Selective, cofactor-mediated catalytic oxidation of alkanethiols in a self-assembled cage host

Article abstract of DOI:10.1039/d0cc05765g

A spacious Fe(ii)-iminopyridine self-assembled cage complex can catalyze the oxidative dimerization of alkanethiols, with air as stoichiometric oxidant. The reaction is aided by selective molecular recognition of the reactants, and the active catalyst is derived from the Fe(ii) centers that provide the structural vertices of the host. The host is even capable of size-selective oxidation and can discriminate between alkanethiols of identical reactivity, based solely on size. This journal is

Full text of DOI:10.1039/d0cc05765g

ChemComm
View Article Online
View Journal
COMMUNICATION
Selective, cofactor-mediated catalytic oxidation
of alkanethiols in a self-assembled cage host†
Cite this:DOI: 10.1039/d0cc05765g
Bryce da Camara, Philip C. Dietz, Kevin R. Chalek, Leonard J. Mueller
Richard J. Hooley
and
Received 25th August 2020,
Accepted 19th October 2020
*
DOI: 10.1039/d0cc05765g
rsc.li/chemcomm
A spacious Fe(II)-iminopyridine self-assembled cage complex can and catalyze polar reactions on the host interior.14 During our
catalyze the oxidative dimerization of alkanethiols, with air as investigations into host-catalyzed thioetherification reactions,14b
stoichiometric oxidant. The reaction is aided by selective molecular we noticed that oxidative dimerization of alkanethiol nucleo-
recognition of the reactants, and the active catalyst is derived from philes was a persistent side reaction that could only be mini-
the Fe(II) centers that provide the structural vertices of the host. The mized under anaerobic conditions. We then investigated how
host is even capable of size-selective oxidation and can discriminate and why this reaction might occur, and the scope of the process.
between alkanethiols of identical reactivity, based solely on size.
The initial test was simple – n-octanethiol (C8-SH) was
refluxed in CD3CN in the presence of 5% Fe4L6 cage complex
1
Self-assembled metal–ligand cages have been used to promote 114a,b for 24 h and monitored by H NMR. After 24 h, all the
and catalyze a variety of reactions,1 from unimolecular rearran- octanethiol was consumed and only n-octyldisulfide could be
gements and cycloadditions,2 to acid and base-catalyzed seen. The cage was mostly intact and the reaction was clean,
additions3 and organometallic transformations.4 Encapsu- with no other obvious byproducts. While oxidative disulfide
lating substrates in host molecules allows a variety of novel formation is simple and well-known,15 the rapid reactivity was
reaction behaviors, including rate accelerations,5 sequestration surprising. As the cage does not decompose, the process must
of reactive intermediates6 and unusual regioselectivity.7 Novel be catalytic, with atmospheric O2 as stoichiometric oxidant –
outcomes such as size-and-shape or positional selectivity often indeed, if the reaction is repeated under N2, minimal reaction
come from strong binding in an internal cavity. This selectivity is seen. The nature of the active catalyst was not obvious,
comes with a price: often, when exquisite size-selectivity in though. While it is obvious that the redox-active Fe(II) ions in
reactions occurs, then the substrates bind too tightly and the Fe4L6 are involved, they are fully saturated in the assembly
turnover can be limited, especially if the hosts are not water- and have no free coordination sites. No change in oxidation
soluble and cannot take advantage of hydrophobic effects.8
state during the reaction can be seen from the NMR analysis
One other rarity in supramolecular catalysis is the host either. The likeliest explanation is that small amounts of Fe(II)
acting as, or delivering, the active reagent for the reaction. ions leach from the assembly, and act as the active catalyst
Hosts are mostly used as tiny (‘‘yoctoliter’’, in some cases9) while the bulk of the cage remains intact. We have seen
flasks. Guests are encapsulated and reaction is accelerated due evidence of this phenomenon before when performing post-
to increased effective concentration. Some cages have internal assembly modifications on Fe-containing cages.16
functional groups,10 some can be exploited as sensitizers for
The next step was to see if this was a common phenomenon
photochemical reactions11 and the walls of the vessels can for Fe-iminopyridine systems, and so we repeated the reaction
sometimes participate,12 but mostly, cage hosts just provide with 1 and two other differently sized cages: Nitschke’s Fe4L6
separate nanophases for the reaction. Enzymes, on the other cage 2,17 and the Fe2L3 helicate 3 (Fig. 1).18 Again, C8-SH was
hand, actively participate in the reaction, and can exploit metal added to a CD3CN solution of 5% cage 1 or 2, or 10% 3
ion cofactors for the catalyzed processes.13
(to ensure the same concentration of Fe) and heated in air.
