Half-sandwich Ir(III) and Rh(III) 2,2′-dipyridylamine complexes: Synthesis, characterization and in vitro cytotoxicity against the ovarian carcinoma cells

Journal of Organometallic Chemistry 872 (2018) 114e122 Contents lists available at ScienceDirect Journal of Organometallic Chemistry journal homepage: www.elsevier.com/locate/jorganchem Half-sandwich Ir(III) and Rh(III) 2,20 -dipyridylamine complexes: Synthesis, characterization and in vitro cytotoxicity against the ovarian carcinoma cells k Dvora k, Zdene k Tra vní Pavel Starha, Zdene cek* Biologically Active Complexes and Molecular Magnets, Regional Centre of Advanced Technologies and Materials, Faculty of Science, Palacký University in Olomouc, Slechtitel u 27, 783 71, Olomouc, Czech Republic a r t i c l e i n f o a b s t r a c t Article history: Received 19 June 2018 Received in revised form 25 July 2018 Accepted 29 July 2018 Available online 1 August 2018 A series of the half-sandwich Ir(III) and Rh(III) complexes [M(h5-Cpx)(dpa)X]PF6 (M ¼ Ir for 1e6 and Rh for dpa), pentam7e12) containing N-(pyridin-2-yl)pyridin-2-amine (2,20 -dipyridylamine, ethylcyclopentadienyl (Cp*; for 1e5 and 7e11) or 1,2,3,4-tetramethyl-5-phenylcyclopentadienyl (Cpph; for 6 and 12) ring, and various monodentate ligands (X), speciﬁcally Cl (for 1, 6, 7 and 12), Br (for 2 and 8), I (for 3 and 9), valproato (VP; for 4 and 10) or 4-phenylbutyrato (PB; for 5 and 11), was prepared. The complexes were thoroughly characterized by elemental analysis, IR and NMR spectroscopy and mass spectrometry. A single-crystal X-ray analysis was performed for complex [Ir(h5-Cpph)(dpa)Cl]PF6 (6), revealing a pseudo-octahedral piano-stool geometry with a bidentate N,N0 -coordinated dpa ligand, a pentahapto coordinated Cpph ring and a monodentate chlorido ligand. The crystal structure of complex 6 is stabilized by NeH/F, CeH/F, CeH/Cl, CeH/C and C/F non-covalent contacts. Complexes 1e12 were screened for their in vitro cytotoxicity against the A2780 human ovarian carcinoma cell line. The bestperforming iridium(III) complex 6 showed markedly higher activity (IC50 ¼ 23.5 mM) than complexes 1 e3, 5, 9 and 12, whose IC50 ranged from 68.7 to 87.1 mM. Iridium(III) complex 4 and rhodium(III) complexes 7, 8, 10 and 11 were inactive against the A2780 cells in the tested concentration range (IC50 > 100.0 mM). The chlorido complexes 1, 6, 7 and 12 were studied by 1H NMR spectroscopy for their hydrolytic stability in the 20% DMF-d7/80% D2O and 20% MeOD-d4/80% D2O mixture of solvents, revealing Ir(III) complexes 1 and 6 as stable, while Rh(III) complexes 7 and 12 partially hydrolysed in the used medium. Moreover, hydrophobicity (lipophilicity) of complexes 1e12 was studied by an octanol/water partition (logP). © 2018 Elsevier B.V. All rights reserved. Keywords: Iridium(III) Rhodium(III) Half-sandwich 2,20 -dipyridylamine Crystal structure Cytotoxicity 1. Introduction In the last decade, various rhodium and iridium complexes have been reported as agents showing considerable cytotoxic activity against human cancer cells [1], thus representing an alternative for the conventional platinum-based drugs (e.g., cisplatin or oxaliplatin), whose clinical application is connected with serious negative side-effects (e.g., nephrotoxicity or myelosuppression) and problems with resistance [2]. Among them, Ir(III) and Rh(III) halfsandwich cyclopentadienyl (Cp) complexes of the [M(h5-Cp)(Y^Z) X]þ/0 type seem to be a promising group of agents for the development of cytotoxic non-platinum complexes [1,3]; Y^Z ¼ a * Corresponding author. vní E-mail address: zdenek.travnicek@upol.cz (Z. Tra cek). https://doi.org/10.1016/j.jorganchem.2018.07.035 0022-328X/© 2018 Elsevier B.V. All rights reserved. bidentate-coordinated ligand, X ¼ a monodentate (usually halogenido) ligand. These compounds offer a different mechanism of action than platinum-based drugs and a different cytotoxicity proﬁle, resulting in their capability to overcome both the acquired and intrinsic resistance of cancer cells against the therapeutic action of platinum-based chemotherapeutics [3a,4]. From the structure-activity relationship point of view, it has been reported that cytotoxicity of Ir(III) half-sandwich complexes can be tuned by a Cp-ring extension. In particular, pentamethylcyclopentadienyl (Cp*) complexes, such as [Ir(h5Cp*)(phen)Cl]PF6 (IC50 > 100 mM), show lower anticancer potency than analogues containing extended Cpph (IC50 ¼ 6.7 mM) and Cpbph (IC50 ¼ 0.7 mM) Cp-rings, studied at the A2780 human ovarian carcinoma cell line; Cpph ¼ 1,2,3,4-tetramethyl-5phenylcyclopentadienyl, Cpbph ¼ 1,2,3,4-tetramethyl-5biphenylcyclopentadienyl, phen ¼ 1,10-phenanthroline [5]. P. Starha et al. / Journal of Organometallic Chemistry 872 (2018) 114e122 115 In this work, we aimed to prepare a series of [Ir(h5-Cpx)(dpa)X] PF6 (1e6) and [Rh(h5-Cpx)(dpa)X]PF6 (7e12) complexes (Fig. 1) with different monodentate ligands (i.e., Cl, Br, I, valproate (VP) and PB) and Cp-rings (Cp* and Cpph) and explore whether these structural variations affect the in vitro cytotoxicity against the A2780 human ovarian carcinoma cell line; dpa ¼ N-(pyridin-2-yl) pyridin-2-amine (2,20 -dipyridylamine). 2. Experimental section 2.1. Materials Fig. 1. A general structural formula of the studied iridium(III) and rhodium(III) complexes 1e12, given together with the speciﬁcations of the Cp-ring substitution and type of the monodentate X ligand. However, the same dependence was not observed for the Rh(III) analogues of the above-mentioned Ir(III) complexes, namely [Rh(h5-Cp*)(phen)Cl]PF6 (IC50 ¼ 17.8 mM), [Rh(h5-Cpph)(phen)Cl] PF6 (IC50 ¼ 57.0 mM) and [Rh(h5-Cpbph)(phen)Cl]PF6 (IC50 ¼ 14.7 mM) [4]. Further, the replacement of the chlorido ligand of complexes [Ir(h5-Cp)(Y^Z)Cl]þ/0 by a different type of a monodentate ligand, such as pyridine or its derivatives or hydrosulﬁde, leads to the cytotoxicity enhancement of the studied agents [6]. However, the effect of the replacement of the chlorido ligand by other halogenido ligands has not been performed yet for half-sandwich Ir(III) and Rh(III) complexes. It is noteworthy that these studies have been performed for half-sandwich Ru(II) and Os(II) complexes [7]. The reported results suggested that a relevant difference of cytotoxicity can be reached by the halogenido ligand variation, as exempliﬁed on complexes [Ru(h6-pcym)(L)Cl]PF6 (IC50 ¼ 16.2 mM) vs. [Ru(h6pcym)(L)I]PF6 (IC50 ¼ 3.0 mM), and [Os(h6-pcym)(L)Cl]PF6 6 (IC50 ¼ 3.0 mM) vs. [Os(h -pcym)(L)I]PF6 (IC50 ¼ 1.2 mM); L ¼ N,Ndimethyl-4-[(E)-pyridin-2-yl-diazenyl]-aniline [7a]. Another possibility how to tune the biological effect (cytotoxicity and/or processes connected with the mechanism of action) of [M(ar)(Y^Z)Cl]þ/0 half-sandwich complexes (ar ¼ a ﬁve- or sixmembered aromatic ring) can be based on the replacement of the chlorido ligand by the bioactive one, as recently reported by us [8] and others [9]. For example, complex [Ir(h5-Cpph)(phen)(PB)]PF6, bearing the O-coordinated histone deacetylase inhibitor 4phenylbutyrate (PB), showed a different cytotoxic proﬁle (RF ¼ 0.8) than its chlorido analogue [Ir(h5-Cpph)(phen)Cl]PF6 (RF ¼ 1.8); RF ¼ IC50(A2780R)/IC50(A2780); A2780R is a cisplatinresistant variant of the A2780 cells [8b]. IrCl3$nH2O and RhCl3$nH2O were purchased from Precious Metals Online. 