Regioselective hydration and deprotection of chiral, dissymmetric iminodinitriles in the scope of an asymmetric strecker strategy

Article abstract of DOI:10.1002/ejoc.200600294

The controlled, selective decomposition of dissymmetric iminodinitriles (DIDN) of formula RCH(CN)-NH-C(CN)R′R″ (considered as N-protected alpha-aminonitriles), is a critical issue for an original asymmetric Strecker strategy previously outlined by us for the enantioselective synthesis of amino acids. This strategy, derived from Harada's work, involves a double sequence of (i) stereoselective Strecker condensation of a chiral ketone R′R″CO with NH₃ and HCN, followed by (ii) stereoselective Strecker condensation with an aldehyde RCHO and HCN, then (iii) regioselective retro-Strecker decomposition of the DIDN intermediate to release the target alpha-aminonitrile. In addition to the use of quite simple, cheap cyclic ketones (e.g. carvone derivatives) as chiral auxiliaries, another great advantage of this strategy is that step (iii) enables the recovery of the chiral ketone and hence its reuse. While our previous investigations on step (iii) under various conditions, either preceded or followed by the hydration of the secondary nitrile group RH(CN)- into an amide, had shown insufficient selectivity, we succeeded in the regioselective hydration of the secondary nitrile of DIDN without significant racemisation, by using a large excess of hydrogen peroxide in methanolic/aqueous ammonia (pH 12.5) at low temperature. The resulting imino nitrile/amide compound was then classically decomposed in acidic medium through a retro-Strecker reaction, affording the chiral alpha-amino amide. Alternately, the regioselective retro-Strecker decomposition of the tertiary moiety of the DIDN was achieved by reaction with silver cation in aqueous nitric acid, also without significant racemisation, thus establishing an original, enantioselective synthesis of alpha-aminonitriles. In both reactions, the chiral ketonic auxiliary resulting from DIDN decomposition was recovered in good yields. Wiley-VCH Verlag GmbH & Co. KGaA, 2007.

Full text of DOI:10.1002/ejoc.200600294

FULL PAPER
DOI: 10.1002/ejoc.200600294
Regioselective Hydration and Deprotection of Chiral, Dissymmetric
Iminodinitriles in the Scope of an Asymmetric Strecker Strategy
Jean-Christophe Rossi,*[a] Marc Marull,[b] Laurent Boiteau,[a] and Jacques Taillades[a]
Keywords: Iminodinitriles / Chiral α-aminonitriles / α-Amino acids / Regioselective hydration / Asymmetric Strecker
synthesis
The controlled, selective decomposition of dissymmetric imi-
nodinitriles (DIDN) of formula RCH(CN)–NH–C(CN)RЈRЈЈ
(considered as N-protected alpha-aminonitriles), is a critical
issue for an original asymmetric Strecker strategy previously
outlined by us for the enantioselective synthesis of amino ac-
ids. This strategy, derived from Harada’s work, involves a
double sequence of (i) stereoselective Strecker condensation
of a chiral ketone RЈRЈЈCO with NH3 and HCN, followed by
(ii) stereoselective Strecker condensation with an aldehyde
RCHO and HCN, then (iii) regioselective retro-Strecker de-
composition of the DIDN intermediate to release the target
alpha-aminonitrile. In addition to the use of quite simple,
cheap cyclic ketones (e.g. carvone derivatives) as chiral aux-
iliaries, another great advantage of this strategy is that step
(iii) enables the recovery of the chiral ketone and hence its
reuse. While our previous investigations on step (iii) under
various conditions, either preceded or followed by the hy-
dration of the secondary nitrile group RH(CN)- into an
amide, had shown insufficient selectivity, we succeeded in
the regioselective hydration of the secondary nitrile of DIDN
without significant racemisation, by using a large excess of
hydrogen peroxide in methanolic/aqueous ammonia (pH
12.5) at low temperature. The resulting imino nitrile/amide
compound was then classically decomposed in acidic me-
dium through a retro-Strecker reaction, affording the chiral
alpha-amino amide. Alternately, the regioselective retro-
Strecker decomposition of the tertiary moiety of the DIDN
was achieved by reaction with silver cation in aqueous nitric
acid, also without significant racemisation, thus establishing
an original, enantioselective synthesis of alpha-aminonitriles.
In both reactions, the chiral ketonic auxiliary resulting from
DIDN decomposition was recovered in good yields.
(© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2007)
Strecker reactions have been more recently developed,[8]
some of which show high efficiency in terms of both yield
and stereoselectivity (e. g. those involving Jacobsen organo-
catalysts), in spite of generally involving either sophisticated
substrates or multistep synthetic routes.
Introduction
Asymmetric Strecker synthesis is an interesting route for
the stereoselective preparation of non-natural amino acids
and derivatives, such as -amino acids. Its principle, initially
proposed in 1963 by Harada,[1] derives from the Strecker
reaction[2] and involves the stereoselective addition of hy-
drogen cyanide to a chiral imine derived from the condensa-
tion of an aldehyde with a chiral, primary amine followed
by the selective cleavage of the N-substituted aminonitrile
intermediate 1 (Scheme 1). Since then, besides the use of
various hydrocyanating agents such as α-hydroxyisobu-
tyronitrile Me2C(CN)OH, trimethylsilylcyanide (TMSCN)
Me3SiCN,[3] diethylaluminium cyanide (C2H5)2AlCN,[4] or
diethyl phosphorocyamidate[5] instead of HCN itself, asym-
metric Strecker synthesis was subject to numerous investi-
gations and extensions such as the recent cyanosilylation of
α-sulfinyl ketimines.[6] Most methods developed since Hara-
da’s work involve the use of a chiral auxiliary usually coval-
ently bound to the amine moiety.[7] Besides, catalytic
Scheme 1.
