An efficient synthesis of 1-naphthylbis(oxazoline) and exploration of the scope in asymmetric catalysis

Article abstract of DOI:10.1002/ejoc.200390036

Both enantiomers of 1-naphthylglycine were obtained in 99% ee by enzymatic resolution of the corresponding racemic amino acid amide, giving access to the novel ligands (R)- and (S)-naphthylbis(oxazoline). Initial studies provided insight into the scope an

Full text of DOI:10.1002/ejoc.200390036

FULL PAPER
An Efficient Synthesis of 1-Naphthylbis(oxazoline) and Exploration of the
Scope in Asymmetric Catalysis
Hester L. van Lingen,[a] Jeroen K. W. van de Mortel,[a] Koen F. W. Hekking,[a]
Floris L. van Delft,[a] Theo Sonke,[b] and Floris P. J. T. Rutjes*[a]
Keywords: Asymmetric catalysis / Amino acids / Enzymatic resolution / Oxazolines / Lewis acids
Both enantiomers of 1-naphthylglycine were obtained in
99% ee by enzymatic resolution of the corresponding ra-
cemic amino acid amide, giving access to the novel ligands
stituted bis(oxazoline) and its steric influence compared to
other bis(oxazolines) in catalytic asymmetric synthesis.
(R)- and (S)-naphthylbis(oxazoline). Initial studies provided ( Wiley-VCH Verlag GmbH & Co KGaA, 69451 Weinheim,
insight into the scope and limitations of the (S)-naphthyl-sub-
Germany, 2003)
Introduction
During the last decade, bis(oxazoline) (box) based metal
complexes[1] have emerged as highly enantioselective cata-
lysts in
a wide range of transformations such as
DielsϪAlder[2] and hetero-DielsϪAlder reactions,[3] aldol
reactions,[4] Michael additions,[5] and FriedelϪCrafts reac-
tions.[6] These conversions generally proceed via in situ
formation of a chiral Lewis acid catalyst through bidentate
complexation of the box ligand to a transition metal. So
far, many variations at the core of the box ligand have been
made, including variation of substituents (R1) on the
bridging methylene group,[7] and the introduction of addi-
tional stereocenters on the oxazoline rings (R2, R3) in such
a way that the C2-symmetry is maintained (type A, Fig-
ure 1).[8] A larger structurally deviating box-ligand features
a bridging pyridine moiety resulting in tridentate coordinat-
ing box ligands (type B).[9] Generally, the commercially
available box ligands of type A (with R2 ϭ R3 ϭ H) are
most often used. Recently, the groups of Pfaltz, Helmchen
and Williams independently developed a new type of non-
C2-symmetric oxazoline-based P,N-ligands (type C), suit-
able for application in palladium-catalyzed asymmetric re-
actions.[10]
Figure 1. General structure of (bis)oxazoline-based ligands
Although many different substituted bis(oxazoline) li-
gands have been synthesized, only a few ligands of type A
(R ϭ isopropyl, tert-butyl, benzyl, phenyl with R2 ϭ R3 ϭ
H) are frequently used. Despite the versatility of these li-
gands, there is not one single ligand that gives optimal re-
sults in every type of conversion; the most suitable ligand
is usually identified by trial-and-error methods.
The frequently observed reversal of facial selectivity ex-
hibited by alkyl- vs. aryl-substituted box ligands is remark-
able. This phenomenon was first reported by the group of
Jørgensen and was ascribed to a change in geometry of the
metal center (from distorted square planar for tert-butyl-
box to tetrahedral for phenyl-box).[11] Alternatively, the
group of Evans performed solid-state crystallization studies
on the phenyl-box and the tert-butyl-box·Cu(H2O)2 com-
plex, both complexes having distorted square planar geo-
metries (with an average distortion of Ϫ9° and ϩ33°, re-
spectively).[12] Therefore, Evans concluded that the phenyl-
box complex exhibits a reduced propensity to deviate from
square planarity than the tert-butyl-box complex. Although
this forms no conclusive evidence for or against JørgensenЈs
proposal, it reduces the possibility of a planar-tetrahedral
equilibrium in solution, based on a series of studies of
bis(N-alkylsalicylaldiminato)copper() complexes.[13] Very
recently, Jørgensen stated that the reversal of selectivity is
mainly caused by differences in the flexibility and dynamics
[a]
Department of Organic Chemistry, NSRIM, University of
Nijmegen,
Toernooiveld 1, 6525 ED Nijmegen, The Netherlands
Phone: (inter.) ϩ31-24/365-3202
Fax: (internat.) ϩ31-24/365-3393
E-mail: rutjes@sci.kun.nl
[b]
DSM Research, Life Science Products, Advanced Synthesis
and Catalysis,
P.O. Box 18, 6160 MD Geleen, The Netherlands
Supporting information for this article is available on the
WWW under http://www.eurjoc.org or from the author.
