Photochemical Reactivity of Naphthol-Naphthalimide Conjugates and Their Biological Activity

Quinone methide precursors 1a–e, with different alkyl linkers between the naphthol and the naphthalimide chromophore, were synthesized. Their photophysical properties and photochemical reactivity were investigated and connected with biological activity. Upon excitation of the naphthol, Förster resonance energy transfer (FRET) to the naphthalimide takes place and the quantum yields of fluorescence are low (ΦF ≈ 10−2). Due to FRET, photodehydration of naphthols to QMs takes place inefficiently (ΦR ≈ 10−5). However, the formation of QMs can also be initiated upon excitation of naphthalimide, the lower energy chromophore, in a process that involves photoinduced electron transfer (PET) from the naphthol to the naphthalimide. Fluorescence titrations revealed that 1a and 1e form complexes with ct-DNA with moderate association constants Ka ≈ 105–106 M−1, as well as with bovine serum albumin (BSA) Ka ≈ 105 M−1 (1:1 complex). The irradiation of the complex 1e@BSA resulted in the alkylation of the protein, probably via QM. The antiproliferative activity of 1a–e against two human cancer cell lines (H460 and MCF 7) was investigated with the cells kept in the dark or irradiated at 350 nm, whereupon cytotoxicity increased, particularly for 1e (>100 times). Although the enhancement of this activity upon UV irradiation has no imminent therapeutic application, the results presented have importance in the rational design of new generations of anticancer phototherapeutics that absorb visible light.


Introduction
Reagents that can alkylate and cross-link DNA molecules are among the most cytotoxic molecules, often employed in anticancer treatment [1]. Among the molecules that have been intensively studied as DNA cross-linking agents are quinone methides (QMs), reactive intermediates in phenol chemistry and photochemistry [2]. The biological activity of QMs [3,4] has been connected to their reactivity with proteins [5][6][7], nucleobases [8], DNA [9][10][11][12] and G-quadruplexes [13][14][15][16]. Moreover, the antiproliferative action of the anticancer antibiotic mitomycin [17][18][19] is rationalized by the intracellular formation of QMs, followed by reversible DNA cross-linking [20][21][22]. An important advantage in the use of QMs in biological systems stems from the fact that QMs can be generated in photochemical reactions under mild conditions [23,24], which allows for the spatial and temporal control of their formation. However, QMs react efficiently with H 2 O molecules [25][26][27], which competes with the reactivity with biological targets such as DNA. Therefore, for efficient 27], which competes with the reactivity with biological targets such as DNA. Therefore, for efficient DNA alkylation and cross-linking, it is important to attach to the QM precursor units to groups that can bind to DNA through noncovalent interactions [10].
We have investigated the photoinduced antiproliferative effect of several classes of QM-precursors, consisting of hydroxybiphenyls [28,29], naphthols [30,31], and anthrols [32,33]. Some anthrol derivatives were particularly interesting since they did not exhibit cytotoxicity in the dark, but upon exposure to near-visible light the cytotoxicity was enhanced 100 times and showed selectivity against cancer stem cells [33]. However, for the ultimate application of QM precursors in biological systems, it is important to develop molecules that can generate QMs upon excitation with visible light [34], as well as to increase their reaction selectivity with biological targets, which can be facilitated via the substitution of QM precursors to DNA binding units [10]. In that context, we have explored the application of QM precursors attached to phenanthridine [35]. Further modifications are required, since in these examples QMs were generated by UV light, precluding their application in living cells.
Herein we present an investigation of the photochemical reactivity and noncovalent and covalent binding of a series of naphthol-naphthalimide conjugates, 1. Naphthol QM precursors are known to undergo efficient photodehydration and deliver QMs (Scheme 1) [36], which can be applied in biological systems and labelling [6,[37][38][39][40]. Our new QM precursors are connected by an alkyl linker of different length to 1,8-naphthalimide, which is a known pharmacophore, exhibiting antiviral [41][42][43], analgesic [44], local anesthetic [45], antitrypanosomal [46] and antagonist activity against 5-HT3 and 5-HT4 receptors [47][48][49]. Of particular interest is the ability of 1,8-naphthalimide derivatives to intercalate into DNA [50,51], which was exploited in the development of anticancer agents [52,53], initiating further clinical studies [54][55][56][57][58]. Furthermore, 1,8-naphthalimides undergo photoinduced electron transfer with guanine-rich polynucleotides [59,60] and induce cleavage of the chains [61,62]. Consequently, newly designed molecules contain two warheads for their anticancer activity. We have investigated the photochemical reactivity of 1 and showed that compounds undergo dehydration, delivering QMs, albeit inefficiently. Furthermore, QMs can be formed upon the excitation of low-energy chromophore naphthalimide, in a process that involves photoinduced electron transfer (PET), as has been demonstrated in a similar system by Freccero et al. [63]. The photophysical properties of 1 were investigated with steady-state and time-resolved fluorescence, whereas reactive intermediates in the photochemistry of 1 were detected through laser flash photolysis (LFP). The biological applicability of 1 was demonstrated by the binding study of 1 to calf thymus DNA (ct-DNA) and bovine serum albumin (BSA). The antiproliferative activity for 1 was investigated on human cancer carcinoma cells, with or without exposure to irradiation, showing an enhancement of the effect upon irradiation. Scheme 1. Photodehydration of Naphth to quinone methide.

Synthesis
The synthesis of naphthol-naphthalimide conjugates is based on a Grignard reaction as the key step in connecting two molecular fragments. The key intermediates, 1,8-naphthalimide derivatives 4, for the coupling, were prepared in three steps (Scheme 2) in good to excellent yields. These steps involved the alkylation of the imide by dibromides to afford 2 [64], followed by hydrolysis by HMPA and H2O to alcohols 3 [65], and oxidation to furnish 4. We tried two oxidation protocols, those of Swern [66] and Dess-Martin [67], Scheme 1. Photodehydration of Naphth to quinone methide.

