CO release from norbornadiene iron(0) tricarbonyl complexes: Importance of ligand dissociation

Article abstract of DOI:10.1021/om300419w

An investigation into the CO-releasing properties of a range of iron tricarbonyl and chromium and molybdenum tetracarbonyl complexes is presented. Iron tricarbonyl complexes containing the 2,5-bicyclo[2.2.1]heptene (norbornadiene) ligand are shown to be effective CO-releasing molecules, in which the rate and extent of CO release may be modulated by modification of the norbornadiene framework. Species containing the parent norbornadiene and those with a substituent at the 7-position of the organic ligand exhibit CO release; those containing ester substituents at the 2- and/or 3-positions do not. A mechanism for CO release in this species is proposed which involves initial norbornadiene dissociation, a suggestion which is supported by the spectroscopic data and the observation that the addition of excess substituted norbornadiene retards the rate of CO release. CO release from the diester-containing norbornadiene complex may be promoted photochemically, and cell viability studies indicate that in the absence of light this complex is nontoxic, making it an excellent candidate for further study as a photo-CO-RM. Both the chromium and molybdenum tetracarbonyl complexes release CO, which in the case of the molybdenum analogue is rapid.

Full text of DOI:10.1021/om300419w

Article
pubs.acs.org/Organometallics
CO Release from Norbornadiene Iron(0) Tricarbonyl Complexes:
Importance of Ligand Dissociation
Anthony J. Atkin, Ian J. S. Fairlamb,* Jonathan S. Ward, and Jason M. Lynam*
Department of Chemistry, University of York, York YO10 5DD, U.K.
S
* Supporting Information
ABSTRACT: An investigation into the CO-releasing proper-
ties of a range of iron tricarbonyl and chromium and
molybdenum tetracarbonyl complexes is presented. Iron
tricarbonyl complexes containing the 2,5-bicyclo[2.2.1]-
heptene (norbornadiene) ligand are shown to be effective
CO-releasing molecules, in which the rate and extent of CO
release may be modulated by modification of the norborna-
diene framework. Species containing the parent norbornadiene and those with a substituent at the 7-position of the organic
ligand exhibit CO release; those containing ester substituents at the 2- and/or 3-positions do not. A mechanism for CO release in
this species is proposed which involves initial norbornadiene dissociation, a suggestion which is supported by the spectroscopic
data and the observation that the addition of excess substituted norbornadiene retards the rate of CO release. CO release from
the diester-containing norbornadiene complex may be promoted photochemically, and cell viability studies indicate that in the
absence of light this complex is nontoxic, making it an excellent candidate for further study as a photo-CO-RM. Both the
chromium and molybdenum tetracarbonyl complexes release CO, which in the case of the molybdenum analogue is rapid.
amino acids,10 sugars,11 and the 2-pyrone motif,12 with a view
to ensuring that potential toxic side effects are minimized. In
addition, much of the development of new CO-RMs has
INTRODUCTION
In apparent contrast to its well-established toxicity, carbon
■
monoxide (CO) has been shown to have a significant
physiological function.1 In mammals, the majority of
focused on the use of metals with a defined biological role such
as iron,10d,e,12,13 manganese,14 and molybdenum.10c,11,12b,13c
endogenously produced CO arises from the catabolism of
heme. Following the discovery that CO plays a well-defined
physiological role, its ability to act as a potential beneficial
therapeutic agent has been extensively investigated. Indeed, CO
However, this has not precluded the use of a wide range of
metals which have no, or little, intrinsic biological role, notably
with ruthenium in CO-RM-310a,b as well as chromium-,10c,13c
tungsten-,10c,13c,15 rhenium-,14h,16 cobalt-,17 and iridium-based18
gas has been shown to be anti-inflammatory,2 to protect
systems being developed.
ischemic damage and reperfusion injury,3 to regulate blood
The mechanistic processes which control CO from a
pressure,4 and to act as an antibacterial agent.5 CO may also
potential therapeutic agent have been investigated. In some
suppress organ graft rejection6 and assist with the treatment of
instances, the rate of release of CO has been compared with
cerebral malaria.4,7
various spectroscopic parameters involving the carbonyl ligand,
Although the beneficial effects of carbon monoxide have
such as the CO stretching frequency in the infrared
been explicitly demonstrated, the gaseous form does not
spectrum,13b,d and in some cases a higher CO stretching
represent a long-term therapeutic solution, principally due to its
inherent toxicity and lack of selectivity. Furthermore, the
precise dosing of the flammable gas is problematic. In order to
frequency correlates with a more rapid rate of release. These
data have been interpreted in terms of reduced metal−carbon
back-bonding, resulting therefore in a more labile CO ligand
circumvent these potential drawbacks, carbon monoxide-
(q.v.) and a mechanistic picture in which initial CO dissociation
is dominant.13d,14f However, in some instances, the infrared
releasing molecules (CO-RMs) have been developed as
therapeutic agents that will act as sources of CO in vivo.1a,8
stretching frequency of the metal carbonyl groups does not
The vast majority of CO-RMs take advantage of the well-
correlate with the rate of CO release and a number of
known affinity of transition-metal compounds for CO, although
alternative mechanistic processes in which release is triggered
a boron-based CO-RM has also been employed9 and it has
been shown that organic compounds such as CH2Cl2 may also
act as sources of carbon monoxide.1a
by, for example, hydrolysis13c or dissociation10c of a coligand or
interaction with the delivery medium19 have been established.
Other methods for the activation of metal carbonyls have
Due to the diverse nature of the applications, the optimum
release rate for CO release is not known; thus, the development
of a variety of CO-RMs with different activities is desirable.