Eur. J. Org. Chem. 2003, 317Ϫ324  2003 WILEY-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim 1434Ϫ193X/03/0102Ϫ0317 $ 20.00ϩ.50/0
317

F. P. J. T. Rutjes et al.
FULL PAPER
of the ligand. This includes not only the ability of the com- limitations. A retrosynthetic analysis shows that ligand 1 is
plex to interconvert between square-planar and tetrahedral accessible from the corresponding enantiopure amino acid
conformations, but also the ability of the oxazoline sub- 2 (Scheme 1), which, in turn, is derived from the racemic
stituent to flip between pseudo-axial and pseudo-equator- amino acid amide 3 through an enzymatic resolution pro-
ial positions.[14]
cess that has been extensively studied in our group.[18] The
For the alkyl-substituents (R), it is generally found that amide was expected to be formed upon treatment of naph-
with an increase of steric bulk of the substituent the enanti- thaldehyde (4) under Strecker conditions, followed by par-
oselectivity is increased. To the best of our knowledge, the tial hydrolysis of the resulting cyanide.
steric influence of aryl-substituents has only once been
extensively studied.[15] It was observed that the use of 2-
naphthyl-box in combination with Mg(OTf)2 induces a
small increase in enantioselectivity in the DielsϪAlder reac-
tion of N-acryloyl-substituted oxazolidinone with cyclo-
pentadiene compared to the use of phenyl-box (from 70 to
77% ee). Hence, the use of a more extended aromatic system
on the chiral ligand appears to have a positive effect on the
enantioselectivity of a tetrahedrally organized chiral cata-
lyst. Moreover, a single example of the use of the bulky 1-
naphthyl-box ligand in asymmetric DielsϪAlder reactions
has been described by the group of Evans.[16] Although the
enantioselectivity was better than the phenyl-box ligand, it
was still moderate and considerably lower than for the
alkyl-substituted ligands. Surprisingly, the synthesis of this
1-naphthyl-substituted ligand has not been reported yet.
As part of a research project aimed at controlling the
enantioselectivity of catalytic processes by variation of the
electronic and steric properties of the substituents on the
bis(oxazoline) ligands, we became interested in the 1-naph-
thyl-box ligand. Therefore, we performed PM3-level calcu-
lations[17] on the (S)-1-naphthyl- and (S)-phenyl-box ligand-
copper() catalyst, choosing methyl pyruvate as the com-
plexing compound since it is often used as the substrate in
different types of reactions (Figure 2). In both cases, the
complex adapted a distorted square-planar geometry in
which one side of methyl pyruvate was blocked from attack
by an incoming nucleophile, suggesting a favorable attack
on the Si-face. A comparison of these complexes clearly
shows an increase of shielding of the Re-face of methyl
pyruvate in the complex with the 1-naphthyl-substituted li-
gand, predicting an enhanced enantioselectivity compared
to the phenyl-box catalyst.
