Anticancer (hexacarbonyldicobalt)propargyl aryl ethers: Synthesis, antiproliferative activity, apoptosis induction, and effect on cellular oxidative stress

Article abstract of DOI:10.1016/j.jinorgbio.2012.10.014

While an increasing number of (hexacarbonyldicobalt)alkynes have been found to possess antiproliferative activity against a number of cancer cell lines, the role of the organometallic moiety in this bioactivity is not well understood. To gain a better und

Full text of DOI:10.1016/j.jinorgbio.2012.10.014

Journal of Inorganic Biochemistry 119 (2013) 28–37
Contents lists available at SciVerse ScienceDirect
Journal of Inorganic Biochemistry
journal homepage: www.elsevier.com/locate/jinorgbio
Anticancer (hexacarbonyldicobalt)propargyl aryl ethers: Synthesis, antiproliferative
activity, apoptosis induction, and effect on cellular oxidative stress
1
⁎
Sydonie D. Schimler , David J. Hall, Stefan L. Debbert
Department of Chemistry, Lawrence University, 711 E. Boldt Way, Appleton, WI 54911, United States
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 25 June 2012
Received in revised form 30 October 2012
Accepted 30 October 2012
Available online 7 November 2012
While an increasing number of (hexacarbonyldicobalt)alkynes have been found to possess antiproliferative
activity against a number of cancer cell lines, the role of the organometallic moiety in this bioactivity is not
well understood. To gain a better understanding of cobalt's role in the medicinal chemistry of these com-
pounds, several simplified analogs of a known organocobalt anticancer compound were synthesized and
assessed for antiproliferative activity against MDA-MB-231 human breast cancer cells. These compounds,
mostly (hexacarbonyldicobalt)propargyl aryl ethers, caused 45–93% growth inhibition of that cell line at
40 μM in a 72 h crystal violet staining assay. The most active analog was the organocobalt nitroaromatic
ether 3a, with an IC50 of 3.3 0.9 μM. Flow cytometric assays on the same cell line demonstrated that 3a
strongly induces apoptosis, arrests the cell cycle at the S phase, increases cellular oxidative stress levels,
and induces permeability of the mitochondrial membrane. While the non-cobalt-containing precursor to
3a also caused an increase in mitochondrial membrane permeability, it did not produce an increase in oxida-
tive stress levels, nor did it have apoptosis-inducing or antiproliferative effects. The induction of oxidative
stress in the cell may be responsible for some of the antiproliferative activity of compound 3a against this
cell line.
Keywords:
Organometallic medicinal chemistry
Oxidative stress
© 2012 Elsevier Inc. All rights reserved.
1. Introduction
owe their activity to the inhibition of kinases, proteases, esterases,
and other enzymes [17]. Still others affect cell proliferation by modu-
The success of platinum-based chemotherapy agents such as
cisplatin has led some medicinal chemists to turn to organometallic com-
pounds in the search for novel, effective anticancer drugs [1–4]. Metals
employed in these anticancer compounds include titanium [5], iron [6],
ruthenium [7,8], osmium [9], gold [10,11] and cobalt [12,13]. Recently,
an aspirin derivative containing a (hexacarbonyldicobalt)alkyne moiety
(1, Fig. 1) was shown to have notable anticancer activity in two breast
cancer cell lines (MCF-7 and MDA-MB-231) [14]. It also exhibited
anti-angiogenic effects in a zebrafish embryo assay, as well as significant
inhibition of matrix metalloprotein-7 (MMP-7) [15]. Among its many
lating cellular oxidative stress, either by releasing carbon monoxide
or by generating reactive oxygen species (ROS) such as hydrogen per-
oxide and superoxide. In some anticancer ferrocene-phenol conju-
gates, it is believed that the organometallic moiety generates ROS,
which then in turn converts the phenol ligand into a reactive quinone
methide; this intriguing effect has been termed kronatropism, after
Kronos, the Titan of myth who devoured his own offspring [18].
Much of the anticancer activity of compound 1 has been attributed
to its ability to efficiently inhibit the cyclooxygenase-2 enzyme
(COX-2) [14,15]. However, analogs of 1 with less COX-inhibitory effi-
cacy, such as the benzoic acid derivative 2, were also found to exhibit
moderate anticancer activity. To probe the mode of action of these
organocobalt anticancer agents further, we synthesized a set of
simplified analogs of 1, assessed their antiproliferative activity, and
investigated their mode of action using flow cytometry.
other demonstrated biological effects was
a strong induction of
caspase-3, which is an important activator in apoptosis. Several other an-
ticancer (hexacarbonyldicobalt)alkyne compounds have also been
reported [16]; in all cases, the bioactivities of these compounds have
been shown to depend on both the hexacarbonyldicobalt moiety and
the organic ligand.
The modes of action observed in bioactive organometallic com-
pounds are as varied as their structures. Although cisplatin and its
analogs target cellular DNA, other organometallics are reported to
2. Results
2.1. Synthesis, characterization and assessment of antiproliferative activity
of (hexacarbonyldicobalt)alkyne compounds
⁎
Corresponding author. Tel.: +1 920 832 6939; fax: +1 920 832 6962.
E-mail address: stefan.debbert@lawrence.edu (S.L. Debbert).
Present address: Department of Chemistry, The University of Michigan, 930 North
In our initial studies, we synthesized a set of (hexacarbonyldicobalt)
propargyl aryl ethers for anticancer testing (Fig. 1). We varied only the
substituent para to the propargyloxy unit, in order to determine
1
University Avenue, Ann Arbor, MI 48109-1055, United States.
0162-0134/$ – see front matter © 2012 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.jinorgbio.2012.10.014

S.D. Schimler et al. / Journal of Inorganic Biochemistry 119 (2013) 28–37
29
tetrahydrofuran (THF) at room temperature to arrive at compounds
3a–j. These compounds are relatively air stable, showing only minimal
decomposition in air over several days.
The structures were confirmed by 1H and 13C NMR spectroscopy,
IR spectroscopy, and mass spectrometry. Upon complexing to the
hexacarbonyldicobalt fragment, the methylene and methine protons
are shifted ~1 and ~3.5 ppm downfield, respectively, in the 1H NMR
spectra. The 13C NMR spectrum of 3a shows the expected influence
of the nitro group on C\O bond polarization; its propargyl methylene
is found at a higher chemical shift than the corresponding carbons in
its analogs, 3b–3j. Minor decomposition occasionally led to poor
peakshapes in NMR spectra unless the compounds were freshly fil-
tered through a Celite plug prior to analysis. The compounds broke
down within hours of dissolving in DMSO, the solutions progressing
from a homogeneous brick-red to a heterogeneous tan-brown. Com-
pound breakdown in N,N-dimethylformamide (DMF) was somewhat
slower, but still observed in solutions stored in a −20 °C freezer for
several days.
Compounds 3a–j, 1, 4a and para-nitrophenol (PNP) were then
assessed for antiproliferative activity against the MDA-MB-231 cell
line, a hormone-independent human breast cancer line, using a crystal
violet assay (Table 1) [14]. Stock solutions of the test compounds in
DMF (100 mM) were prepared freshly immediately prior to testing.
Initially, cells were seeded in a 96 well plate at 6000 cells/well, then
let adhere overnight. The cells were then dosed with 40 μM test
compound, and cell survival relative to control was assessed after
72 h (Table 1). The compounds with the most powerful electron-
withdrawing groups (3a–c) were found to be the most cytotoxic;
Fig. 1. The organocobalt aspirin derivative 1, and analogs thereof.
whether the bioactivity of these compounds depends strongly on the
electron-withdrawing or -donating nature of the organic ligand. Com-
pounds 3a–j were each obtained in two synthetic steps following pub-
lished procedures (Fig. 2) [14,19,20]. Ethers 4a–j, obtained by alkylation
of the respective phenols, were treated with dicobalt octacarbonyl in
however, compound 3j
– with a strongly donating methoxy
substituent – also showed good cytotoxicity at that dose. The most
effective compound in this group was the nitroether 3a, which
caused 91% cell death relative to control after 72 h. This toxicity
is comparable to that found for the organocobalt aspirin derivative
1 [14]. Notably, neither PNP nor 4a, the cobalt-free precursor to 3a,
showed any notable antiproliferative effect in this assay. (Hexa-
carbonyldicobalt)propargyl alcohol (5) itself inhibited cell growth
by only about 32% relative to control in this assay.
We obtained an IC50 value for our lead compound, 3a, by assaying
its cytotoxicity using the same assay over a range of dosages (40 μM–
160 nM) with five duplicate measurements. Fitting the data with a
standard four-parameter model [21], we calculated an IC50 of 3.3
0.9 μM (Fig. 3); this is comparable to the published IC50 of compound
1 against the same cell line (1.9 0.3 μM), and lower than the pub-
lished value for compound 2 (11.3 0.9 μM) [14].
Analogs of compound 3a were then synthesized to further probe
the effect of the ring's electron distribution on cytotoxicity (Fig. 2).
