Influence of supporting ligand microenvironment on the aqueous stability and visible light-induced CO-release reactivity of zinc flavonolato species

The visible light-induced CO-release reactivity of the zinc flavonolato complex [(6-Ph2TPA)Zn(3-Hfl)]ClO4 (1) has been investigated in 1 : 1 H2O : DMSO. Additionally, the effect of ligand secondary microenvironment on the aqueous stability and visible light-induced CO-release reactivity of zinc flavonolato species has been evaluated through the preparation, characterization, and examination of the photochemistry of compounds supported by chelate ligands with differing secondary appendages, [(TPA)Zn(3-Hfl)]ClO4 (3; TPA = tris-2-(pyridylmethyl)amine) and [(bnpapa)Zn(3-Hfl)]ClO4 (4; bnpapa = N,N-bis((6-neopentylamino-2-pyridyl)methyl)-N-((2-pyridyl)methyl)amine)). Compound 3 undergoes reaction in 1 : 1 H2O : DMSO resulting in the release of the free neutral flavonol. Irradiation of acetonitrile solutions of 3 and 4 at 419 nm under aerobic conditions results in quantitative, photoinduced CO-release. However, the reaction quantum yields under these conditions are lower than that exhibited by 1, with 4 exhibiting an especially low quantum yield. Overall, the results of this study indicate that positioning a zinc flavonolato moiety within a hydrophobic microenvironment is an important design strategy toward further developing such compounds as CO-release agents for use in biological systems. Graphical Abstract


Introduction
The development of carbon monoxide-releasing molecules (CORMs) for use in biological systems is of significant current interest [1][2][3][4][5][6][7]. This is due to the identification of several beneficial health effects associated with the controlled administration of small amounts of CO. These include anti-inflammatory and antiapoptotic effects, as well as the promotion of vasodilation and protection of tissues against reperfusion injury [4]. To date, the vast majority of CORMs reported are low-valent metal carbonyl compounds. These compounds, while exhibiting a number of favorable properties for biological use, also have limitations, including the fact that many cannot be controlled in terms of the location and timing of the COrelease. In this regard, metal carbonyl-based CORMs that exhibit photoinduced CO-release reactivity (photoCORMs) using low-energy visible light are of particular current interest [1,2,[8][9][10][11][12][13]. However, an important unresolved issue for these compounds is the potential for toxicity resulting from the remaining metal/ligand fragment following CO-release.
We are interested in developing new types of photoinduced CO-releasing molecules that are not based on the metal-carbonyl unit. In this regard, it is known that quercetin dioxygenases (QDOs) catalyze the oxidative breakdown of flavonols to give CO and an O-benzoylsalicylic acid (O-bs) derivative as products [14]. In studies of synthetic complexes of relevance to the enzyme/substrate adducts of QDOs, we discovered stoichiometric, UV light-induced CO-release reactivity for a family of divalent metal complexes ([(6-Ph 2 TPA) M(3-Hfl)]X, M = Zn(II), Co(II) Cu(II), Ni(II), Mn(II); X = ClO 4 − or OTf − ) containing a deprotonated 3-hydroxy-4-flavonolato ligand  and supported by a tetradentate N 4donor chelate ligand (6-Ph 2 TPA, N,N-bis((6-phenyl-2-pyridyl)methyl)-N-((2-pyridyl)methyl) amine) [15,16]. For these complexes, CO-release occurs via a photoinduced, dioxygenasetype reaction akin to the thermal reaction catalyzed by QDOs. The reaction observed for [(6-Ph 2 TPA)Zn(3-Hfl)]ClO 4 (1) in CH 3 CN under O 2 is shown in scheme 1. The quantum yield for this reaction with irradiation at 300 nm (ɸ = 0.09(1)) [16] is an order of magnitude higher than that exhibited by structurally similar Mn(II), Co(II), Ni(II), Cu(II) complexes (ɸ = 0.005-0.008) [16]. This is the result of quenching of the excited state in the complexes containing an open-shell d n metal ion. The photoinduced reaction pathway for O 2 -dependent CO-release from [(6-Ph 2 TPA)Zn(3-Hfl)]ClO 4 (1) likely involves either the coordinated flavonolato ligand acting as a photosensitizer to generate singlet oxygen ( 1 O 2 ), which can then react with ground state 1, or an excited state singlet to triplet conversion of the zincbound flavonolato anion, which may then undergo reaction with 3 O 2 .
The initial studies outlined above have led us to further examine the chemistry of zinc flavonolato species. In the results described herein, we have evaluated the visible lightinduced CO-release reactivity of 1 in 1 : 1 H 2 O : DMSO. Additionally, we have evaluated how the nature of the supporting chelate ligand, particularly with respect to the secondary microenvironment surrounding the zinc flavonolato moiety, influences the aqueous stability and visible light-induced CO-release reactivity of this unit. These combined results provide evidence that maintaining a hydrophobic microenvironment surrounding the zinc flavonolato moiety is key toward further developing such compounds for use in biological systems.