2,20 -dipyridylamine (dpa), 1,2,3,4,5pentamethylcyclopentadiene, 2,3,4,5-tetramethyl-2cyclopentenone, phenylmagnesium bromide (3.0 M in diethyl ether), MgSO4, KBr, KI, valproic acid, 4-phenylbutyric acid, NaOH, silver triﬂuoromethanesulfonate (AgOTf), silver nitrate, KCl, NH4PF6 and cisplatin were purchased from Sigma-Aldrich, VWR International or Alfa Aesar Ltd. Solvents of a laboratory grade were purchased from Fisher-Scientiﬁc and Litolab, and used without further puriﬁcation, except THF that was dried using 4 Å molecular sieves. DMSO‑d6, DMF-d7, MeOD-d4 and D2O for NMR experiments were supplied by Sigma-Aldrich. Roswell Park Memorial Institute (RPMI-1640) medium, fetal bovine serum, glutamine, penicillin/streptomycin mixture, trypsin and phosphate-buffered saline (PBS) were purchased from SigmaAldrich and Fisher-Scientiﬁc. The starting compounds [Ir(m-Cl)(h5-Cp*)Cl]2 [10], [Ir(m-Cl)(h5Cpph)Cl]2 [11], [Rh(m-Cl)(h5-Cp*)Cl]2 [10], [Rh(m-Cl)(h5-Cpph)Cl]2 [4], silver valproate (AgVP) [12] and silver 4-phenylbutyrate (AgPB) [12] were prepared following the reported protocols. 2.2. Syntheses 2.2.1. Complex [Ir(h5-Cp*)(dpa)Cl]PF6 (1) Complex 1 was prepared by the reaction of [Ir(m-Cl)(h5-Cp*)Cl]2 (0.05 mmol) with dpa (0.10 mmol), stirred in MeOH (5 mL) at ambient temperature for 2 h, leading to a change from an orange suspension to a yellow solution containing complex [Ir(h5Cp*)(dpa)Cl]Cl (1*). Then, an excess of NH4PF6 (0.25 mmol) was added and after 5 min of stirring at ambient temperature the reaction mixture was ﬁltered and the solvent volume was reduced by nitrogen gas until a yellow product precipitated. Complex 1 was collected, washed (1 0.5 mL of MeOH and 3 1.0 mL of diethyl ether) and dried under vacuum. Anal. Calcd. for IrC20H24N3ClPF6 (1): C, 35.37; H, 3.56; N, 6.19%; found: C, 35.42; H, 3.79; N, 6.15%. 1H NMR (CDCl3, ppm): d 12.74 (s, N2eH, 1H), 8.27 (d, JHeH ¼ 6.4 Hz, C6eH, 2H), 8.22 (d, JHeH ¼ 8.3 Hz, C3eH, 2H), 7.74 (m, C4eH, 2H), 7.03 (t, JHeH ¼ 6.4 Hz, C5eH, 2H), 1.44 (s, C16eC20eH, 15H). 13C NMR (DMSO‑d6, ppm): d 153.1 (C2), 150.5 (C6), 140.1 (C4), 120.2 (C5), 116.4 (C3), 87.7 (C11eC15), 8.6 (C16eC20). 1H NMR (DMSO‑d6, ppm): d 10.98 (s, N2eH, 1H), 8.33 (d, JHeH ¼ 6.4 Hz, C6eH, 2H), 8.00 (t, JHeH ¼ 7.8 Hz, C4eH, 2H), 7.33 (d, JHeH ¼ 9.2 Hz, C3eH, 2H), 7.26 (t, JHeH ¼ 6.4 Hz, C5eH, 2H), 1.39 (s, C16eC20eH, 15H). 13C NMR (DMSO‑d6, ppm): d 152.1 (C6), 151.5 (C2), 141.1 (C4), 120.9 (C5), 114.2 (C3), 87.7 (C11eC15), 8.1 (C16eC20). ESI þ MS (MeOH, m/z): 534.1 (calc. 534.1; 100%; [Ir(Cp*)(dpa)Cl]þ), 498.3 (calc. 498.2; 50%; [Ir(Cp*)(dpae)]þ). IR (ATR, n, cm1): 444, 456, 535, 582, 622,651,672, 771, 823, 906, 1024, 1050, 1068, 1105, 1122, 1152, 1234, 1259, 1285, 1306, 1361, 1384, 1428, 1467, 1486, 1526, 1580, 1632, 2905, 2937, 3057, 3118, 3167, 3223, 3341. 116 P. Starha et al. / Journal of Organometallic Chemistry 872 (2018) 114e122 2.2.2. Complexes [Ir(h5-Cp*)(dpa)Br]PF6 (2) and [Ir(h5-Cp*)(dpa)I] PF6 (3) Nearly identical preparative procedures were used for the syntheses of both complexes. The reaction mixture containing complex 1* (0.05 mmol; see above) and AgOTf (0.10 mmol) reacted at ambient temperature in the dark in 5 mL of MeOH for 1 h. The white precipitate of AgCl was removed providing a clear light yellow solution, to which an excess (0.10 mmol) of either KBr (for 2) or KI (for 3) was added. The reaction mixture got darker during 2 h of stirring at ambient temperature. After that, NH4PF6 (0.25 mmol) was added, stirred at ambient temperature for 5 min, ﬁltered and the solvent was evaporated leading to the precipitation of a yellow product, which was removed, washed (1 0.5 mL of MeOH and 3 1.0 mL of diethyl ether) and dried under vacuum. Anal. Calcd. for IrC20H24N3BrPF6 (2): C, 33.20; H, 3.34; N, 5.81%; found: C, 33.03; H, 3.18; N, 5.75%. 1H NMR (CDCl3, ppm): d 9.09 (s, N2eH, 1H), 8.46 (d, JHeH ¼ 5.5 Hz, C6eH, 2H), 7.82 (t, JHeH ¼ 7.3 Hz, C4eH, 2H), 7.53 (d, JHeH ¼ 8.3 Hz, C3eH, 2H), 7.11 (t, JHeH ¼ 6.4 Hz, C5eH, 2H), 1.47 (s, C16eC20eH, 15H). 1H NMR (DMSO‑d6, ppm): d 10.95 (s, N2eH, 1H), 8.42 (d, JHeH ¼ 5.5 Hz, C6eH, 2H), 8.00 (t, JHeH ¼ 7.8 Hz, C4eH, 2H), 7.31 (d, JHeH ¼ 9.2 Hz, C3eH, 2H), 7.24 (d, JHeH ¼ 6.9 Hz, C5eH, 2H), 1.43 (s, C16eC20eH, 15H). ESI þ MS (MeOH, m/z): 578.1 (calc. 578.1; 100%; [Ir(Cp*)(dpa)Br]þ), 498.3 (calc. 498.2; 25%; [Ir(Cp*)(dpae)]þ). IR (ATR, n, cm1): 445, 534, 555, 634, 661, 708, 740, 769, 823, 882, 1030, 1081, 1127, 1164, 1212, 1232, 1291, 1315, 1353, 1378, 1434, 1469, 1491, 1523, 1584, 1627, 2925, 2967, 3065, 3142, 3204, 3249, 3359. Anal. Calcd. for IrC20H24N3IPF6 (3): C, 31.18; H, 3.14; N, 5.45%; found: C, 30.81; H, 2.85; N, 5.05%. 1H NMR (CDCl3, ppm): d 9.66 (s, N2eH, 1H), 8.59 (d, JHeH ¼ 6.4 Hz, C6eH, 2H), 7.80 (m, C4eH, 2H), 7.66 (d, JHeH ¼ 8.3 Hz, C3eH, 2H), 7.05 (t, JHeH ¼ 6.4 Hz, C5eH, 2H), 1.54 (s, C16eC20eH, 15H). 1H NMR (DMSO‑d6, ppm): d 10.97 (s, N2eH, 1H), 8.56 (d, JHeH ¼ 5.5 Hz, C6eH, 2H), 7.99 (t, JHeH ¼ 7.8 Hz, C4eH, 2H), 7.29 (d, JHeH ¼ 8.3 Hz, C3eH, 2H), 7.20 (d, JHeH ¼ 6.4 Hz, C5eH, 2H), 1.49 (s, C16eC20eH, 15H). ESI þ MS (MeOH, m/z): 626.1 (calc. 626.1; 100%; [Ir(Cp*)(dpa)I]þ), 498.3 (calc. 498.2; 25%; [Ir(Cp*)(dpae)]þ). IR (ATR, n, cm1): 446, 533, 554, 620, 634, 660, 706, 740, 767, 822, 880, 1029, 1066, 1080, 1127, 1164, 1211, 1231, 1266, 1291, 1314, 1354, 1380, 1433, 1469, 1490, 1523, 1584, 1627, 2924, 2966, 2981, 3028, 3063, 3139, 3204, 3249, 3361. 2.2.3. Complexes [Ir(h5-Cp*)(dpa)(VP)]PF6 (4) and [Ir(h5Cp*)(dpa)(PB)]PF6 (5) Complex 1 (0.05 mmol; see above) was dissolved in 10 mL of MeOH and 0.15 mmol of AgVP (for 4) or AgPB (for 5) was added. After 2 h of stirring at ambient temperature in the dark, a white solid (AgCl and unreacted silver carboxylates) was ﬁltered off and an excess of NH4PF6 (0.25 mmol) was added. The mixture was stirred meanwhile the solvent volume was reduced by a nitrogen gas to ca. 1 mL. After that, an excess of diethyl ether (5 mL) was poured in, resulting in the precipitation of products. The obtained complexes 4 and 5 were collected by ﬁltration, washed (1 0.5 mL of MeOH and 3 1.0 mL of diethyl ether) and dried under vacuum. Anal. Calcd. for IrC28H39N3O2PF6 (4): C, 42.74; H, 5.00; N, 5.34%; found: C, 42.28; H, 4.95; N, 5.09. 1H NMR (CDCl3, ppm): d 8.80 (s, N2eH, 1H), 8.75 (d, JHeH ¼ 5.5 Hz, C6eH, 2H), 7.82 (m, C4eH, 2H), 7.55 (d, JHeH ¼ 8.3 Hz, C3eH, 2H), 7.12 (t, JHeH ¼ 6.4 Hz, C5eH, 2H), 2.29 (m, C32eH, 1H), 1.44 (s, C16eC20eH, C33eH, 17H), 1.30 (m, C33eH, 2H), 1.19 (m, C34eH, 4H), 0.79 (t, JHeH ¼ 7.3 Hz, C35eH, 6H). 