Synthesis
The synthesis of naphthol-naphthalimide conjugates is based on a Grignard reaction as the key step in connecting two molecular fragments. The key intermediates, 1,8naphthalimide derivatives 4, for the coupling, were prepared in three steps (Scheme 2) in good to excellent yields. These steps involved the alkylation of the imide by dibromides to afford 2 [64], followed by hydrolysis by HMPA and H 2 O to alcohols 3 [65], and oxidation to furnish 4. We tried two oxidation protocols, those of Swern [66] and Dess-Martin [67], both of which produced 4 in good yields. However, the procedure of Dess-Martin is a more elegant procedure, allowing for easier isolation of the products. from the naphthol. The same behavior was observed for all derivatives, 1 ( Figure S3). Since the emission spectrum of the naphthol overlaps with the absorption spectrum of the naphthalimide ( Figure S4 in the ESI), upon excitation of the naphthol, an efficient Förster resonance energy transfer (FRET) to the naphthalimide takes place (vide infra). Quantum yields of fluorescence (Φf) were measured in CH3CN and CH3CN-H2O (4:1 v/v) with the use of quinine sulfate in 0.5 M H2SO4 as a reference (Φf = 0.546) [76] ( Table 1).
The Φf values are generally low, in accordance with the emission properties of 1,8-naphthalimides, which are known to undergo efficient intersystem crossing (ΦISC > 0.9) and populate triplet excited states [77]. The higher Φf measured in the aqueous solvent than in the neat CH3CN is also in accord with the photophysical properties of naphthalimides. Protic solvents form H-bonds with the imide chromophore, which induces bathochromic shifts in the fluorescence spectra. Furthermore, protic solvents perturb the energy levels of the n,π* and π,π* singlet and triplet excited states, and therefore affect the efficiency of the ISC [74]. One generally sees an increase in the Φf by increasing the alkyl linker length between the two chromophores. A mechanism that can involve depopulation of the imide singlet or triplet excited state is photoinduced electron transfer (PET) from the naphthol to the imide, since 1,8-naphthlimides are good oxidizing agents in the excited state [78][79][80].  Quantum yields of fluorescence (Φ f ) were measured in CH 3 CN and CH 3 CN-H 2 O (4:1 v/v) with the use of quinine sulfate in 0.5 M H 2 SO 4 as a reference (Φ f = 0.546) [76] ( Table 1). The Φ f values are generally low, in accordance with the emission properties of 1,8-naphthalimides, which are known to undergo efficient intersystem crossing (Φ ISC > 0.9) and populate triplet excited states [77]. The higher Φ f measured in the aqueous solvent than in the neat CH 3 CN is also in accord with the photophysical properties of naphthalimides. Protic solvents form H-bonds with the imide chromophore, which induces bathochromic shifts in the fluorescence spectra. Furthermore, protic solvents perturb the energy levels of the n,π* and π,π* singlet and triplet excited states, and therefore affect the efficiency of the ISC [74]. One generally sees an increase in the Φ f by increasing the alkyl linker length between the two chromophores. A mechanism that can involve depopulation of the imide singlet or triplet excited state is photoinduced electron transfer (PET) from the naphthol to the imide, since 1,8-naphthlimides are good oxidizing agents in the excited state [78][79][80]. Singlet excited state lifetimes for 1a-1e in CH 3 CN were measured by means of timecorrelated single photon counting (TC-SPC; for all data see Figures S5-S10 and Table S4 in the ESI). The samples were excited at 340 nm where almost only naphthalimide absorbs light, and the emission was detected at 390 nm. Fluorescence decays were fit to single exponential functions, giving lifetimes in the range 80-130 ps, at the limits of detection of the setup used. For some derivatives, fluorescence decay was best described as sum of two exponentials, with a small contribution of a longer decay component ascribed to the emission from the naphthol.
Normalized fluorescence spectra of 1a-1e in CH 3 CN show the strong bands between 350 and 425 nm corresponding to the emission from the naphthalimide, and very weak shoulders at ≈340 nm corresponding to the emission from naphthol, quenched by FRET to the naphthalimide ( Figure S4 in the ESI). To measure the efficiency of FRET, we performed TC-SPC measurements by exciting samples at 280 nm, where both naphthol and naphthalimide absorb light, and by collecting decays at 340 nm, where only naphthol emits. These decays were compared to that of Naphth in CH 3 CN. Fluorescence decays of 1a-1e at 340 nm were all fit to a sum of three exponents, and the average decay time was significantly shorter compared to that of Naphth (τ = 9.07 ns), revealing the efficiency of FRET to be between 69% and 82% ( Table 1). The three-exponential decays were explained by the conformational mobility of molecules 1a-1e, resulting in a distribution of distances between the donor and the acceptor and different rates of FRET.

Photochemistry
Irradiation of Naphth in aqueous CH 3 OH leads to the dehydration and formation of QM, which reacts with nucleophiles, delivering the photomethanolysis product, methyl ether (Naphth-OMe Scheme 1) [36]. To probe for the methanolysis reaction with naphtholnaphthalimide conjugates, 1, we performed irradiations in CH 3 OH-H 2 O (4:1 v/v), where the anticipated methyl ethers would indicate the formation of QMs as intermediates.
Irradiations were performed at 254 nm, exciting mostly naphthol, or at 350 nm, where the naphthalimide was only excited. In both cases, photolyses were not efficient, 1 h irradiation gave photoproducts in 1-2% yield, detected using UPLC-MS. In addition to the anticipated ethers, some unidentified products were also formed. To characterize the anticipated ethers, they were prepared from 1 by means of H 2 SO 4 -catalyzed thermal methanolysis. By comparing UPLC-MS chromatograms of the synthesized ethers and irradiated mixtures, we have shown that the same ethers were formed in the photolyses. Preparative irradiations of 1a, 1c and 1e were performed at 350 nm, since photolyses were more efficient than those conducted at 254 nm. After the irradiations, the mixtures were separated by means of UPLC-MS, and enriched fractions in photoproducts were analyzed using NMR and UPLC-MS. In addition to the anticipated ethers, we detected 1,8-naphthalimide, formed via the fragmentation of 1.
The quantum yields of the photomethanolysis reaction were estimated upon excitation at 254 nm in CH 3 OH-H 2 O (4:1 v/v), which is a monochromatic irradiation source, and KI/KIO 3 was used as an actinometer (Φ R = 0.74) [81]. The results are compiled in Table 2. One generally sees very low values for Φ R , which are due to FRET from the naphthol to the naphthalimide. Such a result was not anticipated in the first instance, since it is known that photodehydration takes place in an ultrafast photochemical reaction [82], which in principle could be in competition with FRET.

