Ruthenium-Catalyzed O- to S-Alkyl Migration: A Pseudoreversible Barton-McCombie Pathway

Article abstract of DOI:10.1002/anie.201505280

A practical ruthenium-catalyzed O- to S-alkyl migration affords structurally diverse thiooxazolidinones in excellent yields. Our studies suggest this catalytic transformation proceeds through a pseudoreversible radical pathway drawing mechanistic parallels to the classic Barton-McCombie reaction. A radical step in a new direction: A practical ruthenium-catalyzed O- to S-alkyl migration affords structurally diverse thiooxazolidinones in excellent yields. Experimental and computational studies suggest a pseudoreversible radical pathway drawing mechanistic parallels to the classic Barton-McCombie reaction.

Full text of DOI:10.1002/anie.201505280

.
Angewandte
Communications
International Edition: DOI: 10.1002/anie.201505280
German Edition: DOI: 10.1002/ange.201505280
Pseudoreversible Reactions
Ruthenium-Catalyzed O- to S-Alkyl Migration: A Pseudoreversible
Barton–McCombie Pathway
William Mahy, Pawel Plucinski, Jesffls Jover, and Christopher G. Frost*
Abstract: A practical ruthenium-catalyzed O- to S-alkyl
migration affords structurally diverse thiooxazolidinones in
excellent yields. Our studies suggest this catalytic transforma-
tion proceeds through a pseudoreversible radical pathway
drawing mechanistic parallels to the classic Barton–McCombie
reaction.
T
he prevalence of sulfur-containing functionality in chemical
biology, pharmaceutical drugs, agrochemicals, and material
science is of fundamental importance and dictates the
development and application of new synthetic methods in
organosulfur chemistry. Historically, O-thiocarbamates have
proved valuable as directing groups for the chemoselective
reduction of alcohol functional groups to the corresponding
alkane (Barton–McCombie reaction)[1] and also in the gen-
À
eration of Ar S compounds from phenols through an OAr to
Scheme 1. Conceptual rationale for a catalytic O- to S-alkyl migration.
Table 1: Selected screening results.
SAr migration (Newman–Kwart rearrangement).[2] The emer-
gence of efficient metal-mediated techniques for the con-
struction and transformation of carbon–sulfur bonds provides
valuable tools for contemporary organic synthesis.[3] In this
context, Lloyd-Jones et al. have reported an elegant palla-
dium-catalyzed variant of the Newman–Kwart rearrangement
facilitating rearrangements at lower temperatures through an
oxidative addition/tautomerization/reductive elimination
sequence.[4] Herein, we report a selective ruthenium-cata-
lyzed Oalkyl to Salkyl migration that occurs when the O-
thiocarbamate functionality is constrained in a readily avail-
able ring structure (Scheme 1).[5] Our studies suggest this
catalytic transformation proceeds through a pseudoreversible
radical pathway drawing mechanistic parallels to the classic
Barton–McCombie reaction.[6]
Entry
Catalyst (mol%)
Solvent
t [h]
Conv. [%][b]
1
–
PhMe
PhMe
PhMe
PhMe
PhMe
THF
MeCN
PhMe
PhMe
PhMe
1
3
1
1
1
3
3
3
3
3
–
2[c]
3
DTBP (10)
Pd(OAc)2 (10)
6 (8)[d]
7 (11)[d]
5 (–)[d]
82 (–)[d]
12
4
5
NiBr2(PPh3)2 (10)
[RuCl2(p-cymene)]2 (5)
[RuCl2(p-cymene)]2 (1)
[RuCl2(p-cymene)]2 (1)
[RuCl2(p-cymene)]2 (1)
[RuCl2(p-cymene)]2 (1)
[RuCl2(p-cymene)]2 (1)
6[e]
7[e]
8[e]
9[f]
10[f,g]
At the onset of our investigation, we explored the O- to S-
alkyl migration of N-phenyloxazolidine-2-thione (1a) under
traditional Barton–McCombie deoxygenation conditions
6
17
79
100
using
a
variety of organic-based radical initiators
(Table 1).[7] The use of a number of transition-metal catalysts
(10 mol% equivalence of metal) known for their catalytic
[a] Reaction conditions: 1a (0.25 mmol), catalyst, toluene (2.5 mL),
1008C, air. Yields of isolated products are given. [b] Conversion of 1b was
1
determined by H NMR spectroscopy. [c] Additional 1.0 equiv Bu3SnH
used. [d] Conversion to N-phenyl oxazolidinone. [e] Reactions performed
at 708C. [f] 2 mol% SPhos. [g] Toluene (1.25 mL). Full screening results
reported in the SI.
