The first catalytic, enantioselective aza-Henry reaction of an unactivated cyclic imine

Article abstract of DOI:10.1002/adsc.201200558

The aza-Henry reaction is an efficient route to important chiral, vicinal diamines. Reports of the asymmetric version typically use electron-deficient acyclic imines. We report the first example of a catalytic, asymmetric aza-Henry reaction on an unfunctionalized cyclic imine that gives access to enantiopure diamine derivatives under experimentally straightforward conditions. The stereocentre is established through the initial nitromethane addition to the imine. The scalemic intermediate may be trapped through acylation, then crystallized. Copyright

Full text of DOI:10.1002/adsc.201200558

COMMUNICATIONS
DOI: 10.1002/adsc.201200558
The First Catalytic, Enantioselective Aza-Henry Reaction of an
Unactivated Cyclic Imine
Nilupa R. Amarasinghe,a Peter Turner,b and Matthew H. Todda,*
a
School of Chemistry, The University of Sydney, Sydney, NSW 2006, Australia
Fax : (+61)-2-9351-3329; e-mail: matthew.todd@sydney.edu.au
Crystal Structure Analysis Facility, School of Chemistry, The University of Sydney, Sydney, NSW 2006, Australia
b
Received: June 26, 2012; Revised: August 11, 2012; Published online: October 30, 2012
Supporting information for this article is available on the WWW under http://dx.doi.org/10.1002/adsc.201200558.
Abstract: The aza-Henry reaction is an efficient
route to important chiral, vicinal diamines. Reports
of the asymmetric version typically use electron-de-
ficient acyclic imines. We report the first example
of a catalytic, asymmetric aza-Henry reaction on an
unfunctionalized cyclic imine that gives access to
enantiopure diamine derivatives under experimen-
tally straightforward conditions. The stereocentre is
established through the initial nitromethane addi-
tion to the imine. The scalemic intermediate may
be trapped through acylation, then crystallized.
Itohꢀs methodology for the allylation of dihydroiso-
quinoline[13] and Ohsawaꢀs Mannich reaction on elec-
tron-deficient dihydro-beta-carbolines[14] which prima
facie require no activation of the imine. The aza-
Henry reaction (Scheme 1, A) employs nitromethane
as nucleophile, a reagent with great potential as a re-
placement for cyanide in asymmetric Strecker-like
processes.[15,16] Yet, beyond our recent report of the
racemic addition of nitromethane to 3,4-dihydroiso-
quinoline (DHIQ)[17] the addition of nitromethane to
cyclic imines (Scheme 1, B) has not been shown
beyond a racemic addition to activated isoquinoline[18]
and a recent addition to electron-deficient trifluoro-
methyl ketimines.[19]
Keywords: asymmetric catalysis; aza-Henry reac-
tion; bifunctional catalysis; nitroalkanes; thioureas
Towards the goal of asymmetric addition of nitro-
methane to a simple cyclic imine, we initially pursued
an approach based on Jacobsenꢀs counterion catalysis
concept[20,21] combining thiourea catalysis and alkyl
Since the first report of diastereo-[1] and enantioselec- salts of DHIQ. Pre-formation of a reactive electro-
tivity[2] in the aza-Henry (nitro-Mannich) reaction, phile was assumed to be necessary to overcome the
there has been a great deal of interest in catalytic, likely reversibility of the reaction between DHIQ and
enantioselective variants.[3] The interest is primarily nitromethane. However, catalysts that had proven to
because the aza-Henry reaction products, b-nitro be successful in related reactions[22] were ineffective
amines, can be transformed into chiral vicinal di- for the present reaction with low yields and enantio-
ACHTUNGTRENNUNGamines that are fundamentally important compounds selectivities being obtained despite a number of varia-
in natural products synthesis, pharmaceuticals and as tions in catalyst structure.
ligands in asymmetric catalysis.[4,5]
Switching to the Takemoto ligand[23] 1a gave
Reports to date of the catalytic, enantioselective a promising result of 60% yield of the aza-Henry re-
aza-Henry reaction have described a limited substrate action product 3a with 68% ee in a convenient one-
scope, in that acyclic, electron-deficient imines are the
ubiquitous electrophile.[3] Cyclic imines present
a greater challenge, and the reaction is unknown in
such cases. There are remarkably few examples of the
catalytic, asymmetric addition of any C-centered nu-
cleophiles to aza-aromatics,[6] but fewer still of addi-
tion to cyclic imines with such reactions typically re-
quiring pre-activation of the imine via acylation[7] (in
some cases oxidatively pre-formed in situ[8]) or alkyla-
tion[9] or the synthesis of the corresponding N-
Scheme 1. Aza-Henry reaction of common (A) and cyclic
oxide[10,11] or azomethine imine.[12] The exception is (B) imines.
