Repairing the thiol-ene coupling reaction

Article abstract of DOI:10.1002/anie.201309984

Thiol-ene coupling (TEC) reactions emerged as one of the most useful processes for coupling different molecular units under reaction mild conditions. However, TEC reactions involving weak C-H bonds (allylic and benzylic fragments) are difficult to run and often low yielding. Mechanistic studies demonstrate that hydrogen-atom transfer processes at allylic and benzylic positions are responsible for the lack of efficiency of the radical-chain process. These competing reactions cannot be prevented, but reported herein is a method to repair the chain process by running the reaction in the presence of triethylborane and catechol. Under these reaction conditions, a unique repair mechanism leads to an efficient chain reaction, which is demonstrated with a broad range of anomeric O-allyl sugar derivatives including mono-, di-, and tetrasaccharides bearing various functionalities and protecting groups. In good repair: Undesired hydrogen-atom transfers are responsible for the lack of efficiency in thiol-ene coupling reactions involving allyl glycosides. This competing reaction cannot be prevented but can be very efficiently repaired by carrying out the reaction in the presence of triethylborane and catechol.

Full text of DOI:10.1002/anie.201309984

.
Angewandte
Communications
DOI: 10.1002/anie.201309984
Radical Reactions
Repairing the Thiol-Ene Coupling Reaction**
Guillaume Povie, Anh-Tuan Tran, David Bonnaffꢀ,* Jacqueline Habegger, Zhaoyu Hu,
Christine Le Narvor, and Philippe Renaud*
In memory of Andrꢀ Lubineau
Abstract: Thiol-ene coupling (TEC) reactions emerged as one
of the most useful processes for coupling different molecular
units under reaction mild conditions. However, TEC reactions
À
involving weak C H bonds (allylic and benzylic fragments)
are difficult to run and often low yielding. Mechanistic studies
demonstrate that hydrogen-atom transfer processes at allylic
and benzylic positions are responsible for the lack of efficiency
of the radical-chain process. These competing reactions cannot
be prevented, but reported herein is a method to repair the
chain process by running the reaction in the presence of
triethylborane and catechol. Under these reaction conditions,
a unique repair mechanism leads to an efficient chain reaction,
which is demonstrated with a broad range of anomeric O-allyl
sugar derivatives including mono-, di-, and tetrasaccharides
bearing various functionalities and protecting groups.
Scheme 1. Thermally initiated TEC reaction with allyl glycosides.
Cbz=benzyloxycarbonyl, THF=tetrahydrofuran.
tion product 3a in only 17% yield. To test whether such
failure was due to the presence of an azide,[6] the reaction was
repeated with the glucopyranoside 1b,[7] thus leading to
a poor 25% yield of 3b (Scheme 1). A closer look at the
literature revealed that TEC reactions involving O-allyl
derivatives or O-benzyl-protected saccharides in organic
solvents require the use of a large excess of thiols (typically
15 equivalents),[4c,8] and moderate to low yields were reported
for simple model systems.[9,10] Therefore, a detailed study of
this reaction was undertaken to develop a more efficient and
reliable procedure. We report herein that the use of Et3B as an
initiator and catechol as a co-reagent allows highly efficient
coupling reactions by an unprecedented chain-reaction repair
process.