[*] W. Mahy, Prof. Dr. C. G. Frost
Department of Chemistry, University of Bath
Bath, BA2 7AY (UK)
E-mail: c.g.frost@bath.ac.uk
activity in single electron transfer processes were also
investigated.[8] [RuCl2(p-cymene)]2 showed exceptional reac-
tivity and selectivity toward O- to S-alkyl migration, giving
rise to 82% 2a after 1 hour, during which no observable
desulfonated N-phenyl oxazolidinone from the competitive
radical pathway was observed. Given that ligation of ruthe-
nium species is well known to improve both physical and
chemical properties in ruthenium-mediated transformations
Dr. P. Plucinski
Department of Chemical Engineering, University of Bath
Bath, BA2 7AY (UK)
Dr. J. Jover
Departament de Química Inorgànica, Universitat de Barcelona
C/Martí i Franqu›s 1–11, 08028 Barcelona (Spain)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201505280.
10944
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2015, 54, 10944 –10948

Angewandte
Chemie
we tested a range of mono- and bidentate ligands in
combination with 1 mol% [RuCl2(p-cymene)]2 (2 mol%
Ru) in toluene at 1008C.[9] Monodentate biaryl phosphines
proved most effective, and SPhos gave the highest improved
reactivity after 3 h, giving 81% conversion of 1a. The effect of
concentration was then explored and higher concentrations
were found to be by far the most effective, allowing
quantitative conversion of 1a at 1008C in 3 h with just
1 mol% of the Ru dimer. In the absence of catalyst there was
no detectable rearrangement.
With the optimized reaction conditions in hand we turned
our attention to the scope and limitations of this ruthenium-
promoted rearrangement. To our delight a wide range of N-
aryl oxazolidine-2-thiones[10] were tolerated under the reac-
tion conditions, affording the corresponding N-aryl thiazoli-
din-2-ones in excellent yields (Scheme 2).
no dehalogenation was observed. A trend in the electronic
nature of the aryl substituents was observed, where strongly
electron-rich and strongly electron-poor aromatics showed
lower rates with respect to the unsubstituted aromatic; this
effect was most pronounced with electron-poor systems, for
which an increase in catalytic loading to 2 mol% of [RuCl2(p-
cymene)]2 (4 mol% Ru) was required to afford improved
yields (products 2 f, 2g, and 2m). Enantiomerically pure 4-
substituted oxazolidine-2-thiones also showed excellent reac-
tivity under the reaction conditions, affording the correspond-
ing enantiopure 4-substituted thiazolidin-2-ones with com-
plete enantiomeric retention (2n–s).
The 5-substituted oxazolidine-2-thiones were significantly
less reactive. Monosubstitution at this position required
elevated temperatures (1308C), increased catalyst loading
(5 mol%), and extended reaction times (24 h) to afford
modest conversion (2t–v), a slight erosion of the enantiopur-
ity (91% ee) was also observed. Geminal substituents showed
an even more pronounced reduction in reactivity (2w–x).
Phenol derivatives (2y), more closely mimicking substrates
used in Newman–Kwart rearrangements showed no conver-
sion even at more forcing conditions.[2]
Notably, the reaction conditions are tolerant of a number
of valuable and reactive functional groups, including Br, CN,
CF3, OBn, and F (products 2e–h, 2l), in the case of 2e and 2l
Subsequently we turned our attention to the rearrange-
ment
of
N-alkyl-substituted
oxazolidine-2-thiones
(Scheme 3). In all cases the corresponding N-alkyl thiazoli-
din-2-ones were formed in good to excellent yields. However,
the reaction using an anthracene derivative gave only
moderate conversion, which could be ascribed to the poor
solubility of the substrate in toluene.
Scheme 2. Synthesis of N-aryl thiazolidin-2-ones from a variety of N-
aryl oxazolidine-2-thiones. Reaction conditions: N-aryl oxazolidine-2-
thiones (0.25 mmol), [RuCl2(p-cymene)]2 (1 mol%), SPhos (2 mol%),
toluene (1.25 mL), 1008C, 3 h, air. Yields of isolated products are
given. [a] [RuCl2(p-cymene)]2 (2 mol%), SPhos (4 mol%) used.
[b] [RuCl2(p-cymene)]2 (5 mol%), SPhos (10 mol%), 1308C, 24 h used.
[c] Conversion calculated from 1H NMR spectra.
