Vanadium(V) Complexes with Siderophore Vitamin E-Hydroxylamino-Triazine Ligands

: Novel vitamin E chelate siderophore derivatives and their V V and Fe III complexes have been synthesised and the chemical and biological properties have been evaluated. In particular, the α - and δ -tocopherol derivatives with bis-methyldroxylamino triazine ( α -tocTHMA) and ( δ -tocDPA) as well their V V complexes, [V 2V O 3 ( α -tocTHMA) 2 ] and [V 2IV O 3 ( δ -tocTHMA) 2 ], have been synthesised and characterised by infrared (IR), nuclear magnetic resonance (NMR), electron paramagnetic resonance (EPR) and ultra violet-visible (UV-Vis) spectroscopies. The dimeric vanadium complexes in solution are in equilibrium with their respefrctive monomers, H 2 O + [V 2V O 2 ( µ -O)] 4+ = 2 [V V O(OH)] 2+ . The two amphiphilic vanadium complexes exhibit enhanced hydrolytic stability. EPR shows that the complexes in lipophilic matrix are mild radical initiators. Evaluation of their biological activity shows that the compounds do not exhibit any signiﬁcant cytotoxicity to cells.

α-Tocopherol acts in biological organisms as a strong lipophilic antioxidant, without any other biological activity. However, the vitamin E (tocopheryl and tocotrienyl) derivatives, such as α-tocopheryl succinate, have anticancer properties [26][27][28][29][30][31][32][33][34]. The hydrophobic domain of the vitamers of vitamin E is responsible for docking the agents in circulating lipoproteins and biological membranes [35]. Conjugate molecules of vitamin E vitamers with pharmaceuticals, such as metal complexes, can be used to transfer the drug in the active site of vitamin E vitamers, inducing biological responses.
Recently, we reported the first study of the synthesis of complexes comprising tocopherol ligating to metals [36]. The ligands in this study are β-tocopherol molecules substituted with chelate groups in o-position derivatives (Scheme 1, H 3 β-tocDEA), thus, enabling coordination of the metal ion from the phenolic oxygen. The [V V O(β-tocDEA)] has been found to be cytotoxic to cancer cells. Some of the features of these amphiphilic Herein, we have attached a siderophore moiety on the phenoxy oxygen of the chromanol (Scheme 1), forming two new ligands, the 2,4-dichloro-6-(((R)-2,5,7,8-tetramethyl-2-((4R,8R)-4,8,12-trimethyltridecyl)chroman-6-yl)oxy)-1,3,5-triazine (H 2 α-tocTHMA) and 2,4dichloro-6-(((R)-2,8-dimethyl-2-((4R,8R)-4,8,12-trimethyltridecyl)chroman-6-yl)oxy)-1,3,5triazine(H 2 δ-tocTHMA). The labile hydrogen atom of the hydroxy group has been replaced with the triazine moiety forming an inert ether bond and, thus, the new organic molecules will act as ligand owing lower antioxidant activity than free tocopherols; the formation of the tocopheryl radical requires deprotonation of chromanol group. In addition, the high lipophilicity of both the tocopherol derivatives and their complexes assures easy penetration in cell membranes [38]. As chelate group for V V we have chosen the siderophore hydroxylamino-triazine (Scheme 1), targeting to enhance the hydrolytic stability of the V V complexes as much as possible. This chelate group, for example in the ligand H 2 bihyat (Scheme 1) forms very strong complexes with hard acids such as Fe III , V V , Mo VI and U VI [41][42][43][44], with Fe III and U VI to exert the higher affinity for this chelate coordination. The V V complexes of this study exhibiting a chromanol hydroxy group unavailable for coordination, present no significant toxicity to cells. These results are in contrast to the high toxicity of the previous reported vanadium complexes, in which the vanadium ion was coordinated directly with the hydroxy group of the chromanol [36].

