Experimental and Theoretical Studies of the Factors Affecting the Cycloplatination of the Chiral Ferrocenylaldimine (S_C)-[(η⁵-C₅H₅)Fe{(η⁵-C₅H₄)–C(H)=N–CH(Me)(C₆H₅)}]

Abstract

The study of the reactivity of the enantiopure ferrocenyl Schiff base (S_C)-[FcCH=N–CH(Me)(C₆H₅)] (1) (Fc = (η⁵-C₅H₅)Fe(η⁵-C₅H₄)) with cis-[PtCl₂(dmso)₂] under different experimental conditions is reported. Four different types of chiral Pt(II) have been isolated and characterized. One of them is the enantiomerically pure trans-(S_C)-[Pt{κ¹-N[FcCH=N–CH(Me)(C₆H₅)]}Cl₂(dmso)] (2a) in which the imine acts as a neutral N-donor ligand; while the other three are the cycloplatinated complexes: [Pt{κ²-C,N[(C₆H₄)–N=CHFc]}Cl(dmso)] (7a) and the two diastereomers {(S_p,S_C) and (R_p,S_C)} of [Pt{κ²-C,N[(η⁵-C₅H₃)–CH=N–{CH(Me)(C₆H₅)}]Fe(η⁵-C₅H₅)}Cl(dmso)] (8a and 9a, respectively). Isomers 7a-9a, differ in the nature of the metallated carbon atom [C^Ph (in 7a) or C^Fc (in 8a and 9a)] or the planar chirality of the 1,2-disubstituted ferrocenyl unit (8a and 9a). Reactions of 7a–9a with PPh₃ gave [Pt{κ²-C,N[(C₆H₄)–N=CHFc]}Cl(PPh₃)] (in 7b) and the diastereomers (S_p,S_C) and (R_p,S_C) of [Pt{κ²-C,N[(η⁵-C₅H₃)–CH=N–{CH(Me)(C₆H₅)}] Fe(η⁵-C₅H₅)}Cl(PPh₃)] (8b and 9b, respectively). Comparative studies of the electrochemical properties and cytotoxic activities on MCF7 and MDA-MB231 breast cancer cell lines of 2a and cycloplatinated complexes 7b-9b are also reported. Theoretical studies based on DFT calculations have also been carried out in order to rationalize the results obtained from the cycloplatination of 1, the stability of the Pt(II) complexes and their electrochemical properties.

1. Introduction

Cyclometallated platinum (II) complexes containing bidentate (C,N)⁻ or terdentate ((C,N,E)^q, (N,C,E)^q, (E = N, S or O and q = −1 or −2) or (C,N,C')²⁻) ligands are nowadays one of the most attractive type of organoplatinum compounds [1,2,3,4,5,6,7,8,9,10]. This is mainly due to their physical, chemical, photo-physical and electrochemical properties, or the reactivity of their ground and excited states that make them extremely useful in a wide range of fields [1,2,3,4,5,6,7,8,9,10]. It is well-known that platinacycles are valuable precursors in homogeneous catalysis, as building blocks in supra- and macromolecular chemistry as well as for the synthesis of organic compounds (including natural products) and other organometallic products (i.e., cyclometallated Pt(IV) derivatives) [11,12,13,14,15]. Sensors, detectors, biomarkers, molecular machines (i.e., switches), imaging probes and different sorts of OLEDs (including white and white polymer light-emitting diodes (WOLED and PGOLED), respectively) containing cycloplatinated cores have been described [16,17,18,19,20,21,22,23,24,25]. Furthermore, cyclometallated Pt(II) complexes with antitumoral activity have been known for a long time, but the idea of using this sort of products in drug discovery is becoming more and more popular [15,26,27,28,29,30,31,32,33]. For illustrative purposes recent examples of platinacycles with bidentate (C^Ph, N)⁻ ligands showing greater cytotoxic activity than cisplatin are shown in Figure 1 [29,32,33].

Figure 1. A selection of cyclometallated Pt(II) complexes with a σ(Pt-C^Ph) bond and potent cytotoxic activity reported recently. For illustrative purposes IC₅₀ values (in µM) for A–C in a specific cancer cell line are given: in HCT-116 colon cell line for A compounds: IC₅₀ = 28 ± 2.4 µM (R¹ = H) and 18 ± 1.9 µM (R¹ = Cl); for type B, IC₅₀ in the range 0.95 µM–7.7 µM (IC₅₀ for cisplatin = 40 ± 4.4 µM). Compound C was tested in the A549 breast cancer cell line, IC₅₀ = 4.60 µM (for cisplatin = 8.0 µM) [29,30,31,32].

Figure 1. A selection of cyclometallated Pt(II) complexes with a σ(Pt-C^Ph) bond and potent cytotoxic activity reported recently. For illustrative purposes IC₅₀ values (in µM) for A–C in a specific cancer cell line are given: in HCT-116 colon cell line for A compounds: IC₅₀ = 28 ± 2.4 µM (R¹ = H) and 18 ± 1.9 µM (R¹ = Cl); for type B, IC₅₀ in the range 0.95 µM–7.7 µM (IC₅₀ for cisplatin = 40 ± 4.4 µM). Compound C was tested in the A549 breast cancer cell line, IC₅₀ = 4.60 µM (for cisplatin = 8.0 µM) [29,30,31,32].

It is well-known that the properties of cycloplatinated compounds, their biological or catalytical activities and their applications are strongly dependent on a wide variety of factors, such as the nature of the donor atoms of the metallated ligand, its denticity, the size of the metallacycle, the electronic and steric effects induced by the presence of substituents and their location and even the nature of the ancillary ligands and their relative arrangement [1,2,3,4,5,6,7,8,9,10].

On the other hand, N-donor ferrocenyl ligands are valuable reagents to achieve heterodimetallic Pt (II) compounds [34,35,36,37,38,39,40,41,42,43]. These ligands may produce different types of products depending on the relative arrangement between the N-donor atom and the presence of: (a) one or more σ(C–H) bonds, with the proper orientation as to undergo metallation, and/or (b) additional heteroatoms with good donor abilities and their location [36,37,40,41]. Moreover, for monosubtituted ferrocene derivatives metallation of the C₅H₄ ring induces planar quirality [34,35,38]. Despite the variety of cyclometallated Pt(II) complexes derived from achiral N-donor ligands described so far, enantio- or diastereomerically pure platinacycles are scarce [39,40,41,42,43,44,45,46,47] and those containing a Pt–C^Fc bond are even less common [39,40,41,42,43]. However, the examples described so far are especially outstanding because of their singular properties, behaviors or activities [39,40,41,42,43]. A few representative examples (A–D) are shown in Figure 2. Compound A is an excellent catalyst on the asymmetric aza-Claisen rearrangement of Z configured trifuoroacetaamidates giving high enantioselectivities [39] and diastereomers B and C (Figure 2), form an acid-base dependent chiral molecular switch [40,41]. Platinacycles D and E are potent cytotoxic agents in lung (A549), breast (MDA-MB231) and colon (HCT-116) cell lines. Moreover, complex E: (a) is less toxic than cisplatin and active in other cancer cell lines (i.e., NCI-H460); (b) induces nuclear translocation of endogenous FOXO3a and a FOXO3a reporter protein in A549 and U2OS cells, respectively; and (c) has a synergic effect with cisplatin [42,43].