D
iscovered more than a century ago,[1] the radical addition
of thiols to alkenes [also called the thiol-ene coupling (TEC)
reaction] became very popular during the last two decades
with applications in polymer science, biology, and bioorganic
chemistry[2] with particular emphasis on glycochemistry.[2,3]
The mild reaction conditions, atom economy, and regioselec-
tivity of the process satisfy essential requirements of the click
concept.[3d] This prompted us to evaluate TEC procedures[4]
for functionalizing the allyl moiety of fully protected frag-
ments of glycosaminoglycans (GAGs)[4a,5] with N-Cbz-pro-
tected 3-aminoethanethiol. Treatment of the allyl azidodisac-
charide 1a with 6 equivalents of the thiol 2 in the presence of
2,2’-azobis(2-methylpropionitrile) (AIBN) afforded the addi-
The TEC reaction between dodecanethiol and two simple
model allyl ethers were investigated first. Under thermal
conditions, trans-4 afforded only 62% of the expected anti-
Markovnikov adduct 5 (Scheme 2a).[11] A significant amount
of the regioisomer 6 was observed, and more surprisingly,
epimerization on the cyclohexyl ring took place to a great
extent in both 5 and 6.[12] Traces of the O,S-acetal 7 were also
isolated and the enolether 8 was observed as a transient
species during the reaction. The allyl benzyl ether 9 was
examined next (Scheme 2b). The addition of dodecanethiol
to 9 proceeded slowly and required large amounts of an
initiator to eventually furnish the adduct 10 in a moderate
55% yield along with 7% of the corresponding regioisomer
11. Benzaldehyde, 1-phenyl-1,1- (didodecylsulfanyl)methane,
and 3-dodecylsulfanylpropan-1-ol were also detected.
The results obtained with trans-4 and 9 demonstrate that
[*] Dr. G. Povie, J. Habegger, Prof. P. Renaud
Universitꢀt Bern, Departement fꢁr Chemie und Biochemie
Freiestrasse 3, 3012 Bern (Switzerland)
E-mail: philippe.renaud@dcb.unibe.ch
Dr. A.-T. Tran, Prof. D. Bonnaffꢂ, Z. Hu, Dr. C. Le Narvor
ICMMO/MS&MT, UMR 8182 (CNRS-UPS), LabEx LERMIT
Universitꢂ Paris-Sud, 91405 Orsay Cedex (France)
E-mail: david.bonnaffe@u-psud.fr
[**] The team at Bern University thanks the Swiss National Science
Foundation (grant 20020_135087) and the University of Bern for
financial support. They are also grateful to BASF Corporation for the
generous gift of Et3B. The group at Orsay, thanks the Paris Sud
University, the CNRS and the “Agence Nationale de la Recherche”
(ANR-09-PIRI-0009) for financial support, as well as the “Ministꢃre
de l’Enseignement Supꢂrieur et de la Recherche” and the “Fonda-
tion Pierre Bergꢂ endowment fun & SIDACTION” for the respective
PhD grants of A.-T.T and Z.H.
the well-established hydrogen atom abstraction from weak
[13]
À
C H bonds by thiyl radicals represents the major issue in
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201309984.
TEC reactions involving allylic ene partners or benzyl-
3894
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 3894 –3898

Angewandte
Chemie
Scheme 2. Thermally initiated TEC reaction with allyl ethers trans-4 and
9. DTBHN=di-tert-butylhyponitrite.
protected substrates. The stable allylic and benzylic radicals
thus generated accumulate and eventually terminate the
chain through recombination or disproportionation reac-
tions.[14] We anticipated that reduction of the undesired
stabilized radicals by an external reagent could regenerate
the starting substrate and repair the radical chain process.
However, this external reagent has to meet very strict
requirements: 1) it should have a weakly bound hydrogen
atom to permit hydrogen-atom transfer to stabilized allylic
and benzylic radicals; 2) the resulting radical should not react
with the alkene but should be readily converted into the
intermediate thiyl radical to regenerate the chain process;
3) if possible, it should possess a low toxicity and be
commercially available and cheap. Based on these require-
ments, we decided to investigate the use of catechols in the
presence of triethylborane.[15]
Scheme 3. The catechol/BEt3 TEC reaction with allyl ethers trans-4, 9,
12, and 14. Reaction conditions: C12H25SH (1.5 equiv), Et3B
(1.2 equiv), TBC (1.2 equiv), CH2Cl2, RT, 4 h.
