Photoinduced Homolysis of Alkyl-Cobalt(III) Bonds in a Cyclodextrin Cage

Article abstract of DOI:10.1002/ejic.201600208

Photodecomposition of methyl- and ethyl-Co^III complexes of meso-tetrakis(4-sulfonatophenyl)porphyrin (CH₃- and C₂H₅-Co^IIITPPSs) was used as a reaction probe to study the cage effect of cyclodextrin capsules formed by two per-O-methylated β-cyclodextrin (TMe-β-CD) molecules and their covalently linked dimer, Ph2CD. The photodecomposition of CH₃-Co^IIITPPS under aerobic conditions was markedly suppressed in the presence of TMe-β-CD and Ph2CD, while C₂H₅-Co^IIITPPS was less affected. Alkyl-Co^IIITPPS formed two types of inclusion complex with Ph2CD, the alkyl groups in Type 1 being located at the opposite side of the phenyl linker of Ph2CD and those in Type 2 being located at the same side. The photodecomposition of C₂H₅-Co^IIITPPS in Type 1 proceeded via an ethylperoxo complex, while that in Type 2 occurred via a radical pair generated in a narrow, rigid cage to form ethylene and Co^IITPPS.

Full text of DOI:10.1002/ejic.201600208

DOI: 10.1002/ejic.201600208
Communication
Host–Guest Chemistry
Photoinduced Homolysis of Alkyl–Cobalt(III) Bonds in a
Cyclodextrin Cage
Kohei Imabeppu,[a] Hiroyuki Kuwano,[a] Eriko Yutani,[a] Hiroaki Kitagishi,[a] and Koji Kano*[a]
Abstract: Photodecomposition of methyl– and ethyl–CoIII com- affected. Alkyl–CoIIITPPS formed two types of inclusion complex
plexes of meso-tetrakis(4-sulfonatophenyl)porphyrin (CH3– and
C2H5–CoIIITPPSs) was used as a reaction probe to study the cage
effect of cyclodextrin capsules formed by two per-O-methylated
ꢀ-cyclodextrin (TMe-ꢀ-CD) molecules and their covalently
linked dimer, Ph2CD. The photodecomposition of CH3–CoIIITPPS
under aerobic conditions was markedly suppressed in the pres-
ence of TMe-ꢀ-CD and Ph2CD, while C2H5–CoIIITPPS was less
with Ph2CD, the alkyl groups in Type 1 being located at the
opposite side of the phenyl linker of Ph2CD and those in Type 2
being located at the same side. The photodecomposition of
C2H5–CoIIITPPS in Type 1 proceeded via an ethylperoxo com-
plex, while that in Type 2 occurred via a radical pair generated
in a narrow, rigid cage to form ethylene and CoIITPPS.
Introduction
alkyl radical and a CoII complex, which, at first, exists in a sol-
vent cage. The alkyl radical, when it escaped from the solvent
cage, would easily react with O2 under aerobic conditions. This
reaction would compete with the recombination reaction
within the cage, to reform the original alkyl–CoIII bond. There-
fore, photoinduced homolysis of the alkyl–CoIII bond is ex-
pected to be used as a probe reaction to investigate the “cage
effect”. Previously, we found that heptakis(2,3,6-tri-O-methyl)-ꢀ-
cyclodextrin (TMe-ꢀ-CD) formed an extremely stable 2:1 inclu-
sion complex with 5,10,15,20-tetrakis(4-sulfonatophenyl)por-
phyrin (TPPS) in aqueous solution.[8] Two sulfonatophenyl
groups at the 5- and 15-positions of TPPS were included with
the TMe-ꢀ-CD molecules to form a trans-type inclusion com-
plex.[8b] The unusually strong interaction between TMe-ꢀ-CD
and TPPS was thought to be derived from the induced fit-type
complexation, due to the flexible nature of TMe-ꢀ-CD.[9] The
FeIII complex of TPPS (FeIIITPPS) also formed a stable 1:2 inclu-
sion complex with TMe-ꢀ-CD (Figure 1).[10] FeIII in the FeIIITPPS/
TMe-ꢀ-CD complex was coordinated by various inorganic
anions that did not bind to FeIIITPPS in aqueous solution with-
out TMe-ꢀ-CD, indicating that the iron center resided in a
hydrophobic environment formed by two cyclodextrin units.