Reagents
All reagents were purchased from Aldrich and Merck, (Kenilworth, NJ, USA). Vanadium complexes used for cell viability studies were dissolved in dimethyl sulfoxide (DMSO). DMSO was also used as vehicle control. Microanalyses for C, H and N were performed using a Euro-Vector EA3000 CHN elemental analyser (Milan, Italy). Infrared (IR) spectra were recorder on a Shimadzu Prestige 21, 7102 Riverwood Drive, Columbia, Maryland 21046, U.S.A. MALDI-TOF mass spectra were recorded on a Bruker Autoflex III Smartbeam (Billerica, MA, USA) instrument using α-Cyano-4-hydroxycinnamic acid (HCCA) as matrix.
2nd method: α-tocTCL (3.00 g, 5.18 mmol) was dissolved in THF (120 mL) at 0 • C. A cooled (0 • C) solution of N-Methylhydroxylamine hydrochloride (1.80 g, 21.5 mmol) and sodium hydroxide (0.86 g, 2.2 mmol) in water (10 mL) was added dropwise to the above solution. The reaction mixture was kept under stirring at room temperature for 4 days. Then, it was evaporated under vacuum to dry, and the residue was dissolved in chloroform and filtrated to remove the insoluble in chloroform NaCl. The solution was evaporated under vacuum to dry yielding H 2 α-tocTHMA, 2.0 g, 66% as an orange-brown oil. 1  Synthesis of N,N -(6-(((R)-2,8-dimethyl-2-((4R,8R)-4,8,12-trimethyltridecyl)chroman-6-yl)oxy)-1,3,5-triazine-2,4-diyl)bis(N-methylhydroxylamine) (H 2 α-tocTHMA). H 2 δ-tocTHMA was synthesised following the same methodology as for the synthesis of H 2 α-tocTHMA. The yields were 82% and 53% for the 1st and the 2nd synthetic methods respectively. 1  α-tocTHMA (0.29 g, 0.49 mmol) dissolved in the minimum amount of methanol, was added to the above methanolic solution resulting in a deep brown solution. The solution was stirred for 24 h at room temperature. Then, it was filtered to remove any precipitation, and the filtrate was kept at room temperature for 5 days. During that time a black solid of 1 was formed, which was filtered and dried under vacuum. The yield was 65 mg, 20%.

Spectroscopic Studies
All NMR samples were prepared from the dissolution of the solids in CDCl 3 or 10% DMSO-d 6 :90% D 2 O at room temperature immediately before NMR spectrometric determinations. NMR spectra were recorded on a Bruker Avance 300 spectrometer at 300 MHz for 1 H, 75.4 MHz for 13 C and 78.9 MHz for 51 V NMR. A 30 • -pulse width was applied for both the 1 H and 51 V NMR measurements, and the spectra were acquired with 3000 and 30,000 Hz spectral window, using 1 and 0.1 s relaxation delay respectively. The spectra were analysed using Topspin 4.0 and MultispecNMR 5.0 (https://sourceforge. net/projects/multispecnmr/, accessed on 1 March 2021). 2D [45] grNOESY spectra were obtained by using standard pulse sequences of Bruker Topspin 3.0 software. These spectra were acquired using 256 increments (with 56 scans each) and mixing time 0.43 s.
UV-Vis measurements were recorded on a Photonics UV-Vis spectrophotometer Model 400, equipped with a CCD array, operating in the range 250 to 1000 nm. The spectra were

Reactivity with DPPH •
The rate of DPPH • consumption was measured by UV-vis spectroscopy at 520 nm for 30 min. Stock solutions of each compound (12 mM) were prepared in dry toluene at room temperature. The final concentrations of the compounds were 80-300 µM, while the concentration of DPPH • was 100 µM. The samples were incubated at 25 • C for 4 min. The reaction was initiated by the addition of the DPPH • solution. The samples were measured in triplicate. Second-order rate constants were calculated to determine the radical scavenging activity (RSC) of antioxidants. The decay of DPPH • from the medium has been assumed to follow pseudo-first-order kinetics, under the conditions of the reaction [DPPH • ] 0 , [AH] 0 . One of the reactants is in large excess compared to the other, so the concentration of the minor component decreased exponentially [46]. The [DPPH • ] concentration is calculated from Equation (1): where [DPPH • ] is the radical concentration at time t, and [DPPH • ] 0 is the radical concentration at time zero, and k obsd is the pseudo-first-order rate constant. The pseudo-first-order rate constant k obsd was linearly dependent on the concentration of antioxidants [AH], and from the slope of their plot, second-order rate constants (k 2 ) were calculated to evaluate the radical scavenging capacity of each compound.

Measurement of Oxidative Inducing Effect of Vanadium Compounds by EPR Spectroscopy
An ELEXSYS E500 Bruker EPR spectrometer operating at cw X-band, resonance fre-quency~9.5 GHz and modulation frequency 100 MHz was used. The resonance frequency was accurately measured with solid DPPH (g = 2.0036). The EPR oxidative inducing effect experiments were conducted by monitoring the evolution of α-tocopheryl radicals versus time [37] at room temperature. The assays were prepared in a 5 mm quartz tube by adding 100 µL or 150 µL of a CHCl 3 stock solution (4.95 mM) of complex to 0.500 g of a commercial extra virgin olive oil. Radical initiators are the 1, 2 and [VO(C18DEA)] whereas the addition step consists of the initial time of the reaction, time = 0 min. EPR spectra were recorded for 25-or 30-time domains, each one consisting of 50 scans. The spectra were processed using appropriate software, MultispecEPR 5.0 (https://sourceforge.net/projects/multispecepr/, accessed on 1 March 2021).

