Self-assembly of a helical zinc-europium complex: speciation in aqueous solution and luminescence

Two new tridentate(NNO)-bidentate(NN) compartmental ligands, HL5 and HL6, are synthesized from pyridine and benzimidazole synthons. They react in aqueous solution under physiological conditions with ZnII, LnIII, or a mixture thereof, to yield complexes of different stoichiometries, 1:3, 2:2, 2:3, 1:1:3, the speciation of which is established by UV-visible titrations and ESI mass spectrometry. Photophysical studies of the EuIII-containing solutions in Tris-HCl 0.1 M (pH = 7.4) show that lanthanide luminescence arises from a unique N6O3 coordination site with pseudo D3 symmetry. Relevant parameters such as crystal field splitting, lifetime, radiative lifetime, and intrinsic quantum yield perfectly match those reported for dinuclear 4f-4f helicates in which the EuIII ion has the same coordination environment.


INTRODUCTION
Helical structures have attracted chemists' attention when Linus Pauling published a seminal series of papers at the beginning of the 1950's dealing with the secondary structures of proteins induced by three-dimensional helical patterns (Pauling et al., 1951). The demonstration that helical structures can also be engineered at the molecular level by taking advantage of stereochemical properties of metal ions had to wait until 1987 when Jean-Marie Lehn isolated 3:2 Cu I :L double-stranded helical complexes that he named helicates, where L is an oligo-bipyridine ligand (Lehn et al., 1987). A few years later, Claude Piguet applied the same concept to trivalent lanthanide ions and successfully self-assembled the first Ln III dinuclear triple-stranded helicate, [Eu 2 (L A ) 3 ] 6+ (Scheme 1), the crystal structure of which evidences a stabilization of the molecular architecture by π-π stacking interactions between the ligand strands . The two 9-coordinate metal ions lie on a pseudo-C 3 axis of symmetry (Figure 1); in solution the average symmetry of the edifice is D 3 on NMR time scale (Piguet et al., 1993). This initial work paved the way for the development of several series of lanthanide polynuclear and polymetallic complexes including heterobimetallic nd-4f (Piguet et al., 1995a) and 4f-4f (André et al., 2004) chelates, as well as tri-and tetranuclear homometallic and heterometallic entities (Piguet et al., 2000;Piguet and Bünzli, 2010). Interestingly, the helicates are quite stable in solution despite large Coulomb repulsion between two neighboring cations which lie about 9 Å apart; careful thermodynamic considerations for 3d-4f and 4f-4f-4f helicates indeed show that the cation-cation repulsive energy (≈700 kJ·mol −1 ) is largely compensated by favorable solvation energy (Canard and Piguet, 2007). Furthermore, soluble helicates [Ln 2 (L C1 ) 3 ] can be assembled in water and are highly stable, with logβ 23 on the order of 26-30 (Elhabiri et al., 1998). Crystal structures revealed triplestranded helicates with 9-coordinate metal ions (Ln = Eu, Tb) well-imbedded into the edifice and displaying interesting photophysical properties (Elhabiri et al., 1998;Gonçalves e Silva et al., 2002). Subsequent molecular engineering led to the series of the more water-soluble [Ln 2 (L C2 ) 3 ] helicates and their bioconjugates which proved to be adequate luminescent bioprobes for live cell staining Chauvin et al., 2013) and for specific detection of biomarkers expressed by cancerous cells (Fernandez-Moreira et al., 2010).
One fascinating aspect of the polymetallic helical molecular edifices is the possibility of controlling the optical and/or magnetic properties of one ion by the other, through communication along the pseudo C 3 axis. Examples are the tuning of the spincrossover temperature in [FeLn(L 1,2 ) 3 ] 5+ (Piguet et al., 1995b;Edder et al., 2000Edder et al., , 2001, (Scheme 1) or the lengthening of the excited state lifetimes of Nd III and Yb III in [CrLn(L 1 ) 3 ] 6+ (Torelli et al., 2005). Such tunability would be of great help in the design of specific biosensors and stains, especially that [EuZn(L 2 ) 3 ] 5+ proved to be quite luminescent in water with a quantum yield of 15% (Edder et al., 1997;Piguet and Bünzli, 2010). Bioprobes need to be water soluble and amenable to bioconjugation; unfortunately, helicates with the carboxylic acid derivatives HL 3 and HL 4 do not show enough water solubility for this purpose. In this paper, we apply to HL 3 the successful strategy used in going from H 2 L C1 to H 2 L C2 in the hope of gaining access to luminescent and soluble 3d-4f helicates with ligands HL 5 and HL 6 . More specifically, and as a first step toward engineering bioprobes based