1 H NMR (DMSO‑d6, ppm): d 10.88 (s, N2eH, 1H), 8.68 (d, JHeH ¼ 5.5 Hz, C6eH, 2H), 8.01 (t, JHeH ¼ 7.8 Hz, C4eH, 2H), 7.36 (d, JHeH ¼ 8.3 Hz, C3eH, 2H), 7.28 (d, JHeH ¼ 5.9 Hz, C5eH, 2H), 2.13 (qui, JHeH ¼ 4.6 Hz, C32eH, 1H), 1.38 (s, C16eC20eH, 15H), 1.18 (m, C33eH, 4H), 1.08 (m, C34eH, 4H), 0.70 (t, JHeH ¼ 7.3 Hz, C35eH, 6H). ESI þ MS (MeOH, m/z): 641.6 (calc. 642.3; 5%; [Ir(Cp*)(dpa)(VP)]þ), 498.3 (calc. 498.2; 100%; [Ir(Cp*)(dpae)]þ). IR (ATR, n, cm1): 446, 556, 652, 674, 772, 830, 1025, 1038, 1117, 1158, 1235, 1284, 1312, 1350, 1370, 1394, 1433, 1445, 1474, 1489, 1535, 1578, 1591, 1618, 1650, 2870, 2930, 2983, 3078, 3192. Anal. Calcd. for IrC30H35N3O2PF6 (5): C, 44.66; H, 4.37; N, 5.21%; found: C, 44.76; H, 4.60; N, 5.31%. 1H NMR (CDCl3, ppm): d 8.79 (m, N2eH, C6eH, 3H), 7.81 (t, JHeH ¼ 7.3 Hz, C4eH, 2H), 7.52 (d, JHeH ¼ 8.3 Hz, C3eH, 2H), 7.23 (m, C36eH, C40eH, 2H), 7.12 (m, C5eH, C37eH, C38eH, C39eH, 5H), 2.56 (t, JHeH ¼ 7.8 Hz, C32eH, 2H), 2.34 (t, JHeH ¼ 7.3 Hz, C34eH, 2H), 1.88 (qui, JHeH ¼ 7.8 Hz, C33eH, 2H), 1.43 (s, C16eC20eH, 15H). 1H NMR (DMSO‑d6, ppm): d 10.86 (N2eH, 1H), 8.75 (d, JHeH ¼ 5.9 Hz, C6eH, 2H), 8.00 (t, JHeH ¼ 7.8 Hz, C4eH, 2H), 7.35 (d, JHeH ¼ 8.6 Hz, C3eH, 2H), 7.28 (t, JHeH ¼ 6.5 Hz, C5eH, 2H), 7.14 (m, C37eH, C38eH, C39eH, 3H), 7.00 (m, C36eH, C40eH, 2H), 2.42 (t, JHeH ¼ 7.6 Hz, C32eH, 2H), 2.15 (t, JHeH ¼ 7.0 Hz, C34eH, 2H), 1.72 (t, JHeH ¼ 7.4 Hz, C33eH, 2H), 1.38 (s, C16eC20eH, 15H). ESI þ MS (MeOH, m/z): 661.7 (calc. 662.2; 10%; [Ir(Cp*)(dpa)(PB)]þ), 498.3 (calc. 498.2; 100%; [Ir(Cp*)(dpae)]þ). IR (ATR, n, cm1): 449, 470, 515, 555, 621, 650, 670, 701, 728, 772, 830, 878, 909, 968, 1028, 1082, 1161, 1186, 1240, 1268, 1314, 1329, 1358, 1380, 1435, 1468, 1490, 1535, 1576, 1590, 1618, 1643, 2789, 2840, 2865, 2928, 2953, 3064, 3088, 3183, 3289, 3372. 2.2.4. Complex [Ir(h5-Cpph)(dpa)Cl]PF6 (6) The preparation of complex 6 was similar to that of complex 1, except that [Ir(m-Cl)(h5-Cpph)Cl]2 was used instead of [Ir(m-Cl)(h5Cp*)Cl]2, and the reaction was performed in the microwave reaction system (100 C, 1 min). Anal. Calcd. for IrC25H26N3ClPF6 (6): C, 40.51; H, 3.54; N, 5.67%; found: C, 40.38; H, 3.67; N, 5.42%. 1H NMR (CDCl3, ppm): d 8.92 (s, N2eH, 1H), 8.34 (d, JHeH ¼ 5.5 Hz, C6eH, 2H), 7.78 (m, C4eH, 2H), 7.60 (m, Cppharom., 2H), 7.52 (d, JHeH ¼ 9.2 Hz, C3eH, 2H), 7.48 (m, Cppharom., 3H), 7.01 (t, JHeH ¼ 6.9 Hz, C5eH, 2H), 1.53 (s, Cpphaliph., 6H), 1.48 (s, Cpphaliph., 6H). 1H NMR (DMSO‑d6, ppm): d 11.05 (s, N2eH, 1H), 8.30 (d, JHeH ¼ 5.9 Hz, C6eH, 2H), 7.98 (t, JHeH ¼ 7.7 Hz, C4eH, 2H), 7.57 (m, Cppharom., 2H), 7.47 (m, Cppharom., 3H), 7.34 (d, JHeH ¼ 8.8 Hz, C3eH, 2H), 7.19 (t, JHeH ¼ 6.6 Hz, C5eH, 2H), 1.48 (s, Cpphaliph., 6H), 1.41 (s, Cpphaliph., 6H). ESI þ MS (MeOH, m/z): 596.1 (calc. 596.1; 100%; [Ir(Cpph)(dpa)Cl]þ), 560.2 (calc. 560.2; 60%; [Ir(Cpph)(dpae)]þ). IR (ATR, n, cm1): 446, 461, 531, 553, 583, 618, 705, 744, 758, 774, 823, 905, 929, 1006, 1031, 1064, 1101, 1128, 1165, 1211, 1236, 1266, 1352, 1380, 1448, 1471, 1521, 1585, 1628, 2919, 2964, 2984, 3063, 3119, 3397. Single-crystals of complex 6 suitable for an X-ray analysis were obtained from the mother liquor after 24 h of standing at ambient temperature. 2.2.5. Rh(III) complexes 7e11 The preparation of complexes [Rh(h5-Cp*)(dpa)Cl]PF6 (7), [Rh(h5-Cp*)(dpa)Br]PF6 (8), [Rh(h5-Cp*)(dpa)I]PF6 (9), [Rh(h5Cp*)(dpa)(VP)]PF6 (10) and [Rh(h5-Cp*)(dpa)(PB)]PF6 (11) were similar to those of their Ir(III) analogues 1e5, except that [Rh(mCl)(h5-Cp*)Cl]2 was used instead of [Ir(m-Cl)(h5-Cp*)Cl]2. Anal. Calcd. for RhC20H24N3ClPF6 (7): C, 40.73; H, 4.10; N, 7.13%; found: C, 40.70; H, 3.99; N, 7.01%. 1H NMR (CDCl3, ppm): d 9.07 (s, N2eH, 1H), 8.39 (d, JHeH ¼ 5.5 Hz, C6eH, 2H), 7.87 (t, JHeH ¼ 7.8 Hz, C4eH, 2H), 7.53 (d, JHeH ¼ 8.3 Hz, C3eH, 2H), 7.21 (d, JHeH ¼ 6.4 Hz, C5eH, 2H), 1.49 (s, C16eC20eH, 15H). 1H NMR (DMSO‑d6, ppm): d 10.94 (s, N2eH, 1H), 8.35 (d, JHeH ¼ 5.5 Hz, C6eH, 2H), 8.03 (t, JHeH ¼ 7.3 Hz, C4eH, 2H), 7.33 (m, C3eH, C5eH, 4H), 1.41 (s, C16eC20eH, 15H). ESI þ MS (MeOH, m/z): 444.0 (calc. 444.1; 80%; [Rh(Cp*)(dpa)Cl]þ), 408.2 (calc. 408.1; 100%; [Rh(Cp*)(dpae)]þ). IR (ATR, n, cm1): 444, 535, 554, 623, 658, 708, 741, 772, 821, 883, 1024, 1081, 1126, 1165, 1209, 1233, 1317, 1351, 1376, 1434, 1466, 1489, 1520, 1583, 1625, 2925, 2968, 2990, 3065, 3105, 3141, 3205, 3246, 3364. P. Starha et al. / Journal of Organometallic Chemistry 872 (2018) 114e122 Anal. Calcd. for RhC20H24N3BrPF6 (8): C, 37.88; H, 3.81; N, 6.63%; found: C, 37.67; H, 3.93; N, 6.63%. 1H NMR (CDCl3, ppm): d 9.30 (s, N2eH, 1H), 8.47 (d, JHeH ¼ 4.6 Hz, C6eH, 2H), 7.85 (t, JHeH ¼ 7.8 Hz, C4eH, 2H), 7.54 (d, JHeH ¼ 8.3 Hz, C3eH, 2H), 7.17 (t, JHeH ¼ 6.4 Hz, C5eH, 2H), 1.52 (s, C16eC20eH, 15H). 1H NMR (DMSO‑d6, ppm): d 10.95 (s, N2eH, 1H), 8.42 (d, JHeH ¼ 4.6 Hz, C6eH, 2H), 8.03 (t, JHeH ¼ 7.8 Hz, C4eH, 2H), 7.31 (m, C3eH, C5eH, 4H), 1.46 (s, C16eC20eH, 15H). ESI þ MS (MeOH, m/z): 488.0 (calc. 488.0; 100%; [Rh(Cp*)(dpa)Br]þ), 408.1 (calc. 408.1; 60%; [Rh(Cp*)(dpae)]þ). IR (ATR, n, cm1): 443, 534, 554, 633, 658, 707, 741, 770, 821, 882, 904, 1023, 1080, 1126, 1163, 1232, 1291, 1317, 1351, 1376, 1433, 1467, 1522, 1552, 1583, 1625, 2925, 2968, 2988, 3064, 3137, 3196, 3230, 3277, 3365. Anal. Calcd. for RhC20H24N3IPF6 (9): C, 35.26; H, 3.55; N, 6.17%; found: C, 35.07; H, 3.80; N, 6.21%. 1H NMR (CDCl3, ppm): d 9.87 (s, N2eH, 1H), 8.58 (d, JHeH ¼ 4.6 Hz, C6eH, 2H), 7.83 (m, C4eH, 2H), 7.77 (d, JHeH ¼ 8.3 Hz, C3eH, 2H), 7.12 (d, JHeH ¼ 6.4 Hz, C5eH, 2H), 1.60 (s, C16eC20eH, 15H). 1H NMR (DMSO‑d6, ppm): d 10.97 (s, N2eH, 1H), 8.54 (d, JHeH ¼ 5.5 Hz, C6eH, 2H), 8.02 (t, JHeH ¼ 7.8 Hz, C4eH, 2H), 7.28 (m, C3eH, C5eH, 4H), 1.55 (s, C16eC20eH, 15H). ESI þ MS (MeOH, m/z): 536.0 (calc. 536.0; 100%; [Rh(Cp*)(dpa)I]þ), 408.1 (calc. 408.1; 10%; [Rh(Cp*)(dpae)]þ). IR (ATR, n, cm1): 445, 533, 555, 622, 740, 772, 831, 871, 904, 1021, 1079, 1123, 1161, 1233, 1265, 1357, 1379, 1433, 1468, 1521, 1554, 1583, 1625, 2915, 2963, 2985, 3024, 3073, 3128, 3195, 3229, 3272, 3377. Anal. Calcd. for RhC28H39N3O2PF6 (10): C, 48.22; H, 5.64; N, 6.02%; found: C, 48.53; H, 5.32; N, 6.16%. 1H NMR (CDCl3, ppm): d 8.84 (d, JHeH ¼ 5.5 Hz, C6eH, 2H), 8.81 (s, N2eH, 1H), 7.85 (m, C4eH, 2H), 7.48 (d, JHeH ¼ 8.3 Hz, C3eH, 2H), 7.20 (t, JHeH ¼ 6.4 Hz, C5eH, 2H), 2.28 (m, C32eH, 1H), 1.46 (m, C16eC20eH, C33eH, 17H), 1.29 (m, C33eH, 2H), 1.18 (m, C34eH, 4H), 0.77 (t, JHeH ¼ 7.3 Hz, C35eH, 6H). 1H NMR (DMSO‑d6, ppm): d 10.89 (s, N2eH, 1H), 8.76 (d, JHeH ¼ 5.5 Hz, C6eH, 2H), 8.04 (t, JHeH ¼ 7.3 Hz, C4eH, 2H), 7.36 (m, C3eH, C5eH, 4H), 2.10 (qui, JHeH ¼ 4.6 Hz, C32eH, 1H), 1.38 (s, C16eC20eH, 15H), 1.17 (m, C33eH, 4H), 1.08 (m, C34eH, 4H), 0.69 (t, JHeH ¼ 7.3 Hz, C35eH, 6H). ESI þ MS (MeOH, m/z): 551.7 (calc. 552.2; 5%; [Rh(Cp*)(dpa)(VP)]þ), 408.2 (calc. 408.1; 100%; [Rh(Cp*)(dpae)]þ). IR (ATR, n, cm1): 451, 555, 587, 652, 671, 775, 828, 1022, 1065, 1116, 1157, 1223, 1246, 1280, 1313, 1367, 1396, 1419, 1433, 1454, 1477, 1489, 1537, 1557, 1595, 1615, 1656, 2873, 2932, 2956, 3079, 3194. Anal. Calcd. for RhC30H35N3O2PF6 (11): C, 50.22; H, 4.92; N, 5.86%; found: C, 50.31; H, 4.68; N, 5.49%. 