Three analogs, 6a–c, were obtained from nucleophilic aromatic sub-
stitution of para-fluoronitrobenzene with 3-butyn-1-ol, 2-methyl-3-
butyn-2-ol, and propargylamine, respectively, followed by complexa-
tion with dicobalt octacarbonyl [22–24]. In a separate synthesis,
para-nitrobenzyl alcohol was deprotonated by sodium hydride in
THF, then alkylated with propargyl bromide to give the benzyl ether;
complexation of dicobalt octacarbonyl with this ether gave compound
7. We selected these analogs to assess whether the polarization of
the ether O\CH2 bond (or, in the extreme case, its heterolytic cleav-
age) is important to the bioactivity of 3a. We predicted that this
polarization would be lessened in 6a due to the greater distance sep-
arating the putative cation from the cobalt center; the ability of a
hexacarbonyldicobalt moiety to stabilize a positive charge on an
adjacent carbon is well known to organic chemists (cf. the Nicholas
reaction) [25,26]. Analogs 6c and 7 lack a para-nitrophenolate leav-
ing group, which should also lessen the O–CH2 (or NH–CH2) polariza-
tion. Alternatively, the dimethyl analog 6b would produce a more
stable cation upon heterolytic C–O cleavage than 3a, which could
lead it to be a more cytotoxic compound if that cleavage is relevant
to the mode of action of these ethers.
Fig. 2. Synthesis of organocobalt compounds 3a–3j, 6a–6c, and 7. Reagents and condi-
tions. (a) K2CO3, propargyl bromide, DMF; (b) Co2(CO)8 (excess), THF. (c) Nucleophile
(alcohol or amine), NaH, DMF. (d) NaH, then propargyl bromide, THF.

30
S.D. Schimler et al. / Journal of Inorganic Biochemistry 119 (2013) 28–37
Table 1
Cell survival of MDA-MB-231 cells treated with organocobalt compounds, relative to vehicle control.
Cell survival
(relative to control)
Co2(CO)6
Y
X
X
Y
40 μM
5 μM
1
0.112 0.023
0.088 0.015
0.116 0.022
0.115 0.015
0.138 0.013
0.143 0.018
0.250 0.048
0.371 0.050
0.553 0.127
0.218 0.036
0.125 0.007
0.994 0.118
0.681 0.138
0.271 0.032
0.411 0.127
0.684 0.127
0.706 0.128
3a
3b
3c
3d
3e
3f
3g
3h
3i
–OCH2−
NO2
CN
CO2Et
Br
Cl
F
"
"
"
"
"
"
"
"
"
H
NHAc
CH3
OCH3
NO2
3j
4a
5
0.660 0.173
(3a w/o Co2(CO)6)
HOCH2−
6a
6b
6c
7
–OC(CH3)2−
–OCH2CH2−
–NHCH2−
NO2
NO2
NO2
NO2
1.244 0.120
0.276 0.145
0.578 0.070
0.455 0.075
–CH2OCH2−
These analogs of 3a, along with those compounds causing over
80% cell death relative to control in the first assay, were then tested
for cytotoxicity at 5 μM in a 72 h assay using crystal violet staining.
At that dose, compound 3a was found to be only slightly less cytotoxic
than the previously studied aspirin derivative 1. Compound 6b showed
no cytotoxicity at that dose, likely due to the steric influence of the
two additional methyl groups. The nitroaniline analog 6c and the
nitrobenzyl ether 7 caused moderate levels of cell death, but were less
active than the lead compound 3a. However, analog 6a, with an extra
methylene between the oxygen atom and the cobalt moiety, was
found to have an even greater cytotoxic effect. This may indicate that
O–CH2 bond polarization due to the cobalt moiety, which would be
lessened in 6a relative to 3a, is not a significant factor in the cytotoxicity
of these compounds.
significantly more or less toxic than the intact compounds. To de-
termine if a relationship exists between compound stability and
antiproliferative activity, the stabilities of some of these compounds
in aqueous buffer were assessed. As the loss of a CO ligand is likely
the first degradation step, we used the extent to which the com-
pounds converted deoxy-myoglobin (deoxy-Mb) to carbonmonoxy-
myoglobin (CO-Mb) after 2 h in buffered aqueous solution [27,28]
as a proxy for the degradation of the compounds. Briefly, a 100 mM
stock solution of a test compound in DMF was freshly prepared,
then added to a solution of myoglobin in phosphate-buffered saline
(PBS, 0.01 M) that had been deoxygenated by the addition of
sodium dithionite (0.1%). The conversion of deoxy-Mb to CO-Mb
was then monitored by UV-visible spectroscopy. Previous work [27]
done using DMSO stock solutions found that similar compounds re-
leased one CO ligand per molecule over that time period, and that
(hexacarbonyldicobalt) propargyl alcohol (5) emitted CO relatively
slowly (about 25% emission after 1 h).
2.2. Assessment of compounds' stability and carbon monoxide emission
in aqueous buffer
We observed even slower CO emission from our compounds; a
40 μM solution of compound 3a produced only 12 μM CO-Mb in
2 h, suggesting that no more than 30% of the compound released a
CO ligand during degradation during that time. Compounds 3b and
3f emitted CO more slowly in the same assay, producing 6.4 μM and
6.8 μM concentrations of CO-Mb, respectively. Other compounds test-
ed (3d, 3h, 3i and 3j) produced less than 4 μM CO-Mb in the assay.
These results suggest that the compounds should be relatively stable
in aqueous buffer for at least 2 h, and that the differences in
antiproliferative activities among the compounds may not be primar-
ily due to their different rates of breakdown in aqueous buffer. The
results do not, however, rule out the possibility that the release of a
CO ligand from 3a contributes in some way to that compound's bioac-
tivity; earlier work has suggested that very slow release of CO may
allow similar compounds to inhibit cellular production of NO [27].
One explanation for the differences in toxicity among these
compounds could involve their relative propensities to break down
in aqueous buffer; the degradation products of these compounds,
such as cobalt ions and carbon monoxide (CO), may themselves be
2.3. Studies of the mode of action of 3a and analogs on MDA-MB-231
cells by flow cytometry
2.3.1. Apoptosis induction
To determine whether the cytotoxicity of these compounds in-
volved the induction of apoptosis, MDA-MB-231 cells were treated
with test compound, stained with both propidium iodide (PI) and
Annexin V-fluorescein isothiocyanate (FITC), and analyzed by flow
cytometry. During the process of apoptosis, cell membrane asymme-
try is lost, and phosphatidylserine becomes accessible on the cell
Fig. 3. Effect of 3a on cell survival, at varying doses. MDA-MB-231 cells were dosed
with the test compound, and cell survival was measured after 72 h by a crystal violet
assay and compared to control.

S.D. Schimler et al. / Journal of Inorganic Biochemistry 119 (2013) 28–37
31
Fig. 4. Apoptosis induction by selected test compounds, as assessed by flow cytometry with Annexin V/PI staining.
surface, where it can bind to FITC-tagged anticoagulant Annexin V.
PI stains late-apoptotic or necrotic cells whose membrane integrity
has been compromised. The orthogonality of these dyes' binding and
fluorescence allow for the simultaneous observation of early apoptosis
and late apoptosis/necrosis [29].
In our assay, MDA-MB-231 cells were seeded in 12-well plates
(4×104 cells in 1 mL media), let attach for 24 h, dosed with a select-
ed test compound at 20 μM, and let incubate with the test compound
for another 18 h at 37 °C before treating with both PI and Annexin
V-FITC. The results of this assay are shown in Fig. 4; cells that
exhibited low fluorescence with either probe were labeled ‘viable,’
cells stained by Annexin V-FITC only were labeled ‘early apoptotic’,
cells stained by both PI and Annexin V-FITC were labeled ‘late apoptotic’,
and cells with low Annexin V-FITC staining but high PI staining were
labeled ‘necrotic.’
Cells dosed with 3a, the most cytotoxic compound in the set,
exhibited nearly five times more Annexin V-FITC staining than control
cells (which were dosed with 0.2% DMF), and slightly more Annexin
V-FITC staining than compound 1. As in our preliminary cytotoxicity
assay, the compounds that induced apoptosis to the greatest extent
were those compounds with strongly electron-withdrawing substitu-
ents on the aromatic ring, the nitro-containing compound 3a and the
nitrile 3b. Other cobalt-containing aryl ethers induced only moderate
levels of apoptosis, and precursors to 3a that lack cobalt moieties
entirely (4a and PNP) induced only minimal levels of apoptosis com-
pared to control. The percentage of necrotic cells was low in each
sample.
partially arrested the cell cycle after DNA duplication but before cell
division. Treatment with the cobalt-free precursor to 3a, 4a, also led
to a decrease in the G2/M fraction (17.1%), but the effect was much
less pronounced than that caused by 3a. Since DNA synthesis brings
cells from the G0/G1 phase into the S phase, this result indicates that
3a, unlike cisplatin, does not target DNA as part of its anticancer
activity.
2.3.3. Effects of 3a and analogs on cellular oxidative stress
Oxidative stress brought on by overproduction of ROS has been
shown to induce apoptosis [30,31]. Organocobalt complexes such as
3a may be contributing to this oxidative stress in a number of ways.
They may be effecting mitochondrial membrane permeability (MMP),
which would lead to increased ROS production in the cytosol, and even-
tually apoptosis. Alternatively, the organocobalt compound itself (or a
product of its breakdown, such as cobalt ions or nanoparticles) could
be directly mediating ROS production by reducing oxygen to superox-
ide and/or generating hydroxyl radicals via Fenton-like reactions [32].