LIGAND SYNTHESIS
The underlying principle of the synthesis of HL 5 and HL 6 is the same as the one adopted for preparing ligands L 1, 2 (Edder et al., 2000), and HL 3 (Edder et al., 1997), namely a multistep strategy based on a modified Phillips reaction for the formation of the benzimidazole rings. However, the diethylamino groups are replaced by 2-[2-(2-methoxyethoxy)ethoxy]-N-{2-[2-(2-methoxyethoxy)ethoxy]ethyl}-ethanamino groups. Following our previous work (Deiters et al., 2009), the latter have been grafted on the key intermediate (6), the synthesis of which is depicted on Scheme 2 while the routes for accessing ligands HL 5 and HL 6 are summarized on Scheme 3. Regarding sulfonation, the absence of directional electronic effects favoring electrophilic substitution in the ditopic ligands and the large number of carbon atoms amenable to sulfonation implies that the corresponding group has to be inserted in one of the starting building blocks, namely (3).
Intermediate (6) was prepared in 7 steps and 23% yield from commercially available 2-picoline, benzylamine and triethylene glycol monomethyl ether (TEGOMe). The first two steps involve the synthesis of 2-[2-(2-methoxyethoxy)ethoxy]-N-{2-[2-(2-methoxyethoxy)-ethoxy]ethyl}ethanamine (2): two TEGOMe arms are grafted on the primary benzylamine by reaction with BrTEGOMe in presence of a weak base (reaction i, Scheme 2) followed by selective and quantitative cleavage of the benzylamino group by continuous flow hydrogenation (reaction ii). Two other steps are necessary for preparing (3) from 2-picoline by treatment with oleum and subsequent oxidation with permanganate, according to a previously described procedure (Delarge, 1965). The sulfonate function of (3) is then activated by chlorination (reaction iii), in presence of a PCl 5 /POCl 3 mixture to give the corresponding sulfonyl chloride (4). Coupling between (2) and (4) is subsequently achieved under standard conditions (reaction iv) to yield sulfonamide (5). Finally, oxidation of the methyl group of (5) into a carboxy group conducted in presence of an excess of SeO 2 in refluxing pyridine affords synthon (6) in almost quantitative yield (reaction v).