2954
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2012, 354, 2954 – 2958

The First Catalytic, Enantioselective Aza-Henry Reaction of an Unactivated Cyclic Imine
Table 1. Optimization of catalyst structure.[a]
finyl substituent being partially tolerated (entry 8).
These results suggested that a Lewis basic group plays
a delicate but important role in the mechanism, most
likely in the activation of the nitromethane. Keeping
the dimethylamino portion constant and varying the
rest of the structure revealed that though there was
some sensitivity to changes in this region, a group ca-
pable of hydrogen bond donation was required
(entry 11). It was found that the simple C2-symmetric
ligand 1l (entry 13), previously only used in a dynamic
kinetic resolution,[24,25] gave good results.
Entry 1 R1
R2
R3
Yield[b] [%] ee[c] [%]
The mechanism of the asymmetric aza-Henry reac-
tion was explored with catalyst 1a. Fewer equivalents
of nitromethane or higher dilution gave poorer yields,
but broadly uniform ee. However, ee was much more
sensitive than yield to solvent polarity (pentane: 76%
yield, 31% ee; 1,2-DCE: 81, 64; DMF: 66, 4) implying
that enantioinduction is mediated by a hydrogen-
bonded assembly. The nature of the catalyst-substrate
interaction was not clear, specifically whether enan-
tioinduction was occurring in the reaction between
DHIQ itself (“unactivated”) or with its in situ-formed
acylated salt.
The screening conditions had involved mixing all
reagents except the acetyl chloride at room tempera-
ture followed by a reduction of the temperature to
À788C, addition of AcCl and three hours of stirring.
In fact the reaction gave broadly similar results after
being stirred for 5 min only at À788C (50% yield,
62% ee) or when the whole process was conducted at
room temperature (85% yield, 52% ee). However if
DHIQ was initially mixed with solvent and catalyst,
then cooled to À788C, followed by addition of nitro-
methane and base, stirred for 3 h then addition of
AcCl and stirred for a further 3 h, no product was
formed, implying that À788C is too low a temperature
for nitromethane addition. If the temperature of both
steps was increased to À108C, the yield and ee were
good (84%, 66%, respectively) and this improved
procedure was used hereafter. Crucially when the
same experiment was performed but the order of addi-
1
2
3
4
5
6
7
8
9
–
a
16
60
34
9
41
17
20
86
49
<1
68
7
12
10
7
-2
31
36
CH3
CH3
H
C2H5
i-Pr
i-Bu
Ac
t-BuSO
CH3
CF3
CF3
CF3
CF3
CF3
CF3
CF3
H
b H
c
d CH3
e
f
C2H5
i-Bu
Ac
H
g
h CH3
10
11
12
i
60
15
47
43
44
4
j
k
l
44
77
13
[a]
Reaction was conducted with 2a (1 equiv.), catalyst
(20 mol%), toluene (0.5 mL), MeNO2 (9 equiv.), Et3N
(1.5 equiv.) and CH3COCl (3 equiv.).
Yields were quantified using a calibration plot based on
the analytical HPLC trace.
[b]
[c]
Enantiomeric excess was determined by HPLC analysis tion was reversed (AcCl first, then MeNO2) no product
of 3a using a ChiralCel OD-H column. The reaction
product was found not to epimerize after 16 h in the
presence of catalytic TFA or Et3N.
