Organometallic Nucleosides: Synthesis and Biological Evaluation of Substituted Dicobalt Hexacarbonyl 2'-Deoxy-5-oxopropynyluridines.

Reactions of dicobalt octacarbonyl [Co2(CO)8] with 2'-deoxy-5-oxopropynyluridines and related compounds gave dicobalt hexacarbonyl nucleoside complexes (83-31 %). The synthetic outcomes were confirmed by X-ray structure determination of dicobalt hexacarbonyl 2'-deoxy-5-(4-hydroxybut-1-yn-1-yl)uridine, which exhibits intermolecular hydrogen bonding between a modified base and ribose. The electronic structure of this compound was characterized by the DFT calculations. The growth inhibition of HeLa and K562 cancer cell lines by organometallic nucleosides was examined and compared to that by alkynyl nucleoside precursors. Coordination of the dicobalt carbonyl moiety to the 2'-deoxy-5-alkynyluridines led to a significant increase in the cytotoxic potency. The cobalt compounds displayed antiproliferative activities with median inhibitory values (IC50) in the range of 20 to 80 μm for the HeLa cell line and 18 to 30 μm for the K562 cell line. Coordination of an acetyl-substituted cobalt nucleoside was expanded by using the 1,1-bis(diphenylphosphino)methane (dppm) ligand, which exhibited cytotoxicity at comparable levels. The formation of reactive oxygen species in the presence of cobalt compounds was determined in K562 cells. The results indicate that the mechanism of action for most antiproliferative cobalt compounds may be related to the induction of oxidative stress.

Introduction

Nucleoside analogues are molecules of high pharmacological interest for the treatment of various conditions, especially cancer and viral diseases.1, 2, 3, 4 These agents behave as antimetabolites and compete with physiologic nucleosides, and consequently, they interact with a large number of intracellular targets to induce cytotoxicity. Substitution at the C‐5 position of the uracil base provides a common framework for potent biological properties.5

In parallel, bioorganometallic chemistry provides new tools to influence biological interactions.6, 7, 8, 9, 10, 11, 12, 13 Among a variety of organometallic compounds, transition‐metal carbonyls forged their presence in medicinal chemistry.14 For example, a ruthenium carbonyl complex, a protein kinase inhibitor, activates p53 and induces apoptosis in human melanoma cells.15 Related complexes were designed distinctly: as tamoxifen‐based anticancer drug derivatives,16, 17 as inhibitors of human carbonic anhydrase (hCA),18 as antibacterial agents,19 and as triazoles displaying antitrypanosomal activity.20 Representative nucleoside‐related examples include rhenium tricarbonyl complexes such as 1 (Figure 1)21 and manganese and chromium tricarbonyl arylalkynyl nucleosides 2 (Figure 1).22 Nucleoside–iron carbonyl complex 3 (Figure 1) was reported to bestow significant apoptosis‐inducing activity against BJAB tumor cells23 and specific cytotoxicity to reactive oxygen species (ROS)‐stressed cancer cells.24

Figure 1

Structures of representative metal–carbonyl complexes with medicinal potential.

Cobalt possesses a diverse array of properties that can be manipulated to yield promising drug candidates.25 The medicinal potential of cobalt carbonyl complexes has been reported.26 Peptide labeling by using a dicobalt hexacarbonyl alkynyl complex27 and cobalt carbonyl complexes encapsulated in a micelle structure (not illustrated), aiming to deliver cobalt pharmaceuticals,28 has been investigated. The activity of a cobalt derivative of 17‐ethynyltestosterone (4; Figure 1) has also been explored.29 Recently, two cobalt‐based hybrid molecules that combine an Nrf2 (a basic leucine zipper protein) inducer with a releaser of carbon monoxide (an anti‐inflammatory product of heme oxygenase‐1) were reported to increase Nrf2/H‐O1 expression markedly and to exert anti‐inflammatory activity in vitro. Compounds 5 and 6 (Figure 1) also up‐regulate tissue heme oxygenase‐1 and deliver CO to the blood after administration in vivo, which supports their potential anti‐inflammatory properties.30

The antiproliferative properties of dicobalt hexacarbonyl complexes have been noted,31, 32, 33, 34, 35 and aspirin–cobalt carbonyl derivative 7 containing a propargyl alcohol unit (Figure 1) was identified as a lead compound during in vitro studies against MCF‐7 and MDA‐MB‐231 human mammary breast cancer cells.33 Additional studies have suggested a mode of action in which cyclooxygenase inhibition plays a major role.34 The aspirin–cobalt carbonyl derivative also exhibited antiangiogenic effects in a zebrafish embryo assay, as well as significant inhibition of matrix metalloproteinase‐7 (MMP‐7).36 Among its many other demonstrated biological effects was a strong induction of caspase‐3, which is an important activator in apoptosis. A related aromatic compound with a nitro substituent, that is, 8 (Figure 1), was shown to strongly induce apoptosis, arrest the cell cycle at the S phase, increase cellular oxidative stress levels, and induce permeability of the mitochondrial membrane.37 Whereas its non‐cobalt‐containing precursor also caused an increase in mitochondrial membrane permeability, it did not produce an increase in oxidative stress levels, nor did it have apoptosis‐inducing or antiproliferative effects.

A review of structures 4–8 (Figure 1) revealed a common cobalt carbonyl propargyloxy ligand (−C≡CCH2O−), which appears in the form of its ester or free alcohol. Thus, we were intrigued as to whether the combination of a nucleoside with the presence of a propargyloxy structural motif would increase the potency of such conjugates. Consequently, the design of the investigated compounds included several propynoxy (propargyloxy) and related units, as discussed below.

Despite numerous advances, the repertoire of metallonucleosides is still limited. Interest in the use of the ethynyl (acetylenic) fragment for the modification of nucleoside bases has resulted in a great number of applications.5 Our earlier studies confirmed activity against MCF‐7 and MDA‐MB‐231 human mammary breast cancer cells of alkyl‐ and aryl‐substituted cobalt hexacarbonyl nucleosides,38 as well as their precursors, 5‐alkynyl uridines.39 We were intrigued to investigate further 2′‐deoxy‐5‐alkynyluridines and their hexacarbonyl dicobalt adducts in particular to pursue the synthesis and biological evaluation of a combination of oxopropynyl (oxopropargyl), cobalt carbonyl, and nucleoside structural features as new target compounds. Enhancement of action through installment of the acetyl group was also explored.

Results and Discussion

2′‐Deoxy‐5‐alkynyluridines 9 a–h (Scheme 1), containing a conjugated alkyne function, represent versatile materials for reactions that lead, among others, to furopyrimidines,40 halofuropyrimidines,41 and alkynyl dimers.42

Scheme 1

Synthesis of 2′‐deoxy‐5‐alkynyluridines 9 a–h from 2′‐deoxy‐5‐ iodouridine (I‐dU).

Sonogashira coupling of alkynes offers an atom‐efficient pathway toward modification of nucleosides.39, 43 Preparative synthesis of a series of 2′‐deoxy‐5‐alkynyluridines 9 was performed from 2′‐deoxy‐5‐iodouridine (I‐dU) and the appropriate terminal alkyne in the presence of catalytic amounts of Pd(PPh3)4, copper(I) iodide, and triethylamine in DMF in a tightly controlled temperature regime (40 °C, oil bath, 22 h; Scheme 1) to avoid subsequent cyclization to furopyrimidines; this protocol did not require protection of the hydroxy groups.39 To explore the structure–reactivity relationship of the alkyne substituents, nucleosides containing a oxopropynyl (oxypropargyl) unit (propargyl alcohol, 9 a, R=CH2OH), alkyl disubstituted [9 b/9 c, R=C(OH)(Me)2/C(OH)(Me)Et], aryl mono‐ and disubstituted [9 d/9 e, R=CH(OH)Ph/C(OH)Ph2], as well as the one‐carbon‐extended homologue (9 f, R=CH2CH2OH), methyl ether (9 g, R=CH2OMe), and acetate (9 h, R=CH2OAc) groups, were obtained in 92–41 % yield by using literature procedures (Table 1). Numerous 2′‐deoxy‐5‐alkynyluridines have been synthesized because of their interesting biological activity. However compounds 9 a–f and 9 h, to our knowledge, have not yet been reported.44 The structures of alkynyl nucleosides 9 were confirmed by 1H NMR and 13C NMR spectroscopy. The high‐resolution mass spectra of 9 a–h exhibit m/z signals of [M+H]+ as molecular ions.