Laser Flash Photolysis (LFP)
To detect intermediates in the photochemistry of naphthol-naphthalimides 1, LFP was used. The samples were excited with a Nd:YAG laser at 266 or 355 nm. The measurements were performed in Ar-and O 2 -purged CH 3 CN solution, where O 2 was expected to quench triplets and radicals, but not QMs (for all transient absorption spectra and the associated decay kinetics, see Figures S11-S16).
Upon excitation at 266 nm, in both Ar-and O 2 -purged solution, a strong transient was detected, with a maximum at 350 nm, with weaker absorption extending over the visible region. Thus, the transient absorption at 400 nm was fit to a sum of two exponents with lifetimes of τ = 2.6 ± 0.1 ns and τ = 23.1 ± 0.8 ns. The transients were tentatively assigned to the naphthol radical cation (λ max = 360 and 460 nm) [83] and the naphthalimide radical anion (λ max = 410 nm) [84], based on comparison with the published spectra in the preceding literature, and the known photochemistry of the phenol-imide conjugates [63]. The short lifetime was tentatively associated to the naphthol radical cation, prone to fast decay via deprotonation to a naphthoxyl radical. The species with longer lifetime probably corresponds to the naphthalimide radical anion, for which the main decay pathway is probably back electron transfer, taking place more slowly than protonation.
LFP measurements allowed for the detection of additional weak transient absorption with a maximum at 470 nm. It decayed with unimolecular kinetics, with the following lifetimes: τ Ar = 460 ± 20 ns; τ air = 290 ± 20 ns and τ O2 = 100 ± 3 ns. Since the transient was quenched by O 2 , k q = (8.7 ± 0.7) × 10 8 M −1 s −1 , and based on the comparison with literature [63,79], it was assigned to the naphthalimide triplet excited state. The naphthalimide triplet was formed using FRET from the naphthol, followed by ISC. Upon excitation at 355 nm, only one strong transient was detected, absorbing with a maximum at 470 nm (Figure 2), decaying with unimolecular kinetics with the lifetimes; τ Ar = 340 ± 2 ns; τ air = 245 ± 3 ns and τ O2 = 105 ± 2 ns. Based on the quenching by O 2 , k q = (7.3 ± 0.3) × 10 8 M −1 s −1 ; and its maximum of absorption [63,79], it was assigned to the naphthalimide triplet excited state. Note the much shorter decay time of the naphthalimide triplet in Ar-purged solution than the published value (τ = 0.25 ± 0.12 ms) [79], which can be explained by the intramolecular PET between the naphthol and the naphthalimide, although characterized with a slow process and the rate constant of k = 2.9 × 10 6 s −1 . Based on the proposed PET between the naphthol and the naphthalimide, the corresponding radical cation and the radical anion should have been detected. However, their fast decay and low absorptivity precluded their detection. Thus, after the decay of the triplet, due to the low intensity of the transient absorption signal, additional transients were not detected.

Laser Flash Photolysis (LFP)
To detect intermediates in the photochemistry of naphthol-naphthalimides 1, LFP was used. The samples were excited with a Nd:YAG laser at 266 or 355 nm. The measurements were performed in Ar-and O2-purged CH3CN solution, where O2 was expected to quench triplets and radicals, but not QMs (for all transient absorption spectra and the associated decay kinetics, see Figures S11-S16).
Upon excitation at 266 nm, in both Ar-and O2-purged solution, a strong transient was detected, with a maximum at 350 nm, with weaker absorption extending over the visible region. Thus, the transient absorption at 400 nm was fit to a sum of two exponents with lifetimes of τ = 2.6 ± 0.1 ns and τ = 23.1 ± 0.8 ns. The transients were tentatively assigned to the naphthol radical cation (λmax = 360 and 460 nm) [83] and the naphthalimide radical anion (λmax = 410 nm) [84], based on comparison with the published spectra in the preceding literature, and the known photochemistry of the phenol-imide conjugates [63]. The short lifetime was tentatively associated to the naphthol radical cation, prone to fast decay via deprotonation to a naphthoxyl radical. The species with longer lifetime probably corresponds to the naphthalimide radical anion, for which the main decay pathway is probably back electron transfer, taking place more slowly than protonation.
LFP measurements allowed for the detection of additional weak transient absorption with a maximum at 470 nm. It decayed with unimolecular kinetics, with the following lifetimes: τAr = 460 ± 20 ns; τair = 290 ± 20 ns and τO2 = 100 ± 3 ns. Since the transient was quenched by O2, kq = (8.7 ± 0.7) × 10 8 M −1 s −1 , and based on the comparison with literature [63,79], it was assigned to the naphthalimide triplet excited state. The naphthalimide triplet was formed using FRET from the naphthol, followed by ISC. Upon excitation at 355 nm, only one strong transient was detected, absorbing with a maximum at 470 nm ( Figure  2), decaying with unimolecular kinetics with the lifetimes; τAr = 340 ± 2 ns; τair = 245 ± 3 ns and τO2 = 105 ± 2 ns. Based on the quenching by O2, kq = (7.3 ± 0.3) × 10 8 M −1 s −1 ; and its maximum of absorption [63,79], it was assigned to the naphthalimide triplet excited state. Note the much shorter decay time of the naphthalimide triplet in Ar-purged solution than the published value (τ = 0.25 ± 0.12 ms) [79], which can be explained by the intramolecular PET between the naphthol and the naphthalimide, although characterized with a slow process and the rate constant of k = 2.9 × 10 6 s −1 . Based on the proposed PET between the naphthol and the naphthalimide, the corresponding radical cation and the radical anion should have been detected. However, their fast decay and low absorptivity precluded their detection. Thus, after the decay of the triplet, due to the low intensity of the transient absorption signal, additional transients were not detected.

Photochemical Reaction Mechanism
A plausible mechanism for the formation of ethers, as well as for the cleavage photoproducts upon excitation at 350 nm, is shown in Scheme 5. The mechanism involves PET from the naphthol to the naphthalimide as the key step, which was indicated from the LFP measurements, and is in accordance with the mechanism proposed by Freccero et al. [63]. Thus, the formation of the QM and methanolysis product upon excitation at 254 nm may proceed via the dehydration mechanism proposed by Popik et al. [36]. However, due to efficient FRET from the naphthol to the naphthalimide, the methanolysis upon excitation at 254 nm is not efficient. On the other hand, excitation at 350 nm leads to naphthalimide S 1 , which, after an efficient ISC, leads to a naphthalimide triplet excited state. The naphthalimide triplet may undergo PET, whereupon the naphthol radical cation and the naphthalimide radical anion are formed (1(CT) in Scheme 5). Efficient back electron transfer (BET) gives 1 in the ground state, and it is probably the main deactivation pathway resulting in the low efficiency of the photochemical processes. In a protic solvent, the acidic radical cation is deprotonated to the phenoxyl radical 1(BRA), whereas the naphthalimide radical anion is protonated to a ketyl radical 1(BR). The formation of radicals most probably leads to combinations of disproportionation and cleavage reactions, whereupon naphthalimide is formed as a detectable product. Note that the formation of biradicals such as 1(BR) may occur via photoinduced H-transfer reactions. Although such processes are not ubiquitous in the naphthalimide photochemistry [77], photochemical hydrogen abstractions are well documented for imide derivatives such as phthalimides [85]. Furthermore, as suggested by Freccero [63], the naphthalimide radical anion may be protonated before the BET, whereupon naphtholate 1 is formed. A subsequent loss of OH − leads to the formation of QM. Note that the QM formation, in this case, was initiated by excitation of the lower-energy chromophore naphthalimide, at 350 nm. Furthermore, the formation of QMs takes place more efficiently upon excitation at 350 nm, although excitation of the naphthol at 254 nm is followed by FRET to the naphthalimide. However, the FRET is not 100% efficient.

