Hydrogen Atom Transfer-Mediated Cyclisations of Nitriles

Article abstract of DOI:10.1002/chem.201805236

Hydrogen atom transfer-mediated intramolecular C?C coupling reactions between alkenes and nitriles, using PhSiH₃ and catalytic Fe(acac)₃, are described. This introduces a new strategic bond disconnection for ring-closing reactions, f

Full text of DOI:10.1002/chem.201805236

A Journal of
Accepted Article
Title: Hydrogen Atom Transfer-Mediated Cyclisations of Nitriles
Authors: John Murphy, Oliver Turner, David Hirst, and Eric Talbot
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Chem. Eur. J. 10.1002/chem.201805236
Link to VoR: http://dx.doi.org/10.1002/chem.201805236
Supported by

10.1002/chem.201805236
Chemistry - A European Journal
COMMUNICATION
Hydrogen Atom Transfer-Mediated Cyclisations of Nitriles
Oliver J. Turner,*[a,b] John A. Murphy,*[b] David. J. Hirst[a] and Eric P. A. Talbot*[a]†
Abstract: Hydrogen atom transfer-mediated intramolecular C-C
coupling reactions between alkenes and nitriles, using PhSiH3 and
catalytic Fe(acac)3, are described. This introduces a new strategic
bond disconnection for ring-closing reactions, forming ketones via
imine intermediates. Of note is the scope of the reaction, including
formation of sterically hindered ketones, spirocycles and fused cyclic
systems.
as radical acceptors, despite potential issues such as β-
scission[24] and competing alkene/nitrile reduction pathways.[25]
In the early 1960s, Kwiatek and Seyler first reported the use of
metal hydrides as catalysts in the hydrogenation of α,β-
unsaturated compounds.[1,2] The discovery by Halpern,[3] later
elegantly developed by Norton,[4] that metal-hydride hydrogen
atom transfer (HAT) proceeded by a free-radical mechanism
opened the door to a wide range of alkene hydrofunctionalisation
reactions. But it was the pioneering work by Mukaiyama[5] on the
catalytic hydration of alkenes, using Co(acac)2 and oxygen, that
sparked wider interest in the field of alkene hydrofunctionalisation.
As a result, there now exists an extensive ‘toolkit’ for the addition
of hydrogen and a functional group to an alkene with Markovnikov
selectivity and high chemo-selectivity using cobalt, manganese
and iron complexes.[6,7] Efforts have also been made to extend
HAT methodologies to C-C bond formation, both in an intra- and
intermolecular fashion: Baran’s group developed a general C-C
coupling reaction, utilising electron-deficient alkenes as capable
radical acceptors (Scheme 1ai).[8–10] Hydropyridylation of alkenes
by intramolecular Minisci reaction was recently demonstrated by
Starr,[11] which allows for the formation of structures such as
dihydropyranopyridines (Scheme 1aii). Furthermore, whilst
conducting the work described in this paper, Bonjoch showed that
ketones were able to undergo radical cyclisation to their tertiary
alcohol counterparts (Scheme 1aiii).[12]
Scheme 1. a) Previous examples of HAT-mediated C-C bond forming
cyclisation reactions, b) the outline of this work and its challenges.
Yet, despite the use of radical acceptors such as
acrylonitrile,[9] the HAT-initiated radical addition to nitriles has
remained unreported. Whilst radical cyclisation onto a C≡N π-
bond is feasible, it is about 50 times slower than its C=C and C≡C
variants.[13] This is highlighted by the numerous reported failed
attempts to cyclise onto a nitrile group;[14–16] although in isolated
examples, radical cyclisations to nitriles have been achieved
utilising tributyltin hydride,[17–19] titanium[20–22] and manganese[23]
reagents.
During studies of conventional hydrofunctionalisation
chemistry with alkene-nitrile 1, an interesting observation was
made (Scheme 2). Rather than the radical intermediate A reacting
with the electrophilic radicophile TsCN in an intermolecular
fashion, leading to product B, the radical was trapped in an 6-exo-
dig cyclisation to afford imine C. Dihydroquinolinone 2 was
subsequently isolated from the reaction mixture upon treatment
with aqueous acid (see below).
In this work (Scheme 1b), we show that the scope of HAT-
mediated cyclisation reactions can be expanded to exploit nitriles
[a]
[b]
O. J. Turner*, Dr. D. J. Hirst and Dr. E. P. Talbot*†
GlaxoSmithKline, Medicines Research Centre, Gunnels Wood
Road, Stevenage, Hertfordshire, SG1 2NY (UK)
O. J. Turner* and Prof. Dr. J. A. Murphy*
Department of Pure and Applied Chemistry, University of
Strathclyde, 295 Cathedral Street, Glasgow, G1 1XL (UK)
Scheme 2. Intramolecular HAT-mediated radical cyclisation onto the proximal
nitrile vs. conventional hydrofunctionalisation reaction.
E-mail: john.murphy@strath.ac.uk
oliver.x.turner@gsk.com
†
Present address: Pharmaron, Hertford Rd, Hoddesdon, EN11 9BU
(UK)
To explore the scope of the reaction, conditions were
screened for conversion of test substrate 1 to ketone 2 (Table 1,
see SI for further details). As oxygen might be needed for the
E-mail: eric.talbot@pharmaron-uk.com
This article is protected by copyright. All rights reserved.