Measurement of Cell Viability
Cells were plated in 96-well plates at density 5 × 10 3 cells/well and treated with the ligands or the complexes for 24 and 48 h. Cell viability was measured after incubation of each well with 50 µL of MTT (stock solution of 3 mg/mL) for 3 h and absorbance was determined at 570 nm (background absorbance measured at 690 nm) using a microplate spectrophotometer (Multiskan Spectrum, Therno Fisher Scientific, Waltham, MA, USA). All experiments were performed in triplicate.
Stock solutions of H 2 α-tocTHMA and H 2 δ-tocTHMA and 1-4 prepared in pure DMSO were diluted into the culture medium so that the final concentration of DMSO was less than 1%. The same amount of DMSO was added to the control sample. Stock solutions were kept at 4 • C.

Synthesis and Characterisation
The triazine tocopherol molecules H 2 α-tocTHMA and H 2 δ-tocTHMA were synthesized by two-step substitution reactions of cyanuric chloride. The synthetic process is summarised in Scheme 2.
Cells were plated in 96-well plates at density 5 × 10 cells/well and treated with the ligands or the complexes for 24 and 48 h. Cell viability was measured after incubation of each well with 50 μL of MTT (stock solution of 3 mg/mL) for 3 h and absorbance was determined at 570 nm (background absorbance measured at 690 nm) using a microplate spectrophotometer (Multiskan Spectrum, Therno Fisher Scientific, Waltham, MA, USA). All experiments were performed in triplicate.
Stock solutions of H2α-tocTHMA and H2δ-tocTHMA and 1-4 prepared in pure DMSO were diluted into the culture medium so that the final concentration of DMSO was less than 1%. The same amount of DMSO was added to the control sample. Stock solutions were kept at 4 °C.

Synthesis and Characterisation
The triazine tocopherol molecules H2α-tocTHMA and H2δ-tocTHMA were synthesized by two-step substitution reactions of cyanuric chloride. The synthetic process is summarised in Scheme 2. Reaction of equimolar quantities of H2α-tocTHMA or H2δ-tocTHMA with [V IV O(acac)2] or V IV OSO4 results in the formation of V V complexes 1 and 2 (Scheme 3). The V IV is oxidised to V V by the atmospheric O2. X-band EPR spectroscopy of frozen CHCl3 solutions of either 1 or 2 did not exhibit any signal supporting that all V IV has been oxidised to V V .
The complexes were characterised by elemental analysis, UV-vis, IR and NMR spectroscopies. The structures of the vanadium complexes are based on the data of the experimental analysis and the X-ray structures of the respective V V -bihyat 2− complexes [42].

Complexes Characterisation by IR
The IR spectra of 1 and 2 are shown in Figure 1. Both 1 and 2 gave a 966 cm −1 attributed to the stretching of the V=O bond. The peaks at 798 cm −1 Scheme 3. Synthetic route for the vanadium complexes. The complexes were characterised by elemental analysis, UV-vis, IR and NMR spectroscopies. The structures of the vanadium complexes are based on the data of the experimental analysis and the X-ray structures of the respective V V -bihyat 2− complexes [42].

Complexes Characterisation by IR
The IR spectra of 1 and 2 are shown in Figure 1. Both 1 and 2 gave a strong peak at 966 cm −1 attributed to the stretching of the V=O bond. The peaks at 798 cm −1 are characteristic to V-O-V stretching vibrations [48,49], thus, confirming the dinuclear structure of the complex. The complexes also show two strong stretching N-O vibrations shifted around 80 cm −1 at higher energy compared to the free ligand suggesting coordination of the metal ion from the hydroxylamine-triazine chelating group.

Complexes Characterisation by IR
The IR spectra of 1 and 2 are shown in Figure 1. Both 1 and 2 gave a strong peak at 966 cm −1 attributed to the stretching of the V=O bond. The peaks at 798 cm −1 are characteristic to V-O-V stretching vibrations [48,49], thus, confirming the dinuclear structure of the complex. The complexes also show two strong stretching N-O vibrations shifted around 80 cm −1 at higher energy compared to the free ligand suggesting coordination of the metal ion from the hydroxylamine-triazine chelating group.