Photochemical Reaction Mechanism
A plausible mechanism for the formation of ethers, as well as for the cleavage photoproducts upon excitation at 350 nm, is shown in Scheme 5. The mechanism involves PET from the naphthol to the naphthalimide as the key step, which was indicated from the LFP measurements, and is in accordance with the mechanism proposed by Freccero et al. [63]. Thus, the formation of the QM and methanolysis product upon excitation at 254 nm may proceed via the dehydration mechanism proposed by Popik et al. [36]. However, due to efficient FRET from the naphthol to the naphthalimide, the methanolysis upon excitation at 254 nm is not efficient. On the other hand, excitation at 350 nm leads to naphthalimide S1, which, after an efficient ISC, leads to a naphthalimide triplet excited state. The naphthalimide triplet may undergo PET, whereupon the naphthol radical cation and the naphthalimide radical anion are formed (1(CT) in Scheme 5). Efficient back electron transfer (BET) gives 1 in the ground state, and it is probably the main deactivation pathway resulting in the low efficiency of the photochemical processes. In a protic solvent, the acidic radical cation is deprotonated to the phenoxyl radical 1(BRA), whereas the naphthalimide radical anion is protonated to a ketyl radical 1(BR). The formation of radicals most probably leads to combinations of disproportionation and cleavage reactions, whereupon naphthalimide is formed as a detectable product. Note that the formation of biradicals such as 1(BR) may occur via photoinduced H-transfer reactions. Although such processes are not ubiquitous in the naphthalimide photochemistry [77], photochemical hydrogen abstractions are well documented for imide derivatives such as phthalimides [85]. Furthermore, as suggested by Freccero [63], the naphthalimide radical anion may be protonated before the BET, whereupon naphtholate 1 is formed. A subsequent loss of OHleads to the formation of QM. Note that the QM formation, in this case, was initiated by excitation of the lower-energy chromophore naphthalimide, at 350 nm. Furthermore, the formation of QMs takes place more efficiently upon excitation at 350 nm, although excitation of the naphthol at 254 nm is followed by FRET to the naphthalimide. However, the FRET is not 100% efficient. The irradiation experiments and mechanistic investigations indicated the formation of QMs from 1, which can induce biological effects, but the reaction is inefficient. However, photolysis of 1 gives rise to a plethora of other reactive intermediates, such as free radicals that can engage in reactions with biomacromolecules and induce cell damage. Most importantly, upon excitation at 350 nm, the naphthalimide triplet excited state is populated, which is quenched by O 2 , giving rise to singlet O 2 and increased ROS, which is probably the main species responsible for the cytotoxic effect.

Noncovalent Binding to ct-DNA
Since naphthalimide derivatives are known to bind to polynucleotides [50,51], we investigated the non-covalent binding of 1 with naturally occurring calf thymus DNA (ct-DNA), characterized by the typical B-helical secondary structure and the almost equimolar ratio of AT-and GC-basepairs. The effect of binding was investigated via thermal denaturation experiments, whereas binding constants were determined by fluorescence titration. To probe for the binding mode (intercalation vs groove binding), we conducted CD measurements. Due to low solubility in H 2 O, stock solutions of 1 were prepared in DMSO and diluted with cacodylate buffer solutions.

Thermal Denaturation
The binding of small molecules to double stranded (ds) polynucleotides affects the stability of the ds-helix, usually by stabilizing its structure. Thus, thermal denaturation and unwinding of the ds-helix into the single stranded (ss) form takes place at a higher or lower temperature, known as the melting temperature (T m ) [86]. The difference between the T m value of a free ds-polynucleotide and a complex with a small molecule (∆T m value) provides information about the interaction between the small molecule and the dspolynucleotides. Moderate to strong stabilization (∆T m > 5 • C) supports the intercalative or minor groove binding interaction [87], whereas weak stabilization (∆T m = 0-5 • C) suggests a binding process driven by hydrophobic effects, accompanied by weak H-bonding or electrostatic interactions.
The thermal denaturation experiments were conducted with 1a and 1f and ct-DNA (Tables S5 and S6 and Figures S17 and S18 in the ESI). However, the compounds in the concentration corresponding to the ratio r = (1)/(ct-DNA) = 0.3 did not affect the thermal stability of the helices, suggesting non-specific binding along the DNA double helix.