One successful strategy to develop a new CO-RM has focused
on the incorporation of biologically compatible ligands such as
Special Issue: Organometallics in Biology and Medicine
Received: May 16, 2012
Published: July 6, 2012
© 2012 American Chemical Society
5894
dx.doi.org/10.1021/om300419w | Organometallics 2012, 31, 5894−5902

Organometallics
Article
Scheme 1. Preparation of Iron Norbornadiene Complexes According to Pettit22a (4a), Watson22b (4b), and Landesberg23 (4d)
employed the use of open-shell metal carbonyl complexes,16
species in which CO release is initiated by interaction with an
enzyme13f (enzyme-triggered or ET-CORMs), and the use of
photochemical methods to promote carbonyl dissociatio-
n.10e,11,13h,14b,c,f,g,i,j,15,20
A clear understanding of the mechanism of CO release for
any given structural type is required in order for structure−
activity relationships to be developed. A distinct advantage of
using transition-metal complexes as CO-RMs is that subtle
ligand alterations may be used to modulate the steric and
electronic properties of the metal center and ultimately the rate
and extent of CO release. However, the modification of a given
ligand framework must be informed by the fundamental steps
that govern the release process. Indeed, the parallels between
designing transition metals for catalytic and therapeutic
applications have recently been discussed.21
Thermal CO-Release Studies. Neither (η4-1,3-
cyclohexadiene)iron(0) tricarbonyl (1) nor [(η5-
cyclohexadienyl)iron(II)tricarbonyl][BF4] (2) exhibited any
CO release over the course of 2 h. This is consistent with
the (diene)iron tricarbonyl based ET-CORMs, which do not
release CO unless in the presence of an enzyme.13f In contrast,
(η4-norbornadiene)iron(0) tricarbonyl (4a) (Scheme 1) was
able to convert deoxy-Mb to Mb-CO with t1/2 = 49.8, 49.8, and
59.5 min for concentrations of 4a of 20, 40, and 60 μM,
respectively (Figure 1). The final concentration of Mb-CO
formed by the two lower concentrations of CO-RM indicates
that only one CO is liberated from the complex.
As our previous investigations into the chemistry of iron
tricarbonyl complexes containing 2-pyrone ligands12 have
demonstrated that substitution of the ligand backbone may
be used to modulate CO release and that the complexes
exhibited low toxicity, we have explored other tunable ligand
frameworks bound to this metal fragment. We now report the
CO-releasing properties of a number of iron tricarbonyl
complexes containing substituted norbornadiene ligands as
well as their chromium and molybdenum tetracarbonyl
analogues.
Figure 1. CO-release profile of 4a showing [Mb-CO] versus time.
DMSO was used as the addition solvent, and the CO-RM
concentration is indicated in the legend.
RESULTS AND DISCUSSION
■
The well-controlled CO-release profiles exhibited by 4a
prompted the exploration of the effects of different substituents
within the norbornadiene framework in order to modulate the
release rate and provide mechanistic insight into the release
process. Previous work has shown the importance of electron-
withdrawing substituents in accelerating CO-release rates,
which is consistent with a reduction in π back-bonding from
the metal to the carbonyl ligand and thus weaker M−C
bonds.13d Therefore, the inclusion of ester substituents onto the
norbornadiene backbone would be expected to increase π back-
donation to this ligand at the expense of the metal carbonyls,
thus weakening the metal−carbon bond. Indeed, the π-acidic
nature of a diester-substituted norbornadiene has been shown
to enhance catalytic activity in palladium-catalyzed cross-
coupling reactions.25 Two strategies were employed to achieve
this goal: first, ester groups were incorporated at the 2- and 3-
positions of the norbornadiene ligands, and second, ester and
alcohol groups were introduced at the 7-position of the ligand.
The complex containing ester substituents at both the 2- and
3-positions of the norbornadiene framework, 4b, was obtained
from the reaction of Fe2(CO)9 with diester 3b (Scheme 1).
The monoester-substituted norbornadiene ligand 3c was
prepared as a racemic mixture via a Suzuki coupling from the
corresponding bromo-substituted precursor using a palladium-
The aim of this study was to investigate the CO-releasing
properties of Fe(0), Cr(0), and Mo(0) complexes containing
norbornadiene and related ligands. A particular goal was to
establish the factors controlling the CO release from the metal.
As shown in Scheme 1, the preparation of the iron complexes is
generally conveniently performed from the reaction of the free
ligand with either Fe(CO)5 or Fe2(CO)9:22,23 the chromium
and molybdenum tetracarbonyl analogues may be prepared
from an analogous reaction using the corresponding
hexacarbonyl complexes as precursors.24
The CO-releasing properties of the complexes described in
this study were assessed with the aid of an assay based on
myoglobin. Treatment of an aqueous solution of deoxygenated
myoglobin (deoxy-Mb) with the complex in question results in
the subsequent capture of any CO released by the complex as
carbonmonoxymyoglobin (Mb-CO), and this process may be
quantified using the changes to the Q bands of the protein in
the UV/vis spectra. This allows for the rate and extent of CO
release to be determined. It is convenient to quote the half-life
for CO-release, t1/2, defined as the time taken for a solution
with CO-RM concentration of X μM to produces a
concentration of Mb-CO of X/2 μM, as a mechanism-
independent measure of the release rate.
5895
dx.doi.org/10.1021/om300419w | Organometallics 2012, 31, 5894−5902

Organometallics
Article
(II) precatalyst (Scheme 2) developed by Fairlamb and
Taylor.26 Reaction of 3c with Fe2(CO)9 results in the formation
of 4c, also as a racemic mixture (Scheme 1).
a
Scheme 2.
a
Legend: (i) PhB(OH)2, [PdBr(N-succinimidyl)(PPh3)2] (1 mol %),
THF, 60 °C; (ii) 2 M Na2CO3 80 °C, 2 h.
An examination of the IR spectra of 4b,c reveals that the
introduction of the electron-withdrawing substituents on the
norbornadiene ligand resulted in a shift in the highest frequency
CO band in comparison with the parent complex 4a (4a, 2028
cm−1; 4b, 2058 cm−1; 4c, 2046 cm−1). However, no CO release
was observed from either 4b or 4c in the myoglobin-based
assay.
The behavior of the complexes in which a substituent had
been incorporated at the 7-position of the norbornadiene
framework was further examined. The reaction of 7-
benzoyloxynorbornadiene (3d) with Fe(CO)5 allowed for the
formation of 4d.23 Reduction of 4d with PhMgCl led to the
formation of the alcohol derivative 4e (Scheme 3). In contrast
Figure 2. CO-release profile for complexes 4d (a) and 4e (b) showing
[Mb-CO] versus time. DMSO was used as the addition solvent, and
the CO-RM concentration is indicated in the legend.
carbon bonds, as σ bonding and changes to the polarization of
the C−O bond are important factors.27 In any case, the
complexes with the highest C−O stretching frequencies do not
release CO thermally. An examination of the 13C NMR
spectroscopic data for 4a,b provides insight into the nature of
the bonding between the metal and the norbornadiene ligand.
