Fluorescent cyclodextrin carriers for a water soluble Zn^II pyrazinoporphyrazine octacation with photosensitizer potential

Abstract

A nitro-benzofurazan-triazolyl carboxymethylated β-cyclodextrin (NBFT-CMβCyD) and an oligomer of carboxymethyl β-cyclodextrin (sodium salt), crosslinked with epichlorohydrin and labeled with rhodaminyl groups (pβCyD-Rh), exhibit very high affinity in aqueous solution for octacationic photosensitizer [(CH₃)₈LZn]⁸⁺ neutralized by I⁻ ions (L = tetrakis-2,3-[5,6-di(2-(pyridiniumyl)pyrazino]porphyrazinato dianion). The photosensitizer (PS) forms complexes with 1 : 2 and 2 : 2 CyD:PS stoichiometry, which were characterized as binding constants and UV-Vis absorption and fluorescence properties. The self-association tendency of [(CH₃)₈LZn]⁸⁺, leading to a monomer–dimer equilibrium shift towards the dimer even at very low concentrations (≈10⁻⁶ M), is not contrasted by either NBFT-CMβCyD or the pβCyD-Rh oligomer, both of which completely convert the [(CH₃)₈LZn]⁸⁺ monomer fraction to the dimer form in the bound state. Quenching of fluorescence observed in [(CH₃)₈LZn]⁸⁺ upon binding with either hosts is consistent with the conversion of the monomer to the negligibly fluorescent dimer. The complexes formed with the CMβCyD units of the pβCyD-Rh oligomer have average association constants, which are larger by 6–7 orders of magnitude than those with the CMβCyD monomer in the NBFT-labeled derivative.

---

1. Introduction

A class of macrocyclic compounds recently attracting significant interest are phthalocyanine analogues with heterocyclic rings annulated to the pyrrole moieties of the porphyrazine core.^1,2 Their UV-Vis spectroscopic and photophysical properties are similar those of the parent phthalocyanines, whereas their redox behavior is affected by the nature of the peripheral heterocyclic rings, causing stepwise one-electron reductions at less negative potentials than the parent phthalocyanines.^3–6 Because of their unique features, such porphyrazine derivatives received considerable attention for technological, catalytic and biomedical applications.²

Several pyrazinoporphyrazines, both free-base and metal complexes, were recently synthesized and characterized.^4–6 Some of these complexes generate singlet oxygen with good yields, therefore they are promising materials for photodynamic therapy (PDT) and antimicrobial treatment.^7,8 Introduction of positively charged groups make them soluble in water mainly in the aggregated form. Multimodal anticancer potentialities emerged for some of these water soluble cationic derivatives.^9–11 While the presence of cationic functional groups favours solubilisation in water, high charge density influences cellular uptake and localization in sub-cellular compartments, not always favouring membrane penetration.^12,13 Thus, the development of suitable carriers is desirable to improve the performances of the photosensitizers, for example, reducing self-association tendency and allowing more selective and efficient accumulation in cancer tissues. The development of photosensitizer-carrier systems (“third generation photosensitizers”) currently represents one of the frontiers in the design of photosensitizers for PDT.¹⁴

In this work, we focused on the water soluble octacationic tetrakis-2,3-[5,6-di(2-(N-methyl)pyridiniumyl)pyrazino]porphyrazinato-Zn^II in the form of iodide salt ([(CH₃)₈LZn]⁸⁺, see Scheme 1A).^5,6 This molecule is an efficient singlet oxygen sensitizer and possesses a large positive charge, which can be beneficial for interaction with negatively charged phosphate groups of nucleic acids and outer membrane of Gram-negative bacteria or negatively charged fungal cell walls.^12,15 Its ability to bind to G-quadruplex and B-DNA has also been proven in solution.^7,10,11 Although aggregation has been observed in water, [(CH₃)₈LZn]⁸⁺ is present in monomeric form in non-aqueous, low-donor solvents.⁵

Scheme 1 (A) The octacation ([(CH₃)₈LZn]⁸⁺) (L = tetrakis-2,3-[5,6-di(2-(pyridiniumyl)pyrazino]porphyrazinato dianion). The cationic porphyrazine is neutralized by I⁻ ions; (B) nitro-benzofurazan-triazolyl carboxymethylated β-cyclodextrin (NBFT-CMβCyD). (C) Oligomer of carboxymethyl β-cyclodextrin sodium salt labeled with rhodaminyl groups (pβCyD-Rh), COOH groups are mainly connected to the secondary side and rhodaminyl moiety is randomly connected to the cyclodextrin unit.

We considered cyclodextrin (CyD) derivatives as complexing agents with carrier function. CyDs are biocompatible cyclic oligosaccharides composed of α-D-glucopyranose units joined by α(1–4) linkages. These complexes possess a torus-shaped structure with a hydrophilic external surface and a hydrophobic cavity capable of forming host–guest complexes. CyD-based host systems have been proposed in both clinical and preclinical studies for improved delivery of drugs¹⁶ and prodrugs, such as porphyrin photosensitizers.¹⁷ In particular, crosslinked CyD polymers can spontaneously assemble into nanoparticles with binding sites amid the 3D macromolecular network and within the CyD cavities, and enhance the apparent solubility as well as regulate the self-association of guests.^18–20 Moreover, these kind of hosts can be implemented for multimodal applications in biomedicine field. For example, an epichlorohydrin crosslinked β-CyD polymer can be made photoresponsive and applicable for both imaging and therapy by co-encapsulation of suitable chromophoric units.²¹