UV-Vis and Fluorescence Titrations
The binding of 1 to ct-DNA was assayed via UV-vis and fluorescence titrations in cacodylate buffer at pH 7. However, UV-vis titrations did not show applicability in determining the association constants due to the low solubility of molecules 1 in the aqueous cacodylate buffer (containing less than 1% DMSO). In the measurement conducted with the use of a 5 cm pathlength probe, allowing titrations at lower concentrations of 1 and DNA, 1a precipitated immediately, whereas 1e precipitated after the addition of ct-DNA at the ratio r = (1)/(ct-DNA) = 0.5, precluding estimation of the constant. However, even three additions of DNA at r = 2-0.5 revealed a weak hypochromic effect at λ < 370 nm, and a hyperchromic effect at λ > 370 nm ( Figure S19   The quenching of fluorescence by ct-DNA may be rationalized by PET between the naphthalimide and guanine, which is the nucleobase with the lowest oxidation potential [59,60,89], and in principle, it can be static (due to the formation of a nonfluorescent complex) or dynamic (due to diffusion). To check if the quenching is static, we performed TC-SPC measurement for 1a and a mixture of 1a and ct-DNA, in the same concentrations as in the fluorescence titration experiments (see Figures S20-S21 and Table S7 in the ESI). The same decay time in the presence of ct-DNA would indicate the static quenching. However, the short decay times at the limits of the detection of the setup used precluded their precise determination. Thus, although shortening of the main decay component from 210 to 170 ps was observed, these values are within experimental error. Consequently, we cannot completely rule out a dynamic quenching mechanism, but the short decay times and quenching in the concentration of DNA at 10 −4 M is in favor of the static quenching mechanism. Namely, the Stern-Volmer analysis of the quenching by DNA would provide the quenching constant kq in the order of 10 12 M −1 s −1 , which is higher than the diffusion limit.
Non-linear fitting of the fluorescence dependence on the ct-DNA concentration using the Scatchard model [88]  The quenching of fluorescence by ct-DNA may be rationalized by PET between the naphthalimide and guanine, which is the nucleobase with the lowest oxidation potential [59,60,89], and in principle, it can be static (due to the formation of a nonfluorescent complex) or dynamic (due to diffusion). To check if the quenching is static, we performed TC-SPC measurement for 1a and a mixture of 1a and ct-DNA, in the same concentrations as in the fluorescence titration experiments (see Figures S20-S21 and Table S7 in the ESI). The same decay time in the presence of ct-DNA would indicate the static quenching. However, the short decay times at the limits of the detection of the setup used precluded their precise determination. Thus, although shortening of the main decay component from 210 to 170 ps was observed, these values are within experimental error. Consequently, we cannot completely rule out a dynamic quenching mechanism, but the short decay times and quenching in the concentration of DNA at 10 −4 M is in favor of the static quenching mechanism. Namely, the Stern-Volmer analysis of the quenching by DNA would provide the quenching constant k q in the order of 10 12 M −1 s −1 , which is higher than the diffusion limit.
Non-linear fitting of the fluorescence dependence on the ct-DNA concentration using the Scatchard model [88] gave a binding constant logK a = 5.8, for the fixed value of the Scatchard ratio n[1a]/[ct-DNA] = 0.2. On the contrary, the addition of polynucleotide to the solution of 1e resulted in an increase in the fluorescence (Figure 3, bottom), further supporting the possibility that the changes in fluorescence are not due to dynamic quenching. A tentative explanation for the fluorescence enhancement is the aggregation of the larger molecule 1e in the aqueous medium, yielding a lower intensity in respect to 1a (Figure 3, see starting intensities of free dyes). Thus, upon the addition of the DNA, de-aggregation of 1e takes place, resulting in the fluorescence increase. The association constant logK a = 6.3 (fixed value n = 0.2) was estimated through the non-linear fitting of data to the Scatchard model [88].

CD Experiments
Although the thermal denaturation experiments indicated that derivatives 1 do not intercalate or bind to grooves of ct-DNA, binding was further investigated by means of circular dichroism (CD) spectroscopy. Namely, CD spectroscopy is a powerful analytical tool for the binding studies of small molecules to chiral macromolecules such as DNA [90], giving rise to the distinctive spectral differences for intercalators and groove binding derivatives [91,92]. Compounds 1 are chiral, but they are present in a racemic mixture, and therefore they do not have measurable CD spectra. However, upon their addition to the solution of ct-DNA, they can affect the helical chirality of DNA, or if they form complexes with the characteristic organization of the chromophores along the chiral axes of DNA, an induced CD signal can show up in the region where molecules 1 absorb light [91,92].
The addition of 1a or 1e to the solution of ct-DNA induced only small changes in the DNA CD signal, decreasing the intensity of the positive band at 275 nm, and that of the negative band at 245 nm ( Figures S22 and S23 in the ESI). The lack of measurable ICD bands >300 nm suggests a non-uniform orientation of the chromophores of 1e in respect to the DNA chiral axis, which is in agreement with the non-specific binding of the small molecule along DNA double helix.
The binding studies with ct-DNA have shown that compounds 1a and 1e form complexes with polynucleotides with moderate binding constants. However, the negligible effect of 1a and 1e in the thermal denaturation of ct-DNA, as well as the absence of ICD bands in the CD experiments, suggest that the studied naphthol-naphthalimides bind to ds-DNA in a nonspecific way, most probably by agglomeration along the double helix, caused by hydrophobic effects, accompanied potentially by weak H-bonding interactions.

Noncovalent and Covalent Binding to BSA Protein
Naphthol naphthalimide derivatives are non-charged species and not very polar, being good substrates for noncovalent binding to serum albumin proteins. We took advantage of the intrinsic fluorescence of 1a and 1e to study their binding to BSA, presuming that the insertion of small molecules into a BSA binding site would change the hydration of chromophores, consequently yielding a measurable change in their emission. The dynamic quenching mechanism was again dismissed due to the short singlet excited state lifetimes of 1a-1e and the low concentration range of protein, which affected the fluorescence.
Accordingly, upon the addition of BSA, the fluorescence of 1a was quenched until a ratio of 1:1 between 1a and BSA was reached (Figure 4, arrow A). Further additions of BSA induced a fluorescence increase in 1a and a considerable broadening of the emission spectrum, accompanied by a significant hypsochromic shift of the emission maximum ( Figure 4, arrow B). Such opposite changes clearly indicate the formation of two types of complexes. The fluorescence quenching upon the addition of BSA to 1a is assigned to the aggregation of two molecules within the BSA binding site, whereas further additions of BSA allowed each molecule of 1a to bind to an independent binding site, thus resulting in the formation of a 1:1 complex (denoted as 1a@BSA) and a consequent fluorescence increase. Multicomponent nonlinear regression analysis of the fluorescence data using HypSpec2014 software [93] gave the best fit to the formation of 1a@BSA complexes in the stoichiometries 2:1 and 1:1, with the association constants logβ (1a 2 @BSA) = 11.38 ± 0.04 and logβ (1a@BSA) = 5.24 ± 0.05, respectively.  Contrary to 1a, the addition of BSA to the solution of 1e led to a fluorescence increase only ( Figure S23). Nonlinear regression analysis using HypSpec2014 software revealed the formation of one complex with the stoichiometric ratio of 1:1 and the association constant logKa (1e@BSA) = 5.05 ± 0.05. It should be noted that 1a and 1e form 1:1 complexes with BSA with similar binding constants, but only the smallest 1a forms a 2:1 complex within a BSA binding site.
Our preliminary binding experiments with BSA undoubtedly indicate the formation of complexes 1a@BSA and 1e@BSA. However, the exact binding site of 1 in the protein remained undetermined and is not in the scope of this report. To demonstrate that the irradiation of complexes of 1 with BSA leads to the alkylation of the protein, we performed irradiations of the 1a@BSA and 1e@BSA complexes. The aqueous buffered solutions of the complexes were irradiated at 350 nm, followed by desalting of the protein and analysis by means of MALDI-MS. Naphthol-naphthalimide with the shortest linker, 1a, did not alkylate the protein, whereas the one with the longest linker, 1e, covalently attached to BSA. The molecular weight of 67.1-67.2 kDa pointed to the addition of 1e to the protein (see Figures S25-S27 and Table S8 in the ESI). It is plausible to assume that the photoinduced mechanism of protein alkylation involves QM intermediates, as seen in the prior literature [6,7], although the photodehydration to QMs from 1 is inefficient.