Scheme 1. Retrosynthesis of (S)-1-naphthylbis(oxazoline)
Results and Discussion
The synthesis of 1-naphthylglycine amide (3) started with
a modified Strecker reaction (HCN prepared in situ from
equimolar amounts of NH4Cl and NaCN in concentrated
ammonia) on freshly distilled 1-naphthaldehyde to give am-
inonitrile 5 (Scheme 2). After the reaction had finished, in
situ reaction with benzaldehyde in the presence of NaOH
(pH ഠ 10) led to conversion of the amino function into
the Schiff base, with concomitant partial hydrolysis of the
cyanide group (presumably via the five-membered ring N,O-
acetal 6).[19] Subsequent hydrolysis of the benzaldimine un-
der the influence of concentrated HCl in acetone resulted
in the crystalline HCl salt of the desired amino acid amide
3. The whole sequence was efficiently carried out in 62%
overall yield with only a single purification step at the end
(precipitation from acetone).[18]
Figure 2. PM3 model of [Cu(S)-1-Np-box] and [Cu(S)-Ph-box]
complex with methyl pyruvate
Considering the few experiments that have been carried
out with the latter ligand, the promising PM3-model and
the anticipated facile access to the required amino acid, we
set out to develop a straightforward and efficient synthesis
of the 1-naphthyl-box ligand and to explore its scope and Scheme 2. Synthesis of 1-naphthylglycine amide
318 Eur. J. Org. Chem. 2003, 317Ϫ324

An Efficient Synthesis of 1-Naphthylbis(oxazoline)
FULL PAPER
Table 1. Enzymatic resolution of amino acid amide 3
Enzyme source
(S)-2
(R)-3
Yield (%)[a]
ee (%)[b]
Yield (%)[a]
ee (%)[b]
P. putida ATCC 12633
O. anthropi NCIMB 40321
86
98
89
99
90
90
99
99
[a]
[b]
Isolated yield. Enantiomeric excess was determined by NMR spectroscopy.[29]
Treatment of racemic 3·HCl with an -specific aminopep-
tidase present in whole cells of the microorganism
Pseudomonas putida ATCC 12633[18] (pH ϭ 8.3, T ϭ 37
°C) led to a mixture of the corresponding (R)-amide 3 and
(S)-acid 2 in good yield and with high enantioselectivity
[45% (99% ee) and 43% (89% ee), respectively; Table 1]. The
amide and acid could be easily separated by a simple extrac-
tion with CH2Cl2 as a result of the low solubility of the
amide in a basic aqueous environment. However, the low
solubility of the amino acid amide caused the resolution
(the enzyme optimum lies around pH 9) to proceed very
slowly. The solubility problem was overcome by applying an
-specific amidase from the microorganism Ochrobactrum
anthropi NCIMB 40321,[20] which has a good activity at pH
6.5 and is able to withstand a higher reaction temperature
(50 °C). The use of this amidase resulted in an improvement
of both yield and enantioselectivity giving the (R)-amide
and (S)-acid in greater than 98% enantiopurity. Beside the
synthesis of the (S)-acid 2, the corresponding (R)-acid
could also be obtained by hydrolysis of (R)-3 with a non-
selective amidase produced by Rhodococcus erythropolis
NCIMB 11540.[21]
Scheme 3. Synthesis of (S)-1-naphthylbis(oxazoline) (S)-1
box 1 (Figure 3). Furthermore, we anticipated being able to
identify reactions that could especially benefit from the use
of this particular ligand. These reactions were carried out
With the required amino acid in hand, the 1-naphthyl-
box ligand (S)-1 was synthesized according to a previously
reported method (Scheme 3).[22] LiAlH4 reduction of (S)-1-
naphthylglycine (2) to the amino alcohol 8 and subsequent
acylation with dimethylmalonyl dichloride resulted in di-
amide (S)-9 in 71% yield. This diamide was then smoothly
Figure 3. Reference box ligands
cyclized to the desired bis(oxazoline) via in situ formation with the [box-CuII] catalyst complexed to methyl pyruvate.
of the bis(tosylate); recrystallization finally afforded As a starting point, we tested the novel 1-Np-box ligand
enantiopure (S)-1 in 60% overall yield starting from 1- in the previously reported Mukaiyama aldol reaction of
naphthylglycine (2) and 37% yield from naphthyldehyde (4). methyl pyruvate with silyl ketene acetal 11 in THF
The scope and limitations of the (S)-1-naphthyl-box li- (Table 2).[24] However, use of this ligand-Cu(OTf)2 complex
gand in catalytic asymmetric synthesis were tested in a resulted only in a decrease of enantioselectivity compared
series of reactions and compared to results obtained with to Ph-box (19% vs. 24% ee). In contrast, a significantly
(R)-phenyl-box, (4S,5S)-diphenyl-box[23] and (S)-tert-butyl- higher ee was observed when using the (4S,5S)-diphenyl-
Eur. J. Org. Chem. 2003, 317Ϫ324
319

F. P. J. T. Rutjes et al.
FULL PAPER
box and (S)-tBu-box ligands (68% and 97% ee, respect- ile would give dioxenone 16, which is an interesting building
ively). Apparently, in this type of aldol reaction, the 4-aryl- block for further transformations.[27] In this reaction, a sig-
substituted ligands do not have a crucial influence on the nificant increase in enantioselectivity was observed by re-
stereochemical outcome of the reaction.
