Structural and Functional Diversity in Rigid Thiosemicarbazones with Extended Aromatic Frameworks: Microwave-Assisted Synthesis and Structural Investigations

The long-standing interest in thiosemicarbazones (TSCs) has been largely driven by their potential toward theranostic applications including cellular imaging assays and multimodality imaging. We focus herein on the results of our new investigations into: (a) the structural chemistry of a family of rigid mono(thiosemicarbazone) ligands characterized by extended and aromatic backbones and (b) the formation of their corresponding thiosemicarbazonato Zn(II) and Cu(II) metal complexes. The synthesis of new ligands and their Zn(II) complexes was performed using a rapid, efficient and straightforward microwave-assisted method which superseded their preparation by conventional heating. We describe hereby new microwave irradiation protocols that are suitable for both imine bond formation reactions in the thiosemicabazone ligand synthesis and for Zn(II) metalation reactions. The new thiosemicarbazone ligands, denoted HL, mono(4-R-3-thiosemicarbazone)quinone, and their corresponding Zn(II) complexes, denoted ZnL2, mono(4-R-3-thiosemicarbazone)quinone, where R = H, Me, Ethyl, Allyl, and Phenyl, quinone = acenapthnenequinone (AN), aceanthrenequinone (AA), phenanthrenequinone (PH), and pyrene-4,5-dione (PY) were isolated and fully characterized spectroscopically and by mass spectrometry. A plethora of single crystal X-ray diffraction structures were obtained and analyzed and the geometries were also validated by DFT calculations. The Zn(II) complexes presented either distorted octahedral geometry or tetrahedral arrangements of the O/N/S donors around the metal center. The modification of the thiosemicarbazide moiety at the exocyclic N atoms with a range of organic linkers was also explored, opening the way to bioconjugation protocols for these compounds. The radiolabeling of these thiosemicarbazones with 64Cu was achieved under mild conditions for the first time: this cyclotron-available radioisotope of copper (t1/2 = 12.7 h; β+ 17.8%; β– 38.4%) is well-known for its proficiency in positron emission tomography (PET) imaging and for its theranostic potential, on the basis of the preclinical and clinical cancer research of established bis(thiosemicarbazones), such as the hypoxia tracer 64Cu-labeled copper(diacetyl-bis(N4-methylthiosemicarbazone)], [64Cu]Cu(ATSM). Our labeling reactions proceeded in high radiochemical incorporation (>80% for the most sterically unencumbered ligands) showing promise of these species as building blocks for theranostics and synthetic scaffolds for multimodality imaging probes. The corresponding “cold” Cu(II) metalations were also performed under the mild conditions mimicking the radiolabeling protocols. Interestingly, room temperature or mild heating led to Cu(II) incorporation in the 1:1, as well as 1:2 metal: ligand ratios in the new complexes, as evident from extensive mass spectrometry investigations backed by EPR measurements, and the formation of Cu(L)2-type species prevails, especially for the AN-Ph thiosemicarbazone ligand (L–). The cytotoxicity levels of a selection of ligands and Zn(II) complexes in this class were further tested in commonly used human cancer cell lines (HeLa, human cervical cancer cells, and PC-3, human prostate cancer cells). Tests showed that their IC50 levels are comparable to that of the clinical drug cis-platin, evaluated under similar conditions. The cellular internalizations of the selected ZnL2-type compounds Zn(AN-Allyl)2, Zn(AA-Allyl)2, Zn(PH-Allyl)2, and Zn(PY-Allyl)2 were evaluated in living PC-3 cells using laser confocal fluorescent spectroscopy and these experiments showed exclusively cytoplasmic distributions.


■ INTRODUCTION
Thiosemicarbazones (TSCs) and corresponding d-block or pblock thiosemicarbazonato metal complexes (MTSCs) have attracted wide research interest due to their antifungal, 1,2 antiviral, 3,4 or antineoplastic 5,6 properties. Their applications in molecular imaging for Positron Emission Tomography have been explored since the observation of hypoxia selectivity in aliphatic derivative 64 Cu-labeled copper(diacetyl-bis(N4-methylthiosemicarbazone)], [ 64 Cu]Cu (ATSM), and several other Cu-based radiotracers (cyclic, or acyclic) have been investigated preclinically as theranostics. 7,8 The occurrence of hypoxia in tumors inversely affects the prognosis and treatment progression in therapy-resistant cases: radiotherapy is generally associated with the production of reactive oxygen species (ROS) and the local pO 2 level. 9 The hypoxia-selectivity of metallo-drugs structurally similar to Cu(ATSM) remains a matter of lively investigations, and it is likely that after decomplexation in vivo, the free ions follow the copper metabolism. 10,11 The research interest in the chemistry of thiosemicarbazones has been sustained over the past two decades, and examples of intrinsically fluorescent derivatives (with higher kinetic stability compared to that observed for other members of the TSCs compounds family, especially those incorporating flexible and aliphatic frameworks) have been reported. 12,13 The inclusion of an intrinsically fluorescent aromatic backbones opens the way for versatile multimodal imaging agents based on thiosemicarbazonato metal complexes. 14 Interestingly, a certain degree of hypoxia selectivity has been previously observed in 68 Galliumradiolabeled bis(thiosemicarbazonato) complexes with general formula 68 [Ga]Ga(BTSC), anchored on ligands with naphthyl groups which provide a rigid, flat and aromatic backbone to the BTSC ligand frameworks. 15 The hypothesis of the reduction of the metal, generally accepted when explaining the mode of action of Cu (ATSM) in hypoxic microenvironments can no longer be applied directly to the case of Ga-substituted BTSCs, as the reduction potentials of Ga(III) are outside of the biological range: as such, its observed trapping in hypoxic cells has been attributed to the targeting of iron species in these cancer cells. 15 This apparent hypoxic behavior in vitro, along with the ability to chelate a wide range of metals, in a plethora of conformations and coordination modes, boosted our interest in thiosemicarbazones as synthetic building blocks for theranostic applications.
A very small number of mono(substituted) aromatic thiosemicarbazones prepared from precursors that include rigid aromatic frameworks have been reported. 12−18 We recently reported the 68 Ga(III) incorporation and cellular uptake behavior of a range of mono(thiosemicarbazones). 15b Additionally, the biological activity of tridentate N/N/S derivatives as antiproliferative agents, such as triapine (3-aminopyridine-2carboxaldehyde thiosemicarbazone) has been reported, and assigned to their ability to generate ROS in the presence of iron, t h u s f u r t h e r e n h a n c i n g t h e i n t e r e s t o f m o n o -(thiosemicarbazonato) ligands for theranostic applications. 16−18 Thiosemicarbazones exhibit a number of coordination modes in binding to metal ions, acting as either bidentate, or tridentate ligands when their structure includes an additional donor atom. Thiosemicarbazones coordinate to the metal ion not only as bidentate and tridentate ligands but also as monodentate ligands. 1,2 In the case of hybrid donors such as O/N/S species, the tridentate mode generally prevails and TSCs have been shown to form highly kinetically stable complexes in a 1:2 (ML 2 ) fashion when M = Zn(II) or Ga(III), which present optical as well as coordination isomerism. 15,19 Of particular interest so far have been the mono(thiosemicarbazonato) complexes of M = Zn(II), Ni(II), Cu(II), and Fe(II) comprising the acenaphthenequinone backbone (denoted AN), and a small number of such MTSCs complexes have been previously reported. 20 A relatively small number of mono(thiosemicarbazone) ligands having the phenanthrenequinone (PH) backbone have been shown to form complexes for applications in catalysis: these were the ruthenium, 21 nickel, 22 or palladium 23 complexes, and the ligands involved were used as agents for water analysis 24−26 or for precious metals recovery. 27 To the best of our knowledge, the published reports on simple phenanthrenequinone-based mono(thiosemicarbazones) are even scarcer, especially in the context of their uses as antiproliferative 28−30 or antibacterial agents. 31 So far the application of rigid aromatic thiosemicarbazonatobased complexes for molecular imaging purposes has generally been limited to bis(substituted) acenaphthenequinone compounds, in a 1:1 metal to ligand fashion, where the bis(thiosemicarbazonato) ligands are coordinating to the metal through their N and S atoms, in a symmetric or asymmetric tetradentate manner. The resulting geometry of the complex changes with the nature of the metal and the occupation of the fifth coordination position is possible for Zn(II), Ga(III), and In(III) giving rise to generally square pyramidal structures, whereas for Cu(II) and Ni(II) analogous complexes, the geometry exhibited by bis(thiosemicarbazones) was square planar. 7 Interestingly, an in vivo imaging study reported a 9,10phenanthrenequinone (PH) thiosemicarbazone labeled with 61 Cu (a positron emitter with t 1/2 = 3.33 h, β+: 62%, E.C: 38%) as a radiolabeled anticancer compound for malignant tissue imaging studies. 32 The activity of octahedral Zn(II) mono-(thiosemicarbazonato) complexes as antineoplastic compounds was evaluated. 33 It was speculated that mechanism of action of these Zn(II) complexes in a number of different cancer cell lines involved the complex uptake followed by a transmetalation of the Zn(II) metal ions by Cu(II) ions present in lysosomes. The resulted Cu(II) complexes, formed intracellularly, then entered in a redox cycle that produced reactive oxygen species (ROS) and induced cellular apoptosis. 34 The rarity of these reports on TSCs, along with the applications for radioactive labeling of phenanthrenequinone-based ligands encouraged us to explore other extended aromatic backbones, and to apply microwave technologies in the synthesis of the ligands as well as in the metal complexation protocols.