1H NMR (CDCl3, ppm): d 8.87 (m, N2eH, C6eH, 3H), 7.85 (m, C4eH, 2H), 7.47 (d, JHeH ¼ 8.3 Hz, C3eH, 2H), 7.18 (m, C36eC40eH, 5H), 7.09 (d, JHeH ¼ 7.3 Hz, C5eH, 2H), 2.56 (t, JHeH ¼ 7.8 Hz, C32eH, 2H), 2.32 (t, JHeH ¼ 7.8 Hz, C34eH, 2H), 1.89 (qui, JHeH ¼ 7.8 Hz, C33eH, 2H), 1.43 (s, C16eC20eH, 15H). 1H NMR (DMSO‑d6, ppm): d 10.90 (N2eH, 1H), 8.83 (d, JHeH ¼ 5.5 Hz, C6eH, 2H), 8.03 (t, JHeH ¼ 7.3 Hz, C4eH, 2H), 7.35 (m, C3eH, C5eH, 4H), 7.17 (m, C37eH, C38eH, C39eH, 3H), 7.01 (d, JHeH ¼ 7.3 Hz, C36eH, C40eH, 2H), 2.42 (t, JHeH ¼ 7.3 Hz, C32eH, 2H), 2.11 (t, JHeH ¼ 6.9 Hz, C34eH, 2H), 1.72 (m, C33eH, 2H), 1.38 (s, C16eC20eH, 15H). ESI þ MS (MeOH, m/z): 517.5 (calc. 572.2; 5%; [Rh(Cp*)(dpa)(PB)]þ), 408.1 (calc. 408.1; 100%; [Rh(Cp*)(dpae)]þ). IR (ATR, n, cm1): 445, 472, 555, 645, 700, 745, 775, 830, 877, 1026, 1080, 1163, 1237, 1273, 1309, 1360, 1395, 1435, 1471, 1533, 1568, 1585, 1630, 1645, 2818, 2857, 2930, 2962, 3084, 3186, 3295, 3374. 2.2.6. Complex [Rh(h5-Cpph)(dpa)Cl]PF6 (12) Complex 12 was prepared similarly as described above for complex 6, using [Rh(m-Cl)(h5-Cpph)Cl]2 instead of [Ir(m-Cl)(h5Cpph)Cl]2. Anal. Calcd. for RhC25H26N3ClPF6 (12): C, 46.07; H, 4.02; N, 6.45%; found: C, 46.25; H, 3.93; N, 6.37%. 1H NMR (CDCl3, ppm): d 8.89 (s, N2eH, 1H), 8.34 (d, JHeH ¼ 6.4 Hz, C6eH, 2H), 7.75 (m, 117 C4eH, 2H), 7.64 (m, Cppharom., 2H), 7.54 (d, JHeH ¼ 8.3 Hz, C3eH, 2H), 7.49 (m, Cppharom., 3H), 7.04 (t, JHeH ¼ 6.9 Hz, C5eH, 2H), 1.54 (s, Cpphaliph., 6H), 1.46 (s, Cpphaliph., 6H). 1H NMR (DMSO‑d6, ppm): d 11.01 (s, N2eH, 1H), 8.30 (d, JHeH ¼ 5.9 Hz, C6eH, 2H), 8.00 (t, JHeH ¼ 7.7 Hz, C4eH, 2H), 7.67 (d, JHeH ¼ 6.6 Hz, Cppharom., 2H), 7.51 (m, Cppharom., 3H), 7.33 (d, JHeH ¼ 8.1 Hz, C3eH, 2H), 7.23 (t, JHeH ¼ 6.6 Hz, C5eH, 2H), 1.48 (s, Cpphaliph., 6H), 1.45 (s, Cpphaliph., 6H). ESI þ MS (MeOH, m/z): 505.9 (calc. 506.1; 70%; [Rh(Cpph)(dpa) Cl]þ), 470.1 (calc. 470.1; 100%; [Rh(Cpph)(dpae)]þ). IR (ATR, n, cm1): 444, 475, 534, 554, 706, 738, 771, 819, 883, 1023, 1081, 1125, 1164, 1208, 1233, 1284, 1351, 1378, 1432, 1466, 1520, 1583, 1624, 2924, 2968, 3065, 3141, 3322, 3364. 2.3. General methods 1 H NMR spectra were recorded using a JEOL JNM-ECA 600II device on CDCl3 and DMSO‑d6 solutions of complexes 1e12 at 300 K, supported by 13C, 1He1H gs-COSY, 1He13C gs-HMQC and 1 He13C gs-HMBC spectra of complex 1; gs ¼ gradient selected, COSY ¼ correlation spectroscopy, HMQC ¼ heteronuclear multiple quantum coherence, HMBC ¼ heteronuclear multiple bond coherence. 1H and 13C NMR spectra were calibrated against the residual signals of the used solvents found at 7.26 ppm (1H) and 77.4 (13C) for CDCl3, and at 2.50 ppm (1H) and 39.5 ppm (13C) for DMSO‑d6 solutions [13]. The splitting of proton resonances in the reported 1H spectra is deﬁned as s ¼ singlet, d ¼ doublet, t ¼ triplet, qui ¼ quintet, m ¼ multiplet. Electrospray ionization mass spectrometry (ESI-MS) was performed by an LCQ Fleet ion trap spectrometer (Thermo Scientiﬁc; QualBrowser software, version 2.0.7) in the positive ionization mode (ESIþ) on the methanolic solutions of the studied complexes. Elemental analyses (C, H, N) were carried out using a Flash 2000 CHNS Elemental Analyser (Thermo Scientiﬁc). Infrared spectroscopy (400e4000 cm1, Attenuated total reﬂectance (ATR) technique) was performed by a Nexus 670 FT-IR spectrometer (Thermo Fisher Scientiﬁc). 2.4. X-ray crystallography The X-ray diffraction data of complex [Ir(h5-Cpph)(dpa)Cl]PF6 (6) were collected by a Bruker D8 QUEST diffractometer (Mo Ka radiation) equipped with a PHOTON 100 CMOS detector. The obtained data were processed and reduced by the APEX3 software package [14] and the molecular structure of complex 6 was solved by direct methods (SHELXS) and reﬁned by a full-matrix least-squares procedure (SHELXL) [15]. Hydrogen atoms of all the structures were found in the difference Fourier maps and reﬁned using a riding model with CeH ¼ 0.95 Å for (CH)aromatic and 0.98 Å for (CH3), and with Uiso(H) ¼ 1.2 Ueq(CH) and 1.5 Ueq(CH3). The X-ray crystallographic data for complex 6 have been deposited in the Cambridge Crystallographic Data Centre under the accession number CCDC 1849574. The crystal data and structure reﬁnements are given in Table 1. The graphics were drawn and additional structural calculations were performed by DIAMOND [16] and Mercury [17] software. 2.5. In vitro cytotoxicity testing An appropriate amount of complexes 1e12 was dissolved in 500 mL of DMF to give the stock solutions of the 100 mM concentration. The stock solutions were diluted by RPMI-1640 medium to the concentrations of 0.1e100.0 mM. The A2780 human ovarian carcinoma cell line (supplied by European Collection of Cell Cultures, ECACC) was cultured according to the ECACC instructions, i.e., as adherent monolayers in a humidiﬁed atmosphere (37 C, 5% CO2), in RPMI-1640 medium 118 P. Starha et al. / Journal of Organometallic Chemistry 872 (2018) 114e122 Table 1 Crystal data and structure reﬁnement for [Ir(h5-Cpph)(dpa)Cl]PF6 (6). Empirical formula C25H26ClIrN3$F6P Formula weight Temperature (K) Wavelength (Å) Crystal system Space group Unit cell dimensions a (Å) b (Å) c (Å) a ( ) b ( ) g ( ) V (Å3) Z, Dcalc (g cm3) Absorption coefﬁcient (mm1) Crystal size (mm) F (000) q range for data collection ( ) Index ranges (h, k, l) 741.11 120(2) 0.71073 Monoclinic P21/n Reﬂections collected Independent reﬂections Data/restraints/parameters Goodnesseofeﬁt on F2 Final R indices [I > 2s(I)] R indices (all data) Largest peak and hole (e Å3) 14.689(3) 10.770 (2) 16.714 (3) 90 104.179(9) 90 2563.7(8) 4, 1.920 5.439 0.22 0.16 0.14 1440 2.271 to 27.560 19 h 19 14 k 14 21 l 21 63423 5903, R(int) ¼ 0.0491 5903/0/341 1.045 R1 ¼ 0.0212, wR2 ¼ 0.0433 R1 ¼ 0.0295, wR2 ¼ 0.0461 1.052 and 1.112 supplemented with 10% of fetal calf serum, 1% of 2 mM glutamine and 1% penicillin/streptomycin. The cells were seeded to 96-well culture plates, preincubated in drug-free media at 37 C for 24 h and treated with complexes 1e12 (and cisplatin involved as the reference drug). After 24 h drug exposure, the supernatants were removed and the cells were washed with drug-free PBS followed by 72 h recovery in a drug-free medium at 37 C. In parallel, the cells were also treated with 0.1% DMF and 1% Triton X-100 to assess the minimal (i.e, 100% cell viability for negative control) and maximal (i.e, 0% cell viability for positive control) cell damage, respectively. The MTT assay (MTT ¼ 3-(4,5-dimethylthiazol-2-yl)-2,5diphenyltetrazolium bromide) was used to determine the cell viability. A concentration of the formed dye was evaluated spectrophotometrically at 540 nm (TECAN, Schoeller Instruments LLC). The data from the cancer cells were acquired from three independent experiments (conducted in triplicate) using the cells from different passages. The data were expressed as the percentage of viability and the resulting IC50 values (mM), calculated from the viability curves, are given as arithmetic mean ± SD. 2.6. 