Scheme 3. Synthesis of N-alkyl thiazolidin-2-ones from a variety of N-
alkyl oxazolidine-2-thiones. Reaction conditions: N-aryl oxazolidine-2-
thiones (0.5 mmol), [RuCl2(p-cymene)]2 (1 mol%), SPhos (2 mol%),
toluene (2.5 mL), 1008C, 4 h, air. Yields of isolated products are given.
Angew. Chem. Int. Ed. 2015, 54, 10944 –10948
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 10945

.
Angewandte
Communications
We next examined the rearrangement methodology on
a multigram scale to probe the scalability of the process. N-
phenyl oxazolidine-2-thione was subjected to the ruthenium-
catalyzed O- to S-alkyl migration with 1 mol% [RuCl2(p-
cymene)]2 under air using the standard conditions
(Scheme 4). Isolation of the target thiazolidin-2-one was
oxo products when reacted under catalytic radically promoted
conditions. Whilst only organic-based radical initiators were
investigated, the choice of radical initiator and promoter were
shown to have a significant effect on reactivity as well as
selectivity. In contrast to ionic-based mechanisms, both
secondary and primary rearrangement products were
observed in addition to the oxo derivative produced by the
reaction of atmospheric oxygen and an intermediate radical.
To probe consistencies with the ruthenium-catalyzed O-
to S-alkyl migration of oxazolidine-2-thiones and this
reported radical mechanism, [RuCl2(p-cymene)]2 and cyclic
thiocarbonate 5a were reacted using the standard conditions
(Scheme 6). As with previous reported works, a distribution
Scheme 4. Multigram synthesis of 2a. Reaction conditions: N-aryl
oxazolidine-2-thiones (11.16 mmol), [RuCl2(p-cymene)]2 (1 mol%),
SPhos (2 mol%), toluene (55.8 mL), 1008C, 3 h, air. Yields of isolated
products are given
achieved in a similarly high yield of 96% following column
chromatography.
To determine the mechanism of the reaction catalyzed by
[RuCl2(p-cymene)]2, a number of kinetic and control experi-
ments were performed. Cross-over experiments showed no
formation of cross-over products when reacted under the
optimized conditions. Reaction rates displayed a first-order
dependency in both [1a] and [Ru] which eliminates a possible
polymerization/depolymerization process reported previ-
ously in the thermally mediated rearrangement of N-phenyl-
oxazolidine-2-thione.[11]
The isomerization of cyclic thiocarbonate esters in the
presence of potassium iodide is known to selectively generate
primary monothiolcarbonates through an ionic pathway.[12] To
investigate the reactivity of N-phenyl oxazolidine-2-thiones
under ionic rearrangement conditions, 1a was heated in the
presence of catalytic potassium iodide (50 mol%) and the
reaction was monitored over time (Scheme 5). Analysis of the
Scheme 6. Reaction of cyclic thiocarbonate 5a. Reaction conditions:
thiocarbonate (0.25 mmol), RuCl2(p-cymene)]2 (1 mol%), SPhos
(2 mol%), toluene (1.25 mL), 1008C, 3 h, air. Ratio of products was
1
determined by H NMR spectroscopy.
of products was observed, including the radical desulfuriza-
tion product (6c, 4%). Interestingly, both secondary and
primary rearrangement products (6a and 6b) were observed
in a 3:1 ratio, well within the ratios observed for classical
radical-promoted rearrangements. The observed consisten-
cies would strongly suggest that the presented ruthenium-
catalyzed system is capable of proceeding through a similar
radical-promoted reaction pathway.[14]
Analysis of the ruthenium-catalyzed O- to S-alkyl migra-
tion by DFT (see the Supporting Information (SI) for details)
supports the proposed reaction proceeding through radical
adducts (Scheme 7).
Thus, the formation of the starting [RuCl2(p-cymene)-
(SPhos)] complex (S0) is followed by a thermoneutral loss of
the p-cymene ligand, generating the catalytic species S1. The
reaction proceeds with the coordination of the N-aryl
oxazolidine-2-thione (1) through the sulfur atom to give rise
to complex S2. The reaction interchanges then from the
singlet to the triplet energy surface through the minimum
energy crossing point MECP_S-T, this transformation
requires approximately 10 kcalmolÀ1. Subsequently, one
electron is transferred from the metal to the substrate
Scheme 5. Treatment of N-phenyl oxazolidine-2-thione under ionic
conditions. Reaction conditions: N-phenyl oxazolidine-2-thione
(0.25 mmol), KI (0.125 mmol), toluene (1.25 mL), 1008C, air. Reaction
1
was monitored by H NMR spectroscopy.
crude reaction mixtures by 1H NMR spectroscopy showed
that even after extended reaction times no rearrangement of
the oxazolidine-2-thione occurred. Therefore it was con-
cluded that it is unlikely that the reaction proceeds through
a nucleophilic ring opening/recombination process.
