A general route to protected quaternary α-amino acids from β-amino alcohols via a stereocontrolled radical approach

Article abstract of DOI:10.1039/b604123j

A radical-based approach facilitates the highly stereocontrolled functionalisation of β-amino alcohols, opening up a new, generally applicable methodology for the preparation of quaternary α-amino acids. The Royal Society of Chemistry 2006.

Full text of DOI:10.1039/b604123j

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
A general route to protected quaternary a-amino acids from b-amino
alcohols via a stereocontrolled radical approach
Mark E. Wood,*a Mark J. Penny,a Jenny S. Steere,a Peter N. Horton,b Mark E. Lightb and
Michael B. Hursthouseb
Received (in Cambridge, UK) 21st March 2006, Accepted 25th May 2006
First published as an Advance Article on the web 7th June 2006
DOI: 10.1039/b604123j
A radical-based approach facilitates the highly stereocontrolled
functionalisation of b-amino alcohols, opening up a new,
generally applicable methodology for the preparation of
quaternary a-amino acids.
Thesignificanceofa,a-disubstituted(orquaternary)a-aminoacids,
particularly in the context of modification/optimisation of the
activity of biologically active peptides and as enzyme inhibitors in
theirownright,hasensuredtheirimportanceassynthetictargets.To
date, the vast majority of asymmetric synthetic approaches to such
compoundshaverelieduponthealkylationofenolatederivativesof
existing amino acids,1 particularly alanine. Stereocontrol has
generally been achieved by the use of chiral auxiliaries, as typified
by the Scho¨llkopf approach,2 or exploitation of the principle of
‘‘self-regenerationofstereocentres(SRS),’’championedbySeebach
et al.3 The sensitivity of many a-amino acid side-chains towards the
strongly basic conditions required for quantitative enolate genera-
tion and the severe steric constraints associated with the efficient
generation of quaternary centres under ‘‘anionic conditions’’,
naturally impose serious restrictions on the amino acid substrates
that can be employed. Reactions involving radical intermediates
suffer from fewer such restrictions and so it is surprising that only a
relatively small number of synthetic approaches to highly
substituted amino acids use such methodology.4
Scheme 1 Preparation of 1,3-oxazolidine radical precursors. Reagents
and conditions: (a) tBuCHO, MgSO4, CHCl3, D; (b) 2-iodobenzoyl-
chloride, pyridine, DMAP (cat.), R.T.
Table 1 Preparation of 1,3-oxazolidine radical precursors
Entry
R
Amino alcohol 2
Product Yield (%)a
1
2
3
4
5
a
CH3
1a
1b
1c
1d
1e
66
21
73
87
12
L-Alaninol
(CH3)2CH
(CH3)2CHCH2
PhCH2
L-Valinol
L-Leucinol
L-Phenylalaninol
L-Methioninol
CH3SCH2CH2
Only single diastereoisomers isolated.
lower yields obtained with both L-valinol and L-methioninol
(entries 3 and 5), can be attributed to the lower efficiency of the
acylation step. In each case, only a single diastereoisomer of the
product 1 could be obtained after chromatographic purification,
the minor stereoisomer possibly being destroyed during the
N-acylation procedure or by hydrolysis on silica gel.
Previously, we have shown that ethanolamine can be alkylated
a-to nitrogen via racemic 1,3-oxazolidine intermediates using a
radical chain sequence involving aryl radical generation, 1,5-
hydrogen atom transfer and diastereoselective trapping of the
a-aminoalkyl radical intermediate with acryronitrile or acrylate
esters. After 1,3-oxazolidine hydrolysis, protected, alkylated
ethanolamine derivatives were obtained.5 Our goal has been to
develop this approach into a generally applicable methodology for
the synthesis of quaternary amino acids, allowing the derivatisa-
tion of a wider range of amine and amino acid derivatives. To this
end, we describe herein, the efficient, stereocontrolled derivatisa-
tion of a number of ‘‘unactivated’’ b-amino alcohols.