SPECIATION IN SOLUTION
Conditional stability constants of both homometallic (M = Zn II , La III , Eu III , Lu III ) and heterometallic (M 1 = Zn II , M 2 = La III , Eu III , Lu III ) complexes have been determined in Tris-HCl 0.1 M (pH 7.4) at 295 K by spectrophotometric titrations of the ligands (1.43 × 10 −5 M for HL 5 and 1.62 × 10 −5 M for HL 6 , corresponding to an absorbance of about 0.5) with concentrated solutions of the metal perchlorates: 2.5-5.0 × 10 −3 M for homometallic titrations and 2 × 10 −4 M for each cation in the case of heterometallic titrations, in view of the poorer solubility of the hetero species. Titrations were performed batch wise for 20-25 [M] t /[HL i ] t (i = 5.6) ratios ranging between 0 and 2 (0 and 4 for Eu III + Zn II , respectively). The first attempts to determine the speciation of the complexes revealed unreliable due to the slow kinetic of formation. Therefore, the time required to reach equilibrium was determined by luminescence spectroscopy, monitoring both the ligand fluorescence (Zn II -HL 6 and Gd III -HL 6 systems) and the phosphorescence of the Eu III ion (Eu III -HL 6 and Eu III -Zn II -HL 6 systems). In the case of the lanthanide solutions with 1:3 and 2:3 Ln:HL 6 ratios, steady states were reached within times not exceeding 1 h at room temperature. On the other hand, equilibrium times could be estimated at about 5-6 days at 313 K for heterometallic solutions. Consequently, solutions were carefully equilibrated before recording spectra, taking these data into consideration: 2 h at room temperature for homometallic solutions and 7 days at 313 K for heterometallic mixtures. These observations are interesting because in previous works, the kinetics of formation of homodinuclear 4f-4f helicates with a ligand similar to HL C1 but with carboxylic acid groups replaced with diethylamide groups (Hamacek et al., 2003) or with a bis(8-hydroxyquinolinate) ligand (Comby et al., 2009) was found to be fast in acetonitrile, equilibrium being reached within minutes. Similarly, [Eu 2 (L C1 ) 3 ] forms within 10 min in water at pH 6.15 . If the observed kinetics for the 1:3 and 2:3 Ln:HL 6 complexes is not too different from the latter observation, formation of the Zn II Eu III helicate is about two orders of magnitude slower.
In the case of ligand HL 5 , the spectra corresponding to the titration with lanthanum could be analyzed with a model including [La(L 5 ) n ] (3 − n)+ (n = 1, 2, 3) and [La 2 (L 5 ) 3 ] 3+ species: the corresponding logβ 1n and logβ 23 being 8.4(3), 14.2(5), 20.4(4), and 29.2(6), pointing to the formation of a very stable helicate, despite the mismatch between the ligand denticity (3 + 2) and the coordination number requirements of La III . On the other hand, data with lutetium gave a less satisfying fit. Moreover, when the ligand was titrated with zinc ions or with an equimolar mixture of zinc and lanthanide ions partial precipitation occurred and the residual absorbance at the end of the titration was markedly smaller than at the beginning. For this reason, no further experiments have been conducted with HL 5 and we have concentrated our efforts on ligand HL 6 .
For the titrations of HL 6 with Zn II and Eu III , factor analysis pointed to the presence of 3-4 species in solution, including the free ligand. However, the fitting procedure was not straightforward and several models were tested. The best convergence and smallest residuals were obtained for the following sets of equilibria (charges are omitted for clarity): Recalculated spectra are heavily correlated (Figures S1, S2, Supplementary Material), which explains the difficulties in the fitting procedure. The corresponding conditional stability constants are listed in Table 1. In the case of Zn II , the main species at 2:3 Zn:L stoichiometric ratio is the dinuclear complex (Figure 2, top) while the 1:3 complex remains a minor species (maximum speciation: 21% at R = 0.20); when R further increases, the 2:3 complex transforms into a 2:2 species. This behavior is in line with our previous results (Piguet and Bünzli, 2010). The tridentate-bidentate compartmental ligand HL 6 is not well-suited for building triple-stranded helicates with Ln III ions (two tridentate coordination units would be required) and this is seen in the corresponding speciation diagram: the dominant species is a 1:3 complex, with a 75% speciation for R = 0.33 (Figure 2, middle). For this ratio, only a small quantity of the 2:3 species is present (8%). Absorbance values extracted at different wavelengths for the titrations with Zn II and Eu III (Figures S1, S2, Supplementary Material) are compatible with the initial formation of 1:3 species. The titration with a mixture of metal ions has been conducted in a slightly different way, due to reduced solubility of the formed products, the concentration of each metal ion has been set to 2.00 × 10 −4 M only. Again fit of the data was difficult in view of the correlated spectra ( Figure S3, Supplementary Material), so that the extracted data and corresponding discussion have to be taken with care. Indeed, logβ 22 for [Zn 2 (L 6 ) 2 ] 2+ extracted from this titration amounts to 18.7(3) whereas a value of 21.7(4) was found from the homometallic titration. We think, however, that the salient features are correct: contrary to what was expected, and found for other Zn II -Ln III helicates in acetonitrile, for instance with L 2 (Edder et al., 2000), the 1:1:3 species is not the dominant one, accounting for only 38% of the speciation at the 1:1:3 stoichiometric ratio. Another species is present in sizeable quantity (18%), namely the Zn II 2:2 complex which is less stable than the 1:1:3 species by less than two orders of magnitude. So it seems there is competition between Zn II and Eu III for the tridentate coordination unit of (L 6 ) − . This competition is further demonstrated by an experiment in which Zn II was added to a 1:3 Eu:(L 6 ) − stoichiometric solution 5.4 μM in Eu III up to an Eu:Zn ratio equal to 1. The Eu III luminescence intensity clearly decreases while ligand fluorescence centered at 450 nm increases ( Figures S4, S5, Supplementary Material). Moreover, the ES-MS spectra discussed below point to other species being present in solution and the low solubility exhibited by this mixed system could well reflect the formation of polymeric (hydroxide?) species as well.
In order to substantiate the speciation determined by UVvisible titrations and, also, to determine if lighter and heavier lanthanides would lead to the same species in solution, ES-MS  Corresponding stability constants are listed in Table 1. spectra of 1:1:3 Zn:Ln:(L 6 ) − solutions in acetonitrile:water (1:1) containing 1% of formic acid and with total ligand concentration equal to 3 mM have been recorded for Ln = Nd and Yb.
The main observed peaks are listed in Table 2. The findings indeed partly corroborate those from UV-visible titrations.
For the Nd-containing sample, the [ZnNd(L 6 ) 3 ] 2+ complex is detected as a quadruple charged (+2H + ) species, along with [Nd(L 6 ) 3 ] and [Zn(L 6 ) 2 ], the latter giving rise to several solvated species and/or adducts with sodium and potassium. Spectra for the Yb-containing samples are simpler; here again, the 1:1:3 species is detected through a peak with a sizeable intensity the high-resolution scan of which matches well the calculated isotopic distribution (Figure 3). As for neodymium, an ytterbium 1:3 species is present, as well as the 1:2 zinc complex. In both cases, no 2:2 zinc complex was identified though, contrary to UV-visible titration data; we note, however that the solvent is different and that the conditions in the spectrometer may lead to dissociation of this species.