Physical methods
1 H NMR spectra were collected using a JEOL ECX-300 or Bruker ARX-400 spectrometer. Chemical shifts (in ppm) are referenced to the residual solvent peaks in CD 2 HCN ( 1 H: 1.94 (quintet) ppm). J values are given in Hz. IR spectra were recorded on a Shimadzu FTIR-8400 spectrometer as KBr pellets. UV-vis spectra were recorded at ambient temperature using a Hewlett-Packard 8453A diode array spectrophotometer. Emission spectra were collected using a Shimadzu RF-5301PC spectrofluorophotometer using a slit width of 4 nm with the excitation wavelength corresponding to the absorption maximum of the complex above 400 nm. A Rayonet photoreactor equipped with RPR-4190A lamps was used for all photochemical reactions. Quantum yields were determined using potassium ferrioxalate actinometry as a standard to measure photon flux [21][22][23]. Carbon monoxide was quantified as previously described [16]. Mass spectral data were collected at the Mass Spectrometry Facility, University of California, Riverside. Elemental analyses were performed by Atlantic Microlab, Inc., Norcross, GA using a PE2400 automatic analyzer.
Caution! Perchlorate salts of metal complexes with organic ligands are potentially explosive. Only small amounts should be prepared, and these should be handled with great care [24].   (3). Under a N 2 atmosphere, a methanol solution (2 mL) of Zn(ClO 4 ) 2 ⋅6H 2 O (40 mg, 0.11 mM) was added to solid TPA (39 mg, 0.13 mM) and the solution was stirred until the chelate ligand had dissolved. The resulting mixture was combined with a methanol solution (2 mL) containing 3-hydroxyflavone (32 mg, 0.13 mM) and Me 4 NOH⋅5H 2 O (24 mg, 0.13 mM). This mixture was stirred for 4 h at ambient temperature. The solvent was then removed under reduced pressure and the residual solid dissolved in CH 2 Cl 2 . The CH 2 Cl 2 solution was filtered through a celite/glass wool plug, and the product was precipitated via the addition of excess hexanes (40 mL

[(bnpapa)Zn(3-Hfl)]ClO 4 (4)
. Prepared using a synthetic procedure similar to that employed for the preparation of [(6-Ph 2 TPA)Zn(3-Hfl)]ClO 4 [20]. Recrystallization of the crude product via slow Et 2 O diffusion into CHCl 3 yielded yellow crystals suitable for single crystal X-ray crystallography (87 mg, 75%). Anal. Calcd for C 43    2.6. X-ray crystallography 2.6.1. Data collection and refinement. Single crystals of 3⋅CH 3 CN and 4⋅CHCl 3 were each mounted on a glass fiber using viscous oil and then transferred to a Nonius Kappa CCD diffractometer for data collection using Mo Kα radiation (λ = 0.71073 Å). Methods of unit cell refinement and determination of final cell constants have been previously reported [25]. The data collected for each compound were corrected for Lorentz polarization and absorption effects using DENZO-SMN and SCALEPAC [26]. Each structure was solved using a combination of direct methods and heavy atoms using SIR 97.
2.6.2. Structure solution. Complexes 3⋅CH 3 CN and 4⋅CHCl 3 crystallize in the space groups P2 1 /n and P-1, respectively. All non-hydrogen atoms were refined with anisotropic displacement coefficients. All hydrogen atoms were assigned isotropic displacement coefficients (U(H) = 1.2U(C) or = 1.5U(C methyl )), and their coordinates were allowed to ride using SHELXL97 [27]. In 3⋅CH 3 CN, the perchlorate anion exhibits disorder of three of the oxygen atoms. In 4⋅CHCl 3 , hydrogen-bonding interactions are found between the secondary amine NH units of O(1) of the coordinated flavonolato ligand.

