Visible light mediated activation and O-glycosylation of thioglycosides

Article abstract of DOI:10.1021/ol302941q

Visible light catalysis allows the efficient construction of single electron transfer (SET) redox cycles that result in minimal formation of byproducts and proceed under exogenous control of a removable light source. The O-glycosylation of thioglycosides via visible light photoredox chemistry is reported. Mechanistic studies show that the reaction is fully light responsive and support a mechanism involving decomposition of an oxidatively generated sulfur radical cation and propagation via reduction of the thiol side product.

Full text of DOI:10.1021/ol302941q

ORGANIC
LETTERS
XXXX
Vol. XX, No. XX
000–000
Visible Light Mediated Activation and
O-Glycosylation of Thioglycosides
Walter J. Wever,† Maris A. Cinelli,† and Albert A. Bowers*
Division of Chemical Biology and Medicinal Chemistry, Eshelman School of Pharmacy,
University of North Carolina at Chapel Hill, Chapel Hill, North Carolina 27599,
United States
abower2@email.unc.edu
Received October 24, 2012
ABSTRACT
Visible light catalysis allows the efficient construction of single electron transfer (SET) redox cycles that result in minimal formation of byproducts
and proceed under exogenous control of a removable light source. The O-glycosylation of thioglycosides via visible light photoredox chemistry is
reported. Mechanistic studies show that the reaction is fully light responsive and support a mechanism involving decomposition of an oxidatively
generated sulfur radical cation and propagation via reduction of the thiol side product.
Visible light catalysis is finding increasing application
in organic synthesis.1,2 Such efforts have been fed by the
development of efficient catalysts for the oxidation of
water to generate hydrogen. A significant appeal of visible
light catalysis is that complementary odd-electron organic
reactions can be joined to construct a viable catalytic cycle.
Becausethe excitedstatemetal complexistypically capable
of giving or receiving an electron, catalytic cycles that
are initiated by oxidation and reduction (termed oxidative
or reductive quenching) have both been formulated
(Figure 1A).3 For example, Stephenson et al.3 have re-
cently reported the use of reductive quenching to convert
alcohols to alkyl halides via intermediate formation of
a VilsmeierꢀHaack-type reagent, while multiple groups
have employed oxidative quenching to generate and nu-
cleophilically trap stabilized iminium ion intermediates.3a,4
Notably, although oxocarbenium ions have been in-
voked mechanistically in visible light reactions,3c they have
not as yet been trapped in a manner similar to the iminium
ions. Moreover, neither species, iminium nor oxocarbe-
nium ions, when generated by visible light, has been reacted
intermolecularly with a stoichiometric alcohol acceptor
to yield the corresponding acetal or aminal linkage.5
Glycosidic bond formation presents a specific example
where such a reaction could be synthetically useful. There-
fore, we initiated the efforts described herein, aimed at
† W.W. and M.A.C. contributed equally.
(1) Recent reviews: (a) Narayanam, J. M. R.; Stephenson, C. R. J.
Chem. Soc. Rev. 2011, 40, 102. (b) Teply, F. Collect. Czech. Chem.
Commun. 2011, 76, 859. (c) Inagaki, A.; Akita, M. Coord. Chem. Rev.
2010, 254, 1220. (d) Yoon, T. P.; Ischay, M. A.; Du, J. Nat. Chem. 2010,
2, 527. (e) Zeitler, K. Angew. Chem., Int. Ed. 2009, 48, 9785.
(2) Representative examples: (a) Nagib, D. A.; MacMillan, D. W. C.