10.1002/chem.201805236
Chemistry - A European Journal
COMMUNICATION
regeneration of the catalyst, but could also potentially
detrimentally intercept crucial organic radicals on our pathway, we
performed experiments in sealed vials with a limited headspace
of air (conditions A), or open to air (conditions B), or occasionally
under inert gas (N2) (conditions C), in all cases using solvents that
were not degassed. To access the ketone, it was found that
hydrolysis of the imine intermediate was most effective under
microwave conditions.
Examining firstly the catalyst, the reaction provided higher
yields when using Fe(acac)3 rather than Mn(dpm)3 (entry 1 vs.
entry 2). Switching from EtOH as solvent (see SI) to iPrOH caused
no change in yields for reactions performed under similar
conditions; however, iPrOH was preferred for reactions
conducted open to the air because of its higher boiling point.
Comparison of entries 3 and 4 showed that for small scale
reactions, loading at 50 mol% of Fe(acac)3 worked better than
20%, but when performed on larger (0.5 mmol) scale, conditions
B with 20 mol% catalyst gave the best isolated yield of desired
ketone 2 (94%) on a 0.5 mmol scale, after just 1 h at 50 ºC (cf.
entry 5 vs 6).
although a superior yield (83%) was observed with EtOH:HFIP
(N2, entry 7). Conducting the reaction under conditions C in pure
EtOH depleted the conversion (entry 8), highlighting the
importance of HFIP for substrate 3 when the volume of air is
limited. The beneficial effect of HFIP may be due to its Lewis acid
character, although its benefits may extend beyond this - oxygen
has a high solubility in fluorinated solvents,[27] which may facilitate
catalyst turnover under sealed conditions (see SI, homogeneous
solution for entry 7 vs. heterogeneous for entry 8).
Table 2. Screening of reaction conditions for conversion of 3 to 4.
Yield (%)
Entry[a]
Solvent
Conditions
(3)
3
(4)
59
77
86
52
11
74
83
33
(5)
8
1
2
EtOH
EtOH:HFIP
EtOH:HFIP
iPrOH
A
A
C
B
B
B
C
C
<1
4
5
Table 1 Screening of reaction conditions for conversion of 1 to 2.
3
3
4
<1
21