Considering the positive performances of β-CyD-based polymers as carriers for photoactivatable prodrugs, we have explored the binding modes of [(CH₃)₈LZn]⁸⁺ with two fluorescent β-CyD systems, a CyD monomer (Scheme 1B) and a CyD crosslinked oligomer (Scheme 1C), for optimizing the transport of cationic pyrazinoporphyrazine photosensitizers. Both the investigated CyD derivatives possess covalently attached units: (i) carboxylate groups, mainly on the CyD secondary side and (ii) fluorescent labels, e.g. a nitro-benzofurazan-triazolyl chromophore on the primary side of the CyD monomer and a rhodaminyl moiety randomly attached to the CyD oligomer. Compared to CyD-based carriers non-covalently associated to fluorescent probes, intrinsically fluorescent CyDs have a more general interest because their detection in solution and in cellular environments is free from drawbacks due to probe dissociation.²² In both the monomeric and the oligomeric CyD hosts the negatively charged carboxylate groups were expected to influence the affinity for the photosensitizer, whereas the fluorescent labels were expected to allow probing.

The complexation process was studied by performing accurate titrations with UV-Vis absorption and fluorescence monitoring. Multiwavelength spectroscopic data obtained at different host concentrations were analyzed with global methods to determine stoichiometry, formation constants and spectral features of the complexes. A further goal of the study was to test the ability of the adopted experimental and computational approach to provide information on the interaction of self-aggregating photosensitizers to the CyD systems. A detailed quantitative and qualitative picture of the binding of [(CH₃)₈LZn]⁸⁺ with the negatively charged CyDs was obtained. Both the hosts were revealed to be unsuited for the monomerization of this guest. However, the adopted methods proved to be useful for shedding light on the interaction of self-associating macrocyclic compounds with CyD systems with either monomeric or polymeric structure.

2. Experimental section

Materials

The octacation ([(CH₃)₈LZn]⁸⁺) (neutralized by I⁻ ions; L = tetrakis-2,3-[5,6-di(2-pyridiniumyl)pyrazino]porphyrazinato dianion) was synthesized as reported previously.⁵ Nitro-benzofurazan-triazolyl carboxymethylated β-cyclodextrin (NBFT-CMβCyD) was prepared according to the procedure described below. The oligomer of carboxymethyl β-cyclodextrin sodium salt crosslinked with epichlorohydrin and labeled with rhodaminyl groups (pβCyD-Rh) was obtained by reacting carboxymethyl β-cyclodextrin with 1-chloro-2,3-epoxy propane and then rhodamine B isothiocyanate, as described by Puskas et al.²³ The oligomer used has a molecular weight of 44 kDa, degree of substitution (DS) for rhodaminyl groups ca. 0.3–0.4 per oligomer, DS for COONa ca. 2.5 per CyD unit. 6-monoazido-6-monodeoxy-β-CyD was a fine chemical product from CycloLab. All the reagents and solvents were from Sigma-Aldrich. Water was purified by Millipore Milli-Q system, Millipore SpA, Milan, Italy.

Synthesis and characterization of β-CyD derivatives

6-Monodeoxy-6-monoazido-carboxy-methyl-β-CyD sodium salt (CMβCyD). Sodium chloroacetate (300 g, 2.57 mol) was dissolved in water (1 L) and 6-monoazido-6-monodeoxy-β-CyD (580 g, 0.5 mol) was gradually added to the stirred solution. The white suspension was heated at 60 °C and water (300 mL) solution of NaOH (120 g, 3 mol) was added dropwise (in a period of 2 h) while maintaining the temperature strictly under 70 °C. The reaction mixture was further heated at 70 °C for 4 h, concentrated under reduced pressure (T = 60 °C) to 1/3 of volume and then poured into EtOH (5 L) under sonication. The suspension was decanted and EtOH (1 L) was added to the sticky white material under sonication. This procedure was repeated three times, the white solid was dissolved in water (1 L), and then anionic (200 g) and cationic (20 g) exchange resins were added and the solution was stirred for 2 h. The resins were filtered, charcoal (50 g) was added to the solution, the suspension was stirred for 1 h, filtered and membrane filtered. The pH of the resulting solution was set to 8–9 with NaOH (3 M) and freeze drying yielded the product as a white powder (518 g, 74%). IR, ¹H-NMR and ¹³C-NMR features are listed below and spectra are given in ESI-1-Fig. S1.† Degree of substitution (DS) for carboxymethyl groups was 3–4.

IR (KBr) ν/cm⁻¹: 3398 (O–H), 2927 (C–H), 2104 (N₃), 1605 (COO⁻), 1419 (C–C/C–O stretch), 1327, 1246, 1161, 1085, 1031, 948, 846, 757, 710, 582, 534.

¹H-NMR (D₂O): δ 3.40–4.40 (br m, 49H, H2, H3, H4, H5, H6, CH₂–COOH), 5.05 (br s, 4H, H1), 5.27 (br s, 3H, H1).

¹³C-NMR (D₂O): δ 54.00 (C6–N3), 60.66 (C6), 70.58 (CH₂–COO⁻), 71.41, 71.61, 71.96, 73.20, 80.86, 99.66 (C1′), 101.63 (C1), 177.75 (CO), 178.91 (CO) (For more detailed assignment see HSQC-DEPT spectrum in ESI-1-Fig. S1†).