We also found that hemoCD1 (Figure 1) strongly binds molec-
ular oxygen even in aqueous solution;[11] though, in general, an
O2–FeII bond immediately cleaves in aqueous solution due to
H2O-induced autoxidation yielding FeIII and a superoxide ion.[12]
The binding of O2 with hemoCD1 suggested that a H2O mol-
ecule hardly penetrated into the cage formed by two per-O-
methylated ꢀ-cyclodextrin units of O2-bound hemoCD1. We ex-
pected the photoreactions of alkyl–CoIIITPPS/TMe-ꢀ-CD com-
plexes to provide the dynamic aspects of the cage formed by
two TMe-ꢀ-CD molecules.
Fundamental chemical studies on vitamin B12 started in the
early 1960s. In 1961, Lenhert and Hodgkin determined the X-
ray structure of 5,6-dimethylbenzimidazolylcobamide (adenos-
ylcobalamin), which confirmed the existence of an alkyl–CoIII
organometallic bond in the vitamin B12 coenzyme.[1] One year
later, methylcobalamin was synthesized from the reaction of
NaBH4-reduced vitamin B12 with CH3I[2] and was proposed as
an intermediate in the reaction of methionine synthase.[3] In
1964, methylcobalamin was first isolated from a natural source
material.[4] Dolphin et al. reported the photolabile nature of
methylcobalamin, which produced formaldehyde and hydroxo-
cobalamin upon photoirradiation under aerobic conditions.[5]
Under anaerobic conditions (10–3 Torr), ethane (61 %), methane
(34.5 %), ethylene (4.4 %), and vitamin B12r were afforded as
photoproducts; however, no photoreaction proceeded under
high vacuum (10–6 Torr). In the photolysis of ethylcobalamin,
the products were acetaldehyde, acetic acid, and hydroxocobal-
amin under aerobic conditions, while ethylene (92 %), ethane,
and butane were produced under anaerobic conditions
(10–3 Torr).[5b] The photoproducts strongly suggested homolysis
of the alkyl–CoIII bonds, yielding alkyl radicals, which react read-
ily with O2, and vitamin B12r
bonds were found to have photoreactivity similar to alkylcobal-
amins in cobalt complexes of dimethylglyoxime and por-
phyrins.[7]
It has been widely known that the photoinduced homolysis
of alkyl–CoIII bonds produces a radical pair, composed of an
[6]
.
In the mid-1960s, alkyl–CoIII
[a] Department of Molecular Chemistry and Biochemistry,
Doshisha University
Kyotanabe, Kyoto 610-0321, Japan
E-mail: kkano@mail.doshisha.ac.jp
http://www1.doshisha.ac.jp/~kkano/
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/ejic.201600208.
Results and Discussion
No studies on the photoreactions of CH3– and C2H5–CoIII-
TPPSs in aqueous solution have been reported.
Eur. J. Inorg. Chem. 2016, 1784–1789
1784
© 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communication
Figure 1. Structures of the supramolecular complexes appearing in this paper.
The photoirradiation of CH3–CoIIITPPS (3.0 μ
rated 0.05
phosphate buffer at pH 7.0 caused the rapid disap- peroxo–CoIIITPPS. Alkylperoxo–CoIII complexes are well-known
M) in air-satu- shown in Scheme 1, where the key intermediate is methyl-
M
pearance of absorption bands centered at 417 and 541 nm, due species.[15] It is highly unlikely that very unstable methyl radi-
to consumption of the starting material, and the appearance of cals could diffuse in aqueous media to couple with each other
new bands at 425 and 539 nm resulting from the formation of and yield ethane. Accordingly, the collision of two methylper-
(H2O)2CoIIITPPS (pKa1 = 8.1, pKa2 = 12.1)[13] with the isosbestic oxo complexes was assumed to be responsible for the forma-
points as shown in Figure 2a. The rapid photodecomposition tion of ethane. Homolysis of the methylperoxo complex in a
of C2H5–CoIIITPPS (λmax = 417 and 541 nm) was also observed solvent cage affords formaldehyde (methanediol) and HO–
(Figure 2b). Products were analyzed for the photolyzed solution CoIIITPPS. The homolysis of alkylperoxo–CoIII complexes has
of 13CH3–CoIIITPPS (1 m
M
) in D2O in an NMR sample tube. Fig- been reported.[15a,15c,16] A somewhat tentative mechanism for
1
ure 3 shows the H NMR spectra of 13CH3–CoIIITPPS before and methanol formation is suggested in Scheme 1.