18
51

5[b]
6[c*]
7[c]
8[d*]
[a]
iPrOH
iPrOH
EtOH:HFIP
EtOH

Yield (%)
Entry[a]
[M] (mol%)
Solvent
Conditions[b]
1
3
2


1
2
Fe(acac)3 (50)
Mn(dpm)3 (50)
Fe(acac)3 (50)
Fe(acac)3 (20)
Fe(acac)3 (20)
Fe(acac)3 (20)
EtOH
EtOH
iPrOH
iPrOH
iPrOH
iPrOH
A
A
B
B
A
B
81
66
70
57
71
94
HAT: 0.1 mmol; solution yield quoted, quantified by HPLC using an
0.5 mmol scale, isolated
yields quoted. *20 mol% Fe(acac)3. 0.5 mmol scale, NMR yield quoted
[b]
[c]
internal standard.
HAT conducted at RT.
14
<1
<1
19
-
[d]
3
as 3 and 4 co-elute during chromatography.
4
The optimised conditions (Table 1, entry 6) were applied to
a range of aromatic substrates (Scheme 3). The 5-exo-dig
cyclisation of alkene-nitrile 6 proceeded excellently under aerobic
conditions (catalyst loading as low as 5 mol% also worked well;
see SI), with only a slight drop in yield for the 6-exo-dig variant for
substrate 3. Benzyl protection of the tethering NH in 7 was well
tolerated (although not necessary, cf. 1) as was the inclusion of
steric hindrance ortho to the nitrile in 8. The presence of an iodide,
(substrate 9), was also well tolerated.
5[c]
6[c]
[a]
HAT: 0.1 mmol; solution yield quoted, quantified by HPLC using an
internal standard. [b] Conditions A, performed in a sealed vial with limited
headspace of air; conditions B, performed in a vial open to air. [c] 0.5 mmol
scale, isolated yields quoted.
We also optimised for conversion of substrate 3, bearing an
all-carbon side-chain, to ketone 4 (Table 2, see SI for further
details), and discovered subtle differences compared to the N-
linked substrate 1. Substrate 3 performed moderately well (59%
yield) on a small scale under the sealed conditions A (entry 1),
with the competitive formation of oxidised side-product 5,
reflecting slower radical cyclisation kinetics than for substrate 1.
To enhance the kinetics, hexafluoroisopropanol (HFIP) was
selected as a co-solvent. HFIP is a known Lewis acid[26] that might
facilitate cyclisation onto the nitrile group. Addition of HFIP as a
co-solvent with EtOH (1:1) increased the yield of desired product
4 to 77% (entry 2). This was increased yet further to 86% under
conditions C (entry 3). Aerobic conditions on a 0.1 mmol scale
(entry 4), were not beneficial, with the undesired tertiary alcohol 5
predominating at lower temperature (entry 5). However, on larger
scale (0.5 mmol), aerobic conditions with 20 mol% catalyst
(optimum for substrate 1) gave a good yield of 4 (entry 6),
The amino-tethered substrate (1) outperforms its all-carbon
variant (3), likely due to the planar N atom, positioning the two
reacting groups in closer proximity resulting in faster cyclisation.
Pleasingly, the HAT reaction of 1 to 2 was performed on a 1g
scale without decrease in yield. Cyclisation proceeded smoothly
with electron-donating aryl substituents para to the nitrile,
operating either by inductive effect [10 gave 21 (77%)] or
mesomeric effect) [11 gave 22 (69%)]. Similarly, electron-
withdrawing groups (CF3, 12) (CO2Et, 13) produced successful
cyclisations (76% and 88% respectively, as did heterocycle 14.
Facile access to spirocycles 26 and 27 was also achieved in good
yields (72% and 82% respectively), providing a new entry to
structurally complex scaffolds.
We next turned our attention to alkene-nitrile cyclisations in
which the alkene and nitrile are not rigidly held by an aromatic ring
This article is protected by copyright. All rights reserved.