Mono-4-(N-propargyl)-7-nitro-benzofuran. The synthesis of this compound was already described.²⁴ We adopted modified conditions in the synthesis. Propargylamine (0.13 mL, 2.0 mmol) was diluted with CH₃CN (10 mL), the solution was cooled down (T = 5 °C) and 4-chloro-7-nitro-benzofurazan (200 mg, 1.0 mmol) dissolved in CH₃CN (2 mL) was added dropwise (10 min addition). The reaction mixture was stirred for 1 h, the crude product was filtrated and extensively washed with CH₃CN to yield mono-4-(N-propargyl)-7-nitro-benzofuran as a brown powder (166 mg, 76%)

¹H NMR (CDCl₃): δ = 2.43 (t, 1H), 4.29 (dd, 2H), 6.30 (broad, 1H, NH), 6.33 (d, 1H), 8.53 (d, 1H) (in accordance with literature data).²⁴

ES-MS in positive ionization mode calcd. m/z = 218.0297, found (M + H)⁺ = 219.0235 (see ESI-1-Fig. S2†).

Nitro-benzofurazan-NH-triazolyl-carboxymethyl-β-CyD (NBFT-CMβCyD). The 6-monodeoxy-6-monoazido-carboxymethyl-β-CyD sodium salt (0.74 g, 0.5 mmol, DS = 3–4 for carboxymethyl) was dissolved in 20 mL H₂O–DMF (96%) 8 : 2 (v/v); mono-4-(N-propargyl)-7-nitro-benzofuran (0.22 g, 1 mmol), CuSO₄·5H₂O (0.25 g, 1 mmol) and sodium ascorbate (0.2 g, 1 mmol) were added and the reaction mixture was stirred for 20 h at room temperature (see Scheme 2).

Scheme 2 Fluorescent labeling of 6-monodeoxy-6-monoazido-carboxymethyl-β-CyD sodium salt. Reaction scheme.

The solution was treated with both cationic and anionic ion exchange resins, the resin was filtered and the filtrate was precipitated with acetone (30 mL). The solid was dissolved in H₂O, the pH was set between 7–8 and then freeze drying yielded the final product as a pale brown solid (0.72 g, 84%). IR, ¹H-NMR, ¹³C-NMR and HR-MS features are listed below and the spectra with the detailed assignments are given in ESI-1-Fig. S3.† Degree of substitution (DS) for the fluorescent tag was ∼1.

IR (KBr) ν/cm⁻¹: 3382 (O–H), 2928 (C–H), 1601 (COO⁻), 1421 (C–C/C–O stretch), 1318, 1157, 1109, 1034, 949, 843, 758, 707, 615, 585.

¹H-NMR (D₂O): δ = 3.34–4.84 (bs, 51H, H2, H3, H4, H5, H6, O–CH₂–COOH, C–CH₂–NH), 5.06 (bs, 5H, 1H), 5.28 (bs, 2H, H1), 6.54 (bs, 1H, aromatic-H), 7.92 (s, 1H, H-triazole), 8.11 (bs, 1H, aromatic-H).

¹³C-NMR (DMSO-d₆): δ = 54.22 (C6–N), 62.04 (C6), 76.20 (CH₂–COO⁻), 79.67, 100.15 (C1), 101.38 (C1′), 123.91, 124.98, 129.44, 132.48, 138.64, 144.85, 164.16, 164.83, 165.70, 181.78 (CO).

(For more detailed assignment see HSQC-DEPT spectrum in ESI-1-Fig. S3†).

Spectroscopic measurements

All the measurements were carried out at 295 K. ¹H-, HSQC-DEPT, ¹³C- and APT-NMR spectra were recorded on a Varian VXR-600 at 400 or 600 MHz. IR spectra were recorded in KBr disk on a Nicolet 205 FTIR. MS spectra were obtained on an Agilent 6230B TOF LC/MS system. UV-Vis absorption spectra were recorded on a Perkin-Elmer Lambda 650 spectrophotometer in 1 cm quartz cuvette with water as the reference. Circular dichroism (CD) spectra were obtained with a Jasco J-715 spectropolarimeter. Each CD spectrum was registered by accumulating 3 scans with an integration time of 1 s; scan rate was 50 nm min⁻¹ for the entire range. Fluorescence spectra were obtained on an Edinburgh FLS 920 Fluorimeter (continuous 450 W Xe lamp for excitation), equipped with a Peltier-cooled Hamamatsu R928 photomultiplier tube for detection in right angle mode. Fluorescence lifetimes were measured in air-equilibrated solutions with a time correlated single photon counting system (IBH Consultants Ltd). Nanosecond LED sources at 373, 465 or 560 nm were used for excitation, and the emission was collected at right angle with suitable long pass cutoff filters. Decay profiles were analysed using a multiexponential function (eqn (1)) and deconvolution of the instrumental response. The software package was provided by IBH Consultants Ltd.

I(t) = Σ_ia_i × exp(−t/τ_i) (1)

Global analysis methods

Titration experiments with absorption or fluorescence monitoring were performed at constant concentration of either component. The best complexation model and the related association constants were determined by the global analysis of multiwavelength data sets corresponding to 10–15 spectra of different mixtures, using the SPECFIT/32 program (version 3.0.40, TgK Scientific). The program uses a multivariate optimization procedure based on singular value decomposition (SVD) and non-linear regression modeling by the Levenberg–Marquardt algorithm. The deviations of the calculated spectroscopic quantities from the experimental values are minimized in a completely numerical procedure. See ESI-3† for further details.