PHOTOPHYSICAL PROPERTIES OF THE SOLUTIONS
Absorption spectra of ligand (L 6 ) − and various solutions containing Ln III ions (Ln = Eu, Gd) or an equimolar Eu III /Zn II mixture are reported on Figure 4. The ligand absorption band at 319 nm can be assigned to a π → π * transition involving intramolecular electron transfer from the benzimidazole units to the pyridine and carboxylic groups. This band is red-shifted to 326.5-327 nm in the solutions containing Ln III ions only while the presence of Zn II results in a slightly larger shift, to 329.5 nm. The molar absorption coefficients of the 1:3 solutions are, within experimental errors, equal to three times the molar absorption coefficient of the free ligand, while they are marginally smaller (−3.5%) for the 2:3 and 1:1:3 solutions ( Table 3).
Upon excitation into the 319-nm absorption band, ligand fluorescence emission is seen as a broad feature with maximum at 466 nm ( Figure 5) and the corresponding excitation spectrum matches the absorption spectrum. At 77 K and upon enforcing a 50-μs delay time, weak phosphorescence is detected with a maximum at 506 nm. For the 1:3 Eu III solution, fluorescence of the ligand is still seen, representing 37% of the total emission of the sample; this is consistent with the fact that the solution contains about 17% of free ligand (Figure 2, middle). In addition characteristic f-f emission from the Eu( 5 D 0 ) level is detected. A striking feature is that this spectrum is quite typical of a species with pseudo D 3 symmetry and is quasi identical to the one recorded for the [Eu 2 (L C2 ) 3 ] helicate . In particular, the branching ratios expressed with respect to the intensity of the magnetic dipole transition, I( 5 D 0 → 7 F J )/I( 5 D 0 → 7 F 1 ) are very similar for the two samples (data for [Eu 2 (L C2 ) 3 ] are between parentheses): 0.02 (0.01), 1.00 (1.00), 0.87 (0.95), 0.16 (0.13), 1.62 (1.72), and 0.05 (n.a.). The splitting of the 5 D 0 → 7 F 1 transition is also very similar, 170 vs. 161 cm −1 . These data point to luminescence arising from a coordination environment made up of 3 NNO moieties and very similar to the sites in [Eu 2 (L C2 ) 3 ]; if some coordination were to occur through the bidentate site, then the coordination sphere would be completed by water molecules, leading to a poorly luminescent species. The solution also contains 8% of the 2:3 species, featuring two different metal ion sites (NNO) 3 and (NN) 3 ; the first one will give a spectrum identical to the one of the 1:3 complex, while the second will be poorly luminescent and therefore its contribution to the emission spectrum can be neglected. As an additional proof, the decay curve of the Eu( 5 D 0 ) luminescence is perfectly monoexponential, with a lifetime of 2.7 ms (2.4 ms for [Eu 2 (L C2 ) 3 ]), confirming that emission essentially originates from very similar coordination environments. Emission spectra of solutions with stoichiometric ratios Eu:(L 6 ) − 2:3 and Zn:Eu:(L 6 ) − 1:1:3 display spectra identical to the one of the 1:3 solution ( Figure S6, Supplementary Material), consistent with the speciation reported in Figure 2; in particular,