Investigations into a new library of aromatic mono-(thiosemicarbazone) ligands including those with aromatic backbones derived from aceanthrenequinone (denoted AA) and pyrene-4,5-dione (denoted PY) are the focus of this work. Here, a new series of ligands and corresponding metal complexes were prepared by a rapid and efficient microwave heating method that allowed us to reduce dramatically the reaction time, whist giving rise to the desired products in superior, or comparable, yields with the cases when conventional heating methods were used. We report hereby on optimized, sustainable, synthetic methods and functionalization protocols for the exocyclic N atom of the TSCs derived from the AA and PY quinones, and the more widely investigated acenaphthenequinone (AN) and phenanthrenequinone (PH), and closely investigate the diverse range of conformations for the TSC framework found in the ligands and complexes ( Figure 1).  terminal NH t Boc group were prepared from the corresponding protected diamines, as described in Scheme 2. Ligand modifications with linkers and protected amine groups were explored using these building blocks and several new thiosemicarbazides containing a terminal NHBoc group and the AN backbone were prepared using the microwave-assisted methodology under optimized conditions. Our previously reported procedure involving ethylenediamine conjugates of aliphatic thiosemicarbazones 15b was simplified and generalized to result in the formation of derivatives with the AN backbone, denoted AN-11, AN-12, and AN-13 (Scheme 2). The synthetic protocol started with the corresponding protected diamines 1− 3 and followed adapted strategies (see Experimental Section), where the reaction proceeded through the condensation of the amine with carbon disulfide in a basic ethanolic medium. The addition of methyl iodide to the reaction mixture formed the corresponding thiocarbamate intermediates that were then isolated on milligram scale (see Experimental Section and SI). The reaction protocol continued with the hydrazinolysis of the intermediate by reflux in ethanol to obtain the desired thiosemicarbazides 7−9 in moderate yields. In these reactions, the main challenge was posed by the hydrazinolysis step, which often led to formation of complex mixtures that required extensive recrystallization and/or chromatography separation, likely due to the occurrence of the well-known cyclization of thiosemicarbazone as highlighted previously for other thiosemicarbazone derivatives. 14,15,37 The optimization of the purification conditions was necessary to obtain the desired product, especially by tuning the nature of the acid catalyst used. The deprotection of the amine by the removal of the t Boc group necessitated just a few drops of conc. HCl, unlike the case of previously studied 2,3-butanedione thiosemicarbazones. 8 Deprotection was confirmed by 1 H NMR spectroscopy which showed the significant diminishing of the characteristic singlet at ca. 1.38 ppm (integrating for 9 protons) characteristic for the t Boc group, and formation of a mixture of products. In the case where the removal of the catalytic HCl was attempted, in our hands, the treatment of the reaction mixture with basic solution (conc. NaOH) led to formation of traces of a urea-type derivative, which were isolated as traces of a paleyellow crystalline byproduct (<5% yield). These single crystals were mechanically separated and characterized only by single crystal X-ray diffraction (CCDC 2218629 and SI).
Therefore, the use of a weak acid as a catalyst for the reaction to avoid the deprotection of the Boc group for the formation of desired product AN-11 was employed. This optimized reaction was repeated several times, for the functional thiosemicarbazones depicted in Scheme 2, including the hexyl, 8 or 2,2′-(ethylenedioxyl) 9 using acetic acid (in a 10% v/v concentration in ethanol) as the acid catalyst. In this case, the desired products compounds AN-11, AN-12, and AN-13 were obtained, in good  yields (see Experimental Section), whereby the deprotection of the amino group did not occur. The presence of the hexyl chain or a PEG unit as a spacer in the last two mono-(thiosemicarbazone) examples significantly changed the phys- ical properties of the product. These linkers and functional groups enhanced the (notoriously limited) solubility of this class of acenaphthenequinone-based compounds in standard organic solvents. The derivatives AN-12 and AN-13 were initially obtained as oils and were separated from the starting materials by column chromatography in CH 2 Cl 2 /MeOH, which led to some hydrolysis, or precipitated as a yellow-colored solid by stirring in pentane overnight, and drying on standing at room temperature.
The linker-functionalized monothiosemicarbazones AN-11, AN-12, and AN-13 (and a range of related TSCs with the derivatized backbones aceanthrenequinone (AA), phenanthrenequinone (PH) and pyrene-4,5-dione (PY)) were characterized by 1 H and 13 C NMR spectroscopy and HR ESI mass spectrometry (see Experimental Section). The 1 H NMR spectra of these compounds in the aromatic region were comparable with those of the alkylic or arylic TSCs described above and also consistent with the previously reported compounds with AN backbones. 6,10 Specifically, the first characteristic hydrazinic proton resonance appears at 12−13 ppm while the second characteristic amino proton appears upfield with respect to it, at 9−10 ppm. This last amino group in the organic chain typically appears as a triplet resonance. The substituent resonances are in the 1−4 ppm region with the presence of the characteristic tbutyl resonances of the protecting group at 1.31 ppm. The 1 H NMR spectrum of compound AN-12 is shown in Figure 3a. Assignment was carried out using 1 H− 1 H COSY experiment (Figure 3b). The NH resonances, notoriously elusive, could be differentiated hereby because the amino proton showed a correlation peak with an alkylic proton (H-9). In addition, in the case of AN-12, crystals suitable for single-crystal X-ray crystallography were obtained by the vapor diffusion method, by dissolving the ligands in THF and layering with hexane and the structural features identified are described below.
Incorporation of Bio-orthogonal Substrates in Functional TSCs. We were interested in the prospect of developing these TSCs as new synthetic scaffolds for the incorporation of peptides of relevance to theranostic applications via the construction of an amide bond within the framework. 19 There is a limited number of TSCs based bioconjugates reported thus far, and these particularly focused on the introduction of a carboxylic acid moiety into a thiosemicarbazonato complex of a 2,3-butanedione moiety to include a benzoic acid in the backbone, which is amenable to then couple the ligand to peptides. 19c To explore the generality of our functionalization methods described above, we adapted the design elements to include a carboxylic acid in the thiosemicarbazonato compound within the AN-backbone functionalized and amine-terminated compounds described in Scheme 2. Our synthetic strategy involved the incorporation of the desired carboxyl functionality by a coupling reaction with a protected glutamine derivative through formation of an amide bond. Additionally, two protected side groups were maintained through a biorthogonal linker with the capability to be subsequently selectively deprotected. The coupling of the protected glutamine derivative was therefore evaluated as a proof of principle hereby using the L-Fmoc-Glu(OtBu)-OH a commercially available protected amino acid derivative (Aldrich). The experimental procedure consisted in the activation of the carboxylic acid group with pyBOP for 2 h at room temperature, followed by the addition of the deprotected amino-functionalized thiosemicarbazone derivative (AN-10), using our standard protocol 15b as shown in Scheme 3, Experimental Section, and SI. The success of the coupling with the protected amino acid expands the scope of the functional TSCs reactivity and opens a route for the attachment to targeting biomolecules and emergence of new bioconjugates.
Structural Highlights in Thiosemicarbazones with Flat, Aromatic, and Extended Backbones. Single crystals suitable for X-ray diffraction were obtained for the ligands by the slow diffusion of hexane into THF solutions of the ligands or from deuterated DMSO or CD 3 CN in NMR tubes. Generally, all TSCs frameworks show highly planar geometries and extensive networks of intramolecular hydrogen bonds are present in their 3D networks (Figures 4−6).