3. Results and discussion

Self-aggregation of [(CH₃)₈LZn]⁸⁺

The UV-visible absorption spectrum of [(CH₃)₈LZn]⁸⁺ in water (ESI-2-Fig. S4†) is characterized by a Q band system with two maxima of comparable intensity at ca. 625 and 655 nm attributed to the existence of a monomer–dimer equilibrium with log(K_d/M⁻¹) = 7.1.^10,11 The peak at the longer wavelength is assigned to the monomeric species. As already reported,¹¹ [(CH₃)₈LZn]⁸⁺ in the presence of sodium dodecyl sulfate (SDS) micelles exhibits a major peak at ca. 660 nm assigned to the monomer, which is dominant due to the association of the octacationic macrocycle to the negatively charged micelle surface. The monomerization of the octacation by the negatively charged micellar environment stimulated us to test the negatively charged cyclodextrin derivatives as carriers.

The fluorescence properties of [(CH₃)₈LZn]⁸⁺ are also strongly affected by self-aggregation. The emission spectrum is characterized by a peak at 670 nm and a shoulder at 750 nm and is assigned mainly to the monomeric species with a lifetime of ca. 2.52 ns, whereas the dimer has a subnanosecond lifetime and does not contribute appreciably to the steady state emission intensity (see Table 1).¹¹

Table 1 Fluorescence properties of NBFT-CMβCyD, pβCyD-Rh oligomer and [(CH₃)₈LZn]⁸⁺ in water, T = 22 °C

	Exc/nm	Em/nm	(a1) τ1^a/ns	(a2)τ2^a/ns	Notes
a The relative amplitude of the fluorescence component is given between parentheses.
NBFT-CMβCyD	465	539	(1.00) 1.40		χ^2 = 0.80
[(CH3)8LZn]^8+	373	670	(0.15) 0.37	(0.85) 2.52	From ref. 11
pβCyD-Rh	560	575	(0.65) 1.03	(0.35) 2.60	χ^2 = 1.09

---

Spectroscopic properties of NBFT-CMβCyD

The UV-Vis absorption, fluorescence and circular dichroism (CD) spectra of the NBFT-CMβCyD host in water are shown in Fig. 1. The absorption and emission spectra are characterized by the maxima at 467 and 537 nm, respectively. The measured fluorescence lifetime is 1.4 ns (Table 1). The CD signal observed above 240 nm and extending into the visible indicates the presence of a CyD-induced rotational strength in the lowest energy electronic transitions of the dye.

Fig. 1 NBFT-CMβCyD 4.0 × 10⁻⁵ M in water: (A) UV-Vis absorption (solid, cell path length 1 cm) and emission (dashed, λ_exc = 470 nm, uncorrected); (B) circular dichroism spectrum (200–350 nm, cell path length 2 cm, 350–560 nm, cell path length 1 cm).

Interaction of NBFT-CMβCyD with [(CH₃)₈LZn]⁸⁺

UV-Vis absorption and fluorescence. UV-Vis absorption titration in water was performed by maintaining the [(CH₃)₈LZn]⁸⁺ concentration (8 × 10⁻⁶ M) constant and varying the NBFT-CMβCyD concentration (from 1 × 10⁻⁶ M to 2 × 10⁻⁵ M). In these conditions, the initial monomer fraction of the octacation is 6%. With increasing NBFT-CMβCyD concentration, the 655 nm/625 nm peak ratio decreases and an hypochromic effect is produced in the Soret absorption region of [(CH₃)₈LZn]⁸⁺. Although the spectral variations are small, they are meaningful and clearly indicate that an interaction with the CyD derivative occurs, and the [(CH₃)₈LZn]⁸⁺ dimer is further stabilized in the bound state (Fig. 2).

Fig. 2 UV-visible absorption spectra of 8 × 10⁻⁶ M [(CH₃)₈LZn]⁸⁺ (black) upon titration with NBFT-CMβCyD in the concentration range 1 × 10⁻⁶ M (green) to 2 × 10⁻⁵ M (pink); 1.0 cm cell, 22 °C.

The same mixtures were analyzed by measuring the steady state fluorescence with the selective excitation of [(CH₃)₈LZn]⁸⁺ at 630 nm (Fig. 3A), i.e. in a fairly isosbestic region (see Fig. 2).²⁵ With increasing NBFT-CMβCyD concentrations the fluorescence intensity shows a progressive decrease. Considering the short lifetime of [(CH₃)₈LZn]⁸⁺ we can exclude quenching due to partner diffusional encounter during the lifetime of the excited molecule and attribute the decreased fluorescence intensity to quenching in a ground state complex with the CyD host. The almost complete suppression of the [(CH₃)₈LZn]⁸⁺ emission at the end of the titration is perfectly consistent with a further shift of the monomer–dimer equilibrium toward the dimer, as inferred from the evolution of the absorption spectra.

Fig. 3 Emission spectra of (A) 8 × 10⁻⁶ M [(CH₃)₈LZn]⁸⁺ upon titration with NBFT-CMβCyD in the concentration range 1 × 10⁻⁶ to 2 × 10⁻⁵ M, λ_exc = 630 nm, and (B) 5 × 10⁻⁶ M NBFT-CMβCyD upon titration with [(CH₃)₈LZn]⁸⁺ in the concentration range 5 × 10⁻⁷ to 1 × 10⁻⁵ M, λ_exc = 480 nm. T = 22 °C. Emission spectra are uncorrected. Inset (B): calculated fluorescence ε × Φ amplitudes of the free host (black) and 2 : 2 complex (green).