Synthesis and characterization of 3 and 4
To investigate the influence of the chelate ligand on the aqueous stability and photoinduced CO-release reactivity of 1, analogs were prepared and characterized using tetradentate chelate ligands that either lacked secondary appendages (TPA) or contained hydrogen-bond donor moieties (bnpapa).  (4) were characterized by elemental analysis, X-ray crystallography, absorption and emission spectroscopy, IR, 1 H NMR, and mass spectrometry.

Elemental analysis.
The elemental analysis data for bulk samples of 3 and 4 are consistent with the dried powder form of each complex having an analytical formulation that lacks the CH 3 CN and CHCl 3 solvate molecules, respectively, that are found in the X-ray structures of these complexes (vide infra).    Table 3. Selected bond distances (Å) and angles (°) for 3·CH 3 CN and 4·CHCl 3 . a which exhibits an overall coordination number of five, a distorted square pyramidal geometry (τ = 0.35) [28], and a dissociated phenyl-appended pyridyl donor. All three complexes exhibit bidentate coordination of the flavonolato ligand, with the shorter zinc-oxygen distance in each involving the deprotonated hydroxyl donor (1⋅2CH 2 Cl 2 : 1.951 (2) [29][30][31]. Notably, both Zn-O distances elongate across the supporting chelate ligand series in the order 6-Ph 2 TPA < TPA < bnpapa. The same trend is found in the average Zn-N py distance, which increases from 2.11 Å in the 6-Ph 2 TPA-ligated complex to 2.16 and 2.17 Å in the TPA-and bnpapa-ligated complexes, respectively. This latter trend is likely due to the effect of overall coordination number, as well as the presence of the secondary amine donor appendages in the bnpapa-ligated system, which introduces additional steric hindrance.
Examination  S4 and S6, respectively). Each exhibits an isotopic cluster molecular ion that is consistent with the proposed cationic formulation.

3.2.4.4.
Absorption and emission spectra in CH 3 CN and 1 : 1 H 2 O : DMSO. Free 3-hydroxyflavone (3-HflH) exhibits absorption bands at 304 and 343 nm, respectively, when dissolved in methanol [35,36]. The lower energy feature (Band I) is assigned to a π → π* HOMO to LUMO transition, while the higher energy Band II corresponds to a HOMO-1 to LUMO transition. The absorption spectra of 3 and 4 in CH 3 CN are shown in figure 3. The Band I absorption feature for 1, 3, and 4 is red-shifted relative to the neutral flavonol, with the degree of shift depending on the nature of the supporting chelate ligand. Specifically, the most red-shifted Band I absorption is found for the 6-Ph 2 TPA-ligated complex 1 (λ max = 420 nm (21,000 M −1 cm −1 )) [16], followed by those of 3 (λ max = 415 nm (16,500 M −1 cm −1 )) and 4 (λ max = 401 nm (16, similar Stokes shift (~72 nm) to that found for 1, but with an overall lower intensity fluorescent emission. Essentially, no fluorescent emission is observed for 4 upon excitation at its absorption maximum (401 nm). Thus, the fluorescence intensity decreases in the order of 1 > 3 > 4 in CH 3 CN. We propose that this difference relates to the microenvironment surrounding the coordinated flavonolato ligand. Specifically, the presence of the phenyl appendages in the 6-Ph 2 TPA chelate in 1 results in the flavonolato ligand being positioned in a hydrophobic microenvironment. This motif limits solvent access to the flavonolato ligand relative to that in 3 thereby reducing collisional quenching. That being said, in 1 : 1 H 2 O : DMSO, the emission of 1 decreases in intensity relative to that found in CH 3 CN [figure 1(b)], suggesting that there is some solvent access to the flavonolato moiety. For complex 4, the presence of the rigid intramolecular hydrogen-bond donors, which may undergo excited state lengthening or shortening, results in significant fluorescence quenching [37].