Table 1

Preparation of 2′‐deoxy‐5‐alkynyluridines 9 a–h and conversion into hexacarbonyl dicobalt derivatives 10 a–h.

R	Alkynyl uridine 9	Yield [%]	Cobalt complex 10	Yield [%]
CH2OH	9 a	50	10 a	58
C(OH)Me2	9 b	41	10 b	31
C(OH)(Me)Et	9 c	56	10 c	64
CH(OH)Ph	9 d	46	10 d	71
C(OH)Ph2	9 e	92	10 e	62
CH2CH2OH	9 f	54	10 f	54
CH2OMe	9 g 44	69	10 g	83
CH2OAc	9 h	66	10 h	83

The conversion of alkynyl nucleosides 9 into 10, which are the corresponding dicobalt hexacarbonyl complexes of the 2′‐deoxy‐5‐alkynyluridines, was accomplished at room temperature [Co2(CO)8, 22 °C, 1 h], and these compounds were obtained in 83–31 % yield after silica gel column chromatography (Scheme 1, Table 1).

The structures of new nucleosides 10 a–h were confirmed by NMR and IR spectroscopy and HRMS. In most cases, the 1H NMR spectra exhibit one signal in the region of H6, as expected (δ=8.35–8.11 ppm). However, two signals are observed for compounds 10 c and 10 d (δ=8.30/8.27 and 8.17/8.11 ppm, respectively), presumably due to the epimers at the stereocenter located at the propargyl carbon atom. It can be assumed that most molecules, due to steric reasons, would exist in solution in the conformation that resembles the one observed in the crystal structure of 10 f. This includes the planar assembly of C6−C5−C7−C8−C9, leading to potential hydrogen‐bonding engagement of the propargyl and H‐O5′ hydroxy groups, which would lead to restricted rotation across the C8−C9 bond. Under such circumstances, H6 would be positioned gauche to the hydroxy group and one of the substituents of the stereocenter, which would lead to nonmagnetically equivalent environments for each of the propargyl (C9) epimers. The H6 signals for 10 c/10 d are separated by 21.4/30.9 Hz, respectively, in line with the anticipated lower magnetic impact of the alkyl group versus the phenyl group. The variable‐temperature 1H NMR spectra show significant line broadening at 60 °C, but incomplete coalescence (CDCl3; higher temperature leads to decomposition). The reformation of well‐separated signals is observed upon returning to 20 °C. In the 13C NMR spectrum, nonequivalency of the cobalt carbonyl signals (CoCO) is observed only for 10 d (δ=199.45 and 198.93 ppm). The IR spectra exhibit characteristic bands for the alkyne [Co2(CO)6] group [ around 2092 (m), 2052 (s), and 2016 cm−1 (vs)]. Selected representative compounds as well as their alkynyl counterparts (i.e. compounds 9 a/10 a, 9 g/10 g, and 9 h/10 h) gave acceptable elemental analyses. The high‐resolution mass spectra exhibit m/z signal of ions [M−OH]+ for 10 a–f and 10 g and [M−OAc]+ for 10 h, which are presumably formed due to a dehydration or deacylation reaction during acquisition; the ability of the dicobalt hexacarbonyl moiety to stabilize a positive charge at an adjacent carbon atom is well known. Nucleoside 10 f containing a homopropargyl group, which lacks an option to form a well‐stabilized carbocation, shows the m/z [M+H]+ signal for the molecular ion.

Efforts to obtain diffraction‐quality crystals were so far only successful in the case of 10 f by diffusion of pentane into a THF solution (−20 °C) under a dry nitrogen atmosphere. X‐ray crystallography confirms the structure of dicobalt hexacarbonyl 2′‐deoxy‐5‐(4‐hydroxybut‐1‐yn‐1‐yl)uridine (10 f). Although nucleosides are often resistant to crystallization, intramolecular hydrogen bonding O10⋅⋅⋅H‐O5′ forms a 14‐membered ring that rigidifies the structure and presumably facilitates formation of the X‐ray‐quality crystals (Figure 2).45 Coordination of the alkyne to the dicobalt changes the position of the CH2CH2OH group relative to the pyrimidine base through the C−C≡C−C planar unit of the cobalt complex with angles of 142–143°. The R group is directed towards the ribose unit and is long enough to create a contact of O10 with the hydrogen atom of the 5′‐hydroxy group O10⋅⋅⋅H‐O5′ 2.809(7) Å [calcd 2.756 Å].46

Figure 2

ORTEP view of 10 f with the atom‐labeling scheme. Thermal ellipsoids are drawn at the 50 % probability level. Selected interatomic distances [Å] (calcd values are given in square brackets): C5−C7 1.452(10) [1.442], C7−C8 1.333(9) [1.359], C8−C9 1.520(9) [1.490], Co1−Co2 2.4758(16) [2.404], O10⋅⋅⋅H‐O5′ 2.809(7) [2.756]; key angles [°]: C5−C7−C8 142.9(7) [140.937], C7−C8−C9 142.1(7) [141.257].

In the crystalline form, the hydrogen atoms of the remaining hydroxy groups (H‐O3′ and CH2CH2OH) are each stabilized by a molecule of THF (O3′⋅⋅⋅O17 2.791 Å, calcd 2.754 Å; O18⋅⋅⋅O10 2.750 Å; O18A⋅⋅⋅O10 2.703 Å, calcd 2.693 Å). The C2 carbonyl group of 10 f adopts an anti orientation towards the ribose ring: the glycosidic bond torsion angle χ (O4′−C1′−N1−C2) is 108.3(6)°. The Co−Co bond is perpendicular to the uracil plane, and the dicobalt carbonyl unit is located syn to the ribose ring.

Computational Studies

DFT calculations were conducted on 10 f in the gas phase to optimize the ground‐state structure.47 The energies were acquired by using PBE0/6‐31G*. Selected calculated metric parameters for the geometry‐optimized structure were compared to the experimental results (Figure 2). The largest difference between the experimental and calculated bond lengths was 0.07 Å (Co1−Co2), whereas the bond angles were in good agreement. Despite slight differences, the calculated structure is quite close to the experimental structure (even reproducing hydrogen bonds), allowing electronic properties to be confidently extracted. The calculated Mulliken charge value for Co1 and Co2 in 10 f is −0.08, consistent with neutrality. Figure 3 illustrates selected molecular orbitals: HOMO, HOMO−1, LUMO, and LUMO+1 for 10 f. The HOMO (α−187) is largely distributed over the cobalt d orbitals and the π system of the alkyne and pyrimidine base. The LUMO (α−188) is primarily distributed over the cobalt d orbitals and the π system of the alkyne group and represents an antibonding orbital between the cobalt centers. HOMO−1 (α−186) represents the bonding interaction between the cobalt d orbitals. The HOMO–LUMO gap for 10 f is large (4.25 eV) consistent with high kinetic stability.48, 49

Figure 3

Plots of molecular orbitals: HOMO−1, HOMO, LUMO, and LUMO+1 for 10 f. Orbital energies [eV] are indicated.

Synthesis of Co2(CO)4(dppm) Complex

We were intrigued to investigate the effect of coordinating a phosphine ligand to the cobalt complex. Phosphine ligands have been employed in complexes screened for anticancer properties.50 The affinity of the 1,1‐bis(diphenylphosphino)methane (dppm) ligand towards cobalt carbonyls is well known,51 and to our knowledge, no biological studies have been reported so far for (alkyne)Co2(CO)4(dppm)‐connected compounds. Dicobalt hexacarbonyl alkyne complexes are known to react with dppm; however, we elected first to coordinate dppm to the cobalt carbonyl ligand and, subsequently, to react the product with a free alkyne nucleoside. Accordingly, acyl‐containing nucleoside 9 h was combined with Co2(CO)6(dppm), which was prepared from dicobalt octacarbonyl and dppm in toluene following a known procedure.52 Workup gave nucleoside 11 in 25 % yield (Scheme 2). Although the presence of dppm in the molecule could enhance the crystallinity, efforts to obtain an X‐ray‐quality crystal of 11 have not been successful thus far.