Nature 2011, 480, 224. (b) Pham, P. V.; Nagib, D. A.; MacMillan,
D. W. C. Angew. Chem., Int. Ed. 2011, 50, 6119. (c) Nicewicz, D. A.;
MacMillan, D. W. C. Science 2008, 322, 77. (d) Lin, S.; Ischay, M. A.;
Fry, C. G.; Yoon, T. P. J. Am. Chem. Soc. 2011, 133, 19350. (e) Du, J.;
Espelt, L. R.; Guzei, I. A.; Yoon, T. P. Chem. Sci. 2011, 2, 2115. (f)
Hurtley, A. E.; Cismesia, M. A.; Ischay, M. A.; Yoon, T. P. Tetrahedron
2011, 67, 4442. (g) Lu, Z.; Shen, M.; Yoon, T. P. J. Am. Chem. Soc. 2011,
133, 1162. (h) Du, J.; Yoon, T. P. J. Am. Chem. Soc. 2009, 131, 14604. (i)
Furst, L.; Narayanam, J. M. R.; Stephenson, C. R. J. Angew. Chem., Int.
Ed. 2011, 50, 9655. (j) Tucker, J. W.; Narayanam, J. M. R.; Shah, P. S.;
Stephenson, C. R. J. Chem. Commun. 2011, 47, 5040. (k) Nguyen, J. D.;
Tucker, J. W.; Konieczynska, M. D.; Stephenson, C. R. J. J. Am. Chem.
Soc. 2011, 133, 4160. (l) Dai, C.; Narayanam, J. M. R.; Stephenson,
C. R. J. Nat. Chem. 2011, 3, 140. (m) Tucker, J. W.; Nguyen, J. D.;
Narayanam, J. M. R.; Krabbe, S. W.; Stephenson, C. R. J. Chem.
Commun. 2010, 46, 4985.
(3) (a) Esswein, A. J.; Nocera, D. G. Chem. Rev. 2007, 107, 4022. (b)
Kalyanasundaram, K. Coord. Chem. Rev. 1982, 46, 159. (c) Juris, A.;
Balzani, V.; Barigelletti, F.; Campagna, S.; Belser, P.; von Zelewsky, A.
Coord. Chem. Rev. 1988, 84, 85. (d) Balzani, V.; Bergamini, G.;
Marchioni, F.; Ceroni, P. Coord. Chem. Rev. 2006, 250, 1254.
(4) Iminium ions have also been generated and trapped by alcohols
intramolecularly or else with alcohol as cosolvent: (a) Courant, T.;
Masson, G. Chem.;Eur. J. 2012, 18, 423. (b) Xuan, J.; Feng, Z.; Duan,
S.-W.; Xiao, W.-J. RSC Advances 2012, 2, 4065.
(5) Visible light catalysis has recently been applied to a free radical
mediated synthesis of C-glycosides from glycosyl bromides by Gagne
et al.: (a) Andrews, R. S.; Becker, J. J.; Gagne, M. R. Angew. Chem., Int.
Ed. 2010, 49, 7274–76. (b) Andrews, R. S.; Becker, J. J.; Gagne, M. R.
Org. Lett. 2011, 13, 2406–09. (c) Andrews, R. S.; Becker, J. J.; Gagne,
M. R. Angew. Chem., Int. Ed. 2012, 51, 4140–43.
r
10.1021/ol302941q
XXXX American Chemical Society

development of a visible light mediated O-glycosylation
from thioglycoside donors.5
We first sought to determine if oxidative quenching by
commonly deployed co-oxidants, such as bromotrichloro-
methane or carbon tetrabromide, could be coupled to
thioglycoside activation and alcohol trapping. We sub-
mitted S-phenyl- and S-(p-methoxy)phenyl-tetra-O-ben-
zyl-thioglycosides (3 and 4, respectively) to visible light
irradiation under the various conditions described in Table 1.
As anticipated, based on the known oxidation poten-
tials, neither 3 nor 4 was activated by Ru(bpy)3Cl2, and
only 4 wasactivatedby Ir[dF(CF3)ppy]2(dtbbpy)PF6 (2).10
The oxidation potential of 3 has been measured at 1.31 eV
(vs Ag/AgCl), and that of 4, at 1.16 eV, just above and
below the reported oxidation potential of 2 (1.21 eV), and
well above that of 1. In the event, a 96% yield of the methyl
glycoside (5) was obtained in the presence of 2 equiv of
bromotrichloromethane and 10 equiv of methanol (Table 1,
entry 4). In addition, the symmetric bis(p-methoxyphenyl)
disulfide could be isolated in near-quantitative yield. We
further sought to reduce the number of equivalents of alcohol
to make the reaction amenable to couplings with more
precious alcohol acceptors. The efficiency of the reaction
diminished appreciably, however, and only trace amounts of
5could be obtained from our standard 6 h reaction time in the
presence of 2 equiv of alcohol (Table 1, entry 5).