1H NMR studies of hydrolytic stability The MeOD-d4 solutions (120 mL) were prepared from 0.81 mg (for 1), 0.89 mg (for 6), 0.71 mg (for 7) and 0.78 mg (for 12) of the studied chlorido complexes and 480 mL of D2O was added to each solution to provide 600 mL of 2 mM solutions in 20% MeOD-d4/80% D2O. 1H NMR spectra were acquired on the fresh solutions (t z 0 h) and after 1, 2, 6, 24 and 48 h of standing at ambient temperature. Spectra were calibrated against the residual solvent signal of methanol (3.34 ppm) [13]. After 48 h of standing at ambient temperature, an excess (5 mol equiv.) of KCl was added to the equilibrated solutions of complexes 1, 6, 7 and 12 in 20% MeOD-d4/80% D2O and 1H NMR spectra were recorded. For comparative purposes, 1 H NMR spectra were also acquired for the 1h, 6h, 7h and 12h species in the same medium, which were prepared from the fresh solutions of complexes 1, 6, 7 and 12 in 20% MeOD-d4/80% D2O by the addition of the stoichiometric amount of silver nitrate. The mixtures were shaken under an aluminium foil (25 C) for 1 h, then the formed precipitate of AgCl was centrifuged and the obtained solutions were used for 1H NMR experiments. The same experiments were carried out for complexes 1, 6, 7 and 12 in 20% DMF-d7/80% D2O mixture of solvents. Spectra were calibrated against the residual solvent signal of DMF (8.03 ppm) [13]. 1 H NMR experiments at different pH were performed for Rh(III) complexes 7 and 12 to determine the composition of their hydrolysates (i.e., aqua vs. hydroxido species). The solutions of the 7h and 12h species were performed in 5% MeOD-d4/95% D2O (note: solubility of complexes 7 and 12 in D2O was under the detection limit of the used device). pH* was adjusted by DClO4 and KOD solutions (0.1, 0.01 and 0.001 M) to the values of 4, 7 and 10 and 1 H NMR spectra were recorded at the individual pH* points (pH* ¼ pH in D2O). 2.7. Hydrophobicity studies Octanol-saturated water (OSW) and water-saturated octanol (WSO) were prepared from octanol and 0.2 M water solution of KCl (for chlorido complexes 1, 6, 7 and 12), KBr (for bromido complexes 2 and 8), KI (for iodido complexes 3 and 9), NaVP (for valproato complexes 4 and 10) or NaPB (for 4-phenylbutyrato complexes 5 and 11) by the overnight stirring. The stock solutions were prepared by shaking (Vibramax 100, Heidolph Instruments) of 1 mmol of complexes 1e12 in 11 mL of OSW for 1 h. Then the mixtures were centrifuged (5 min, 11,000 rpm) and supernatant was collected. 5 mL of this solution was studied by ICP-MS (the obtained values were corrected for the adsorption effects) for the Ir or Rh content (e.g., [M]OSWb, see below), while other 5 mL of this solution was added to 5 mL of WSO and shaken for 2 h at ambient temperature. After that, these mixtures were centrifuged and aqueous layers were carefully separated. The Ir or Rh concentrations were determined by ICP-MS (the obtained value was corrected for the adsorption effects). logP ¼ log([M]WSO/[M]OSWa) equation was used for the partition coefﬁcient calculation; [M]OSWb and [M] OSWa stands for the Ir or Rh concentration before and after partition, respectively, and [M]WSO ¼ [M]OSWb e [M]OSWa. The experiment was conducted in triplicate and the results are presented as arithmetic mean ± SD. 3. Results and discussion 3.1. Synthesis and general characterization In this work, a series of ionic iridium(III) (1e6) and rhodium(III) (7e12) complexes of the general composition [M(h5-Cpx)(dpa)X] PF6 was studied (Fig. 1). The herein used chelating N,N-donor dpa ligand is a known organic ligand offering various coordination modes [18], and it has recently been reported as a chelating ligand of complexes [M(h5-Cp*)(dpa)Cl]Cl$2H2O (M ¼ Ir or Rh), representing the hydrated chloride salts of complexes 1 and 7 reported herein [19]. Chlorido complexes 1, 6, 7 and 12 were prepared according the conventional protocol widely used for similar half-sandwich Ir(III) and Rh(III) complexes, which uses the dimeric compounds [M(mCl)(h5-Cpx)Cl]2 (M ¼ Ir or Rh, Cpx ¼ Cp* or Cpph) as the Ir(III) and Rh(III) starting materials (e.g., ref. [3,4]). The [Ir(h5-Cp*)(dpa)Cl]þ and [Rh(h5-Cp*)(dpa)Cl]þ species, representing the cations of complexes 1 and 7, were recently prepared by the same method (but in dichloromethane with markedly longer reaction time of 16 h) as employed in this work, and isolated as the hydrated chloride salts [Ir(h5-Cp*)(dpa)Cl]Cl$2H2O and [Rh(h5-Cp*)(dpa)Cl] Cl$2H2O [19]. Bromido (2, 8) and iodido 3, 9) complexes were P. Starha et al. / Journal of Organometallic Chemistry 872 (2018) 114e122 prepared from the [Ir(h5-Cp*)(dpa)Cl]þ chlorido species through their dechlorination by AgOTf, followed by the addition of appropriate potassium halogenide. The syntheses of carboxylato complexes 4, 5, 10 and 11 were similar to those recently reported for the half-sandwich complexes [Os(h6-pcym)(dpa)(VP)]PF6 [8a], [Ir(h5Cpph)(phen)(PB)]PF6 [8b], [Ru(h6-pcym)(dpa)(VP)]PF6 and [Ru(h6pcym)(dpa)(PB)]PF6 [12]. The purity and composition of complexes 1e12 (Fig. 1) were studied by elemental analysis (<0.4% differences between the theoretical and experimental C, H, and N contents), 1H NMR spectroscopy, ESI þ mass spectrometry and FTIR spectroscopy. The infrared spectra of complexes 1e12 showed the peaks of the characteristic vibrations of the Cpx ligand at 2905e2990 (for ns(CeH)aliphatic and nas(CeH)aliphatic) and at 1440e1470 and ca 770 cm1 for nas(CeC), and nas(CeCH3), respectively [19,20]. The peaks at 3030e3200 cm1 belong to the n(CeH)aromatic vibrations of the dpa (for 1e12) and Cpph (for 6 and 12) ligands. Further for the dpa ligand, the peaks centred around 3350 cm1 are assignable to the n(NeH) vibration [19,21]. For carboxylato complexes 4, 5, 10 and 11, the characteristic peaks belonging to the nas(C¼O) and ns(C¼O) vibrations were observed at ca 1645, and 1390 cm1, respectively, for the VP and PB ligands [22]. The FTIR spectra of all complexes contained the peaks assignable to the n(PeF) vibrations of the PF 6 counterion centred at ca. 830, 770 and 555 cm1 [23]. ESI þ mass spectra showed the peaks of the [M(Cpx)(dpa)X]þ species, corresponding to the cations of the studied complexes 1e12, and the [M(Cpx)(dpae)]þ species, which formed from the complex cations of complexes 1e12 by a release of the corresponding halogenido or carboxylato ligand (X) together with a proton (most likely from the eNH group of the dpa ligand). All the signals of both Cpx and dpa ligands were detected in 1H NMR spectra of complexes 1e12 dissolved in CDCl3 (Figs. S1eS7) and DMSO‑d6. For example, the characteristic