after photoirradiation (60 min). The 13CH3 signal at δ = –5.03
(d, J = 120 Hz) ppm disappeared, while peaks attributed to
ethane and methanol appeared at δ = 0.82 (d, J = 120 Hz) and
3.355 (d, J = 145 Hz) ppm,[14] respectively, upon irradiation. In
addition, a correlation peak was observed in the heteronuclear
multiple-quantum correlation (HMQC) spectrum of the photo-
lyzed sample between signals at δ = 4.82 (1H) and 81.49 (13C)
ppm, indicating the formation of methanediol (Figure S1 in the
Supporting Information). The signal intensities suggested that
the relative yields of the products were in the order: methane-
diol (formaldehyde) > methanol >> ethane. If the formation of
ethane was due to the coupling of the methyl radicals that had
escaped from the solvent cage, such a coupling reaction should
occur even under anaerobic conditions; however, no photo-
decomposition of CH3–CoIIITPPS occurred under an Ar atmos-
phere. Therefore, molecular oxygen was essential for promoting
the reactions. Thus, we proposed a reaction mechanism,
Scheme 1. Plausible mechanism for the photodecomposition of CH3–
CoIIITPPS.
Eur. J. Inorg. Chem. 2016, 1784–1789
www.eurjic.org
1785
© 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communication
Figure 2. Photodecomposition of 3 μ
(a,b) and the presence (c,d) of 7.2 μ
M
CH3–CoIIITPPS (a, c) and C2H5–CoIIITPPS (b, d) in aerobic 0.05
M
phosphate buffer at pH 7.0 and 25 °C in the absence
M
TMe-ꢀ-CD. The effects of TMe-ꢀ-CD on the photodecomposition rates are depicted in (e) and (f).
The products of the photoreaction (30 min) of C2H5–CoIII- TPPS/TMe-ꢀ-CD was difficult because of serious overlap of the
1
TPPS (2 m
M
) in aerobic D2O were analyzed by H NMR spectro- weak product signals with the intense signals of TMe-ꢀ-CD (Fig-
scopy, which indicated the formation of ethanol [δ = 1.19 (t, J = ure S7). In addition, some product signals were broad because
7.0 Hz) and 3.65 (q) ppm], 1,1-ethanediol [δ = 1.33 (d, J = 6.0 Hz) of diamagnetic impurity (probably CoIITPPS). However, new
and 5.25 (q) ppm], acetaldehyde [δ = 2.24 (d, J = 2.0 Hz) and NMR signals were clearly observed at δ = 1.24 and 2.27 ppm as
9.68 (q) ppm], and ethylene (δ = 5.44 ppm) (Figure S2). Signal broad singlets, which might be ascribed to 1,1-ethanediol and
intensities suggested that the relative product yields were in acetaldehyde, respectively. Ethylene was detected at δ =
the order: acetaldehyde (1,1-ethanediol) > ethylene > ethanol. 5.43 ppm. From these results, we envisaged the effects of the
Ethylene is a known product of the photolysis of ethylcobal- cyclodextrin cage on the photolysis of alkyl–CoIIITPPSs. The
amine[5b] and ethylcobaloxime.[17] Ethylene formation was inter- methyl group of CH3–CoIIITPPS was covered by the capsule
preted to result from hydrogen transfer from an ethyl radical to formed from two TMe-ꢀ-CD molecules (Figure S7), allowing im-
the CoII complex, both species generated from the photolysis of mediate recombination of the photochemically produced radi-
ethylcobaloxime.[17] In contrast to CH3–CoIIITPPS, C2H5–CoIIITPPS cal pair (a CH3 radical and CoIITPPS) due to the instability of
reacted photochemically even under an Ar atmosphere, afford- the methyl radical. Meanwhile, the photolysis of C2H5–CoIIITPPS
ing CoIITPPS and ethylene. The formation of CoIITPPS was con- yielded a radical pair composed of an ethyl radical and CoIITPPS
firmed by comparing the absorption (λmax = 412 and 524 nm) in a cage, where hydride transfer occurred to yield ethylene and
and EPR spectra (Figures S3 and S4) with those of the authentic H–CoIIITPPS. H–CoIIITPPS is so reactive that a coupling reaction
sample prepared by electrophotochemical reduction of CoIII- yielded H2 and CoIITPPS. The ethyl radical is more stable than
TPPS.