Figure 1. (A) General scheme of commonly employed visible
light photoredox cycles and (B) electrochemical activation of
thioglycosides.
We speculated that the metalꢀligand charge complex
generated by oxidative quenching of an excited state visible
light catalyst, such as Ru(bpy)3Cl2 (1) or Ir[dF(CF3)ppy]2-
(dtbbpy)PF6 (2), would be strong enough to oxidize
an electron-rich aryl thioglycoside, yielding a stabilized
oxocarbenium ion intermediate via fragmentation of the
radical cation.6 Thioglycosides are stable, easily manipul-
able anomeric protecting groups that are widely employed
in oligosaccharide synthesis.7 Oxidation potentials of a
number of thioglycoside donors are known and Yoshida
et al. have demonstrated that these substrates can be
cleanly oxidized in an electrochemical cell (Figure 1B).8
Subsequent condensation with an alcohol acceptor would
provide an O-glycosidic linkage with the byproduct for-
mation of acid and the symmetric disulfide. Although
photoinitiated glycosylations by UV irradiation of
selenoglycosides are known, this would provide a unique
example of a visible light mediated glycosylation from
thioglycosides.9
Table 1. Optimization of Glycosylation Conditionsa
a Reactions carried out at ∼0.1 M thioglycoside in MeCN with
5 mol % catalyst, 10 equiv HFIP, 0.25 equiv X3CBr, and irradiated
for ∼6 h with blue LEDs. PMP = p-methoxyphenyl. Yields are reported
for isolated products as mixtures of anomers.
(6) To date, visible light photoredox chemistry of thiols has been
limited to direct oxidation to sulfoxides: Zen, J. M.; Liou, S. L.; Kumar,
A. S.; Hsia, M. S. Angew. Chem., Int. Ed. 2003, 42, 577–579. bUses as
terminal reductants: Okada, K.; Okubo, K.; Morita, N.; Oda, M.
Tetrahedron Lett. 1992, 33, 7377–7380.
(7) For example: (a) Crich, D.; Smith, M. J. Am. Chem. Soc. 2001,
123, 9015–16. (b) Yamago, S.; Yamada, T.; Maruyama, T.; Yoshida, J.
Angew. Chem., Int. Ed. 2004, 43, 2145 and references therein.
(8) (a) Amatore, C.; Jutand, A.; Mallet, J.-M.; Meyer, G.; Sinay, P.
Chem. Commun. 1990, 718. (b) Yamago, S.; Kokubo, K.; Hara, O.;
Masuda, S.; Yoshida, J.-I. J. Org. Chem. 2002, 67, 8584–92. (c) Suzuki,
S.; Matsumoto, K.; Kawamura, K.; Suga, S.; Yoshida, J.-I. Org. Lett.
2004, 6, 3755–58. (d) Nokami, T.; Shibuya, A.; Tsuyama, H.; Suga, S.;
Bowers, A. A.; Crich, D.; Yoshida, J.-I. J. Am. Chem. Soc. 2007, 129,
10922–28. (e) Nokami, T.; Shibuya, A.; Yoshida, J.-i. Trends Glycosci.
Glycotechnol. 2008, 20, 175–85. (f) Saito, K.; Ueoka, K.; Matsumoto, K.;
Suga, S.; Nokami, T.; Yoshida, J.-i. Angew. Chem., Int. Ed. 2011, 50,
5153–56.