2′,7′-Dichlorodihydrofluorescein diacetate (H2DCFDA) is a com-
monly used probe for oxidative stress [33,34]. H2DCFDA itself is
non-fluorescent, and is converted to 2′,7′-dichlorodihydrofluorescein
(H2DCF) once in the cell by esterases. There, H2DCF can be oxidized to
highly fluorescent 2′,7′-dichlorofluorescein (DCF); the increase in cel-
lular fluorescence, observed by cytometry, is often interpreted as an
increase in intracellular ROS or RNS (reactive nitrogen species) levels.
In order to probe the effect of 3a on the oxidative stress of
MDA-MB-231 cells, cells were incubated with 25 μM 2′,7′-
dichlorodihydrofluorescein diacetate (H2DCFDA), treated with 3a or
one of its analogs, and analyzed by flow cytometry. Cells treated with
10 μM 3a showed increased levels of DCF fluorescence compared
to control (4.06 in the treated cells vs. 3.41 in the untreated cells)
(Fig. 5). This effect was observed to be dose dependent, as an increase
in 3a dosing from 10 μM to 40 μM was accompanied by a rise in DCF
fluorescence levels. Compound 4a, the cobalt-free precursor to 3a, did
not elicit this effect — in fact, DCF fluorescence was observed to drop
below control levels in 4a-treated cells, with a geometric mean of fluo-
rescence of 2.35; the significance of this result is unclear, but perhaps
the methylene CH2 of 4a was oxidized in the cell, depleting the cell of
some endogenous oxidant. When cells were treated with both 40 μM
3a and 50 μM N-acetylcysteine (NAC), a known antioxidant, DCF fluo-
rescence again dropped below control levels. This is expected because
NAC scavenges ROS in an attempt to reduce the oxidative stress on
the cell. In our positive control experiment, treatment with 10 mM
H2O2 produced the expected large increase in DCF fluorescence. These
results suggest that the ability of 3a to strongly induce apoptosis in
MDA-MB-231 cells may be due to its effect on the oxidation state of
the cell.
2.3.2. Cell cycle analysis
In order to determine the effect of the cobalt compounds on the
cell cycle, MDA-MB-231 cells were treated with 20 μM 3a for 18 h,
and then the DNA was stained with PI. In comparing 3a-treated
cells with our control cells, we observed little change in the G0/G1
fraction of the populations (59.5% vs. 58.3% for control), but a marked
difference in the relative amounts of the G2/M and S fractions be-
tween the two populations (Table 2). The 3a-treated cells exhibited
a larger S fraction (30%) and a smaller G2/M fraction (10.5%) than
the control cells (18.5% and 23.2%, respectively), indicating that 3a
Table 2
Cell cycle analysis. The average percentages of three readings are shown.
Compound
G0/G1
S
G2/M
Blank
3a (20 μM)
4a (20 μM)
58.34 0.03
59.53 0.02
63.68 0.01
18.51 0.003
29.97 0.04
19.19 0.02
23.15 0.03
10.50 0.05
17.11 0.03

32
S.D. Schimler et al. / Journal of Inorganic Biochemistry 119 (2013) 28–37
Fig. 5. DCF fluorescence measured, by flow cytometry, in cells doped with H2DCFDA; the fluorescence observed is proportional to cellular oxidative stress. The fluorescence is given
as the mean fluorescence intensity (MFI).
2.3.4. Effects of 3a and analogs on mitochondrial membrane permeability
During the process of apoptosis, the mitochondrial membrane po-
tential is disrupted through the opening of permeability transition
pores, which leads to elevated oxidative stress [29,35–37]. Nitrophenols
are well known to effect this depolarization of the mitochondrial mem-
brane, which uncouples oxidative phosphorylation [38]. In order to
determine if the mitochondrial membrane is disrupted by incubation
with 3a, MDA-MB-231 cells were treated with a test compound or
vehicle control for 18 h, then incubated with rhodamine 123 (R123), a
fluorescent lipophilic cation that accumulates in mitochondrial propor-
tionally to the electrical potential across the inner mitochondrial mem-
brane (ΔΨ) [39]. Cells in which this membrane has becomes permeable
show reduced fluorescence due to R123, because of the decrease in ΔΨ.
The compounds assayed were 3a, 4a, and anisomycin (included as a
positive control).
solely due to the depolarization of the mitochondrial membrane,
since oxidative stress is not increased by treatment with 4a, whose
effect on the mitochondrial membrane is similar to that of 3a.
3. Discussion
Compound 3a, a organocobalt derivative of para-nitrophenol, was
found to have antiproliferative activity comparable to 1, a cytotoxic
organocobalt aspirin derivative previously described in the literature
[14], and cisplatin, a commonly used inorganic anticancer agent,
on MDA-MB-231 cells. This cytotoxicity was found to be dependent
on both the presence of the cobalt and the ligand; without the
Table 3
Flow cytometry data for the mitochondrial transmembrane potential of MDA-MB-231
cells using rhodamine 123. Values were obtained from the average of three measure-
ments, using 20 μM dosing.
We observed a decrease in the fluorescence corresponding to
R123 compared to control cells upon incubation with either 3a or
4a; the effects induced by the two compounds are the same to within
experimental error (Table 3). Anisomycin, our positive control, in-
duced greater depolarization than either 3a or 4a. Similar results
were observed when tetramethylrhodamine methyl ester (TMRM)
was used as the dye instead of R123. These results indicate that the
increase in cellular oxidative stress upon treatment with 3a is not
Compound
Mean R123 fluorescence intensity
Control
3a
Anisomycin
4a
7.63 0.47
5.17 0.10
4.39 0.14
4.99 0.06

S.D. Schimler et al. / Journal of Inorganic Biochemistry 119 (2013) 28–37
33
coordination of cobalt or the nitro substituent, the antiproliferative
activity was significantly decreased. Furthermore, most modifications
to this compound led to a decrease in antiproliferative activity.
A number of studies were conducted to elucidate the possible
modes of action of these compounds. One possibility is that the bioac-
tivity of these (hexacarbonyldicobalt)alkynes is primarily due to the
products of their degradation in aqueous solution, including CO, an
important mammalian signaling molecule [40–43], and cobalt ions.
Several compounds, including (hexacarbonyldicobalt)alkynes such
as 5, have been studied as CO-releasing molecules (CORMs), and
have been shown to inhibit both cell viability and cellular nitric
oxide (NO) production [27]. A myoglobin assay indicated that the
compounds described herein release CO only slowly in aqueous buffer;
3a released only about 30 mol% of CO after 2 h, and the other analogs
studied released considerably less. As in earlier work [27], it appears
that initial CO-release rates are not strongly correlated with cytotoxicity.
However, the very slow release of CO from these compounds may yet
have a hidden contribution to the compounds' bioactivity.
To determine whether these hexacarbonyldicobalt compounds
induced apoptosis, a flow cytometric assay was done using Annexin
V-FITC and PI staining. The results from this experiment indicate
that incubation with 3a led to an increased number of apoptotic
cells compared to untreated cells, without significantly increasing the
percentage of necrotic cells. Both the ligand and the cobalt complex
are necessary to induce apoptosis in breast cancer cells. The parent
compound of 3a, 4a, did not induce apoptosis. These results, coupled
with those of the cytotoxicity assays, suggest that 3a may exert its
antiproliferative activities through the induction of apoptosis in breast
cancer cells.
Studies on the cell cycle progression of cancer cells treated with 3a
also shed light on the apoptotic events that are induced by this com-
pound. Upon incubation with 3a, there is no significant increase in the
percentage of cells in the nondividing phase G0/G1, but there is a
significant difference observed in the percentage of cells in both the
S phase and M phase. The percentage of cells in the mitotic phase is
significantly decreased upon incubation with 3a, indicating disrup-
tion of the process of cellular division. Compound 3a does not, how-
ever, arrest the cell in G0/G1 phase, indicating that it does not target
DNA. Cells incubated with the parent compound of 3a, 4a, did not
exhibit significant differences in the number of cells in S and M
phase. This indicates that the (hexacarbonyldicobalt)alkyne moiety
is necessary to exert changes on the cell cycle in these phases.
If cells are undergoing a significant amount of DNA fragmentation,
there will be a visible sub-G1 peak corresponding to a lower concen-
tration of DNA than nondividing cells. While we observed no signifi-
cant increase in the percentage of cells in G0/G1 phase, we did see a
shift in the peak intensity of this region to lower intensity. This indi-
cates that there is a lower average percentage of DNA in the cells that
are in nondividing phases. This further suggests that this compound is
inducing apoptosis in cancer cells. Furthermore, the shift in average
intensity is not observed in cells incubated with 4a, which did not
induce apoptosis at the dose tested.
Many factors could contribute to this induction of apoptosis; one
important factor may be an increase in cellular oxidative stress [31].
A cytofluorometric assay using H2DCFDA as a fluorescent probe indi-
cated that 3a increases oxidative stress in a dose-dependent manner.