It has been reported that the treatment of cyclic thiocar-
bonates under Barton–McCombie-promoted conditions
results in the formation of the O- to S-rearrangement product
when catalytic amounts of the radical promoter is used.[13]
Tsuda et al. found that cyclic thiocarbonates derived from
glycosides gave a distribution of rearrangement, deoxy, and
À
(SET1), automatically provoking the cleavage of the C O
bond and generating the RuIII diradical species T1. The single-
electron transfer process requires 12 kcalmolÀ1 but still
remains at a quite reasonable height.
Once T1 is obtained, the pendant radical rotates through
the corresponding transition state (Rot_TS), which is less
than 7 kcalmolÀ1 higher than the previous intermediate, to
form the diradical complex T2. The second electron transfer
À
process (SET2) generates the C S bond, the reaction then
10946 www.angewandte.org
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2015, 54, 10944 –10948

Angewandte
Chemie
Acknowledgements
We are grateful to the University of Bath, EPSRC DTC in
Sustainable Chemical Technologies for funding (W.M.). We
are also grateful for the valuable assistance of Anneke
Lubben (mass spectrometry) and John Lowe (NMR spec-
troscopy), and Dr. Dave Carbery and Christina Gulacsy
(Chiral HPLC, University of Bath).
Keywords: alkyl migration · Barton–McCombie reaction ·
catalysis · O-thiocarbamates · pseudoreversible reaction
How to cite: Angew. Chem. Int. Ed. 2015, 54, 10944–10948
Angew. Chem. 2015, 127, 11094–11098
Scheme 7. Proposed model for ruthenium-catalyzed O- to S-alkyl
migration. Numbers are DFT-derived free energies (in kcalmolÀ1) at
1008C.
[1] Original paper: a) D. H. R. Barton, S. W. McCombie, J. Chem.
Soc. Perkin Trans. 1 1975, 1574 – 1585; Selected reviews: b) D.
Crich, L. Quintero, Chem. Rev. 1989, 89, 1413 – 1432; c) S. W.
McCombie, W. B. Motherwell, M. J. Tozer, Org. React. 2012, 77,
161 – 432; d) S. W. McCombie in Comprehensive Organic Syn-
thesis, Vol. 8 (Ed.: B. M. Trost), Pergamon, New York, 1991,
chap. 4.2; e) L. Chenneberg, J.-P. Goddard, L. Fensterbank in
Comprehensive Organic Synthesis II (Eds.: P. Knochel, G. A.
Molander), Elsevier, Amsterdam, 2014, pp. 1011 – 1030.
[2] a) C. Zonta, O. De Lucchi, R. Volpicelli, Top. Curr. Chem. 2007,
275, 131; b) G. C. Lloyd-Jones, J. D. Moseley, J. S. Renny,
Synthesis 2008, 661.
[3] Selected reviews: a) T. Kondo, T. Mitsudo, Chem. Rev. 2000, 100,
3205 – 3220; b) S. V. Ley, A. W. Thomas, Angew. Chem. Int. Ed.
2003, 42, 5400 – 5449; Angew. Chem. 2003, 115, 5558 – 5607;
c) I. P. Beletskaya, V. Ananikov, Chem. Rev. 2011, 111, 1596 –
1636; d) A. S. Deeming, E. J. Emmett, C. S. Richards-Taylor,
M. C. Willis, Synthesis 2014, 2701 – 2710.
returns to the singlet energy surface through the correspond-
ing minimum energy crossing point (MECP_T-S), giving rise
to the RuII intermediate S3. Finally, the product is liberated
into the reaction mixture and the starting catalytic species is
regenerated. The computed Gibbs free energy of the catalytic
cycle is À14.4 kcalmolÀ1, indicating that the whole process is
thermodynamically favored. The computed overall barrier for
the reaction, obtained as the free-energy difference between
S2 and SET1, is approximately 23 kcalmolÀ1, which is highly
plausible for a reaction occurring at 1008C.
In addition, a classical, non-radical, oxidative addition/
À
[4] J. N. Harvey, J. Jover, G. C. Lloyd-Jones, J. D. Moseley, P.
Murray, J. S. Renny, Angew. Chem. Int. Ed. 2009, 48, 7612 –
7615; Angew. Chem. 2009, 121, 7748 – 7751.
[5] For a ring-opening polymerization – depolymerization process,
see: T. Mukaiyama, I. Kuwajima, K. Mizui, J. Org. Chem. 1966,
31, 32 – 34.