Visible light-induced reactivity of 3 and 4
As 3 undergoes reaction in 1 : 1 H 2 O : DMSO to produce free 3-hydroxyflavone, we have only examined the visible light-induced CO-release reactivity of this compound in aerobic CH 3 CN using 419 nm irradiation. Complex 4 was also examined in aerobic CH 3 CN and was found to exhibit such low photochemical efficiency (vide infra) that further studies in 1 : 1 H 2 O : DMSO were not performed. Each reaction mixture was irradiated until judged to be complete by the disappearance of the~400-415-nm absorption band. 3.3.1. Product Identification. GC headspace gas analysis, 1 H NMR, IR, and mass spectrometry were used to characterize the products generated upon photoirradiation of 3 and 4 in CH 3 CN at 419 nm. For each complex, the photoreaction resulted in the formation of one equivalent of CO (table 1) as well as a [(L)Zn(O-bs)]ClO 4 complex (TPA: 5; bnpapa: 6). The latter complexes were identified by their 1 H NMR spectral features (S7 and S10), molecular ions in MALDI mass spectral experiments (figures S8, S9, S11, and S12), and by their characteristic ν C=O ester vibration at 1740-1770 cm −1 for the O-benzoylsalicylato ligand. Quantitative 18 O incorporation from 18 O 2 for the reactions of 3 and 4 provides conclusive evidence for a dioxygenase-type reaction. Appropriate dark and anaerobic control reactions demonstrate that the observed reactions for 3 and 4 are O 2 -dependent photoinduced processes.
3.3.2. Reaction Quantum Yields. The quantum yields for the visible light-induced COrelease reactions of 3 and 4 in CH 3 CN are shown in table 1. The structure of the supporting chelate ligand has a significant effect on the photochemical efficiency of the CO-release reactions, with the quantum yield for 1 being twofold higher than that of 3, and both of these values being >20-fold higher than the quantum yield for the reaction of 4. These results demonstrate that positioning a zinc flavonolato moiety within a rigid hydrogen-bond donor environment significantly stabilizes the compound with respect to light-induced reactivity.

Conclusions
In the studies outlined herein, we have examined the aqueous, visible light-induced COrelease reactivity properties of 1. The results indicate that this complex is stable in an aqueous, aerobic environment in the dark, but will exhibit quantitative, visible light-induced CO-release to generate a single zinc-containing product. The quantum yield for CO-release in 1 : 1 H 2 O : DMSO is a factor of two lower than that observed in organic solvent.
We have also examined how the supporting chelate ligand influences the aqueous stability and visible light-induced CO-release reactivity of zinc flavonolato species. We were particularly interested in examining the influence of the aryl appendages of the 6-Ph 2 TPA ligand in 1 as it is well known that the photochemical reactivity of organic and inorganic molecules can be modulated via encapsulation within a confined microenvironment such as a cyclodextrin cavity, hydrogen-bonded capsules, or hollow molecular structure [38,39]. Our results demonstrate that the secondary environment introduced by the tetradentate chelate ligand influences the chemistry of zinc flavonolato complexes in multiple ways. First, the presence of the hydrophobic microenvironment in 1 stabilizes the zinc flavonolato moiety with respect to water and differentiates the chemistry of 1 from that of 3 in aqueous DMSO solution. The observation of flavonol displacement from 3 in water is noteworthy not only to this study but also to investigations being performed in other laboratories wherein simple zinc flavonolato species such as [Zn(3-Hfl)(acetate)] are being evaluated in aqueous media for their antidiabetic and DNA binding properties [40][41][42]. Second, the lower quantum yields observed for photoinduced CO-release from 3 and 4 in CH 3 CN relative to 1 is likely a consequence of the enhanced accessibility of solvent to the flavonolato moiety and/or secondary hydrogen bonding. These interactions offer additional pathways for non-radiative decay thus leading to the observed excited state quenching and less efficient CO-release processes relative to 1.
Overall, these investigations revealed that the presence of a hydrophobic microenvironment is important toward providing aqueous stability and maximizing the photochemical efficiency of the CO-release reaction of a zinc flavonolato moiety. Therefore, if zinc flavonolato species are to be advanced for use as photoinduced CO-release agents in biological systems, careful consideration will need to be given to the supporting microenvironment within the compound.

Supplementary material
1 H NMR data for 1 and 3-6; MALDI-MS data for 3-6. CCDC 1006159 and 1006160 contain the supplementary crystallographic data for this paper. This data can be obtained free of charge from The Cambridge Crystallographic Data Center via www.ccdc.cam.ac.uk/ data_request/cif.