], whereas weak or negligible stabilization (∆T m = 0-5 • C) suggests a binding process driven mostly by hydrophobic effect accompanied by weak H-bonding and/or electrostatic interactionsusually excluding classical intercalation as a binding mode.
Since only 3×HCl showed measurable interaction with ds-DNA/RNA in fluorescence titrations, we investigated the impact of 3×HCl on thermal denaturation of ct-DNA. The compound had only negligible effect to the stability of double helix ( Figure S49 and Table S3 in the SI), increasing the melting temperature (T m ) by 0.5 • C only, which is within the experimental error. However, upon irradiation (300 nm, 5 min) of ct-DNA in the presence of the compound, the thermal stability of ds-DNA decreased (∆T m = −2.2 • C), indicating destabilization of double helix, which may be due to photochemical alkylation of a single nucleobase, consequently resulting in steric hindrance of base pair recognition and also adjacent base pair stacking and thus decreasing the thermal stability of polynucleotide double helical structure.

CD Experiments
To investigate the mode of binding of peptides 1×HCl-3×HCl to ct-DNA, circular dichroism (CD) spectroscopy was used. CD spectroscopy is an useful analytical tool in the binding study of small molecules to chiral macromolecules such as DNA [75] since it can provide information on the binding mode to polynucleotide, with distinctive spectral differences for intercalators and groove-binding derivatives [76,77]. Compounds are chiral and possess CD spectra, and measurable positive signals are located around 235 nm, thus not interfering with the positive CD bands of DNA/RNA in the 260-290 nm range.
Thus, the small decrease of ct-DNA positive band at 275 nm caused by the addition of 3×HCl ( Figure 5) can be attributed to binding of small molecule to DNA and consequent unwinding of DNA double helix to accommodate small molecule, resulting in minor loss of helicity (noted as CD band decrease). Changes in 230-260 nm range are mostly due to the additive impact of CD of 3×HCl (Figure 5 UP). Analogously, upon addition of 3×HCl to GC-DNA and AU-RNA minor decrease of positive bands at 260-290 nm range was observed ( Figures S53 and S55 in the SI), while for AT-DNA the change was negligible ( Figure S52 in the SI). Similar trend was observed in the CD spectra of ct-DNA with 1×HCl (with 20% DMSO) and 2×HCl ( Figures S50 and S51 in the SI).
To investigate the mode of binding of peptides 1×HCl-3×HCl to ct-DNA, circular dichroism (CD) spectroscopy was used. CD spectroscopy is an useful analytical tool in the binding study of small molecules to chiral macromolecules such as DNA [75] since it can provide information on the binding mode to polynucleotide, with distinctive spectral differences for intercalators and groove-binding derivatives [76,77]. Compounds are chiral and possess CD spectra, and measurable positive signals are located around 235 nm, thus not interfering with the positive CD bands of DNA/RNA in the 260-290 nm range.
Thus, the small decrease of ct-DNA positive band at 275 nm caused by the addition of 3×HCl ( Figure 5) can be attributed to binding of small molecule to DNA and consequent unwinding of DNA double helix to accommodate small molecule, resulting in minor loss of helicity (noted as CD band decrease). Changes in 230-260 nm range are mostly due to the additive impact of CD of 3×HCl (Figure 5 UP). Analogously, upon addition of 3×HCl to GC-DNA and AU-RNA minor decrease of positive bands at 260-290 nm range was observed ( Figures S53 and S55 in the ESI), while for AT-DNA the change was negligible ( Figure S52 in the ESI). Similar trend was observed in the CD spectra of ct-DNA with 1×HCl (with 20% DMSO) and 2×HCl (Figures S50 and S51 in the ESI). In summary, CD and fluorescence titration results, as well as thermal denaturation experiments, support binding of 1-3 into minor groove of DNAs and major groove of RNA. Sterically more restricted GC-DNA groove allows only binding of single molecules, at variance to AT-DNA minor and AU-RNA major groove, which also allow binding of dimers (only at excess of peptide over DNA/RNA-binding sites).