The importance of a protic solvent to aid solvation and
disruption of the charge transfer complex has been well
documented for visible light catalysts such as 1 and 2.5,11
Thus, we anticipated that by supplementing with a non-
nucleophilic, protic solvent, we could effect glycosylation
with e2 equiv of acceptor alcohol. After some testing,
hexafluoroisopropanol (HFIP) proved an effective surro-
gate and a 51% yield of methyl glycoside could be
(10) (a) Lowry, M. S.; Goldsmith, J. I.; Slinker, J. D.; Rohl, R.;
Pascal, R. A., Jr.; Malliaras, G. G.; Bernhard, S. Chem. Mater. 2005, 17,
5712. (b) Goldsmith, J. I.; Hudson, W. R.; Lowry, M. S.; Anderson,
T. H.; Bernhard, S. J. Am. Chem. Soc. 2005, 127, 7502.
(11) (a) DeLaive, P. J.; Lee, J. T.; Sprintschnik, H. W.; Abruna, H.;
Meyer, T. J.; Whitten, D. G. J. Am. Chem. Soc. 1977, 99, 7094. (b)
DeLaive, P. J.; Foreman, T. K.; Giannotti, C.; Whitten, D. G. J. Am.
Chem. Soc. 1980, 102, 5627.
(9) (a) Furuta, T.; Takeuchi, K.; Iwamura, M. Chem. Commun. 1996,
157. (b) Cumptsey, I.; Crich, D. J. Carbohydr. Chem. 2011, 30, 469.
Although one example of UV (350 nm) activated glycosylation from
thioglycosides has been reported [(c) Griffin, G. W.; Bandara, N. C.;
Clarke, M. A.; Tsang, W.-S.; Garegg, P. J.; Oscarson, S.; Silwanis, B. A.
Heterocycles 1990, 30, 939–47], reaction times are long (24ꢀ48 h) and
substrate compatibility is poor (no protecting groups could be used and
the only reported acceptor was methanol used as a cosolvent).
B
Org. Lett., Vol. XX, No. XX, XXXX

obtained. Yields could be buffered by use of CBr4 in place
of BrCCl3. However, a more significant increase in yield
could be obtained when co-oxidant was employed in
substoichiometric quantities. 0.25 equiv of either CBr4 or
BrCCl3 provided good yields of methyl glycosides and
reduced side products. The mass balance in this case was
comprised of a complex mixture of decomposition prod-
ucts, which we presume to result from unchecked reac-
tivity of a poorly stabilized oxacarbeniumion under the
currentconditionsofreaction. TheconditionsfromTable1
(entries 9 and 10) were employed to further investigate
substituent effects on this intermediate, as well as substrate
compatibility for the visible light glycosylation.
variant 4 also afforded good yields. The tetra-acetyl thio-
glycoside (7) did not activate under the present conditions
(entry 3), presumably due to the higher oxidation potential
of this disarmed glycosyl donor. Acetates, however, are
tolerated, as demonstrated by reactions employing the
2-deoxy-L-rhamnothioglycoside (8). Additionally, the reac-
tion tolerates appreciable steric bulk (Table 2, entries 7ꢀ8),
aryl functionalities (entries 9, 10, and 13), and an amide
containing an enolizable stereocenter (entry 10). In the
2-deoxy series, yields with the tribenzyl-glucose-6-OH ac-
ceptor (9) were improved relative to the glucose series. We
speculate that this is due to the heightened reactivity of an
assumed glycosyl cation-like intermediate.
Primary, secondary, and tertiary alcohols (Table 2)
were all permissible nucleophiles and yielded O-glycosides as
mixtures of anomers. Although the majority of exploratory
work was carried out with tetramethyl donor 6, tetrabenzyl
We next sought to interrogate the mechanism of thio-
glycoside activation. We systematically removed each of
the individual components, light, alkyl halide, and catalyst,
to demonstrate that all are necessary for conversion of the
thioglycoside donor (Table 3, entries 1ꢀ3). In the absence of
catalyst, no reaction is observed (Table 3, entry 2). However,
without alkyl halide, the reaction still proceeds, albeit very
slowly (entry 3). As expected, the reaction is completely
responsive to light and shuts down in its absence. This was
rigorously demonstrated in a time course experiment, in
which the light source was varied on and off for 1 h intervals
over a 6 h period and then allowed to remain on until the
reaction had reached completion (see SI Figure 4A). At each
time point, 1H NMR of crude aliquots were examined and
reaction progress was determined by integration of protons
from the S-(p-methoxy)phenyl group of 6, versus those of the
disulfide byproduct. The reaction showed no progress during
intervals when the light source was turned off.