The cobalt-alkyne moiety was necessary to cause this effect, as no in-
crease in DCF fluorescence was seen in cells treated with 4a. Although
H2DCFDA is commonly used as a probe for ROS (and, more specifically,
H2O2), its specificity is a matter of some debate [44]. It has been noted
that H2DCF is not directly oxidized to DCF in the cytosol by either
H2O2 or superoxide (O2−•), but rather by mitochondrial cytochrome c
and/or Fenton and Fenton-like reactions mediated by transition metals
such as iron. Since cobalt is known to mediate Fenton-like reactions in
the cell [32], 3a may be causing the increased DCF fluorescence seen
in the assay either by increasing intracellular ROS levels or by simply
oxidizing H2DCF directly. In either case, 3a appears to act as an oxidant
in the cell, and its contribution to oxidative stress may play a significant
role in its induction of apoptosis in this cell line.
One of the great advantages of metal-containing drugs is their
potential effect on the oxidation state of the cell. Ferroquine, a
ferrocene-containing derivative of the antimalarial drug chloroquine,
is believed to cause the death of Plasmodium falciparium parasites
by both inhibiting hemozoin formation and generating ROS inside
the digestive vacuole via a one-electron redox reaction [6,17,45,46].
Ferrocifen, an organoiron derivative of the front-line breast cancer
drug tamoxifen, has also been found to induce oxidative stress in
MDA-MB-231 cells [47,48]. Some organic drugs are also thought to
derive their anticancer activity from their ability to induce elevated
ROS levels in the cell, including elesclomol, a potential melanoma treat-
ment [49], and the anti-androgenic nitroaromatic drugs flutamide and
nilutamide [50]. Cobalt and its ions are known to react with oxygen in
water to generate ROS species [32], so the bioactivity of the organocobalt
compounds described herein may be due to the cobalt atoms them-
selves, whose entrance into the cell is mediated by the ligand.
One potential source of oxidative stress is mitochondrial mem-
brane permeability [36], which may allow oxidants such as cyto-
chrome c and ROS to leak into the cytosol. A flow cytometric assay
using R123 was used to probe the mitochondrial membrane poten-
tial; the results of this assay indicate that the mitochondrial mem-
brane is compromised by treatment with 3a. However, there is also
a decrease in mitochondrial transmembrane potential observed in
cells that were incubated with 4a. This implies that the nitroaromatic
ligand but not the (hexacarbonyldicobalt)alkyne moiety is necessary for
effecting mitochondrial membrane permeability; since nitrophenols
are well known to depolarize mitochondrial membranes, this is perhaps
unsurprising. However, treatment with 4a did not lead to increased
oxidative stress levels in our H2DCFDA assay, implying that the increase
in these levels caused by treatment with 3a is not necessarily associated
with the permeability of the mitochondrial membrane.
4. Conclusions
In conclusion, we report herein the development of a new series of
anticancer organocobalt compounds. The most active of this series,
the nitroaromatic ether 3a, has been found to induce apoptosis and
disrupt the cell cycle at the S phase. The anticancer activity of 3a
may be due to its ability to induce elevated levels of oxidative stress
in the cell. Further work will examine the cellular uptake of 3a and
its bioactive analogs. We will also elaborate the effects of these com-
pounds on the oxidation state of the cell, including their effects on
cellular levels of NO and glutathione [51]. More broadly, we will
seek to develop similar anticancer compounds and elaborate their
modes of action.
5. Experimental
Commercially available chemicals and solvents were used without
further purification, with the exception of DMF, which was dried over
4 Å molecular sieves, and THF, which was purified using an SPBT-1
benchtop solvent purification system from LC Technology Solutions,
Inc. (Seabrook, NH). MDA-MB-231 cells were purchased from ATCC
(Manassas, VA). Myoglobin (lyophilized equine heart, Sigma), H2DCFDA
(Cayman Chemical), anisomycin (Sigma), PI (Biotium), FITC-Annexin V
(Biotium), rhodamine 123 (Aldrich), and RPMI and HBSS media (Cellgro)
were used as received. Products were purified by column chromatog-
raphy on either silica gel (32–63 μM, Dynamic Absorbents, Inc.,
Norcross, GA) or alumina (Aldrich). Melting points were taken on
either an Electrothermal 9100 digital melting point apparatus or a
simple Mel-Temp, and are uncorrected. IR spectra were acquired on a
Thermo Scientific Nicolet iS10 instrument fitted with a Specac attenuated
total reflectance (ATR) accessory. UV–visible spectroscopy was carried

34
S.D. Schimler et al. / Journal of Inorganic Biochemistry 119 (2013) 28–37
out with an Ocean Optics USB2000. All NMR spectra were taken on a
Varian 400-MR instrument operating at 400 MHz (1H) or 100 MHz
(
1026, 821. 1H NMR (CDCl3): 7.40 (dd, 2H, J=2.4 Hz, 8.0 Hz), 6.87
(dd, 2H, J=2.4 Hz, 8.0 Hz), 4.67 (d, 2H, J=2.4 Hz), 2.52 (t, 1H, J=
2.4 Hz). 13C NMR (CDCl3): 156.6, 132.3, 116.8, 113.9, 78.1, 75.9,
56.0. MS (EI): 210/212 (M+•), 171/173, 143/145, 131 (base peak),
103, 77, 63, 39.
13C), using chloroform-d (Aldrich) as the solvent and tetramethylsilane
as an internal standard. Gas chromatography was carried out on a
Hewlett-Packard 6890 instrument that was connected to a HP5973
mass selective detector. Mass spectra of our cobalt-containing com-
pounds were obtained by the Mass Spectrometry Lab of the University
of Illinois-Urbana-Champaign, on either a Waters 70VSE (electron impact
(EI)) or a Waters Q-Tof Ultima (electrospray ionization (ESI)) instrument.
Spectrophotometric assays were done using a Dynex LT plate reader. Flow
cytometric assays were carried out with a Benton-Dickinson FACSCalibur
instrument, and histogram statistics were analyzed by the program
CellQuest (Becton Dickinson, San Jose, CA). Compounds 1 and 5 were
prepared as described in the literature [14].
5.1.1.7. Propargyl para-chlorophenyl ether (4e). Method B. Yield: 43%
(clear oil). IR (ATR, cm−1): 3295, 2124, 2594, 2583, 1488, 1215,
1172, 1093. 1H NMR (CDCl3): 7.24 (2 H, d, J=9.2 Hz), 6.90 (2H, d,
J=9.2 Hz), 4.65 (2H, d, J=2.3 Hz), 2.52 (1H, t, J=2.3 Hz). 13C NMR
(CDCl3): 156.1, 129.4, 126.6, 116.3, 78.2, 75.9, 56.1. MS (EI): 168/
166 (M+•), 131 (base peak), 129/127, 103, 99.
5.1.1.8. Propargyl para-fluorophenyl ether (4f). Method B. Yield: 95%
(pale yellow oil). IR (ATR, cm−1): 3298, 2865, 2125, 1602, 1501,
1194, 1026. 1H NMR (CDCl3): 6.95 (m, 4 H, Ar–H), 4.66 (d, 2H, J=
2.4 Hz, CH2), 2.52 (t, 1H, J=2.4 Hz, CH). 13C NMR (CDCl3): 157.0
(J=239 Hz), 154.0 (J=2.3 Hz), 116.0 (J=7.8 Hz), 115.5 (J=
23.4 Hz), 78.4, 75.6, 56.5. MS (EI): 150 (M+•), 149, 111, 83 (base
peak), 57, 39.
5.1. Preparation of 3a–j, 4a–j, 6b–c, and 7
5.1.1. Synthesis of propargyl aryl ethers 4a–4j
5.1.1.1. Method A [20]. The respective phenol derivative (10 mmol), po-
tassium carbonate (4.14 g, 30 mmol), propargyl bromide (1.30 mL of
an 80% solution in toluene, 12 mmol), and tetrabutylammonium bro-
mide (0.332 g, 1 mmol) were stirred and refluxed in 50 mL acetone
overnight. After removal of acetone, water and ethyl acetate (EtOAc)
were added, and the organic layer was washed with 10% NaOH, water
and brine, then dried over MgSO4. Removal of the solvent by rotary
evaporation yielded crystals of sufficient purity in most cases; in other
cases (i.e., 4a), the crude solid was recrystallized from hexanes/ethyl
acetate.
5.1.1.9. Propargyl phenyl ether (4 g). Method A. Yield: 19% (clear oil).
IR (ATR, cm−1): 3289, 3066, 3042, 2870, 2121, 1598, 1589, 1494,
1237, 1213, 1174, 1036, 1021. 1H NMR (CDCl3): 7.30 (dt, 2H, J=
2.4 Hz, 6.4 Hz); 6.98 (m, 3H), 4.69 (d, 2 H, J=2.4 Hz), 2.51 (t, 1H,
J=2.4 Hz). 13C NMR (CDCl3): 157.5, 129.5, 121.6, 114.9, 78.6, 75.4,
55.7. MS (EI): 131 (M+•, base peak), 103, 78, 65, 39.
5.1.1.10. para-(Propargyloxy)acetanilide (4 h). Method B. Yield: 33%
(white needles, mp 119–120 °C). IR (ATR, cm−1): 3311, 3245, 3192,
3132, 3069, 2930, 2867, 2822, 1649, 1610, 1560, 1504, 1447, 1408,
1372, 1271, 1238, 1172, 1115, 1021, 1007. 1H NMR (CDCl3): 7.41
(2H, d, J=9.1 Hz), 7.12 (1H, br s), 6.94 (2H, d, J=9.1 Hz), 4.67 (2H,
d, J=2.4 Hz), 2.51 (1H, t, J=2.4 Hz), 2.15 (3H, s). 13C NMR
(CDCl3): 168.1, 154.4, 131.8, 121.7, 115.4, 78.5, 75.5, 56.2, 24.4. MS
(EI): 189 (M+•), 150, 108 (base peak), 80, 43.