Figure 2. Representative examples of diastereomerically pure cyclometallated Pt(II) complexes containing a σ(Pt–C^Fc) bond [39,40,41,42,43]. (For types B and C, R is Me or ⁱPr).

Figure 2. Representative examples of diastereomerically pure cyclometallated Pt(II) complexes containing a σ(Pt–C^Fc) bond [39,40,41,42,43]. (For types B and C, R is Me or ⁱPr).

All these findings suggest that compounds containing simultaneously ferrocenyl units and cycloplatinated rings are promising candidates for the design of multifunctional products with both biological activity and interesting electrochemical properties (due to the presence of a redox active center). This is still an emerging topic and further work is needed to clarify some key points. Unfortunately, isomeric platinacycles differing exclusively in the nature of the metallated carbon have not been described so far, and thus the effect produced by the characteristics of the carbon atom bound to the Pt(II) (i.e., C^Fc versus C^Ph) on their properties and activities still remains unknown.

Since during the preparation of D and E (Figure 2), no evidences of the formation of platinacycles arising from the metallation of the phenyl ring were detected [42], we decided to investigate whether for (S_C)-[FcC(H)=N–CH(Me)(C₆H₅)] (1) (Figure 3) [48], the relative orientation between the phenyl ring and the N atom and their proximity could also allow the isolation of platinacycles with a Pt–C^Ph bond.

Figure 3. Schiff base 1 selected for this study and schematic view of the different types of Pt(II) complexes (2–11) that may be formed. Compounds 2-11 differ in the conformation of the ligand (anti-(E) in 2, 4, 6, 8–11 or syn-(Z) in 3, 5 and 7), the binding mode of the imine ((N) in 2–5, (C,N)⁻ in 6–9 or (C,N,C')²⁻ in 10 and 11), the type of metallated C atom (C^Ph in 6 and 7 or C^Fc in 8 and 9), the number of Pt–C bonds contained (one in 6–9 or two in 10 and 11) or the planar chirality of the Fc unit (S_p in 8 and 10 or R_p in 9 and 11). (X, L = monoanionic and neutral ligand, respectively).

Figure 3. Schiff base 1 selected for this study and schematic view of the different types of Pt(II) complexes (2–11) that may be formed. Compounds 2-11 differ in the conformation of the ligand (anti-(E) in 2, 4, 6, 8–11 or syn-(Z) in 3, 5 and 7), the binding mode of the imine ((N) in 2–5, (C,N)⁻ in 6–9 or (C,N,C')²⁻ in 10 and 11), the type of metallated C atom (C^Ph in 6 and 7 or C^Fc in 8 and 9), the number of Pt–C bonds contained (one in 6–9 or two in 10 and 11) or the planar chirality of the Fc unit (S_p in 8 and 10 or R_p in 9 and 11). (X, L = monoanionic and neutral ligand, respectively).

Ligand 1 may exhibit different binding modes and hapticities giving a wide variety of Pt(II) complexes (Figure 3). If 1 binds to the Pt(II) atom through the imine nitrogen exclusively, different isomeric forms of the optically active complexes: [Pt{κ¹-N[FcC(H)=N–CH(Me)(C₆H₅)]}X₂(L)] could be expected (2–4 in Figure 3). They differ in the anti-(E) (in 2 and 3) or the syn-(Z) (in 4 and 5) form of the imine and a trans- (in 2 and 4) or cis- (in 3 and 4) disposition of the X⁻ ligands. Moreover, 1 has several carbon atoms susceptible to undergo metallation. Four types of platinacycles (Figure 3, 6–9) could arise from monometallation of 1. In two of them (6 and 7) there is a σ(Pt–C^Ph) bond, and the imine adopts the anti-(E) (in 6) or syn-(Z) (in 7) forms; while the other two (8 and 9) are diastereomers (S_p,S_C) and (R_p,S_C) arising from platination of the C₅H₃ ring that induces planar chirality [34,35,38].

A few examples of cyclometallated Pt(II) compounds with (C,N,C)²⁻ pincer ligands have been reported [49,50,51,52], and thus the activation of a σ(C^Ph–H] and a σ(C^Fc–H] bond to give compounds 10 or 11 (Figure 3) cannot be disregarded. Furthermore, and since several groups have achieved cylometallated Pt(IV) complexes in the reaction of cis-[PtCl₂(dmso)₂] with N-donor ligands [23,53,54], the formation of the platinum(IV) derivatives cannot be ruled out either.

In this work, we report: (a) the synthesis of chiral Pt(II) complexes in which 1 adopts different binding modes {(N), (C^Fc, N)⁻ or (C^Ph, N)⁻}, (b) a comparative study of their electrochemical and cytotoxic activities in two breast cancer cell lines, and (c) a theoretical study based on DFT calculations to explain the selectivity of the metallation process and the properties of the new Pt(II) compounds.

2. Results and Discussion

2.1. Synthesis and Characterization of the Compounds

Treatment of equimolar amounts of (S_C)-[FcC(H)=N–CH(Me)(C₆H₅)] (1) and cis-[PtCl₂(dmso)₂] under reflux for 2 h, followed by work up of a SiO₂ column chromatography, gave traces of ferrocenecarboxaldehyde (FcCHO, 12 mg) and two new Pt(II) compounds in a molar ratio 1.00:0.13 (Scheme 1, step A).

Scheme 1. Synthesis of the new Pt(II) complexes: (i) cis-[PtCl₂(dmso)₂] in refluxing MeOH (2 h) followed by SiO₂ column chromatography; (ii) cis-[PtCl₂(dmso)₂] and NaOAc in a toluene:MeOH (4:1) mixture {molar ratios 1:Pt(II):OAc⁻ = 1:1:2} under reflux (see text) followed by SiO₂ chromatography; (iii) PPh₃ in CH₂Cl₂ under reflux for 1 h; (iv) separation of the diastereomers by SiO₂ column chromatography.

Scheme 1. Synthesis of the new Pt(II) complexes: (i) cis-[PtCl₂(dmso)₂] in refluxing MeOH (2 h) followed by SiO₂ column chromatography; (ii) cis-[PtCl₂(dmso)₂] and NaOAc in a toluene:MeOH (4:1) mixture {molar ratios 1:Pt(II):OAc⁻ = 1:1:2} under reflux (see text) followed by SiO₂ chromatography; (iii) PPh₃ in CH₂Cl₂ under reflux for 1 h; (iv) separation of the diastereomers by SiO₂ column chromatography.

Elemental analyses and mass spectra of the major component (Supporting Information) were consistent with those expected for any of the four isomers of [Pt(1)Cl₂(dmso)] (Figure 3, types 2–5 with X = Cl and L = dmso (2a–5a)). In the ¹H-NMR spectrum the signals due to the protons of the dmso ligand appeared as one singlet at δ = 3.40 ppm (³J_H,Pt = 19 Hz). This is characteristic of trans-[Pt{κ¹-N[FcC(R¹)=N–R²]}Cl₂(dmso)] (R¹ = H or Me) complexes [20,36,40,41,42]. Moreover, in its (¹H–¹H) NOESY spectrum, the signal of the imine proton showed cross peaks with those of the H² and H⁵ protons of the Fc unit, and also with the resonances of the −CH(Me)-moiety. Thus indicating that imine 1 adopts the anti-(E) form. All these features suggested that in the major product the arrangement of the ligands corresponded to type 2 shown in Figure 1 (with X = Cl and L = dmso, herein after referred to as 2a).