Antiproliferative Activity
We have demonstrated that compounds 1a-1e bind to a model DNA and protein by means of noncovalent interactions, where, after photochemical excitation, they may induce damage. Therefore, we investigated the antiproliferative activity of compounds 1a-1e on two human cancer cell lines, H-460 (lung) and MCF-7 (breast), with the cells kept in the dark, or irradiated at 350 nm (for all data, see Table S9 in the ESI). Control experiments were performed with a psoralene derivative (trioxsalen), which is known to induce photoactivable cross-linking [12,94] and Naphth (Scheme 1). The analysis of results in Table  S9 showed decreasing cytotoxicity under dark conditions (expressed by IC50 values) proportional to the linker length between the two chromophores, with 1e bearing the longest linker, being non-toxic for the H-460 cell line even at the highest tested concentration. However, the enhancement of the antiproliferative effect upon UV irradiation (Table S9, 350 nm) shows the opposite trend-the strongest effects were observed for the compounds with longer linkers (1d, 1e). This result suggests that UV irradiation of 1e inside the cell leads to the production of a higher amount of cytotoxic agent (possibly QMs, as Contrary to 1a, the addition of BSA to the solution of 1e led to a fluorescence increase only ( Figure S23). Nonlinear regression analysis using HypSpec2014 software revealed the formation of one complex with the stoichiometric ratio of 1:1 and the association constant logK a (1e@BSA) = 5.05 ± 0.05. It should be noted that 1a and 1e form 1:1 complexes with BSA with similar binding constants, but only the smallest 1a forms a 2:1 complex within a BSA binding site.
Our preliminary binding experiments with BSA undoubtedly indicate the formation of complexes 1a@BSA and 1e@BSA. However, the exact binding site of 1 in the protein remained undetermined and is not in the scope of this report. To demonstrate that the irradiation of complexes of 1 with BSA leads to the alkylation of the protein, we performed irradiations of the 1a@BSA and 1e@BSA complexes. The aqueous buffered solutions of the complexes were irradiated at 350 nm, followed by desalting of the protein and analysis by means of MALDI-MS. Naphthol-naphthalimide with the shortest linker, 1a, did not alkylate the protein, whereas the one with the longest linker, 1e, covalently attached to BSA. The molecular weight of 67.1-67.2 kDa pointed to the addition of 1e to the protein (see Figures S25-S27 and Table S8 in the ESI). It is plausible to assume that the photoinduced mechanism of protein alkylation involves QM intermediates, as seen in the prior literature [6,7], although the photodehydration to QMs from 1 is inefficient.

Antiproliferative Activity
We have demonstrated that compounds 1a-1e bind to a model DNA and protein by means of noncovalent interactions, where, after photochemical excitation, they may induce damage. Therefore, we investigated the antiproliferative activity of compounds 1a-1e on two human cancer cell lines, H-460 (lung) and MCF-7 (breast), with the cells kept in the dark, or irradiated at 350 nm (for all data, see Table S9 in the ESI). Control experiments were performed with a psoralene derivative (trioxsalen), which is known to induce photoactivable cross-linking [12,94] and Naphth (Scheme 1). The analysis of results in Table S9 showed decreasing cytotoxicity under dark conditions (expressed by IC 50 values) proportional to the linker length between the two chromophores, with 1e bearing the longest linker, being non-toxic for the H-460 cell line even at the highest tested concentration. However, the enhancement of the antiproliferative effect upon UV irradiation (Table S9, 350 nm) shows the opposite trend-the strongest effects were observed for the compounds with longer linkers (1d, 1e). This result suggests that UV irradiation of 1e inside the cell leads to the production of a higher amount of cytotoxic agent (possibly QMs, as well as singlet oxygen/reactive oxygen species (ROS)) in comparison to the smallest, 1a (showing an order of magnitude weaker effect). Furthermore, compounds 1a-e exhibit higher cytotoxicity than Naphth, indicating that the naphthalimide moiety is important for this activity. Although we do not have clear evidence as to which of the reactive species formed in the photochemical reactions led to the observed photoinduced antiproliferative effect, it is plausible that singlet O 2 and subsequent ROS formed by the quenching of the naphthalimide triplet played an important role, as well as QMs and reactive radicals formed in the photoinduced processes. It should be noted that the investigated molecules cannot have therapeutic value, since irradiation with UV light was used. However, this proof of principle is important for the future rational design of photoactivable compounds with antiproliferative activity.

Conclusions
Naphthol-naphthalimide conjugates, QM precursors 1 undergo photodehydration to the corresponding QMs very inefficiently due to FRET from the naphthol to the naphthalimide. In addition to the elimination of H 2 O, QMs can also be formed upon excitation of the naphthalimide in a process that involves PET from the naphthol to the naphthalimide. The linker length variation between the naphthol and the naphthalimide in the 1a-1e series mildly affects their photophysical properties and the photochemical reactivity, but had a much more obvious impact on bio-relevant properties, DNA-fluorimetric response, protein-binding stoichiometry and cytotoxicity on cell lines. The naphthol-naphthalimide conjugates showed moderate affinity toward ds-DNA (K a ≈ 10 5 -10 6 M −1 ), due to their non-specific agglomeration along the DNA helix. Comparison of the shortest (1a) and the longest (1e) derivative revealed opposite fluorimetric responses, attributed to 1e aggregation in a free state and de-aggregation upon DNA binding. Furthermore, the studied compounds also efficiently bind to the BSA protein, whereby the shortest analogue, 1a, can form dimers inside a BSA binding site, at variance to the longest, 1e, which binds only as a monomer. Upon irradiation of the complex 1e@BSA, photoinduced alkylation of the protein takes place, probably via the QM. Derivatives 1a-e exhibit cytotoxic effects against cancer cell lines H-460 and MCF-7 that is reversely proportional to the linker length and correlated to the aggregation propensity of the longest derivative, 1e. All compounds showed enhanced cytotoxicity upon exposure to irradiation at 350 nm, and this was particularly pronounced for the longest analogue, 1e, which not only increased by two orders of magnitude in H-460 cells but also was an order of a magnitude stronger than the shortest analogue, 1a. The enhancement of this activity probably stems from the formation of singlet oxygen (or ROS-related species) but the effect of QMs and reactive radicals formed in the photoinduced processes cannot be disregarded. The particularly promising theragnostic properties of 1e, unifying photoinduced-bioactivity and fluorescence enhancement upon binding to biorelevant targets, encourage the further study of its close analogues.