In the scope of Harada’s principle, however, it would be
highly profitable to develop an asymmetric Strecker synthe-
sis involving only simple, easily accessible compounds and
one-pot processes, with the possibility of recycling the chiral
auxiliary. Following our previous investigations on the
mechanism of the Strecker reaction,[9,10] we already out-
lined a promising case where the chiral amine is a tertiary α-
aminonitrile 4 derived from the Strecker reaction of a chiral
ketone 3 with hydrogen cyanide and ammonia.[11a] The key
asymmetric intermediate is then a dissymmetric iminodin-
itrile 6, obtained by the subsequent Strecker reaction of 4
with an aldehyde RCHO and HCN. The diastereoselectivity
of the addition of HCN on the prochiral imine 5 depends
[a] Dynamique des Systèmes Biomoléculaires Complexes,
UMR 5073 CNRS, Université Montpellier II,
Place Eugène Bataillon, 34095 Montpellier Cedex 5, France
[b] Institut de Chimie Moléculaire et Biologique, BCh, Ecole Poly-
technique Fédérale,
Lausanne, Switzerland
662
© 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2007, 662–668

An Asymmetric Strecker Strategy
FULL PAPER
both on the nature of the substituents R, RЈ, RЈЈ, on the
HCN donor and on the reaction conditions.[11b] In the in-
vestigated cases where 3 is a substituted cyclohexanone, the
most stable β configuration of the pericyclic nitrile favours
an eclipsed imine–nitrile configuration in 5.[11a]
Results and Discussion
The investigated iminodinitriles 10a–d (with R = Me, iPr,
iBu, Bn respectively, Scheme 3) were prepared according to
our previously described procedure,[11a] starting from
enantiopure 1R,2R,3R,5R(–)-3-cyanodihydrocarvone 8 ob-
tained by straightforward, stereoselective hydrocyanation of
(–)-carvone. The presence of the 3-cyano group on 8 provid-
ing an anchor for easy chemical transformation (through its
reduction into an amine), was chosen in view of prospective
covalent binding of this chiral auxiliary onto a polymer
resin. Isolation of the intermediate tertiary α-aminonitrile 9
was not necessary (whether 9 was isolated or not, overall
yields and stereoselectivities for 10 were identical in all
cases). Separation of diastereoisomer pairs 10-R and 10-S
(obtained in a ratio of ca. 4:1) was carried out by liquid
chromatography. In this paper, stereoisomers of either one
of 2, 7, 10–14, will be referred to after the configuration of
the stereogenic centre of the target amino acid or derivative
(the carbon-atom-bearing substituent R), e.g. 2-R, 10a-S.
The main advantage of this pathway is the possibility of
easily and selectively decomposing 6 through a retro-
Strecker process, either before or after regioselective hy-
dration of the secondary nitrile group, thus enabling the
recovery of the starting chiral ketone 3. Indeed, our pre-
vious investigations on the reactivity of α-aminonitriles 4
and 7 in aqueous solution[10] showed that they are subject
to two competing processes: (i) decomposition into alde-
hyde/ketone, through elimination of HCN followed by NH3
(retro-Strecker) and (ii) hydration of the nitrile group into
an amide. The most important parameter for the selectivity
of this reaction is the tertiary (faster decomposition) or sec-
ondary (faster hydration) nature of the α-aminonitrile. In
addition, in moderately alkaline media, α-aminonitrile hy-
dration is often subject to autocatalysis by either the ketone
or the aldehyde released by the decomposition of the sub-
strate.[10]
The reactivity of iminodinitriles in aqueous solution is in
theory subject to the same kind of selectivity: faster hy-
dration of secondary nitrile groups and faster decomposi-
tion of tertiary nitrile groups. However, the higher steric
hindrance and the presence of both secondary and tertiary
nitriles on the same substrate require the experimental vali-
dation of this relation to dissymmetric iminodinitrile reac-
tivity. The main concern addressed in this paper is thus the
controlled decomposition of iminodinitrile 6 to release the
targeted α-amino acid 2 (or nitrile 7 or amide 10) without
loss of chirality and with good recovery of the chiral
auxiliary.
Fulfilling this requirement would lead to further interest-
ing developments of this synthetic pathway including the
covalent binding of the chiral ketone 3 onto a polymer resin
to facilitate its recycling (a method already described else-
where[12] for other purposes), or the enantioselective synthe-
sis of non-proteogenic amino acids (e.g. α-methylated
amino acids, by replacing the aldehyde RCHO in Scheme 2
by a methyl ketone, as it seems possible on the basis of
preliminary experiments). Moreover, it will also be valuable
to extend this synthetic scheme to the preparation of chiral
α-aminophosphinic acids, known for their antibacterial
properties as mimics of α-amino acids or for their anti-
tumour or herbicidal properties.[13]
Scheme 3.