www.frontiersin.org
September 2013 | Volume 1 | Article 15 | 5 the emission intensity of the heterometallic solution is weak, due to the low concentration in 1:1:3 species (38%, which translates in 19% with respect to the europium site). Low-temperature emission spectra are presented on Figure 6 for the ligand, a Gd III -containing solution and the Eu:Zn 1:1 sample. Upon enforcing a 50-μs delay time, fluorescence of the ligand almost disappears to the benefit of a phosphorescence band centered at 509 nm. In the Gd:(L 6 ) − 1:3 sample, this band is red shifted at 525 nm and presents a vibrational structure (463, 494, 525, 559 nm) with ≈1200 cm −1 spacing, typical of a ring breathing mode. The Eu:Zn sample also displays ligand fluorescence and phosphorescence, again consistent with the speciation of the solution. The ligand phosphorescence band for this sample is identical to the one exhibited by the Gd sample. From these spectra, the 0-phonon energy of the triplet state of the bound ligand can be estimated to be 21,600 cm −1 , while the 0-phonon energy of the singlet state lies at about 26,000 cm −1 , as estimated  from the onset of the fluorescence band. These energies are close to those reported for (L C2 ) 2−  and are adequate for ensuring efficient intersystem crossing and energy transfer onto the 5 D 1 (19,030 cm −1 ) and 5 D 0 (17,230 cm −1 ) levels of Eu III ; on the other hand the quasi resonance between 3 ππ * and 5 D 2 (21,500 cm −1 ) may generate some back energy transfer. Lifetimes and quantum yields of the Eu( 5 D 0 ) level are reported in Table 4. Despite the different speciation of the solutions, the luminescence decays can be fitted with monoexponential functions and the resulting lifetimes lie within a narrow range, 2.46-2.69 ms, the shortest figure corresponding to the solution with the largest concentration of the 2:3 (or 1:1:3) species, in line with the somewhat shorter lifetime reported for [Eu 2 (L Ci ) 3 ], 2.43 ms for i = 1 (Elhabiri et al., 1999) and 2  under similar experimental conditions. At low temperature, the lifetimes are much longer, pointing to temperaturedependent quenching mechanism(s) operating, e.g., back transfer (see above) or photo-induced electron transfer (PET). The radiative lifetimes can be estimated from the following equation: where A MD, 0 = 14.65 s −1 is the decay rate of the magnetic dipole 5 D 0 → 7 F 1 transition, n is the refractive index of the medium, and I MD and I tot are the integrated emission areas of the 5 D 0 → 7 F 1 and 5 D 0 → 7 F J (J = 0-6) transitions. The corresponding data lead to evaluation of the intrinsic quantum yield, i.e., the quantum yield upon direct excitation onto the 5 D 0 level, since it is very difficult to determine experimentally in view of the faint molar absorption coefficients of the f-f transitions.
The sensitization efficiency of the ligand can subsequently be calculated: Both sets of data, radiative lifetimes and intrinsic quantum yields, are consistent for the three samples and, moreover they compare very well with data reported for the [Eu 2 (L C1, 2 ) 3 ] helicates: τ rad = 6.8−6.9 ± 0.3 ms and Q Eu Eu = 36-37 ± 4% . On the other hand, absolute quantum yields and sensitization efficiencies are about three-fold smaller compared to the reference helicates (Q L Eu = 21-24%, η sens = 58-67%) due to the non-quantitative formation of the species. It is however noteworthy that the quantum yields of the two solutions containing Eu III only, which feature approximately the same concentration of EuN 6 O 3 sites (79 and 73% for 1:2 and 2:3 stoichiometric ratios, respectively), are equal, within experimental errors. For the heterometallic solution, the quantum yield is smaller, due to the small concentration of the 1:1:3 species, which, in addition features only one EuN 6 O 3 coordination site.