In the case of complex 4a the coordination shift of the
norbornadiene ligand (defined as the difference in chemical
shift between the free and coordinated ligands) is 105.8 ppm.
In the case of 4b, however, the ester-containing alkene has a far
greater coordination shift (116.9 ppm), although the shift for
the unsubstituted alkene is not as great.28
a
Scheme 3. Preparation of 4e according to Landesberg23
a
Legend: (i) PhMgCl, THF, 20 °C, 18 h.
to 4b,c 4d,e both released CO in the myoglobin-based assay.
However, as shown in Figure 2, the CO-release profiles for
both of these complexes exhibited some unusual behavior.
Taking complex 4e, for example, the release rate for the 60 μM
concentration of CO-RM release is slower than the rate when
the assay was performed with a CO-RM concentration of 40
μM. In addition, the 40 μM profile has a plateau in the
evolution of CO at approximately 7−17 min, followed by
resumption of CO release.
Mechanistic Considerations. The nontrivial behavior
exhibited by the iron norbornadiene complexes prompted a
series of studies to gain insight into the fundamental process
underpinning the CO-release mechanism. A consideration of
the spectroscopic data for complexes 4 provided some insight
into the ground-state properties of these species. Selected IR
and 13C NMR spectroscopic data for the complexes are
presented in Table 1. As detailed above, the IR spectroscopic
data for ester-containing complexes 4b,c are at a higher
frequency than the substituted derivatives. A traditional
interpretation of these data would be that there is less
metal−carbon back-bonding to these carbonyl ligands, due to
the electron-withdrawing nature of the ester-containing
norbornadiene ligand. However, recent studies on metal
carbonyl complexes have indicated that high CO stretching
frequencies may not necessarily be indicative of weak metal−
These data indicate that the ester-substituted alkene in the
norbornadiene ligands is more strongly bound to the metal
than in the equivalent unsubstituted complex, which points to a
mechanism of CO release that is triggered by the initial loss of
the norbornadiene ligand rather than CO under thermal
conditions.29
A further series of experiments were undertaken to probe the
unusual CO-release profile exhibited by these complexes. The
CO release exhibited by complex 4d was reassessed using a
modified myoglobin assay with added 3d. To this end, 0, 5, 10,
and 20 μM of ligand 3d were added to four different myoglobin
assay solutions prior to the addition of 4d (40 μM). The four
assays were run in parallel, and the amount of Mb-CO formed
was calculated in the standard manner. The results of this
experiment (Figure 3) show that the release rate of the CO-RM
is retarded in the presence of the free ligand, supporting the
supposition that the release of CO from this CO-RM is
controlled by dissociation of the norbornadiene ligand from the
complex.
It is therefore proposed that the key mechanistic step in the
thermal CO release from complexes 4 is dissociation of the
norbornadiene ligand. This is supported by the spectroscopic
data, which suggested that when norbornadiene is more
strongly bound, then CO release is slow, and vice versa. In
addition, this mechanistic proposal provides an explanation for
5896
dx.doi.org/10.1021/om300419w | Organometallics 2012, 31, 5894−5902

Organometallics
Article
a
Table 1.
IR data ν(CO)/cm−1
13C NMR data/ppm
complex
ν1
ν2
ν3
δf1
δf2
δc1
δc2
Δδ1
Δδ2
4a
4b
4c
4d
2028
2058
2046
2034
2032
1954
2003
1992
1963
1959
143.4
153.3
37.6
36.4
105.8
116.9
1976
1962
143.1
60.5
82.6
4e
a
Definitions: δf = frequency of uncoordinated alkene; δc = frequency of coordinated alkene; Δδ1 = δf1 − δc1; Δδ2 = δf2 − δc2.
activate the complex was explored. A solution of 4b (final
concentration 40 μM) was prepared and administered to the
Mb assay: as expected, only a trace amount of Mb-CO was
formed over the course of ca. 40 min. After this time the sample
was exposed to irradiation from an LED light source (2.4 W,
400 nm, irradiation for 2 min during each 5 min period), which
promoted the release of CO (Figure 4a) and saturation of the
Figure 3. CO-release profile of 4d (40 μM) with added free
norbornadiene ligand (3d, concentration of norbornadiene indicated
in legend) showing [Mb-CO] versus time. DMSO was used as the
addition solvent.
the unusual CO-release behavior exhibited by these complexes.
For example, in the case of 4d,e the most rapid release is when
the concentration of CO-RM is 40 μM and the rate and extent
of CO release by 4d is the slowest for the most concentrated
solution. If the CO-release process is controlled by the loss of
the norbornadiene, then as complete substitution at the metal
by a suitable donor such as DMSO occurs, the concentration of
the free ligand in solution will increase with time. This liberated
ligand would then be expected to inhibit further loss of
coordinated norbornadiene. Inhibition will be more rapid for
higher concentrations of CO-RMs, which explains the
unusually slow release of the more concentrated solutions of
4d,e. Although a full kinetic analysis of the CO-release profiles
has not been performed, the presence of free ligand inhibiting
norbornadiene dissociation may explain the stepwise nature of
the release profile of a number of these species. No evidence for
interaction of the free ligands 3a,d (40 μmol) with either
deoxy-Mb or Mb-CO was obtained under the conditions of the
myoglobin-based assay. Finally, it is important to note that the
CO-release behavior exhibited by the parent compound 4a
shows fewer deviations from the release profiles typically
observed than do either 4d or 4e. This may simply reflect the
lower solubility of unsubstituted norbornadiene in the aqueous
medium in comparison to the ester- or alcohol-substituted
compounds; thus, inhibition in the case of 4a will not be as
pronounced. It should also be noted that the half-life for CO
release in the case of 4a is slowest for the most concentrated
solution.
Figure 4. CO-release tests on 4b showing [Mb-CO] versus time: (a)
2.4 W, 400 nm irradiation, 2 min on, 3 min off, [4b] = 40 μM; (b) 2.4
W, 400 nm irradiation, 2 min on, 3 min off, [4b] = 10 μM. The
asterisk marks the start of irradiation.
assay within ca. 1 h of irradiation. An identical assay performed
on a solution of CO-RM (10 μM) resulted in a final Mb-CO
concentration of ca. 20 μM, indicating that each molecule was
releasing on average 2 equiv of CO (Figure 4b). These data
illustrate that 4b is a photochemically activated CO-releasing
molecule (photo-CO-RM), and the fact that two molecules of
CO are released as opposed to one for the thermally activated
analogues may point to a different mechanistic process (i.e.,
direct CO expulsion from the metal) occurring in this case.