Strong conditions, either alkaline or acidic, are inappro-
priate, because of the lability of the secondary nitrile group
or the high sensitivity toward racemisation of the com-
pounds involved (Scheme 6). Hydrolysis of 10b (either-R
or -S) with KOH in either aqueous or tert-butyl alcohol
solution affords the fully racemised amino amide 11b. The
catalysis of nitrile hydration with acetone under alkaline
conditions developed by us[14] is poorly efficient with imin-
odinitrile 10 because of the difficulty of formation of the
key intermediate by addition of acetone on the hindered
secondary amine. The need for high temperature then re-
sults in both the partial racemisation of 11 and the degrada-
tion of the chiral auxiliary 8.
Although the hydrolysis of symmetrical secondary imin-
odinitriles (10 with RЈ = R and RЈЈ = H) in sulfuric acid
yields the corresponding imino diacids,[15] our investigation
on acidic hydrolysis/decomposition of iminodinitriles 10[11a]
did not provide good results because of the competing de-
composition reaction. Hydrolysis of iminodinitriles 10a–d
(either -R or -S) at 110 °C in 6  HCl afforded the expected
Scheme 2.
Scheme 4.
Eur. J. Org. Chem. 2007, 662–668
© 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
663

J.-C. Rossi, M. Marull, L. Boiteau, J. Taillades
FULL PAPER
Table 1. Yields and optical purity of α-amino acids or derivatives and recovery of chiral ketonic auxiliary 8 under various conditions: ee
measured by GC analysis of the (–)-menthyloxycarbonyl derivative of amino acids 2[22] (entries 1–8 and 14–16 after HCl hydrolysis of 7)
or by polarimetry (entries 9–13). Experimental conditions are detailed in the text.
Entry
Substrate
R
Conditions
Product
AA
Yield (%)
ee (%)
Recovered 8 (%)
1
2
3
4
5
6
7
8
10a-R
10a-S
10b-R
10b-S
10c-R
10c-S
10d-R
10d-S
10d-R
13b-S
10b-R
10c-R
10d-R
10b-R
10b-S
10d-R
Me
Me
iPr
iPr
iBu
iBu
Bn
Bn
Bn
iPr
iPr
iBu
Bn
iPr
iPr
Bn
HCl 110 °C
HCl 110 °C
HCl 110 °C
HCl 110 °C
HCl 110 °C
HCl 110 °C
HCl 110 °C
HCl 110 °C
2a-R
2a-S
2b-R
2b-S
2c-R
2c-S
2d-R
2d-S
-Ala
-Ala
-Val
-Val
37
54
23
62
17
52
41
57
97.8
99.0
98.2
97.0
81.2
95.1
96.4
97.0
77.3
81.0
95.6[b]
93.8[b]
93.2[b]
96.6
96.6
95.4
–
–
–
–
–
–
–
–
-Leu
-Leu
-Phe
-Phe
(-Phe)
(-Val)
(-Val)
(-Leu)
(-Phe)
(-Val)
(-Val)
(-Phe)
9
HCl/HCO2H
H2O 100 °C
11d-R
11b-S
11b-R
11c-R
11d-R
7b-R
7b-S
78
71
–
–
10
11
12
13
14
15
16
H2O2/NH3/HCl 50 °C[a]
H2O2/NH3/HCl 50 °C[a]
H2O2/NH3/HCl 50 °C[a]
AgNO3/HNO3
AgNO3/HNO3
AgNO3/HNO3
50[b]
72[b]
66[b]
40
37[c]
75[c]
79[c]
68
75
82
61
57
7d-R
[a] Two steps. [b] Overall yields and ee. [c] Recovered after the second step.
amino acids 2a–d with good optical purity (ee ≈ 98%) but S) with HCl/HCOOH affords -valinamide (11b-S) without
with poor yields, especially for the -R series (entries 1–8 in racemisation, the hydrolysis of enantiopure 13b-S (prepared
Table 1), together with important degradation of the cyano- by condensing 8 with enantiopure 11b-S in ethanol) in hot
dihydrocarvone moiety (Scheme 4).
water (100 °C, 12 h) yields 11b-S with only 81% ee (entry
Under softer acidic conditions, namely saturated hydro- 10 in Table 1).