CONCLUSION
Reaction of ligands HL 5 and HL 6 (Scheme 3) under physiological conditions with equimolar quantities of Zn II and Ln III ions did not lead to a thermodynamically controlled assembly of the desired 3d-4f helicates, as expected from work with HL 4 , the 1:1:3 complex representing only 40% of the speciation. This can be traced back to stability of the Zn II complexes with these ligands, even with respect to coordination to the tridentate unit, which is close to that of the Ln III complexes. Attempts to isolate solid state samples of the helicates with transition metal ions such as Zn II , Cr III , Ru II also afforded mixtures which we did not succeed to purify to an acceptable level. There is no doubt that the NNO moiety of the ligand should be remodeled to get better recognition of the Ln III ions. An encouraging aspect, however, is that all luminescent data gathered for Eu III solutions with different compositions point to the formation of either 1:3 or 2:3, or 1:1:3 complexes in which the lanthanide ion is coordinated to the tridentate chelating unit of ligand HL 6 , a judged by the crystal field splitting and other photophysical parameters which reflect the peculiar signature of the Eu(NNO) 3 environment. In particular, the radiative lifetimes and intrinsic quantum yields match those of the previously reported helicates [Eu 2 (L C1 ) 3 ] and [Eu 2 (L C2 ) 3 ] with bis(tridentate) ligands.
On the other hand, the synthetic strategy applied for the preparation of tridentate-bidentate compartmental ligands aimed at assembling 3d-4f binuclear complexes proved to be valuable in that the ligands are obtained in reasonable yields given the number of steps needed. Moreover, the strategy can be adapted to graft other substituents on the sulfonate groups through modification of the key building block 6 (Scheme 2), so that this class of ligands represent a valuable addition to the chemistry of 3d-4f complexes.

Starting materials and general procedures
Chemicals and solvents were purchased from Fluka A.G or Aldrich. Solvents were purified by passing them through activated alumina columns from Innovative Technology Inc. (Pangborn et al., 1996). Complexes were studied in solution only. Stock solutions of lanthanides were prepared just before use in freshly boiled, doubly distilled water from the corresponding Ln(ClO 4 ) 3 ·xH 2 O salts (Ln = La, Eu, Gd, Lu; x = 2.5-4.5). These salts were prepared from their oxides (Rhône-Poulenc, 99.99% and Catalysis or Research Chemicals, Phoenix, AZ) in the usual way (Bünzli and Mabillard, 1986). The concentrations of the solutions were determined by complexometric titrations using a standardized Na 2 H 2 EDTA in urotropine buffered medium and with xylenol orange as indicator (Schwarzenbach, 1957).

5-(bis{2-[2-(2-methoxyethoxy)ethoxy]ethyl}sulfamoyl)pyridine-2carboxylic acid (6)
Compound (5) (1.80 g, 3.88 mmol) was added to a suspension of SeO 2 (1.94 g, 17.45 mmol) in dry pyridine (60 mL) maintained under an N 2 stream. The heterogeneous mixture was refluxed for 24 h and filtered through Celite® after cooling. Celite® was further washed with Et 2 O (about 100 mL) and the solvents were removed under reduced pressure. The residue was dissolved in distilled H 2 O (about 20 mL) and the pH was increased to 10 by addition of aqueous NaOH (5%). The aqueous phase was then extracted with CH 2 Cl 2 (3 × 100 mL). The aqueous phase was acidified to pH 3 by adding aqueous hydrochloric acid (25%) and the resulting solution was extracted again with CH 2 Cl 2 (3 × 100 mL). The organic phases were combined, reduced to a volume of about 100 mL, dried over Na 2 SO 4 , filtered, and rotor-evaporated under reduced pressure. After drying, (6) was obtained as an amber oil (1.83 g, 96% yield
After completion of the reaction, the solvents were evaporated. The residue was dissolved in distilled water (50 mL) and the resulting aqueous solution was acidified to pH = 2 by addition of 0.02 M hydrochloric acid. The acidic solution was then extracted with CH 2 Cl 2 (5 × 100 mL), dried over Na 2 SO 4 and evaporated. The crude product was triturated with hexane (100 mL), filtered, and dried under vacuum to give a pale yellow solid (498 mg, 98% yield). 1 H NMR (400 MHz,298 K,

ACKNOWLEDGMENTS
This project was funded by the Swiss National Science Foundation (grant 200020_119866/1). Jean-Claude G. Bünzli thanks the WCU program from the National Science Foundation of Korea (grant R31-2012-000-10035-0) for support.

SUPPLEMENTARY MATERIAL
The Supplementary Material for this article can be found online at: http://www.frontiersin.org/Inorganic_Chemistry/10.3389/ fchem.2013.00015/abstract Figure S1 | (Top) Re-calculated spectra from the titration of HL 6 with zinc perchlorate at 295 K and pH 7.4. (Bottom) Absorbance values extracted at different wavelengths during the titration compared with theoretical prediction from the stability extracted from the fit procedure (Table 1).