Complex Characterisation by 51 V NMR, 2D { 1 H} grNOESY
The 51 V spectra of each of the V V complexes (1, 2) in CDCl 3 solutions gave two signals at −216 and −387 ppm ( Figure 2). The intensity of the peaks is dependent on the concentration of the complexes in solution. At low concentration (i.e., 1 mM) the component at −216 ppm is the major, whereas at more concentrate solutions (i.e., 7 mM) the spectra of each of the complexes shows only the peak at −387 ppm. For more concentrated solutions three broad additional peaks of equal intensity appear at higher field (−402, −439 and −648 ppm), presumably originated from a higher nuclearity compound. The 51 V NMR spectra changes observed by the variation of the concentration are attributed to the equilibrium between the monomer (1m), dimer (1)  each of the complexes shows only the peak at −387 ppm. For more concentrated solutions three broad additional peaks of equal intensity appear at higher field (−402, −439 and −648 ppm), presumably originated from a higher nuclearity compound. The 51 V NMR spectra changes observed by the variation of the concentration are attributed to the equilibrium between the monomer (1m), dimer (1)    The 2D { 1 H} grNOESY of 1 is shown in Figure 3. The two methyl groups show difference in chemical shifts due to the different chemical environment. 2D { 1 H} grNOESY shows positive cross peaks between the protons of the two methyl groups assigned to the slow rotation methylhydroxylamine giving two peaks at 2.869 and 2.573 ppm in proton NMR. The rotation of tocopherol is performed around the ether bond between tocopherol and triazine moieties (Figure 3). The 2D { 1 H} grNOESY of 1 is shown in Figure 3. The two methyl groups show difference in chemical shifts due to the different chemical environment. 2D { 1 H} grNOESY shows positive cross peaks between the protons of the two methyl groups assigned to the slow rotation methylhydroxylamine giving two peaks at 2.869 and 2.573 ppm in proton NMR. The rotation of tocopherol is performed around the ether bond between tocopherol and triazine moieties (Figure 3).
The 2D { 1 H} grNOESY of 1 is shown in Figure 3. The two methyl groups show difference in chemical shifts due to the different chemical environment. 2D { 1 H} grNOESY shows positive cross peaks between the protons of the two methyl groups assigned to the slow rotation methylhydroxylamine giving two peaks at 2.869 and 2.573 ppm in proton NMR. The rotation of tocopherol is performed around the ether bond between tocopherol and triazine moieties (Figure 3).

Complexes Characterisation by UV-Vis
The UV-vis spectra of the CHCl 3 solutions of 1 and 2 are shown in  1m and 2m, it does not exhibit any strong absorption peaks in the visible region. Thus, the strong colour of 1m and 2m is due to electron transitions from the chromanol ring to the metal. Chromanol ring can contribute electronically to the metal ion through the resonance of the triazine ring (Scheme 5) [42]. The shift of the UV-vis peaks of 1 to lower energy compared to those of 2 agrees with the higher electron density of αthan δ-tocopherol, supporting our hypothesis regarding the significance of the chromanol role on the LMCT effect.

Complexes Characterisation by UV-Vis
The UV-vis spectra of the CHCl3 solutions of 1 and 2 are show strong peaks in the visible region [λ(ε) of 1 = 493 nm (3600 M −1 cm −1 ), λ(ε) of 2 = 484 nm (2300 M −1 cm −1 ), 645 nm (860 M −1 cm −1 )] a metal charge transfer transitions (LMCT). The concentration of th were 0.500 mM, and according to 51 V NMR the species in the solut structure, 1m and 2m (Scheme 4). Although [V V O2(bihyat)] − has the monomers 1m and 2m, it does not exhibit any strong absorp region. Thus, the strong colour of 1m and 2m is due to electron tr manol ring to the metal. Chromanol ring can contribute electro through the resonance of the triazine ring (Scheme 5) [42]. The s of 1 to lower energy compared to those of 2 agrees with the high than δ-tocopherol, supporting our hypothesis regarding the signi role on the LMCT effect.  The CHCl 3 solutions of Fe III complexes 3 and 4 gave peaks at 560 nm (3200 M −1 cm −1 ) and 535 nm (2200 M −1 cm −1 ) respectively ( Figure 5A). These spectra are similar to other hydroxylamine-triazin iron complexes, for example the Fe III -bihyat compounds [41]. Complexes 3 and 4 exhibit the same pattern as the vanadate complexes; the α-tocopherol complex 3 absorbs at lower energy than the δ-tocopherol complex 4. This is in line with the proposed electron transfer resonance mechanism proposed in Scheme 5. plex 3 absorbs at lower energy than the δ-tocopherol complex 4. This is in line with the proposed electron transfer resonance mechanism proposed in Scheme 5.
Addition of various quantities of either H2α-tocTHMA or H2δ-tocTHMA to a CHCl3 solution of Fe III Cl3 gave the same spectra with 3 and 4 respectively. Titration of the CHCl3 solution of Fe III Cl3 with either H2α-tocTHMA ( Figure 5B) or H2δ-tocTHMA reveal that only the 1:2 Fe III -Ligand complexes are formed in the solution.