A close inspection of the selected molecular parameters for a range of compounds featuring ethyl groups at the exocyclic N atoms (Table 1) highlighted a subtle trend in key structural parameters, according to their collection in two groups: ligands featuring a 5-membered fused ring to the aromatic rings (TSCs with aceanthrenequinone and acenaphthenequinone backbones, denoted AA and AN, respectively), and those containing a 6-membered fused ring to the aromatic rings, i.e., with phenanthrenequinone (PH) and pyrene-4,5-dione (PY) backbones. The structural analysis structures of these ligands shows that generally the O1−C1 and C2−N1 distances are shorter in the first group (with the acenaphthenequinone AN and aceanthrenequinone AA backbones) while the corresponding C1−C2 and O1−N2 distances are larger. The shorter O1−N2 distance in the AN and AA-based derivatives (ca. 2.76 Å) with respect to the values found in phenanthrenequinone and pyrene-4,5-dione (ca. 2.56 Å) is most probably the reason of the deshielding of the hydrazinic proton as observed in the 1 H NMR spectra shown in Figure 1. The values for O−C1−C2 and C1− C2−N1 angles are larger in the case of those compounds incorporating either the AN or the AA backbone.
Scheme 3. Protocol Applied at the Coupling of a Deprotected and Amino-Functionalized Thiosemicarbazone with L-Fmoc-Glu(O)O t Bu a a General conditions applied for the synthesis of the deprotected AN-10 were: conventional heating (HCl cat., reflux, EtOH, 3.5h, 58% yield) or microwave-assisted irradiation (HCl cat., 90°C, 20 min, both in ca. 60% yields). Note: Use of AcOH as a catalyst (10% v/v) in this step led to retention of t Boc group and isolation of AN-11 in 38% yield. The final product AN-11-GLU was obtained from the known compound AN-10 as a yellow/orange solid in ca. 50% yield after purification, as described in the Experimental Section.
A selection of parameters of the new thiosemicarbazones with expanded frameworks are compared in Tables 1−3. The C1−C2 distance is of an average of ca. 1.48−1.49 Å across the series, and no significant differences were observed; the slight increase appears to be related to the bulkiness of the substituent in the R group of the thiosemicarbazone. As expected for sp 2 carbon's hybridization atom, the O−C1−C2 angles have a value in the region of 121°for all the compounds. The complementary angle between the C1−C2−N1 atoms have a value of ca. 124°and a trend to increase as seen for the C1−C2 distance, above. These molecules are all highly planar with the thiosemicarbazone, the exocyclic N substituent and the backbone all within the same plane with negligible deviations from planarity. The exception is the phenyl derivative where the PY-Ph aromatic ring is out of the plane formed by the backbone and the thiosemicarbazone unit by an angle of 56.62°. The distance between planes in the solid packing is close to 3.2 Å except for the PY-Ph derivative where this distance is larger, ca. 3.56 Å, due to the displacement of the Ph ring out of the plane. For the allyl-substituted compounds, the bond distances and angles are within the expected range (Table 3) and compare well with the crystallographic data for the mono(4-allyl-3-thiosemicarbazone) butane-2,3-dione 20 and with our previously reported structures of AN-Ethyl, AN-Allyl, and AN-Phenyl. 15b The O−C1−C2 and C1−C2−N1 angles are, however, smaller for mono(4-allyl-3-thiosemicarbazone) butane-2,3-dione than in any of the aromatic analogues investigated hereby. This can be attributed to the relative E configuration of the keto and imino groups along the C1−C2 bond, as the free rotation is allowed in the butane-2,3-dione derivative and rigidified in this series of mono-(thiosemicarbazones). Overall, the O−C1 and N1−C2    the six-membered cyclic ring fused to the aromatic group. The crystal structures for the functional derivative denoted AN-12 was also obtained ( Figure 4). The disposition of the thiosemicarbazone substituent with respect to the backbone is analogous to that of the ethyl-functionalized compounds, above, and the intramolecular hydrogen bond also features between O1 and N2 as in the entire range of monoTSCs derivatives. The nature of the R group does not seem to affect significantly the structural parameters in this series, and the observed values are all highly comparable to other acenaphthenequinone derivatives previously investigated. 15 The presence of an intramolecular hydrogen bond between the oxygen and nitrogen (N1) atoms can also be observed in AN-12. The bond distances are in the same range as those highlighted above compounds and especially close for AN-12 and its known AN-Et analogue. 15 A close inspection of the unit cell fragments and corresponding 3D packing diagrams (as highlighted for some representative examples depicted in Figures 4e and 6), all the ligands are arranged in the solid state in zigzag orientations and present interactions with molecules in the planes above and below the aromatic core. The solid-state packing of these compounds revealed the presence of short contact interactions between the different molecules. The distance between planes and the aromatic character of these compounds point to the presence of π−π interactions and observing the structures, the character could be attributed to parallel displaced π−π interactions except for the phenyl derivative with the PYbackbone, that also presents perpendicular y-shaped interactions, analogous to those found for other pyrene-based derivatives of interest to targeting cell nucleus and acting as intercalators in DNA. 21−23 The packing diagram of AN-12 showed that the CO groups in the acenaphthenequinone units are facing each other in a zigzag disposition with short contacts between the sulfur and C2, and there are close, and extended, intermolecular hydrogen bonds between the CO group of the Boc group and the NH of the NH t Boc group of a "neighboring" molecule in the unit cell.
Microwave-Assisted Metalation Reactions with Zn(II) Acetate. The formation of a small number of mono-(thiosemicarbazone) complexes of d-block metals has been reported to proceed under conventional heating, often involving prolonged reflux conditions, as highlighted in the Introduction. 20−22 Furthermore, we showed the preferential formation of acenaphthenequinone mono(thiosemicarbazonato) complexes of AN-Et, AN-Allyl, and AN-Ph in a 2:1 ligand: metal fashion for M = Zn(II) and Ga(III), where the ligand coordinated to the metal in a tridentate manner arranged in a distorted mer−mer configuration. Those Gallium(III) compounds have been described by us in the context of "cold" and "hot" gallium complexes formation, i.e., under thermodynamic vs kinetic control, respectively, and some of their analogous Zn(II) complexes were characterized structurally. 15b We adopted new metalation strategies based on conventional as well as microwave-assisted irradiation protocols for the formation of a range of new thiosemicarbazonato complexes of Zn(II) for the new ligands featuring rigid and extended aromatic backbones. The microwave-assisted metalation was successfully applied, and optimized with respect to our earlier studies, 15b to yield the thiosemicarbazone ligand featuring H as the substituent of the exocyclic N's as additionally to the ethyl and allyl mono(thiosemicarbazones) ligands incorporating the extended backbones denoted AA, PH, or PY. For selected ligands (with R = H, Me, Et, and Allyl), the corresponding Zn(II) complexes were also obtained by applying both conventional heating and microwave irradiation methods sideby-side, and in a range of ligand: metal ratios, to optimize the Zn(OAc) 2 metalation as shown in the Experimental Section and SI. The use of microwave irradiation for the Zn(II) metalation reactions technique reduced considerably the reaction time needed in the preparation of these derivatives, and the ZnL 2 species emerged preferentially after some extremely straightforward protocol. For the AA-Ph, PH-Ph and PY-Ph, the low solubility of the resulting metal complexes in common organic solvents prevented the spectrochemical characterization and unequivocal identification, and further studies are in progress in our laboratories. For the case of the TSC ligands featuring AN, AA, PH, and PY backbones and substituted with R = H, Ethyl, and Allyl, the Zn(II) complexes obtained after the microwave reaction emerged generally as orange to red colored solids in yields, ranging from 50−95%, with minimum purification being necessary (see Experimental Section). We found that mild and highly reproducible synthetic routes developed here led to a new class of Zn(II) complexes in ca. 90−95% purity by HPLC. A color change was observed during the synthetic process and HPLC analysis (with UV detection at 280 nm, as well as 450 nm), as well as UV−vis spectroscopy, were used to monitor the complex formation (Scheme 4 and Figure 7a−b). The final products were fully characterized by ESI + mass spectrometry and 1 H NMR. Figure 7d shows a comparison of the 1 H NMR spectroscopy and free ligand spectroscopy of PH-Et given for this ligand of type HL and its corresponding Zn(II) complex (of type ZnL 2 ), and Figure 7e 2 , in all cases, the disappearance of the NH proton from the hydrazine group was observed, indicating metal coordination. The aromatic region showed several overlapping multiplets for the resonances assignable to the H's in the ligand backbones. The interpretation of 1 H NMR spectra, however, often proved challenging: for AN backbone spectra showed the inequivalence of the two ligand units by NMR in d 6 -DMSO, which we assigned to optical isomerism in previous studies. 15b As previously described for Ga(III) and Zn(II) compounds, formation of coordination isomers for Zn(II) in C.N. Four with a distorted tetrahedral geometry (in the N/S/N/ S environment), as well as in a pseudo-octahedral environment (O/N/S/O/N/S or O/N/S/S/N/O), are also possible; however, HPLC did not indicate any differences in solution for any of the Zn(II) complexes investigated, and only one single dominant species was found, which we assigned to the ZnL 2 type derivatives. The general low solubility of this class of compounds (due to aggregation behavior in common organic solvents) prevented detailed NMR investigations and hampered full 13 C{ 1 H} NMR assignments, especially for the quaternary carbon resonances (see Experimental Section and Supporting Information).