10.1002/chem.201805236
Chemistry - A European Journal
COMMUNICATION
(Scheme 4). Pleasingly, cis-fused aliphatic ring system 34 was
formed in very good yield (73%) from 28. Substrate 29
Bonjoch[12] (alkoxyl radicals) and Baran[8–10] (radicals α- to an
electron-withdrawing group), single electron transfer (SET), is
proposed to convert the radicals to the
Scheme 3. Aromatic substrate scope for the HAT-mediated cyclisation of
Scheme 4. Construction of aliphatic ring systems by HAT-mediated cyclisation
alkene-nitriles.
of alkene-nitriles
.
Isolated yields quoted. All reactions performed on a 0.5 mmol scale using
conditions from Table 1 entry 6, unless otherwise stated. If ‘X‘ is undefined,
assume it is CH. [a] Conditions taken from entry 7, Table 2. [b] Hydrolysis for 5 h.
*94% yield on 1g (5.8 mmol) scale.
Isolated yield quoted, NMR yield determined with an internal standard given in
parenthesis were applicable. [a] Aerobic conditions taken from Table 1, entry 6.
Inert conditions taken from Table 2, entry 7. * 30 mins reaction time and
[b]
hydrolysis omitted, isolated product is impure. ^ 75 mol% Fe(acac)3 and 4.5 eq
PhSiH3. +NMR yield, some fractions of 39 co-elute on silica with the
hydrogenated starting material.
derived from diethyl malonate cyclised in excellent yield (93%) to
form 5-membered saturated ring 38 and the analogous 6-
membered product 36 was obtained from 30. Ethyl cyanoacetate-
derived 31 was cyclised in good yield to form 37 (61%).
Interestingly, substrate 32 underwent 5-exo-dig-cyclisation
followed by reversible nitrile translocation. In aerobic conditions,
the resulting benzyl radical is trapped by oxygen leading to the
formation of benzoyl derivative 39.[28] However, under conditions
C, the iminyl radical was preferentially trapped and the resulting
imine was hydrolysed on work-up to the expected ketone 38
(70%). The excellent selectivity shown likely reflects the strength
of the Si-H bond, impeding abstraction by a stabilised benzylic
radical; in contrast, the electrophilic iminyl radical may more
rapidly abstract an H atom from the hydridic Si-H bond, due to
polarity matching. Meanwhile, trapping of molecular oxygen
occurs rapidly for the benzylic radical, while the electrophilic iminyl
radical should be slow to form a weak N-O bond through coupling
to O2. The diphenyl variant 33 underwent cyclisation to yield the
highly sterically hindered, and consequently hydrolytically stable,
imine 40.
corresponding anions, with FeII being simultaneously oxidised to
FeIII. The iminyl radical present in our reactions may not be so
easily reduced by electron transfer, and instead may abstract H
from PhSiH3 (supported by large drop off in yield when only 1.5eq
of PhSiH3 are used, see SI). The resulting imine is then
hydrolysed in situ with aqueous acid to the corresponding ketone.
The FeII species can be oxidised to FeIII in the presence of oxygen
to complete the catalytic cycle.
Scheme 5. Proposed mechanism for HAT-mediated alkene-nitrile
cyclisation
We envisage a simplified mechanism[10] for coupling of alkenes to
nitriles (Scheme 5). HAT from in situ-generated HFe(acac)2 (41)
to the alkene sets up the exo-dig cyclisation. We then considered
the fate of the resulting iminyl radical. In the examples by
If the iminyl radical intermediate is not being converted
rapidly to the anion, then it should be possible to intercept the
This article is protected by copyright. All rights reserved.