Characterisation of the Complexes in 10% DMSO:90% D2O Solutions by 51 V NMR
The 51 V NMR spectra of 10% DMSO:90% D2O solutions of inorganic vanadate with either 1 or 2 at pD = 5.0-7.5 clearly shows very different chemical shifts for the peaks of vanadate oligomers from those of the complexes, undoubtedly assigning the peaks at −560 ppm to the new vanadium complexes (Figures 6 and 7). The addition of D2O in the DMSO solutions (10% DMSO:90% D2O, pD = 6.0-7.5) of 1 or 2 at concentrations 0.10 mM do not hydrolyse the complexes as evidenced by the 51 V NMR spectroscopy (the spectra show only one peak originated from the complex and there is no formation of any inorganic vanadate species (Figure 6)). In the 51 V NMR spectra, a shift from −387 ppm in CDCl3 to −560 ppm in 10% DMSO:90% D2O observed for 1 originated from the structural change of the complex from tetragonal pyramidal to dioxido octahedral geometry. A similar shift was observed upon changing from CDCl3 to D2O solutions for [VO2(bihyat)] − as well [42]. The 10% DMSO:90% D2O solutions of 1 or 2 were stable at these conditions for more than 72 h. The high hydrolytic stability of 1 and 2 is attributed to their amphiphilic nature Addition of various quantities of either H 2 α-tocTHMA or H 2 δ-tocTHMA to a CHCl 3 solution of Fe III Cl 3 gave the same spectra with 3 and 4 respectively. Titration of the CHCl 3 solution of Fe III Cl 3 with either H 2 α-tocTHMA ( Figure 5B) or H 2 δ-tocTHMA reveal that only the 1:2 Fe III -Ligand complexes are formed in the solution.

Characterisation of the Complexes in 10% DMSO:90% D 2 O Solutions by 51 V NMR
The 51 V NMR spectra of 10% DMSO:90% D 2 O solutions of inorganic vanadate with either 1 or 2 at pD = 5.0-7.5 clearly shows very different chemical shifts for the peaks of vanadate oligomers from those of the complexes, undoubtedly assigning the peaks at −560 ppm to the new vanadium complexes (Figures 6 and 7). The addition of D 2 O in the DMSO solutions (10% DMSO:90% D 2 O, pD = 6.0-7.5) of 1 or 2 at concentrations 0.10 mM do not hydrolyse the complexes as evidenced by the 51 V NMR spectroscopy (the spectra show only one peak originated from the complex and there is no formation of any inorganic vanadate species (Figure 6)). In the 51 V NMR spectra, a shift from −387 ppm in CDCl 3 to −560 ppm in 10% DMSO:90% D 2 O observed for 1 originated from the structural change of the complex from tetragonal pyramidal to dioxido octahedral geometry. A similar shift was observed upon changing from CDCl 3 to D 2 O solutions for [VO 2 (bihyat)] − as well [42]. The 10% DMSO:90% D 2 O solutions of 1 or 2 were stable at these conditions for more than 72 h. The high hydrolytic stability of 1 and 2 is attributed to their amphiphilic nature [36,38,39]. The lipophilicity of 1 and 2 may enhance the hydrolytic stabilisation over the non-lipophilic vanadium complexes, exhibiting the same coordination environment, through a more favourable solvation [40].

Characterisation of the Complexes in 10% DMSO:90% D2O Solutions by UV-Vis Spectroscopy
The UV-vis spectra of the 10%DMSO: 90%D2O solutions of 1 and 2 were simila the spectra of the complexes in CHCl3 (Figure 8). The only difference between the spec  [36,38,39]. The lipophilicity of 1 and 2 may enhance the hydrolytic stabilisation over the non-lipophilic vanadium complexes, exhibiting the same coordination environment, through a more favourable solvation [40].