We have recently shown that self-assembled Fe4L6 cage com- As can be seen in Fig. 2, the reactivity difference is stark – in the
plexes can act as hosts for neutral molecules in organic solution, presence of 1, 54% conversion of C8-SH to the corresponding
disulfide (C8-S)2 is seen after 7.5 h at 80 1C, whereas even at
Department of Chemistry and the UCR Center for Catalysis, University of
80 1C for 19 h, minimal (o5%) conversion occurs with either
hosts 2 or 3. In addition, when 25% Fe(NTf2)2 was added to a
solution of C8-SH in CD3CN and heated for 36 h, no oxidation
California-Riverside, Riverside, CA 92521, USA. E-mail: richard.hooley@ucr.edu
† Electronic supplementary information (ESI) available: Spectroscopic data,
including guest binding and reactivity profiles. See DOI: 10.1039/d0cc05765g
Chem. Commun.
This journal is © The Royal Society of Chemistry 2020

View Article Online
Communication
ChemComm
conditions were not sufficient for effective oxidation. Further-
more, adding extra Fe(NTf2)2 to the cage-catalyzed reaction
caused a reduction in conversion (Fig. S7, ESI†). Adding 10,
25 or 50% (with respect to C8-SH) to the C8-SH dimerization
reaction with 5% cage 1 gave 35%, 33% and 20% conversion
respectively, after 11.5 h at 50 1C.
The large cage 1 evidently displays unusual reactivity, and
this is most likely due to its molecular recognition capabilities.
The cavity in 1 is much larger than those in 2 or 3 (Fe–Fe
distances are shown in Fig. 1), and we have previously shown
that it is a strong host for small neutral molecules.14 Molecular
recognition could allow size-selectivity, so we analyzed the
relative rate of reaction for differently sized thiols. Six different
n-alkanethiols were reacted with 5% 1 for 11.5 h at 50 1C in
CD3CN in air, and the observed conversions are shown in
Table 1. These conditions were chosen to allow a comparison
of the relative rates of reaction. Maximal (490%) conversion to
disulfide product was possible after 22 h reflux in CD3CN
(80 1C), although some decomposition of the cage did occur
when heated for extensive periods of time at this temperature.
The conversion of the small and medium-sized thiols (C5-SH–
C
10-SH) under the less forcing 50 1C/11.5 h reaction conditions
was essentially identical, with 50–60% conversion observed in
each case. As the thiol increased in size, however, the efficacy of
the process reduced sharply. C11-SH was oxidized, albeit slower
than C5–C10, but dodecanethiol was oxidized far more slowly,
with only 15% conversion under the conditions.
The binding properties of the different alkanethiols in 1
were analyzed by UV-Vis absorbance spectroscopy. This is the
optimal method of determining the association constant and
binding stoichiometry for cages such as 1,14 which show rapid
in/out exchange of neutral small molecule guests on the NMR
timescale, making quantitative NMR analysis of the recognition
challenging. The alkanethiol guests were no different, and
showed rapid in/out exchange by NMR. Each guest was titrated
into a 3 mM solution of 1 (or 2) in CH3CN, and the changes in
absorbance at both 330 and 370 nm (or 275/335 nm for 2) were
recorded and analyzed. In each case, the binding isotherms
were fit to both the 1 : 1 and unbiased 1 : 2 binding models and
the variances calculated.19 The significance of the 1 : 2 model
was judged based on the inverse ratio of the squared residuals
compared to the 1 : 1 model, and quantified via p-value. The
results are summarized in Table 2: for the full fitting details,
including fitting curves, variances and error analysis, see ESI.†
Fig. 1 Self-assembled cage complexes tested. (a) Large Fe4L6 host 1;
(b) medium-sized Fe4L6 tetrahedron 2; (c) Fe4L6 helicate 3. (d) Illustration
of the oxidation process.
Table 1 Relative reactivity of alkanethiols in cage 1a
Fig. 2 1H NMR analysis of the reaction catalyzed by various Fe-containing
species. Expansion of the CH2–S region of the 1H NMR spectra of (a)
reaction mixture after reaction for the indicated time; (b) purified disulfide
product (C8-S)2; (c) purified thiol starting material C8-SH. CD3CN, 400 MHz,
spectra acquired at 298 K.
Reactant
Conversion, %
Reactant
Conversion, %
C5-SH
C6-SH
C8-SH
49
47
54
C10-SH
C11-SH
C12-SH
63
43
15
a
1
product was seen. Oxidation can occur non-catalytically with
Reactions performed at 50 1C, 11.5 h, CD3CN and analyzed by H NMR,
free FeII salts under more forcing conditions, but these mild concentrations determined using dioxane standard. [Cx-SH] = 18.2 mM.