Scheme 2

Synthesis of dppm–dicobalt tetracarbonyl nucleoside 11 from alkyne 9 h.

The structure of new nucleoside 11 was confirmed by 1H NMR, 13C NMR, 31P NMR, and IR spectroscopy and HRMS. The 31P NMR spectrum features a characteristic resonance at δ=41.05 ppm in [D6]DMSO (δ=40.84 ppm in CDCl3), and this clearly differs from the spectrum of Co2(CO)6(dppm).52 The IR spectrum exhibits the characteristic νCO pattern at =2022, 1990, and 1961 cm−1.53 Similar to the high‐resolution mass spectrum of 10 h, an m/z signal for [M−OAc]+ is observed.

Inhibition of Cell Proliferation

The cytotoxic properties of compounds 9 a–h, 10 a–h, and 11 were tested for their activity in HeLa (human cervix carcinoma), K562 (chronic myelogenous leukemia), and HUVEC (human umbilical vein endothelial cells) cells. As the control (100 % viability in the MTT [3‐(4,5‐dimethylthiazol‐2‐yl)‐2,5‐diphenyltetrazolium bromide] assay), cells treated with DMSO (1 %) were used. The viability of cells was determined at six different drug concentrations: 2×10−1, 1×10−1, 5×10−2, 1×10−2, 1×10−3, and 1×10−4 mm. As the control of the whole experiment, staurosporine (1 μm) was used.

All cobalt compounds displayed significant antiproliferative effects with median inhibitory concentration (IC50) values reaching 20(±5.1) μm (for 10 e in HeLa cells) and 16(±3.5) μm (for 11 in K562 cells). Thus, the potency of the more active target compounds is well within the range of established anticancer drugs such as cisplatin, 20(±6.0) and 40(±7.0) μm in HeLa and K562 cells, respectively (Table 2).

Table 2

Cytotoxic activity (effect of the substituents) of cobalt nucleosides 10 a–h and 11 for the proliferation of the HeLa, K562, and HUVEC cell lines after incubating for 48 h.

Compound	IC50 [a] [μm]
	HeLa	K562	HUVEC
10 a	34±3.4	28±4.3	40±7.6
10 a+NAC (1 mm)	140±8.4	90±7.5
10 a+NAC (2 mm)	>200	>200
10 b	28±4.5	18±3.6	20±2.0
10 c	80±6.1	29±4.7	40±4.8
10 d	80±6.5	25±3.6	31±3.1
10 e	20±5.1	18±1.9	16±2.6
10 f	80±5.7	30±3.7	35±5.1
10 g	50±4.6	28±4.8	20±3.8
10 h	30±5.0	22±4.7	37±5.5
11	25±3.2	16±3.5	7±5.2
cisplatin	20±6.0	40±7.0	30±6.5

[a] Results were obtained in two separate experiments, each n=6.

The structure–reactivity relationship seems not to be straightforward. The lead compound for cytotoxic hexacarbonyl dicobalt complexes was diphenyl‐substituted 10 e (IC50 values in the range of 18 to 20 μm in this assay). HeLa cells were less sensitive towards the structure of metallonucleosides than K562 cells (20–80 vs. 18–30 μm, respectively). Cobalt nucleosides 10 c, 10 d, and 10 f showed selectivity towards K562 cells and were less active in HeLa cells. This selectivity almost disappeared for derivative 10 e, as good antiproliferative effects could be noted in both cells.

For corresponding cobalt carbonyl species 10 h and its dppm homologue 11, the antiproliferative activity was slightly increased for the latter (IC50 values from 30 to 25 μm and from 22 to 16 μm); however, toxicity against HUVEC cells also increased. Interestingly, dppm derivative 11, containing fewer carbonyl groups than 10 h, was more active in both tumor cell cultures investigated (Table 2).

In regard to the alkyne precursors, preliminary results obtained for noncoordinated alkynyluridines 9 a–h were all above >100 μm; thus, no further screening was attempted. Clearly, coordination of the alkynes to Co2(CO)6 has a strong influence on the biological activity of the respective alkyne compounds. In general, the coordination process led to a significant increase in the cytotoxic potency for all substituents at the non‐nucleoside side of the alkyne.

Oxidative Stress

Uncontrolled production of reactive oxygen species (ROS) in cells results in oxidative stress that impairs cellular functions and contributes to the development of cancer, chronic disease, and toxicity.54, 55 Growing evidence suggests that cancer cells exhibit increased intrinsic ROS stress, due in part to oncogenic stimulation, increased metabolic activity, and mitochondrial malfunction.54 Given that the mitochondrial respiratory chain (electron‐transport complexes) is a major source of ROS generation in cells, the vulnerability of mitochondrial DNA to ROS‐mediated damage appears to be a mechanism to amplify ROS stress in cancer cells. As this state of oxidative stress makes cancer cells vulnerable to agents that further augment ROS levels, the use of pro‐oxidant agents is emerging as an exciting strategy to target tumor cells selectively.55, 56 Accumulation of peroxides and free radicals leads to death of cancer cells, for example, by apoptosis. This biochemical aspect can be exploited to develop novel therapeutic drugs to target cancer cells preferentially and selectively through ROS‐mediated mechanisms.56, 57 The contribution of reactive oxygen species to the antitumor activity of many chemotherapeutic agents commonly used in treatment has already been documented.56

To gain insight into the induction of oxidative stress, which may be responsible for part of the antiproliferative activity of the investigated compounds against human cancer cells, the levels of reactive oxygen species in the presence of cytotoxic compounds 10 a–h and 11, as well as representative free‐alkyne corresponding nucleoside 9 c, were determined in K562 (chronic myelogenous leukemia) cells by using DCFDA (2,7‐dichlorofluorescein diacetate) dye assay. After staining, the cells were incubated for 4 h with the test compounds at concentrations of 5, 10, 20, 50, 100, and 200 μm. Cells treated with 50 and 100 μm H2O2 served as positive controls. All cobalt compounds showed a significant increase in the level of reactive oxygen species, which is confirmed by the DCF (2,7‐dichlorofluorescein) fluorescence intensity values shown in Table 3.

Table 3

Oxidative stress of cobalt nucleosides 10 a–h and 9 c (bar plot for 100 μm).

Compound	DCF fluorescence intensity[a]
	5 μm	10 μm	20 μm	50 μm	100 μm	200 μm
DMSO+DCFDA	1±0.0[b]
H2O2				3.7±0.10[c]	6.1±0.17[d]
9 c	0.9±0.01	0.9±0.03	0.9±0.02	0.9±0.02	0.9±0.03	0.9±0.04
10 a	1.2±0.06	1.7±0.19	1.9±0.03	3.2±0.02	3.6±0.09	3.8±0.04
10 b	1.2±0.03	1.6±0.42	1.7±0.04	3.0±0.04	3.6±0.06	4.2±0.09
10 c	1.2±0.05	1.7±0.19	2.1±0.14	3.5±0.04	4.3±0.02	4.7±0.01
10 d	1.3±0.15	1.6±0.13	2.1±0.17	3.5±0.05	4.2±0.07	4.7±0.09
10 e	1.2±0.03	1.5±0.01	1.9±0.04	3.5±0.09	4.0±0.04	4.9±0.03
10 f	1.3±0.18	1.6±0.17	2.0±0.12	3.2±0.02	3.8±0.04	4.1±0.02
10 g	1.2±0.03	1.6±0.18	1.7±0.04	3.2±0.05	3.6±0.04	4.1±0.04
10 h	1.1±0.07	1.3±0.03	1.7±0.04	3.0±0.04	3.4±0.01	3.9±0.03
11	1.4±0.10	1.6±0.01	2.5±0.02	3.7±0.06	4.2±0.02	5.2±0.06

[a] Change normalized to control sample. [b] Control, concentration 1 % DMSO, 20 μm DCFDA. [c] Positive control, concentration 50 μm. [d] Positive control, concentration 100 μm.