X-Ray crystallographic analysis{ of 1a gave not only confirma-
tion of the absolute and relative stereochemistry but also revealed
the preferred rotameric structure of the amide bond, which
provides much closer proximity of the future radical-bearing
carbon atom of the PRT group (C9 in Fig. 1) to the key hydrogen
˚
˚
atom at C-4 compared with that at C-2 (2.9 A vs. 4.8 A, Fig. 1).
This amide geometry and resulting iodide placement are essential
for regiospecific generation of the desired C-4 a-aminoalkyl
radical, given that the rate of 1,5-hydrogen atom transfer is much
faster than the rate of rotation around the C(O)–N bond.7
NOESY studies also supported the C-2/C-4 cis-relative stereo-
chemistry for L-phenylalaninol-derived product 1d (entry 4).
Unfortunately, crystallisation could not be induced in any of the
other products and NOESY experiments proved inconclusive for
determination of relative stereochemistry. Given that there is
considerable literature precedent for the predominance of the
thermodynamically favoured cis-product in the formation of such
1,3-oxazolidines8,9 it was assumed that the remaining products
possessed the same relative stereochemistry.
The requisite radical reaction substrates 1 were prepared using a
conventional two-step approach involving 1,3-oxazolidine forma-
tion from enantiomerically pure L-b-amino alcohols 2,6 followed
by N-acylation for introduction of the 2-iodobenzoyl ‘‘protecting-
radical-translocating (PRT)’’ group (Scheme 1 and Table 1). The
aSchool of Biosciences (originally Department of Chemistry), University
of Exeter, Geoffrey Pope Building, Stocker Road, Exeter, UK EX4 4QD.
E-mail: m.e.wood@exeter.ac.uk; Tel: +44 1392 263450
bEPSRC X-ray Crystallography Service, Department of Chemistry,
University of Southampton, Highfield, Southampton, UK SO17 1BJ
This journal is ß The Royal Society of Chemistry 2006
Chem. Commun., 2006, 2983–2985 | 2983

View Article Online
Fig. 1 ORTEP representation of 1a; ellipsoids drawn at 30% probability
level. (Note: The asymmetric unit of 1a contains two independent
molecules. For clarity, only one is shown.)
Fig. 2 ORTEP representation of 3a; ellipsoids drawn at 30% probability
level.
Scheme 3 Hydrolysis of functionalised 1,3-oxazolines. Reagents and
conditions: CF3CO2H, CH2Cl2, R.T.
Scheme 2 Radical reactions of 1,3-oxazolidines. Reagents and condi-
tions: Bu3SnCl (0.15 equiv.), NaBH3CN (2 equiv.), acrylonitrile or methyl
t
stereochemistry. There are however, many literature examples of
closely related enolate alkylations of serine-derived 1,3-oxazoli-
dines that occur predominantly on the opposite face of the
heterocycle to the C-2 tert-butyl group,3,10 hence it was thought
reasonable to assume that the remaining radical-trapping products
possessed the same relative stereochemistry as 3a.
acrylate (5 equiv.), AIBN (4 6 0.2 equiv.), BuOH, D.
Compounds 1a–e were subjected to radical generating/trapping
using in situ generation of tributyltin hydride (from 15 mol%
tributyltin chloride and sodium cyanoborohydride) and acryloni-
trile and methyl acrylate in 5-fold excess as the radicalphile
(Scheme 2). The results from these experiments are summarised in
Table 2. In all cases, a mixture of both trapped product 3 and
reduced material 4 was obtained, with only a single diastereo-
isomer of 3 being detectable by 1H NMR spectroscopy after
chromatographic isolation. (The stereochemical integrity of the
reduction product 4 was not determined.) The expected trans-
relationship between the C-2 tert-butyl substituent and the newly
added substituent at the quaternary C-4 centre was revealed by
X-ray crystallographic analysis{ of alaninol-derived product 3a
(Fig. 2).