Materials and Methods
General: 1 H and 13 C NMR spectra were recorded at 300, 400, 500 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 an ATR spectrometer and the characteristic peak values were given in cm −1 . HRMS was carried out on a MALDI TOF/TOF instrument. Details on semipreparative MPLC and UPLC separations and analyses are reported in the ESI. Chemicals were purchased from the usual commercial sources and were used as received. Solvents for conducting moisture-sensitive reactions were further dried over sodium metal. Solvents for chromatographic separations were used as delivered from the supplier (p.a. or HPLCgrade). Preparations of known compounds 2, 3, 5 and 6 are described in the ESI. ct-DNA was dissolved in Na-cacodylate buffer, I = 0.05 mol dm −3 , pH 7.0, and it was additionally sonicated and filtered through a 0.45-µm filter. The polynucleotide concentration was determined spectroscopically as the concentration per nucleotide/phosphate, using molar absorption coefficient for calf thymus DNA of 6600 M −1 cm −1 at 260 nm.

• General procedure for the preparation of N-(ω-formylalkyl)-1,8-naphthalimide 4-Swern oxidation
A dry two-neck round-bottom flask, equipped with a dropping funnel and a septum, under Ar inert atmosphere, was charged with oxalyl chloride (5 mmol) and dry CH 2 Cl 2 (1 mL/mmol oxalyl chloride). The reaction mixture was cooled down to −78 • C, and a solution of DMSO (10 mmol) in dry CH 2 Cl 2 (0.2 mL/mmol DMSO) was added dropwise. After 1 h of stirring at −78 • C to the reaction mixture, a solution of 3 (1 mmol) in dry CH 2 Cl 2 (1 mL/mmol 3) was added dropwise. The reaction mixture was stirred at −78 • C for 2 h. Dry triethylamine (15 mmol) was then added dropwise, and the reaction was stirred at −78 • C for 1 h and then was left to spontaneously reach rt over 17 h. The reaction mixture was transferred to a separation funnel, diluted with H 2 O and extracted with CH 2 Cl 2 (3 × 20 mL). The combined organic layers were washed with saturated aqueous NH 4 Cl (3 × 20 mL) and brine (2 × 20 mL). The solution was dried over anhydrous Na 2 SO 4 , filtered and the solvent was removed on a rotary evaporator. The residue was purified on a silica gel column using EtOAc (0 to 50%)/cyclohexane as eluent to afford the pure product.

Dess-Martin oxidation [95]
A round-bottom flask, equipped with a dropping funnel, was charged with Dess-Martin periodinane (DMP, 1.2 mmol) and CH 2 Cl 2 (2.5 mL/mmol DMP). A solution of 3 (1 mmol) in CH 2 Cl 2 (2.5 mL/mmol 3) was added dropwise at rt. The reaction mixture was stirred at rt for 1.5 h, and then a solution of Na 2 S 2 O 3 (9.5 mmol) in a saturated aqueous solution of NaHCO 3 (1 mL/mmol Na 2 S 2 O 3 ) was added. The resulting suspension was stirred for 20 min after it was transferred to a separation funnel. The layers were separated and the aqueous layer was extracted with CH 2 Cl 2 (3 × 20 mL). The combined organic layers were washed with saturated aqueous NaHCO 3 (3 × 10 mL) and brine (2 × 10 mL). The solution was dried over anhydrous Na 2 SO 4 , filtered and the solvent was removed on a rotary evaporator. The residue was purified on a silica gel column using EtOAc (0 to 50%)/cyclohexane as eluent to afford the pure product.
N-(3-formylprop-1-yl)-1,8-naphthalimide (4a) was prepared according to the general procedure for Swern oxidation from alcohol 3a (2.15 g, 7.9 mmol). The reaction after work-up and chromatography furnished 1.17 g (55%) of the pure product in the form of a colorless solid. The NMR characterization was in accordance with the data from the preceding literature [96]. N-(7-formylhept-1-yl)-1,8-naphthalimide (4e) was prepared according to the general procedure for Swern oxidation from alcohol 3e (5.85 g, 17.9 mmol). The reaction after work-up and chromatography furnished 3.57 g (61%) of the pure product in the form of a colorless solid.  1 mmol) was added dropwise via a syringe. The reaction mixture was stirred at 0 • C for 30 min, and then it was allowed to reach rt over 5 h. To the reaction mixture, a saturated aqueous solution of NH 4 Cl (10 mL) was added, and after 15 min of stirring, the mixture was transferred to a separation funnel. Extraction with EtOAc (3 × 20 mL) was carried out. The combined organic layers were dried over anhydrous MgSO 4 and filtered and the solvent was removed on a rotary evaporator. The residue was purified on a silica gel column using EtOAc (0 to 30%)/cyclohexane as eluent to afford the pure product.

Removal of the benzyl group-general procedure
A round-bottom flask, equipped with a septum, was charged with benzyl ether 7 (1 mmol), THF (5 mL/mmol 7), methanol (5 mL/mmol 7), Pd/C (10%, 0.1 mmol) and ammonium formate (5.0 mmol). The reaction mixture was heated at 65 • C for 2-6 h and the progress of the reaction was monitored by means of UPLC-MS/UV (method 2 , see the ESI). After all the benzyl ether was consumed, the reaction mixture was cooled down to rt, filtered through a celite plug, and the solvent was removed on a rotary evaporator. The residue was purified by means of preparative HPLC-MS/UV or on a column of silica gel using EtOAc (0 to 30%)/cyclohexane as eluent to afford the pure product.

•
General procedure for the preparation of 1-OMe ethers using the acid-catalyzed thermal method A round-bottom flask (10 mL) was charged with a derivative 1 (≈10 mg), methanol (3.0 mL) and concentrated sulfuric acid (50 µL). The reaction mixture was stirred at rt for 15 h and then diluted with H 2 O (10 mL). The reaction mixture was neutralized with a saturated aqueous solution of NaHCO 3 (≈2 mL), transferred to a separation funnel and extracted with diethyl ether (3 × 5 mL). The combined organic layers were dried over anhydrous MgSO 4 and filtered and the solvent was removed on a rotary evaporator. The residue was chromatographed on a TLC using cyclohexane-EtOAc (7:3 v/v) as eluent to afford the pure product.