was formed. These results show that the catalyst is in-
teracting with DHIQ itself and that a scalemic nitroa-
mine intermediate formed between DHIQ and nitro-
methane is undergoing acylation. Hence the enan-
tioinduction in this case is via the addition of nitrome-
pot protocol (Table 1). Increasing catalyst loading to thane to the unfunctionalized imine that is presuma-
40 mol% gave a 10% increase in yield, and decreasing bly transiently activated in the catalyst-substrate
the loading to 5 mol% gave a 10% decrease in yield complex. The conclusion that the stereochemistry is
of product (with 57% ee). Extensive variation of the set by nitromethane addition to the imine was sup-
catalystꢀs dimethylamino group revealed a very high ported by the following preliminary experiments: (i)
dependence of product yield and ee on the amine sub- the initially formed nitro amine does not undergo
stituents, with small changes bringing about significant cross-over upon addition of an excess of nitroethane,
changes in product profile. Changing the methyl as determined by monitoring the reaction with
groups to other alkyl or acyl substituents invariably 1H NMR spectroscopy, and (ii) racemic nitro amine
eroded both yield and ee (entries 2–7), with only a sul- (formed in situ without catalyst) to which was added
Adv. Synth. Catal. 2012, 354, 2954 – 2958
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
2955

Nilupa R. Amarasinghe et al.
COMMUNICATIONS
Table 2. Substrate scope of the reaction.
Entry
2
R1
R2
R3
3
Yield[a] [%] ee [%]
1
2
3
4
5
6
a
b
c
d
a
e
H
H
Ph
CH3
H
H
OCH3
H
H
H
H
H
H
H
a
b
c
–
f
84
81
trace
–
66
Figure 1. Dual activation of DHIQ and nitromethane by bi-
functional thiourea catalyst for the asymmetric aza-Henry
reaction.
À49
nd[c]
–
CH3
58
dr[b]
–
isoquinoline
–
sm[c]
catalyst gave, after quenching with acetyl chloride,
racemic 3a implying that there is no dynamic kinetic
resolution of the nitro amine by the catalyst on the
time scale of the reaction.
These data suggest the aza-Henry reaction proceeds
via an unusual mechanism (Figure 1) whereby the cat-
alyst simultaneously activates DHIQ and nitrome-
[a]
Isolated yield.
[b]
DS1:DS2=1:0.8, DS1=44:56 (racemic reaction); 48:52
(catalytic reaction). DS2=46:54 (racemic reaction);
48:52 (catalytic reaction).
[c]
nd=not determined; sm=starting material.
thane, and the enantioenriched nitro amine is ulti- elimination of HCl from acylated 2d (SI-9, entry 4).
mately acylated, presumably off-catalyst, allowing its Isoquinoline did not show any reaction with MeNO2
isolation.
and only starting material was recovered from the re-
The crucial role of the catalyst dimethylamino action (entry 6). When nitromethane was replaced
moiety is explained if nitromethane exists as its azinic with nitroethane, the yield dropped giving the product
acid tautomer in the catalyst-substrate complex and (3f) diastereomers in a 1:0.8 ratio but as racemates
its deprotonation by the dimethylamino moiety pro- (entry 5).
motes nucleophilic attack; the activation of the imine
Remarkably the ee for compound 3a increased
occurs during the reaction mechanism via transient from 66% to 99% after a single recrystallization from
non-covalent interactions.[15] (The role of the exoge- ethyl acetate. The scope of the process was demon-
nous base (triethylamine) is therefore to react with strated through the successful use of four other acylat-
the HCl produced.) Although the covalent activation ing agents. In each case the scalemic initially-formed
of the imine through acylation might be the more in- compound (3a, g–j) could be obtained enantiopure
tuitively obvious way of promoting the reaction, such after one crystallization (Table 3). The ee of the dime-
a method would bring the complication of (i) diaste- thoxy derivative increased to 99% after the second re-
reomeric acyliminium ions and (ii) competing electro- crystallization. Single crystal X-ray diffraction analysis
philic sites (imine carbon vs. carbonyl carbon) be- of compounds 3a and 3g confirmed the anticipated
tween which the catalyst would need to distin- (S)-configuration (see Supporting Information). The
guish.[26,27] For the carbon-carbon bond to be formed relative stereochemistry for the major diastereomer
the DHIQ would need to be oriented as shown to of 3f (also determined by X-ray crystallography, see
bring the imine carbon into close proximity to the Supporting Information) was found to be syn, that is,
azinic acid carbon, implying the formation of an (S)- (R,R) and (S,S).
stereocentre.