placing Ph-box by 1-Np-box (16% to 51% ee in CH2Cl2 and
17% to 43% ee in THF). The best result, however, for this
addition was obtained with tBu-box in THF (74% ee). In
this reaction, a reversal of enantioselectivity in the presence
of aryl-substituted box ligands compared to the alkyl-sub-
stituted analogues was observed, as discussed before. Thus,
in this case the enantioselectivity of the 1-Np-box-induced
addition closely approaches the result of the tBu-box-Cu(-
OTf)2 complex, particularly in CH2Cl2, affording the op-
posite enantiomer.
Table 2. Results of the Mukaiyama aldol reaction
Ligand
THF
Yield (%)[a]
ee (%)[b]
Table 4. Results of the addition reaction of silyl enol ether 15
(S)-1-Np-box
(R)-Ph-box[c]
n.d.
73
82
19 (S)
24 (R)
68 (S)
97 (S)
(4S,5S)-diPh-box
(S)-tBu-box[c]
89
[a]
[b]
Isolated yield. Enantiomeric excess was determined by chiral
HPLC on a Chiralcel OD column (hexane/IPA, 99:1, 0.5 mL/min).
[c]
Data correspond to literature values.[24]
[a]
Ligand
CH2Cl2
THF
Yield (%)[b] ee (%)[c] Yield (%)[b] ee (%)[c]
(S)-1-Np-box 73
51 (R)
16 (S)
82
68
43 (R)
17 (S)
74 (S)
Another type of reaction in which methyl pyruvate was
applied as the substrate in combination with box-Cu(OTf)2
complexes is the hetero-DielsϪAlder reaction with (E)-1-
methoxy-3-(trimethylsilyloxy)-1,3-butadiene (Danishefsky’s
(R)-Ph-box
(S)-tBu-box
99
90
49 (S)[d] 76
[a]
In CH2Cl2 a mixture of Me3Si-protected and unprotected alco-
diene, 13; Table 3).[25] In both solvents, we observed no sig- hol was observed. Isolated yield. Enantiomeric excess was de-
[b]
[c]
termined by HPLC on a Chiralpak AD column (heptane/2-pro-
nificant differences between the Ph- and the 1-Np-box li-
panol, 95:5, 0.5 mL/min). [d] The geometry of the major enantiomer
gand; a slight increase in CH2Cl2 (13% to 19% ee, respect-
was tentatively assigned based on the outcome of the Mukaiyama
ively) and a slight decrease in THF (35% to 22% ee, respect-
aldol and hetero-DielsϪAlder reaction with methyl pyruvate, with
ively). In the case of the aryl-substituted box ligands, the
selectivity was dependent on the solvent. This effect is in
line with the results of Jørgensen.[14] In conclusion, the use
of the 1-Np-box did not give the expected enhancement in
enantioselectivity of this reaction.
the use of (S)-tBu-box as ligand.
Finally, we decided to look into reactions where the reac-
tion center is relatively remote from the coordination site
and therefore may require an extended steric effect of the
ligand, such as the MukaiyamaϪMichael reaction
(Table 5). Knoevenagel adduct 17 was used as the substrate
in an addition reaction with silyl ketene acetal 11 catalyzed
Table 3. Results of the hetero-DielsϪAlder reaction
Table 5. Results of the MukaiyamaϪMichael addition
Ligand
CH2Cl2
THF
Yield (%)[a] ee (%)[b] Yield (%)[a] ee (%)[b]
(S)-1-Np-box n.d.
(R)-Ph-box[c]
n.d.
(S)-tBu-box[c] n.d.
19 (S) 90
13 (R) 99
Ͼ98 (S) 99
22 (R)
35 (S)
Ͼ95 (S)
Ligand
CH2Cl2
Yield (%)[a]
ee (%)[b]
[a]
[b]
Isolated yield. Enantiomeric excess was determined by chiral
[c]
(S)-1-Np-box
(R)-Ph-box[c]
(S)-tBu-box[c]
55
67
99
70 (S)
56 (S)
Ͼ90 (R)
GC on a γ-CD column (110 °C). Data correspond to literature
values.[25]
[a]
Isolated yield after chromatography.[b] Enantiomeric excess was
In a search for factors that could favor the use of the 1-
Np-box, we came across addition reactions with the silyl
determined by HPLC on a Chiralpak AD column (heptane/2-pro-
[c]
panol, 98:2, 1.0 mL/min).