Emission spectra in Fig. 3B show titration of 5 × 10⁻⁶ M NBFT-CMβCyD with [(CH₃)₈LZn]⁸⁺ from 5 × 10⁻⁷ M to 1 × 10⁻⁵ M. Excitation at 480 nm was almost selective for the CyD host. The emission intensity of the NBFT chromophore decreased with the complexation progress. Diffusive quenching of the chromophore fluorescent state with lifetime 1.4 ns (see 3.3) is again excluded at the low [(CH₃)₈LZn]⁸⁺ concentrations used in this experiment.

Global analysis of the NBFT-CMβCyD–[(CH₃)₈LZn]⁸⁺ spectroscopic titrations. The set of spectra of Fig. 2 were globally analysed with SPECFIT/32 (see ESI-3† for details) in the interval 550–700 nm, where [(CH₃)₈LZn]⁸⁺ exclusively contributes to the absorption. The monomer–dimer equilibrium constant log(K_d/M⁻¹) = 7.1 and the absolute absorption spectra of the [(CH₃)₈LZn]⁸⁺ monomer and dimer in water were taken from a previous study¹¹ and were maintained constant in the calculation. The formation of two complexes with 1 : 2 and 2 : 2 NBFT-CMβCyD–[(CH₃)₈LZn]⁸⁺ stoichiometry was assessed with log(K₁₂/M⁻²) = 13.4 ± 0.5 and log(K₂₂/M⁻³) = 19.5 ± 0.5 (DW 1.4). Fig. 4A shows the absolute spectra of all the species in solution. The spectral features of the complexes clearly point to guest dimeric structure. The concentration profiles in Fig. 4B show almost complete association of the [(CH₃)₈LZn]⁸⁺ macrocycle to the CyD host in a 2 : 2 complex. The quality of the agreement between calculated and experimental absorbances at representative wavelengths is reported in ESI-3-Fig. S5.†

Fig. 4 (A) Absolute absorption spectra of [(CH₃)₈LZn]⁸⁺ monomer (red)¹¹ and [(CH₃)₈LZn]⁸⁺ dimer (green)¹¹ and of the NBFT-CMβCyD–[(CH₃)₈LZn]⁸⁺ 1 : 2 complex (black) and 2 : 2 complex (blue) for association constants log(K₁₂/M⁻²) = 13.4 and log(K₂₂/M⁻³) = 19.5; (B) concentration profiles of all the species in solution (same color code).

Analysis of the fluorescence titration data in Fig. 3A in the interval 640–800 nm was performed assuming that the only emitting species in the system is the [(CH₃)₈LZn]⁸⁺ monomer (excitation at 630 nm is selective). Similar values of the association constants were obtained, log(K₁₂/M⁻²) = 14.3 ± 0.3 and log(K₂₂/M⁻³) = 20.5 ± 0.6 (DW 2.9). We also analyzed the data of Fig. 3B, relevant to the emission intensity changes of NBFT-CMβCyD titrated with [(CH₃)₈LZn]⁸⁺, in the region 500–640 nm, where the NBFT chromophore exclusively contributes. The solutions were optically thin (A at λ_exc < 0.1 over 1 cm path), and thus a linear dependence of fluorescence signal on the concentration is operative for all the species involved. The complexation model was essentially confirmed with log(K₁₂/M⁻²) = 14.2 ± 0.08 and log(K₂₂/M⁻³) = 20.5 fixed (DW 1.2). To obtain convergence it was necessary to assume that the emission of the NBFT chromophore is completely quenched in the 1 : 2 complex (i.e. this complex does not contribute to total intensity, see ESI-3†), whereas emission survives in the 2 : 2 complex. According to the results of the calculations, the molar emission contribution of the 2 : 2 complex is ca. 52% that of the free host (from the integrated relative amplitudes represented in the inset of Fig. 3B, each proportional to ε × Φ with ε molar absorption coefficient at the excitation wavelength). Considering that the ε value is almost double in the 2 : 2 complex the average Φ is ca. 26% that of the free NBFT-CMβCyD host. The emission decay measured with excitation at 465 nm in a sample of NBFT-CMβCyD ca. 95% bound in the 2 : 2 complex (2 × 10⁻⁵ M NBFT-CMβCyD in presence of 8 × 10⁻⁶ M [(CH₃)₈LZn]⁸⁺) is monoexponential with a lifetime of ∼1.4 ns, the same as that of the free host. This fact indicates that, as expected, one of the two NBFT chromophores is strongly quenched whereas the other one is fairly unperturbed.