The cytotoxicity of 4b in the absence of irradiation was
assessed with the aid of two assays. Both lactate dehydrogenase
(LDH) and Alamar Blue cell viability assays were performed on
RAW264.7 macrophages.12b,13a,b,d,14f,19 The LDH assays allow
Photochemical CO Release and Toxicity Study on 4b.
Given that no CO release was observed from 4b under thermal
conditions, the possibility of using photochemical methods to
5897
dx.doi.org/10.1021/om300419w | Organometallics 2012, 31, 5894−5902

Organometallics
Article
a
for an evaluation of the damage to a cell membrane, whereas
the Alamar Blue assay is an index of cell metabolism. The
results of both assays demonstrate that under the conditions
employed 4b is essentially nontoxic to the cells. No LDH is
released from the cells up to a 4b concentration of 140 μM
(Figure 5), indicating that no detectable damage to the cell wall
Scheme 4.
a
n
Legend: (i) Bu2O, hexane, reflux (5-Cr, 2 days; 5-Mo, 20 h).
of 0, 2.3, and 12.0 min at CO-RM concentrations of 20, 40, and
60 μM, and in DMSO, t1/2 values of <1 min at 20 and 40 μM
and 0.7 min at 60 μM.
An examination of the behavior exhibited by 5-Mo
demonstrated an important feature that must be considered
when evaluating fast-releasing CO-RMs. The stock solution of
CO-RM was prepared in EtOH and then added to six different
solutions of deoxy-Mb at intervals of ca. 10 s; the evolution of
Mb-CO was then monitored over the course of ca. 30 min. As
shown in Figure 7, the incubation time in the solvent before
Figure 5. LDH release from RAW264.7 cells treated for 24 h with
increasing concentrations of 4b. LDH is expressed as a percentage of
the total LDH released from Triton-treated cells.
was occurring. Furthermore, the addition of 4b up to
concentrations of 140 μM did not significantly affect cell
viability (Figure 6). The inherently low toxicity of 4b makes it
an attractive photo-CO-RM, as it is predicted that no adverse
side effects will occur in areas which are not exposed to
irradiation.
Figure 7. CO-release profile of 5-Mo (20 μM) showing [Mb-CO]
versus time: (a) EtOH as the addition solvent; (b) DMSO as the
addition solvent. The order of addition of the CO-RM to deoxy-Mb is
indicated in the legend and occurred at 10 s intervals.
Figure 6. Cell viability of RAW264.7 macrophages exposed to
increasing concentrations of 4b for 24 h. Viability is expressed as a
percentage relative to the control.
Chromium and Molybdenum Norbornadiene Com-
plexes. Given the diversity of CO-release profiles that are
exhibited by the iron complexes 4, the related chromium- and
molybdenum-containing complexes 5-Cr and 5-Mo were
prepared from the reaction of the corresponding metal
hexacarbonyl and norbornadiene (Scheme 4). Both 5-Cr and
5-Mo were found to convert deoxy-Mb to Mb-CO when added
as either EtOH or DMSO solutions to the assay, indicating that
both complexes are active CO-RMs. The half-lives of CO
release for 5-Cr were found to be 3.2, 4.2, and 2.8 min at CO-
RM concentrations of 20, 40, and 60 μM, respectively, in EtOH
and 0.8, 1.6, and 0.7 min at the same concentrations,
respectively, in DMSO. The molybdenum analogue was
found to release CO extremely rapidly: in EtOH, t1/2 values
addition to the protein had a profound effect on the release
profile. Longer incubation times resulted in lower conversion to
Mb-COsimilar observations were made when DMSO was
used as the delivery solvent, although the effect was not as
pronounced. These data demonstrate that 5-Mo undergoes a
rapid reaction on dissolution in EtOH or DMSO solutions
which triggers CO release. Hence, the longer the residence time
in the solvent before addition to deoxy-Mb, the more CO is
released to the surroundings and the less is sequestered as Mb-
CO. Similar incubation effects have been observed in the case
of the [Co2(CO)6(μ-alkyne)] complexes.19 In the case of these
cobalt complexes differences in CO-release profiles were
observed on incubation over a period of minutes, rather than
seconds. The results from both sets of complexes illustrate the
5898
dx.doi.org/10.1021/om300419w | Organometallics 2012, 31, 5894−5902

Organometallics
Article
importance of establishing the nature of any incubation effects
exhibited by any given CO-RM and that consistent residence
times before administration to deoxy-Mb are used.
The behavior of 5-Cr and 5-Mo in EtOH and DMSO
solutions was probed by in situ IR analysis using a
ReactIR instrument. IR spectra of 5-Cr recorded immediately
on dissolution in DMSO exhibited three bands at 2022, 1923,
and 1882 cm−1, similar to those observed in CH2Cl2 (2027,
1932, 1893 cm−1). Over the period of ca. 40 min these bands
decreased in intensity, to be replaced by a weak peak at 2077
cm−1 and a strong band at 1951 cm−1 (Figure 8). After addition
Figure 9. IR spectrum of 5-Mo in DMSO over time.
Figure 8. (a) IR spectrum of 5-Cr in DMSO over time. (b) Time
dependence of the bands at 1951 and 1925 cm−1.
Figure 10. IR spectra of (a) 5-Cr and (b) 5-Mo in EtOH recorded
over time.
of CH2Cl2 to this solution and washing with water, the resulting
organic layer was dried with MgSO4. Removal of the solvent in
vacuo afforded a yellow solid which exhibited a strong band at
1953 cm−1 and weak peaks at 2004 and 2080 cm−1. This species
has been observed previously in the reaction of [NEt4][CrBr-
(CO)5] with DMSO and was assigned to [Cr-
(CO)5(DMSO)].13c Unfortunately, the behavior of the
molybdenum analogue was far more complicated and mean-
ingful information could not be obtained. Immediately on
addition of DMSO a number of new bands were observed
(Figure 9) which could not be assigned; given that bands were
observed at frequencies as low as 1750 cm−1, it is proposed that
a number of CO ligands were substituted by DMSO and
complexes containing both terminal and bridging CO ligands
were formed. However, mechanistic information on this
process could not be obtained on this rapid reaction.