chloric acid in aqueous formic acid at room temperature,[16]
Hydrogen peroxide. We finally examined the reaction of
10 was converted into a mixture of monohydrated adduct
12, imidazolidinone 13 and amino amide 11. Heating this
mixture at 100 °C for 12 h after neutralisation, enabled the
hydrolysis of both intermediates 12 and 13, affording the
amino amide 11 with better yields than above (41–75%
yield depending on R) but with partial racemisation (ee av-
eraging 75%, entry 9 in Table 1), while the chiral ketone 8
was recovered in an average yield of 65%. The stability of
12 is mostly due to the weak basicity of the amine group
(pKa around 4), which prevents the elimination of HCN
and thus the formation of the α-ketimino amide intermedi-
ate 14 (Scheme 5). Intermediate 13 results from the cyclis-
ation of imine 14.[17] The racemisation occurs during the
late steps of the above process and mainly involves the im-
ine 14 intermediate, as already reported by us for other
cases.[18] Indeed, while the hydrolysis of -valinonitrile (7b-
10 with hydrogen peroxide in methanol under moderately
alkaline conditions (ammonia). The ability of the hydroper-
oxide anion HOO– to convert nitriles into amides much
faster than HO– (discovered by Radziszewski[19]) had been
already extended by us to α-aminonitriles.[20] Whereas 10
can be directly converted into amide 11 and chiral ketonic
auxiliary 8 after sufficient reaction time (ca. 24 h at room
temperature), this method turned out to be somewhat ra-
cemising and insufficiently selective. It appeared much more
advantageous to limit the reaction to its first step, namely
the regioselective hydration of the secondary nitrile group,
to yield 12, then decomposing 12 by another, non-racemis-
ing method. The regioselectivity of iminodinitrile 10 hy-
dration is clearly determined by the blocked axial configu-
ration of the tertiary nitrile on the cyclohexyl ring, where
1,3-diaxial steric hindrance limits the access of the nucleo-
phile (either HOO– or HO–), as already observed in the hy-
dration of 1-amino-4-tert-butylcyclohexane(β)carbonitrile
mediated by HOO–, which is ca. 2000 times slower than that
of an acyclic aminonitrile (either secondary or tertiary).[20]
Optimisation. To find out optimal conditions for this re-
action, a systematic study was carried out on 10b-R. This
showed the efficiency and selectivity of hydration mediated
by HOO– to be dependent on the pH, temperature and the
[H2O2]/[10b-R] stoichiometric ratio. The reaction rate in-
creases with pH; however, it varies very little above pH 12.5
(Figure 1), roughly following the concentration of HOO– in
the medium (the pKa of H2O2 is 11.6 at 25 °C), thus con-
firming that HOO– is the active species. Meanwhile, the
selectivity was observed to decrease with increasing pH, so
that the optimum was found at pH 12.5, where the rate con-
stant of nitrile hydration with HOO– is ca. 200 times higher
Scheme 5.
664
www.eurjoc.org
© 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2007, 662–668

An Asymmetric Strecker Strategy
FULL PAPER
than that involving HO–.[20] At room temperature, increas- 12b-R, up to 90% within 10 h of reaction for a 40-fold stoi-
ing the H2O2 stoichiometry (and concentration) resulted in chiometric excess. This temperature effect can be rational-
a faster reaction; however, the selectivity of the reaction was ised by comparing the activation energies of nitrile hy-
not improved with regard to the hydration product 12b-R, dration induced by HOO– (13 kcalmol–1) with that of cya-
whose yield never exceeded 30–35% at the end of the reac- nide elimination (about 20 kcalmol–1), the former reaction
tion (Figure 2, solid symbols), because of side reactions thus being favoured at lower temperature.[20]
such as retro-Strecker decomposition and further degrada-
In summary, optimised conditions are pH 12.5, 10 °C,
tion of the cyanodihydrocarvone moiety (especially above [H2O2]/[10b-R] = 40 and a reaction time of ca. 10–15 h. No
sixfold excess in H2O2). At lower temperature, the reaction racemisation occurred at the secondary moiety (only the
was much more selective in favour of 12, and the inverse expected stereoisomer 12b-R was observed by NMR). The
trend was observed (Figure 2, open symbols): increasing the above conditions were then successfully applied to com-
H2O2 stoichiometric ratio resulted in increased yields of pounds 10c–d-R, affording 12c–d-R respectively in ca. 90%
yield, also without observable epimerisation.
Cleavage of intermediates 12b–d. The subsequent retro-
Strecker decomposition of 12 into α-amino amide 11 and
ketone 8 was then achieved by classical, straightforward
acidic hydrolysis in HCl at pH 1 and 50 °C, without signifi-
cant racemisation (Scheme 6 and entries 11–13 in Table 1).
Iminodinitriles 12b–d-R thus led to comparable overall
yields in α-amino amides 11b–d-R, respectively (with ee
around 95% in the studied cases where R = iPr, iBu, Bn)
and cyanodihydrocarvone 8.
Figure 1. Hydration of 10b with H2O2/NH3 in methanol: pH de-
pendence of reaction half time t1/2 (logscale) at 25 °C (diamonds)
for a stoichiometric ratio [H2O2]/[10b] = 2 and comparison with
the inverse of HOO– concentration {curve, calculated from the
equation 1/[HOO–] = ([H+] + Ka)/(Ka ϫc) with arbitrary c and
pKa(H2O2) = 11.6 at 25 °C}.
Scheme 6.
Ag+ treatment. Alternately, we showed that the treatment
of iminodinitrile 10 with silver nitrate in nitric acid
(Scheme 6), leads to the α-aminonitrile 7 through regiose-
lective decomposition of the tertiary moiety with good yield
and ee (measured after hydrolysis of 7 into amino acid 2)
and with acceptable recovery of the chiral auxiliary 8 (en-
tries 14–16 in Table 1). While acidic conditions prevent the
racemisation of the intermediates, the presence of Ag+ cat-
ions (known to catalyse the elimination of HCN) just fav-
ours the natural trend of tertiary aminonitriles to decom-
pose faster than secondary ones, thus providing an original
asymmetric synthetic route to α-aminonitriles 7. It is worth
mentioning that the asymmetric synthesis of N-unsubsti-
tuted α-aminonitriles is otherwise quite difficult, usually re-
quiring the dehydration of the N-protected α-amino amide,
e.g. with the method by Kawashiro starting from N-(o-ni-
trophenylsulfoxy)-α-amino amides.[21]
Conclusions
Figure 2. Hydration of 10b with H2O2/NH3 in methanol: yield of
12b against time at pH 12.5 at 10 °C (open symbols connected by
solid lines) and 25 °C (solid symbols connected by dashed lines)
for various stoichiometric ratios [H2O2]/[10b]: 2 (diamonds), 6 (tri-
angles), 10 (squares), 20 (circles), 40 (crosses). The data points in
the same series are connected for clarity.