Interaction of pβCyD-Rh with [(CH₃)₈LZn]⁸⁺

UV-Vis absorption and fluorescence. The absorption spectrum of 3.25 × 10⁻⁶ M pβCyD-Rh in water reveals the occurrence of aggregation in the solution. Considering one rhodaminyl group every three oligomers we get a molar absorption coefficient ε = 55 383 M⁻¹ cm⁻¹ at 557 nm for the rhodaminyl chromophore, which is quite low when compared to that of a free rhodamine species. This clearly indicates that the oligomer is aggregated. The emission of the rhodaminyl moiety in pβCyD-Rh has a biexponential decay (Table 1) with lifetimes of 1.0 ns (60%) and 2.6 ns (40%). The two components are assigned either to monomeric and dimeric rhodaminyl species or two different environments for rhodaminyl groups in the oligomer frame

Titration of 2 × 10⁻⁶ M [(CH₃)₈LZn]⁸⁺ (14% as monomer) in water with increasing amounts of pβCyD-Rh from 1.6 × 10⁻⁸ M to 8 × 10⁻⁷ M leads to changes in the absorption intensity and profile. Two main phases are observed: a progressive decrease of the [(CH₃)₈LZn]⁸⁺ absorption bands (both the Soret and the Q-band system) up to 6.4 × 10⁻⁸ M oligomer; at higher oligomer concentrations a subsequent increase of the absorption, reaching a maximum at ca. 4.8 × 10⁻⁷ M (Fig. 5). Progressive increase in the ratio of the [(CH₃)₈LZn]⁸⁺ 625 nm peak to the 655 nm peak on increasing the oligomer concentration points to an increase of the dimer fraction in the bound state.

Fig. 5 UV-visible absorption titration of 2 × 10⁻⁶ M [(CH₃)₈LZn]⁸⁺ with pβCyD-Rh (from 1.6 × 10⁻⁸ M to 8 × 10⁻⁷ M) in water, cell path length 1 cm, 22 °C, water as reference.

The fluorescence spectra of the same mixtures with excitation at 630 nm, selective for [(CH₃)₈LZn]⁸⁺, are shown in Fig. 6. The emission of the [(CH₃)₈LZn]⁸⁺ monomer is progressively quenched by the oligomer and almost completely disappears at 8 × 10⁻⁸ M pβCyD-Rh. The emission of the rhodaminyl pendant on selective excitation at 530 nm is also partially quenched by [(CH₃)₈LZn]⁸⁺. Fig. 7A shows the effect of [(CH₃)₈LZn]⁸⁺ on pβCyD-Rh at various concentrations. The quenching shows two phases, reasonably assigned to the progress of two different complexation steps (Fig. 7B).

Fig. 6 Emission spectra of 2 × 10⁻⁶ M [(CH₃)₈LZn]⁸⁺ upon titration with pβCyD-Rh (from 1.6 × 10⁻⁸ M to 8 × 10⁻⁷ M) in water, λ_exc = 630 nm and T = 22 °C. Spectra are uncorrected.

Fig. 7 (A) Quenching of the fluorescence intensity of pβCyD-Rh at various concentrations in water by 2 × 10⁻⁶ M [(CH₃)₈LZn]⁸⁺, λ_exc = 530 nm, T = 22 °C. Emission spectra are uncorrected. (B) Dependence of the rhodaminyl fluorescence intensity on the concentration of pβCyD-Rh alone (black) and in presence of 2 × 10⁻⁶ M [(CH₃)₈LZn]⁸⁺ (red).

Global analysis of the pβCyD-Rh–[(CH₃)₈LZn]⁸⁺spectroscopic titrations. The best equilibrium model and the association constants were determined by the global analysis of the absorption data of Fig. 5 in the interval 550–700 nm, expressing the host concentration in terms of CMβCyD units. The monomer–dimer equilibrium constant in water (log(K_d/M⁻¹) = 7.1) and the absorption spectra of the [(CH₃)₈LZn]⁸⁺ monomer and dimer¹¹ were fixed. The absorption spectrum of the rhodaminyl moiety was also fixed, averaged on the concentration of CMβCyD units (we recall that the CMβCyD content of the oligomer is 60% and on average every third oligomer molecule contains one rhodaminyl group). Formation of two complexes with 1 : 2 and 2 : 2 CMβCyD–[(CH₃)₈LZn]⁸⁺ stoichiometry and log(K₁₂/M⁻²) = 20.7 ± 0.2 and log(K₂₂/M⁻³) ≈ 27 was found to reasonably reproduce the absorption data. The 2 : 2 association constant requires to be fixed in the calculation but only pK₂₂ ≈ 27 allows convergence. Fig. 8A and B show the absolute spectra and the concentration profiles of all the species.

Fig. 8 (A) Absolute spectra of [(CH₃)₈LZn]⁸⁺ monomer (red) and dimer (green),¹¹ and CMβCyD–[(CH₃)₈LZn]⁸⁺ 1 : 2 (black) and 2 : 2 (blue) complexes in water. See text for the association constants. (B) Concentration profiles of all the species in solution.

Global analysis of the data in Fig. 6 assuming that [(CH₃)₈LZn]⁸⁺ monomer is the only emitting species, confirmed the complexation model with 1 : 2 and 2 : 2 CMβCyD–[(CH₃)₈LZn]⁸⁺ stoichiometry and the association constants as log(K₁₂/M⁻²) = 18.1 ± 1.7 and log(K₂₂/M⁻³) ≈ 26.7 ± 2 (DW 2.3). As noticed with the NBFT-CMβCyD host, the absorption spectra of the complexes in Fig. 8A clearly show the dimeric structure for [(CH₃)₈LZn]⁸⁺ in the bound state. Almost complete association of dimeric pyrazinoporphyrazine occurs with either one or two negatively charged CMβCyD units of pβCyD-Rh.

4. Conclusions

The combination of spectroscopic titrations and global analysis of multiwavelength data has provided a reliable description of the interaction of a self-associating porphyrazine macrocycle with CyD systems of both monomeric and polymeric structure. This approach has a good potential as an analytical and characterization tool for the investigation and optimization of new third generation macrocyclic photosensitizers, and in general of host–guest systems involving complex equilibria.