The IR spectrum of 5-Cr recorded on immediate dissociation
in EtOH exhibited three bands at ν 2029, 1936, and 1902 cm−1,
which are similar to those observed in CH2Cl2 solution. Over a
period of time the bands at 2029 and 1902 cm−1 decreased in
intensity and a new peak for Cr(CO)6 was observed at 1983
cm−1: the band at 1936 cm−1 does not alter in intensity,
suggesting that an intermediate species with a band coincident
with that of 5-Cr has been formed (Figure 10a). The
corresponding experiment with 5-Mo showed different
behavior, as the spectrum recorded on initial dissolution in
EtOH was markedly different from that observed in CH2Cl2
solution (EtOH, 1942 cm−1; CH2Cl2, 2041, 1952, 1888
cm−1)however, on standing, this band decreased in intensity
to be replaced by a band due to Mo(CO)6 at 1985 cm−1
(Figure 10b). Repeating the experiment with 5-Mo in the
presence of 100 equiv of free norbornadiene resulted in a far
slower degradation. In this case, the initial IR spectra recorded
in EtOH closely matched those acquired in CH2Cl2 solution
and a slower conversion to the intermediate with a band at
1942 cm−1 and subsequently Mo(CO)6 occurred.
These in situ IR spectroscopic studies demonstrate that both
5-Cr and 5-Mo are intrinsically unstable in both EtOH and
DMSO and, as would be expected on the basis of the CO-
release profiles obtained from the myoglobin assay, the
molybdenum complex degrades more rapidly than its
chromium congener. As is the case with related studies on
other group 6 metal carbonyls such as [NEt4][CrCl(CO)5], the
formation of the corresponding hexacarbonyl species and
[Cr(CO)5(DMSO)] is observed,13c indicating that in the
absence of myoglobin the group 6 metal acts as an effective CO
scavenger. Given that Mb-CO is formed in the appropriate
assays, the binding constant between the group 6 metal and the
5899
dx.doi.org/10.1021/om300419w | Organometallics 2012, 31, 5894−5902

Organometallics
Article
Coupling constants are reported in hertz (Hz), and spin multiplicities
are indicated by the following symbols: s (singlet), d (doublet), t
(triplet), q (quartet), m (multiplet), br (broad). 13C spectra were
recorded with full proton decoupling. Photoinitiated carbon monoxide
release was carried out using a 5 W 400 nm LED attached to the top of
a UV cuvette directly above the solution, drawing 2.4 W of power.
Tricarbonyl(η4-1,3-cyclohexadiene)iron(0) (1),31 tricarbonyl(η4-
1,3-cyclohexadiene)iron(II) tetrafluoroborate (2),31 methyl 3-
bromobicyclo[2.2.1]hepta-2,5-diene-2-carboxylate (3f),32 tricarbonyl-
(η4-bicyclo[2.2.1]hepta-2,5-diene)iron(0) (4a),22a tricarbonyl(η4-
dimethylbicyclo[2.2.1]hepta-2,5-diene-2,3-dicarboxylate)iron(0)
(4b),22b tricarbonyl(η4-7-benzoyloxybicyclo[2.2.1]hepta-2,5-diene)-
iron(0) (4d),23 Tricarbonyl(η4-7-bicyclo[2.2.1]hepta-2,5-dieneol)iron-
(0) (4e),23 tetracarbonyl(η4-bicyclo[2.2.1]hepta-2,5-diene)chromium-
(0) (5-Cr),24 and tetracarbonyl(η4-bicyclo[2.2.1]hepta-2,5-diene)-
molybdenum(0) (5-Mo)24 were prepared by literature methods.
Additional preparative, spectroscopic, and analytical data for these
known compounds and organic precursors are provided in the
Supporting Information.
liberated CO must be lower than that of the iron in myoglobin
(ca. 107 M−1).30 Although norbornadiene dissociation does
evidently occur in these group 6 systems, it is less clear if this is
the rate-controlling step, as appears to be the case for the iron-
containing analogues.
CONCLUSIONS
■
The nature of the norbornadiene ligand may dramatically
modulate the nature of the CO-release process in a series of
Fe(0) complexes. Indeed, the incorporation of ester sub-
stituents at the 2- and/or 3-positions of the norbornadiene
framework provides a sufficient perturbation to inhibit thermal
CO release. The spectroscopic and mechanistic data indicate
that CO release is controlled by dissociation of norbornadiene
rather than direct CO loss from the metal center. The
inhibition of thermal release in the ester-substituted com-
pounds provided the opportunity to develop a well-controlled
photo-CO-RM. A particularly beneficial feature is that the
parent CO-RM does not display any toxicity, ensuring that side
effects in areas which are not irradiated could potentially be
minimized. Therefore, the ability to modulate the CO-release
behavior of this series of molecules is an attractive feature,
indicating that higher-yielding synthetic routes should be
pursued.
rac-Methyl 4-Phenylbicyclo[2.2.1]hepta-2,5-diene-2-carbox-
ylate (3c). Compound 3f (0.5 mmol, 115 mg), phenylboronic acid
(0.5 mmol, 61 mg), and bromo(N-succinimidyl)bis-
(triphenylphosphine)palladium(II) (5 μmol, 1 mol %, 4 mg)26 were
added to a Schlenk tube under N2, dissolved in THF (2 mL), and
heated to 60 °C. Aqueous Na2CO3 (2 M, 1 mL) was added, and the
reaction mixture was heated at 80 °C for a further 2 h. The reaction
mixture was cooled to room temperature, extracted with Et2O, and
concentrated in vacuo. The residue was purified by chromatography
on silica gel using hexane/Et2O (9/1 v/v), giving 3c as a colorless oil
The behavior of 5-Mo demonstrates that considerable care
must be taken when assessing complexes which undergo
extremely rapid active solvent-induced CO release. In these
cases the incubation time in the delivery solvent may result in
artifacts in the myoglobin-based assay being obtained, mirroring
previous observations on [Co2(CO)6(μ-alkyne)] complexes.19
In the most extreme case, no CO release would be detected if
the residence time in the delivery solvent is too long in
comparison to the rate of release. It is suggested that for
complexes which do undergo such rapid release the residence
time should be minimized and varied to establish the effects of
this variable.