Its X-ray crystal structure (Figure 4) consists of discrete molecules of trans-[Pt{κ¹-N[(η⁵-C₅H₅)Fe[(η⁵-C₅H₄)–CH=N–CH(Me)(C₆H₅)]}Cl₂(dmso)] separated by van der Waals contacts. In each molecule, the Pt(II) is bound to the imine nitrogen, the sulphur of the dmso ligand and to the Cl(1) and Cl(2) atoms, in a trans- arrangement [bond angle Cl(1)–Pt–Cl(2) = 176.7(1)°], giving a slightly distorted square-planar environment around the Pt(II) atom. The Pt-ligands’ bond lengths agree with those reported for trans-[Pt(imine)Cl₂(dmso)] compounds [20,41,55,56].

Figure 4. Molecular structure and atom labelling scheme for trans-(S_c)-[Pt{(η⁵-C₅H₅)Fe[(η⁵-C₅H₄)–CH=N–CH(Me)(C₆H₅)]Fe(η⁵-C₅H₅)}Cl₂(dmso)] (2). Selected bond lengths (in Å) and angles (in deg.): Pt–N, 2.017(5); Pt–S, 2.206(2); Pt–Cl(1), 2.266(3); Pt–Cl(2), 2.280(3); C(10)–C(11), 1.424(12); C(11)–N, 1.261(9); N–C(12), 1.481(11); C(12)–C(13), 1.516(15); C(12)–C(14), 1.532(13); C(14)–C(15), 1.321(15); C(15)–C(16), 1.409(17); C(16)–C(17), 1.32(2); C(17)–C(18), 1.45(2); C(18)–C(19), 1.377(19); C(19)–C(14), 1.404(13); S–O, 1.444(8); S–Pt–Cl(1), 92.82(10); S–Pt–Cl(2) ,90.0(1);N–Pt–Cl(1), 88.4(2); N–Pt–Cl(2), 88.6(2); C(6)–C(10)–C(11), 118.9(8); C(9)–C(10)–C(11), 132.1(7); C(10)–C(11)–N, 128.6(7); C(11)–N–C(12), 116.0(6); N–C(12)–C(13), 111.2(8); N–C(12)–C(14), 112.9(8); O–S–C(20), 107.3(7); O–S–C(21), 107.6(6) and C(20)–S–C(21), 102.7(8).

Figure 4. Molecular structure and atom labelling scheme for trans-(S_c)-[Pt{(η⁵-C₅H₅)Fe[(η⁵-C₅H₄)–CH=N–CH(Me)(C₆H₅)]Fe(η⁵-C₅H₅)}Cl₂(dmso)] (2). Selected bond lengths (in Å) and angles (in deg.): Pt–N, 2.017(5); Pt–S, 2.206(2); Pt–Cl(1), 2.266(3); Pt–Cl(2), 2.280(3); C(10)–C(11), 1.424(12); C(11)–N, 1.261(9); N–C(12), 1.481(11); C(12)–C(13), 1.516(15); C(12)–C(14), 1.532(13); C(14)–C(15), 1.321(15); C(15)–C(16), 1.409(17); C(16)–C(17), 1.32(2); C(17)–C(18), 1.45(2); C(18)–C(19), 1.377(19); C(19)–C(14), 1.404(13); S–O, 1.444(8); S–Pt–Cl(1), 92.82(10); S–Pt–Cl(2) ,90.0(1);N–Pt–Cl(1), 88.4(2); N–Pt–Cl(2), 88.6(2); C(6)–C(10)–C(11), 118.9(8); C(9)–C(10)–C(11), 132.1(7); C(10)–C(11)–N, 128.6(7); C(11)–N–C(12), 116.0(6); N–C(12)–C(13), 111.2(8); N–C(12)–C(14), 112.9(8); O–S–C(20), 107.3(7); O–S–C(21), 107.6(6) and C(20)–S–C(21), 102.7(8).

The imine is in anti-(E) form [torsion angle C(10)–C(11)–N(1)–C(12) = 175.9°] in good agreement with the conclusions reached by NMR and the C(10)–N(1) bond length (1.261(9) Å) is similar to that reported for [Pd(1)(η³-C₃H₅)Cl] [1.255(6) Å] [48]. Bond lengths and angles of the Fc moiety are similar to those reported for most monosustituted ferrocenes [55,57]; the two pentagonal rings are planar, nearly parallel (tilt angle = 2.9°) and they deviate by 10.6(5)° from the ideal eclipsed conformation. The Fe(II)···Pt(II) distance (4.748 Å) precludes the existence of any direct interaction.

In the crystals, molecules are connected by weak Cl···H intermolecular interactions between: (a) the Cl(1) and the H(3) of a close molecule at (1 − x, −1/2 − y, 1/2 − z) (distance Cl(1)···H(3) = 2.894(8) Å) and (b) the Cl(2) and the H(15) atoms of another unit located at (1/2 + x, 5/2 − y, −z), (Cl(2)···H(15) = 2.690(7) Å) forming chains. These chains are connected through additional C(17)−H(17)···O and C(21)−H(21)··· π (C(6)−C(10) ring) interactions.

Elemental analyses and mass spectra of the minor component (I) were consistent with those expected for any of the platinacycles with (C,N)⁻ bidentate ligands (types 6–9, in Figure 2 with X = Cl and L = dmso (6a–9a)}. In the ¹H-NMR spectrum: (a) the resonances of the protons of the dmso appeared as a singlet, this is consistent with trans-arrangement of the dmso and imine nitrogen; and (b) the signal of the imine proton (at δ = 8.91 ppm, (⁴J_H,Pt = 51 Hz)) was low-field shifted in relation to the free ligand (δ = 8.20 ppm) [48]. This suggested, according to the bibliography [20,36,37,40,41,42], that 1 adopted the syn-(Z) form. Besides that, only four cross peaks were detected in the aromatic region of the {¹H–¹³C}-HSQC spectra, indicating the existence of a Pt−C^Ph bond in the complex. The ¹⁹⁵Pt chemical shift was similar to those of complexes with “Pt(C^Ph ,N)ClS(dmso)” cores. All these findings indicate that I was [Pt{κ²-C,N-[(C₆H₄)−CH(Me)−N=C(H)Fc]}Cl(dmso)] (Figure 3, type 7, with X = Cl and L = dmso (7a)). Moreover, since it has been demonstrated that cyclometallation of enantio- or diastereomerically pure N-donor ferrocenyl ligands does not modify the chirality of the stereogenic center [40,41,42,58,59,60], the absolute configuration of 7a is S_C.