The [(CH₃)₈LZn]⁸⁺ guest binds almost exclusively as a dimer to both NBFT-CMβCyD and pβCyD-Rh in water. CyD–guest complexes with 1 : 2 and 2 : 2 stoichiometry are formed with both the host systems. The pyrazinoporphyrazine macrocycle did not penetrate deeply into the CyD cavity because of its large size and peripheral positive charges. The stabilization of the [(CH₃)₈LZn]⁸⁺ dimer by the CyD units is reasonably attributed to the interaction with the negative carboxylate groups, capable of determining complete conversion of the [(CH₃)₈LZn]⁸⁺ free monomer fraction to the dimer state in the bound condition. Indeed the width of the macrocycle and the diameter of the cyclodextrin secondary face are very similar to each other. Therefore, it can be hypothesized that the distribution of the positive charges in the two [(CH₃)₈LZn]⁸⁺ units of the dimer and of the carboxylate negative charges of the CMβCyD units allow the establishment of multiple host–guest ionic bonds in the complexes, strongly favouring 1 : 2 and 2 : 2 structures. Because the [(CH₃)₈LZn]⁸⁺ dimer is negligibly fluorescent, host-induced dimerization is the most likely mechanism for the [(CH₃)₈LZn]⁸⁺ fluorescence quenching in the presence of either NBFT-CMβCyD or the pβCyD-Rh oligomer. At low CMβCyD concentrations one unit binds to the [(CH₃)₈LZn]⁸⁺ in dimeric form. At higher concentrations one more CyD unit adds, contributing to the establishment of a final population with the 2 : 2 complex as largely predominant species.

Both 1 : 2 and 2 : 2 complexes with the CMβCyD units of the pβCyD-Rh oligomer have average stability constants by 6–7 orders of magnitude larger than those with the monomeric CMβCyD moiety of the NBFT labelled derivative. The different chemical nature of the fluorescent label is not expected to play a key role. In contrast, according to previous findings relevant to the binding affinities of monomeric and polymeric CyD hosts vs. the same guest,²⁰ the high local CyD concentration and the 3D spatial organization of the oligomer are the most likely reasons for such a large increase in the [(CH₃)₈LZn]⁸⁺ binding constants. This is a positive outcome for the application of similar CyD oligomers as carriers of highly charged singlet oxygen photosensitizers. From a biological point of view, e.g. for cell penetration, we reasonably expect a more favourable behaviour of the oligomer complexed [(CH₃)₈LZn]⁸⁺, partly neutralized, compared to the isolated octacation. Unfortunately, the carrier is not capable of disrupting the [(CH₃)₈LZn]⁸⁺ dimer in aqueous environment, which could represent a drawback in the frame of singlet oxygen sensitization. This behaviour differs from that of other CyD derivatives with polymeric structure, either neutral or negatively charged, which are able to contrast, for example, the self-association of doxorubicin in neutral buffer.^19,20 However, we envisage that biological actors can promote the monomerization of [(CH₃)₈LZn]⁸⁺ once the photosensitizer was vehicled inside the cell. In this context it is worth recalling that the [(CH₃)₈LZn]⁸⁺ dimer is partially disrupted upon binding to G4-DNA.¹⁰

Acknowledgements

We thank Professor Claudio Ercolani (University of Rome “La Sapienza”) for helpful discussions and the FP7 PEOPLE-ITN project no. 237962-CYCLON for the funding of the research. We also thank CNR project NANOMAX for financial support.