This exploratory work demonstrates the feasibility of the
enantioselective synthesis of unprotected α-amino amides
and acids from aldehydes through a Strecker reaction, with
use of a chiral ketonic auxiliary that is recycled, by im-
proving the selective hydrolysis of the dissymmetric imino-
Eur. J. Org. Chem. 2007, 662–668
© 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
665

J.-C. Rossi, M. Marull, L. Boiteau, J. Taillades
FULL PAPER
was opened, and the reaction mixture was washed with chloroform
(3ϫ2 mL). After removal of the solvent in vacuo, the residue was
dissolved in HCl (1 mL, 0.1 ).
dinitrile intermediate 10. This was achieved through regiose-
lective hydration followed by cleavage of a dissymmetric im-
inodinitrile under non-racemising conditions. The regiose-
lective hydration of a secondary nitrile is achieved with hy- Determination of ee:[22] An aliquot of the above solution (50 µL)
was mixed in a vial tube with a pyridine/methanol mixture (40 µL,
20:80 v/v) and (–)-menthyl chloroformate (10 µL), stirred for
20 min, and then extracted with chloroform (100 µL). Aliquots of
the lower (chloroform) layer (2 µL) were then analysed by gas
chromatography.
drogen peroxide in aqueous ammonia, while the subsequent
cleavage of the resulting N-protected α-amino amide is
achieved in dilute hydrochloric acid. An alternative, also
non-racemising route involving silver nitrate in aqueous ni-
tric acid directly converts the iminodinitrile 10 into an N-
unprotected α-aminonitrile 7, thus representing an original,
enantioselective synthesis of the latter. In both routes, over-
all yields and ee from iminodinitrile 10 are quite good, as
well as the yield of the recovered chiral auxiliary 8.
Another good asset of the iminodinitrile strategy out-
lined in this paper is its ability to use alkyl aldehydes as
substrates, thus involving alkylimines as intermediates; the
latter were otherwise known to give rise to serious problems
Hydration of Iminodinitrile 10d-R into Amino Amide 11d-R with
HCl/HCO2H: Gaseous HCl was bubbled through a stirred solution
of iminodinitrile 10d-R (1 g, 3 mmol) in formic acid (15 mL) for
18 h. Water (100 mL) was then added to the mixture, which was
further stirred for 24 h at room temperature. After removal of both
formic and hydrochloric acids in vacuo, chloroform was added, and
the resulting solution was bubbled with ammonia. Precipitated salts
were filtered off, and chloroform was removed in vacuo. The resi-
due was dissolved in water, and the pH was adjusted to 7 with
when used in classical Strecker reactions. Our next goal is KOH. The solution was then heated under reflux for 12 h to com-
plete the hydrolysis of the intermediates. After acidification, the
resulting solution was washed with dichloromethane to recover
ketone 8. The pH of the aqueous layer was adjusted to 11, the
solution was saturated with sodium chloride and extracted with
dichloromethane. Bubbling gaseous HCl through the organic layer
caused the α-amino amide 11d-R (-phenylalanylamide) to precipi-
tate as the hydrochloride salt. Yield: 78%. Enantiomeric excess:
77.3% (based on [α]D = –34.8, c 0.5, MeOH).
now the investigation of more efficient chiral ketonic auxil-
iaries in order to improve the diastereoselectivity of dissym-
metric iminodinitrile 10 formation through the Strecker re-
action. This new perspective in asymmetric Strecker synthe-
sis that uses small auxiliaries bearing only few stereogenic
centres opens new potential routes for the asymmetric syn-
thesis of aminophosphinic acids and for the Ugi reaction.
Hydration of
L-2-Amino-3-methylbutanenitrile (7b-S) with HCl/
HCO2H: Gaseous HCl was bubbled through a stirred solution of
-valinonitrile (7b-S) (0.5 g, 3.27 mmol) in formic acid (15 mL) for
4 h. The reaction medium was treated with water (100 mL) to re-
move residual HCl by steam stripping in vacuo. After drying, -
valinamide hydrochloride (11b-S·HCl) was obtained with an op-
tical purity of 100% ([α]D = –34.8, c 0.5, MeOH).
Experimental Section
1H NMR spectra were recorded with a Bruker AC250 (250 MHz)
or a Varian EM360 (60 MHz) spectrometer. Chemical shifts are
expressed in parts per million downfield from TMS, which is used
as an internal standard. 13C NMR spectra were recorded with a
Bruker AM300 spectrometer (75 MHz) using broadband decoup-
ling. Chemical shifts are expressed in parts per million downfield
from TMS, using the middle resonance of CDCl3 (δ =77 ppm) as
an internal standard. Infrared spectra were recorded with a Perkin–
Elmer 1420 spectrophotometer. Mass spectra were obtained with a
Jeol JMS D100 spectrometer. Elemental analyses were performed
with a Perkin–Elmer 2400 CHN recorder. Optical rotations were
measured at 589 nm (Na) with a Perkin–Elmer 241 polarimeter.