Chem. Commun.
This journal is © The Royal Society of Chemistry 2020

View Article Online
ChemComm
Communication
Table 2 Binding affinities of alkanethiols in cage 1a
the same order, there is minimal product inhibition seen, as
one species does not dominate the binding.
The strong binding of the targets in host 1, and the oppor-
tunity for encapsulation, provides an explanation for the
unusual reactivity of 1 when compared to other Fe-based
assemblies. We have no evidence that the intact cage itself acts
II
as the catalyst, so the theory that a small amount of Fe released
from cage 1 in solution is the active catalyst in the reaction is
the most plausible, with atmospheric oxygen as stoichiometric
oxidant. This process could obviously occur in free solution,
but coencapsulation of the guests increases their effective
concentration, smoothing the reaction process. Cages with
small (or no) cavities such as 2 or 3 are not capable of this
reactivity. Interestingly, even though cage 2 can bind thiols in a
1 : 1 manner, no reactivity is seen, indicating that coencapsula-
tion is needed. Addition of superstoichiometric (with respect to
cage) amounts of Fe(NTf2)2 slowed the reaction down, which
suggests that in this case the additional ions are competitive
guests for the cage, displacing the thiol guests and slowing
reaction. There appears to be a ‘‘sweet spot’’ in [Fe] that allows
both the guests and the active Fe(II) ion catalyst to bind in host 1.
The small amounts of Fe(II) active catalyst act as ‘‘cofactors’’ for
this biomimetic reaction in the host. Cofactor-mediated catalysis,
1 : 2 substrate (1)
K1 Â 103 MÀ1
2150 Æ 650
540 Æ 130
174 Æ 43
K2 Â 103 MÀ1
a (4K2/K1)
C5-SH
C6-SH
C8-SH
1.2 Æ 3.0
8.7 Â 10À4
2.4 Æ 1.5
0.018
0.78 Æ 0.53
0.018
1 : 1 substrate (1) Ka  103 MÀ1 1 : 1 Substrate (1) Ka  103 MÀ1
C10-SH
C11-SH
C12-SH
(C3-S)2
(C5-S)2
19.7 Æ 6.4
40.0 Æ 19
2.7 Æ 0.6
16.6 Æ 2.4
38.8 Æ 7.1
(C6-S)2
(C8-S)2
(C10-S)2
(C11-S)2
(C12-S)2
71.0 Æ 14
76.1 Æ 3.8
27.9 Æ 9.4
5.5 Æ 0.5
8.4 Æ 0.9
1 : 1 substrate (2)
Ka  103 MÀ1
420 Æ 130
C6-SH
a
In CH3CN, [1], [2] = 1.5 mM, absorbance changes measured at 330 nm
and 370 nm for 1, and 278/335 nm for 2.19
The binding affinities of all the alkanethiols for cage 1 were namely the use of an additional reactant bound inside the parent
quite high (as we have seen for other guests),14b in the range ‘‘apoenzyme’’ host to effect reactivity, is usually only seen with
of 2000–40 000 MÀ1. Most interestingly, the medium-sized large superstructures.20
(C5-SH–C8-SH) thiols fit best to a 2 : 1 model, with negative
cooperativity. C5-SH had the strongest affinity, but the
affinities are broadly similar. As the thiols increase in size
(C10-SH–C12-SH), error analysis indicates that the 1 : 1 binding
motif is more favored, and C12-SH has by far the lowest affinity
for 1. As the analysis is based on variance analysis to a fitting
model, it is important to state that both modes of binding are
possible in each case, just less favored: the cavity of 1 is
theoretically big enough to fit two copies of C12-SH.
The larger products only fit to a 1 : 1 model, as might be
expected. The binding affinities of the products correlate nicely
with the observation that ‘‘mid-sized’’ thiols react fastest in 1:
the strongest affinities are for (C6-S)2 and (C8-S)2, whereas
smaller (C3, C5) and larger (C10–C12) products are less favored.
Importantly, when thiols, even small ones such as C6-SH, were
titrated into xylene cage 2, the binding only fit to a 1 : 1 model,
and 2 : 1 binding was highly unlikely. Minimized structures
(SPARTAN, AM 1 forcefield) are shown in Fig. 3, and support
this observed selectivity: cage 1 has a large cavity and can
encapsulate two molecules of C8-SH (Fig. 3d), whereas cage 2
is much smaller and only one guest can fit (Fig. 3e). Larger
guests (C10-SH–C12-SH) would fill the cavity of 1, disfavoring
2 : 1 binding. The models show that the large panel gaps do not
prevent ingress/egress of the reactants and allow protrusion of
the alkyl arms of the guest(s), if needs be, and that the exact
orientation of the guest(s) in cavity will be quite variable.