The proportion of reduction product 4 obtained was found to
increase with the use of methyl acrylate rather than acrylonitrile as
the radicalphile, a finding for which there is literature precedent.11
These results highlight the excellent efficiency of stereocontrolled
quaternary centre generation under radical conditions and the
successful result obtained with L-methioninol-derived 1,3-oxazoli-
dine 1e, shows the highly desirable tolerance for functionalised
side-chains using this methodology.
Hydrolysis of acrylonitrile-derived products 3a–d was carried
out using trifluoroacetic acid in dichloromethane, in order to avoid
simultaneous hydrolysis of the nitrile and amide (Scheme 3).
Lower yields were obtained with more highly substituted 1,3-
oxazolidines such as L-valinol-derived 3b (Table 3) and further
optimisation of this step is required.
For structures 3b to 3i, crystals suitable for X-ray analysis could
not be obtained and peak overlap and line broadening in some
key 1H NMR signals from rotameric structures, precluded the use
of NOESY experiments for determination of the relative
Table 2 Radical reactions of 1,3-oxazolidines
Entry
Substrate
R
Radicalphile
Product 3
Yield of 3a (%)
Yield of 4 (%)
1
2
3
4
5
6
7
8
9
a
1a
1b
1c
1d
1e
1a
1b
1c
1d
CH3
Acrylonitrile
Acrylonitrile
Acrylonitrile
Acrylonitrile
Acrylonitrile
Methyl acrylate
Methyl acrylate
Methyl acrylate
Methyl acrylate
3a
3b
3c
3d
3e
3f
3g
3h
3i
85
59
32
63
50
48
38
34
22
14
41
46
31
34
23
36
28
18
(CH3)2CH
(CH3)2CHCH2
PhCH2
CH3SCH2CH2
CH3
(CH3)2CH
(CH3)2CHCH2
PhCH2
Only single diastereoisomers isolated.
2984 | Chem. Commun., 2006, 2983–2985
This journal is ß The Royal Society of Chemistry 2006

View Article Online
Service Centre (Swansea, UK) for mass spectra and Dr S. J.
Simpson (Exeter, UK) for invaluable assistance with NOESY
experiments and interpretation of crystallographic data.
Table 3 Results for hydrolysis of functionalised 1,3-oxazolidines
Entry
Substrate
R
Product
Yield (%)
1
2
3
4
3a
3b
3c
3d
CH3
5a
5b
5c
5d
75
24
44
21
(CH3)2CH
(CH3)2CHCH2
PhCH2
Notes and references
{ Suitable crystals were selected and data collected on a Bru¨ker Nonius
KappaCCD Area Detector at the window of a Bru¨ker Nonius FR591
12
rotating anode (lMo-Ka 5 0.71073 A) driven by COLLECT and
˚
DENZO13 software at 120 K. Structures were determined in SHELXS-
9714 and refined using SHELXL-97.15
Crystal data for 1a: C15H20INO2,
M 5 373.22, monoclinic,
˚
˚
˚
a 5 12.7729(7) A, b 5 7.4551(4) A, c 5 17.0731(8) A, b 5 107.236(4)u,
3
˚
U 5 1552.75(14) A , T 5 120(2) K, space group P21, Z 5 4, m(Mo-
Ka) 5 1.597 mm21, 15953 reflections measured, 6846 unique (Rint 5 0.0352)
which were used in all calculations. Final R1 5 0.0370, wR2 5 0.0728 [F2 .
2s(F2)], R1 5 0.0526, wR2 5 0.0787 (all data). Absolute structure
parameter 5 0.005(18).
Scheme 4 Oxidation/esterification of functionalised b-amino alcohols.
Reagents and conditions: CrO3, H2SO4, H2O, (CH3)2CO; (b)
(CH3)3SiCHN2, C6H6, CH3OH (4 :1 v/v).