10.1002/chem.201805236
Chemistry - A European Journal
COMMUNICATION
radical in a tandem cyclisation reaction (Scheme 6). To test this
hypothesis, we designed substrate 42 (see S2 in SI also), capable
of undergoing a second cyclisation and subjected it to our
standard reaction conditions (Scheme 6). The isolated tandem
product was the 7-membered ring heterocycle 43 (14%),
(rationalised in the SI file). The mono-cyclised product 44 was
also isolated (10%). This initial observation of tandem radical
cyclisations supports our proposal that the lifetime of the iminyl
radical is not negligible.
Acknowledgements
Financial support for this work was provided by GSK via the
GSK/University of Strathclyde Centre for Doctoral Training in
Synthetic and Medicinal Chemistry. We also thank Stephen
Richards (GSK) for his assistance with NMR analysis.
Keywords: HAT • nitriles • radical • cyclisation • Iron (III)
[1]
J. Kwlatek, I. L. Mador, J. K. Seyler, J. Am. Chem. Soc. 1962, 84,
304–305.
Scheme 6. HAT-mediated alkene-nitrile-alkene tandem cyclisation reaction.
[2]
[3]
[4]
J. Kwlatek, J. K. Seyler, J. Organomet. Chem. 1965, 3, 421–432.
J. Halpern, R. L. Sweany, J. Am. Chem. Soc. 1977, 99, 8335–8337.
J. Choi, L. Tang, J. R. Norton, J. Am. Chem. Soc. 2007, 129, 234–
240.
[5]
[6]
[7]
[8]
[9]
[10]
T. Mukaiyama, S. Isayama, S. Inoki, K. Kato, T. Yamada, T. Takai,
Chem. Lett. 1989, 449–452.
NMR yields quoted were determined with an internal standard (products
isolated by MDAP and characterised).
S. W. M. Crossley, C. Obradors, R. M. Martinez, R. A. Shenvi,
Chem. Rev. 2016, 116, 8912–9000.
C. Obradors, R. M. Martinez, R. A. Shenvi, J. Am. Chem. Soc.
2016, 138, 4962–4971.
Encouraged by this result, we next applied our methodology
to substrates bearing a cyanamide moiety that could undergo a
tandem cyclisation with an (hetero)aromatic ring (Scheme 7). To
our delight, benzamide 45 underwent the desired transformation
in very good yield (70%) to form spiro-quinazolinone 48. We also
pursued challenging targets 49 and 50, of direct relevance29-31 to
medicinal chemistry programmes. Nicotinic acid-derived
substrate 46 gave 49 as the major product (30%), along with a
small amount of 4-substituted regioisomer (5%); analogously,
pyrazole 47 was converted to the complex fused heterocycle 50,
an otherwise challenging target.
J. C. Lo, Y. Yabe, P. S. Baran, J. Am. Chem. Soc. 2014, 136, 1304–
1307.
J. C. Lo, J. Gui, Y. Yabe, C.-M. Pan, P. S. Baran, Nature 2014, 516,
343–348.
J. C. Lo, Y. Yabe, P. S. Baran, J. Am. Chem. Soc. 2017, 139, 2484–
2503.
[11]
[12]
S. Bordi, J. T. Starr, Org. Lett. 2017, 19, 2290–2293.
M. Saladrigas, C. Bosch, G. V. Saborit, J. Bonjoch, B. Bradshaw,
Angew. Chem. Int. Ed. 2018, 57, 182–186.
[13]
[14]
[15]
D. Griller, P. Schmid, K. U. Ingold, Can. J. Chem. 1979, 57, 831–
834.
B.-W. Anissa Yeung, J. L. M. Contelles, B. Fraser-Reid, J. Chem.
Soc., Chem. Commun. 1989, 1160–1162.
Scheme 7. HAT-mediated alkene-cyanamide-(hetero)aryl tandem cyclisation
reactions.