small amount of the bis(addition) products 16 (11%) and 17
(3%).[16]
The optimized reaction conditions were tested with a wide
range of allyl glycosides including mono-, di-, and tetrasac-
charides (Scheme 4).[17] Within 4 hours, full conversion and
excellent yields of the isolated products were obtained[18] by
using only two equivalents of the thiol on a reaction scale
ranging from 40–80 mg of the allyl glycosides to 1.16 g of
disaccharide 1 f. This procedure proved to be superior to the
thermal reactions (compare to results in Scheme 1: 1a: 91%
versus 17%; 1b: 83% versus 25%).[19,20]
The TEC reaction between dodecanethiol and trans-4 was
investigated first (Scheme 3a). For this study, 4-tert-butylca-
techol (TBC) was preferred to catechol because of its better
solubility in organic solvents.[15] To our great satisfaction, the
reaction proceeded in almost quantitative yields with good
selectivity for the anti-Markovnikov adduct when the reaction
was run in the presence of TBC (1.2 equiv) and Et3B
(1.3 equiv). Under these reaction conditions (room temper-
ature), the trans/cis isomerization process was almost com-
pletely suppressed (trans/cis > 200:1). A similar result was
observed with the model substrate 9. By running the reaction
under the optimized Et3B/catechol conditions, a yield of 83%
for the formation of 10 was obtained and only 2% of the
regioisomer 11 were detected (Scheme 3b). The reaction was
also tested with 2-methylallyl benzyl ether (12), which was
converted into 13 in 93% yield as a single product (Scheme
3c). The radical nature of the reaction was unambiguously
demonstrated by running the reaction with the radical probe
14 (Scheme 3d). The reaction afforded the furan 15, resulting
from a 5-exo-trig cyclization process, in 80% together with
Alkylthiyl radicals add to nonactivated terminal olefins at
rates close to kadd = 106 mÀ1 sÀ1 (ca. 300 K).[21] The reverse
fragmentation process is much slower (kfrag = 105 sÀ1; ca.
300 K).[22] Thiols are excellent hydrogen-atom donors
toward alkyl radicals (kH = 107 mÀ1 sÀ1, ca. 300 K).[23] Thus,
assuming a moderately high concentration of the thiol
(ꢀ0.1m), the reduction of the 2-(alkylthio)alkyl radical will
be faster than its fragmentation. Moreover, the high rate
constants for the two elementary steps of the transformation
predict a long-chain mechanism leading to fast, selective, and
efficient reactions.[24] However, hydrogen atom abstraction by
thiyl radicals from allylic[25] and benzylic positions[26] may
cause the unexpected failures of TEC reactions. Hydrogen-
À
atom transfer from C H bonds adjacent to a heteroatom is
particularly efficient (rate constants 103–107 mÀ1 sÀ1).[27] The
À
S H bond dissociation energies (BDE) of alkanethiols are
around 87–88 kcalmolÀ1. Thus, the hydrogen abstraction by
À
an alkanethiyl radical from allylic and benzylic C H bonds
(BDE 79–87 kcalmolÀ1)[28] is exothermic and the equilibrium
Angew. Chem. Int. Ed. 2014, 53, 3894 –3898
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3895

.
Angewandte
Communications
solution are sufficient to drive all reactions to completion.
Our data suggest that an efficient repair process is occurring
in which the undesired allylic radical and the thiol are
converted back into the starting allyl ether and the thiyl
radical (Scheme 5, repair process). The repair process de-
scribed in Scheme 5 for allylic radicals is expected to be valid
for any kind of radicals resulting from undesired hydrogen-
atom transfer such as benzyl radical and a-heteroatom-
substituted alkyl radicals.
Addition of triethylborane to a solution of alkene,
dodecanethiol, and 4-tert-butylcatechol starts an important
ethane evolution, even in carefully degased solutions.[30]
When a RSH/TBC/Et3B 1:1:1 ratio is used, a new borinic
ester attributed to the structure 18 is the sole compound
observed by 11B NMR spectroscopy (d = 56.3 ppm). Under
our reaction conditions, all the catechol is converted into 18
according to Equations (1) and (2) (Scheme 6). The mono-
Scheme 4. The catechol/BEt3 TEC reaction with allyl glycosides.