N2eH singlet of the dpa ligand was detected at 8.79e12.74 ppm (for CDCl3 solutions) and at 10.86e11.05 ppm (for DMSO‑d6 solutions). The aromatic CeH signals of the dpa ligand showed between 8.87 and 7.01 ppm (for CDCl3 solutions). Moreover, 1H NMR spectra of carboxylato complexes 4, 5, 10 and 11 (both CDCl3 and DMSO‑d6 solutions) revealed the signals of the VP and PB ligands, such as the CaeH (i.e., C32eH) signal detected for VP at ca 2.10 ppm (for complexes 4 and 10) and for PB at 2.42 ppm (for complexes 5 and 11). Further for the carboxylato complexes 4, 5, 10 and 11 (both CDCl3 and DMSO‑d6 solutions), the d values of the VP and PB ligands differ markedly from free valproate (e.g., d ¼ 2.39 ppm (CDCl3) and 2.21 ppm (DMSO‑d6) for C32eH) and 4-phenylbutyrate anions (e.g., d ¼ 2.68 ppm (CDCl3) and 2.53 ppm (DMSO‑d6) for C32eH) (Fig. S8). The chemical shifts of the dpa, VP and PB ligands were consistent with the values observed for complexes [Os(h6-pcym)(dpa)(VP)] PF6 [8a], [Ru(h6-pcym)(dpa)(VP)]PF6 and [Ru(h6-pcym)(dpa)(PB)] PF6 [12]. The presence of the Cp* ligand in the structure of complexes 1e5 and 7e11 was proved by its characteristic 1H NMR singlet detected at 1.43e1.60 (for CDCl3 solutions) and at 1.38e1.55 ppm (for DMSO‑d6 solutions), while 1H NMR spectra of Cpph-complexes 6 and 12 contain two singlets 1.46e1.54 (for CDCl3 solutions) and 1.41e1.48 ppm for DMSO‑d6 solutions) in the aliphatic region, belonging to the C16eC19 methyl groups (see Fig. 1). The 1H NMR studies revealed that complexes 1e12 are stable in the CDCl3 and DMSO‑d6 solution, because no chemical shift changes and/or new signals were detected after 72 h of standing at ambient temperature. 119 Cpph)(dpa)Cl]PF6 (6) adopts a pseudo-octahedral piano-stool coordination geometry, a typical one for similar cyclopentadienyl half-sandwich Ir(III) complexes (Fig. 2). The central Ir(III) atom is coordinated by the extended cyclopentadienyl ring (Cpph), chlorido ligand and by two nitrogen atoms (i.e, N(1) and N(1A)) of the bidentate-coordinated dpa ligand. The determined bond lengths and angles around the central Ir(III) atom are summarized in Table 2. Complex 6 is structurally similar to the previously reported Cpph complexes [Ir(h5-Cpph)(phen)Cl]PF6 (CCDC 802288), [Ir(h5Cpph)(bpy)Cl]PF6 (CCDC 802287) and [Ir(h5-Cpph)(en)Cl]BPh4 (CCDC 802290), which represent the only three deposited X-ray structures of Ir(III) chlorido complexes containing the Cpph ring and a chelating N,N-donor ligand; bpy ¼ 2,20 -bipyridyl, en ¼ ethylene1,2-diamine [3a]. The IreCl and IreN bond lengths of complex 6 (Table 2) are comparable with those detected for three abovementioned X-ray structures of similar Ir(III) half-sandwich Cpph complexes, for which the average values equalled 2.395(11) Å (for IreCl bond) and 2.10(2) Å (for IreN bonds). Also of interest for cyclopentadienyl half-sandwich complexes, the IreCg distance in complex 6 (Table 2) falls into the range of 1.7827e1.7945 Å reported for the Ir(III) Cpph chlorido complexes containing phen, bpy and en [3a]. The dihedral angle between both pyridines of the dpa ligand equals 33.75(9) , which is somehow higher than 28.00(7) calculated for the Rh(III) complex [Rh(h5-Cp*)(dpa)Cl]Cl$2H2O (CCDC 1012354) [19]. Another intraligand dihedral angle of complex 6 is formed by the cyclopentadienyl ring and its phenyl substituent and its value is 54.12(10) . Concerning the interligand dihedral angles between the cyclopentadienyl ring and both pyridines of dpa, they were calculated as 21.53(11) (for the N(1)-containing pyridine ring) and 23.64(10) (for the N(1A)-containing pyridine ring). As for the molecular structure of the cationic complex 6 (Fig. S9 and Table S1), it also contains the PF 6 anion, with the shortest Ir/P distance of 6.038(2) Å for Ir/P(1)vii (symmetry code: vii, -x, 1-y, 1z), connected with the complex cation via NeH/F, CeH/F and C/F non-covalent intermolecular contacts. In addition, there are also CeH/Cl contacts between the dpa and chlorido ligands of the adjacent complex cations stabilizing the crystal structure, and intermolecular CeH/C contacts detected between the Cpph and dpa ligands (see Table S1). The shortest Ir/Ir distance is 7.1049(10) for Ir$$$Irv (symmetry code: v, 1-x, -y, 1-z). 3.3. In vitro cytotoxicity testing In vitro cytotoxicity of complexes 1e12 was studied against the 3.2. X-ray structure of complex 6 The crystallographically characterized complex [Ir(h5- Fig. 2. The molecular structure of complex [Ir(h5-Cpph)(dpa)Cl]PF6 (6) showing the atom labelling scheme. Displacement ellipsoids are drawn at the 50% probability level. P. Starha et al. / Journal of Organometallic Chemistry 872 (2018) 114e122 120 Table 2 Selected bond lengths (Å) and angles ( ) determined by a single-crystal X-ray analysis for complex [Ir(h5-Cpph)(dpa)Cl]PF6 (6). Cg ¼ centroid of the cyclopentadienyl ring of the Cpph ligand. Bond angles ( ) Bond lengths (Å) IreN(1) IreN(1A) IreCl(1) IreC(20) IreC(21) IreC(22) IreC(23) IreC(24) IreCg 2.101(2) 2.097(2) 2.3877(7) 2.186(3) 2.151(3) 2.159(3) 2.157(3) 2.177(3) 1.7858(4) N(1)eIreN(1A) N(1)eIreCl(1) N(1)eIreCg N(1A)eIreCl(1) N(1A)eIreCg Cl(1)eIreCg C(2)eN(2)eC(2)A 83.63(9) 87.52(6) 126.52(7) 85.30(6) 126.93(7) 131.44(2) 126.1(2) complex) and ca. 0.9 (for Rh(III) complex) at ovarian carcinoma cells (cisplatin was used as a reference drug). Further, in contrast to complexes 7 and 12 (see Table 3), Cpph complex [Rh(h5-Cpph)(phen) Cl]PF6 (IC50 ¼ 57.0 mM) was less cytotoxic than its Cp* analogue [Rh(h5-Cp*)(phen)Cl]PF6 (IC50 ¼ 17.8 mM) at A2780 cells [4]. A2780 human ovarian carcinoma cells, which were treated by the studied agents for 24 h, followed by 72 h recovery time in a drugfree medium. The A2780 cell viability was evaluated by an MTT assay, providing the cell viability curves, from which the resulting IC50 values were calculated (see Table 3). The obtained results indicated that Ir(III) chlorido complexes 1 and 6 are markedly more in vitro cytotoxic than their Rh(III) analogues 7 and 12. Importantly, the impact of the Cp-ring extension from Cp* (complexes 1 and 7) to Cpph (complexes 6 and 12) on the in vitro cytotoxicity was positive in the case of both Ir(III) and Rh(III) pairs of agents. The replacement of the chlorido ligand by different halogenido ones did not led to any general trend of in vitro cytotoxicity. In particular, Ir(III) complexes 2 and 3 were less potent than chlorido complex 1, while Rh(III) iodido complex 9 exceeded the biological effect of both its halogenido analogues 7 and 8. It is noteworthy, that Rh(III) iodido complex 9 showed comparable cytotoxicity with Ir(III) chlorido complex 1. Concerning the carboxylato complexes, they were (complexes 4, 10 and 11) inactive up to the highest tested concentration in most cases, meaning that in the case of Ir(III) complexes the replacement of the chlorido ligand of complex 1 by the VP