5.1.1.2. Method B [19]. The respective phenol derivative (7.2 mmol)
and potassium carbonate (3 g, 21.6 mmol) were combined in DMF
(15 mL) and heated to 55 °C for 30 min. The reaction was then cooled
to room temperature, and propargyl bromide (0.93 mL of an 80% so-
lution in toluene, 8.6 mmol) was added. The reaction was then stirred
at room temperature for 4 h. Water and EtOAc were then added, and
the organic layer was washed with 10% NaOH, water and brine, dried
over MgSO4, concentrated in vacuo and purified by column chroma-
tography (9:1 hexanes/EtOAc) or recrystallization.
5.1.1.11. Propargyl para-tolyl ether (4i). Method B. Yield: 9% (clear oil).
IR (ATR, cm−1): 3282, 2960, 2927, 2316, 1605, 1510, 1273, 1215. 1H
NMR (CDCl3): 7.09 (2H, d, J=8.4 Hz), 6.87 (2H, d, J=8.4 Hz), 4.65
(2H, d, J=2.3 Hz), 2.49 (1H, t, J=2.3 Hz), 2.28 (3H, s). 13C NMR
(CDCl3): 155.4, 130.9, 129.9, 114.8, 78.8, 75.3, 55.9, 20.5. MS (EI):
146 (M+•), 145, 131, 117, 107, 79, 77 (base peak).
5.1.1.3. Propargyl para-nitrophenyl ether (4a). Method A. Yield: 44%
(pale yellow solid, mp 113–114 °C). IR (ATR, cm−1): 3235, 1651,
1607, 1552, 1504, 1369. 1H NMR (CDCl3): 8.23 (dd, 2H, J=2.2 Hz,
7.4 Hz), 7.06 (dd, 2H, J=2.2 Hz, 7.4 Hz), 4.80 (d, 2H, J=2.4 Hz),
2.58 (t, 1H, J=2.4 Hz). 13C NMR (CDCl3): 161.4, 127.2, 125.8, 115.0,
78.3, 76.3, 56.3. MS (EI): 177 (M+•), 176, 160, 131 (base peak), 130,
103, 77, 63, 39.
5.1.1.12. Propargyl para-anisyl ether (4j). Method B. Yield: 46% (clear
oil). IR (ATR, cm−1): 3287, 2953, 2835, 1505, 1204, 1181, 1035. 1H
NMR (CDCl3): 6.92 (dd, 2H, J=2.8 Hz, 6.8 Hz), 6.83 (dd, 2H, J=
2.8 Hz, 6.8 Hz), 4.62 (d, 2H, J=2.0 Hz), 3.76 (s, 3 H), 2.49 (t, 1H,
J=2.8 Hz). 13C NMR (CDCl3): 154.5, 151.7, 116.2, 114.6, 78.9, 75.3,
56.6, 55.7. MS (EI): 162 (M+•), 147, 123 (base peak), 95, 39.
5.1.1.4. Propargyl para-cyanophenyl ether (4b). Method A. Yield: 40%
(off-white flaky solid, mp 112–113 °C). IR (ATR, cm−1): 3249, 2221,
1604, 1505. 1H NMR (CDCl3): 7.62 (dd, 2H, J=2.4 Hz, 6.8 Hz), 7.05
(dd, 2H, J=2.0 Hz, 6.8 Hz), 4.76 (d, 2H, J=2.4 Hz), 2.57 (t, 1H, J=
2.4 Hz). 13C NMR (CDCl3): 160.7, 133.9, 119.0, 115.6, 105.0, 79.6,
76.5, 55.9. MS (EI): 156 (M+•, base peak), 128, 102, 90, 39.
5.1.2. Synthesis of (hexacarbonyldicobalt)propargyl aryl ethers 3a–3j
5.1.2.1. General method for preparation of cobalt–alkyne complexes [12].
The alkyne and dicobalt octacarbonyl (at least 1.5 equivalents) were
dissolved in THF. The solution was stirred at room temperature
under nitrogen atmosphere, venting to an oil bubbler, until no
starting material was observed by thin-layer chromatography. The
products, dark red crystals in all cases, were purified by column chro-
matography (hexanes:EtOAc 15:1).
5.1.1.5. Ethyl para-(propargyloxy)benzoate (4c). Method B. Yield: 63%
(white needles, mp 43–44 °C). IR (ATR, cm−1): 3248, 2988, 2131,
1695, 1508, 1276, 1245, 1169, 1104, 1024. 1H NMR (CDCl3): 8.02
(dd, 2H, J=2.0, 7.2 Hz), 7.00 (dd, 2H, J=2.0 Hz, 7.2 Hz), 4.75 (d,
2H, J=2.8 Hz), 4.35 (q, 2H, J=7.2 Hz), 2.54 (t, 1 H, J=2.4 Hz), 1.38
(t, 3H, J=7.2 Hz). 13C NMR (CDCl3): 166.2, 161.1, 131.5, 123.8,
114.4, 77.8, 76.0, 60.7, 55.8, 14.3. MS (EI): 204 (M+•), 175, 159, 131
(base peak), 103, 77, 64.
5.1.2.2. (Hexacarbonyldicobalt)propargyl para-nitrophenyl ether (3a).
Yield: 33% mp: 108–110 °C. IR (ATR, cm−1): 2098, 2051, 2032,
1997, 1591, 1504, 1492, 1340, 1253, 1168, 1110, 1000. 1H NMR
(CDCl3): 8.24 (d, 2H, J=9.2 Hz), 7.02 (d, 2H, J=9.2 Hz), 6.10 (s,
5.1.1.6. Propargyl para-bromophenyl ether (4d). Method A. Yield: 36%
(clear oil). IR (ATR, cm−1): 3293, 2979, 1580, 1486, 1071, 1219,

S.D. Schimler et al. / Journal of Inorganic Biochemistry 119 (2013) 28–37
35
1H), 5.29 (s, 2H). 13C NMR (CDCl3): 199.9, 163.1, 142.0, 126.0, 114.7,
87.5, 71.8, 69.2.
154.2, 152.3, 115.6, 114.7, 89.7, 71.8, 68.8, 55.7. HR-MS (EI): calcd.
for C15H10O7Co2 [M+-CO] 419.9091; found 419.9087.
HR-MS (ESI-TOF): calcd. for C15H7NO9Co2Na [M+Na] 485.8682;
found 485.8697.
5.1.3. Synthesis of analogs of 3a (6a–c, 7)
5.1.2.3. (Hexacarbonyldicobalt)propargyl para-cyanophenyl ether (3b).
Yield: 84%; mp: 103 °C (decomp). IR (ATR, cm−1): 2220, 2098,
2051, 2032, 2000, 1602, 1505, 1350, 1252, 1168, 1003. 1H NMR
(CDCl3): 7.62 (d, 2 H, J=8.4 Hz), 7.01 (s, 2H, J=8.4 Hz), 6.08 (s, 1H),
5.24 (s, 2H). 13C NMR (CDCl3): 198.9, 161.3, 134.1, 119.0, 115.3, 104.6,
87.8, 71.7, 68.7. HR-MS (EI): calcd. for C15H7NO6Co2 [M+-CO]
414.8937; found 414.8934.
5.1.3.1. (Hexacarbonyldicobalt)-3-butynyl para-nitrophenyl ether (6a)
[22,23]. 3-butyn-1-ol (0.27 mL, 3.6 mmol) was dissolved in dry THF
(25 mL). Sodium hydride (138 mg of a 60% suspension, 3.6 mmol)
was
then
added
to
the
reaction,
followed
by
para-fluoronitrobenzene (0.32 mL, 3.0 mmol). The reaction was
stirred at room temperature overnight, then concentrated under re-
duced pressure, redissolved in ether, and washed with aqueous HCl
(1 N), saturated aqueous NaHCO3, and water. The organic layer was
concentrated again under reduced pressure, and the crude product
was recrystallized from hexanes/ethyl acetate (9:1) to obtain
222 mg (39%) of white needles. mp: 90–91 °C. IR (ATR, cm−1):
3291, 2117, 2951, 1915. 1H NMR (CDCl3): 8.21 (dd, 2H, J=2.2 Hz,
7.2 Hz), 6.98 (dd, 2H, J=2.2 Hz, J=7.2 Hz), 4.19 (t, 2H), 2.74 (m,
2H), 2.07 (t, 1H, J=2.8 Hz) 13C NMR (CDCl3): 163.4, 141.8, 125.9,
114.5, 79.6, 70.4, 66.6, 19.4. MS (EI): 191 (M+•), 152, 145, 139, 109,
53 (base peak). The cobalt–alkyne complex was then synthesized
using the usual method and purified by column chromatography
(hexanes:ethyl acetate 15:1), yielding a dark red crystalline com-
pound. Yield: 58%. mp: 69–70 °C. IR (ATR, cm−1): 3113, 2907, 2088,
2046, 1986, 1609, 1592, 1507, 1338, 1253, 1178, 1111, 1030. 1H
NMR (CDCl3): 8.22 (2H, d, J=9.4 Hz), 6.97 (2H, d, J=9.4 Hz), 6.10
(1 H, t, J=1.2 Hz), 4.30 (2H, t, J=5.8 Hz), 3.37 (2H, dt, J=1.2 Hz,
5.8 Hz). 13C NMR (CDCl3): 199.6, 163.4, 141.8, 126.0, 114.3, 91.1,
73.8, 68.6, 33.0. HR-MS (ESI-TOF): calcd. for C16H10NO9Co2 [M+H]
477.9020; found 477.9019.