PMB=para-methoxybenzyl.
lies on the side of the C-centered radical. The main
termination reactions involve coupling and disproportiona-
tion of stabilized allylic and benzylic radicals (Scheme 5,
chain disruption process).[29] The long-chain mechanism
operating under the optimized Et3B/catechol conditions is
particularly striking since only traces of oxygen in a degased
Scheme 6. Repair mechanism operating in the catechol/BEt3 TEC
reaction.
borinate 18 acts as a hydrogen donor with allylic (or benzylic)
radicals [Eq. (3)]. This step is close to thermoneutral or
À
slightly exothermic (BDE of O H bonds in catechols lie close
to 80 kcalmolÀ1).[31] After the hydrogen-atom transfer step,
the resulting aryloxyl radical undergoes a rapid 5-exo
cyclization which produces an ethyl radical [Eq. (4)], which
reinitiates the chain upon reaction with the thiol [Eq. (5)].[15]
At the end of the reaction with trans-4, the 11B NMR
spectrum reveals that around 15% of 18 has been converted
into the B-ethylcatecholborane 19 (d = 35.8 ppm).[32]
The repair mechanism proposed here works best with 1-
alkoxyallyl radicals if the reduction regioselectively provides
the allylether over the enolether (Scheme 7, top). The
reduction of the allyl radical is expected to be faster at the
most-electron-rich carbon atom, thus favoring the regener-
ation of the allylether over the isomerization to the eno-
lether.[33] Interestingly, according to the Curtin–Hammett
principle, a nonregioselective hydrogen-transfer could also
explain the repair mechanism (Scheme 7, bottom).
In conclusion, the high reactivity of thiyl radicals induces
unwanted degradations by hydrogen abstraction, thus result-
ing in disruption of the chain reaction. To overcome this
Scheme 5. Disruption and repair processes.
3896 www.angewandte.org
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 3894 –3898

Angewandte
Chemie
[9] a) M. Gꢁmez-Garcꢂa, J. M. Benito, R. Gutiꢀrrez-Gallego, A.
Maestre, C. Ortiz Mellet, J. M. Garcꢂa Fernꢃndez, J. L. Jimꢀnez
Blanco, Org. Biomol. Chem. 2010, 8, 1849 – 1860; b) A. Siriwar-
dena, M. R. Jørgensen, M. A. Wolfert, M. L. Vandenplas, J. N.
Moore, G. Boons, J. Am. Chem. Soc. 2001, 123, 8145 – 8146; c) G.
Triola, L. Brunsveld, H. Waldmann, J. Org. Chem. 2008, 73,
3646 – 3649; d) N. Merbouh, F. K. Wallner, O. M. Cociorva, P. H.
Seeberger, Org. Lett. 2007, 9, 651 – 653.
[10] The best results involving allyl ethers were obtained under
photoinitiation: M. Fiore, A. Chambery, A. Marra, A. Dondoni,
Org. Biomol. Chem. 2009, 7, 3910 – 3913; M. Lo Conte, S.
Staderini, A. Chambery, N. Berthet, P. Dumy, O. Renaudet, A.
Marra, A. Dondoni, Org. Biomol. Chem. 2012, 10, 3269 – 3277.
Under these reaction conditions, the TEC reaction involving
substrates 1c and 1 f with thiol 2 afforded the desired products
3c and 3 f in only 50% and 48% yields, respectively, along with
degradation products. These results were anticipated since
benzyl ethers are reported to be unstable under such photo-
initiation conditions (see Ref. [4b]).
Scheme 7. Regioselectivity issue and the origin of the repair mecha-
nism.
[11] At the dilution of 0.3m used in all experiments, the concentration
in thiol at the end of the reaction is 0.15m, thus giving a pseudo-
first-order rate of hydrogen-atom transfer to an alkyl radical of
around 3 ꢄ 106 sÀ1, which is still more than ten times faster than
the fragmentation of the 2-(alkylthio)alkyl radical intermediate.
[12] The partial epimerization was attributed to a reversible hydro-
gen-atom transfer at either the benzylic or the a-O position.
[13] F. Dꢀnꢅs, M. Pichowicz, G. Povie, P. Renaud, Chem. Rev. 2014,
DOI: 10.1021/cr400441m.
problem, we developed a new TEC procedure which allows
restoration of the chain process. Although the concept of the
repair process is well known in biological systems, the
procedure described here represents a unique example
where a repair process is used to increase the efficiency of
a synthetic method involving a radical chain reaction.