The formation of 1:2 trans-type inclusion complexes of CH3–
the methyl radical, and some ethyl radicals therefore preserved
in the cyclodextrin cage until they collided with O2 penetrating
into the cyclodextrin capsule. In any event, the deceleration in
the photodecomposition clearly indicated that the TMe-ꢀ-CD
dimer markedly prevented the penetration of O2 into the cap-
sule.
and C2H5–CoIIITPPSs with TMe-ꢀ-CD was verified by using COSY
spectra (Figure S5). The effects of TMe-ꢀ-CD on the photodec-
omposition of CH3– and C2H5–CoIIITPPSs under aerobic condi-
tions are shown in Figure 2c and 2d, respectively, which demon-
strate a marked inhibition of the photolysis of CH3–CoIIITPPS
Next, we wanted to collect information on a cage formed by
(Figure 2e) and a weak inhibition of that of C2H5–CoIIITPPS (Fig- the covalently linked cyclodextrin dimer, hemoCD1. However,
ure 2f). The HMQC spectrum of the sample obtained by the coordination with the pyridine moiety in Py3CD (Figure 1) inter-
photolysis of 13CH3–CoIIITPPS/TMe-ꢀ-CD indicates the formation fered with the use of alkyl–CoIIITPPS photodecomposition as a
of methanediol and methanol (Figure S6). Ethane was not de- probe for studying the cage effect. As an alternative, we used
tected. Analysis of the products of the photolysis of C2H5–CoIII- Ph2CD complexes of CH3– and C2H5–CoIIITPPSs. Ph2CD is a per-
Eur. J. Inorg. Chem. 2016, 1784–1789
www.eurjic.org
1786
© 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communication
Figure 3. 1H NMR spectra of 13CH3–CoIIITPPS (1 m
of the photolyzed sample.
M) in D2O (a) before and (b) after photoirradiation for 60 min, and (c) the proton-decoupled NMR spectrum
O-methylated ꢀ-cyclodextrin dimer with a phenyl linker (Fig- CH2 protons were also observed in the Type 1 complex, though
ure 1), which might provide information on the cage formed not in the C2H5–CoIIITPPS/TMe-ꢀ-CD complex. These results
by cyclodextrin dimers, in which the monomer units are bound strongly suggest that Ph2CD forms a tighter cyclodextrin cap-
to each other with a covalent bond. UV/Vis spectral titrations sule than TMe-ꢀ-CD. Integration of NMR signals indicate that
revealed the formation of very stable 1:1 inclusion complexes the ratio of formation of Type 1 to Type 2 is 60:40. CH3–CoIIITPPS
of Ph2CD with CH3– and C2H5–CoIIITPPSs in aqueous solution also forms similar Type 1 and 2 inclusion complexes with
1
(Figure S9). The H NMR spectrum of the CH3–CoIIITPPS/Ph2CD Ph2CD, in a 64:36 ratio.