B. Chenera, C.-P. Chuang, D. J. Hart, L.-Y. Hsu, J. Org. Chem.
1985, 50, 5409–5410.
[16]
[17]
[18]
J. D. Kilburn, Tetrahedron Lett. 1990, 31, 2193–2196.
D. P. Curran, W. Liu, Synlett 1999, 1, 117–119.
R. A. Alonso, C. S. Burgey, B. Venkateswara Rao, G. D. Vite, R.
Vollerthun, M. A. Zottola, B. Fraser-Reid, J. Am. Chem. Soc. 1993,
115, 6666–6672.
[19]
[20]
[21]
[22]
D. L. J. Clive, P. L. Beaulieu, L. Set, J. Org. Chem. 1984, 49, 1313–
1314.
J. Streuff, M. Feurer, P. Bichovski, G. Frey, U. Gellrich, Angew.
Chem. Int. Ed. 2012, 51, 8661–8664.
A. Fernꢀndez-Mateos, S. E. Madrazo, P. H. Teijꢁn, R. R. Gonzꢀlez,
J. Org. Chem. 2009, 74, 3913–3918.
A. Gansäuer, F. Piestert, I. Huth, T. Lauterbach, Synthesis (Stuttg).
2008, 3509–3515.
Isolated yields quoted. [a] Fe(acac)3 20 mol%, 1.05 eq PhSiH3, iPrOH, 50 ºC, 1h,
air. [b] Fe(acac)3 20 mol%, 1.5 eq PhSiH3, 2 eq TFA, iPrOH, 80 ºC, 14h, air - note
5% 4-substituted regioisomer isolated also. 24 h. Reduced yields for the
[23]
[24]
B. B. Snider, B. O. Buckman, J. Org. Chem. 1992, 57, 322–326.
R. Bowman, C. Bridge, P. Brookes, Tetrahedron Lett. 2000, 41,
8989–8994.
*
nitrogen containing heterocycles is due to esterification of the starting
cyanamide with isopropanol.
[25]
[26]
[27]
[28]
[29]
[30]
N. Gandhamsetty, J. Jeong, J. Park, S. Park, S. Chang, J. Org.
Chem. 2015, 80, 7281–7287.
D. Lebœuf, L. Marin, B. Michelet, A. Perez-Luna, R. Guillot, E.
Schulz, V. Gandon, Chem. Eur. J. 2016, 22, 16165–16171.
In summary, we have developed an iron-mediated HAT
reaction between alkenes and nitriles. This work allows for the
formation of hindered ketones, spirocycles and fused bicyclic
systems. The reaction has been optimised to perform catalytically
under air and has been shown to scale-up without significant loss
of yield. Further investigations on cyanamides and other novel
HAT substrates are currently ongoing.
A. M. A. Dias, M. Freire, J. A. P. Coutinho, I. M. Marrucho, Fluid
Phase Equilib. 2004, 222–223, 325–330.
X.-W. Lan, N.-X. Wang, C.-B. Bai, C.-L. Lan, T. Zhang, S.-L. Chen,
Y. Xing, Org. Lett. 2016, 18, 5986–5989.
J. Zheng, Z. Deng, Y. Zhang, S. Cui, Adv. Synth. Catal. 2016, 358,
746–751
J. Zheng, Y. Zhang, D. Wang, S. Cui, Org. Lett. 2016, 18, 1768–
This article is protected by copyright. All rights reserved.

10.1002/chem.201805236
Chemistry - A European Journal
COMMUNICATION
1771.
[31].
N. Brugger et al. WO2014117920A1
This article is protected by copyright. All rights reserved.

10.1002/chem.201805236
Chemistry - A European Journal
COMMUNICATION
COMMUNICATION
O. J. Turner,* J. A. Murphy,* D. J. Hirst
and E. P. A. Talbot*
Page No. – Page No.
Hydrogen Atom Transfer-Mediated
Cyclisations of Nitriles
Hydrogen atom transfer (HAT) methodology has been expanded to an intramolecular
C-C coupling reaction between alkenes and nitriles, using PhSiH3 and catalytic
Fe(acac)3. This introduces a new strategic bond disconnection for ring-closing
reactions, forming ketones via imine intermediates.
This article is protected by copyright. All rights reserved.