Received: November 17, 2013
Revised: January 11, 2014
Published online: March 11, 2014
[14] Dimers and oligomers are not observed in gas chromatography
1
analyses. However the H NMR spectra of some side products
supports dimeric structures although none of them could be
isolated in pure form for characterization.
[15] G. Villa, G. Povie, P. Renaud, J. Am. Chem. Soc. 2011, 133, 5913 –
5920.
[16] The TEC reaction with a second radical probe, (1-cyclopropyl-
vinyl)cyclopropane, was also investigated and confirmed further
the radical nature of the process when the reaction was
performed in presence of catechol and triethylborane (see the
Supporting Information).
[17] The allyl glycosides 1a–i were prepared according to reported
procedures. See the Supporting Information for references.
[18] See the Supporting Information.
Keywords: allylic compounds · boranes · glycosides ·
radical reactions · synthetic methods
.
[1] T. Posner, Chem. Ber. Dtsch. Chem. Ges. 1905, 38, 646 – 657.
[2] a) C. E. Hoyle, T. Y. Lee, T. M. Roper, J. Polym. Sci. Part A 2004,
42, 5301 – 5338; b) A. Dondoni, Angew. Chem. 2008, 120, 9133 –
9135; Angew. Chem. Int. Ed. 2008, 47, 8995 – 8997; c) K. L.
Killops, L. M. Campos, C. J. Hawker, J. Am. Chem. Soc. 2008,
130, 5062 – 5064; d) C. E. Hoyle, A. B. Lowe, C. N. Bowman,
Chem. Soc. Rev. 2010, 39, 1355 – 1387; e) C. Hoyle, C. Bowman,
Angew. Chem. 2010, 122, 1584 – 1617; Angew. Chem. Int. Ed.
2010, 49, 1540 – 1573.
[3] a) J. J. Lundquist, E. J. Toone, Chem. Rev. 2002, 102, 555 – 578;
b) J. M. de La Fuente, S. Penades, Biochim. Biophys. Acta Gen.
Subj. 2006, 1760, 636 – 651; c) A. Dondoni, A. Marra, Chem. Soc.
Rev. 2012, 41, 573 – 586; d) H. C. Kolb, M. G. Finn, K. B.
Sharpless, Angew. Chem. 2001, 113, 2056 – 2075; Angew. Chem.
Int. Ed. 2001, 40, 2004 – 2021.
[4] a) A. Lubineau, H. Lortat-Jacob, O. Gavard, S. Sarrazin, D.
Bonnaffꢀ, Chem. Eur. J. 2004, 10, 4265 – 4282; b) M. Fiore, A.
Marra, A. Dondoni, J. Org. Chem. 2009, 74, 4422 – 4425; c) P. B.
van Seeventer, J. van Dorst, J. F. Siemerink, J. P. Kamerling,
J. F. G. Vliegenthart, Carbohydr. Res. 1997, 300, 369 – 373.
[5] a) A. Dilhas, R. Lucas, L. Loureiro-Morais, Y. Hersant, D.
Bonnaffꢀ, J. Comb. Chem. 2008, 10, 166 – 169; b) D. Bonnaffꢀ, C.
R. Chim. 2011, 14, 59 – 73.
[19] See the procedure from Oshima et al. (see Ref. [20a]) as it
involves Et3B initiation in methanol to give yields between 30%
and 50%. Under these reaction conditions, full conversion
required a much longer reaction time (ꢀ 24 h) compared to the
reaction in the presence of catechol (ꢁ 4 h).
[20] a) K. Oshima, Y. Ishinose, K. Wakamatsu, K. Nozaki, J. L.
Birbaum, K. Utimoto, Chem. Lett. 1987, 1647 – 1650; b) J.
Taillades, P. Rey, M. Garrait, G. Gros, Fr. Patent FR2 824 823,
2002; J. Taillades, P. Rey, M. Garrait, G. Gros, Eur. Patent EPI
260 500, 2002.
[21] a) C. Sivertz, J. Phys. Chem. 1959, 63, 34 – 38; b) D. J. McPhee,
M. Campredon, M. Lesage, D. Griller, J. Am. Chem. Soc. 1989,
111, 7563 – 7567.