Melting points (uncorrected) were measured with a Totoli instru-
ment. HPLC separations were performed with a Varian system
equipped with a UV detector (Varian 2550), a pump (Varian 2510),
a Shimadzu C-R6A Chromatopac integrator and a Kromasil C18
column (25 cm, 5 µm). GC analyses were recorded with a Dani
3900 system equipped with a silica capillary column OV1701 (30 m,
0.25-mm i.d., 0.25-µm stationary phase) and a Shimadzu C-R6A
Chromatopac integrator. TLC was performed with silica 60 F254
(Merck) plates and revealed with iodine and/or ninhydrin. pH mea-
surements were performed with a Metrohm Titrino SM 702 pH-
meter/pH-stat.
Synthesis of Compound 13b-S: -Valinamide (11b-S) (1.5 g,
12.4 mmol) and 8 (2.2 g, 13 mmol) were dissolved in absolute etha-
nol (40 mL) and heated under reflux for 5 d. The solvent was then
evaporated in vacuo, and the residue was dissolved in hydrochloric
acid (10 mL, 1 ). The aqueous phase was extracted with dichloro-
methane. Organic layers were joined, dried with Na2SO4, and the
solvent was removed in vacuo to give the crude product, which was
purified by column chromatography (Silica, ethyl acetate) to afford
13b-S. Yield: 0.4 g, 11%. IR (CHCl3): 3410 (NH), 3080 (CH2=CH),
1
2240 (CN), 1700 (C=O) cm–1. H NMR (60 MHz, CDCl3, 25 °C,
TMS): δ = 4.8 (s, 2 H, CH2), 3.6 (d, 1 H, HC-iPr), 1.8 (s, 3 H,
CH3), 1.4 (t, 6 H, iPr), 1.1 (d, 3 H, CH3) ppm.
Hydrolysis of Imidazolidin-4-one 13b-S in Neutral Aqueous Medium:
A solution of 13b-S (0.4 g, 1.45 mmol) in water (30 mL) was heated
under reflux for 36 h. Ketone 8 was recovered after extraction with
diethyl ether. After removal of the solvent in vacuo, CH2Cl2
(30 mL) was added to the crude product and HCl was bubbled
through it. -Valinamide hydrochloride (11b-S·HCl) was obtained
with an optical purity of 81% based on [α]D = 34.8, (c 0.5, MeOH).
(–)-(1R,2R,5R)-2-Methyl-5-(1-methylethenyl)-3-oxocyclohexane
carbonitrile (8) and iminodinitriles 10a–d (both -R and -S) were
prepared according to our previously described procedures.[11a]
Hydration of Iminodinitriles 10b–d-R into 12b–d-R with Hydrogen
Peroxide
Compound 10b-R (227 mg, 0.8 mmol) was suspended in MeOH
(30 mL) in a magnetically stirred, thermostatted (10 or 25 °C) reac-
tor. After complete dissolution of 10b-R, the pH was adjusted to
the desired value by addition of aqueous NH3 (32% w/w); the pH
was then maintained at the same value along the reaction, by fur-
ther NH3 addition controlled by means of a pH-stat. Aqueous hy-
Hydrolysis of Iminodinitriles 10a–d into Amino Acids 2a–d with Hot
HCl (Entries 1–8 in Table 1)
A solution of iminodinitrile 10a, 10b, 10c or 10d (-R or -S) (10 mg)
in hydrochloric acid (1 mL, 6 ) was heated in a sealed tube at
110 °C for 12 h. After being cooled to room temperature, the tube
666
www.eurjoc.org
© 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2007, 662–668

An Asymmetric Strecker Strategy
FULL PAPER
drogen peroxide (33% w/w and 2, 6, 10, 20 or 40 equiv. with respect
to 10b-R) was then added to the medium (time t = 0). The reaction
kinetics was monitored by HPLC analysis of 0.1-mL aliquots di-
luted in HPLC eluent (1 mL): MeOH/KH2PO4 buffer (0.05 , pH
6) 55:45.
H, α-H), 7.4 (br., 5 H, C6H5), 7.5–8.0 (s, 2 H, CONH2) ppm (iden-
tical to literature data).
Decomposition of Iminodinitriles 10 (10b-R, 10b-S, 10d-R) with
AgNO3/HNO3 (Entries 14–16 in Table 1)
A solution of iminodinitrile 10 (1 mmol) in methanol was added to
aqueous AgNO3 (2 mmol, 20 mL, 0.1 ) adjusted to pH 1 by ad-
dition of HNO3 (63% w/w). This solution was heated at 50 °C for
8 h (completion monitored by TLC, eluent chloroform/acetone
97:3). Insoluble silver cyanide salts were filtered off, and then meth-
anol was removed in vacuo. The resulting aqueous solution was
washed with diethyl ether (2ϫ20 mL) to recover the chiral ketone
8. The aqueous layer was then neutralised to pH 7 by addition of
NaOH (4 ) and extracted with dichloromethane (4ϫ10 mL). The
combined organic layers were dried with Na2SO4, and the solvent
was removed in vacuo to obtain the α-aminonitrile 7. The ee was
measured as described above by GC analysis of (–)-menthyloxycar-
bonyl derivatives after hydrolysis of 7 with HCl (6 ) at 110 °C.
Optimised Conditions: Iminodinitrile 10 in methanol (0.8 mmol,
16 mL, 0.05 molL–1), reaction temperature 10 °C, pH 12.5 (pH-
stat controlled by addition of aqueous NH3 32% w/w), aqueous
hydrogen peroxide (1.5 mL, 33% w/w, 40 equiv.). After completion
of the reaction (consumption of compound 10 monitored by TLC,
eluent CHCl3/acetone, 97:3), the solvent was removed in vacuo to
afford 12 as white crystals in 90% yield.