5.1.2.4. Ethyl para-((hexacarbonyldicobalt)propargyloxy)benzoate (3c).
Yield: 73%; mp: 49–50 °C. IR (ATR, cm−1): 2989, 2904, 2094, 2019,
2009, 1997, 1706, 1606, 1508, 1289, 1240, 1170, 1106, 1010. 1H
NMR (CDCl3): 8.03 (d, 2H, J=8.0 Hz), 6.97 (d, 2H, J=8.0 Hz), 6.07
(s, 1H), 5.24 (s, 2H), 4.36 (q, 2H, J=6.4 Hz), 1.39 (t, 3H, J=6.4 Hz). 13C
NMR (CDCl3): 199.1, 166.3, 161.8, 131.7, 123.6, 114.2, 88.5, 71.8, 68.5,
60.7, 14.4. HR-MS (ESI-TOF): calcd. for C18H13O9Co2 [M+H] 490.9224;
found 490.9225.
5.1.2.5. (Hexacarbonyldicobalt)propargyl para-bromophenyl ether (3d).
Yield: 62%; mp: 87–90 °C. IR (ATR, cm−1): 2097, 2048, 2031, 1990,
1580, 1486, 1238, 1012, 998. 1H NMR (CDCl3): 7.39 (d, 2H, J=
8.4 Hz), 6.84 (d, 2H, J=9.2 Hz), 6.05 (s, 1H), 5.16 (s, 2H). 13C NMR
(CDCl3): 199.2, 157.3, 132.4, 116.5, 113.4, 88.8, 71.8, 68.5.
5.1.2.6. (Hexacarbonyldicobalt)propargyl para-chlorophenyl ether (3e).
Yield: 68% mp: 87–88 °C. IR (ATR, cm−1): 3092, 2839, 2097, 2048,
2031, 1989, 1583, 1489, 1347, 1238, 1169, 1014, 999. 1H NMR
(CDCl3): 7.26 (2H, d, J=8.0 Hz), 6.87 (2H, d, J=8.0 Hz), 6.04 (1H, s),
5.15 (s, 2H). 13C NMR (CDCl3): 199.1, 156.7, 129.4, 126.1, 115.9, 88.8,
71.7, 68.5.
5.1.3.2. (Hexacarbonyldicobalt)-2-methyl-3-butyn-2-yl para-nitrophenyl
ether (6b). Obtained by the same method used to synthesize 6a, except
that the crude ether intermediate was purified by column chromatog-
raphy (hexanes:ethyl acetate 30:1) instead of recrystallization. Char-
acterization data for ether intermediate: Yield: 14% (yellow oil). IR
(ATR, cm−1): 3289, 2992, 2113, 1607, 1590, 1511, 1339, 1254, 1133,
1110. 1H NMR (CDCl3): 8.17 (2H, d, J=9.0 Hz), 7.30 (2H, d, J=
9.0 Hz), 2.71 (1H, s), 1.73 (6H, s). 13C NMR (CDCl3): 161.3, 142.1,
125.2, 119.0, 84.5, 75.5, 73.0, 29.5. MS(EI): 205 (M+•), 190, 139, 123,
109, 107, 67 (base peak). Characterization data for 6b: Yield: 45%.
mp: 79–80 °C. IR (ATR, cm−1): 3129, 2993, 2093, 1999, 1590, 1508,
1491, 1342, 1257, 1112. 1H NMR (CDCl3): 8.19 (2H, d, J=9.0 Hz),
7.13 (2H, d, J=9.0 Hz), 6.08 (1 H, s), 1.79 (6H, s). 13C NMR (CDCl3):
199.4, 161.1, 142.6, 125.3, 120.9, 101.7, 83.0, 72.2, 30.8. HR-MS
(ESI-TOF): calcd. for C17H11NO9Co2Na [M+Na] 513.8995; found
513.8998.
5.1.2.7. (Hexacarbonyldicobalt)propargyl para-fluorophenyl ether (3f).
Yield: 72% mp: 58–59 °C. IR (ATR, cm−1): 3091, 2839, 2097, 2030,
1971, 1504, 1343, 1243, 1204, 1097, 1023, 1000. 1H NMR (CDCl3):
7.00 (2H, m), 6.89 (2H, m), 6.05 (1H, s), 5.15 (2H, s). 13C NMR
(CDCl3): 199.2, 157.8 (d, J=239.0 Hz), 154.2 (d, J=2.3 Hz), 115.9
(d, J=22.5 Hz), 115.6 (d, J=7.8 Hz), 89.1, 71.7, 68.8.
5.1.2.8. (Hexacarbonyldicobalt)propargyl phenyl ether (3g). Yield: 20%,
mp: 45–46 °C. IR (ATR, cm−1): 2096, 2048, 2029, 1989, 1598, 1586,
1494, 1242, 1031. 1H NMR (CDCl3): 7.31 (m, 2H), 6.98 (m, 3H), 6.06
(s, 1H), 5.19 (s, 2H). 13C NMR (CDCl3): 199.3, 158.2, 129.6, 121.2,
114.6, 89.6, 71.8, 68.1.
5.1.2.9. para-((Hexacarbonyldicobalt)propargyloxy)acetanilide (3h).
Yield: 74% mp: 177–178 °C. IR (ATR, cm−1): 2097, 2050, 2033,
2000, 1657, 1556, 1507, 1267, 1237, 1019. 1H NMR (CDCl3): 7.42
(d, 2H, J=7.6 Hz), 6.91 (d, 2H, J=8.0 Hz), 6.05 (s, 1H), 5.17 (s, 2H),
2.17 (s, 3H). 13C NMR (CDCl3): 199.2, 168.1, 155.0, 131.5, 121.8,
115.0, 89.3, 71.8, 68.5, 24.4.
5.1.3.3. (Hexacarbonyldicobalt)-N-(para-nitrophenyl)propargylamine (6c)
[24]. para-Fluoronitrobenzene (0.32 mL, 3.0 mmol), propargylamine
(0.38 mL, 6.0 mmol), potassium carbonate (0.41 g, 3.0 mmol) and po-
tassium fluoride (0.17 g, 3.0 mmol) were combined in DMSO (10 mL)
and stirred at room temperature overnight. The reaction was then
diluted with water (30 mL) and filtered; the precipitate was washed
with water (2×) and isopropyl alcohol, and let dry under vacuum.
Yield: 400 mg (76%) of a yellow powder. mp: 151–153 °C. IR (ATR,
cm−1): 3407, 3242, 2853, 2124, 1592, 1523, 1500, 1451, 1291, 1261,
1130, 1090. 1H NMR (CDCl3): 8.13 (2H, d, J=9.4 Hz), 6.64 (2H, d, J=
9.4 Hz), 4.69 (1H, br s), 2.29 (1 H, t, 2.4 Hz), 1.58 (2H, s). 13C NMR
5.1.2.10. (Hexacarbonyldicobalt)propargyl para-tolyl ether (3i). Yield:
78%. mp: 45–47 °C. IR (ATR, cm−1): 3087, 2983, 2836, 2097, 2051,
2032, 1990, 1586, 1509, 1441, 1347, 1238, 1018, 994. 1H NMR
(CDCl3): 7.10 (s, 2H), 6.85 (s, 2H), 6.04 (s, 1H), 5.15 (s, 2H), 2.30
(s, 3H). 13C NMR (CDCl3): 199.3, 156.0, 130.5, 130.0, 114.5, 89.8,
71.8, 68.2, 20.5.
(CDCl3): 151.9, 139.2, 126.2, 111.9, 79.0, 72.4, 33.1. MS (EI): 176 (M+•
,
base peak), 146, 130, 129, 103, 91, 77. The cobalt–alkyne complex was
then synthesized using the usual method and purified by column chro-
matography (hexanes:EtOAc 15:1), yielding a dark red crystalline com-
pound. Yield: 41% mp: 82 (decomp.) IR (ATR, cm−1): 3348, 2904, 2092,
2035, 2009, 1994, 1598, 1473, 1299, 1272, 1180, 1106. 1H NMR (CDCl3):
8.13 (2 H, d, J=9.2 Hz), 6.63 (2H, d, J=9.2 Hz), 6.09 (1H, s), 4.88 (br t,
5.1.2.11. (Hexacarbonyldicobalt)propargyl para-anisyl ether (3j). Yield:
70%. mp: 62–63 °C. IR (ATR, cm−1): 3095, 2967, 2891, 2836, 2097,
2030, 1982, 1594, 1229, 1041, 1025. 1H NMR (CDCl3): 6.87 (s, 4H),
6.04 (s, 1H), 5.14 (s, 2H), 3.78 (s, 3H). 13C NMR (CDCl3): 199.2,

36
S.D. Schimler et al. / Journal of Inorganic Biochemistry 119 (2013) 28–37
J≈5.5 Hz), 4.62 (2H, d, J=6.3 Hz). 13C NMR (CDCl3): 199.1, 152.0,
was washed once in PBS, and the cells were fixed in 70% ethanol.