Characterisation of the Complexes in 10% DMSO:90% D2O Solutions by UV-Vis Spectroscopy
The UV-vis spectra of the 10%DMSO: 90%D2O solutions of 1 and 2 were similar to the spectra of the complexes in CHCl3 (Figure 8). The only difference between the spectra

Characterisation of the Complexes in 10% DMSO:90% D 2 O Solutions by UV-Vis Spectroscopy
The UV-vis spectra of the 10%DMSO: 90%D 2 O solutions of 1 and 2 were similar to the spectra of the complexes in CHCl 3 (Figure 8). The only difference between the spectra in the two different solvents is the lower intensity of the peaks in 10%DMSO: 90%D 2 O than CHCl 3 . However, complexes 1 and 2 show significant different absorption coefficients compared to those of [VO 2 (bihyat)] − in various solvents [42]. The extinction coefficients of 1 and 2 are lower in protic polar solvents than in the non-polar ones in the same manner as [VO 2 (bihyat)] − . The absorbance values from the UV spectra of 1, 2 solutions appear to obey Beer's law, even at low concentrations at 50 µM, suggesting that the complexes are hydrolytically stable in those solutions (Figure 9), in agreement with 51 V NMR spectroscopy. The spectra of the 10% DMSO:90% D 2 O solutions of the iron complexes 3 and 4 are also similar with their spectra in CHCl 3 solutions.
Inorganics 2021, 9, x 13 of 2 in the two different solvents is the lower intensity of the peaks in 10%DMSO: 90%D2 than CHCl3. However, complexes 1 and 2 show significant different absorption coeff cients compared to those of [VO2(bihyat)] − in various solvents [42]. The extinction coeff cients of 1 and 2 are lower in protic polar solvents than in the non-polar ones in the sam manner as [VO2(bihyat)] − . The absorbance values from the UV spectra of 1, 2 solution appear to obey Beer's law, even at low concentrations at 50 μM, suggesting that the com plexes are hydrolytically stable in those solutions (Figure 9), in agreement with 51 V NM spectroscopy. The spectra of the 10% DMSO:90% D2O solutions of the iron complexes and 4 are also similar with their spectra in CHCl3 solutions.

Reactivity with DPPH •
The radical scavenging activity (RSC) values of the organic compounds and the com plexes towards scavenging the DPPH • radical are shown in Table 1. α-tocTHMA and δ tocTHMA exhibit very low antioxidant activity, much lower than free α-tocopherol. Th reason for this low activity is the replacement of the labile phenoxy proton of α-tocophero in the two different solvents is the lower intensity of the peaks in 10%DMSO: 90%D2O than CHCl3. However, complexes 1 and 2 show significant different absorption coefficients compared to those of [VO2(bihyat)] − in various solvents [42]. The extinction coefficients of 1 and 2 are lower in protic polar solvents than in the non-polar ones in the same manner as [VO2(bihyat)] − . The absorbance values from the UV spectra of 1, 2 solutions appear to obey Beer's law, even at low concentrations at 50 μM, suggesting that the complexes are hydrolytically stable in those solutions (Figure 9), in agreement with 51 V NMR spectroscopy. The spectra of the 10% DMSO:90% D2O solutions of the iron complexes 3 and 4 are also similar with their spectra in CHCl3 solutions.

Reactivity with DPPH •
The radical scavenging activity (RSC) values of the organic compounds and the complexes towards scavenging the DPPH • radical are shown in Table 1. α-tocTHMA and δ-tocTHMA exhibit very low antioxidant activity, much lower than free α-tocopherol. The reason for this low activity is the replacement of the labile phenoxy proton of α-tocopherol

Reactivity with DPPH •
The radical scavenging activity (RSC) values of the organic compounds and the complexes towards scavenging the DPPH • radical are shown in Table 1. α-tocTHMA and δ-tocTHMA exhibit very low antioxidant activity, much lower than free α-tocopherol. The reason for this low activity is the replacement of the labile phenoxy proton of α-tocopherol with an inert ether bond of α-tocTHMA and δ-tocTHMA. Large number of vanadium complexes exhibit radical scavenging activity [50]. However, complexes 1-4 either did not show any or very little decrease of the peak intensity at 520 nm, resulting in the conclusion that they do not have any antioxidant activity. 6.8 ± 0.3 α-tocopherol [33] 560 ± 80