The DCF fluorescence intensity in the presence of the tested cobalt compounds (1.1±0.07 to 5.2±0.06‐fold change in relation to control cells) was comparable to the value for the positive control (i.e. H2O2, 3.7±0.10‐ to 6.1±0.17‐fold change in relation to control cells), which is one of the reactive forms of oxygen. The DCF fluorescence intensity increased for all of the tested cobalt compounds upon increasing the concentration of the cobalt complexes. Studies showed that compound 11 (1.4±0.10‐ to 5.2±0.06‐fold change in relation to control cells) was the most potent inducer of oxidative stress in K562 cells.

Confirmation that one of the major mechanisms of action of cobalt nucleoside derivatives involved overproduction of ROS was sought. Therefore, the effect of representative compound 10 a in the presence of a generalized intracellular ROS inhibitor was investigated. HeLa and K562 cancer cells were incubated in the presence of N‐acetyl‐l‐cysteine (NAC). Pretreatment of HeLa and K562 cells (30 min) with NAC solutions (1 or 2 mm) enhanced their viability by 20–51 and 10–56 %, respectively, compared to treatment with 10 a alone (Figure 4). Thus, the role of ROS induction in the potency of the cobalt complexes was confirmed.

Figure 4

Survival rate of HeLa (top) and K562 (bottom) cells after incubating with compounds 10 a and 10 a+NAC (1 or 2 mm) for 48 h.

Also, non‐cobalt nucleoside derivative 9 c was studied, and this compound did not increase the level of reactive oxygen species (0.9±0.03‐fold change in relation to control cells). This result confirms that the cobalt atom present in the structure is responsible for the pro‐oxidative properties of the cobalt nucleosides. In conclusion, the results show that the mechanism of action of most antiproliferative cobalt compounds in leukemia cells (K562) may be related to the induction of oxidative stress.

Conclusions

Novel 5‐oxopropynyl‐substituted 2′‐deoxyuridines containing alkyl and aryl substituents at the propargyl carbon atom and hydroxy, methoxy, and acetoxy groups were synthesized and converted into their dicobalt hexacarbonyl derivatives. X‐ray crystallography confirmed the structure of the cobalt nucleoside containing the homopropargyl alcohol (but‐3‐yn‐1‐ol) unit (see compound 10 f) and determined the conformation and hydrogen bonding present in the solid state. The acetoxy cobalt derivative was converted into the corresponding complex containing a bidentate 1,1‐bis‐(diphenylphosphino)methane (dppm) ligand. It was demonstrated that the presence of the cobalt carbonyl was essential to achieve a cytotoxic effect, as the alkynyl precursor did not exhibit pronounced activity against HeLa and K562 human cancer cells in vitro, with higher efficacy seen in the last one. Similar to the results obtained for other cobalt carbonyl complexes, the activity depended on the presence of the cobalt carbonyl moiety, suggesting that metal carbonyls are useful functional groups for modifying or inducing biological activity.

Experimental Section

General Instrumentation

All NMR spectroscopy measurements were performed with a Bruker Avance III 500 spectrometer operating at 500.13 MHz for 1H, 125.75 MHz for 13C, and 202.45 MHz for 31P at 22 °C. Mass spectra were recorded with an Agilent 6520 Q‐TOF LC–MS (HRMS). FTIR spectra were recorded with ATI Mattson Infinity Series AR60, ThermoScientific Nicolet 6700 ATR, and Bruker Alpha‐P ATR spectrometers. All products were stored in a refrigerator (4 °C).

Syntheses

General Procedure for the Synthesis of 2′‐Deoxy‐5‐ alkynyluridines 9 a–h

A round‐bottomed flask was charged with 2′‐deoxy‐5‐iodouridine (I‐dU; 2.09 g, 5.90 mmol), Pd(PPh3)4 (0.690 g, 0.597 mmol), CuI (0.120 g, 0.628 mmol), DMF (15 mL), Et3N (1.8 mL, 12 mmol), and the respective terminal alkyne. The mixture was stirred at 40 °C for 22 h (oil bath). The solvent was removed (by oil pump vacuum), and the residue was dissolved in chloroform (10 mL) and kept in the freezer (−4 °C) for 12 h. The precipitate was filtered off and washed with cold chloroform (3×3 mL). The mother liquor was concentrated and subjected to column chromatography (230–400 mesh silica gel, 0→12 % MeOH/CHCl3). The product was dried by oil pump vacuum for 2 h to give 9 a–h.

2′‐Deoxy‐5‐(3‐hydroxyprop‐1‐yn‐1‐yl)uridine (9 a): From 2‐propyn‐1‐ol (propargyl alcohol, 0.86 mL, 14.8 mmol). Yield: 0.830 g, 2.94 mmol (50 %). R f=0.25 (CHCl3/MeOH 7:1). 1H NMR ([D6]DMSO): δ=11.57 (s, 1 H, N‐H), 8.14 (s, 1 H, H‐6), 6.07 (t, J=6.5 Hz, 1 H, H‐1′), 5.26 (t, J=6.1 Hz, 1 H, OH‐3′), 5.21 (d, J=4.51 Hz, 1 H, OH), 5.08 (t, J=5.15 Hz, 1 H, OH‐5′), 4.20 (s, 2 H, CH2), 4.19 (s, 1 H, H‐3′), 3.77–3.74 (m, 1 H, H‐4′), 3.60–3.55 (m, 1 H, H‐5′), 3.54–3.49 (m, 1 H, H‐5“), 2.10–2.06 ppm (m, 2 H, H‐2′, H‐2”); 13C NMR ([D6]DMSO): δ=163.10, 150.91, 144.99, 99.69, 93.97, 89.06, 86.15, 77.80, 71.67, 62.44, 50.96, 40.38 ppm; IR (KBr): =3401 (br m), 2924 (m), 1726 (s), 1667 (s), 1468 (m), 1279 (m), 1051 (m) cm−1; HRMS (ESI‐TOF): m/z: calcd for C12H15N2O6: 283.0925 [M+H]+; found: 283.0927; elemental analysis calcd (%) for C12H14N2O6 (282.25): C 51.07, H 5.00, N 9.93; found: C 51.18, H 5.49, N 9.79.

2′‐Deoxy‐5‐(3‐hydroxy‐3‐methylbut‐1‐yn‐1‐yl)uridine (9 b): From 2‐methyl‐3‐butyn‐2‐ol (1.43 mL, 14.8 mmol). Yield: 0.750 g, 2.42 mmol (41 %). R f=0.27 (CHCl3/MeOH 8:1). 1H NMR ([D6]DMSO): δ=11.54 (s, 1 H, N‐H), 8.10 (s, 1 H, H‐6), 6.07 (t, J=6.9 Hz, 1 H, H‐1′), 5.36 (s, 1 H, OH), 5.23–5.20 (m, 1 H, OH‐3′), 5.09–5.05 (m, 1 H, OH‐5′), 4.21–4.17 (m, 1 H, H‐3′), 3.77–3.74 (m, 1 H, H‐4′), 3.60–3.55 (m, 1 H, H‐5′), 3.54–3.50 (m, 1 H, H‐5“), 2.09–2.05 (m, 2 H, H‐2′, H‐2”), 1.37 ppm (s, 6 H, CH3); 13C NMR ([D6]DMSO): δ=161.59, 149.54, 143.28, 98.61, 98.49, 87.60, 84.70, 79.23, 73.15, 70.19, 63.71, 60.99, 31.68 ppm; IR (KBr): =3418 (br m), 2982 (m), 1706 (s), 1467 (m), 1283 (m), 1093 cm−1 (m); HRMS (ESI‐TOF): m/z: calcd for C14H19N2O6: 311.1238 [M+H]+; found: 311.1236.