Table 2. Substrate Compatibility of Glycosylationa
Notably, in all instances, the final anomeric selectivity
is inconsistent with a solvent “nitrile effect.” The time
course data show that there is in fact initial β-selectivity,
presumably from a solvent effect, but that the selectivity
erodes over the course of the reaction, particularly dur-
ing instances of no light and, thus, no reaction (see SI
Figure 4B).12 This is presumably due to the combined
acidity of HFIP and any other byproduct acid evolved
during the reaction (pH ∼4). Presumably an acid cata-
lyzed background process is responsible for the dimin-
ished yields with acceptor 9. Efforts to buffer the reaction
with a non-nucleophilic base, such as DTBMP or TTBP,
were unsuccessful.
19F NMR of reaction mixtures before and after irradia-
tion showed that HFIP is not consumed during the reac-
tion. Similar experiments carried out for BrCCl3 showed
that it is only partially consumed when present at 1.0 equiv,
reinforcing the need for only substoichiometric amounts
(see SI).13 In fact, the reaction proceeds, albeit slowly,
(12) (a) Sinay, P.; Pougny, J. R. Tetrahedron. Lett. 1976, 17, 4073.
(b) Lemieux, R. U.; Ratcliffe, R. M. Can. J. Chem. 1979, 57, 1244. (c)
Hashimoto, S.; Hayashi, M.; Noyori, R. Tetrahedron. Lett. 1984, 25,
1379. (d) Schmidt, R. R.; Behrendt, M.; Toepfer, A. Synlett 1990, 694.
(e) Satoh, H.; Hansen, H. S.; Manabe, S.; van Gunsteren, W. F.;
a Reactions carried out at 0.1 M thioglycoside with 5 mol % catalyst
and irradiated for 6 h with blue LEDs unless otherwise stated. Yields are
reported for isolated products. b CBr4 used. c BrCCl3 used. d Irradiated
for 23 h. e Irradiated for 18 h. f Irradiated for 9 h. g Irradiated for 10 h.
€
Hunenberger, P. H. J. Chem. Theory Comput. 2010, 6, 1783.
(13) Likely alkyl halide byproducts, such as chloroform, hexachlor-
ethylene, tetrachlorethane, and a number of possible products from
reaction of a presumed trichlorocarbon radical, were absent from
NMRs of reactions run in CD3CN/CH3CN and from GC/MS traces.
Org. Lett., Vol. XX, No. XX, XXXX
C

Table 3. Investigation of Glycosylation Mechanism
Figure 2. Proposed mechanism for visible light mediated
O-glycosylation of thioglycosides and H-detection. A = un-
treated polymer, B = reaction with no light, C = polymer
treated with H2, D = reaction with light, E = reaction with
light, no cat., F = reaction with HBr, no light.
Reactions with donor 4 except where noted. a 0.25 equiv. b Negligible
change after an additional 6 h of no light. c Negligible change (∼3%) after
an additional 2 h of no light. d Reaction was performed using tetra-OMe-
S-PMP (6). Reactions carried out with 2.0 equiv of MeOH.
without any co-oxidant (Table 3, entry 3). The addition of
acid, either substoichiometric HBr (entry 4) or 2 equiv of
TFA (entry 5), enhances the BrCCl3-independent mechan-
ism. However, no conversion is observed in the absence of
light, underscoring the fact that a strong acid itself does not
catalyze activation of the thioglycoside. The reaction does
not appear to be oxygen sensitive or dependent (Table 3,
entries 6 and 7), and peroxide and halide tests were both
negative under standard reaction conditions.
Based on these data we propose the mechanism illustrated
in Figure 2, in which the catalytic cycle is initiated by the co-
oxidant and propagated by reduction of a thiol radical
generated from collapse of the thioglycoside radical cation.