The substrate scope of the reaction was explored to a range of structures based on chiral vicinal di-
using four other unactivated cyclic imine substrates amines. Reduction of 3a to the corresponding amine
The enantiopure beta-nitro amines are entry points
AHCTUNGTRENNUNG
(Table 2): 6,7-dimethoxy-3,4-dihydroisoquinoline (2b), could be achieved in 85% yield with Raney nickel
1-phenyl-3,4-dihydroisoquinoline (2c), 1-methyl-3,4- under mild conditions with no loss of ee. The enantio-
dihydroisoquinoline (2d) and isoquinoline (2e).
pure amine 4a could be further functionalized by al-
The 6,7-dimethoxy-DHIQ afforded the product in kylation (to give 5) or conversion to the thiourea 6,
good yield with relatively low (and reversed) enantio- again with no loss of ee in either case. These are
selectivity (entry 2). The substitution on C-1 has a sig- novel chiral diamine motifs that are interesting proto-
nificant impact on the reaction outcome. The 1- types for new classes of asymmetric catalyst ligands
phenyl derivative gave a trace amount of product currently under investigation in our laboratory.
(entry 3) whereas substrate having CH3 at the C-1 po-
In summary, we have reported the first catalytic,
sition did not give the nitromethane addition product asymmetric aza-Henry reaction on an unactivated
but instead generated the enamide resulting from cyclic imine. The reaction successfully operates open
2956
asc.wiley-vch.de
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2012, 354, 2954 – 2958

The First Catalytic, Enantioselective Aza-Henry Reaction of an Unactivated Cyclic Imine
Table 3. Recrystallization of aza-Henry reaction products and subsequent modification.[a]
Series
R
Initially formed 3
Yield[b] [%]
Recrystallized 3
Yield[c] [%]
ee [%]
ee [%]
a
g
h
i
CH3
CH2Cl
Ph
p-Br, Ph
OCH2CH3
CH3
84
92
76
69
89
81
66
64
64
70
64
49
55
58
48
34
99
99
99
99
j
55
99
b[d]
25[e]
À99[e]
[a]
Reaction conditions: i) a. DCE, catalyst 1a (20 mol%), À108C, 10 min, b. MeNO2, Et3N at À108C, 3 h, c. RCOCl at
À108C, 1 h, ii) H2 (balloon), Raney Ni, NH3/MeOH (1:10), room temperataure, 3 h. iii) HCHO (aqueous), NaBH
DCE, room temperature, 2 h. iv) 3,5-Bis(trifluoromethyl)phenyl isothiocyanate, Et3N, DCM, 08C, 1 h.
Isolated yield.
ACHTUNGTRENNUNG(OAc)3,
[b]
[c]
[d]
[e]
Yield based on the starting material.
6,7-Dimethoxy-3,4-dihydroisoquinoline.
Yield and ee after second recrystallization.
to the air and using nitromethane that has not been Acknowledgements
dried. The reaction does not require a strong base for
We thank The University of Sydney for a USydIS (University
of Sydney International Scholarship) to N.R.A., Dr. Chris
McErlean (The Universiy of Sydney) and Dr. Nick Tyrrell
(Almac Sciences) for valuable advice.
the generation of a nitronate. The nitro amine prod-
ucts may be crystallized to enantiopurity and used in
the synthesis of new chiral amine-derived scaffolds.
Experimental Section
References
General Procedure for the Synthesis and
Crystallization of Enantiopure Aza-Henry Products
[1] H. Adams, J. C. Anderson, S. Peace, A. M. K. Pennell,
J. Org. Chem. 1998, 63, 9932–9934.
[2] K. Yamada, S. J. Harwood, H. Grçger, M. Shibasaki,
Angew. Chem. 1999, 111, 3713–3715; Angew. Chem.
Int. Ed. 1999, 38, 3504–3506.
Synthesis and crystallization of 3a: To a solution of 2a
(200 mg, 1.50 mmol) in DCE (5.0 mL) was added catalyst 1a
(124 mg, 0.30 mmol) at room temperature. The reaction
flask was capped with a rubber septum and was cooled to
À108C and the mixture stirred for 10 min. Nitromethane
(725 mL, 13.5 mmol) and Et3N (310 mL, 2.20 mmol) were
added and the reaction mixture was stirred for 3 h at
À108C. Acetyl chloride (320 mL, 4.50 mmol) was added and
stirring was continued for 1 hour at À108C. Solvents were
removed under vacuum and the residue was purified by
flash chromatography on silica gel (EtOAc:hexane=1:1) to
give 3a as a colourless solid; yield: 302 mg (84%, 66% ee),
which was recrystallized from EtOAc (approx. 6 mL) at
room temperature to give enantiopure 3a as colourless
tablet-shaped crystals; yield: 196 mg (55%, 99% ee).