2′‐Deoxy‐5‐(3‐hydroxy‐3‐methylpent‐1‐yn‐1‐yl)uridine (9 c): From 3‐methyl‐1‐pentyn‐3‐ol (1.68 mL, 14.8 mmol). Yield: 1.07 g, 3.30 mmol (56 %). R f=0.39 (CHCl3/MeOH 8:1). 1H NMR ([D6]DMSO): δ=11.56 (s, 1 H, N‐H), 8.13 (s, 1 H, H‐6), 6.10 (t, J=6.6 Hz, 1 H, H‐1′), 5.27–5.22 (m, 2 H, CH2), 5.09 (br s, 1 H, OH‐3′), 4.22 (br s, 1 H, OH‐5′), 3.78 (br s, 1 H, H‐3′), 3.62–3.58 (m, 1 H, H‐5′), 3.57–3.53 (m, 1 H, H‐5“), 2.11 (dd, J=5.8, 5.4 Hz, 1 H, H‐4′), 1.62–1.51 (m, 2 H, H‐2′, H‐2”), 1.34 (s, 3 H, CH3), 0.95 ppm (dd, J=7.4 Hz, 3 H, CH3); 13C NMR ([D6]DMSO): δ=161.57, 149.51, 143.17, 98.53, 97.47, 87.59, 84.69, 79.23, 74.29, 70.13, 67.22, 60.94, 36.47, 29.31, 9.10 ppm; IR (KBr): =3348 (br m), 2973 (m), 1668 (s), 1461 (s), 1275 (m), 1089 cm−1 (m); HRMS (ESI‐TOF): m/z: calcd for C15H21N2O6: 325.1395 [M+H]+; found: 325.1397.

2′‐Deoxy‐5‐(3‐hydroxy‐3‐phenylprop‐1‐yn‐1‐yl)uridine (9 d): From 1‐phenyl‐2‐propyn‐1‐ol (1.80 mL, 14.8 mmol). Yield: 0.972 g, 2.72 mmol (46 %). R f=0.59 (CHCl3/MeOH 8:1). 1H NMR ([D6]DMSO): δ=11.61 (s, 1 H, N‐H), 8.20 and 8.19 (2s, 1 H, H‐6), 7.53–7.40 (m, 2 H, Ph), 7.36–7.28 (m, 2 H, Ph), 7.27–7.21 (m, 1 H, Ph), 6.12–6.06 (m, 2 H, H‐1′, OH), 5.51 (d, J=6.22 Hz, 1 H, OH‐3′), 5.23 (d, J=4.48 Hz, 1 H, OH‐5′), 5.10 (t, J=4.98 Hz, 1 H, CH), 4.23–4.19 (m, 1 H, H‐3′), 3.79–3.76 (m, 1 H, H‐4′), 3.62–3.57 (m, 1 H, H‐5′), 3.57–3.52 (m, 1 H, H‐5“), 2.13–2.07 ppm (m, 2 H, H‐2′, H‐2”); 13C NMR ([D6]DMSO): δ=161.69, 149.56, 143.69, 141.97, 128.48, 128.31, 127.69, 126.76, 126.19, 98.23, 93.96, 87.73, 84.88, 79.26, 77.47, 70.25, 63.13, 61.06 ppm; IR (KBr): =3345 (br m), 3061 (m), 2824 (m), 1668 (s), 1454 (s), 1278 (m), 1088 cm−1 (m); HRMS (ESI‐TOF): m/z: calcd for C18H17N2O5: 341.1132 [M−OH]+; found: 341.1133.

2′‐Deoxy‐5‐(3‐hydroxy‐3,3‐diphenylprop‐1‐yn‐1‐yl)uridine (9 e): From 1,1‐diphenyl‐2‐propyn‐1‐ol (3.10 g, 14.8 mmol). Yield: 2.36 g, 5.43 mmol (92 %). R f=0.23 (CHCl3/MeOH 9:1). 1H NMR ([D6]DMSO): δ=11.68 (s, 1 H, N‐H), 8.29 (s, 1 H, H‐6), 7.63–7.59 (m, 4 H, Ph), 7.33–7.28 (m, 4 H, Ph), 7.23–7.19 (m, 2 H, Ph), 6.85 (s, 1 H, OH), 6.14 (t, J=6.7 Hz, 1 H, H‐1′), 5.26 (br s, 1 H, OH‐3′), 5.12 (br s, 1 H, OH‐5′), 4.26 (br s, 1 H, H‐3′), 3.83–3.80 (m, 1 H, H‐4′), 3.65–3.61 (m, 1 H, H‐5′), 3.60–3.56 (m, 1 H, H‐5“), 2.18–2.13 ppm (m, 2 H, H‐2′, H‐2”); 13C NMR ([D6]DMSO): δ=161.72, 149.51, 146.36, 143.47, 128.01, 127.07, 125.77, 98.12, 96.11, 87.72, 84.92, 79.26, 78.40, 73.28, 70.15, 60.97, 45.79, 8.74 ppm; IR (KBr): =3342 (br m), 3056 (m), 2931 (m), 1668 (s), 1449 (s), 1276 (m), 1029 cm−1 (m); HRMS (ESI‐TOF): m/z: calcd for C24H21N2O5: 417.1445 [M−OH]+; found: 417.1447.

2′‐Deoxy‐5‐(4‐hydroxybut‐1‐yn‐1‐yl)uridine (9 f): From 3‐butyn‐1‐ol (1.12 mL, 14.8 mmol). Yield: 0.943 g, 3.18 mmol (54 %). R f=0.31 (CHCl3/MeOH 8:1). 1H NMR ([D6]DMSO): δ=11.50 (s, 1 H, N‐H), 8.07 (s, 1 H, H‐6), 6.07 (t, J=6.7 Hz, 1 H, H‐1′), 5.20 (br s, 1 H, OH‐3′), 5.04 (br s, 1 H, OH‐5′), 4.81 (br s, 1 H, OH), 4.19 (br s, 1 H, H‐3′), 3.77–3.72 (m, 1 H, H‐4′), 3.59–3.46 (m, 4 H, H‐5′, H‐5“, CH2), 2.48–2.42 (m, 2 H, CH2), 2.10–2.03 ppm (m, 2 H, H‐2′, H‐2”); 13C NMR ([D6]DMSO): δ=161.89, 149.55, 143.00, 99.03, 91.08, 87.61, 84.65, 73.51, 70.29, 61.08, 59.80, 56.14, 45.84, 23.49 ppm; IR (KBr): =3333 (br m), 3060 (m), 2931 (m), 1668 (s), 1463 (s), 1277 (m), 1037 cm−1 (m); HRMS (ESI‐TOF): m/z: calcd for C13H17N2O6: 297.1082 [M+H]+; found: 297.1080.

2′‐Deoxy‐5‐(3‐methoxyprop‐1‐yn‐1‐yl)uridine (9 g):47 From methyl propargyl ether (1.25 mL, 14.8 mmol). Yield: 1.20 g, 4.05 mmol (69 %). R f=0.43 (CHCl3/MeOH 7:1). 1H NMR ([D6]DMSO): δ=11.59 (s, 1 H, N‐H), 8.22 (s, 1 H, H‐6), 6.07 (t, J=6.4 Hz, 1 H, H‐1′), 5.21 (br s, 1 H, OH‐3′), 5.08 (br s, 1 H, OH‐5′), 4.21 (s, 2 H, CH2), 4.19 (br s, 1 H, H‐3′), 3.76 (dd, J=3.6 Hz, 1 H, H‐4′), 3.61–3.56 (m, 1 H, H‐5′), 3.55–3.50 (m, 1 H, H‐5“), 3.24 (s, 3 H, OCH3), 2.11–2.07 ppm (m, 2 H, H‐2′, H‐2”); 13C NMR ([D6]DMSO): δ=161.67, 149.51, 144.13, 97.82, 88.65, 87.66, 84.85, 78.77, 70.09, 60.93, 59.63, 56.97, 40.23 ppm; IR (KBr): =3394 (br m), 2993 (m), 1703 (s), 1680 (s), 1468 (m), 1294 (m), 1080 cm−1 (m); HRMS (ESI‐TOF): m/z: calcd for C13H17N2O6: 297.1082 [M+H]+; found: 297.1080; elemental analysis calcd (%) for C13H16N2O6 (296.275): C 52.70, H 5.44, N 9.46; found: C 52.49, H 5.87, N 9.51.