Covalent Binding to Polynucleotides
Covalent binding of tripeptide 3 (1 mM) to nucleotides was essayed by irradiation (300 nm) in the presence of the following nucleotides (10 mM): 2′-deoxyadenosine, 2′-deoxycytidine, 2′-deoxyguanidine, 2′-deoxyadenosine-5′-diphosphate, 2′-deoxycytidine-5′monophosphate, and 2′-deoxyguanidine-5′-monophosphate. However, HPLC-MS and NMR analyses revealed no trace of any adduct to nucleotide. The finding is in accord with In summary, CD and fluorescence titration results, as well as thermal denaturation experiments, support binding of 1-3 into minor groove of DNAs and major groove of RNA. Sterically more restricted GC-DNA groove allows only binding of single molecules, at variance to AT-DNA minor and AU-RNA major groove, which also allow binding of dimers (only at excess of peptide over DNA/RNA-binding sites).

Covalent Binding to Polynucleotides
Covalent binding of tripeptide 3 (1 mM) to nucleotides was essayed by irradiation (300 nm) in the presence of the following nucleotides (10 mM): 2 -deoxyadenosine, 2deoxycytidine, 2 -deoxyguanidine, 2 -deoxyadenosine-5 -diphosphate, 2 -deoxycytidine-5 -monophosphate, and 2 -deoxyguanidine-5 -monophosphate. However, HPLC-MS and NMR analyses revealed no trace of any adduct to nucleotide. The finding is in accord with the Rokita's results that non-covalent binding of the QM precursors was essential for the alkylation to occur [35]. The affinity of peptides toward mononucleotides was too low to ensure close contact in the moment of QM generation. Therefore, we investigated aptitude of photogenerated QM from 3 to alkylate double stranded DNA strains.
Models for double stranded (ds) DNA were obtained by annealing complementary oligonucleotides (dG 12 with dC 12 , or dA 10 with dT 10 ). After mixing oligonucleotides in the appropriate concentrations their solutions were heated to 90 • C and allowed to cool slowly to rt to initiate the annealing. Formation of the ds-dG 12 -dC 12 and dA 10 -dT 10 was demonstrated by the thermal denaturation experiment ( Figure S55 in the SI) and CD spectroscopy ( Figure S56 in the SI). The solutions containing the annealed ds-oligonucleotides in ammonium acetate buffer were analyzed by HPLC (for details see Table S4 in the SI and Figures S57-S75), where dG 12 -dC 12 shows up as one peak with the retention time of 7 min ( Figure S57) and dA 10 -dT 10 as two signals with the retention times of 7 min and 19 min, corresponding to the non-protonated and protonated form, respectively.
The addition of 3×HCl to the solutions of ds-oligonucleotides affected their retention times on HPLC due to the non-covalent binding and formation of complexes. Thus, in the presence of 3×HCl the retention time of dG 12 -dC 12 was retarded by ≈0.3 min (Figure S60), whereas the effect was opposite for the dA 10 -dT 10 where the signal at 7 min appeared ≈0.5 min earlier ( Figure S72). Further, we performed irradiations of the mixtures of 3×HCl with ds-oligonucleotides (300 nm, 15 or 60 min), and the composition of the irradiated solutions was analyzed by HPLC. As control experiments we performed irradiation of 3×HCl under the same conditions in the same buffered solutions, as well as irradiations of ds-oligonucleotides. Interestingly, overlapped HPLC chromatograms after the irradiation 3×HCl and dG 12 -dC 12 do not show any new signal which could be attributed to the photoinduced alkylation of the oligonucleotides ( Figures S61-S64), whereas upon the irradiation of 3×HCl and dA 10 -dT 10 a new broad signal at ≈16.5 min appeared, not seen in the control irradiation experiments ( Figures S71, S73 and S74). This finding is at variance to previous reports on alkylation of polynucleotides by o-QMs, where the reaction at the exocyclic guanine NH 2 group predominated [33]. The observed preferable alkylation of AT may be explained by specific binding of 3×HCl to AT-sequence, probably as an aggregate inserted deeply within the minor groove, whereupon the adenine residues are the site of attack by QMs. On the contrary, binding of single molecules 3×HCl to the minor groove of GC-sequence sterically blocked by protruding amino groups left the peptide much more exposed to the aqueous environment, which upon exposure to irradiation and formation of QMs lead primarily to hydrolysis [48,49], instead of the alkylation. The results further indicate that positioning of the QM-precursor in the proximity of the ds-helices probably plays more important role than the intrinsic reactivity of the QM.