Crystal data for 3a: C18H24N2O2, M 5 300.39, orthorhombic,
3
˚
˚
˚
˚
a 5 7.6733(6) A, b 5 13.630(2) A, c 5 15.8871(19) A, U 5 1661.6(3) A ,
T 5 120(2) K, space group P212121, Z 5 4, m(Mo-Ka) 5 0.079 mm21, 9839
reflections measured, 2926 unique (Rint 5 0.0326) which were used in all
calculations. Final R1 5 0.0304, wR2 5 0.0723 [F2 . 2s(F2)], R1 5 0.0345,
wR2 5 0.0751 (all data). CCDC 602435 and 602436. For crystallographic
data in CIF or other electronic format see DOI: 10.1039/b604123j
Table 4 Results for b-amino alcohol oxidation/esterification
Entry
Substrate
R
Product
Yield (%)
1
2
3
4
5a
5b
5c
5d
CH3
6a
6b
6c
6d
55
70
71
52
1 C. Cativiela and M. D. D´ıaz-de-Villegas, Tetrahedron: Asymmetry,
1998, 9, 3517.
2 U. Scho¨llkopf, W. Hartwig and U. Groth, Angew. Chem., Int. Ed. Engl.,
1979, 18, 863.
(CH3)2CH
(CH3)2CHCH2
PhCH2
3 D. Seebach, A. R. Sting and M. Hoffmann, Angew. Chem., Int. Ed.
Engl., 1996, 35, 2708.
The highly functionalised b-amino alcohols 5 thus obtained
were oxidised efficiently to the corresponding N-benzoyl quatern-
ary a-amino acids using Jones’ reagent followed by treatment with
(trimethylsilyl)diazomethane to give the methyl esters 6 as single
enantiomers (Scheme 4). The yields given in Table 4 cover the
combination of oxidation and esterification steps.
4 C. J. Easton, Chem. Rev., 1997, 97, 53; L. Giraud and P. Renaud, J. Org.
Chem., 1998, 63, 9162.
5 R. Gosain, A. M. Norrish and M. E. Wood, Tetrahedron Lett., 1999,
40, 6673; R. Gosain, A. M. Norrish and M. E. Wood, Tetrahedron,
2001, 57, 1399.
6 C. Agami, S. Comesse and C. Kadouri-Puchot, J. Org. Chem., 2002, 67,
1496.
7 V. Snieckus, J.-C. Cuevas, C. P. Sloan, H. Liu and D. P. Curran, J. Am.
Chem. Soc., 1990, 112, 896.
8 D. Seebach and J. D. Aebi, Tetrahedron Lett., 1984, 25, 2545.
9 C. Agami and F. Couty, Eur. J. Org. Chem., 2004, 677.
10 For a recent, typical example see: M. Brunner and A. M. P. Koskinen,
Tetrahedron Lett., 2004, 45, 3063.
11 F. Beaulieu, J. Arora, U. Veith, N. J. Taylor, B. J. Chapell and
V. Snieckus, J. Am. Chem. Soc., 1996, 118, 8727.
12 R. Hooft, Collect: Data collection software, Nonius B. V., Delft, The
Netherlands, 1998.
13 Z. Otwinowski and W. Minor, Methods Enzymol., 1997, 276,
307.
14 G. M. Sheldrick, Acta Crystallogr., Sect. A: Found. Crystallogr. , 1990,
46, 467.
15 G. M. Sheldrick, University of Go¨ttingen, Germany, 1997.
In summary, we have demonstrated a straightforward radical-
based methodology for the asymmetric derivatisation of b-amino
alcohols a-to nitrogen. The methodology represents a much-
needed, versatile route to highly functionalised quaternary a-amino
acids which will be applicable to a wide range of biologically
important substrates. The carboxylic acid oxidation level of the
nitrile and ester substituents, introduced via the acrylonitrile or
methyl acrylate radicalphiles, allows the reported products 6 to be
regarded as highly functionalised glutamic acid analogues.
We thank EPRSC for a Quota studentship (J. S. S.) and Pfizer
Global Research and Development (Sandwich, UK) for a Summer
Scholarship (M. J. P.), the EPSRC National Mass Spectrometry
This journal is ß The Royal Society of Chemistry 2006
Chem. Commun., 2006, 2983–2985 | 2985