[22] a) C. Walling, W. Helmreich, J. Am. Chem. Soc. 1959, 81, 1144 –
1148; b) C. Chatgilialoglu, A. Altieri, H. Fischer, J. Am. Chem.
Soc. 2002, 124, 12816 – 12823.
[23] E. T. Denisov, C. Chatgilialoglu, A. Shestakov, T. G. Denisova,
Int. J. Chem. Kinet. 2009, 41, 284 – 293.
[6] D. Horton, A. E. Luetzow, J. C. Wease, Carbohydr. Res. 1968, 8,
366 – 367.
[24] B. H. Northrop, R. N. Coffey, J. Am. Chem. Soc. 2012, 134,
13804 – 13817.
[7] A. Lubineau, S. Escher, J. Alais, D. Bonnaffꢀ, Tetrahedron Lett.
1997, 38, 4087 – 4090.
[25] a) M. P. Bertrand, S. Escoubet, S. Gastaldi, V. I. Timokhin,
Chem. Commun. 2002, 216 – 217; b) S. Escoubet, S. Gastaldi,
V. I. Timokhin, M. P. Bertrand, D. Siri, J. Am. Chem. Soc. 2004,
126, 12343 – 12352.
[8] a) C. Heidecke, T. Lindhorst, Chem. Eur. J. 2007, 13, 9056 – 9067;
b) K. Matsuoka, H. Oka, T. Koyama, Y. Esumi, D. Terunuma,
Tetrahedron Lett. 2001, 42, 3327 – 3330; c) R. T. Lee, Y. C. Lee,
Carbohydr. Res. 1974, 37, 193 – 201.
Angew. Chem. Int. Ed. 2014, 53, 3894 –3898
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3897

.
Angewandte
Communications
[26] a) S. Escoubet, S. Gastaldi, N. Vanthuyne, G. Gil, D. Siri, M. P.
Bertrand, Eur. J. Org. Chem. 2006, 3242 – 3250; b) K. Qvortrup,
D. A. Rankic, D. W. C. MacMillan, J. Am. Chem. Soc. 2013, 135,
18323 – 18326.
[30] a) B. M. Mikhailov, Y. N. Bubnov, Zh. Obshch. Khim. 1961, 31,
150 – 155; b) B. M. Mikhailov, Prog. Boron Chem. 1970, 3, 313 –
370; c) A. G. Davies, B. P. Roberts, J. Chem. Soc. B 1971, 1830 –
1837.
[27] a) W. A. Pryor, G. Gojon, D. F. Church, J. Org. Chem. 1978, 43,
793 – 800; b) C. Schçneich, K. D. Asmus, J. Chem. Soc. Faraday
Trans. 1995, 91, 1923 – 1930; c) T. Nauser, C. Schçneich, J. Am.
Chem. Soc. 2003, 125, 2042 – 2043; d) T. Nauser, C. Schçneich,
Chem. Res. Toxicol. 2003, 16, 1056 – 1061; e) A. J. Fielding, B. P.
Roberts, Tetrahedron Lett. 2001, 42, 4061 – 4064.
[28] Y.-R. Luo, Handbook of Bond Dissociation Energies in Organic
Compounds, CRC Press, Boca Raton, 2002.
[29] The observation of the enolether 8 as a transient intermediate
during the reaction described in Scheme 3 indicates that the
formation of allylic radicals is also taking place in the presence of
tert-butylcatechol.
[31] M. Lucarini, G. F. Pedulli, M. Guerra, Chem. Eur. J. 2004, 10,
933 – 939.
[32] The boron-containing compounds were identified by compar-
ison of their 11B NMR signals with those of authentic samples
prepared from catechol.
[33] Based on spin density distribution, no regioselectivity is
expected since both allylic positions of the 1-methoxyallyl
radical are nearly equally populated: Landolt-Bçrnstein Numer-
ical Data and Functional Relationships in Science and Technol-
ogy, Vol. 9, part b (Eds.: H. Fischer, K.-H. Hellwege), Springer,
Berlin, 1977, p. 362.
3898
www.angewandte.org
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 3894 –3898