General
1 H and 13 C NMR spectra were recorded at 300 or 600 MHz at rt, using TMS as a reference and chemical shifts were reported in ppm. Melting points were determined using a Mikroheiztisch apparatus and were not corrected. IR spectra were recorded on a spectrophotometer in KBr and the characteristic peak values were given in cm −1 . HRMS were obtained on a MALDI TOF/TOF instrument. Semipreparative HPLC separations were performed on a Varian Pro Star instrument equipped with a Phenomenex Jupiter C18 5µ 300A column, using CH 3 OH/H 2 O + TFA as an eluent. Analysis of samples was performed on a Shimadzu HPLC equipped with a diode-array detector, and a Phenomenex Luna 3u C18(2) column was used. Mobile phase was CH 3 OH/H 2 O + TFA. Analyses of the solutions containing annealed oligonucleotides and 3×HCl were performed on a HPLC equipped with a PLRP-S 5 µm column and the mobile phase was CH 3 CN/ammonium acetate aqueous buffer. Details on methods for the sample analysis are given in the SI. Irradiation experiments were performed in a reactor equipped with 8 lamps with the output at 254 or 300 nm (1 lamp 8 W). Solvents for the irradiations were of HPLC purity. Chemicals were purchased from the usual commercial sources and were used as received. Solvents for chromatographic separations were used as they are delivered from supplier (p.a. or HPLC grade) or purified by distillation (CH 2 Cl 2 ). Preparation of known compounds, N-Boc-L-Trp-OSu [78], N-Boc-L-Trp-L-Trp-OH [21], N-Boc-L-Trp-L-Trp-OSu [78], N-Boc-L-Trp-L-Trp-L-Phe-OBn [21], N-Boc-L-Trp-L-Trp-L-Tyr-OBn [21] are given in the supporting information. Polynucleotides were purchased as noted: poly A-poly U, poly (dGdC) 2 , poly dG-poly dC, poly (dAdT) 2 , ct-DNA (Sigma, St. Louis, MO, USA). Polynucleotides were dissolved in Na-cacodylate buffer, I = 0.05 M, pH 7.0. ct-DNA was additionally sonicated and filtered through a 0.45-mm filter. Polynucleotide concentration was determined spectroscopically as the concentration of phosphates (nucleobases), as described by producer.

Preparation of N-Boc-L-Trp-L-Trp-L-Tyr[CH 2 N(CH 3 ) 2 ]-OBn (3)
A solution of N-Boc-L-Trp-L-Trp-OSu (320 mg, 0.65 mmol) in THF-u (6 mL) was added slowly to a suspension of the salt TFA×H-Tyr[CH 2 N(CH 3 ) 2 ]-OBn (315 mg, 0.6 mmol) and NaHCO 3 (252 mg, 3.0 mmol) in a mixture of THF and H 2 O (1:1, 12 mL). The reaction mixture was stirred at rt over 3 days. THF was removed on a rotary evaporator and the aqueous residue was acidified by HCl (0.5 M) to the pH value of 2-3. The extraction with EtOAc (3 × 50 mL) was carried out and the extracts were dried over anhydrous sodium sulfate. After filtration, the solvent was removed on a rotary evaporator and the crude residue was purified on a column of silica gel using CH 3 OH/CH 2 Cl 2 (1→20%) as eluent to afford the pure oily product (0.14 g, 29%). 1  Four quartz test tubes were filled with a solution of 3 (4 × 10 mg, 4 × 0.013 mmol) in CH 3 OH (4 × 15 mL). The solutions were purged with N 2 for 30 min, sealed and irradiated in a reactor equipped with 8 lamps with the output at 300 nm over 15 min. The composition of the irradiated solution was analyzed by HPLC. After the irradiation, the solvent was removed on a rotary evaporator and the residue was purified by TLC on silica gel using 10% CH 3 OH/CH 2 Cl 2 as an eluent. The pure photoproduct was isolated in the form of oily substance (9 mg, 22%), and some starting compound was regenerated (8 mg, 20%). 1