complex showed two singlet signals attributed to CH3 protons
The photodecomposition of the CH3–CoIIITPPS/Ph2CD com-
at δ = –4.74 and –4.53 ppm. The COSY spectrum of the C2H5– plex under aerobic conditions was followed by UV/Vis spectro-
CoIIITPPS/Ph2CD complex is shown in Figure 4 and indicates scopy (Figure S10). The photoreaction occurred more slowly
that two types of ethyl group exist in the inclusion complex, than that with the TMe-ꢀ-CD complex, affording the CoIIITPPS/
the two CH2 protons existing in different environments. The Ph2CD complex as a final product. Meanwhile, for the C2H5–
structures derived from these results are shown in Figure 4. The CoIIITPPS (3 μ
M) / Ph2CD (3.2 μM) complex (λmax = 406 and
ethyl group in the Type 1 complex is positioned opposite to 522 nm), the reaction ceased after 12 min irradiation, and ab-
the phenyl linker, while in the Type 2 complex it is on the same sorption maxima of the photolyzed sample were observed at
side. The porphyrin ring in the Ph2CD capsule cannot rotate 409, 427, and 525 nm (Figure S11). It was found by comparing
freely because of the steric barrier resulting from the phenyl with the authentic samples that the final absorption spectrum
linker. The ethyl group in the Type 2 complex is stuck in a nar- was a combination of spectra of Ph2CD complexes of CoIITPPS
row space where it cannot rotate freely, resulting in the distin- and CoIIITPPS. The formation of CoIITPPS was confirmed by EPR
guishable CH2 protons in the ethyl group. These distinguishable spectroscopy (Figure S12). Although CoIITPPS was always ob-
Eur. J. Inorg. Chem. 2016, 1784–1789
www.eurjic.org
1787
© 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communication
Figure 4. COSY spectrum (500 MHz) of the C2H5–CoIIITPPS/Ph2CD complex ([C2H5–CoIIITPPS] = 2.0 m
M, [Py2CD] = 2.4 mM) in D2O.
served as a product in the photoreactions of C2H5–CoIIITPPS yielded ethylene and CoIITPPS in a strictly closed cage formed
under anaerobic conditions, photolysis of the C2H5–CoIIITPPS/ by sandwiched phenyl and porphyrin rings.
Ph2CD complex was the only case in which CoIITPPS formation
was evidently measured by means of UV/Vis spectroscopy even
Conclusions
Photochemical reactions of CH3– and C2H5–CoIIITPPSs were con-
under aerobic conditions. The absorption spectrum suggested
the ratio of CoIITPPS to CoIIITPPS to be 60:40. The formation of
CoIITPPS involved a CoIII hydride complex, as shown in
venient probes for investigating the “cage effect”. More quanti-
Scheme 2.[17]
tative studies will be required to establish this method as an
accurate probe.
Experimental Section
Methyl– and ethyl–CoIIITPPSs were prepared by the reactions of
NaBH4-reduced CoTPPS with the corresponding alkyl iodides in
methanol containing sodium methoxide under an Ar atmosphere.
Scheme 2. Formation of CoIITPPS in the photolysis of C2H5–CoIIITPPS.
Each R–CoIIITPPS was purified on Sephadex LH-20 with methanol.
The products were identified by 1H MNR spectroscopy (Figures 3
and S2). The synthesis of Ph2CD is almost the same as that of the
cyclodextrin dimer whose linker was pyrid-3,5-diyl in place of 3,5-
phenyl.[19] Photolysis was performed by using the light from a
150 W Xe lamp, which was passed through a HOYA B-390 cut-off
filter (310–510 nm). The distance between the sample cell and the
lamp was 20 cm.
Although we attempted to detect the hydride complex of
CoIIITPPS by using the cage effect of Ph2CD and/or TMe-ꢀ-CD,
we could not do so reproducibly by using NMR spectroscopy
in the photolysis of C2H5–CoIIITPPS in anaerobic D2O-containing
cyclodextrin at 77 K. The hydride complex of CoIIITPPS could
produce CoIITPPS and H2.[18] CoIITPPS was stable to air oxidation
and the formation of CoIITPPS was therefore a measure of the
extent to which ethylene formation was involved in the reac-
tion. Because no hydride transfer reaction occurred in the pho-
tolysis of the CH3–CoIIITPPS/Ph2CD system, the reaction of the
methyl radical with O2 seemed to proceed slowly along with
recombination of the radical pair as the main event. Meanwhile,
the reactions of C2H5–CoIIITPPS included in the Ph2CD capsule
proceeded by two pathways. One of them is the reaction start-
ing from the Type 1 complex, where the photochemically gen-
erated ethyl radical mainly reacted with O2 to form an ethylper-
oxo complex affording CoIIITPPS as a final product. Another
pathway is the reaction starting from the Type 2 complex that
Acknowledgments
The authors acknowledge the financial support by Doshisha
University and the MEXT-Supported Program for Strategic Re-
search Foundation at Private Universities, 2015–2019.