Fixed cells were then washed twice in PBS, stained with 50 μM PI
containing 5ug/mL DNAsefree RNase, incubated 30 min, and analyzed
by flow cytometry (10,000 cells).
139.0, 126.4, 111.9, 91.0, 72.8, 45.8. HR-MS (ESI-TOF): calcd. for
C15H9N2O8Co2 [M+H] 462.9023; found 462.9025.
5.1.3.4. (Hexacarbonyldicobalt)propargyl (para-nitrophenyl)methyl ether
(7). Sodium hydride (0.29 g of a 60% oil suspension, 7.2 mmol) was
washed twice with hexanes and suspended in THF (12 mL).
4-nitrobenzyl alcohol (0.92 g, 6.0 mmol) and propargyl bromide
(0.78 mL of an 80% toluene solution, 7.2 mmol) were then added to the
reaction, which was subsequently stirred at RT overnight. The mixture
was then filtered through a Celite plug, concentrated by rotary evapora-
tion, and purified by column chromatography (hexanes:EtOAc 9:1) to a
white crystalline solid (311 mg, 27%). mp: 55–56 °C. IR (ATR, cm−1):
3282, 2923, 2860, 2115, 1603, 1510, 1450, 1343, 1300, 1258, 1091,
1030, 1017, 953. 1H NMR (CDCl3): 8.17 (2H, dd, J=2.4, 6.8 Hz), 7.30
(2H, dd, J=2.4, 6.8 Hz), 2.71 (1H, s), 1.73 (6H, s). 13C NMR (CDCl3):
161.2, 142.0, 125.1, 118.9, 84.4, 75.5, 72.9, 29.3. MS (EI): 191 (M+•),
190, 174, 161, 152 (base peak), 150, 137, 115, 107, 89, 78, 77. The
cobalt–alkyne complex was then synthesized using the usual method
and purified by column chromatography (hexanes:EtOAc 15:1), yield-
ing a dark red crystalline compound. Yield: 71%. mp: 71–72 °C. IR
(ATR, cm−1): 2872, 2092, 1990, 1603, 1519, 1345, 1107. 1H NMR
(CDCl3): 8.22 (2H, d, J=8.6 Hz), 7.54 (2H, d, J=8.6 Hz), 6.07 (1H, s),
4.81 (2H, s), 4.74 (2H, s). 13C NMR (CDCl3): 198.4, 146.4, 144.5, 126.5,
122.7, 89.3, 76.2, 70.5, 70.1. MS (ESI-TOF): 448.9 (M+-CO).
5.6. Myoglobin assay for compound stability in aqueous buffer [27]
A solution of myoglobin in PBS (0.01 M) was prepared (45–55 μM,
about 1 mg/mL), degassed by bubbling nitrogen for 15 min, and re-
duced to deoxy-Mb with sodium dithionite (0.1% w/w). In a quartz
cuvette, a 0.8 μL aliquot of either DMF or a freshly prepared solution
of a compound in DMF (100 mM) was briefly mixed with 2 mL of
the deoxy-Mb solution (40 μM final concentration) and the sample
was overlaid with mineral oil to prevent oxidation of deoxy-Mb
by the air. In another cuvette, 2 mL of the deoxy-MB solution was
saturated with CO, either from a cylinder or produced separately by
the slow addition of sulfuric acid to formic acid [28]. The measurement
of the absorbance of this solution at 540 nm allowed the calculation
of the total CO-Mb concentration ([CO-Mb]max=ACO,540/ε, where ε,
the extinction coefficient of CO-Mb, is 15,400 μM−1 cm−1).
The conversion of deoxy-Mb to CO-Mb in each sample was then
monitored by UV–visible spectroscopy. The concentration of CO-Mb
at 2 h ([CO-Mb]f) was then calculated for each sample from its ab-
sorbances at 540 nm and 510 nm (an isosbestic point between
deoxy-Mb and CO-Mb):
hꢀ
ꢁ
ꢀ
ꢁ
i
ꢂ
ꢃ
½CO−Mbꢀf ¼ ½CO−Mbꢀmax Af −Ai
−
Af −Ai
= ðACO−AiÞ540−ðACO−AiÞ510
5.2. Cell culture
540
510
where the subscripts indicate the initial (i) and final (f) values for the
sample and the value in the spectrum of the CO-saturated solution
(CO). Each set of runs was accompanied by a control run, which simply
added DMF, but not any test compound, to the deoxy-Mb solution;
these control runs showed minimal absorbance changes during the
2 h run (b1 μM calculated [CO-Mb]f).
MDA-MB-231 cells were cultured in RPMI 1640 (Cellgro) with 10%
cosmic calf serum (Cellgro, Manassa, VA) and 1% penicillin/streptomycin
(Gibco) at 37 °C in a humidified incubator with 5% CO2, and passaged
upon reaching plate confluence.
5.3. Assessment of antiproliferative activity by crystal violet staining
5.7. Measurement of oxidative stress by flow cytometry
MDA-MB-231 cells were plated on 96 well plates (TPP) at a density
of 6×104 cells/mL, incubated overnight, treated with either a test
compound or vehicle control (0.4% DMF) (five replicates), further in-
cubated for 72 h, fixed with 1% glutaraldehyde in PBS (100 μL/well),
and stained with 0.02 M crystal violet (100 μL/well). After the plates
were thoroughly washed with water, the stain was solubilized with
aqueous ethanol (70%, 200 μL/well) on a rotary shaker for 3 h, and
the absorbance at 620 nm was read with a plate reader and normal-
ized to control.
MDA-MB-231 cells were disassociated by trypsin and centrifuged
for two min at 4000 rpm. The supernatant was replaced with HBSS,
and the pellet resuspended in medium by gentle agitation at a con-
centration of ~1.0×106 cells/mL and kept on ice. The suspension
was then loaded with 25 μM H2DCFDA and incubated for 30 min at
37 °C. After 30 min, aliquots were transferred to flow cytometry
tubes (500,000 cells/tube) and centrifuged, and the supernatant was
replaced by media containing the test compounds (or the vehicle
control). After incubating for 30 min further, flow cytometry was
performed by measuring the DCF fluorescence of 10,000 cells. Data
were expressed as mean fluorescent intensity (MFI).
5.4. Measurement of apoptosis induction by flow cytometry
MDA-MB-231 cells were seeded in 12-well plates at 4×104 cells/
well and incubated overnight. Wells were then dosed with test com-
pound at 20 μM and incubated at 37 °C 18 h further. A 0.2% DMF
solution was used as the vehicle control. The supernatant from each
well was removed to a labeled tube, and the adherent cells were
trypsinized and added to the appropriate tube. The tubes were then
centrifuged, and the supernatant replaced with 0.5 mL HBSS
containing 40 mM CaCl2. To each tube was then added 1 μL PI and
5 μL Annexin V-FITC. The cells were then immediately analyzed by
flow cytometry, measuring both the PI and Annexin V-FITC fluores-
cence of 10,000 cells.
5.8. Measurement of mitochondrial membrane permeability by flow
cytometry
MDA-MB-231 cells were treated with 20 μM of test compound in
media for 18 h at 37 °C. Cells were then disassociated by trypsin
and centrifuged for 2 min at 4000 rpm. The supernatant was re-
moved, and the pellet resuspended in medium by gentle agitation,
transferred to flow cytometry tubes (500,000 cells/tube), and treated
with R123 for 10 min on ice. Flow cytometry was performed by mea-
suring 10,000 cells. Data were expressed as mean fluorescent intensity
(MFI).
5.5. Cell cycle analysis
Abbreviations
MDA-MB-231 cells were grown in a 6-well plate at 37 °C in a 5%
CO2/95% air atmosphere for 24 h. The cells were then dosed with
test compounds at 20 μM in DMF and incubated for an additional
18 h. After incubation with the drugs, cells were trypsinized and
spun at 7000 rpm for 6 min. The supernatant was removed, the pellet
ATR
attenuated total reflectance
CO
carbon monoxide
CORM
carbon monoxide releasing molecule

S.D. Schimler et al. / Journal of Inorganic Biochemistry 119 (2013) 28–37
37
[10] J.L. Hickey, R.A. Ruhayel, P.J. Barnard, M.V. Baker, S. Berners-Price, A. Filipovska,
J. Am. Chem. Soc. 130 (2008) 12570–12571.
[11] P.J. Barnard, M.V. Baker, S.J. Berners-Price, D.A. Day, J. Inorg. Biochem. 98 (2004)
1642–1647.
[12] K. Schmidt, M. Jung, R. Keilitz, B. Schnurr, R. Gust, Inorg. Chim. Acta 306 (2000)
6–16.
[13] M. Jung, D.E. Kerr, P.D. Senter, Arch. Pharm. (Weinheim) 330 (1997) 173–176.
[14] I. Ott, K. Schmidt, B. Kircher, P. Schumacher, T. Wiglenda, R. Gust, J. Med. Chem. 48
(2005) 622–629.