[3] E. Marques-Lopez, P. Merino, T. Tejero, R. P. Herrera,
Eur. J. Org. Chem. 2009, 2401–2420.
[4] D. Lucet, T. Le Gall, C. Mioskowski, Angew. Chem.
1998, 110, 2724–2772; Angew. Chem. Int. Ed. 1998, 37,
2581–2627.
[5] S. Kotti, C. Timmons, G. G. Li, Chem. Biol. Drug Des.
2006, 67, 101–114.
[6] M. Ahamed, M. H. Todd, Eur. J. Org. Chem. 2010,
5935–5942.
[7] T. Kanemitsu, E. Toyoshima, M. Miyazaki, K. Nagata,
T. Itoh, Heterocycles 2010, 81, 2781–2792.
Adv. Synth. Catal. 2012, 354, 2954 – 2958
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
2957

Nilupa R. Amarasinghe et al.
COMMUNICATIONS
[8] C. Dubs, Y. Hamashima, N. Sasamoto, T. M. Seidel, S.
Suzuki, D. Hashizume, M. Sodeoka, J. Org. Chem.
2008, 73, 5859–5871.
[9] A. M. Taylor, S. L. Schreiber, Org. Lett. 2006, 8, 143–
146.
[18] J. S. Yadav, B. V. S. Reddy, N. N. Yadav, M. K. Gupta,
Synthesis 2009, 1131–1136.
[19] H. X. Xie, Y. A. Zhang, S. L. Zhang, X. B. Chen, W.
Wang, Angew. Chem. 2011, 123, 11977–11980; Angew.
Chem. Int. Ed. 2011, 50, 11773–11776.
[10] S. Wang, C. T. Seto, Org. Lett. 2006, 8, 3979–3982.
[11] S. Murahashi, Y. Imada, T. Kawakami, K. Harada, Y.
Yonemushi, N. Tomita, J. Am. Chem. Soc. 2002, 124,
2888–2889.
[12] T. Hashimoto, M. Omote, K. Maruoka, Angew. Chem.
2011, 123, 9114–9117; Angew. Chem. Int. Ed. 2011, 50,
8952–8955.
[13] M. Miyazaki, N. Ando, K. Sugai, Y. Seito, H. Fukuoka,
T. Kanemitsu, K. Nagata, Y. Odanaka, K. T. Nakamura,
T. Itoh, J. Org. Chem. 2011, 76, 534–542.
[20] R. R. Knowles, E. N. Jacobsen, Proc. Natl. Acad. Sci.
USA 2010, 107, 20678–20685.
[21] S. E. Reisman, A. G. Doyle, E. N. Jacobsen, J. Am.
Chem. Soc. 2008, 130, 7198–7199.
[22] R. S. Klausen, E. N. Jacobsen, Org. Lett. 2009, 11, 887–
890.
[23] T. Okino, Y. Hoashi, Y. Takemoto, J. Am. Chem. Soc.
2003, 125, 12672–12673.
[24] A. Berkessel, S. Mukherjee, T. N. Muller, F. Cleemann,
K. Roland, M. Brandenburg, J. M. Neudorfl, J. Lex,
Org. Biomol. Chem. 2006, 4, 4319–4330.
[14] T. Itoh, M. Yokoya, K. Miyauchi, K. Nagata, A.
Ohsawa, Org. Lett. 2003, 5, 4301–4304.
[15] S. J. Zuend, M. P. Coughlin, M. P. Lalonde, E. N. Jacob-
sen, Nature 2009, 461, 968–970.
[16] S. C. Pan, J. Zhou, B. List, Angew. Chem. 2007, 119,
618; Angew. Chem. Int. Ed. 2007, 46, 612.
[17] M. Ahamed, T. Thirukkumaran, W. Y. Leung, P.
Jensen, J. Schroers, M. H. Todd, Eur. J. Org. Chem.
2010, 5980–5988.
[25] A. Berkessel, S. Mukherjee, F. Cleemann, T. N. Muller,
J. Lex, Chem. Commun. 2005, 1898–1900.
[26] M. Takamura, K. Funabashi, M. Kanai, M. Shibasaki, J.
Am. Chem. Soc. 2000, 122, 6327–6328.
[27] C. K. De, N. Mittal, D. Seidel, J. Am. Chem. Soc. 2011,
133, 16802–16805.
2958
asc.wiley-vch.de
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2012, 354, 2954 – 2958