It should be noted that: (a) the molar ratio 2a:7a did not vary significantly when the reaction period (t) increased from 2 h (1.00:0.13) to 2 days (1.00:0.15), and (b) no evidences of the presence of platinacycles with σ[Pt−C(ferrocenyl)] bonds were detected by ¹H-NMR of the crude of the reactions. In view of these and since the formation of platinacycles with bidentate (C^Ph or C^Fc, N)]⁻ ligands commonly requires the presence of a base (i.e., NaOAc or the ligand itself), [20,36,37,40,41,42], we also studied the reactivity of 1 toward Pt(II) in the presence of NaOAc.

Treatment of ligand 1 with cis-[PtCl₂(dmso)₂] and NaAcO (in a molar ratio 1:Pt(II):OAc⁻ = (1:1:2)) in a mixture of toluene: MeOH (4:1) under reflux for 2 h (Scheme 1, step B) followed by the work-up of a chromatographic column, gave small amounts of FcCHO (18 mg), compounds 2a, 7a and traces (ca. 9 mg) of a purple solid (hereinafter referred to as II). Elemental analyses of II agreed with those expected for the platinacycles with 1 acting as a (C,N)⁻ ligand. Its ¹H-NMR spectrum showed two sets of superimposed signals (of relative intensities 1.00:0.90) with the typical pattern of 1,2-disubstituted ferrocene derivatives. Thus, indicating the presence of two species with a Pt−C^Fc bond in solution. Since metallation of the C₅H₄ ring induces planar chirality [20,38,40,41,42,58,59,60], two diastereomers ((S_p, S_C) and (R_p,S_C)) could be expected in principle (Figure 3, types 8 and 9 with X = Cl and L = dmso (8a and 9a, respectively)). Consequently II contains a mixture of both isomers. Longer reaction periods (up to 6 days) produced significant variations of the relative abundance of 2a, 7a and II, (from 1.00:0.21:0.09 (for t = 2 h) to 1.00:0.46:0.64 (for t = 2 days)); and after 6 days, the typical resonances of 2a were not detected in the ¹H-NMR spectrum of the crude of the reaction and only FcCHO, 7a and II were isolated after the work up of the column. These results suggested that the formation of isomers (8a and 9a) was slow.

Since: (a) the crystal structure of 2a revealed that the C−H bond of the ferrocenyl unit is quite close to the Pt(II) ion (distance Pt···H(9) = 2.866 Å); and (b) the relative proportion 2a: II decreased with time; we tentatively assumed that probably 2a was the precursor of the mixture II. In order to test this hypothesis, compound 2a was treated with NaOAc (in a 1:2 molar ratio) in toluene:MeOH (4:1) under reflux for 6 days and then filtered and concentrated to dryness. NMR studies revealed the presence of FcCHO and the two diastereomers of II in a relative ratio (1.00:0.86) which is close to that obtained when the reaction was performed using ligand 1, cis-[PtCl₂(dmso)₂] and NaAcO (1.00:0.90).

All the results presented here reveal that despite it is well-known that ferrocene derivatives are more prone to undergo electrophilic attacks than their phenyl analogues [34,35,58,61], for 1, that has different types of σ(C−H) bonds susceptible to metallate, the formation of the platinacycle 7a with a σ[Pt−C(phenyl)] is achieved under milder experimental conditions than those required to isolate the 8a and 9a with a σ[Pt−C(ferrocenyl)] bond. Unfortunately, attempts to separate the two components of II by fractional crystallisation or SiO₂ column chromatography failed and only solids enriched in ca. 70%–80% could be isolated by successive chromatographic columns.

Since it is well-known that platinacycles with PPh₃ ligands are more cytotoxic than their dmso analogues, we decided to study the reactivity of the platinacycles with PPh₃. As expected, treatment of II with the equimolar amount of PPh₃ gave the diastereomers (R_p, S_C) and (S_p, S_C) of [Pt{{κ²-C,N-[(η⁵-C₅H₃)−CH=N−CH(Me)(C₆H₄)]Fe(η⁵-C₅H₅)}Cl(PPh₃)] (8b and 9b, respectively) that could be isolated after a careful work-up of a SiO₂ column (Scheme 1, step C). In order to compare the electrochemical properties and the cytotoxic activities of isomeric compounds differing only in the nature of the metallated carbon, we also transformed platinacycle 7a in CH₂Cl₂ following identical procedure and we isolated the (S_C)-[Pt{κ²-C,N-[(C₆H₄)−CH(Me)−N=C(H)Fc}Cl(PPh₃)] (7b) (Scheme 1, step D).

Characterization data of 7b–9b (Supplementary Information) agreed with the proposed formulae. ³¹P{¹H}-NMR data for 7b (δ = 19.1 (¹J_Pt−P = 3801 Hz)), 8b (δ = 15.9 (¹J_Pt−P= 4191 Hz)) and 9b (δ = 15.7 (¹J_Pt−P = 4202 Hz)) are similar to those reported for related platinacycles with “(C^Ph or C^Fc, N,P,Cl)” cores and a trans- arrangement of the PPh₃ ligand and the imine nitrogen in good agreement with the transphobia effect [61]. ¹⁹⁵Pt NMR spectra of compounds 7b–9b showed a doublet due to coupling with the ³¹P-nucleus since it is well-known that an upfield shift in ¹⁹⁵Pt NMR is related to a strong donor interaction [62,63], Thus, differences detected in the ¹⁹⁵Pt chemical shifts of 7b and those of 8b and 9b can be taken as a measure of the different donor abilities of imine 1 adopting the (C^Ph,N)⁻ or the (C^Fc,N)⁻ binding mode.

At this point and due to the increasing interest in platinacycles with (C,N,C')²⁻ pincer ligands we turned our attention to 8b and 9b. Molecular models revealed that in both isomers the C−H bond on the ortho site of the phenyl ring has the proper orientation as undergo metallation to give platinacycles with 1 acting as a (C^Fc,N, C^Ph)²⁻ pincer ligand (Figure 2, types 10 and 11 with X = Cl and L = PPh₃ (10b and 11b)) and having central and planar chirality. In view of this, we also studied the reactions of the 8b (or 9b) with equimolar amounts of a NaOAc under reflux for 2 days in the toluene: MeOH (4:1) mixtures. However, no significant change was observed and ¹H-NMR spectra of the raw materials did not reveal the presence of any new Pt(II) complex.

As mentioned above, isomeric cycloplatinated complexes differing exclusively in the nature of the metallated carbon or the planar chirality of the metallated ferrocenyl moiety are scarce and studies on chiral derivatives are even less common. In view of this and mainly due to the ongoing interest in: (a) the potential antitumoral activity of platinacycles [1,15,26,27,28,29,30,31,32,42,43] and (b) electrochemical properties of ferrocene derivatives and heteropolymetallic complexes with ferrocenyl ligands [34,35,40,41,59], we decided to use compounds 7b–9b for preliminary studies on these fields.

2.2. Study of the Cytotoxic Activity

Two human breast cancer cell lines (MCF7 and MDA-MB231) were used to test the cytotoxic activity of complex 2a and the platinacycles (7b–9b). For comparison purposes, cisplatin (as positive control) was also evaluated under the same experimental conditions. The IC₅₀ values corresponding to the inhibition of cancer cell growth at 50% level are listed in Table 1.