Keywords: Cage effect · Host–guest systems · Cobalt ·
Photochemistry · Vitamins
[1] P. G. Lenhert, D. C. Hodgkin, Nature 1961, 192, 937–938.
[2] E. L. Smith, L. Mervyn, A. W. Johnson, N. Shaw, Nature 1962, 194, 1175.
Eur. J. Inorg. Chem. 2016, 1784–1789
www.eurjic.org
1788
© 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communication
[3]
Dagallier, M. Kodera, T. Matsuo, T. Hayashi, Y. Hisaeda, S. Hirota, Inorg.
Chem. 2006, 45, 4448–4460.
J. R. Guest, S. Friedman, D. D. Woods, E. L. Smith, Nature 1962, 195, 340–
342.
[12] K. Shikama, Chem. Rev. 1998, 98, 1357–1373.
[4]
[5]
K. Lindstrand, Nature 1964, 204, 188–189.
[13] S. Bhagan, B. B. Wayland, Inorg. Chem. 2011, 50, 11011–11020.
[14] G. R. Fulmer, A. J. M. Miller, N. H. Sherden, H. E. Gottlieb, A. Nudelman,
B. M. Stoltz, J. E. Bercaw, K. I. Goldberg, Organometallics 2010, 29, 2176–
2179.
[15] For example: a) F. A. Chavez, P. K. Mascharak, Acc. Chem. Res. 2000, 33,
539–545; b) M. J. Kendrick, W. Al-Akhdar, Inorg. Chem. 1987, 26, 3971–
3972; c) T. Hayashi, K. Okazaki, N. Urakawa, H. Shimakoshi, J. L. Sessler,
E. Vogel, Y. Hisaeda, Organometallics 2001, 20, 3074–3078.
[16] J. Watanabe, J. Setsune, J. Organomet. Chem. 1999, 575, 21–32.
[17] G. N. Schrauzer, J. W. Sibert, R. J. Windgassen, J. Am. Chem. Soc. 1968,
90, 6681–6688.
a) D. Dolphin, A. W. Johnson, R. Rodrigo, Ann. N. Y. Acad. Sci. 1964, 112,
590–600; b) D. Dolphin, A. W. Johnson, R. Rodrigo, J. Chem. Soc. 1964,
3186–3193.
[6]
R. Yamada, S. Shimizu, S. Fukui, Biochim. Biophys. Acta 1966, 124, 195–
197.
[7] For example: a) G. N. Schrauzer, R. J. Windgassen, J. Am. Chem. Soc. 1966,
88, 3738–3743; b) G. N. Schrauzer, Acc. Chem. Res. 1968, 1, 97–103; c)
D. A. Clark, D. Dolphin, R. Grigg, A. W. Johnson, H. A. Pinnock, J. Chem.
Soc. C 1968, 881–885.
[8] a) K. Kano, N. Tanaka, H. Minamizono, Y. Kawakita, Chem. Lett. 1996, 925–
926; b) K. Kano, R. Nishiyabu, T. Asada, Y. Kuroda, J. Am. Chem. Soc. 2002,
124, 9937–9944.
[18] J. L. Dempsey, B. S. Brunschwig, J. R. Winkle, H. B. Gray, Acc. Chem. Res.
2009, 42, 1995–2004.
[9] K. Kano, R. Nishiyabu, R. Doi, J. Org. Chem. 2005, 70, 3667–3673.
[10] K. Kano, H. Kitagishi, S. Tamura, A. Yamada, J. Am. Chem. Soc. 2004, 126,
15202–15210.
[19] K. Kano, Y. Itoh, H. Kitagishi, T. Hayashi, S. Hirota, J. Am. Chem. Soc. 2008,
130, 8006–8015.
[11] a) K. Kano, H. Kitagishi, M. Kodera, S. Hirota, Angew. Chem. Int. Ed. 2005,
44, 435–438; Angew. Chem. 2005, 117, 439; b) K. Kano, H. Kitagishi, C.
Received: February 29, 2016
Published Online: April 5, 2016
Eur. J. Inorg. Chem. 2016, 1784–1789
www.eurjic.org
1789
© 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim