Molecular Recognition and Scavenging of Arsenate from Aqueous Solution Using Dimetallic Receptors

A series of copper(II), nickel(II) and zinc(II) dimetallic complexes were prepared and their affinities towards arsenate investigated. Indicator displacement assays (IDAs) were carried out to establish the complexes with best affinities towards arsenate. A di-zinc complex (3) was selected and its arsenate-binding abilities investigated by isothermal titration calorimetry (ITC). The X-ray crystal structure of this metallo-receptor bound to arsenate is also reported, which allowed us to establish the binding mode between 3 and this oxyanion. Immobilising 3 onto HypoGel resin yielded a novel adsorbent (Zn–HypoGel) with high affinity for arsenate. Adsorption of arsenate from competitive solutions and natural groundwater was greater than that of the commercially used iron oxide Bayoxide E33. Zn–HypoGel could be efficiently and simply regenerated by washing with sodium acetate solution.


Binding of PV by complex 3 -Job's Plot
The method of continuous variance (Job's Plot) was used to confirm the stoichiometry of the receptor: interaction. A series of solutions containing 3 and PV in varying molar ratios (0:1, 1:9…9:1, 1:0) were prepared in a 96 well plate, and the UV/vis absorbance at 445 nm was read. As shown in Figure S1, a plot of mole fraction vs ΔAbsorbance yields a parabolic curve. According to the Job's method, the maximum (or minimum) of this plot occurs at the mole fraction that results in the maximum formation of the receptor: indicator complex. In this case, the minimum occurs at

Binding of PV by complex 3 -Titration
The strength of the interaction between 3 and PV was investigated by UV/vis titration. Increasing amounts of 3 were added to 25 µM solutions of PV in a 96 well plate. A plot of receptor concentration vs absorbance at 445 nm yields a binding curve, which was fit in Origin using a 1:1 binding model described by Anslyn et al [1] and the binding constant was determined to be (2.

Binding of PV by complex 6 -Titration
The strength of the interaction between 6 and PV was investigated by UV/vis titration. Increasing amounts of 6 were added to 25 µM solutions of PV in a 96 well plate. A plot of receptor concentration vs absorbance at 445 nm yields a binding curve, which was fit in Origin using a 1:1 binding model described by Anslyn et al [1] and the binding constant was determined to be (7.5 ± 2.2) x 10 3 M -1 Figure S3, binding curve obtained upon titration of PV with complex 6 in 100 mM HEPES at pH 7.5, with fit curve shown in black

Arsenate and Phosphate Binding by complex 3 -displacement of PV
Displacement Assays could then be used to determine the strength of binding between 3 and arsenate and phosphate. Solutions containing a 1:1 ratio of 3 and PV were prepared in 100 mM HEPES at pH 7.5, and then increasing amounts of either arsenate or phosphate were added. Again, a plot of anion concentration vs absorbance at 445 nm yielded a binding curve for each interaction. These data could be fit in Origin using a displacement assay script reported by Anslyn et al and the anion binding constants were determined. Figure S4, displacement curves obtained upon titrating a 1:1 mixture of 3 and PV with arsenate and phosphate in 100 mM HEPES at pH 7.5, with fit lines shown in black The same method was also used to determine the strength of sulphate binding, however no displacement occurred even up to 10 equivalents of anion, and therefore the binding constant could not be determined.  Figure S5, showing that no displacement occurred upon titrating a 1:1 mixture of 3 and PV with sulphate in 100 mM HEPES at pH 7.5

Displacement of PV by Acetate
As crystal structures of 3 have been reported with acetate counter-ion bridging the zinc(II) centres, and acetate (or other carboxylates) could be present in natural waters, displacement assays were also used to determine whether binding of acetate would be significant. Acetate was added to a solution containing a 1:1 mixture of 3 and PV.  As shown in Figure S5, even addition of 10 equivalents of acetate to the solution induced only a tiny change in the UV-vis spectrum. This shows that binding of acetate is insignificant compared with arsenate or phosphate.

ITC -binding of Sulphate by complex 3
ITC was also used to investigate sulphate binding by 3, however as shown in figure S5, no binding interaction was observed.

Zinc Loading -Quantification
Zinc uptake by the functionalised HypoGel beads was determined by quantification of the zinc present in the initial and final reaction solutions, as well as the buffer washes. To this end, pyrocatechol violet was used as a zinc indicator. First, a zinc standard was used to produce a calibration curve for the response of PV to zinc in 10 mM HEPES at pH 7, as shown in figure S8.

Adsorption -Kinetics Experiment
In order to confirm that arsenate equilibrium was reached during the batch adsorption studies, arsenate uptake by Zn-HypoGel was monitored over 24 hours. 5 mg of sorbent was added to a 50 ml solution containing 1800 ppb arsenate in 10 mM HEPES at pH 7. 100 µL aliquots were removed at each time point for subsequent voltammetric analysis. As can be seen in figure S9, there is no further change in concentration after 8 hours, and therefore equilibrium is reached well within the 24 time period of the batch experiments.

Crystallography
The X-ray crystal structure of 5. The OH hydrogen atom of the included methanol solvent molecule in the structure of 5 was located from a ΔF map and refined freely subject to an O- The included solvent was found to be highly disordered, and the best approach to handling this electron density was found to be the SQUEEZE routine of PLATON. [ The O(94)-based methoxy group was found to be disordered. Two orientations were identified of ca. 75 and 25% occupancy, their geometries optimised, the thermal parameters of adjacent atoms restrained to be similar, and only the non-hydrogen atoms of the major occupancy orientation was refined anisotropically (the remainder were refined isotropically).
The B(1) and B(2)-based BF 4 anions were both found to be disordered. In each case three partial occupancy orientations were identified, (of ca. 53:25:22 and 59:24:17% occupancy respectively), their geometries optimised, the thermal parameters of adjacent atoms restrained to be similar, and only the atoms of the major occupancy orientations were refined anisotropically (the remainder were refined isotropically).

Fig. S11
The structure of the di-nickel(II) cation present in the crystal of 5 (50% probability ellipsoids)

Fig. S12
The structure of the di-cation present in the crystal of 7 (50% probability ellipsoids)

Table S1
Selected bond lengths (Å) for the crystal structure of 5.