[6] For a review of complementary carbon – sulfur bond formation
processes by catalytic allylation of sulfur nucleophiles, see: W.
Liu, X. Zhao, Synthesis 2013, 2051 – 2069.
[7] a) D. H. R. Barton, D. O. Jang, J. C. Jaszberenyi, Tetrahedron
1993, 49, 2793 – 2804; b) D. H. R. Barton, D. O. Jang, J. C.
Jaszberenyi, Tetrahedron Lett. 1992, 33, 2311 – 2314; c) D. H. R.
Barton, D. Crich, A. Lobberding, S. Z. Zard, Tetrahedron 1986,
42, 2329 – 2338.
[8] Selected reviews: a) G. J. Rowlands, Tetrahedron 2009, 65, 8603 –
8655; b) J. Iqbal, B. Bhatia, N. K. Nayyar, Chem. Rev. 1994, 94,
519 – 564; c) M. Malacria, Chem. Rev. 1996, 96, 289 – 306; d) M.
Ouchi, T. Terashima, M. Sawamoto, Chem. Rev. 2009, 109, 4963 –
5050; e) N. Zhang, S. R. Samanta, B. M. Rosen, V. Percec, Chem.
Rev. 2014, 114, 5848 – 5958.
[9] a) C. A. Tolman, Chem. Rev. 1977, 77, 313 – 348; b) B. Li, S. Lee,
K. Shin, S. Chang, Org. Lett. 2014, 16, 2010 – 2013; c) C. Six, K.
Beck, A. Wegner, W. Leitner, Organometallics 2000, 19, 4639 –
4642; d) S. Manzini, C. A. U. Blanco, A. M. Z. Slawin, S. P.
Nolan, Organometallics 2012, 31, 6514 – 6517.
reductive elimination catalytic cycle through ruthenium C O
insertion has been also computed (see SI for details). DFT
calculations showed that the energy requirements for this
pathway are much higher than those obtained for the radical
mechanism; in this case, the massive barrier obtained for the
concerted oxidative addition step (+ 52.2 kcalmolÀ1) would
completely shut down the reaction. In addition, the rotational
transition-state barrier is also very high (+ 42.7 kcalmolÀ1
from the lowest species), likely due to the charge separation
produced in the ligand replacement process on ruthenium.
In summary, we have developed a robust and efficient
ruthenium-catalyzed O- to S-alkyl migration process for the
practical preparation of N-substituted thiazolidine-2-thiones
from readily accessible N-substituted oxazolidine-2-thiones.
The products from the reaction are useful as sulfur-implanted
heterocycles or as precursors to aminothiols which are
significantly more challenging to access than amino alcohols.
Further studies are underway to explore new applications of
catalytic O- to S-alkyl migrations in organic synthesis. Initial
experimental and computational investigations into the
mechanism of this transformation suggest a pseudoreversible
Barton–McCombie-type pathway is plausible.
[10] W. Mahy, P. Plucinski, C. G. Frost, Org. Lett. 2014, 16, 5020 –
5023.
[11] See Ref. [5].
[12] a) D. Trimnell, W. M. Doane, C. R. Russell, C. E. Rist, Carbo-
hydr. Res. 1971, 17, 319 – 326; b) D. Trimnell, W. M. Doane,
Methods in Carbohydrate Chemistry, Vol. VII, (Eds.: R. L.
Angew. Chem. Int. Ed. 2015, 54, 10944 –10948
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 10947

.
Angewandte
Communications
Whistler, J. N. BeMiller), Acadamic Press, New York, 1976,
pp. 42 – 43.
[14] For a photocatalytic Barton – McCombie deoxygenation that is
proposed to proceed through a carbon-centered radical, see: L.
Chenneberg, A. Baralle, M. Daniel, L. Fensterbank, J.-P. God-
dard, C. Ollivier, Adv. Synth. Catal. 2014, 356, 2756 – 2762.
[13] a) Y. Tsuda, Y. Sato, K. Kanemitsu, S. Hosoi, K. Shibayama, K.
Nakao, Y. Ishikawa, Chem. Pharm. Bull. 1996, 44, 1465 – 1475;
b) Y. Tsuda, S. Noguchi, K. Kanemitsu, Y. Sato, K. Kakimoto, Y.
Iwakura, S. Hosoi, Chem. Pharm. Bull. 1997, 45, 971 – 980; c) Y.
Tsuda, S. Noguchi, H. Niino, Chem. Pharm. Bull. 2001, 49, 1210 –
1213.
Received: June 9, 2015
Published online: July 27, 2015
10948 www.angewandte.org
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2015, 54, 10944 –10948