one (complex 4) caused a decrease of in vitro cytotoxicity against the A2780 cells. PB complex 5 was comparably cytotoxic with complex 1 against the A2780 cells, which is consistent with the results observed for a pair of chlorido and 4phenylbutyrato complexes [Ir(h5-Cpph)(phen)Cl]PF6 and [Ir(h5Cpph)(phen)(PB)]PF6, studied at the same human cancer cells [8]. All the complexes, including the best-performing complex 6, were less cytotoxic than the reference-drug cisplatin (IC50 ¼ 5.9 ± 1.2 mM) and the relative activity (RA) of complex 6 equalled ca. 0.3, which is comparable with the formerly reported Cpph complex [Ir(h5-Cpph)(phen)Cl]PF6 (RA ~ 0.2) [3a,5], and slightly higher than for [Ir(h5-Cpph)(bpy)Cl]PF6 (RA ~ 0.1) [3a,5] and the clinically investigated Ru(III) complex KP1019 (RA ~ 0.1) [24], studied at the same human cancer cells against cisplatin as the reference. A superior cytotoxicity of Ir(III) complex within pairs of analogical Ir(III) and Rh(III) chlorido complexes, as observed for Cp* complexes 1 and 7, and Cpph complexes 6 and 12 in this work, was also reported for a pair of half-sandwich Cp* Ir(III) and Rh(III) complexes containing 1,3,5-tris(di-2-pyridylaminomethyl)benzene [25]. In particular, RA of these agents equals ca. 0.5 (for Ir(III) 3.4. 1H NMR studies of hydrolytic stability Because chlorido complexes 1, 6, 7 and 12 were the most cytotoxic type of the studied iridium(III) and rhodium(III) complexes (except for the inactive complex 7 involved in the hydrolysis studies for comparative purposes), their hydrolytic stability in watercontaining media (20% DMF-d7/80% D2O and 20% MeOD-d4/80% D2O) was studied by 1H NMR; note: DMF-d7 and MeOD-d4 ensured the dissolution of low water-soluble complexes in the sufﬁcient concentration (i.e. 2 mM). 1 H NMR spectra of the best-performing complex 1 and 6 did not change in time in both the used mixtures of solvents (Fig. 3 and Fig. S10). In particular, only one set of the CeH signals of dpa was detected at 8.47 (C6eH), 8.05 (C4eH), 7.43 (C3eH) and 7.37 (C5eH) ppm in the spectra acquired for complex 1 on the fresh 20% DMFd7/80% D2O solution and after 48 h of standing at ambient temperature (Fig. 3). The position of the 1H NMR signals detected in the spectra of complex 1 did not change even after the addition of excess KCl (5 mol equiv.), proving these complexes as hydrolytically stable under the used conditions (Fig. 3). This is also evidenced by a difference in the signal positions of complex 1 in comparison with its dechlorinated species 1h (Table 4 and Fig. 3). The same was observed also for hydrolytically stable complex 1, whose 1H NMR spectra in 20% DMF-d7/80% D2O contained only one set of signals (even after 48 h of standing at 25 C) with different d values than observed for 1h (Table 4). In contrast to the above discussed iridium(III) chlorido complexes 1 and 6, their rhodium(III) analogues 7 and 12 hydrolysed rapidly, which is evidenced by the presence of two sets of 1H NMR signals of both dpa and Cpx (Cpx ¼ Cp* for 7 or Cpph for 12) ligands detected in their spectra (Table 4, Fig. 3 and Fig. S10), as observed in both the used water-containing media. The hydrolysis rate of ca. 30% (for 7) and ca. 25% (for 12) detected on the fresh 20% DMF-d7/ 80% D2O solution did not change in time up to 48 h of standing at Table 3 The results of in vitro cytotoxicity (IC50±SD; mM) of [M(h5-Cpx)(dpa)X]PF6 complexes 1e12 against the A2780 human ovarian carcinoma cells treated for 24 h followed by 72 h recovery in a drug-free media, given together with logP values of complexes 1e12. X Cpx M ¼ Ir IC50 (mM) logP M ¼ Rh IC50 (mM) logP Cl Br I VP PB Cl Cp* Cp* Cp* Cp* Cp* Cpph 1 2 3 4 5 6 70.6 ± 2.6 87.1 ± 5.3 83.0 ± 4.1 >100.0 76.9 ± 11.4 23.5 ± 1.1 0.52 ± 0.09 0.29 ± 0.03 1.69 ± 0.03 0.19 ± 0.01 0.17 ± 0.02 0.37.±0.02 7 8 9 10 11 12 >100.0 >100.0 70.1 ± 4.6 >100.0 >100.0 68.7 ± 7.2 1.45 ± 0.05 0.35 ± 0.02 1.29 ± 0.03 1.87 ± 0.00 1.52 ± 0.19 1.31 ± 0.15 P. Starha et al. / Journal of Organometallic Chemistry 872 (2018) 114e122 121 The formation of the hydroxido species rather than the aqua ones was proved for complexes 7 and 12 by the results of pH*dependent 1H NMR experiments, because the same chemical shifts were observed for the dechlorinated species at pH* ¼ equal to 7 and 10 (e.g., 8.36 and 7.96 ppm for C6eH, and C4eH, respectively, for complex 12h), while the d values changed as a consequence of pH* decrease to 4 (e.g., 8.43 and 8.04 ppm for C6eH and C4eH, respectively) [4]. 3.5. Hydrophobicity studies Hydrophobicity (lipophilicity) is an important feature of cytotoxic complexes known to be related to cellular accumulation and cytotoxicity itself. In this work, the logP values were determined by an octanol/water partition (Table 3). The obtained results showed, that iridium(III) complexes are more lipophilic than their rhodium(III) analogues. Unfortunately, no general conclusion can be drawn as for the comparison of lipophilicity and cytotoxicity. However, it can be seen in the case of iridium(III) complexes 1 and 6 that the Cpph chlorido complex 6 was more lipophilic than its Cp* analogue 1, which correlates with a higher anticancer potency of complex 6 over complex 1. Similar trend of lipophilicity and cytotoxicity of iridium(III) Cpph and Cp* complexes was reported for structurally very close complexes containing 1,10-phenanthroline or deprotonated 2-phenylpyridine [3a]. On the other hand, similar logP values were found in the case of rhodium(III) chlorido complexes 7 and 12, although they differ in cytotoxicity. Moreover, the following trend in lipophilicity of halogenido Cp* complexes 1e3 and 7e9 was observed for both iridium(III) and rhodium(III) complexes: I > Br > Cl. However, in the case of iridium(III) complexes, the least lipophilic complex 1 exceeded cytotoxicity of more lipophilic complexes 2 and 3. Fig. 3. 1H NMR studies of the representative complexes 1 and 7 in 20% DMF-d7/80% D2O, as observed at different time points (t ¼ 0 or 48 h) and at various conditions (þKCl stands for the addition of 5 M equiv. of KCl). Spectra of the dechlorinated species 1h and 7h were depicted as well for comparative purposes. ambient temperature. As a consequence of the addition of excess KCl, one set of signals disappeared in the 1H NMR spectra of complexes 7 and 12 (with d corresponding to the dechlorinated species 7h and 12h, respectively). Thus, the signals remaining in the spectra after the KCl addition can be unambiguously assigned to the studied chlorido complexes (Table 4, Fig. 3 and Fig. S10). 