Quantum Yield of Methanolysis
Quantum yields for the photomethanolysis reactions for 3 and 3×HCl were determined by using KI/KIO 3 (Φ 254 = 0.74) actinometer [58,60], as recently described by us [79]. Solutions of peptides in CH 3 OH and actinometer were freshly prepared and their concentrations were adjusted to have absorbances of 0.4-0.8 at 254 nm. After adjustment of the concentrations and measurement of the corresponding UV-vis spectra, the solutions were purged with a stream of N 2 (20 min), and then, sealed with a cap. The cells were placed in a holder which ensured equal distance of all samples from the lamp and were irradiated at the same time in the reactor with 1 lamp at 254 nm for 10 min. Before and after the irradiation, the samples were taken from the cells using a syringe and analyzed by HPLC to determine the photochemical conversions. The conversion did not exceed 30% to avoid a change of the absorbance, or filtering of the light by the product. From the conversion of actinometer, irradiance was calculated according to Equations (S1)-(S5) reported in the SI. The average value of three measurements was reported.

Absorption and Fluorescence Measurements
Absorption spectra were recorded on a PG T80/T80+ or a Varian Cary 100 Bio spectrophotometer at rt. Fluorescence measurements were performed on an Agilent Cary Eclipse fluorometer by using slits corresponding to the bandpass of 10 nm for the excitation and the emission. The samples were dissolved in EtOAc, THF, CH 3 CN, or CH 3 CN-H 2 O (1:9) and the concentrations were adjusted to absorbances of less than 0.1 at the excitation wavelengths of 280, 290, or 300 nm. Fluorescence quantum yields were determined by comparison of the integral of the emission bands with the one of NATA in H 2 O (Φ f = 0.14) [58]. One fluorescence measurement was performed by exciting sample at three different wavelengths and the average value was calculated (Equation (S5) in the SI). Prior to the measurements, the solutions were purged with Ar for 15 min. The measurement was performed at rt (25 • C).
TC-SPC measurements were performed on an Edinburgh FS5 spectrometer equipped with a pulsed LED at 280 nm. The duration of the pulse was ≈800 ps. Fluorescence signals at 370 nm were monitored over 1023 channels with the time increment of ≈20 ps/channel. The decays were collected until they reached 3000 counts in the peak channel. A suspension of silica gel in H 2 O was used as a scattering solution to obtain instrument response function (IRF). Absorbances at 280 nm were 0.07-0.09. Prior to the measurements the solutions were purged with a stream of nitrogen for 20 min. The measurement was performed at rt (25 • C). Decays of fluorescence were fit to a sum of exponentials according to Equation (S7) in the SI, using software implemented with the instrument.

Preparation of the Annealed Double Stranded Oligonucleotides
Stock solutions of dG 12 , dC 12 , dA 10 , and dT 10 (c = 1.0 × 10 −3 M), were prepared in ammonium acetate buffer (100 mM, pH 7.0). The solution of dG 12 (250 µL) was mixed with the solution of dC 12 (250 µL) in a UV cuvette and their mixture was heated to 90 • C for 5 min, and slowly cooled to rt and stored for 24 h at 4 • C before the use in experiments. Analogously, the solution of dA 10 (180 µL) was mixed with the solution of dT 10 (180 µL) in a UV cuvette, heated to 90 • C for 5 min, slowly cooled to rt and stored for 24 h at 4 • C before using. The solutions of annealed oligonucleotides were used in the thermal denaturation experiments, CD and photochemical alkylation experiments.  [72]. ∆T m values were calculated by subtracting T m of the free nucleic acid from that of the respective complex with ∆T m values (Equation (S7) in the SI) are the average of at least two independent measurements and the error in ∆T m is ca. ± 0.5 • C.