Under thermal conditions for both complexes 4 (Fe) and 5
(Cr and Mo) the active CO-releasing molecules are generated
in the reaction medium through either loss of norbornadiene
and/or interaction with the solvent. This is an important factor
to consider when deriving structure−activity relationships for
CO-RMs and indeed other metallo-drugs and may be
compared to activation of a precatalyst, which is frequently
required in homogeneous catalysis.
1
(67 mg, 59%). H NMR (270 MHz, CDCl3): δ 2.07 (app dt, J = 6.6,
1.6 Hz, 1H), 2.26 (app dt, J = 6.6, 1.6 Hz, 1H), 3.68 (s, 3H), 3.86 (m,
1H), 4.07 (m, 1H), 6.92 (ddd, J = 5.0, 3.0, 0.8, 1H), 6.99 (ddd, J = 5.0,
3.0, 0.7, 1H), 7.31−7.41 (m, 3H), 7.52−7.55 (m, 1H). 13C NMR (68
MHz, CDCl3): δ 51.2, 58.5, 70.6, 127.7, 128.6, 135.6, 138.9, 140.8,
143.6, 165.9, 166.8. HR (ESI-MS): m/z calcd for C15H15O2 227.1067,
found 227.1068 [M + H]+.
rac-Tricarbonyl(η4-methyl 4-phenylbicyclo[2.2.1]hepta-2,5-
diene-2-carboxylate)iron(0) (4c). To a solution of Fe2(CO)9
(1.39 mmol, 507 mg) in THF (15 mL) under N2 was added 3c
(0.93 mmol, 210 mg). The reaction mixture was heated at reflux for 18
h. The solution was concentrated in vacuo and the residues were
purified by chromatography on silica gel using hexane/Et2O (19/1 v/
v) to give 4c as a yellow solid (16.1 mg, 5%). Mp: 53−54 °C. 1H NMR
(500 MHz, CD2Cl2): δ 1.32 (app dt, J = 9.6, 1.6 Hz, 1H), 1.49 (app dt,
J = 9.6, 1.6 Hz, 1H), 3.60 (m, 1H), 3.65 (s, 3H), 3.83 (m, 1H), 3.95
(app t, J = 3.7 Hz, 1H), 4.11 (app t, J = 3.7 Hz, 1H), 7.22 (t, J = 7.4
Hz, 1H), 7.30 (t, J = 7.4 Hz, 2H), 7.57 (d, J = 7.4 Hz, 2H). 13C NMR
(126 MHz, CD2Cl2): δ 34.7, 45.5, 51.0, 53.3, 55.0, 55.1, 55.2, 56.1,
126.3, 127.5, 130.7, 141.5, 173.6, 212.9. IR (CH2Cl2): ν 2046 (s), 1992
(s), 1962 (s) cm−1. HR(ESI-MS): m/z calculated for C18H15FeO5
367.0257; found 367.0257 [M + H]+.
CO-release tests were performed as described previously.10c,13c,17,19
Details of the Almar blue and LDH assays are provided in the
Supporting Information.
EXPERIMENTAL SECTION
■
General Details. THF, benzene, and diethyl ether were dried over
sodium-benzophenone ketyl and were distilled prior to use. DCM was
purified with aid of an Innovative Technology anhydrous solvent
engineering system. TLC analysis was carried out on Merck TLC
aluminum sheets (silica gel 60 F254), and flash chromatography was
run on silica gel (Fluka). Compounds were visualized using UV light
(254 nm) and a basic aqueous solution of potassium permanganate.
Solution cell infrared spectroscopy was undertaken using a Thermo
Nicolet Avatar 370 FT-IR spectrometer (2 cm−1 resolution). Reactions
were monitored using a Mettler Toledo ReactIR ic10 instrument with
K6 Conduit, SiComp (silicon) probe, and MCT detector: resolution 4
cm−1, range 4000−650 cm−1, and gain adjustment at 1×. Ultraviolet−
visible spectra were recorded using a JASCO V-560 (bandwidth 1 nm).
1H and 13C NMR spectra were obtained, in the appropriate deuterated
solvent indicated, using Bruker AMX500, JEOL EXC400, and JEOL
EX270 spectrometers at 500, 400, and 270 MHz for 1H and 126, 100,
and 68 MHz for 13C, respectively. Chemical shifts are reported in parts
per million (δ) downfield from an internal tetramethylsilane reference.
ASSOCIATED CONTENT
■
S
* Supporting Information
Text and figures giving preparative and spectroscopic details for
known compounds and experimental details for the Alamar
blue and LDH assays. This material is available free of charge
via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
■
Corresponding Author
*Fax: +44 (0)1904 432516. Tel: +44 (0)1904 432534. E-mail:
ian.fairlamb@york.ac.uk (I.J.S.F.); jason.lynam@york.ac.uk
(J.M.L.).
5900
dx.doi.org/10.1021/om300419w | Organometallics 2012, 31, 5894−5902

Organometallics
Article
(13) (a) Sawle, P.; Hammad, J.; Fairlamb, I. J. S.; Moulton, B.;
O’Brien, C. T.; Lynam, J. M.; Duhme-Klair, A. K.; Foresti, R.;
Motterlini, R. J. Pharmacol. Exp. Ther. 2006, 318, 403−410.
(b) Scapens, D.; Adams, H.; Johnson, T. R.; Mann, B. E.; Sawle, P.;
Aqil, R.; Perrior, T.; Motterlini, R. Dalton Trans. 2007, 4962−4973.
(c) Zhang, W. Q.; Atkin, A. J.; Thatcher, R. J.; Whitwood, A. C.;
Fairlamb, I. J. S.; Lynam, J. M. Dalton Trans. 2009, 4351−4358.
(d) Hewison, L.; Crook, S. H.; Johnson, T. R.; Mann, B. E.; Adams,
H.; Plant, S. E.; Sawle, P.; Motterlini, R. Dalton Trans. 2010, 39,
8967−8975. (e) Gonzalez, M. A.; Fry, N. L.; Burt, R.; Davda, R.;
Hobbs, A.; Mascharak, P. K. Inorg. Chem. 2011, 50, 3127−3134.
(f) Romanski, S.; Kraus, B.; Schatzschneider, U.; Neudorfl, J. M.;
Amslinger, S.; Schmalz, H. G. Angew. Chem., Int. Ed. 2011, 50, 2392−
2396. (g) Jackson, C. S.; Schmitt, S.; Dou, Q. P.; Kodanko, J. J. Inorg.