2′‐Deoxy‐5‐[3‐(acetyloxy)prop‐1‐yn‐1‐yl]uridine (9 h): From propargyl acetate (1.47 mL, 14.8 mmol). Yield: 1.26 g, 3.89 mmol (66 %). R f=0.54 (CHCl3/MeOH 7:1). 1H NMR ([D6]DMSO): δ=11.62 (s, 1 H, N‐H), 8.22 (s, 1 H, H‐6), 6.06 (t, J=6.5 Hz, 1 H, H‐1′), 5.21 (br s, 1 H, OH‐3′), 5.08 (br s, 1 H, OH‐5′), 4.83 (s, 2 H, CH2), 4.21–4.16 (m, 1 H, H‐3′), 3.77–3.74 (m, 1 H, H‐4′), 3.61–3.55 (m, 1 H, H‐5′), 3.54–3.50 (m, 1 H, H‐5“), 2.11–2.06 (m, 2 H, H‐2′, H‐2”), 2.02 ppm (s, 3 H, CH3); 13C NMR ([D6]DMSO): δ=169.83, 161.50, 149.48, 144.73, 97.36, 87.70, 86.84, 84.92, 78.93, 70.13, 60.96, 52.37, 20.54 ppm; IR (KBr): =3442 (m), 3389 (m), 2987 (m), 2823 (m), 1701 (s), 1627 (s), 1467 (m), 1288 (m), 1052 cm−1 (m); HRMS (ESI‐TOF): m/z: calcd for C14H17N2O7: 325.1031 [M+H]+; found: 325.1028; elemental analysis calcd (%) for C14H16N2O7 (324.29): C 51.85, H 4.97, N 8.64; found: C 51.58, H 5.19, N 8.73.

General Procedure for the Synthesis of Dicobalt Hexacarbonyl 2′‐Deoxy‐5‐alkynyluridines 10 a–h

Under a nitrogen atmosphere, a round‐bottomed flask was charged with Co2(CO)8 (0.690 g, 2.02 mmol), 9 a–h (1.69 mmol), and THF (18 mL). The mixture was stirred at room temperature (22 °C) for 1 h. The solvent was removed by rotary evaporation. Column chromatography (230–400 mesh silica gel, 0→5 % MeOH/CHCl3) gave 10 a–h as a red‐brown compound.

Dicobalt hexacarbonyl 2′‐deoxy‐5‐(3‐hydroxyprop‐1‐yn‐1‐yl)uridine (10 a): From compound 9 a (0.477 g, 1.69 mmol). Yield: 0.557 g, 0.98 mmol (58 %). R f=0.75 (CHCl3/MeOH 7:1). 1H NMR ([D6]DMSO): δ=11.60 (s, 1 H, NH), 8.18 (s, 1 H, H‐6), 6.17 (br s, 1 H, H‐1′), 5.53 (br s, 1 H, OH‐3′), 5.24 (br s, 1 H, OH‐5′), 5.04 (br s 1 H, OH), 4.73 (br s, 2 H, CH2), 4.23 (br s, 1 H, H‐3′), 3.84 (br s, 1 H, H‐4′), 3.59–3.44 (m, 2 H, H‐5′, H‐5“), 2.16 (br s, 1 H, H‐2′), 2.02 ppm (br s, 1 H, H‐2”); 13C NMR ([D6]DMSO): δ=200.25, 161.27, 150.21, 139.69, 111.64, 103.36, 88.49, 85.52, 82.63, 79.63, 71.49, 62.52, 62.08, 40.83, 40.20 ppm; IR (KBr): =3421 (br m), 2927 (w), 2094 (s), 2054 (s), 2026 (s), 1693 (m), 1456 (m), 1276 (m), 1095 cm−1 (m); HRMS (ESI‐TOF): m/z: calcd for C18H13Co2N2O11: 550.9178 [M−OH]+; found: 550.9173; elemental analysis calcd (%) for C18H14Co2N2O12 (568.18): C 38.05, H 2.48, N 4.93; found: C 37.27, H 2.52, N 5.13.

Dicobalt hexacarbonyl 2′‐deoxy‐5‐(3‐hydroxy‐3‐methylbut‐1‐yn‐1‐yl)uridine (10 b): From compound 9 b (0.524 g, 1.69 mmol). Yield: 0.312 g, 0.52 mmol (31 %). R f=0.53 (CHCl3/MeOH 8:1). 1H NMR ([D6]DMSO): δ=11.79 (s, 1 H, NH), 8.27 (s, 1 H, H‐6), 6.20–6.15 (m, 1 H, H‐1′), 5.74 (br s, 1 H, OH), 5.25 (br s, 1 H, OH‐3′), 5.02 (br s, 1 H, OH‐5′), 4.23 (br s, 1 H, H‐3′), 3.84 (br s, 1 H, H‐4′), 3.58–3.45 (m, 2 H, H‐5′, H‐5“), 2.21–2.14 (m, 1 H, H‐2′), 2.02–1.94 (m, 1 H, H‐2”), 1.46 ppm (s, 6 H, CH3); 13C NMR ([D6]DMSO): δ=199.88, 161.55, 149.68, 140.12, 111.23, 110.50, 88.08, 85.02, 83.38, 79.24, 71.78, 71.23, 61.83, 31.84, 31.64 ppm; IR (KBr): =3383 (br m), 2975 w, 2093 (s), 2056 (s), 2019 (s), 1651 (m), 1461 (m), 1287 (m), 1101 cm−1 (m); HRMS (ESI‐TOF): m/z: calcd for C20H17Co2N2O11: 578.9491 [M−OH]+; found: 578.9494; elemental analysis calcd (%) for C20H18Co2N2O12 (596.23): C 40.29, H 3.04, N 4.70; found: C 39.90, H 3.29, N 4.29.

Dicobalt hexacarbonyl 2′‐deoxy‐5‐(3‐hydroxy‐3‐methylpent‐1‐yn‐1‐yl)uridine (10 c): From compound 9 c (0.548 g, 1.69 mmol). Yield: 0.660 g, 1.08 mmol (64 %). R f=0.78 (CHCl3/MeOH 8:1). 1H NMR ([D6]DMSO): δ=11.86 (s, 1 H, N‐H), 8.30 and 8.27 (2s, 1 H, H‐6), 6.26–6.17 (m, 1 H, H‐1′), 5.69 (d, J=5.7 Hz, 1 H, OH), 5.27 (br s, 1 H, OH‐3′), 5.06 (br s, 1 H, OH‐5′), 4.26 (br s, 1 H, H‐3′), 3.90–3.84 (m, 1 H, H‐4′), 3.61–3.55 (m, 1 H, H‐5′), 3.54–3.46 (m, 1 H, H‐5“), 2.23–2.17 (m, 1 H, CH2), 2.05–1.96 (m, 1 H, CH2), 1.74–1.61 (m, 2 H, H‐2′, H‐2”), 1.42 (s, 3 H, CH3), 0.90 ppm (d, J=7.0 Hz, 3 H, CH3); 13C NMR ([D6]DMSO): δ=199.57, 161.42, 149.33, 139.95, 111.23, 110.61, 87.82, 84.94, 83.60, 79.01, 73.44, 71.03, 61.55, 36.28, 28.98, 8.11 ppm; IR (KBr): =3353 (br m), 2979 (m), 2091 (m), 2050 (s), 2004 (vs), 1665 (s), 1459 (s), 1273 (m), 1098 cm−1 (m); HRMS (ESI‐TOF): m/z: calcd for C21H19Co2N2O11: 592.9648 [M−OH]+; found: 592.9650.