Compound 12b-R: M.p. = 179 °C. IR (CHCCl3): 3305–3380 (NH),
3080 (CH2=CH), 2240 (CN), 1680 (C=O), 1645 (C=C) cm–1. 1H
NMR (250 MHz, CDCl3, 25 °C): δ = 1.04 (dd, 3JHH = 6.9 Hz, 4JHH
3
= 2.7 Hz, 6 H, iPr), 1.35 (t, JHH = 12.9 Hz, 1 H, 6-Hax), 1.48 [d,
3JHH = 6.8 Hz, 3 H, (CH3)a], 1.55 (m, 1 H, 4-Hax), 1.74 (s, 3 H,
CH3), 1.82 (m, 1 H, 2-Hax), 2.13 (m, 1 H, β-H), 2.22 (m, 1 H, 6-
2-Amino-3-methylbutanenitrile (D/L-Valinonitrile) (7b-R and 7b-S):
3
3
1H NMR (250 MHz, CDCl3, 25 °C): δ = 1.04 (d, JHH = 6.7 Hz, 6
Heq), 2.54 (m, 1 H, 5-Hax), 3.12 (m, 1 H, 3-Heq), 3.24 (t, JHH
=
2
3
4.6 Hz, 1 H, α-H), 4.77 (d, JHH = 24 Hz, 2 H, CH2), 6.1 (s, 1 H,
CONH-H), 6.35 (s, 1 H, CONH-H) ppm. [α]D = +12.4 (c 0.5,
MeOH). C17H26N4O (302): calcd. C 67.55, H 8.61, N 18.54; found
C 67.45, H 8.80, N 18.23.
H, iPr), 1.56 (s, 2 H, NH2), 1.90 (m, 1 H, β-H), 3.49 (d, JHH
=
1
5.7 Hz, 1 H, α-H) ppm. H NMR (250 MHz, CD3OD, 25 °C): δ =
1.15 [d, JHH = 8 Hz, 6 H, (γ-Me)2], 2.28 (m, 1 H, β-H), 4.45 (d,
3
3JHH = 5.7 Hz, 1 H, α-H) ppm.
1
2-Amino-3-phenylpropanenitrile (
D
-Phenylalaninonitrile) (7d-R): H
Compound 12c-R: M.p. = 144 °C. IR (CHCCl3): 3305–3380 (NH),
3080 (CH2=CH), 2240 (CN), 1680 (C=O), 1645 (C=C) cm–1. 1H
NMR (250 MHz, CDCl3, 25 °C): δ = 1.0 (dd, 3JHH = 6.8 Hz, 4JHH
3
NMR (250 MHz, CDCl3): δ = 1.60 (s, 2 H, NH2), 3.00 (d, JHH
13.8 Hz, 2 H, CH2), 3.90 (t, JHH = 7.1 Hz, 1 H, α-H), 7.30 (m, 5
H, C6H5) ppm. H NMR 250 MHz (CD3OD): δ = 3.20 (d, JHH
13.8 Hz, 2 H, CH2), 4.70 (t, JHH = 7.1 Hz 1 H, α-H), 7.40 (m, 5
=
3
1
3
3
=
= 2.5 Hz, 6 H, iPr), 1.4 [d, JHH = 13.1 Hz, 3 H, (CH3)a], 1.7 [s, 3
3
H, (CH3)b], 2.20–2.57 (m, 9 H, Hcyclic, β-γ-H), 3.1 (m, 1 H, 3-Heq),
3.45 (m, 1 H, α-H), 4.75 [d, 2 H, (CH2)c], 5.8 (s, 1 H, CONH-H),
6.5 (s, 1 H, CONH-H) ppm. [α]D = +11.7 (c 0.5, MeOH).
H, C6H5).
Compound 12d-R: M.p. = 145 °C. IR (CHCCl3): 3305–3380 (NH),
3080 (CH2=CH), 2240 (CN), 1680 (C=O), 1645 (C=C) cm–1. 1H
NMR (250 MHz, CDCl3, 25 °C): δ = 1.07 [d, 3 H, (CH3)a], 1.31 (t,
3JHH = 12.9 Hz, 1 H, 6-Hax), 1.45 (m, 1 H, 4-Hax), 1.72 [s, 4 H,
(CH3)a, Hax], 2.17 (m, 3 H, NH, 4-Heq, 6-Heq), 2.95 (q, 1 H, CH-
H), 3.05 (m, 1 H, 3-Heq), 3.30 (m, 1 H, CH-H), 3.70 (t, 1 H, α-H),
4.75 [d, 2 H, 3JHH = 26.5 Hz (CH2)c], 6.01 (s, 1 H, CONH-H), 6.65
(s, 1 H, CONH-H), 7.30 (m, 5 H, C6H5) ppm. [α]D = –17.7 (c 0.5,
CHCl3).
Acknowledgments
We are grateful to the French Ministry of Education and Research
for a grant to one of us (M. M.).
[1] K. Harada, Nature 1963, 200, 1201.
[2] A. Strecker, Justus Liebigs Ann. Chem. 1850, 75, 27–45.
[3] a) H. Kunz, W. Sager, Angew. Chem. Int. Engl. 1987, 26, 557–
559; b) H. Kunz, W. Sager, W. Pfrengle, D. Schanzenbach, Tet-
rahedron Lett. 1988, 29, 4397–4400; c) H. Kunz, W. Sager, D.