Table 1. Summary of: (a) the cytotoxic activities (IC₅₀ (µM)) of compound 2a and the platinacycles 7b–9b in the MCF7 and MDA-MB231 breast cancer cell lines ^a; and (b) their electrochemical properties (Electrochemical data (anodic (E_pa) and cathodic (E_pc) potentials, separation between peaks (ΔE = E_pa – E_pc) and the ratio between the intensities of the anodic and cathodic peaks I_pa/I_pc)) ^b.

Compound	Binding mode of 1	Cytotoxic activity, IC50		Electrochemical data
		MCF7	MDA-MB231	Epa	Epc	ΔE	Ipa/Ipc
2a	N	>100	76 ± nd	0.076	0.006	0.070	1.08
7b	(CPh, N)−	37 ± 4	29 ± 3	0.101	0.038	0.063	1.10
8b	(CFc, N)−	11.4 ± 1.1	5.8 ± 0.8	−0.127	−0.197	0.070	0.98
9b	(CFc, N)−	9.2 ± 0.9	4.3 ± 0.6	−0.157	−0.228	0.071	0.97

^a IC₅₀ values for cisplatin under identical experimental conditions = 19.4 ± 4.5 µM (in MCF7) and 6.3 ± 2.4 µM (in MDA-MB231); ^b Potentials (in V) are referred to the ferrocene/ferricinium couple.

Table 1. Summary of: (a) the cytotoxic activities (IC₅₀ (µM)) of compound 2a and the platinacycles 7b–9b in the MCF7 and MDA-MB231 breast cancer cell lines ^a; and (b) their electrochemical properties (Electrochemical data (anodic (E_pa) and cathodic (E_pc) potentials, separation between peaks (ΔE = E_pa – E_pc) and the ratio between the intensities of the anodic and cathodic peaks I_pa/I_pc)) ^b.

Compound	Binding mode of 1	Cytotoxic activity, IC50		Electrochemical data
		MCF7	MDA-MB231	Epa	Epc	ΔE	Ipa/Ipc
2a	N	>100	76 ± nd	0.076	0.006	0.070	1.08
7b	(CPh, N)−	37 ± 4	29 ± 3	0.101	0.038	0.063	1.10
8b	(CFc, N)−	11.4 ± 1.1	5.8 ± 0.8	−0.127	−0.197	0.070	0.98
9b	(CFc, N)−	9.2 ± 0.9	4.3 ± 0.6	−0.157	−0.228	0.071	0.97

^a IC₅₀ values for cisplatin under identical experimental conditions = 19.4 ± 4.5 µM (in MCF7) and 6.3 ± 2.4 µM (in MDA-MB231); ^b Potentials (in V) are referred to the ferrocene/ferricinium couple.

The comparison of in vitro cytotoxic activities of 2a, 7b, 8b and 9b against the MCF7 and MDA-MB231 breast cancer cell lines (Table 1 and Figure 5a) shows that complex 2a, where 1 acts as an N-donor group, did not show any significant effect on the MCF7 cell line and only a moderate activity in MDA-MB231 cell line and platinacycles 7b–9b were more potent than 2a in the two cell lines assayed. In fact, their potencies increase according to sequence (1).

2a < 7b < cisplatin < 8b ≤ 9b (for MCF7 and MDA-MB231)

(1)

Figure 5. (a) Comparative plot of the cytotoxic activities (IC₅₀ values (µM)) of complexes 2a, the platinacycles 7b–9b and cisplatin in the MCF7 and MDA-MB231 breast cancer cell lines and (b) cyclic voltammograms (at a scan rate of 0.1 V × s⁻¹) of the new Pt(II) complexes containing the aldimine 1 acting as a neutral N-donor (in 2a) or as a monoanionic ((C^Ph, N)⁻ (in 7b) or (C^Fc, N)⁻ (in 8b and 9b)) ligand.

Figure 5. (a) Comparative plot of the cytotoxic activities (IC₅₀ values (µM)) of complexes 2a, the platinacycles 7b–9b and cisplatin in the MCF7 and MDA-MB231 breast cancer cell lines and (b) cyclic voltammograms (at a scan rate of 0.1 V × s⁻¹) of the new Pt(II) complexes containing the aldimine 1 acting as a neutral N-donor (in 2a) or as a monoanionic ((C^Ph, N)⁻ (in 7b) or (C^Fc, N)⁻ (in 8b and 9b)) ligand.

Among the tested compounds, complex 2a is a weaker cytotoxic agent than 7b; while platinacycle 9b, (and in a lesser extent, 8b) are the most potent ones (IC₅₀ in the ranges: 9–11 µM for MCF7 and 5.8–6.3 µM for the triple negative (ER, PR and no HER2 over expression) MDA-MB231 cell line). In particular, in particular the activity of 9b is ca. 1.3 times (in MDA-MB231) or 2.1 times (in the MCF7) bigger than that of cisplatin. These findings indicate that platinacycles with Pt–C^Fc bond (8b and 9b) are more active than 7b, where the Pt(II) ion is bound to a C^Ph atom. The cytotoxic activity of the two diastereomers (8b and 9b) is a bit different, if significant. This agrees with results obtained for compounds D and E (Figure 2) and suggest that the planar quirality of the metallated ring is not as important as the type of metallated carbon. Some authors have also found that for substituted ferrocene derivatives with stereogenic carbon atoms and their transition metal complexes, the absolute configuration does not affect significantly their cytotoxic activity [64,65,66].

Taking into account these results, trend (1) can be re-written as

(N) < (C^Ph, N)⁻ < (C^Fc, N)⁻

(2)

that is a relationship between binding mode of 1 to the Pt(II) and the cytotoxic activity of the complexes.

2.3. Electrochemical Properties

In a first attempt to elucidate the effects produced by the mode of binding of the Schiff base to the Pt(II) atoms, and the nature of the metallated carbon atom on the electronic environment of the iron (II), we also studied the electrochemical properties of 2a,7b, 8b and 9b. It should be noted that ¹H-NMR spectra of these products in CD₃CN revealed they were stable in this solvent. The electrochemical studies were carried out by cyclic voltammetry of freshly prepared solutions (10⁻³ M) in acetonitrile using (Bu₄N)[PF₆] as supporting electrolyte. All these experiments were carried out at different scan rates υ (υ = from 0.05 to 1.0 V × s⁻¹). A summary of the electrochemical data for all the compounds under study is presented in Table 1.