Dicobalt hexacarbonyl 2′‐deoxy‐5‐(3‐hydroxy‐3‐phenylprop‐1‐yn‐1‐yl)uridine (10 d): From compound 9 d (0.605 g, 1.69 mmol). Yield: 0.773 g, 1.20 mmol (71 %). R f=0.71 (CHCl3/MeOH 8:1). 1H NMR ([D6]DMSO): δ=11.75 (s, 1 H, N‐H), 8.17 and 8.11 (2s, 1 H, H‐6), 7.34–7.21 (m, 4 H, Ph), 7.18–7.11 (m, 1 H, Ph), 6.24–6.15 (m, 2 H, H‐1′, OH), 6.12 (br s, 1 H, OH‐3′), 5.23 (br s, 1 H, OH‐5′), 4.97 (br s, 1 H, CH), 4.27–4.17 (m, 1 H, H‐3′), 3.87–3.77 (m, 1 H, H‐4′), 3.55–3.44 (m, 1 H, H‐5′), 3.42–3.34 (m, 1 H, H‐5“), 2.19–2.12 (m, 1 H, H‐2′), 2.04–1.88 ppm (m, 1 H, H‐2”); 13C NMR ([D6]DMSO): δ=199.45, 198.93, 160.78, 149.48, 145.60, 139.22, 127.79, 127.74, 126.87, 125.16, 111.91, 111.64, 108.26, 87.70, 84.41, 83.06, 78.94, 71.88, 71.75, 61.54 ppm; IR (KBr): =3364 (br m), 3030 (br m), 2092 (m), 2052 (s), 2004 (vs), 1659 (s), 1453 (s), 1272 (m), 1034 cm−1 (m); HRMS (ESI‐TOF): m/z: calcd for C24H17Co2N2O11: 626.9491 [M−OH]+; found: 626.9489; elemental analysis calcd (%) for C24H18Co2N2O12 (644.27): C 44.74, H 2.82, N 4.35; found: C 44.35, H 2.73, N 4.44.

Dicobalt hexacarbonyl 2′‐deoxy‐5‐(3‐hydroxy‐3,3‐diphenylprop‐1‐yn‐1‐yl)uridine (10 e): From compound 9 e (0.734 g, 1.69 mmol). Yield: 0.754 g, 1.05 mmol (62 %). R f=0.53 (CHCl3/MeOH 9:1). 1H NMR ([D6]DMSO): δ=12.02 (s, 1 H, N‐H), 8.35 (s, 1 H, H‐6), 7.41–7.23 (m, 10 H, 2 Ph), 7.21 (s, 1 H, OH), 6.24 (t, J=6.7 Hz, 1 H, H‐1′), 5.29 (br s, 1 H, OH‐3′), 5.07 (br s, 1 H, OH‐5′), 4.27 (br s, 1 H, H‐3′), 3.89 (br s, 1 H, H‐4′), 3.57–3.51 (m, 1 H, H‐5′), 3.49–3.43 (m, 1 H, H‐5“), 2.29–2.22 (m, 1 H, H‐2′), 2.09–2.02 ppm (m, 1 H, H‐2”); 13C NMR ([D6]DMSO): δ=198.82, 161.80, 149.27, 146.63, 140.13, 127.75, 127.16, 127.05, 126.87, 111.61, 110.06, 87.92, 85.77, 85.04, 79.50, 78.96, 70.95, 61.38 ppm; IR (KBr): =3341 (br m), 3054 (m), 2092 (m), 2052 (s), 2007 (vs), 1661 (s), 1447 (s), 1273 (m), 1091 cm−1 (m); HRMS (ESI‐TOF): m/z: calcd for C30H21Co2N2O11: 702.9804 [M−OH]+; found: 702.9800.

Dicobalt hexacarbonyl 2′‐deoxy‐5‐(4‐hydroxybut‐1‐yn‐1‐yl)uridine (10 f): From compound 9 f (0.500 g, 1.69 mmol). Yield: 0.531 g, 0.91 mmol (54 %). R f=0.54 (CHCl3/MeOH 8:1). 1H NMR ([D6]DMSO): δ=11.56 (s, 1 H, N‐H), 8.19 (s, 1 H, H‐6), 6.22–6.14 (m, 1 H, H‐1′), 5.22 (br s, 1 H, OH‐3′), 5.04 (br s, 1 H, OH‐5′), 4.77 (br s, 1 H, H‐3′), 4.24 (br s, 1 H, H‐4′), 3.83 (br s, 1 H, OH), 3.67–3.45 (m, 4 H, H‐5′, H‐5“, CH2), 3.10 (br s, 2 H, CH2), 2.18–2.09 (m, 1 H, H‐2′), 2.06–1.97 ppm (m, 1 H, H‐2”); 13C NMR ([D6]DMSO): δ=199.35, 160.19, 149.36, 138.34, 111.06, 98.59, 87.57, 84.50, 83.80, 78.73, 70.72, 61.41, 61.23, 36.25 ppm; IR (KBr): =3358 (br m), 2945 (m), 2089 (m), 2047 (s), 1994 (vs), 1668 (s), 1455 (s), 1273 (m), 1052 cm−1 (m); HRMS (ESI‐TOF): m/z: calcd for C19H17Co2N2O12: 582.9440 [M+H]+; found: 582.9441; elemental analysis calcd (%) for C19H16Co2N2O12 (582.20): C 39.20, H 2.77, N 4.81; found: C 38.97, H 3.08, N 4.98.

Dicobalt hexacarbonyl 2′‐deoxy‐5‐(3‐methoxyprop‐1‐yn‐1‐yl)uridine (10 g): From compound 9 g (0.500 g, 1.69 mmol). Yield: 0.812 g, 1.39 mmol (83 %). R f=0.51 (CHCl3/MeOH 7:1). 1H NMR ([D6]DMSO): δ=11.59 (s, 1 H, N‐3), 8.19 (s, 1 H, H‐6), 6.17 (br s, 1 H, H‐1′), 5.24 (br s, 1 H, OH‐3′), 5.06 (br s, 1 H, OH‐5′), 4.67 (s, 2 H, CH2), 4.24 (br s, 1 H, H‐3′), 3.84 (br s, 1 H, H‐4′), 3.63–3.46 (m, 2 H, H‐5′, H‐5“), 3.39 (s, 3 H, OCH3), 2.15 (br s, 1 H, H‐2′), 2.03 ppm (br s, 1 H, H‐2”); 13C NMR ([D6]DMSO): δ=199.60, 160.73, 149.83, 139.42, 111.04, 97.01, 88.12, 85.13, 82.54, 72.63, 71.09, 61.66, 58.41 40.51 ppm; IR (KBr): =3425 (br m), 2933 (w), 2094 (s), 2053 (s), 2026 (s), 1690 (m), 1453 (m), 1275 (m), 1093 cm−1 (m); HRMS (ESI‐TOF): m/z: calcd for C19H17Co2N2O12: 582.9440 [M+H]+; found: 582.9438; elemental analysis calcd (%) for C19H16Co2N2O12 (582.20): C 39.20, H 2.77, N 4.81; found: C 39.17, H 2.43, N 5.11.

Dicobalt hexacarbonyl 2′‐deoxy‐5‐[3‐(acetyloxy)prop‐1‐yn‐1‐yl]uridine (10 h): From compound 9 h (0.548 g, 1.69 mmol). Yield: 0.856 g, 1.40 mmol (83 %). R f=0.79 (CHCl3/MeOH 7:1). 1H NMR ([D6]DMSO): δ=11.63 (s, 1 H, NH), 8.22 (s, 1 H, H‐6), 6.17 (t, J=6.7 Hz, 1 H, H‐1′), 5.47–5.35 (m, 2 H, CH2), 5.24 (br s, 1 H, OH‐3′), 5.09–5.04 (m, 1 H, OH‐5′), 4.24 (br s, 1 H, H‐3′), 3.84 (br s, 1 H, H‐4′), 3.59–3.53 (m, 1 H, H‐5′), 3.53–3.48 (m, 1 H, H‐5“), 2.20–2.14 (m, 1 H, H‐2′), 2.08–2.04 (m, 1 H, H‐2”), 2.03 ppm (s, 3 H, CH3); 13C NMR ([D6]DMSO): δ=199.22, 170.12, 160.77, 149.79, 139.83, 110.92, 94.91, 88.19, 85.30, 82.43, 79.24, 71.09, 64.92, 61.62, 20.33 ppm; IR (KBr): =3423 (br m), 2936 (w), 2095 (s), 2056 (s), 2026 (s), 1689 (m), 1454 (m), 1273 (m), 1096 cm−1 (m); HRMS (ESI‐TOF): m/z: calcd for C18H13Co2N2O11: 550.9178 [M−OAc]+; found: 550.9181; elemental analysis calcd (%) for C20H16Co2N2O13 (610.21): C 39.37, H 2.64, N 4.59; found: C 39.17, H 2.43, N 5.11.