Chem. 2011, 50, 5336−5338. (h) Kretschmer, R.; Gessner, G.; Gorls,
H.; Heinemann, S. H.; Westerhausen, M. J. Inorg. Biochem 2011, 105,
6−9.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We thank the EPSRC (Grant EP/E005152/1, Ph.D. student-
ship to A.J.A.), the BBSRC (DTA studentship to J.S.W.), and
the University of York (purchase of the ReactIR instrument)
for funding. The assistance of Dr. Adrian Mountford and Mr.
David Sanin (Department of Biology, University of York) with
the Alamar blue and LDH assays and the contribution of Dr.
James Moir (Department of Biology, University of York) to the
development of research into CO-RMs at York are gratefully
acknowledged.
REFERENCES
■
(14) (a) Nobre, L. S.; Seixas, J. D.; Romao, C. C.; Saraiva, L. M.
Antimicrob. Agents Chemother. 2007, 51, 4303−4307. (b) Niesel, J.;
Pinto, A.; N’Dongo, H. W. P.; Merz, K.; Ott, I.; Gust, R.;
Schatzschneider, U. Chem. Commun. 2008, 1798−1800. (c) Kunz, P.
C.; Huber, W.; Rojas, A.; Schatzschneider, U.; Spingler, B. Eur. J. Inorg.
Chem. 2009, 5358−5366. (d) Pfeiffer, H.; Rojas, A.; Niesel, J.;
(1) (a) Johnson, T. R.; Mann, B. E.; Clark, J. E.; Foresti, R.; Green,
C. J.; Motterlini, R. Angew. Chem., Int. Ed. 2003, 42, 3722−3729.
(b) Motterlini, R.; Mann, B. E.; Foresti, R. Expert Opin. Invest. Drugs
2005, 14, 1305−1318. (c) Mann, B. E.; Motterlini, R. Chem. Commun.
2007, 4197−4208.
(2) Chapman, J. T.; Otterbein, L. E.; Elias, J. A.; Choi, A. M. K. Am. J.
Physiol.-Lung Cell. Mol. Physiol. 2001, 281, L209−L216.
(3) Amersi, F.; Shen, X. D.; Anselmo, D.; Melinek, J.; Iyer, S.;
Southard, D. J.; Katori, M.; Volk, H. D.; Busuttil, R. W.; Buelow, R.;
Kupiec-Weglinski, J. W. Hepatology 2002, 35, 815−823.
Schatzschneider, U. Dalton Trans. 2009, 4292−4298. (e) Bruckmann,
̈
N. E.; Wahl, M.; Reiss, G. J.; Kohns, M.; Watjen, W.; Kunz, P. C. Eur.
̈
J. Inorg. Chem. 2011, 4571−4577. (f) Crook, S. H.; Mann, B. E.;
Meijer, A. J. H. M.; Adams, H.; Sawle, P.; Scapens, D.; Motterlini, R.
Dalton Trans. 2011, 40, 4230−4235. (g) Gonzalez, M. A.; Yim, M. A.;
Cheng, S.; Moyes, A.; Hobbs, A. J.; Mascharak, P. K. Inorg. Chem.
(4) (a) Motterlini, R.; Gonzales, A.; Foresti, R.; Clark, J. E.; Green, C.
J.; Winslow, R. M. Circ. Res. 1998, 83, 568−577. (b) Sammut, I. A.;
Foresti, R.; Clark, J. E.; Exon, D. J.; Vesely, M. J. J.; Sarathchandra, P.;
Green, C. J.; Motterlini, R. Br. J. Pharmacol. 1998, 125, 1437−1444.
(5) (a) Chin, B. Y.; Otterbein, L. E. Curr. Opin. Pharmacol. 2009, 9,
490−500. (b) Chung, S. W.; Liu, X.; Macias, A. A.; Baron, R. M.;
Perrella, M. A. J. Clin. Invest. 2008, 118, 239−247.
(6) Sato, K.; Balla, J.; Otterbein, L.; Smith, R. N.; Brouard, S.; Lin, Y.;
Csizmadia, E.; Sevigny, J.; Robson, S. C.; Vercellotti, G.; Choi, A. M.;
Bach, F. H.; Soares, M. P. J. Immunol. 2001, 166, 4185−4194.
(7) Pamplona, A.; Ferreira, A.; Balla, J.; Jeney, V.; Balla, G.;
Epiphanio, S.; Chora, A.; Rodrigues, C. D.; Gregoire, I. P.; Cunha-
Rodrigues, M.; Portugal, S.; Soares, M. P.; Mota, M. M.. Nat. Med.
2007, 13, 703−710.
(8) (a) Mann, B. E. Carbon Monoxide: An Essential Signalling
Molecule. In Medicinal Organometallic Chemistry; Springer-Verlag:
Berlin, 2010; Topics in Organometallic Chemistry 32, pp 247−285.
(b) Romao, C. C.; Blattler, W. A.; Seixas, J. D.; Bernardes, G. J. L.
Chem. Soc. Rev. 2012, 41, 3571−3583.
(9) (a) Motterlini, R.; Sawle, P.; Hammad, J.; Bains, S.; Alberto, R.;
Foresti, R.; Green, C. J. FASEB J. 2005, 19, 284−6. (b) Alberto, R.;
Motterlini, R. Dalton Trans. 2007, 1651−1660. (c) Pitchumony, T. S.;
Spingler, B.; Motterlini, R.; Alberto, R. Org. Biomol. Chem. 2010, 8,
4849−4854.
(10) (a) Foresti, R.; Hammad, J.; Clark, J. E.; Johnson, T. R.; Mann,
B. E.; Friebe, A.; Green, C. J.; Motterlini, R. Br. J. Pharmacol. 2004,
142, 453−460. (b) Johnson, T. R.; Mann, B. E.; Teasdale, I. P.; Adams,
H.; Foresti, R.; Green, C. J.; Motterlini, R. Dalton Trans. 2007, 1500−
1508. (c) Zhang, W.-Q.; Whitwood, A. C.; Fairlamb, I. J. S.; Lynam, J.