The latter thiol radical may be present as the naked radical or
the thiol radical cation formed by combination with in situ
generated acid. Isolation of the disulfide could then be ex-
plained by background oxidation of 2 equiv of this thiol,
possibly with concomitant formation of molecular hydrogen.
Indeed, subjecting free thiol to our standard reaction
conditions (no thioglycoside present) resulted in clean,
rapid (<4 h) formation of the disulfide. Disulfide forma-
tion occurred only in the presence of catalyst and light.
Although necessary quantities of thioglycoside made it
difficult to assess hydrogen production in a glycosylation
reaction, the disulfide formation could be run on a 200-fold
scale in the presence of a H-detecting polymer (Figure 2).14
Color change occurred only in the presence of light and a
catalyst and was consistent with the formation of hydrogen.
Typically, photocatalytic procedures for hydrogen produc-
tion utilize both a photosensitizer, such as Ru(bpy)3Cl2 (1) or
Ir[dF(CF3)ppy]2(dtbbpy)PF6 (2), and an electron relay cat-
alyst, such as Co(bpy)3 or methyl viologen.15 However,
2þ
direct hydrogen production from aqueous solution, without
the need for an electron relay, has been observed in the
presence of strong acids such as HCl or HBr.16,17
In summary, we have developed a visible light mediated
photoredox O-glycosylation reaction. The mechanism in-
volves a unique visible light activated single-electron oxida-
tion on sulfur, followed by collapse to an oxocarbenium ion.
In its present iteration, this visible light glycosylation pro-
ceeds with limited stereoselectivity due to thermodynamic
equilibration under the reaction conditions. However, re-
cent advances in visible light catalysis in flow suggest a likely
strategy for conducting the reaction under kinetic control.18
Efforts to this end are ongoing in our laboratory.
(14) Molybdenum trioxide based polymers provided by Element One
Inc. Chemistry are described at http://www.elem1.com/NHA2006final.pdf.
(15) Examples: (a) Kirch, M.; Lehn, J. M.; Sauvage, J. P. Helv. Chim.
Acta 1979, 62, 1345. (b) Chan, S.-F.; Chou, M.; Creutz, C.; Matsubara,
T.; Sutin, N. J. Am. Chem. Soc. 1981, 103, 369. (c) Krishnan, C. V.;
Sutin, N. J. Am. Chem. Soc. 1981, 103, 2141. (d) Ziessel, R.; Hawecker,
J.; Lehn, J. M. Helv. Chim. Acta 1986, 69, 1065.
(16) (a) Eidem, P. K.; Maverick, A. W.; Gray, H. B. Inorg. Chim.
Acta 1981, 50, 59. (b) Gray, H. B.;Maverick, A. W. Science 1981, 214, 1201.
(17) One example was recently reported: DiRocco, D. A.; Rovis, T.
J. Am. Chem. Soc. 2012, 134, 8094.
(18) Photoflow reactors allow shorter reaction times and lower
temperatures: (a) Andrews, R. S.; Becker, J. J.; Gagne, M. R. Angew.
Chem., Int. Ed. 2012, 51, 4140. (b) Tucker, J. W.; Zhang, Y.; Jamison,
T. F.; Stephenson, C. R. J. Angew. Chem., Int. Ed. 2012, 51, 4144.
(c) Bou-Hamdan, F. R.; Seeberger, P. H. Chem. Sci. 2012, 3, 1612.
Supporting Information Available. Procedures and data
for all new compounds. This material is available free of
charge via the Internet at http://pubs.acs.org.
Acknowledgment. David Colby (Purdue), Stefan
Bernhard (Carnegie-Mellon), and Christopher Neumann
(University of Washington) are thanked for helpful dis-
cussions. Dr. Karl Wood (Purdue) provided valuable
assistance with GC/MS. David Benson of Element One
Inc. generously donated H-sensing polymers.
The authors declare no competing financial interest.
D
Org. Lett., Vol. XX, No. XX, XXXX