The cyclic voltammograms (CVs) (Figure 5b) exhibited one anodic peak with a directly associated reduction in the reverse scan, intensities ratio I_pa/I_pc was close to 1.0 and the relationship between the I_pa and υ^½ was lineal. These findings are consistent with those expected for a simple reversible one-electron process [67]. The main difference between the CVs of 2a,7b,8b and 9b is the position of the waves, that shift to the anodic region as follows:

9b < 8b < < 2a < 7b

(3)

For the platinacycles 7b–9b the trend is similar to that obtained for cyclopalladated compounds arising from the metallation of the Ph ring or the Fc unit of (1S_C,2R_C)-[(η⁵-C₅H₅)Fe{(η⁵-C₅H₄)−CH=N −CH(Me)-CH(OH)(C₆H₅)}] [59]. For the diastereomers 8b and 9b, the differences are small and fall in the ranges reported for the pairs of isomers (S_p,S_C) and (R_p,S_C) of [Pt{κ²-C,N[(η⁵-C₅H₃)−CH=N−CH(R)−CH₂(OH)]Fe(η⁵-C₅H₅)}Cl(PPh₃)] (R = Me or ⁱPr) or (S_p,1S_C,2R_C) and (R_p,1S_C,2R_C) of [Pt{κ²-C,N[(η⁵-C₅H₃)−CH=N−CH(Me)−CH(OH)(C₆H₅)]Fe(η⁵-C₅H₅)}Cl(PPh₃)] [40,41,59].

It is well-known that the proclivity of ferrocene derivatives to oxidize is strongly dependent on the nature of the substituents [41,59,61,67,68]. In general, the presence of electron withdrawing groups produces an increase of the E_pa value; while donor groups have the opposite effect. On these basis, sequence (4) provides conclusive evidences of the influence of the binding mode of the ligand 1 to the Pt(II) on their the electrochemical properties of the complexes.

Compounds 8b and 9b with the (C^Fc,N)⁻ ligand are more prone to oxidize than 2a where 1 binds to the Pt(II) through the N atom exclusively; while platinacycle 7b is even more resistant to oxidation. Thus indicating that in 7b the metallacycle has greater electron pulling effect on the ferrocenyl array than in 8b, 9b. Comparison of electrochemical data of platinacycles 7b–9b and their biological activity, reveals that 7b is simultaneously less prone to oxidize and a weaker cytotoxic agent than 8b and 9b. This relation shift has also been found for other ferrocene-containing compounds [69,70].

2.4. Theoretical Studies

In a first attempt to rationalise the experimental results and in particular, to explain: (a) why for 1 the cycloplatination of the phenyl ring occurs; and (b) the electrochemical properties of the Pt(II) complexes, we decided to undertake DFT calculations of the imine 1 [in the anti-(E) and syn-(Z) forms, herein after referred to as 1^E and 1^Z, respectively], the isomers of [Pt{FcC(H)=CH(Me) (C₆H₅)}Cl₂(dmso)] (2a and 3a) and platinacycles 6–9 of Figure 2 (X = Cl and L = dmso (6a–9a)).

All the calculations were carried out using the B3LYP hybrid functional [70,71] and the LANL2DZ basis set [72,73] implemented in the Gaussian03 program [74]. The geometries of compounds and were optimized without imposing any restriction. The bond lengths and angles of the optimized geometry of 2a were consistent with those obtained from the X-ray studies (the differences did not clearly exceed 3σ).

For each one of the molecules under study in their optimized geometries the total energy (E_T) was calculated. The results obtained showed that the free ligand in the syn-(Z) conformation (1^Z) is 5.29 kcal/mol higher in energy than the anti-(E) isomer (1^E). Thus, suggesting that the energy required for the process 1^E → 1^Z (Scheme 2, step A) is expected to be small, but slightly greater than for the anti-(E)→syn-(Z) isomerization of the ligand in trans-[Pt{FcC(H)=CH(Me)−(C₆H₅)]}Cl₂(dmso)] (2a) (Scheme 2, B).

Complex 2a has several sites susceptible to metallate which would produce either 6a (Scheme 2, step C) or the isomers 8a and 9a (Scheme 2, steps D and E); but for 3a only the metallation of the phenyl ring, to give 7a, is possible (Scheme 2, step F). In steps C–F the formation of the platinacycles requires the release of one of the Cl⁻ ligands and the activation of the corresponding σ(C−H) bond. On the other hand, the experimental work shows that despite the Fc unit is more prone to undergo electrophillic attacks than the phenyl ring [22], the formation of 7a is achieved under milder experimental conditions than for 8a and 9a.

As a first approach to rationalize experimental results, first we determined the energy for the optimized geometries of 6a–9a. The minimum value was obtained for 9a; for 8a, it was only 0.71 kcal/mol higher; but for 7a and 6a the calculated energies exceed, (by 4.05 and 16.42 kcal/mol, respectively), that obtained for 9a.

We also calculated the energy variations for reactions C–F, where complex 2a is transformed into the platinacycles 6a, 8a and 9a; or 3a into 7a. These processes involve the activation of the C–H bonds and the formation of the monocycloplatinated products. Data presented in Scheme 2, indicate that from an energetic point of view: (a) for 2a, metallation of the ferrrocenyl moiety to give 8a and 9a is preferred over platination of the phenyl ring; and (b) the reaction 3a → 7a, requires a smaller energy income than steps C, D and E.

Scheme 2. Summary of the different processes studied by DFT calculations that involve: (a) the anti-(E) → syn-(Z) isomerization of the free ligand (1) or of compound 2a (steps A and B, respectively); (b) the transformation of complex 2a into platinacycles 6a, 8a and 9a by activation of the σ(C^Ph−H) (step C) or the σ(C^Ph−H) bonds (steps D and E); (c) the conversion of complex 3a into the platinacycle 7a (step F) and d) the platination of the phenyl ring of 8a or 9a (steps G and H, respectively) to give the biscycloplatinated derivatives 10a and 11a. (The values below or next to the arrows are the variations of energies (in kcal/mol) for the corresponding step).

Scheme 2. Summary of the different processes studied by DFT calculations that involve: (a) the anti-(E) → syn-(Z) isomerization of the free ligand (1) or of compound 2a (steps A and B, respectively); (b) the transformation of complex 2a into platinacycles 6a, 8a and 9a by activation of the σ(C^Ph−H) (step C) or the σ(C^Ph−H) bonds (steps D and E); (c) the conversion of complex 3a into the platinacycle 7a (step F) and d) the platination of the phenyl ring of 8a or 9a (steps G and H, respectively) to give the biscycloplatinated derivatives 10a and 11a. (The values below or next to the arrows are the variations of energies (in kcal/mol) for the corresponding step).

Using the E_T values for isomers 6a–9a and T = 298 K, we also calculated the Boltzmann’s distribution to estimate the relative proportion of the four isomers: 51.3 (for 8a), 38.6 (for 9a), 10.0 (for 7a) and 0.1% (for 6a). Except for 6a, that was neither isolated nor detected in any of the reactions studied, the relative amounts of the isomers 7a–9a formed when the reaction was performed using cis-[PtCl₂(dmso)₂], the ligand 1 and NaOAc (in a 1:1:2 molar ratio) follow the trend 7a << 9a < 8a that is in good agreement with the results reached from the Boltzman’s analyses.

Experimental work revealed that 2a could be converted into 8a and 9a, but this process required the presence of a two fold excess of NaOAc and long refluxing periods (6 days). However, attempts to achieve platinacycles with 1 acting as (C^Fc,N,C^Ph)²⁻ from 8a or 9a failed. ΔE_T calculated for steps G and H of Scheme 2 is high (more than 2.7 times the ΔE_T of D or of E). This may explain why no evidences of the formation of compounds 10a or 11a were detected in any of the reactions investigated.