Fig. 3 Size-selective reactivity. Expansions of the GC traces obtained
after reaction between different thiols (25 1C, 7 days, CD3CN, 5% 1).
(a) C3-SH and C10-SH; (b) C6-SH and C7-SH; (c) C6-SH and C12-SH.
Notably, all of the products have some affinity for the cage,
even (C12-S)2, which shows that while large reactants favor 1 : 1
binding, 2 : 1 binding is possible and reaction can still occur. As
the binding affinities of reactant and product are generally of
Minimized structures of (d) 1Á(C8-SH)2 (e) and 2ÁC8-SH (SPARTAN).
(i) dodecane; (ii) unreacted C10-SH; (iii) impurity in the GC column.
Chem. Commun.
This journal is © The Royal Society of Chemistry 2020

View Article Online
Communication
ChemComm
Table 3 Heterodimerization selectivity of alkanethiols in cage 1a
size, but identical reactivity, all while showing good turnover.
This selectivity is unusual, and we are currently investigating its
applications in dynamic combinatorial libraries.
The authors would like to thank the National Science
R2R2 Foundation (CHE-1708019 and CHE-2002619 to R. J. H.), and
R1
R2
Conversion, %
R1R1
R1R2
the National Institutes of Health (GM097569 to L. J. M.) for
C3
C3
C6
C6
C6
C8
C10
C7
C10
C12
19
20
20
18
12
7
10
23
19
28
35
34
49
42
72
58
56
28
39
0
support. Prof. Jocelyn Millar and Kyle Arriola are grate-
fully acknowledged for assistance with GC, and we thank
Prof. Darren Johnson for initial discussions.
a
Reactions performed at 25 1C, 7 d, CD3CN and analyzed by GC,
concentrations determined using dodecane as internal standard.
Equimolar amounts of each thiol used, overall [Cx-SH] = 18.2 mM.
Conflicts of interest
There are no conflicts to declare.
While the size-selectivity of the reaction in 1 is modest when
comparing homodimerization of n-alkanethiols, we were inter-
ested in determining whether any selectivity could be seen
when reacting two different thiols. The heterodimerization
products of reaction between n-alkanethiols of different length
cannot be distinguished by NMR, as might be expected, so a GC
method is required. Initial tests run at 80 1C for 24 h were not
encouraging, as statistical mixtures were seen. However, mix-
tures of RSH + RSSR are well-known to equilibrate over time,
especially at high temperature.21 To remove this equilibration,
we analyzed the reactions between sets of equimolar amounts
of two different alkanethiols in the presence of 5% 1 at 25 1C for
7 days. The observed conversions were o20% in each case, and
allow a view of the initial selectivity. The combinations tested
were C3/C8, C3/C10, C6/C7, C6/C10 and C6/C12 – Fig. 3a–c shows
GC data for three of these reactions (see ESI† for full,
uncropped GC traces), and Table 3 shows the product
distributions.
Notes and references
1 (a) Y. Fang, J. A. Powell, E. Li, Q. Wang, Z. Perry, A. Kirchon, X. Yang,
Z. Xiao, C. Zhu, L. Zhang, F. Huang and H.-C. Zhou, Chem. Soc. Rev.,
2019, 48, 4707; (b) I. Sindha and P. S. Mukherjee, Inorg. Chem., 2018,
57, 4205; (c) C. J. Brown, F. D. Toste, R. G. Bergman and
K. N. Raymond, Chem. Rev., 2015, 115, 3012.
´
2 (a) V. Martı-Centelles, A. L. Lawrence and P. J. Lusby, J. Am. Chem.
´
Soc., 2018, 140, 2862; (b) T. A. Young, V. Martı-Centelles, J. Wang,
P. J. Lusby and F. Duarte, J. Am. Chem. Soc., 2020, 142, 1300.
3 (a) J. Jiao, Z. Li, Z. Qiao, X. Li, Y. Liu, J. Dong, J. Jiang and Y. Cui, Nat.
Commun., 2018, 9, 4423; (b) M. W. Cullen, M. C. Misuraca,
C. A. Hunter, N. H. Williams and M. D. Ward, Nat. Chem., 2016,
8, 231.