References

1. P. A. Stuzhin and C. Ercolani, in The Porphyrin Handbook; Phthalocyanines: Synthesis, ed. K. M. Kadish, K. M. Smith and R. Guilard, Academic Press, New York, 2003, vol. 15, pp. 263–364; M. S. Rodriguez-Morgade and P. A. Stuzhin, J. Porphyrins Phthalocyanines, 2004, 8, 1129–1165; T. Fukuda and N. Kobayashi, Dalton Trans., 2008, 4685–4704.
2. M. J. Fuchter, C. Zhong, H. Zong, B. M. Hoffman and A. G. M. Barrett, Aust. J. Chem., 2008, 61, 235–255.
3. E. J. Baerends, G. Ricciardi, A. Rosa and S. J. A. van Gisbergen, Coord. Chem. Rev., 2002, 230, 5–27; M. P. Donzello, D. Dini, G. D'Arcangelo, C. Ercolani, R. Q. Zhan, Z. P. Ou, P. A. Stuzhin and K. M. Kadish, J. Am. Chem. Soc., 2003, 125, 14190–14204; M. P. Donzello, R. Agostinetto, S. S. Ivanova, M. Fujimori, Y. Suzuki, H. Yoshikawa, J. Shen, K. Awaga, C. Ercolani, K. M. Kadish and P. A. Stuzhin, Inorg. Chem., 2005, 44, 8539–8551.
4. M. P. Donzello, Z. P. Ou, D. Dini, M. Meneghetti, C. Ercolani and K. M. Kadish, Inorg. Chem., 2004, 43, 8637–8648; M. P. Donzello, Z. P. Ou, F. Monacelli, G. Ricciardi, C. Rizzoli, C. Ercolani and K. M. Kadish, Inorg. Chem., 2004, 43, 8626–8636.
5. C. Bergami, M. P. Donzello, C. Ercolani, F. Monacelli, K. M. Kadish and C. Rizzoli, Inorg. Chem., 2005, 44, 9852–9861.
6. C. Bergami, M. P. Donzello, F. Monacelli, C. Ercolani and K. M. Kadish, Inorg. Chem., 2005, 44, 9862–9873.
7. M. P. Donzello, D. Vittori, E. Viola, I. Manet, L. Mannina, L. Cellai, S. Monti and C. Ercolani, Inorg. Chem., 2011, 50, 7391–7402.
8. M. P. Donzello, E. Viola, M. Giustini, C. Ercolani and F. Monacelli, Dalton Trans., 2012, 41, 6112–6121.
9. I. Manet, F. Manoli, M. P. Donzello, C. Ercolani, D. Vittori, L. Cellai, A. Masi and S. Monti, Inorg. Chem., 2011, 50, 7403–7411.
10. I. Manet, F. Manoli, M. P. Donzello, E. Viola, G. Andreano, A. Masi, L. Cellai and S. Monti, Org. Biomol. Chem., 2011, 9, 684–688.
11. I. Manet, F. Manoli, M. P. Donzello, E. Viola, A. Masi, G. Andreano, G. Ricciardi, A. Rosa, L. Cellai, C. Ercolani and S. Monti, Inorg. Chem., 2013, 52, 321–328.
12. T. J. Jensen, M. G. a. H. Vicente, R. Luguya, J. Norton, F. R. Fronczek and K. M. Smith, J. Photochem. Photobiol., B, 2010, 100, 100–111.
13. D. Kessel, R. Luguya and M. G. a. H. Vicente, Photochem. Photobiol., 2003, 78, 431–435.
14. A. E. O'Connor, W. M. Gallagher and A. T. Byrne, Photochem. Photobiol., 2009, 85, 1053–1074.
15. Y.-Y. Huang, P. Mroz, T. Zhiyentayev, S. K. Sharma, T. Balasubramanian, C. RuzieÌ, M. Krayer, D. Fan, K. E. Borbas, E. Yang, H. L. Kee, C. Kirmaier, J. R. Diers, D. F. Bocian, D. Holten, J. S. Lindsey and M. R. Hamblin, J. Med. Chem., 2010, 53, 4018–4027.
16. M. E. Davis and M. E. Brewster, Nat. Rev. Drug Discovery, 2004, 3, 1023–1035; T. Loftsson and M. E. Brewster, J. Pharm. Sci., 1996, 85, 1017–1025; M. E. Brewster and T. Loftsson, Adv. Drug Delivery Rev., 2007, 59, 645–666; F. van de Manakker, T. Vermonden, C. F. van Nostrum and W. E. Hennink, Biomacromolecules, 2009, 10, 3157–3175; A. Ueno, Supramol. Sci., 1996, 3, 31–36.
17. A. Mazzaglia, N. Micali, L. M. Scolaro, M. T. Sciortino, S. Sortino and V. Villari, J. Porphyrins Phthalocyanines, 2010, 14, 661–677; N. Kandoth, E. Vittorino, M. T. Sciortino, T. Parisi, I. Colao, A. Mazzaglia and S. Sortino, Chem.–Eur. J., 2012, 18, 1684–1690.
18. E. Renard, A. Deratani, G. Volet and B. Sebille, Eur. Polym. J., 1997, 33, 49–57; R. Gref, C. Amiel, K. Molinard, S. Daoud-Mahammed, B. Sebille, B. Gillet, J. C. Beloeil, C. Ringard, V. Rosilio, J. Poupaert and P. Couvreur, J. Controlled Release, 2006, 111, 316–324; S. Daoud-Mahammed, P. Couvreur, K. Bouchemal, M. Cheron, G. Lebas, C. Amiel and R. Gref, Biomacromolecules, 2009, 10, 547–554.
19. R. Anand, F. Manoli, I. Manet, S. Daoud-Mahammed, V. Agostoni, R. Gref and S. Monti, Photochem. Photobiol. Sci., 2012, 11, 1285–1292.
20. R. Anand, M. Malanga, I. Manet, F. Manoli, K. Tuza, A. Aykaç, L. C. E. Fenyvesi, A. Vargas-Berenguel, R. Gref and S. Monti, Photochem. Photobiol. Sci., 2013, 12, 1841–1854.
21. A. Fraix, N. Kandoth, I. Manet, V. Cardile, A. C. E. Graziano, R. Gref and S. Sortino, Chem. Commun., 2013, 49, 4459–4461; E. Deniz, N. Kandoth, A. Fraix, V. Cardile, A. C. E. Graziano, D. Lo Furno, R. Gref, F. M. Raymo and S. Sortino, Chem.–Eur. J., 2012, 18, 15782–15787.
22. Y. F. He, P. Fu, X. H. Shen and H. C. Gao, Micron, 2008, 39, 495–516; R. N. Dsouza, U. Pischel and W. M. Nau, Chem. Rev., 2011, 111, 7941–7980.
23. I. Puskas, A. Szemjonov, E. Fenyvesi, M. Malanga and L. Szente, Carbohydr. Polym., 2013, 94, 124–128.
24. N. W. McGill and S. J. Williams, J. Org. Chem., 2009, 74, 9388–9398.
25. The absorbance of the [(CH₃)₈LZn]⁸⁺ solution at 630 nm in the experiment of Fig. 3A was ca. 0.4 over 1 cm. This high value did not introduce error, because it is kept fairly constant in the course of the titration (see Fig. 2).

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c3ra47034b

---