On the other hand, experimental results indicate the nature of the C atom bound to the Pt(II) in platinacycles 7b–9b produces great changes in their electrochemical properties. Since the oxidation requires the removal of the electron from the highest occupied molecular orbital (HOMO), molecular orbital calculations were also performed for platinacycles 7a–9a, for comparison purposes we also included complex 6a in these calculations.

For 6a and 7a, the HOMO (Figure 6), is mainly formed by the contribution of a 5d orbital of the Pt(II), a 3p orbital of the chloride, and a π orbital of the metallated phenyl ring. The HOMOs of 8a and 9a are markedly different since they are mainly centered on the Fe(II) and the Pt(II) atoms. These findings confirm that for 6a–9a the HOMO is not only iron based. For 6a–9a LUMO orbitals (Figure 6) have a similar shape that results from the contribution of a π* orbital of the >C=N− group and to a lesser extent an atomic orbital of the Fe(II).

Figure 6. Highest occupied molecular orbital (HOMO) and LUMO orbitals for the cycloplatinated compounds 6a–9a.

Figure 6. Highest occupied molecular orbital (HOMO) and LUMO orbitals for the cycloplatinated compounds 6a–9a.

In order to explain the different proclivity of compounds 7b–9b to undergo oxidation, we determined the energy of the HOMO (E_HOMO). As shown in Table 2 E_HOMO increases according to the sequence 6a ≈ 7a << 8a ≤ 9a, indicating that platinacycles 6a and 7a, with the (C^Ph, N)⁻ ligand, are less prone to oxidize than their analogues formed by metallation of the Fc unit. This trend agrees with the results obtained from CV that revealed that 7b less more prone to oxidize than 8b or 9b.

Table 2. Calculated energies (in hartrees) of the HOMO and LUMO (E_HOMO and E_LUMO, respectively) and energy gap (E_LUMO–E_HOMO) for the isomers 6a–9a (shown in Scheme 2).

Compound	Binding Mode of 1	EHOMO	ELUMO	ELUMO–EHOMO
6a	(CPh, N)−	−0.204	−0.058	0.146
7a	(CPh, N)−	−0.204	−0.062	0.142
8a	(CFc, N)−	−0.198	−0.054	0.144
9a	(CFc, N)−	−0.193	−0.057	0.136

Table 2. Calculated energies (in hartrees) of the HOMO and LUMO (E_HOMO and E_LUMO, respectively) and energy gap (E_LUMO–E_HOMO) for the isomers 6a–9a (shown in Scheme 2).

Compound	Binding Mode of 1	EHOMO	ELUMO	ELUMO–EHOMO
6a	(CPh, N)−	−0.204	−0.058	0.146
7a	(CPh, N)−	−0.204	−0.062	0.142
8a	(CFc, N)−	−0.198	−0.054	0.144
9a	(CFc, N)−	−0.193	−0.057	0.136

3. Experimental Section

3.1. Preparation of the Compounds

3.1.1. trans-(S_C)-[Pt{κ¹-N[FcC(H)=N−CH(Me)(C₆H₅)]}Cl₂(dmso)] (2a)

A suspension containing 250 mg (5.92 × 10⁻⁴ mol) of cis-[PtCl₂(dmso)₂] and 40 mL of MeOH (HPLC-grade) was refluxed until complete dissolution of the Pt(II) complex. Then, the hot solution was filtered out and filtrate was treated with 188 mg (5.92 × 10⁻⁴ mol) of ligand 1. The reaction mixture was protected from the light with aluminium foil and refluxed for 2 h. After this period the solution was concentrated to dryness on a rotary evaporator, Afterwards the residue was dissolved in the minimum amont of CH₂Cl₂ and passed through a SiO₂-column chromatography (13.5 cm × 2.3 cm) using CH₂Cl₂ as eluant. The first band collected gave, after concentration to dryness small amounts of FcCHO (18 mg). Once this band was collected the subsequent elution with CH₂Cl₂ produced the release of an adidional and wide band, that contained complex 2a. This band was collected and concentrated to ca. 5 mL. Slow evaporation of the solvent at 282 K produced the formation of small crystals of 2a that were collected and air-dried. (yield: 275 mg, 70.2%).

3.1.2. (S_C)-[Pt{κ²-C,N[(C₆H₄)(Me)CHN=C(H)Fc]}Cl(dmso)] (7a)

This compound was obtained as a minor side-product during the preparation of 2a. and it was isolated from the column described in the preceding paragraph as follows. After the elution of the band containing 2a, the column was eluted with a CH₂Cl₂: MeOH (100:0.5) mixture. This produced the release of a tiny and narrow band that was collected and concentrated to dryness on a rotary evaporator. The solid formed was air-dried and then dried in vacuum for 2 days (yield: 34 mg, 9.3%).

3.1.3. Synthesis of II {a Mixture of the (S_p, S_C) and (R_p, S_C) Diastereomers [Pt{κ²-C,N[(η⁵-C₅H₃)−CH=N−{CH(Me)(C₆H₅)}]Fe(η⁵-C₅H₅)}Cl(dmso)] (8a and 9a, respectively)

This mixture was obtained by using two alternative procedures (a and b) using either 1 and cis-[PtCl₂(dmso)₂] as starting materials (method a) or complex 2a (method b).

Method a. Cis-[PtCl₂(dmso)₂] (250 mg, 5.92 × 10⁻⁴ mol) and imine 1 (188 mg, 5.92 × 10⁻⁴ mol) were treated with 40 mL of toluene. Afterwards, a solution formed by NaOAc (98 mg, 1.19 × 10⁻³ mol) and 10 mL of MeOH was added. The reaction mixture weas protected from the light with aluminum foil and refluxed for 6 days. After this period, the hot solution was filtered out and the filtrate was concentrated to dryness on a rotary evaporator. The residue was kept in a SiO₂ dissicator overnight. Then it was dissolved in the minimum amount of CH₂Cl₂ (ca. 10 mL) and passed through a SiO₂-column chromatography (13.5 cm × 2.3 cm) using CH₂Cl₂ as eluant. Concentration to dryness of the yellowish band eluted gave small amounts of FcCHO (22 mg). Then elution with a CH₂Cl₂: MeOH (100:0.5) mixture, produced two bands of which the first eluted one gave after evaporation of the solvent complex 7a (42 mg). The second band that was purple and remained on the top of the column was later on collected increasing the polarity of the eluant to 100:1. The solid II, containing the two diastereomers (8a and 9a, in a molar ratio 1.00:0.89), was isolated after concentration and the subsequent drying in vacuum for 2 days (140 mg).

Method b. Compound 2a (120 mg, 1.8 × 10⁻⁴ mol) was dissolved in 20 mL of toluene, then a solution formed by 30 mg of NaOAc (3.6 × 10⁻⁴ mol) and 5 mL of MeOH was added. Then the flask containing the reaction mixture was protected from the light with aluminum foil and refluxed for 6 days. After this period, the nearly black solution was filtered through a Celite pad and the filtrate was concentrated to dryness on a rotary evaporator. The purple solid formed, that contained 8a and 9a (1.00:0.86), was washed with (3 × 5 mL portions) of n-hexane, air-dried and then dried in vacuum for 2 days (75 mg, 66%).