M. Inorg. Chem. 2010, 49, 8941−8952. (d) Hewison, L.; Johnson, T.
R.; Mann, B. E.; Meijer, A. J. H. M.; Sawle, P.; Motterlini, R. Dalton
Trans. 2011, 40, 8328−8334. (e) Velasquez, V. P. L.; Jazzazi, T. M. A.;
Malassa, A.; Goerls, H.; Gessner, G.; Heinemann, S. H.; Westerhausen,
M. Eur. J. Inorg. Chem. 2012, 1072−1078.
(11) Zhang, W.-Q.; Atkin, A. J.; Fairlamb, I. J. S.; Whitwood, A. C.;
Lynam, J. M. Organometallics 2011, 30, 4643−4654.
(12) (a) Fairlamb, I. J. S.; Duhme-Klair, A. K.; Lynam, J. M.;
Moulton, B. E.; O’Brien, C. T.; Sawle, P.; Hammad, J.; Motterlini, R.
Bioorg. Med. Chem. Lett. 2006, 16, 995−998. (b) Fairlamb, I. J. S.;
Lynam, J. M.; Moulton, B. E.; Taylor, I. E.; Duhme-Klair, A. K.; Sawle,
P.; Motterlini, R. Dalton Trans. 2007, 3603−3605.
2012, 51, 601−608. (h) Dordelmann, G.; Pfeiffer, H.; Birkner, A.;
̈
Schatzschneider, U. Inorg. Chem. 2011, 50, 4362−4367. (i) Huber, W.;
Linder, R.; Niesel, J.; Schatzschneider, U.; Spingler, B.; Kunz, P. C.
Eur. J. Inorg. Chem. 2012, 3140−3146. (j) Mohr, F.; Niesel, J.;
Schatzschneider, U.; Lehmann, C. W. Z. Anorg. Allg. Chem. 2012, 638,
543−546.
(15) Rimmer, R. D.; Richter, H.; Ford, P. C. Inorg. Chem. 2010, 49,
1180−1185.
(16) (a) Zobi, F.; Degonda, A.; Schaub, M. C.; Bogdanova, A. Y.
Inorg. Chem. 2010, 49, 7313−7322. (b) Zobi, F.; Blacque, O. Dalton
Trans. 2011, 40, 4994−5001. (c) Zobi, F.; Blacque, O.; Jacobs, R. A.;
Schaub, M. C.; Bogdanova, A. Y. Dalton Trans. 2012, 41, 370−378.
(17) Atkin, A. J.; Lynam, J. M.; Moulton, B. E.; Sawle, P.; Motterlini,
R.; Boyle, N. M.; Pryce, M. T.; Fairlamb, I. J. S. Dalton Trans. 2011,
5755−5761.
(18) Bikiel, D. E.; Solveyra, E. G.; Di Salvo, F.; Milagre, H. M. S.;
Eberlin, M. N.; Correa, R. S.; Ellena, J.; Estrin, D. A.; Doctorovich, F.
Inorg. Chem. 2011, 50, 2334−2345.
(19) Atkin, A. J.; Williams, S.; Sawle, P.; Motterlini, R.; Lynam, J. M.;
Fairlamb, I. J. S. Dalton Trans. 2009, 3653−3656.
(20) (a) Crespy, D.; Landfester, K.; Schubert, U. S.; Schiller, A.
Chem. Commun. 2010, 46, 6651−6662. (b) Schatzschneider, U. Eur. J.
Inorg. Chem. 2010, 1451−1467. (c) Jackson, C. S.; Schmitt, S.; Dou,
Q. P.; Kodanko, J. J. Inorg. Chem. 2011, 50, 5336−5338.
(d) Schatzschneider, U. Inorg. Chim. Acta 2011, 374, 19−23.
(e) Berends, H.-M.; Kurz, P. Inorg. Chim. Acta 2012, 380, 141−147.
(21) Noffke, A. L.; Habtemariam, A.; Pizarro, A. M.; Sadler, P. J.
Chem. Commun. 2012, 48, 5219−5246.
(22) (a) Pettit, R. J. Am. Chem. Soc. 1959, 81, 1266−1266.
(b) Watson, W. H.; Nagl, A.; Kashyap, R. P.; Marchand, A. P.;
Dave, P. R. Acta Crystallogr., Sect. C 1990, 46, 24−27.
(23) Landesberg, J. M.; Sieczkowski, J. J. Am. Chem. Soc. 1971, 93,
972−980.
(24) Bennett, M. A.; Pratt, L.; Wilkinson, G. J. Chem. Soc. 1961,
2037−2044.
(25) Liu, Q.; Duan, H.; Luo, X.; Tang, Y.; Li, G.; Huang, R.; Lei, A.
Adv. Synth. Catal. 2008, 350, 1349−1354.
(26) Burns, M. J.; Fairlamb, I. J. S.; Kapdi, A. R.; Sehnal, P.; Taylor,
R. J. K. Org. Lett. 2007, 9, 5397−5400.
5901
dx.doi.org/10.1021/om300419w | Organometallics 2012, 31, 5894−5902

Organometallics
Article
(27) (a) Frenking, G.; Frohlich, N. Chem. Rev. 2000, 100, 717−774.
̈
(b) Zobi, F. Inorg. Chem. 2009, 48, 10845−10855. (c) Zobi, F. Inorg.
Chem. 2010, 49, 10370−10377.
(28) The coordination shifts for complexes 4c,d could not be
determined, as precise assignments for each alkene carbon in both the
free and coordinated ligands could not be made.
(29) In a recent publication it has been shown that the sulfur
compounds in the myoglobin assay may trigger CO release from CO-
RM-3. In the case of the photochemically activated CO-RMs this is
clearly not the case, as no CO is released in the dark and the group 6
norbornadiene complexes release CO in pure EtOH or DMSO
solution, again indicating that the process is not intitiated by the
conditions of the assay. See: McLean, S.; Mann, B. E.; Poole, R. K.
Anal. Biochem. 2012, 427, 36−40.
(30) Balasubramanian, S.; Lambright, D. G.; Simmons, J. H.; Gill, S.
J.; Boxer, S. G. Biochemistry 1994, 33, 8355−8360.
(31) Knolker, H. J.; Baum, G.; Foitzik, N.; Goesmann, H.; Gonser,
P.; Jones, P. G.; Rottele, H. Eur. J. Inorg. Chem. 1998, 993−1007.
(32) Yoo, W.-J.; Tsui, G. C.; Tam, W. Eur. J. Org. Chem. 2005, 2005,
1044−1051.
5902
dx.doi.org/10.1021/om300419w | Organometallics 2012, 31, 5894−5902