The UV−vis and fluorescence spectroscopy in highly diluted solutions were performed to evaluate their potential as optical imaging agents ( Figure 7). The summary of the UV−vis and fluorescence properties such as maximum absorption wave-  length, maximum emission wavelength, Stokes shift (Δλ), and quantum yields (Φ), are given in the SI (Tables S2−S3 and Figures S10−S18). All the complexes with the exception of the Zn(AN-Allyl) 2 , which was reported earlier, 15 and included hereby for a comparison, show absorption wavelength maxima in the visible region. The emission wavelengths are in the visible region at ca. 600 nm except for the complex having the pyrene-4,5-dione backbone, denoted Zn(PY-Allyl) 2 , which is at 542 nm. The Stokes shifts are large for all the complexes. As expected, the quantum yields are low for all the mono-(thiosemicarbazonato) complexes with respect to other organic or inorganic fluorophores. 15d,e Structural Investigations of Zn(II) Complexes of Thiosemicarbazides with Extended Backbones. Crystals suitable for X-ray diffraction for the Zn(II) mono(4-ethyl-3thiosemicarbazonato) acenaphthenequinone and phenanthrenequinone complexes were obtained by slow diffusion of pentane in a THF/DMSO solution of the complexes, or from concentrated d 6 -DMSO solutions. For the analysis of Zn(AN-H) 2 complex, crystallography studies indicated that the two structural isomers were present in the same asymmetric unit. These presented two different coordination geometries around the zinc center. The crystal structure of this zinc complex showed the expected planar geometry for each mono-(thiosemicarbazone) ligand unit AN-H and a heavily distorted tetrahedral geometry (i.e., with Zn(II) in a N/S/N/S environment) vs the corresponding pseudo-octahedral geometry (where Zn(II) ion was found in the mer−mer O/N/S/S/N/O environment), as shown in Figures 8− Figure 10. DFT calculations (vide infra, and Supporting Information) showed that the optimized, equilibrated structure for the Zn(II) in the environment of two ligands (L − ) displays an octahedral environment. Our previous studies for the structure determi-nation on Zn(AN-Et) 2 indicated that in the solid state, both a tetrahedral environment and an octahedral environment at the metal centre occurred for Zn(II) complexes, 15b and the occurrence of a distorted tetrahedral environment at the Zn(II) center was confirmed hereby for Zn(AN-H) 2 . The possibility of both tetrahedral and octahedral donor arrangement around the Zn(II) center here is similar to the case of the two isomers of the   (Tables S5−S19).
The structures of a range of Zn(II) complexes are depicted in Figure 10 along with a fragment of the unit cell showing the 3Dpacking arrangement for Zn(PH-Et) 2 in the solid state ( Figure  10b). Generally, the observed geometry corresponds to a distorted octahedral disposition of the O/N/S donor atoms with the Zn(II) metal in the center in the expected mer-mer geometry. This is similar to the octahedral isomer found for the synthesis of Zn(AN-Et) 2 and no evidence of a tetrahedral isomer was found hereby by crystallography for any of the species analyzed and where the extended backbone was AA, PH, or PYtype. An overview of the structural parameters indicated that the estimated angle between the ligands' mean planes is close to 90°( e.g., 89.9°for Zn(PH-Et) 2 ), more so than the ca. 85.6°found in the previously reported complex Zn(AN-Et) 2 which showed a heavily distorted octahedral geometryf around the metal center, as shown in the corresponding X-ray structure. 15b Optical Spectroscopy and Cellular Imaging with Zn(II) Complexes. The cellular uptake and cytotoxicity were evaluated for several ligands and Zn(II) complexes in two commonly used, cancer cells lines, HeLa and PC-3 cells. These are well established, commercially available from ATCC as obtained from human cervical cancer and human metastatic prostate cancer, respectively, and routinely used for cancer pathological mechanism studies and drug testing, including in our own previous investigation on related thiosemicarbazones. 15 These tests were carried out for a subset of compounds which showed intrinsic fluorescence and most promising solubility in aqueous media (with 1% DMSO), aiming to ascertain their relevance for bioimaging assays using laser scanning confocal microscopy ( Figure 11) and MTT assays following our standard protocols. 13−15 The Zn(II) compounds investigated were Zn(AN-Allyl) 2 , Zn(AA-Allyl) 2 , Zn(PH-Allyl) 2 , and Zn(PY-Allyl) 2, (e.g., each incorporating the allyl substituent at the exocycic N's and different aromatic backbones. In each case, it was observed that at 100 μM concentration (1% DMSO) these showed cellular uptake and fluorescent emissions were visible in the cells' cytoplasm. The highest fluorescence intensity emission was obtained in the green channel (λ em 516−530 nm), for excitation wavelengths of 405 or 488 nm ( Figure 11). For the Zn(II) compounds with R = Et or Ph substituents at the exocyclic N's, the relatively high concentrations needed to achieve sufficiently observable fluorescent emission led to considerable cellular damage, as well as precipitation. None of the free ligands show sufficient intrinsic fluorescence in cellular media at comparable concentrations. Meanwhile, the fluorescence intensity in the series Zn(AN-Allyl) 2 , Zn(AA-Allyl) 2 , Zn(PH-Allyl) 2 , and Zn(PY-Allyl) 2 increased with the increase of the number of aromatic rings in the backbone. A certain degree of precipitation of the complex was still observed upon addition in serum-free and phenol-free RPMI medium, and therefore PBS washing protocols needed to be employed. This challenge could be potentially solved by enhancing the solubility by introducing the functional groups at the exocyclic N's, which will be pursued in further investigations in our laboratories.
Furthermore, for a subset of free ligands with AN-backbone and several representative Zn(II) complexes, the 48 h MTT assay was performed in PC-3 and HeLa cell lines to evaluate the IC 50 values and to compare these with the well-established cytotoxic behavior, previously reported for related bis-(thiosemicarbazonato) complexes and the clinical drug cisplatin. 13,14 The IC 50 values of the ligands featuring the ANbackbone and simple substituents as well as those of the cisplatin treatment groups were obtained following incubation for 48 h (Table S4, Figure S70). The Zn(II) compounds investigated generally showed lower cytotoxicity than the free ligands with the IC 50 2 ] (synthesized via the microwave protocol described in Experimental Section, and discussed below) showed a IC 50 value of (2.25 ± 0.01) μM, highly comparable to that seen in the free ligands. The IC 50 values of the cis-platin treatment groups were (31.28 ± 9.38) μM in HeLa cells, and (30.64 ± 3.26) μM in PC-3 cells. Therefore this subset of compounds, analyzed for proof-of-concept (whether free ligands or Zn(II) or Cu(II) complexes), showed a significant cytotoxicity in line with previous observations of related TSCs, 15 and further, more detailed biological investigations are underway in our laboratories.
Metalation Reactions with Cu(OAc) 2 and Investigations by EPR Spectroscopy. To generalize the synthetic approach to other d-block metal ion incorporation into these TCSs, room temperature, as well as analogous microwaveassisted conditions, were applied to the reactions of the ANbackboned TSCs with Cu(OAc) 2 in a variety of organic solvents (DMSO, MeOH or THF). However, the reaction outcome was not as straightforward as seen with Zn(II), as discussed below. A color change toward red-brown was observed in all cases upon mixing the starting materials and the products were observed by mass spectrometry but the reaction mixture contained decomposition products observed as dark residues in the product and several peaks in the mass spectrometry (see Supporting Information and below). This result would indicate that the synthetic method is metal-dependent, and it is clear that the Zn(II) complexation is significantly more thermodynamically and kinetically favorable. In this work, reactions of selected ligands (of relevance for the 64 Cu-radiolabeling experiments, vide infra) were carried out with anhydrous Cu(OAc) 2 and were explored at the room temperature, under the microwave irradiation or using mild conventional heating.