Schanzenbach, M. Decker, Liebigs Ann. Chem. 1991, 649–654.
[4] F. A. Davis, P. S. Portonovo, R. E. Reddy, Y. Chiu, J. Org.
Chem. 1996, 61, 440–441.
Hydrolysis of Compounds 12b–d-R (Entries 11–13 in Table 1)
Compound 12 (1 mmol) was dissolved in a mixture of methanol
(25 mL) and water (25 mL). The pH was adjusted to 1 by addition
of HCl (6 ), and then the solution was heated to 50 °C for 4 h.
After completion of the reaction (monitored by TLC with chloro-
form/acetone 97:3), the solvent was removed in vacuo. The re-
sulting aqueous phase was washed with diethyl ether (3ϫ10 mL)
to recover the chiral ketonic auxiliary 8. Evaporation of the aque-
ous layer in vacuo afforded α-amino amide 11-R as the hydrochlo-
ride salt (ee measured by polarimetry).
[5] T. Shiori, S. Harusawa, Y. Hamada, Tetrahedron Lett. 1979, 20,
4663–4666.
[6] H. Acherki, C. Alvarez-Ibarra, A. De-Dios, M. Quiroga, Tetra-
hedron 2002, 58, 3217–3327.
[7] R. O. Duthaler, Tetrahedron 1994, 50, 1539–1650.
[8] H. Gröger, Chem. Rev. 2003, 103, 2795–2827.
[9] M. Béjaud, L. Mion, A. Commeyras, Tetrahedron Lett. 1975,
16, 2985–2986.
[10] R. Pascal, J. Taillades, A. Commeyras, Tetrahedron 1978, 34,
D
-Valinamide (11b-R): [α]D = –33.3 (c 0.5, MeOH). 1H NMR
2275–2280.
(60 MHz, CD3OD, 25 °C): δ = 1.1 (d, 6 H, iPr), 2.2 (m, 1 H, β-H),
3.8 (d, 1 H, α-H), 2.2 (m, 1 H, β-H), 3.8 (d, 1 H, α-H), 7.5–8.0 (s,
2 H, CONH2) ppm (identical to literature data).
[11] a) C. Bousquet, Z. Tadros, J. Tonnel, L. Mion, J. Taillades,
Bull. Soc. Chim. Fr. 1993, 130, 513–520; b) J. Taillades, M.
Marull, J. C. Rossi, unpublished results.
[12] A. Commeyras, J. Taillades, J. Brugidou, R. Sola, A. Previero,
R. Pascal, M. Laspéras, A. Rousset, U. S. patent 4.851576,
1983.
[13] V. P. Kukhar, H. R. Hudson, Aminophosphonic and Aminophos-
phinic Acids, Wiley, 2000.
[14] A. Commeyras, J. Taillades, L. Mion, R. Pascal, M. Lasperas,
A. Rousset A U. S. patent, 4.2423814, 1981.
D
-Leucinamide (11c-R): [α]D = –11.0 (c 0.5, MeOH). 1H NMR
(60 MHz, CD3OD, 25 °C): δ = 1.15 (d, 6 H, δ-H), 1.8 (m, 3 H, β-
γ-H), 4.0 (t, 1 H, α-H), 7.5–8 (s, 2 H, CONH2) ppm (identical to
literature data).
D
-Phenylalaninamide (11d-R): [α]D = –20.7 (c 0.5, MeOH). 1H
NMR (60 MHz, CD3OD, 25 °C): δ = 3.2 (m, 2 H, β-H), 4.2 (m, 1 [15] R. Sudo, S. Ichihara, Bull. Chem. Soc. Jpn. 1963, 36, 34–37.
Eur. J. Org. Chem. 2007, 662–668
© 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
667

J.-C. Rossi, M. Marull, L. Boiteau, J. Taillades
FULL PAPER
[16] F. Becke, H. Fleig, P. Pässler, Justus Liebigs Ann. Chem. 1971,
749, 198–201.
[17] a) Y. Ariyoshi, N. Sato, Bull. Chem. Soc. Jpn. 1972, 45, 2015–
2018; b) A. Tsujita, T. Yamana, Chem. Pharm. Bull. 1974, 22,
2434–2443; c) U. Zehavi, D. Ben-Ishai, J. Org. Chem. 1961, 26,
1097–1101.
[20] a) J. C. Rossi, L. Garrel, J. Taillades, A. Commeyras, C. R.
Acad. Sci., Ser. II 1996, 322(IIb), 767–773; b) J. Taillades, J. C.
Rossi, L. Garrel, M. Marull, A. Commeyras, Bull. Soc. Chim.
Fr. 1996, 133, 89–100.
[21] K. Kawashiro, H. Yoshida, S. Moritomo, Chem. Lett. 1976, 5,
417–418.
[18] a) P. M. Hardy, D. J. Samworth, J. Chem. Soc., Perkin Trans. 1
1977, 1954–1960; b) R. Polt, D. Seebach, Helv. Chim. Acta
1987, 70, 1930–1936.
[22] N. Domergue, M. Pugnière, A. Previero, Anal. Biochem. 1993,
214, 420–425.
Received: April 4, 2006
Published Online: November 27, 2006
[19] B. Radziszewski, Ber. Dtsch. Chem. Ges. 1884, 17, 1289.
668
www.eurjoc.org
© 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2007, 662–668