3.1.4. (S_C)-[Pt{κ²-C,N[(C₆H₄)CH(Me)−N=C(H)Fc]}Cl(PPh₃)] (7b)

Complex 7a (21 mg, 3.4 × 10⁻⁵ mol) was dissolved in 15 mL of CH₂Cl₂, then, PPh₃ (9 mg, 3.4 × 10⁻⁴ mol) was added. The reaction mixture was protected from the light with aluminum foil and refluxed for 1 h. After this period, the undisolved materials were removed by filtration and the pale orange filtrate was concentrated on a rotary evaporator to ca. 5 mL. The container was kept overnight on a refrigerator. After this period the solution was passed through a fast and short SiO₂ column (1.0 cm × 0.5 cm) using CH₂Cl₂ as eluant. The orange band released was collected and concentrated to dryness on a rotary evaporator. The solid formed was collected, air-dried and afterwards dried in vacuum for 2 days (yield: 27 mg, 69%).

3.1.5. Preparation of the Mixture of the Diastereomers of [Pt{κ²-C,N[(η⁵C₅H₃)−CH=N−CH(Me)(C₆H₅)]Fe(η⁵-C₅H₅)}Cl(PPh₃)] (8b and 9b)

To a solution containing 66 mg (1.02 × 10⁻⁴ mol) of the mixture of two diastereomers of [Pt{κ²-C,N[(η⁵-C₅H₃)−CH=N−CH(Me)(C₆H₅)]Fe(η⁵-C₅H₅)}Cl(dmso)] (8a and 9a) and 50 mL of CH₂Cl₂, PPh₃ (27 mg, 1.02 × 10⁻⁴ mol) was added. The reaction mixture was refluxed for 1 h and then filtered out. The filtrate was concentrated to ca. 5 mL and passed through a short SiO₂-column chromatography. The use of a CH₂Cl₂:MeOH (100: 0.2) mixture as eluant produced the release of a deep-orange band which was collected and concentrated to dryness on a rotary evaporator. Finally, the solid was dried in vacuum for 3 days and the two diastereomers in a ca 1.0:0.9 molar ratio (58 mg, 70%).

3.1.6. Separation of the Diastereomers of [Pt{κ²-C,N[(η⁵-C₅H₃)−CH=N−CH(Me)(C₆H₄)]Fe(η⁵-C₅H₅)}Cl(PPh₃)] (8b and 9b)

A 50 mg amount of the mixture of diastereomers (8b and 9b) was dissolved in the minimum amount of CH₂Cl₂ and passed through a SiO₂ column (2.5 cm × 3.0 cm) using a CH₂Cl₂:MeOH (100:0.01) mixture as eluant. The purple band released was collected in ca. 20 mL portions up to a total volume of 180 mL and then concentrated to dryness on a rotary evaporator giving 9b (23 mg). The solution obtained with the subsequent elution with 160 mL, lead after concentration, a mixture of the two isomers; while the final nine fractions collected produced isomer 8b (19 mg).

4. Conclusions

The results presented in this work reveal that the enantiopure imine 1 is a valuable ligand to achieve different types of chiral heterodimetallic complexes containing Pt(II) and Fe(II). Among the organic compounds that exhibit a similar rich and versatile chemistry in the presence of Pt(II) atoms, ligand 1 is, to the best of our knowledge, the first that can: (a) generate a variety of chiral complexes with ferrocenyl units and (b) adopt three different binding modes ((N) (in 2a), (C^Ph,N)⁻ (in 7a and 7b) or (C^Ph,N)⁻ (in 8a, 8b, 9a and 9b)), and thus leading to isomeric forms of cyclometallated Pt(II) complexes differing in Pt–C bond (C^Ph (in 7a or 7b) or C^Fc (in 8a, 8b, 9a and 9b)) or the planar chirality {S_p (in 8a and 8b) or R_p (in 9a and 9b)}. It has been reported that cyclopalladation of (1S_C, 2R_C) [(η⁵-C₅H₅)Fe{(η⁵-C₅H₄)−CH=N−CH(R)−CH(OH)(C₆H₅)}] (R = Me or ⁱPr) gives palladacycles in which this imine acts as a (C^Fc,N)⁻ or (C^Ph,N)⁻ ligand [59], but for Pt(II) only the diastereomers (S_p,1S_C,2R_C) and (R_p,1S_C,2R_C) of [Pt{κ²-C,N[(η⁵-C₅H₃)−CH=N−CH(Me)−CH(OH)(C₆H₅)]Fe(η⁵-C₅H₅)}Cl(dmso)] (D and E in Figure 2) were isolated. Thus, indicating that the intercalation of the “(R_C) –CH(OH)–” unit between the stereogenic carbon of 1 and phenyl ring is important as to hinder the formation of platinacycles with a Pt−C^Ph bond and a [6.6] bicyclic system (the six-membered metallacycle fused with the Ph ring).

Moreover, we have also demonstrated that an accurate control of the experimental conditions permits to control the preferential formation of a specific type of product as well as the nature of the C−H bond involved in the cycloplatination process. Although the selectivity of the cyclometallation process is low, it provides a simple method for achieving enantio- or diastereomerically pure organometallic Pt(II) complexes with one stereogenic carbon center and with (in 8a, 8b, 9a and 9b) or without (in 2a, 7a and 7b) planar chirality which are extremely scarce, but attracting more and more interest due to their potential utilities and applications.

The new Pt(II) complexes presented herein appear to be excellent candidates for use not only as precursors in asymmetric synthesis or catalysis, but also as electrochemical reagents or as cytotoxic agents. For instance, an accurate selection of the product allows the fine-tuning of the anodic and cathodic potentials in a wide range (i.e., E_pa(minimum) = −0.157 V (for 9b) and E_pa(maximum) = 0.101 V (for 7b)). Thus, these compounds are also attractive in view of their potential utility as selective and specific electrochemical sensors or detectors.

Computational studies have also allowed us to explain the effects produced by the nature of the metallated carbon atom on the relative stability of platinacycles with “Pt(C,N)(X)(L) cores” (Figure 2 6–9), the regioselectivity of the process and also the shift of the waves detected in the cyclic voltammograms.

On the other hand, the biological studies reveal that platinacycle 7b is a weaker cytotoxic agent than cisplatin, but 8b and 9b are even more potent agents. This indicates that the binding mode of 1 in the platinacycles ((C^Ph,N)⁻ or (C^Fc,N)⁻) is important as to modify their biological activity. Moreover, the IC₅₀ values of 8b and 9b in the MDA-MBD231 breast cancer cell line fall in the range obtained for compounds D and E (Figure 2), which exhibit a synergic effect with cisplatin [43]. These findings open up new research paths mainly focused on the following areas: (a) the evaluation of the cytotoxic effect of 8b and 9b in other cancer cell lines including some cisplatin resistant ones, such as HCT-116 colon cancer cell line; (b) additional studies to bring information about their mechanism of action; (c) the chemical modification of the complexes i.e. by replacement of one or the two ancillary ligands to get neutral of ionic derivatives, to be tested on these two breast cancer cell lines.