Several reaction setups were explored, using either 1:1 or 1:2 molar ratios of metal to ligand, in order to obtain Cu(II) complexes from Cu(OAc) 2 , i.e., in processes carried out under thermodynamic control. The reaction mixtures and purified compounds were monitored by HPLC and extensive mass spectrometry. Copper complexes formed seem to have significantly lower kinetic stability in solution and with respect to acidic environment with respect to their Zn(II) counterparts. In the absence of X-ray diffraction, mass spectrometry was especially instrumental in pointing out the possibility of complexes showing 1:1, as well as 1:2, metal: ligand ratios, and this was consistent with observations from radio-HPLC investigations at the formation of new 64 Cu complexes (under kinetic control). Scheme 5 gives a representation of our postulated formation of Cu(II) compounds (and corresponding isomers) under mild conditions, and Figure 12 shows the "relaxed" gas-phase calculations for the DFT calculated geometries for the AN-Ph ligand.
The postulated structures of the three main species formed, likely in equilibrium in solution additionally to the free ligands, as shown in Scheme 5, and evidence for the presence of these species was found by extensive high resolution ESI + mass spectrometry investigations. The gas-phase DFT optimizations of the proposed monometallic Cu(II) species identified by HRMS ESI + were carried out to identify whether or not such species would show thermodynamic stability in gas phase. Extensive molecular parameters data for the optimized geometries, their energies, and the corresponding HOMO−LUMO levels and corresponding Cu and Zn complexes of the AN-Ph ligand and the simplified AN-H variant are given in the Supporting Information. We postulate that the major component for reactions carried out under thermodynamic control is the CuL 2 -type species, while under kinetic control, the equilibrium between 1:1 and 1:2 metal:ligand species cannot be discounted (vide infra), additionally to the formation of CuL 2type species (and corresponding isomers/stereoisomers).
While the possibility of isomerism in CuL 2 species is expected by analogy with Zn(II) structures discussed above, where the octahedral vs tetrahedral geometries are possible in the solid state, the structures of the Cu(II) complexes formed in these reactions could not be determined unequivocally in the absence of X-ray structure determinations for the compounds synthesized. Therefore, extensive EPR spectroscopy was used to shed light into the nature of these species in solution, as well as in the solid state, with an aim to probe for the possibility of the coexistence of multiple Cu(II) centers. The products of the reactions carried out under mild conditions between Cu(OAc) 2 and HL ligands (in a 1:2 ratio, for AN backbone, and R = Me, Et, Allyl, Ph), as well as those emerging from the reaction conducted in a 1:1 ratio of Cu(OAc) 2 :ligand HL (where HL was AN-Et, AN-Ph) were analyzed, and corresponding EPR parameters (Figures 12 and S57, Supporting Information) and magnetic susceptibility behavior were evaluated ( Figure S58, Supporting Information).
The species of interest for analysis by detailed EPR spectroscopy emerged from reactions carried out at the room temperature using either 1:1 or 1:2 metal:ligand ratios, as described in the Experimental Section. Samples analyzed by EPR spectra were first obtained for the powdered solids (A−F) listed above, as well as in fluid and frozen solutions, as described below. Spectra in neat DMSO as solvent showed no Cu hyperfine splitting when frozen. All samples above gave poorly resolved fluid solution spectra, which may derive from the inclusion of the acenaphthenequinone backbone of these ligands, and which could serve to slow the tumbling rate of the molecule in the relatively viscous DMSO solvent and cause a broadening of the spectrum.
We assigned this to the possibility that in solution an equilibrium between a number of Cu(II)-ligand species can occur, with the proposed geometries shown in Scheme 5. Additionally, a 7:1 v/v EtOH/DMSO mixture was used to produce well-resolved frozen glass spectra. These spectra exhibited three of four lower field, low intensity Cu-hyperfine resonances and more intense higher field features with no obvious Cu-hyperfine structure, which is consistent with a tetragonally elongated electronic structure for the Cu(II) center in each complex. The "parallel" region seems qualitatively diagnostic. The polycrystalline powder spectra do not provide much additional information. The absence of 63,65 Cu hyperfine splitting indicates that the samples are magnetically concentrated.
The spectra of frozen solution and polycrystalline powder spectra of samples (A)-(F), denoted Cu-AN-Me, Cu-AN-Et-a, Cu-AN-Et-b, Cu-AN-Allyl, Cu-AN-Ph-a, and Cu-AN-Ph-b, as shown in Figure 13. All the frozen glass spectra are of the tetragonally distorted type, and all except the 1:2 Cu:ligand complex of AN-Ph ligand which seemed to suggest that at least two species are present, based on the patterns in the A ∥ region. Furthermore, Table 4 shows that the numbers of copper(II)containing species simultaneously present as indicated by extensive simulations are as follows: (  There is no well-defined 14 N superhyperfine splitting. The powder CW EPR spectra of Cu-AN-Me, Cu-AN-Et-a, Cu-AN-Et-b, Cu-AN-Allyl, Cu-AN-Ph-a display a single broad line that is centered near the middle of the equivalent frozen glass spectrum, and this is consistent with a magnetically broadened spectrum. The powder spectrum of Cu-AN-Ph-a is resolved into two g-value components, with the more intense g ⊥ at lower field than g ∥ , which might imply a reversal of g-values for the isolated Cu sites, which would be consistent with a (3d z 2 ) 1 ground state configuration being trapped in the solid lattice, which then relaxes to the more common (3d x 2 −y 2 ) 1 configuration on dissolution.
These observations are consistent with the ESI + mass spectrometry results, which indicated that in all cases, fragments consistent with [LCu(OAc) + H + ] and [LCu(DMSO)] + (where L − corresponds to the deprotonated ligand, i.e., AN-Me, AN-Et, or AN-Ph) can be identified additionally to the [CuL 2 ] + . The DFT calculations for this simple system indicated that formation of such 1:1 Cu: L − species is plausible, additionally to the expected [CuL 2 ] species, which dominates the ESI + mass spectra for all investigated Cu(II) compounds irrespective of the reactants ratios used. We already suggested in earlier studies that these versatile tridentate ONS ligands could adopt mergeometry in the 1:2 Cu: ligand complexes of type ML 2 ; however, if a distorted octahedral geometry is adopted in the solid state, the arrangement of the exocyclic substituents could also give rise to several geometric isomers additionally to the presence of optical isomerism, which has been proposed for similar molecules, and supported by DFT calculations. 15b The crystallography of Zn(II) complexes of the type ZnL 2 (with R = H, Et, and backbones AN, PH, and PY) indicated the possibility of tetrahedral (N/S/N/S), as well as highly distorted trigonal pyramidal and octahedral arrangements of the ligands around the metal center (O/N/S/O/N/S for R = Et), as shown above, and in previous reports. 15 Furthermore, formation of dimeric copper(II) complexes derived from the related monothiosemicarbazone anchored on 2-formylpyridine, HFo-PyTSC seems ubiquitous whereby tridentate NNS and the fourth basal positions are occupied by acetate oxygen that are strongly coordinated (Cu−O bonds of ca. 1.95 Å). Related structures with centrosymmetric dimers with more weakly bound, axial placed acetate dimers are also possible (and with distances of Cu−O 2.42 Å) more closely resembling monomeric species which are weakly associated in the solid state. 36 We also obtained the temperature and field dependent magnetization measurements for the samples A−F immobilized in eicosane ( Figure S57, Supporting Information), all of which behave as paramagnets, as expected. Samples E, Cu-AN-Ph-a, and F, Cu-AN-Ph-b, showed nearly horizontal lines, however one would expect better correspondence than observed hereby if each of the samples investigated were to be considered as simple paramagnets, and the plots for Cu-AN-Et-a and especially Cu-AN-Me seem to be indicative of strong antiferromagnets.
The powder CW EPR spectra of Cu-AN-Me, Cu-AN-Et-a, Cu-AN-Et-b, Cu-AN-Allyl, Cu-AN-Ph-a, and Cu-AN-Ph-b all displayed the characteristic single broad line that is centered near the middle of the equivalent frozen glass spectrum, consistent with a magnetically broadened spectrum. Deconvolution of CW EPR spectrum of Cu(AN-Me) (i.e., emerging from the 1:2 reaction of Cu(OAc) 2 with the HL-type ligand AN-Me) indicate the presence of 3 different copper(II) environments in this sample. If the structures we propose all show a strong component from species exhibiting distorted octahedral environments for the Cu(II) in all these samples, whether emerging from 1:1 reactions or 1:2, there must either be some very strong intermolecular interaction, or maybe the compounds are coupled ligand radicals in the solid state. For the low T magnetization, for a simple s = 1/2 paramagnet the value of molar magnetization expected when the curve plateaus at high field is g × S, which would be ca. 1.05 μ B (assuming g av = 2.1). Magnetisation data (Supporting Information, Figure S58) show all four compounds tend to reach saturation at high field. Since the EPR measurements also indicated a number of species present, possibly in equilibrium, the correction for diamagnetism was not deemed feasible: the decrease in magnetic moment with temperature might imply that the solid state structures resemble the supramolecular aggregation analogous to that already seen in Zn(II) complexes of type ML 2 , in that they stack extensively in the solid state and as such these Cu(II) samples would display overall antiferromagnetic interactions.
From all analytical and spectroscopy data, taken together, we speculate that the possibility of Cu(II) dimers, linked by one or even two acetate ligands in a bridging mode, similarly to the case of the literature-reported, related monothiosemicarbazone anchored on 2-formylpyridine, HFoPyTSC seems plausible, e.g., whereby tridentate NNS and the fourth basal positions are occupied by acetate oxygen that are strongly coordinated could not be discounted. However, the frozen solution EPR spectra, where it could be resolved, are all consistent with the occurrence of monometallic species being present, rather than dimers. If something like a paddlewheel dimer structure could be found then there would be seven hyperfine lines, but there are only four observed hereby, and the simulation account for all the features in the spectra. The powder EPR spectra cannot determine the exact nature of these species, although the breadth suggests there is a copper component, and on dissolution a typical Cu pattern is present: we eliminated the possibility of production of organic radical species under the mild conditions in which the reactions were conducted. As stated above the presence of a minor (inseparable) component Cu: L 1:1 additionally to the Cu: L 2 dominant component cannot be discounted, and the extensive mass spectrometry investigations carried out (ESI) pinpoints to a range of species being feasibly present in solution, possibly in equilibrium as shown in Scheme 5, and so do the frozen solution EPR spectra. Furthermore acetate-bridged dimers with the general formula [(TSC)Cu(OAc) 2 Cu(TSC)] have been reportedly isolated for other thiosemicarbazone complexes of Cu(II), however we did not see evidence for such dimers in mass spectrometry of the species analyzed hereby. The EPR determinations in such dimeric compounds have been scarce and a direct comparison of this work, with previously investigated TSCs has not thus far been possible. 36 Radiochemistry Assays for the 64 Cu Incorporation under Mild Conditions. Metalation reactions under kinetic control were carried out using 64 Cu(OAc) 2 , as described in the Experimental Section and in Supporting Information. Overall room temperature radiolabeling carried out at pH 5.5 generally proceeded with ca. 50% incorporation yield, whereas moderate heating for ca. 30  Accurate determination of the g x , g y , |A x |, and | A y | values was not possible owing to the second-order nature of the perpendicular region, although it was noted that satisfactory simulation could only be achieved with the particular set of values reported in the simulation. Furthermore, it was noted that the superhyperfine splitting due to 14 N/ 1 H-nuclei along the g x , g y , regions was poorly resolved/not clearly visible to the naked eye; however, given the ambiguity in the number of 14 N/ 1 H nuclei coupled to the electron spin, these were not included in the simulations for selective cases to remove the overparameterization. Simulations, which included the 63,65 Cu-hyperfine matrix is given in the table above. c For a selected experimental spectrum, the simulation involves inclusion of two/three EPR-active species, whose population is provided next to the g-matrix values. d The sign of the hyperfine coupling is not determined, so absolute values are given. e The line shape of the spectra was reproduced by considering an isotropic Voigtian line shape and an anisotropic broadening(H-Strain) respectively. Supporting Information). The UV detection HPLC traces of the corresponding "cold" Cu(II) complexes were difficult to assign as the degradation of the complex to free ligand occurs in the presence of TFA, and precipitation also occurs at the concentrations needed to record these HPLCs. The radioHPLCs of the 64 Cu complexes indicate consistent behavior at the formation of copper-64 species in solution in all compounds studied. Furthermore, an increase in 64 Cu activity used at the start of the radiolabeling experiments (from 10 mBq to 100 MBq activity in starting materials samples for the radioreaction, see Supporting Information) showed that it is possible to resolve the 64 Cu species present, and up to three distinguishable peaks occur within 90 min experiment time under conventional heating (i.e., for the optimizations performed at the 64 Cu labeling of AN-Et and AN-Ph).
The radioHPLC traces obtained for the reactions at room temperature pointed out the presence of two different 64 Cu(II) species in solution, similar with the case of the analogous 68 Ga chemistry, reported by us earlier. 15b Analogous to the gallium radiochemistry assays carried out under similar conditions, we also observed hereby the formation of two main 64 Cu-based species by radio-HPLC, that cannot be separated. Reactions carried out overnight seem to lead to one major species by radioHPLC however broadening of this peak is also observed. This unusual feature is currently assigned to differences regarding the synthesis protocol under thermodynamic vs kinetic control, or decomposition of the CuL 2 species to a Cu(II)LX species under radiosynthetic conditions where excess of NaOAc and other ligands may also be present in aqueous environment (X = OAc − , Cl − , or OH − and L = monoanionic mono(thiosemicarbazide) ligand). Through optimized radiolabeling experiments sought to shed light into the nature of these compounds, which we assigned either to isomerism in the octahedral vs tetrahedral compounds, optical isomerism in the isomer with octahedral geometry and/or the simultaneous formation of a 1:1 M:L species, possibly in exchange in diluted solutions with the ML 2 -type compounds. Under kinetic control formation of species of type [CuL(DMSO] + and [CuL(OAc)] could not be ruled out in the presence of competing ligands, such as (OAc) − and coordinating solvents (DMSO).
Although radiochemical yields were moderate for the ligands with extended backbones, such as PH and AA (which we assigned to steric hindering and extensive aromatic stacking of the free ligands involved), the radioincorporation yield (estimated by integrating radioHPLC) was well above >90% for the AN derivatives. By similarity with what was reported for a 61 Cu radiolabeling of a PH-backbone TSC (monitored by iTLC 41 ) and our own observations from the 68 Ga radiolabeling assays reported earlier, 15b we propose that the generally consistent occurrence of two different 64 Cu species in reactions carried out at the room temperature may also be assignable to the presence of optical isomers for 64 CuL 2 additionally to the presence of 64 CuLX species, that these species occur simultaneously under kinetic control, and they are generally detectable at ca. 1 min difference by radioHPLC in fresh solutions, yet not fully separable at more than two half-lives for the radio-reaction. Collection and reinjection of samples consistently led to formation of these "twin" peaks, and employment of high radioactivity starting material 64 Cu, and slightly harsher conditions (90 min with heating in DMSO) led to observation that 3 different peaks occur, which can be distinguished within 1−2 min r.t. in reverse phase HPLC.
The addition of a base (e.g., NH 4 OH or LiOH) to deprotonate the ligand prior to 64 Cu radiolabeling did not appear to have a significant effect in the radiolabeling yield or number of species detected by radioHPLC. Validating the nature of the species emerging from the 64 Cu radiolabeling of monothiosemicarbazones of this family of compounds (using 64 Cu(OAc) 2 as the precursor of choice) proved challenging, yet in many respects analogous to the behavior observed for the analogous 68 Ga radiochemistry. The occurrence of the "twin" peaks features are consistent with our observations from 68 Ga incorporation in TSCs, 15b and we also assign this either to the simultaneous 1:1 and 1:2 Cu:Ligand association, or to isomerism in Cu(L) 2 -type complexes.
As indicated above, the analogous Cu(II) coordination chemistry, whereby reactions between the ligands investigated (AN-H, AN-Me, AN-Et, AN-Allyl, and AN-Ph) and Cu(OAc) 2 were carried out under thermodynamic control (in ratio of ligands to metal of either 1:1 or 2:1) did not appear to proceed efficiently. We suggested that these resulted in complex mixtures of species that undergo equilibrium reactions in solution and seem to revert to the free ligand under the HPLC conditions in the presence TFA. Thus, the "cold" standards for analytical chemistry comparisons were unavailable for 64 Cu(II) radiochemistry. In the case of the Zn(II) complexation reactions, these reactions invariably lead to the preferential formation of ML 2 -type derivatives. The HPLCs of the "cold" Cu(II) compounds isolated and purified from thermodynamically controlled reactions seem to indicate loss of ligand in the presence of TFA which was used under the standard HPLC conditions successfully applied for their Zn(II) analogues. Further details and corresponding data is given in Supporting Information.
We suggest that in each case the two broad signals correspond to at least two distinct complexes of copper(II) in different coordination environments and which we hypothesize to feature an earlier eluting time, presumably due to an MLX 2 type compound together with the ML 2 complex, and which equilibrate with time supporting our postulated structures from Scheme 5 and Figure 12. These findings are also consistent with our observations from EPR. All compounds radiolabeled rapidly with 64 Cu(OAc) 2 at room temperature or with moderate heating, although some challenges remain to be addressed, in line with our earlier observations from 68 Ga chemistry: dual or triple peaks with rather close retention times were identified by radioHPLC for all these copper derivatives. These may be assignable to isomerism in ML 2 species, or monosubstituted MLX/ML(DMSO) + type thiosemicarbazones: the precise identity of all the species present in solution and their behavior at varying pH remain to be investigated in future studies in our laboratories.

■ CONCLUSIONS
In summary, we developed an efficient general method for the synthesis of novel highly planar and rigid mono-(thiosemicarbazone) ligands with extended aromatic backbones reliant on microwave-assisted irradiation. A new family of aromatic mono(thiosemicarbazones) was obtained by varying their exocyclic N group R substituents, including aliphatic or aromatic groups from in-house prepared or commercially available thiosemicarbazides.
The modification of the thiosemicarbazide functional groups was explored, opening new routes for the future applications of these compounds, such as true-theranostic probes for dual imaging and sensing applications. For example, the functionalization of the R group in the thiosemicarbazide opens up the possibility to further bioconjugation with targeting biomolecules. Selected ligands were used in the preparation of Zn complexes also applying a microwave method that reduced the reported procedure through conventional methods by several hours. All new compounds were fully characterized by spectroscopic techniques ( 1 H and 13 C{ 1 H} NMR, IR, UV−vis, and fluorescence spectroscopies, also EPR spectroscopy in the case of the Cu(II) compounds synthesized). Their structures were demonstrated by single crystal X-ray diffraction, and a variety of conformations was highlighted for the free ligands (denoted HL) or the corresponding Zn(II) complexes of the monodeprotonated monothiosemicarbazones studied, denoted Zn(L) 2 . A selection of ligands and metal complexes was also carried forward to perform cytotoxicity assays in standard cancer cells, and for radiolabeling experiments to incorporate 64 Cu. The biological evaluation by MTT cytotoxicity assays was pursued by in vitro cellular imaging experiments by laser confocal microscopy in two commonly used human cancer cell lines, HeLa (human cervical cancer cells) and PC-3 (human prostate cancer cells), and the compounds analyzed show consistently higher cytotoxic activity by comparison with cis-platin, which is in line with our previous investigations into related species. 12−15 The potential of these mono(thiosemicarbazones) to act as synthetic scaffolds for new molecular imaging agents was explored by performing 64 Cu radiolabeling assays analogous to those developed for [ 64 Cu]Cu(ATSM) and related bis-(thiosemicarbazones). 10,11,19 Our experiments gave rise to new, longer-lived radiotracer analogues with respect to our previously investigated 68 Ga and 18 F labeled mono-(thiosemicarbazones) in this family. We suggest that these new ( nat Cu or 64 Cu-labeled) copper(II) compounds, while very interesting structurally, are less kinetically stable than their Ga(III) mono-or bis(thiosemicarbazonato) complexes in aqueous, acidic solutions, especially under acidic conditions, whereas the corresponding Zn(II) compounds, which were used for optical imaging in living cells, are the most kinetically robust in this series of metal complexes. Their cytotoxicity, fluorescent emissive properties and their radiolabeling versatility with several different radioisotopes renders these mono-(thiosemicarbazones) as versatile synthetic scaffolds for future theranostic agents. These findings pave the way for their more in-depth testing in vitro and in vivo: this class of compounds could be of relevance in the design and synthesis of new tracers with theranostic potential for preclinical and clinical biomedical research.

■ EXPERIMENTAL SECTION
All chemicals and solvents were reagent grade and used as received (Sigma, Aldrich) unless otherwise specified. Highpurity or HPLC grade solvents were obtained from Aldrich Chemical Co. (Gillingham, UK) and/or VWR (Radnor, PA, USA). Milli-Q water was obtained from a Millipore Milli-Q purification system and anhydrous solvents were obtained from a PS-400-7 Innovative technologies SPS drying system. The deuterated solvents were purchased from Aldrich and dried over 4 Å molecular sieves.
Microwave reactions were conducted in a Biotage (Uppsala, Sweden) Initiator 2.5 reactor (0−450 W depending on T) in 20 mL glass capped vials. The reaction mixture was prestirred for 30 s and then heated for the selected time. Generally, if the irradiation power is not set, it reaches its maximum (300 W from magnetron at 2.45 GHz) at the start of the reaction until the target temperature is reached, decreasing to lower values afterward.
General Procedure A for the Synthesis of Aromatic Mono(Thiosemicarbazone) Ligands by Microwave-Assisted Heating. Aromatic diketone (1 equiv) and thiosemicarbazide (0.9−1 equiv) were charged in a microwave vial in ethanol (5−10 mL). The mixture was sonicated for 3 min to homogenize the dispersion and 3 drops of concentrated hydrochloric acid added. The vial was capped and heated under microwave irradiation at 90°C for 10 min. The solid was filtered while hot, washed with ethanol and diethyl ether, and dried under vacuum.
General Procedure B for the Synthesis of Zn(II) Mono(thiosemicarbazonato) Ligands by Microwave-Assisted Heating. The corresponding ligand (1 equiv, exact quantities given in each case, below) and anhydrous zinc acetate (1 equiv, exact quantities given in each case, below) were suspended in ethanol (5 mL) and the mixture was homogenized by ultrasonication. The reaction mixture was heated for 1 h at 90°C under microwave irradiation and subsequently filtered while hot. The resulting solid was washed with ethanol, then CH 2 Cl 2 and dried under vacuum. Further details are given below and in SI.
HPLC method A was performed in a Dionex Ultimate 3000 HPLC instrument with a UV−vis diode array detector measuring at eight wavelengths between 200 and 800 nm. NMR spectroscopy was performed using a Bruker (Banner Lane, UK) Advance NMR spectrometer and/or a 500 MHz Agilent automated system. Spectra were acquired at 500 MHz for 1 H NMR, at 125 MHz for 13 C{ 1 H}NMR at 298 K, unless otherwise stated. Chemical shifts δ are reported in ppm and coupling constants (J) are reported in Hertz (Hz) with a possible discrepancy ≥0.2 Hz. Chemical shifts of solvent residues were identified as follows: CDCl 3 : 1 H, δ = 7.26, 13 C, δ = 77.0; d 6 -DMSO 1H, δ = 2.50; 13 C, δ = 39.5; D 2 O: 1 H, δ = 4.79). Peak multiplicities in the assignments hereby are as follows: s, singlet; d, doublet; t, triplet; q, quartet; m, multiplet; brs, broad signal.
Accurate Mass Spectrometry was carried out at the EPSRC National Mass Spectrometry Centre of Swansea University, UK, using MALDI, ESI and EI modes, also Atmospheric solids analysis probe (ASAP) using API ionization method.
The IR spectra were recorded on a PerkinElmer (Waltham, Massachusetts) Frontier FTIR spectrometer, in the range between 650 and 4000 cm −1 with a resolution of 4 cm −1 . UV−visible spectra were obtained using a Lamda 650 PerkinElmer Spectrometer in DMSO and processed using UV Winlab 3 software. The orientation of the 1.00 cm quartz cuvette was the same for each experiment for consistency. Fluorescence spectra and excitation−emission maps were measured in a LS55 PerkinElmer luminescence spectrophotometer using a 1.00 cm ACS Omega http://pubs.acs.org/journal/acsodf Article quartz cuvette. A scan from 250−750 nm with increments of 50 nm was initially carried out to discover excitation wavelength of maximum emission (λ ex-max ).
Synthesis of Zinc Mono(4-Allyl-3-thiosemicarbazone) Aceanthrenequinone Zn(AA-Allyl) 2 . The ligand AA-Allyl (0.048 g, 0.14 mmol) and anhydrous zinc acetate (0.027 g, 0.14 mmol) were heated in ethanol for 1 h at 90°C under microwave irradiation. Resuspension of the solid in diethyl ether yielded a red colored compound that was further washed with Et 2 O to remove impurities and dried under reduced pressure. The product was obtained as a red solid in quantitative yield. Synthesis of Zn Mono(4-allyl-3-thiosemicarbazonato) Pyrene-4,5-dione (ML 2 ). Compound PY-Allyl (0.050 g, 0.14 mmol) and anhydrous zinc acetate (0.026 g, 0.14 mmol) were heated in ethanol for 1 h at 90°C under microwave irradiation, according to the general procedure, above. The product was obtained as a dark red solid, in quantitative yield.  2 ]. The ligand AN-H (0.035 g, 0.14 mmol) and anhydrous zinc acetate (0.027 g, 0.14 mmol) were heated in ethanol for 1 h at 90°C under microwave irradiation. The mixture was filtered; then, it was resuspended in diethyl ether and filtered. This yielded a red colored compound that was further washed with Et 2 O to remove impurities and dried under reduced pressure. The product was obtained as a red solid after the concentration of volatiles and washing with Et 2 O in quantitative yield.
Mass spectrum: NSI-MS calc. for C 26 2 ]. The ligand AA-H (0.043 g, 0.14 mmol) and anhydrous zinc acetate (0.027 g, 0.14 mmol) were heated in ethanol for 1 h at 90°C under microwave irradiation. The mixture was filtered, then resuspended in diethyl ether and filtered. This yielded a red colored compound that was further washed with Et 2 O to remove impurities and dried under reduced pressure. The desired product was obtained in quantitative yield. 1 H NMR (400 MHz, d 6

Zn(II)(Mono(4-Ethyl-3-thiosemicarbazonato) Aceanthrenequinone) 2 [Zn(AA-Et) 2 ].
The ligand AA-Ethyl (0.047 g, 0.14 mmol) and anhydrous zinc acetate (0.027 g, 0.14 mmol) were heated in ethanol for 1 h at 90°C under microwave irradiation, according to the general procedure, given above. The mixture was filtered, then resuspended in diethyl ether and filtered. This yielded a red colored compound that was further washed with Et 2 O to remove impurities and dried under reduced pressure. The product was obtained as a red solid (0.030 g, 63%).  2 ]. The compound PH-H (0.045 g, 0.16 mmol) and anhydrous zinc acetate (0.028 g, 0.16 mmol) were heated together in ethanol for 1 h at 90°C under microwave irradiation following the general procedure given above. Resuspension of the solid in diethyl ether yielded a red colored compound that was further washed with Et 2 O to remove impurities and dried under reduced pressure. The product was obtained as an orange solid in quantitative yield.
wave vial was filled with the ligand AN-Ph (0.050, 0.15 mMol), anydrous copper acetate (0.03g, 0.15 mMol), and 10 mL of EtOH. Immediately after the addition of EtOH, the reaction mixture turned red-brown and a dark brown precipitate started to form after ca. 2 minutes under mild sonication. Then, the sample was subjected to microwave irradiation (according to the general procedure) at 90°C for 1 h. The slurry was left to cool down to room temperature filtered under air, washed with Et 2 O, and dried under reduced pressure. (Note: For the reaction carried out in the 2:1 Ligand:Cu(II) ratio, 0.10 g (0.30 mMol) of AN-Ph ligand and anhydrous Cu(OAc) 2 (0.03 g, 0.15 mMol) together with 10 mL of EtOH were used, and the reaction under microwave-assisted irradiation conditions proceeded according to the general protocol given above. The desired product, denoted Cu(AN-Ph) 2 was obtained in quantitative yield). Mass spectrometry for the products from both reactions gave rise to almost identical spectra. Alternative Procedure for the Synthesis of Copper Mono(4-Phenyl-3-thiosemicarbazone) Acenaphthenequinone [Cu(AN-Ph) 2 ]. One equivalent of compound AN-Phenyl (0.033 g, 0.100 mmol) and one equivalent of copper acetate (0.018 g, 0.100 mmol) were suspended in 20 mL of MeOH and stirred for 3 h at room temperature. The majority of the solvent was removed using the rotavapor; then diethyl ether was dropwise added until a precipitate formed, which was filtered, washed with diethyl ether, and then dried under vacuum. The final product was obtained as a dark red color solid, which was investigated by EPR. Alternative Procedure for the Synthesis of Copper Mono(4-Phenyl-3-thiosemicarbazone) Acenaphthenequinone [Cu(AN-Ph) 2 ]. Two equivalent of compound AN-Phenyl (0.066 g, 0.200 mmol) and one equivalent of copper acetate (0.018 g, 0.100 mmol) were suspended in 20 mL of MeOH and stirred for 3 h at room temperature. The majority solvent was removed using the rotavapor; then, diethyl ether was dropwise added until a precipitate formed, which was then filtered, washed with diethyl ether, and then dried under vacuum. The final product was obtained as a dark red solid, which was investigated by EPR.
Mass spectrum: ESI-MS calc. for C 38  final concentration of Cu(II) complex) in an aerobic condition. Samples containing ∼5 mM/625 μM of Cu(II) complexes and polycrystalline powders of samples were transferred into 4 mm outer diameter/3 mm inner diameter Suprasil quartz EPR tubes (Wilmad LabGlass) and frozen in liquid N 2 . All EPR samples were measured on a Bruker EMXplus EPR spectrometer equipped with a Bruker ER 4112SHQ X-band resonator. Sample cooling was achieved using a Bruker Stinger cryogen free system mated to an Oxford Instruments ESR900 cryostat, temperature control was maintained using an Oxford Instruments MercuryITC. 38−40 The optimum conditions used for recording the spectra are given below: microwave power 30 dB (0.219 mW), modulation amplitude 5 G, time constant 82 ms, conversion time 16.67 ms, sweep time 60 s, receiver gain 30 dB, and an average microwave frequency of 9.368 GHz. All EPR spectra were measured as frozen solutions at 20 K, respectively. The analysis of the continuous wave EPR spectra and simulations were performed using EasySpin toolbox (5.2.35) for the Matlab (MATLAB_R2022a) program package. 35 Computational Details. All calculations were carried out using density functional theory (DFT) as implemented in the Gaussian 09 package. 42 A variety of exchange correlation functionals and basis sets were used (see ESI for structural parameters together with available experimental details). The optimum exchange correlation functional and the basis set found for ligands incorporating only C, N, O, S, and H were Perdew− Burke−Ernzerhof (PBE) and 6-31G**. For the compounds incorporating Cu and Zn centers, we used PBE exchange correlation functional and aug-cc-PVTZ-pp basis set as implemented in this code. This combination resembled well the experimental structures from X-ray diffraction studies and the Mulliken analysis 43 was used to estimate the charges on the atoms in ligands and metal complexes well. Supporting Information contains main structural parameters of the metal complexes modelled with a range of different functionals and basis sets. The final structures were optimized using PBE exchange correlation functional. For Cu and Zn, aug-cc-PVTZpp basis sets were used. For ligands consisting of C, N, O, S, and H, 6-31G** basis sets were used, and corresponding.xyz files are also provided as Supporting Information.
In Vitro Assays. The human prostate cancer cells (PC-3) and the human cervical cancer cells (HeLa) were purchased from American Type Culture Collection (ATCC). Cell culture was performed in Eagle's Minimum Essential Medium (EMEM) for HeLa, RPMI-1640 medium for PC-3. The media contained fetal calf serum (FCS) (10% for HeLa and PC-3, and 15% for FEK-4), 0.5% penicillin/streptomycin (10,000 IU mL −1 /10,000 mg mL −1 ), and 200 mM L-glutamine (5 mL). All steps were performed in absence of phenol red. Cells were cultured at 37°C in 5% CO 2 incubator in T75 flasks until 60−70% confluency and passaged by trypsinization. Cells were then counted using a hemocytometer and then seeded as appropriate for the necessary assays, as follows: Cellular Imaging. Experiments were carried out in the PC-3 cell line. Cells were cultured as above, then seeded in 35 mm glass bottom Petri-dishes at a density of 2 × 10 5 cells/dish and cultured at least 48 h prior to the recording of control data in untreated cells, and cells incubated with the compounds: Zn(AN-Allyl) 2 , Zn(AA-Allyl) 2 , Zn(PH-Allyl) 2 and Zn(PY-Allyl) 2 , at a final concentration of 100 μM (1% DMSO and 99% conditioned media) incubated for 20 min. The media was replaced by a phenol free serum-free medium before the image capturing.
Systems Inc., Florida, USA) and a Laura 3 software (LabLogic, Sheffield, UK). The gradient elution was 0.1% TFA in milli-Q water as solvent A and 0.1% TFA in MeCN as solvent B. A reverse gradient was applied starting with A at 95% for 2 min, going up to 5% A at 12 min, isocratic level until 14 min and gradient until 95% A at 16 min, then hold to 25 min (Method D).
In most cases, the radiotraces obtained after purification indicated the presence of two new major copper-64 species and that of one other minor species, with Rt generally ranging between 10 and 17 min . The analysis of peak integrals indicated that the overall radio-incorporation yield was generally high for all ligands featuring the AN backbone investigated hereby, and, for these, virtually no traces of the unbound 64-Copper were found in the expected region (Rt ca. 2.5 min).

■ ASSOCIATED CONTENT Data Availability Statement
The data that supports the findings of this study are available in the supplementary material of this article or from the authors.