Fluorescence Titrations
Stock solutions of solution 1×HCl-3×HCl were prepared in DMSO (c = 5 × 10 −3 M). For the titrations, the stock solutions were diluted in a fluorescence cell (3 mL) with cacodylate (pH 7.0, 50 mM) to reach the concentration c = 2.0-6.0 × 10 −6 M, thus having <1% DMSO. Polynucleotide stock solutions were c = 1-2 × 10 −4 M. The fluorescence spectra were measured on a Cary Eclipse (Agilent Technologies, Santa Clara, CA, USA) at 25 • C. The samples were excited at 295 nm, and the emission was recorded in the range 300-600 nm. The excitation slit was set to the bandpass of 10 nm, and the emission slit to 20 nm. Small aliquots of the solutions of polynucleotides were added to the solution of 1×HCl-3×HCl and after an incubation time of 2 min, fluorescence spectra were taken. Data obtained by fluorescence titrations were processed by nonlinear regression analysis according to the Scatchard model, McGhee, von Hippel formalism [68]. Each titration with ct-DNA was performed five times, and with other polynucleotides at least twice.  (Table S4). Experiments were performed twice.

Laser Flash Photolysis (LFP)
The measurements were performed on a LP980 Edinburgh Instruments spectrophotometer. For the excitation the fourth harmonic of a Q-smart Q450 Quantel YAG laser at 266 nm was used. The energy of the laser pulse at 266 nm was set to 20 mJ and the pulse duration was 7 ns. Absorbances at the excitation wavelength were set to 0.3-0.4. The static cells were used and they were frequently exchanged to assure no absorption of light by photoproducts. At least three decays were collected to determine the decay kinetics, and the average values were reported, whereas the quoted errors correspond to maximal standard deviations. The solutions were purged for 15 min with Ar or O 2 prior to the measurements, which were conducted at 25 • C.

Conclusions
We have demonstrated that photochemically reactive modified tyrosine can be incorporated in peptides containing Trp amino acids, where it remains photochemically reactive (Φ R = 0.11-0.13), whereas the peptide retains the Trp fluorescence (Φ f = 0.03-0.06). The intrinsic fluorescence properties of Trp residues can be used for the study of peptide interactions with polynucleotides. The affinities of 2×HCl and 3×HCl to all studied dspolynucleotides were similar (logK a = 6.0-6.8). However, 3×HCl fluorescence spectral responses were strongly dependent on the base pair composition: the GC-containing polynucleotides efficiently quenched Trp emission, at variance to AT-or AU-polynucleotides, which induced bisignate response. Namely, addition of AT(U) polynucleotides at excess over studied peptide induced emission quenching (attributed to Trp-aggregation in the grooves of polynucleotides), whereas at excess of DNA/RNA over peptide, Trpfluorescence increase was observed. This result suggested that studied peptides bind into the DNA minor groove, deep inserted into the AT-sequence and much superficially inserted into the sterically blocked GC-sequence.
Upon UV-irradiation of the photochemically reactive peptide 3×HCl bound to dsoligonucleotides, only QM-induced alkylation of dA 10 -dT 10 was detected, whereas no covalent reaction with dG 12 -dC 12 was observed. This reactivity pattern is not the typical for QMs, showing that non-covalent binding and positioning of the QM precursor within the polynucleotide-binding site might be the deciding factor for the QM reactivity. It seems that just deep insertion into the AT-DNA minor groove ensured effective alkylation, at variance to the GC-DNA binding, in which peptide is significantly more exposed to aqueous environment. Therefore, our investigations show a proof of principle that QM precursor can be introduced into a peptide and used as a photochemical switch to enable primarily noncovalent interaction and consequently light-induced alkylation of ds-DNA, enabling further applications and important impact in chemistry and biology.