• Irradiation experiments-general
A quartz test tube (20 mL) was filled with a solution of 1 in CH 3 OH (7 mL/1 mg of 1), and the solution was diluted with H 2 O. The final CH 3 OH-H 2 O ratio was 4:1 (v/v). The solution was sealed with a septum and purged with Ar for 15 min and irradiated in a reactor equipped with 8 or 14 lamps (1 lamp = 8 W) with the output at 254 nm, or ≈350 nm. CH 3 OH for the irradiations was of HPLC purity, and Milli-Q H 2 O was used. The solutions were irradiated from 60 to 180 min, and the composition of the irradiated solutions was analyzed using UPLC-MS/UV.

Irradiation of 1c
According to the general procedure, 1c (7 mg, 13 µmol) was dissolved in CH 3

• Quantum yield of methanolysis
Quantum yields for the photomethanolysis reactions for 1 were determined using a KI/KIO 3 (Φ 254 = 0.74) actinometer [81], as recently described by us [97]. Solutions of 1 in CH 3 OH-H 2 O (4:1 v/v) and the actinometer were freshly prepared and their concentrations were adjusted to have absorbances of 0.4-0.8 at 254 nm. After the adjustment of the concentrations and measurement of the corresponding UV-vis spectra, the solutions were purged with a stream of Ar (20 min), and then, sealed with a cap. The cells were placed in a holder, which ensured an 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 6 h (1 min for the actinometer). Before and after the irradiation, the samples were taken from the cells with the use of a syringe and were analyzed using UPLC-MS/UV to determine photochemical conversions. The conversion did not exceed 30% to avoid a change in the absorbance or filtering of the light by the product. Based on the actinometer conversion, irradiance was calculated according to Equations (S1)-(S5), reported in the ESI. 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 fluorimeter with slits corresponding to the bandpass of 10 nm for the excitation and 20 nm for the emission. The samples were dissolved in CH 3 CN or CH 3 CN-H 2 O (4:1 v/v) and the concentrations were adjusted to absorbances of less than 0.1 at the excitation wavelengths of 300, 310 and 320 nm. Fluorescence quantum yields were determined through the comparison of the integral of the emission bands with that of quinine sulfate in 0.05 M H 2 SO 4 (Φ f = 0.546) [76]. One fluorescence measurement was performed by exciting sample at three different wavelengths and the average value was calculated (Equation (S6) in the ESI). Prior to the measurements, the solutions were purged with Ar for 15 min. The measurement was performed at rt (25 • C). Fluorescence decays were measured using the TC-SPC method on an Edinburgh FS5 spectrometer equipped with pulsed LEDs at 280 or 340 nm. The duration of the pulse was ≈800 ps. Fluorescence signals were monitored over 1023 channels with the time increment of ≈20 ps/channel. The decays were collected until they reached 1000 counts in the peak channel. The histograms were analyzed by means of a nonlinear least-squares deconvolution method using Equation (S7) in the ESI.

•
Thermal denaturation experiments with ct-DNA [86] A stock solution of 1a or 1e was prepared in DMSO (c = 2.0 × 10 −2 M), whereas the stock solution of ct-DNA was prepared in aqueous cacodylate buffer (pH = 7.0, 50 mM) c(ct-DNA) = 1.1 × 10 −3 M. In the denaturation experiments, the ct-DNA solution was diluted in a quartz UV-vis cell (with the optical path of 1.0 cm) with cacodylate buffer to the concentration of c = 3.0 × 10 −5 M, and the appropriate amount of the solution of 1a or 1e was added to reach the desired ratio r ((1)/(ct-DNA)) = 0.3 (containing <1% DMSO). The dependence of the absorbance at 260 nm as a function of temperature was measured on a Cary 100 Bio (Agilent Varian) UV-vis spectrometer. The temperature was varied from 25 • C to 98 • C in intervals of 0.5 • C. The denaturation temperature T m values are the midpoints of the transition curves, determined from the maximum of the first derivative. ∆T m values were calculated by subtracting the T m of the free nucleic acid from that of the respective complex, with ∆T m values being the average of at least two independent measurements and the error in ∆T m being ca. ±0.5 • C.

• Fluorescence titrations with BSA
Stock solutions of solution 1a or 1e were prepared in DMSO (c = 1.0 × 10 −3 M). For the titrations, the stock solutions were diluted in a fluorescence cell (3 mL) with cacodylate buffer (pH 7.0, 50 mM) to reach the concentration c = 1.0 × 10 −6 M (containing <1% DMSO). A stock solution of BSA in cacodylate buffer was prepared c = 1.0 × 10 −3 M. The fluorescence spectra were measured on a Cary Eclipse (Agilent Technologies) at 25 • C. The samples were excited at 320 nm, and the emission was recorded in the range 340-700 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 BSA were added to the solution and, after an incubation time of 2 min, fluorescence spectra were measured. Data obtained via fluorescence titrations were processed by means of multicomponent nonlinear regression analysis in HypSpec2014 software [93].

Photochemical alkylation of BSA
A fluorescence cell containing an aqueous solution of BSA (c = 1.4 × 10 −5 M) in cacodylate buffer (2 mL, 50 mM, pH 7.0) and 1a or 1e (c = 1.0 × 10 −6 M) in DMSO (total amount < 1%) was irradiated in a Luzchem photoreactor with 8 lamps at 350 nm over 5 min. Prior to the MS analysis, 100 µL of the irradiated solution was purified using Aspire RP30 Desalting Tips (ThermoFisher Scientific, MA, USA). After the purification, the solvent was evaporated in Speed-Vac (Eppendorf, Switzerland) and re-constituted in 10 µL of the sinapinic acid matrix (5 mg/mL of sinapinic acid in 1:1, CH 3 CN:0.1% TFA, v/v). The mixture of the analyte and the matrix was deposited onto the MALDI plate (1 µL) and dried slowly in a stream of cold air. MALDI mass spectra were recorded on an AB SCIEX MALDI-TOF/TOF 4800+ (MA, USA).

•
Laser flash photolysis (LFP) Data Availability Statement: Electronic supporting information (ESI) of this article contains synthetic procedures for the preparation of known compounds, UV-vis and fluorescence data, noncovalent binding to ct-DNA, noncovalent and covalent binding to BSA, data on antiproliferatve activity and copes of 1 H and 13 C NMR spectra of all new compounds.