4 C. M. Hong, R. G. Bergman and K. N. Raymond, Acc. Chem. Res.,
2018, 51, 2447.
5 D. M. Kaphan, M. D. Levin, R. G. Bergman, K. N. Raymond and
F. D. Toste, Science, 2015, 350, 1235.
6 Y. Ueda, H. Ito, D. Fujita and M. Fujita, J. Am. Chem. Soc., 2017,
139, 6090.
7 Y. Nishioka, T. Yamaguchi, M. Yoshizawa and M. Fujita, J. Am.
Chem. Soc., 2007, 129, 7000.
In these non-equilibrated kinetic experiments, the selectivity
for differently sized alkanethiols is obvious, and quite impressive.
While minimal selectivity is seen when C6-SH and C7-SH are
combined, as might be expected, other combinations showed
significant excesses of one product. For example, when C3-SH
and C8-SH were reacted, (C8-S)2 was favored in an 8.6 : 5.5 : 1
ratio over C8-S–S-C3 and (C3-S)2, respectively. Similar product
ratios were observed for the C3/C10 combination, but the
selectivity towards (C10-S)2 was slightly lower. Consistent with
the observation that larger alkanethiols (ZC10) were not favorably
coencapsulated, the combination of C6-SH and C12-SH gave only
two products, with the C6-S–S-C12 heterodimer being formed in a
3.6 : 1 excess over (C6-S)2, and no (C12-S)2 was observed at all. The
8 S. Liu, H. Gan, A. T. Hermann, S. W. Rick and B. C. Gibb, Nat. Chem.,
2010, 2, 847.
9 X. Cai, R. Kataria and B. C. Gibb, J. Am. Chem. Soc., 2020, 142, 8291.
´
10 F. Yu, D. Poole, S. Mathew, N. Yan, J. Hessels, N. Orth, I. Ivanovic-
´
Burmazovic and J. N. H. Reek, Angew. Chem., Int. Ed., 2018, 57,
11247.
11 J. Guo, Y.-Z. Fan, Y.-L. Lu, S.-P. Zheng and C.-Y. Su, Angew. Chem.,
Int. Ed., 2020, 59, 8661.
12 C. Zhao, F. D. Toste, R. G. Bergman and K. N. Raymond, J. Am. Chem.
Soc., 2014, 136, 14409.
13 R. Banerjee and S. W. Ragsdale, Annu. Rev. Biochem., 2003, 72, 209.
14 (a) L. R. Holloway, P. M. Bogie, Y. Lyon, C. Ngai, T. F. Miller,
R. R. Julian and R. J. Hooley, J. Am. Chem. Soc., 2018, 140, 8078;
(b) C. Ngai, P. M. Bogie, L. R. Holloway, P. C. Dietz, L. J. Mueller and
R. J. Hooley, J. Org. Chem., 2019, 84, 12000.
15 S. Otto, R. L. E. Furlan and J. K. M. Sanders, J. Am. Chem. Soc., 2000,
122, 12063.
most favored combinations are those with approximately 13–18 16 (a) L. R. Holloway, H. H. McGarraugh, M. C. Young, W. Sontising,
G. J. O. Beran and R. J. Hooley, Chem. Sci., 2016, 7, 4423;
(b) L. R. Holloway, P. M. Bogie, Y. Lyon, R. R. Julian and
R. J. Hooley, Inorg. Chem., 2017, 56, 11435.
carbon atoms, i.e. C3-S–S-C10, C6-S–S-C10 or C6-S–S-C12, which
corresponds with the observation that medium-sized thiols such
as C8-SH are favorably coencapsulated and the reaction rate drops 17 W. Meng, J. K. Clegg, J. D. Thoburn and J. R. Nitschke, J. Am. Chem.
Soc., 2011, 133, 13652.
18 M. C. Young, A. M. Johnson and R. J. Hooley, Chem. Commun., 2014,
off as the guests increase in size.
In conclusion, we have shown that small amounts of a self-
50, 1378.
assembled host are capable of the catalytic oxidation of 19 P. Thordarson, Chem. Soc. Rev., 2011, 40, 1305.
20 Q. Zhang, L. Catti, J. Pleiss and K. Tiefenbacher, J. Am. Chem. Soc.,
alkanethiols to their corresponding disulfides. The reaction
requires coencapsulation to proceed effectively, and the host
2017, 139, 11482.
21 N.-M. Phan, E. P. K. L. Choy, L. N. Zakharov and D. W. Johnson,
is capable of distinguishing between alkanethiols of differing
Chem. Commun., 2019, 55, 11840.
Chem. Commun.
This journal is © The Royal Society of Chemistry 2020