Oxidative Inducing Effect of Vanadium Compounds by EPR Spectroscopy
The ability of the new vanadium compounds to produce radicals was examined by monitoring the generation of α-tocopheryl radicals in olive oil by cw X-band EPR using 2D intensity vs. time experiments (Figure 10). The ability of complexes 1 and 2 were compared with that of the [V V O(C18DEA)] used in a previous study [36,37]. [V V O(C18DEA)] has been studied for its activity towards the production of radicals in olive oil, therefore, it is used in this work as a reference. Based on previous studies, it has been reported that V V and/or V IV coordinated catalytic sites are able to activate phenolics in the lipophilic matrix of oil mediated by dioxygen activation; in this oxidative environment free radicals are trapped by α-tocopherol to give α-tocopheryl radical. The generation of α-tocopheryl radicals is monitored by X-band cw-EPR vs. time. The graph of the signal intensity vs. time is a very useful quantification tool to determine the ability of the complexes to initiate radicals. Experiments were run for two different quantities of each radical initiator for the study, (0.490 µmole or 0.720 µmole). The intensity of the EPR peaks, at the same time period after addition of the radical initiator in olive oil is higher for [V V O(C18DEA)] than 1, meaning that [V V O(C18DEA)] produces more α-tocopheryl radicals than 1. The intensity of the EPR signal is lower at higher concentrations of the radical initiator due to the faster oxidation of the polar antioxidants that regenerate α-tocopherol in olive oil; the mechanism has been previously investigated [37]. Apparently, [V V O(C18DEA)], and consequently [36], are by far much more potent radical initiators than 1.
[V V O(C18DEA)] and [V V O(β-toc)DEA] vanadium complexes have been reported to have high cytotoxicity [36]. If the cytotoxicity of the complexes is related to their oxidative power measured by EPR, then 1 is expected to be less cytotoxic than

Cytotoxic Activity
None of the complexes exerted cytotoxic activity against the three cell lines, a fact that differentiated them significantly from the ligands. As seen in Figure 11, exposure of Cal33 cells for 24 h to increasing concentrations of the ligands and the complexes had no severe effect on the ability of the cells to proliferate ( Figure 11A,C). Prolongation of the incubation time revealed that complexes 1 and 4 exerted a no-dose-dependent cytotoxicity across the different doses (1 to 100 µM) leading to a 40% reduction of cell population. On the contrary, the cytotoxic activity of the ligands H 2 α-tocTHMA and H 2 δ-tocTHMA as well as the complexes 2 and 3 was depicted mainly at doses higher than 25 µM. Order of cytotoxic activity (100 µM) was found as following: H 2 δ-tocTHMA < H 2 α-tocTHMA < 4 < 1 < 3 < 2 (24 h), H 2 δ-tocTHMA = H 2 α-tocTHMA < 2 < 3 < 1 < 4 (48 h).

Cytotoxic Activity
None of the complexes exerted cytotoxic activity against the three cell lines, a fac that differentiated them significantly from the ligands. As seen in Figure 11, exposure o Cal33 cells for 24 h to increasing concentrations of the ligands and the complexes had n severe effect on the ability of the cells to proliferate ( Figure 11A,C). Prolongation of th incubation time revealed that complexes 1 and 4 exerted a no-dose-dependent cytotoxicit across the different doses (1 to 100 μM) leading to a 40% reduction of cell population. O the contrary, the cytotoxic activity of the ligands H2α-tocTHMA and H2δ-tocTHMA a well as the complexes 2 and 3 was depicted mainly at doses higher than 25 μΜ. Order o A similar cytotoxic profile was also seen against Hela cells ( Figure 12). Twenty-four hour of exposure to the ligands and the complexes exerted a mild effect on cell viability even at th highest dose. At 48 h a slightly greater reduction in cell viability was recorded for both ligand and the complexes. Order of cytotoxic activity (100 μΜ): H2α-tocTHMA < 3 < 2 < H2δ-toc THMA < 1 < 4 (24 h), H2α-tocTHMA < 1 < 3 = H2δ-tocTHMA < 2 < 4 (48 h). Against embryonic mouse fibroblasts (NIH/3T3), the complexes exerted minimal tox icity after 24 h of incubation ( Figure 13A,C). H2α-tocTHMA presented a strong cytotoxi A similar cytotoxic profile was also seen against Hela cells ( Figure 12). Twenty-four hours of exposure to the ligands and the complexes exerted a mild effect on cell viability even at the highest dose. At 48 h a slightly greater reduction in cell viability was recorded for both ligands and the complexes. Order of cytotoxic activity (100 µM): H 2 α-tocTHMA < 3 < 2 < H 2 δ-tocTHMA < 1 < 4 (24 h), H 2 α-tocTHMA < 1 < 3 = H 2 δ-tocTHMA < 2 < 4 (48 h). A similar cytotoxic profile was also seen against Hela cells ( Figure 12). Twenty-four hours of exposure to the ligands and the complexes exerted a mild effect on cell viability even at the highest dose. At 48 h a slightly greater reduction in cell viability was recorded for both ligands and the complexes. Order of cytotoxic activity (100 μΜ): H2α-tocTHMA < 3 < 2 < H2δ-toc-THMA < 1 < 4 (24 h), H2α-tocTHMA < 1 < 3 = H2δ-tocTHMA < 2 < 4 (48 h). Against embryonic mouse fibroblasts (NIH/3T3), the complexes exerted minimal toxicity after 24 h of incubation ( Figure 13A,C). H2α-tocTHMA presented a strong cytotoxic effect after 48 h, a pattern similar to that seen against Cal33 cells. On the contrary, Against embryonic mouse fibroblasts (NIH/3T3), the complexes exerted minimal toxicity after 24 h of incubation ( Figure 13A,C). H 2 α-tocTHMA presented a strong cytotoxic effect after 48 h, a pattern similar to that seen against Cal33 cells. On the contrary, complexes 1 and 3 had no effect on cell viability and remained relatively non-toxic ( Figure 13B). Complex 4 maintains the same cytotoxic profile, as seen in HeLa and Cal33 cells, exerting a mild reduction in cell viability across the range of doses (1-100 µM) whereas H 2 δ-tocTHMA and 2 were cytotoxic at concentrations higher than 25 µM ( Figure 13D). Order of cytotoxic activity (100 µM) was found as follows: 2 > H 2 δ-tocTHMA > 1 = H 2 α-tocTHMA > 3 > 4 (24 h), H 2 δ-tocTHMA = 2 > H 2 α-tocTHMA > 1 > 4 > 3 (48 h).

Conclusions
Stepwise substitution reactions of cyanuric chloride with α-or δ-tocopherol and then with N-methylhydroxylamine resulted in the synthesis of the amphiphilic H2α-tocTHMA and H2δ-tocTHMA ligands. Reaction of the ligands with either the V IV starting materials [V IV O(acac)2] or V IV OSO4 afforded the complexes 1 and 2. Reaction of Fe III Cl3 with H2α-tocTHMA or H2δ-tocTHMA resulted in the formation of complexes 3 and 4 respectively. The new compounds have been characterised by NMR, UV/Vis and infrared spectroscopies. The RSC activities for all compounds have been determined by the DPPH • assay and the results showed than none of the molecules, ligands or complexes, exhibits antioxidant activity. On the contrary, EPR spectroscopy showed that 1 and 2 are radical initiators. All complexes exhibit high hydrolytic stability even at low concentrations similar to those used in cell viability studies. All complexes, 1-4, do not exerted significant cytotoxic activity against NIH/3T3, Cal33 and HeLa cell lines. The low cytotoxic activity is attributed to the low antioxidant-prooxidant activity of the tocopherol-triazine conjugate molecules. This is in line with the fact that 1 and 2 are moderate radical initiators. Previous studies support that the tocopherol-metal complexes with the hydroxy group of chromanol accessible to metal ion coordination, are stronger radical initiators than 1 and 2, and they exert high cytotoxic activity. However, we cannot exclude the structural differences of the chelate moieties that might induce various biological responses. Currently, the vitamin E metal complexes are rare, and more work is required, including synthesis of new compounds with specific structural features, in order to understand the mechanism of their reactivity.

Conclusions
Stepwise substitution reactions of cyanuric chloride with αor δ-tocopherol and then with N-methylhydroxylamine resulted in the synthesis of the amphiphilic H 2 α-tocTHMA and H 2 δ-tocTHMA ligands. Reaction of the ligands with either the V IV starting materials [V IV O(acac) 2 ] or V IV OSO 4 afforded the complexes 1 and 2. Reaction of Fe III Cl 3 with H 2 α-tocTHMA or H 2 δ-tocTHMA resulted in the formation of complexes 3 and 4 respectively. The new compounds have been characterised by NMR, UV/Vis and infrared spectroscopies. The RSC activities for all compounds have been determined by the DPPH • assay and the results showed than none of the molecules, ligands or complexes, exhibits antioxidant activity. On the contrary, EPR spectroscopy showed that 1 and 2 are radical initiators. All complexes exhibit high hydrolytic stability even at low concentrations similar to those used in cell viability studies. All complexes, 1-4, do not exerted significant cytotoxic activity against NIH/3T3, Cal33 and HeLa cell lines. The low cytotoxic activity is attributed to the low antioxidant-prooxidant activity of the tocopherol-triazine conjugate molecules. This is in line with the fact that 1 and 2 are moderate radical initiators. Previous studies support that the tocopherol-metal complexes with the hydroxy group of chromanol accessible to metal ion coordination, are stronger radical initiators than 1 and 2, and they exert high cytotoxic activity. However, we cannot exclude the structural differences of the chelate moieties that might induce various biological responses. Currently, the vitamin E metal complexes are rare, and more work is required, including synthesis of new compounds with specific structural features, in order to understand the mechanism of their reactivity.