[15] I. Ott, B. Kircher, C.P. Bagowski, D.H. Vlecken, E.B. Ott, J. Will, K. Bensdorf, W.S.
Sheldrick, R. Gust, Angew. Chem. Int. Ed Engl. 48 (2009) 1160–1163.
[16] I. Ott, B. Kircher, R. Dembinski, R. Gust, Expert Opin. Ther. Patents 18 (2008)
327–337.
[17] F. Dubar, J. Khalife, J. Brocard, D. Dive, C. Biot, Molecules 13 (2008) 2900–2907.
[18] E.A. Hilliard, A. Vessiéres, G. Jaouen, in: G. Jaouen, N. Metzler-Nolte (Eds.), Medicinal
Organometallic Chemistry, Springer, Berlin, 2010, pp. 81–117.
[19] M. Pal, K. Parasuraman, K.R. Yeleswarapu, Org. Lett. 5 (2003) 349–352.
[20] Q. Chen, Y. Cui, T. Zhang, J. Cao, B. Han, Biomacromolecules 11 (2010) 13–19.
[21] BioDataFit, Chang BioScience, http://www.changbioscience.com/stat/ec50.html
(accessed May 31, 2012).
[22] B.M. Trost, B.L. Ashfeld, Org. Lett. 10 (2008) 1893–1896.
[23] O. Yamada, H. Matsumoto, T. Shimizu, United States Patent 2006/0041031
(2003).
[24] D.J. McNamara, E.M. Berman, D.W. Fry, L.M. Werbel, J. Med. Chem. 33 (1990)
2045–2051.
COX
DCF
DMF
EI-MS
cyclooxygenase
2′,7′-dichlorofluorescein
N,N-dimethylformamide
electron impact ionization mass spectrometry
ESI-MS electrospray ionization mass spectrometry
FITC
H2DCF
fluorescein isothiocyanate
2′,7′-dichlorodihydrofluorescein
H2DCFDA 2′,7′ dichlorodihydrofluorescein diacetate
HBSS Hanks' buffered salt solution
HR-MS high resolution mass spectrometry
MMP mitochondrial membrane permeability
MMP-7 matrix metalloprotein-7
NAC
NO
PBS
PI
PNP
R123
RNS
ROS
RPMI
THF
N-acetylcysteine
nitric oxide
phosphate-buffered saline
propidium iodide
para-nitrophenol
rhodamine 123
reactive nitrogen species
reactive oxygen species
Roswell Park Memorial Institute medium
tetrahydrofuran
[25] R.F. Lockwood, K.M. Nicholas, Tetrahedron Lett. 18 (1977) 4163–4165.
[26] B.J. Teobald, Tetrahedron 58 (2002) 4133–4170.
[27] A.J. Atkin, S. Williams, P. Sawle, R. Motterlini, J.M. Lynam, I.J.S. Fairlamb, Dalton
Trans. (2009) 3653–3656.
[28] J.A. Bailey, J. Chem. Educ. 88 (2011) 995–998.
[29] N. Zamzami, G. Kroemer, Curr. Biol. 13 (2003) R71–R73.
[30] T. Ozben, J. Pharm. Sci. 96 (2007) 2181–2196.
TMRM tetramethylrhodamine methyl ester
[31] S. Reuter, S.C. Gupta, M.M. Chaturvedi, B.B. Aggarwal, Free Radic. Biol. Med. 49
(2010) 1603–1616.
Acknowledgements
[32] S. Leonard, P.M. Gannett, Y. Rojanasakul, D. Schwegler-Berry, V. Castranova, V.
Vallyathan, X. Shi, J. Inorg. Biochem. 70 (1998) 239–244.
[33] A. Gomes, E. Fernandes, J.L. Lima, J. Biochem, Biophys. Methods 65 (2005) 45–80.
[34] X. Chen, Z. Zhong, Z. Xu, L. Chen, Y. Wang, Free Radic. Res. 44 (2010) 587–604.
[35] G. Kroemer, Curr. Med. Chem. 10 (2003) 1469–1472.
[36] G. Kroemer, L. Galluzzi, C. Brenner, Physiol. Rev. 87 (2007) 99–163.
[37] L. Galluzzi, N. Larochette, N. Zamzami, G. Kroemer, Oncogene 25 (2006)
4812–4830.
[38] G. Edsall, N. Engl. J. Med. 211 (1934) 385–390.
[39] L. Galluzzi, N. Zamzami, T. de La Motte Rouge, C. Lemaire, C. Brenner, G. Kroemer,
Apoptosis 12 (2007) 803–813.
[40] T.R. Johnson, B.E. Mann, J.E. Clark, R. Foresti, C.J. Green, R. Motterlini, Angew.
Chem. Int. Ed Engl. 42 (2003) 3722–3729.
[41] R. Motterlini, J.E. Clark, R. Foresti, P. Sarathchandra, B.E. Mann, C.J. Green, Circ. Res.
90 (2002) E17–E24.
This work was supported by a National Science Foundation-Major
Research Instrumentation grant (CHE-0923473). Intramural support
for was received from the Chester Hill Fund, the Excellence in Science
Fund and the Steitz Fund. Mr. Daniel Goetz, Ms. N. Jasmine Hamid and
Ms. Meareg G. Amare are thanked for synthetic contributions. The
Q-Tof Ultima mass spectrometer was purchased in part with a grant
from the National Science Foundation, Division of Biological Infra-
structure (DBI-0100085). The 70-VSE mass spectrometer was pur-
chased in part with a grant from the Division of Research Resources,
National Institutes of Health (RR 04648). The reviewers of this work
are thanked for helpful comments and suggestions.
[42] R. Motterlini, L.E. Otterbein, Nat. Rev. Drug Discov. 9 (2010) 728–743.
[43] C. Peers, D.S. Steele, J. Mol. Cell. Cardiol. 52 (2012) 359–365.
[44] M. Karlsson, T. Kurz, U.T. Brunk, S.E. Nilsson, C.I. Frennesson, Biochem. J. 428
(2010) 183–190.
References
[45] C. Biot, D. Taramelli, I. Forfar-Bares, L.A. Maciejewski, M. Boyce, G. Nowogrocki, J.S.
Brocard, N. Basilico, P. Olliaro, T.J. Egan, Mol. Pharm. 2 (2005) 185–193.
[46] C. Biot, D. Dive, in: G. Jaouen, N. Metzler-Nolte (Eds.), Bioorganic Chemistry and
Malaria, Medicinal Organometallic Chemistry, vol. 32, Springer, Berlin, 2010,
pp. 155–193.
[47] A. Nguyen, A. Vessières, E.A. Hillard, S. Top, P. Pigeon, G. Jaouen, Chim. Int. J.
Chem. 61 (2007) 716–724.
[48] D. Osella, H. Mahboobi, D. Colangelo, G. Cavigiolio, A. Vessières, G. Jaouen, Inorg.
Chim. Acta 358 (2005) 1993–1998.
[49] J.R. Kirshner, S. He, V. Balasubramanyam, J. Kepros, C.Y. Yang, M. Zhang, Z. Du, J.
Barsoum, J. Bertin, Mol. Cancer Ther. 7 (2008) 2319–2327.
[50] K.J. Coe, Y. Jia, H.K. Ho, P. Rademacher, T.K. Bammler, R.P. Beyer, F.M. Farin, L.
Woodke, S.R. Plymate, N. Fausto, S.D. Nelson, Chem. Res. Toxicol. 20 (2007)
1277–1290.
[1] G. Jaouen, N. Metzler-Nolte, Medicinal Organometallic Chemistry, Springer,
Berlin, 2010.
[2] G. Gasser, I. Ott, N. Metzler-Nolte, J. Med. Chem. 54 (2011) 3–25.
[3] E. Meggers, Curr. Opin. Chem. Biol. 11 (2007) 287–292.
[4] S.H. van Rijt, P.J. Sadler, Drug Discov. Today 14 (2009) 1089–1097.
[5] M. Hogan, M. Tacke, in: G. Jaouen, N. Metzler-Nolte (Eds.), Medicinal Organometallic
Chemistry, Springer, Berlin, 2010, pp. 119–140.
[6] C. Biot, G. Glorian, L.A. Maciejewski, J.S. Brocard, O. Domarle, G. Blampain, P.
Millet, A.J. Georges, H. Abessolo, D. Dive, J. Lebibi, J. Med. Chem. 40 (1997)
3715–3718.
[7] J.M. Rademaker-Lakhai, D. van den Bongard, D. Pluim, J.H. Beijnen, J.H. Schellens,
Clin. Cancer Res. 10 (2004) 3717–3727.
[8] M.A. Jakupec, V.B. Arion, S. Kapitza, E. Reisner, A. Eichinger, M. Pongratz, B.
Marian, N. Graf von Keyserlingk, B.K. Keppler, Int. J. Clin. Pharmacol. Ther. 43
(2005) 595–596.
[51] I. Ott, R. Gust, Biometals 18 (2005) 171–177.
[9] Y. Fu, A. Habtemariam, A.M. Pizarro, S.H. van Rijt, D.J. Healey, P.A. Cooper, S.D.
Shnyder, G.J. Clarkson, P.J. Sadler, J. Med. Chem. 53 (2010) 8192–8196.