4. Conclusions A series of iridium(III) and rhodium(III) [M(h5-Cpx)(dpa)X]PF6 complexes 1e12 containing 2,20 -dipyridylamine (dpa), a variously substituted cyclopentadienyl ring (Cp* or Cpph), and halogenido or carboxylato monodentate ligands (X) was synthesized and thoroughly characterized. The crystallographically characterized complex [Ir(h5-Cpph)(dpa)Cl]PF6 (6) adopted a pseudo-octahedral piano-stool geometry, a typical one for similar half-sandwich complexes. Complexes 1e3, 5, 6, 9 and 12 showed moderate Table 4 The chemical shifts (in ppm) of the CeH signals of the dpa ligand, as observed by 1H NMR spectroscopy on the fresh solutions (t ¼ 0 h) of the chlorido complexes 1, 6, 7 and 12 in 20% DMF-d7/80% D2O or 20% MeOD-d4/80% D2O and after 48 h of standing at ambient temperature. For comparative purposes, the results acquired after the addition of excess KCl (labelled as “þ KCl”) and the results obtained for the dechlorinated species 1h, 6h, 7h and 12h in the same medium are given as well. Sample 1 1 (þKCl) 1h 6 6 (þKCl) 6h 7 7 (þKCl) 7h 12 12 (þKCl) 12h 20% DMF-d7/80% D2O 20% MeOD-d4/80% D2O C3eH C4eH C5eH C6eH Content 0h 48 h 7.43 7.44 7.49 7.44 7.46 7.50 7.42 7.49 7.43 7.48 7.42 7.49 7.43 7.49 8.05 8.06 8.10 8.04 8.05 8.09 8.09 8.15 8.10 8.15 8.06 8.13 8.06 8.14 7.37 7.37 7.43 7.27 7.28 7.34 7.44 7.52 7.45 7.52 7.32 7.42 7.33 7.43 8.47 8.48 8.54 8.47 8.47 8.52 8.50 8.57 8.50 8.57 8.47 8.55 8.47 8.55 100% 100% 100% 100% 70% 30% 70% 30% 75% 25% 75% 25% C3eH C4eH C5eH C6eH Content 0h 48 h 7.32 7.32 7.40 7.33 7.33 7.40 7.30 7.35 7.31 7.35 7.32 7.37 7.32 7.38 7.95 7.95 8.02 7.93 7.93 8.00 7.98 8.05 7.99 8.04 7.96 8.02 7.96 8.03 7.25 7.25 7.30 7.15 7.15 7.24 7.33 7.38 7.33 7.38 7.20 7.30 7.20 7.31 8.38 8.38 8.42 8.35 8.35 8.43 8.40 8.45 8.41 8.45 8.36 8.42 8.36 8.42 100% 100% 100% 100% 75% 25% 75% 25% 80% 20% 80% 20% P. Starha et al. / Journal of Organometallic Chemistry 872 (2018) 114e122 122 in vitro cytotoxicity at the A2780 human ovarian carcinoma cells (IC50 ¼ 23.5e87.1 mM). The obtained cytotoxicity results suggest that the replacement of the chlorido ligand of half-sandwich iridium(III) and rhodium(III) complexes by a different halogenido ligand (Br, I) or by a simple releasable bioactive O-donor ligand (valproato or 4-phenylbutyrato) has, in most cases, a negligible impact on the resulting potency of such agents. On the other hand, the positive effect of an extension of the cyclopentadienyl ring on in vitro cytotoxicity, recently reported for iridium(III) complexes, can be also accepted for half-sandwich rhodium(III) complexes, because complex [Rh(h5-Cpph)(dpa)Cl]PF6 (12; IC50 ¼ 68.7 mM) exceeds in vitro cytotoxicity of its Cp* analogue 7 (IC50 ˃ 100 mM) at A2780 human cancer cells. Notes Conﬂicts of interest None. Acknowledgements The authors gratefully thank the Ministry of Education, Youth and Sports of the Czech Republic (projects LO1305 and CZ.1.05/ 17-08512Y) 2.1.00/19.0377), the Czech Science Foundation (GACR and Palacký University Olomouc (PrF_2018_011) for the ﬁnancial support. The authors also thank Ms. Lucie Hanouskov a for her help for with syntheses of complexes 1e12, Ms. Katerina Kubesova in vitro cytotoxicity testing, Dr. Peter Antal for recording the NMR n Van spectra, Assoc. Prof. Ja co for performing ICP-MS experiments, Dr. Bohuslav Drahos for recording the ESI-MS spectra, Mrs. Pavla for performing elemental analyses and Dr. Alena Richterova for recording the FTIR spectra. Klanicova Appendix A. Supplementary data CCDC 1849574 contains the supplementary crystallographic data for this paper, and these data can be obtained free of charge from The Cambridge Crystallographic Data Centre via https://www. ccdc.cam.ac.uk. Supplementary data related to this article can be found at https://doi.org/10.1016. Appendix B. Supplementary data Supplementary data related to this article can be found at https://doi.org/10.1016/j.jorganchem.2018.07.035. References [1] C.H. Leung, H.J. Zhong, D.S.H. Chan, D.L. Ma, Coord. Chem. Rev. 257 (2013) 1764e1776. [2] L. Kelland, Nat. Rev. Canc. 7 (2007) 573e584. [3] (a) Z. Liu, P.J. Sadler, Acc. Chem. Res. 47 (2014) 1174e1185; (b) J. Ruiz, V. Rodríguez, N. Cutillas, K.G. Samper, M. Capdevila, O. Palacios, A. Espinosa, Dalton Trans. 41 (2012) 12847e12856; (c) P. Zhang, P.J. Sadler, J. Organomet. Chem. 839 (2017) 5e14; (d) C.C. Konkankit, S.C. Marker, K.M. Knopf, J.J. Wilson, Dalton Trans. (2018), https://doi.org/10.1039/C8DT01858H. Advance Article. n, P.J. Sadler, J. Inorg. [4] J.J. Soldevila-Barreda, A. Habtemariam, I. Romero-Canelo Biochem. 153 (2015) 322e333. [5] Z. Liu, A. Habtemariam, A.M. Pizarro, S.A. Fletcher, A. Kisova, O. Vrana, L. Salassa, P.C.A. Bruijnincx, G.J. Clarkson, V. Brabec, P.J. Sadler, J. Med. Chem. 54 (2011) 3011e3026. n, A. Habtemariam, G.J. Clarkson, P.J. Sadler, Or[6] (a) Z. Liu, I. Romero-Canelo ganometallics 33 (2014) 5324e5333; n, G.J. Clarkson, P.J. Sadler, (b) P. Starha, A. Habtemariam, I. Romero-Canelo Inorg. Chem. 55 (2016) 2324e2331. n, L. Salassa, P.J. Sadler, J. Med. Chem. 56 (2013) [7] (a) I. Romero-Canelo 1291e1300; (b) M. Kubanik, H. Holtkamp, T. Sohnel, S.M.F. Jamieson, C.G. Hartinger, Organometallics 34 (2015) 5658e5668. vní k, Dalton Trans. 47 [8] (a) P. Starha, Z. Tra cek, R. Herchel, P. Jewula, Z. Dvora (2018) 5714e5724; vní k, Appl. Organomet. (b) P. Starha, Z. Tra cek, B. Drahos, R. Herchel, Z. Dvora Chem. 32 (2018) e4246; vní k, Molecules 23 (2018) 420. (c) P. Starha, Z. Tra cek, J. Van co, Z. Dvora [9] G. Agonigi, T. Riedel, M.P. Gay, L. Biancalana, E. Oneate, P.J. Dyson, G. Pampaloni, E. Paunescu, M.A. Esteruelas, F. Marchetti, Organometallics 35 (2016) 1046e1056. € nnemann, J. Risse, Z. Grote, R. Scopelliti, K. Severin, Eur. J. Inorg. Chem. [10] J. To 2013 (2013) 4558e4562. ~ a, J.S. Merola, Polyhedron 84 (2014) [11] D.M. Morris, M. McGeagh, D. De Pen 120e135. vní , Z. Dvor [12] P. Starha, Z. Tra cek, R. Krikavova ak, Molecules 21 (2016) 1725. [13] H.E. Gottlieb, V. Kotlyar, A. Nudelman, J. Org. Chem. 62 (1997) 7512e7515. [14] Apex3, Bruker AXS Inc, Madison, WI, 2015. [15] G.M. Sheldrick, Acta Crystallogr. C71 (2015) 3e8. [16] K. Brandenburg, Diamond Version 4.0.3, Crystal Impact GbR, Bonn, Germany, 2015. [17] C.F. Macrae, I.J. Bruno, J.A. Chisholm, P.R. Edgington, P. McCabe, E. Pidcock, L. Rodriguez-Monge, R. Taylor, J. van de Streek, P.A. Wood, J. Appl. Crystallogr. 41 (2008) 466e470. [18] D.W. Brogden, J.F. Berry, Comments Mod. Chem. 36 (2016) 17e37. [19] P.A. Vekariya, P.S. Karia, J.V. Vaghasiya, S. Soni, E. Suresh, M.N. Patel, Polyhedron 110 (2016) 73e84. [20] E. Bencze, B.V. Lokshin, J. Mink, W.A. Herrmann, F.E. Kühn, J. Organomet. Chem. 55 (2001) 55e66. [21] C.J. Pouchert, The Aldrich Library of Infrared Spectra (Ed. III), Aldrich Chemical Co., Milwaukee, WI, USA, 1981. [22] (a) H.G. Brittain, Cryst. Growth Des. 10 (2010) 1990e2003; (b) G. Petrusevski, P. Naumov, G. Jovanovski, G. Bogoeva-Gaceva, S.W. Ng, ChemMedChem 3 (2008) 1377e1386. [23] P. Govindaswamy, Y.A. Mozharivskyj, M.R. Kollipara, Polyhedron 24 (2005) ̆ 1710e1716. [24] M. Groessl, O. Zava, P.J. Dyson, Metall 3 (2011) 591e599. [25] A.B.P. Rao, A. Uma, T. Chiranjeevi, M.S. Bethu, J.V. Rao, D.K. Deb, B. Sarkar, W. Kaminsky, M.K. Rao, J. Organomet. Chem. 824 (2016) 131e139.