[Bis(diphenylphosphino)methane] dicobalt tetracarbonyl 2′‐deoxy‐5‐[3‐(acetyloxy)prop‐1‐yn‐1‐yl]uridine (11): Under a nitrogen atmosphere, a 50 mL flask was charged with Co2(CO)8 (0.500 g, 1.46 mmol), toluene (20 mL), and 1,1‐bis(diphenylphosphino)methane (dppm; 0.561 g, 1.46 mmol). The mixture was stirred at room temperature for 3 h. The solvent was removed by rotary evaporation. Column chromatography (230–400 mesh silica gel, 0→2 % MeOH/CHCl3) gave Co2(CO)6(dppm)52 (0.842 g, 1.26 mmol, 86 %).

A 50 mL flask was charged with Co2(CO)6(dppm) (0.402 g, 0.600 mmol), nucleoside 9 h (0.194 g, 0.600 mmol), and THF (12 mL). The solution was stirred at room temperature (22 °C) for 24 h. The solvent was removed by rotary evaporation. Column chromatography (230–400 mesh silica gel, 0→10 % MeOH/CHCl3) gave compound 11 (0.141 g, 0.150 mmol, 25 %). R f=0.83 (CHCl3/MeOH 9:1). 1H NMR ([D6]DMSO): δ=11.42 (s, 1 H, NH), 7.88 (br s, 1 H, H‐6), 7.52–7.09 (m, 20 H, 4 Ph), 6.26–6.17 (m, 1 H, H‐1′), 5.29 (br s, 2 H, CH2), 4.91 (br s, 1 H, OH‐3′), 4.23 (br s, 1 H, OH‐5′), 3.98–3.84 (m, 2 H, CH2), 3.81 (br s, 1 H, H‐3′), 3.49 (br s, 1 H, H‐4′), 3.43 (br s, 1 H, H‐5′), 3.33 (br s, 1 H, H‐5′′), 2.23–2.16 (m, 1 H, H‐2′), 2.04–1.96 (m, 1 H, H‐2′′), 1.67 ppm (s, 3 H, CH3); 13C NMR ([D6]DMSO): δ=205.96, 203.42, 201.20, 200.76, 169.55, 160.23, 149.70, 137.20, 137.01, 136.85, 135.85, 131.46, 131.04, 129.20, 129.42, 128.07, 127.87, 116.21, 96.99, 87.40, 84.16, 80.00, 70.94, 67.08, 61.74, 32.97, 23.11, 20.07 ppm; 31P NMR ([D6]DMSO): δ=41.05 ppm; 31P NMR (CDCl3): δ=40.84 ppm; IR (KBr): =3394 (br m), 3054 (w), 2022 (s), 1990 (s), 1961 (vs), 1674 (m), 1433 (m), 1218 (m), 1092 cm−1 (m); HRMS (ESI‐TOF): m/z: calcd for C41H35Co2N2O9P2: 879.0477 [M−OAc]+; found: 879.0476.

Cells and Cytotoxicity Assay

HeLa (human cervix carcinoma) and K562 (chronic myelogenous leukemia) cells were cultured in RPMI 1640 medium supplemented with antibiotics and 10 % fetal calf serum (HeLa, K562) under a 5 % CO2/95 % air atmosphere. A total of 7×103 cells were seeded on each well of a 96‐well plate (Nunc). After 24 h, cells were exposed to the test compounds for an additional 48 h. Stock solutions of test compounds were freshly prepared in DMSO. The final concentrations of the compounds tested in the cell cultures were: 2×10−1, 1×10−1, 5×10−2, 1×10−2, 1×10−3, and 1×10−4 mm. The concentration of DMSO in the cell culture medium was 1 %.

Human umbilical vein endothelial cells (HUVEC) were acquired from Life Technologies and were cultured in Medium 200 with low serum growth supplement (Life Technologies) according to the manufacturer's instructions. A total of 10×103 cells were seeded on each well of a 96‐well plate (Nunc) in the presence of antibiotics (100 U mL−1 penicillin and 100 μg mL−1 streptomycin).

The cytotoxicity of each compound was determined by the MTT [3‐(4,5‐dimethylthiazol‐2‐yl)‐2,5‐diphenyltetrazolium bromide, Sigma] assay, as previously described.58 Briefly, after 24 or 48 h of incubation with the drug, the cells were treated with the MTT reagent and incubation was continued for 2 h. MTT–formazan crystals were dissolved in 20 % sodium dodecyl sulfate (SDS) and 50 % DMF at pH 4.7, and the absorbance was read at λ=570 and 650 nm with an ELISA‐PLATE READER (FLUOstar Omega). As control (100 % viability), cells grown in the presence of only vehicle (1 % DMSO) were used.

In a separate assay, the HeLa and K562 cells were pretreated with 1 or 2 mm N‐acetyl‐l‐cysteine (NAC), a ROS inhibitor, for 30 min before completing the above steps, following the previously described procedure.59

The IC50 values (the concentration of test compound required to reduce the cell survival fraction to 50 % of the control) were calculated from dose–response curves and were used as a measure of cellular sensitivity to a given treatment.

Intracellular ROS Measurement in Living Cells

Intracellular ROS levels were assessed with a 2,7‐dichlorofluorescein diacetate (DCFDA) dye assay. DCFDA (Sigma–Aldrich) is a cell‐permeant reagent, a fluorogenic dye that measures the activity of hydroxyl, peroxyl, and other ROS within the cell. After diffusion into the cell, DCFDA was deacetylated by cellular esterases to a nonfluorescent compound, which was later oxidized by the ROS into 2′,7′‐dichlorofluorescein (DCF). DCF is a highly fluorescent compound that can be detected by fluorescence spectroscopy with maximum excitation and emission bands at λ=485 and 520 nm, respectively.

K562 (chronic myelogenous leukemia) cells were cultured in RPMI 1640 medium supplemented with antibiotics and 10 % fetal calf serum under a 5 % CO2/95 % air atmosphere. Next, 10×104 cells per well were stained with 20 μm DCFDA in the culture media and were then incubated for 30 min at 37 °C. After staining, the cells were centrifuged and suspended in complete medium without phenol red, and 10×104 cells were seeded on each well of a 96‐well plate (PerkinElmer). Next, the cells were incubated for 4 h with the test compounds at concentrations of 5, 10, 20, 50, 100, and 200 μm. Stock solutions of the test compounds were freshly prepared in DMSO. The concentration of DMSO in the cell culture medium was 1 %. Cells treated with 50 and 100 μm H2O2 served as positive controls. After incubation, fluorescence in each well was measured (excitation at λ=485 nm, emission measured at λ=520 nm) by using a microplate reader FLUOStar Omega (BMG‐Labtech, Germany). The increase in the fluorescence intensity of DCF was a sign of an increase in ROS levels. For normalization of the data, the DCF fluorescence intensity level in the control cells (exposed to 1 % DMSO and 20 μm DCFDA) was taken as 1.0.

Computational Details

Quantum‐chemical calculations providing molecular orbitals (HOMO–LUMO), energy‐minimized molecular geometries, and vibrational spectra of 10 f were obtained by using density functional theory (DFT) as implemented in the Gaussian 09 (Revision C.01) program package.60 We utilized the PBE0 hybrid functional61 and 6‐31G* basis set.62 Complete ground‐state geometry optimization was afforded without symmetry constraints. Only default convergence criteria were used during geometry optimization. Initial atomic coordinates were imported from the crystal structure. Optimized structures in the singlet state were indicated to be local minima (no imaginary frequencies). Selected theoretical and experimental metric parameters are provided in Figure 2. Molecular orbitals were acquired by using Avogadro (an open‐source molecular builder and visualization tool, Version 1.1.0).63

As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re‐organized for online delivery, but are not copy‐edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.

Supplementary

Click here for additional data file.