Data correspond to literature
enol ether 15[26] (Table 4). Use of this rather large nucleoph- values.[28]
320
Eur. J. Org. Chem. 2003, 317Ϫ324

An Efficient Synthesis of 1-Naphthylbis(oxazoline)
FULL PAPER
by a box ligand with Cu(SbF6)2.[28] Indeed, we were pleased
to find that catalysis with the 1-Np-box ligand afforded the
addition product with good enantioselectivity (70% ee) and
significant improvement of the ee compared to Ph-box
(52% ee). Surprisingly, we observed a reversal in stereo-
chemistry on going from the phenyl- to the naphthyl-box
ligand. Again, the highest enantioselectivity so far was
achieved with the tBu-box ligand (Ͼ90% ee), but also in
this case the product of opposite stereochemistry was ob-
tained with the 1-Np-box ligand. Moreover, we expect that
optimization studies on the use of 1-Np-box will further
increase the enantioselectivity. From these results, it may be
postulated that the use of 1-Np-box as a ligand is especially
promising in reactions where the addition takes place at a
functional group that is relatively remote from the coor-
dination site.
Experimental Section
General: All reactions were carried out under an atmosphere of dry
nitrogen or argon. Standard syringe techniques were applied for
transfer of dry solvents and air- or moisture-sensitive reagents.
Solvents were distilled from the appropriate drying agents immedi-
ately prior to use. Unless stated otherwise, all chemicals were pur-
chased and used as such. 1-Naphthaldehyde and 2,2,6-trimethyl-
1,3-dioxen-4-one were distilled before use. S,S-DBTAAN, whole
cells of Pseudomonas putida ATCC 12633 and a cell-free extract
containing the Ochrobactrum anthropi NCIMB 40321 -amidase
were kindly provided by DSM. Dowex 50 W ϫ 4 (Fluka) was used
for ion exchange with 2  NH4OH as the eluent and 2  HCl as
regeneration agent. IR spectra were recorded on an ATI Mattson
Genesis Series FTIP spectrometer. Absorptions are reported in
cmϪ1. NMR spectra were recorded with Bruker DPX200 and
DMX300 spectrometers from CDCl3 solutions (unless otherwise
reported) using TMS as internal standard. Shifts are given in ppm.
Optical rotations were determined with a PerkinϪElmer 241 po-
larimeter. Mass spectra were measured with a Fisons (VG) Micro-
mass 7070E apparatus. Melting points are uncorrected.
Conclusions
N-Benzylidene-1-naphthylglycine Amide (7): Concentrated ammo-
nia (90 mL, 25% in H2O), NH4Cl (6.50 g, 121 mmol) and 1-naph-
thaldehyde (19.90 g, 121 mmol) were added to a solution of NaCN
(5.97 g, 122 mmol) in methanol (100 mL). After stirring for 5 h at
room temperature, the reaction was complete according to TLC
(Et2O/heptane, 1:1). Toluene (100 mL), aqueous NaOH (122 mL,
122 mmol, 1.0 equiv., 1 ) and benzaldehyde (12.9 g, 121 mmol,
0.99 equiv.) were then added to the reaction mixture. After stirring
for 1 h, a precipitate was formed, which was collected by filtration,
washed with heptane and dried in vacuo. Benzaldimine 7 was isol-
ated as a white solid (23.0 g, 66% from 1-naphthaldehyde) and used
directly for the next step. 1H NMR (CDCl3, 298 K, 200 MHz): δ ϭ
8.25 (s, 1 H, CHϭN), 7.88Ϫ7.47 (m, 12 H, Ar-H), 7.24 (br., 1 H,
NH), 5.94 (br., 1 H, NH), 5.67 (s, 1 H, CH) ppm. 13C NMR
(CDCl3, 298 K, 50 MHz): δ ϭ 174.52 (CϭO), 163.25 (CϭN),
135.61, 135.41, 134.28, 131.57, 131.12, 128.99 (Ar-C), 128.90 (2C,
Ar-C), 128.77 (Ar-C), 128.56 (2C, Ar-C), 127.27, 126.63, 125.95,
125.53, 124.58 (Ar-C), 73.67 (CϪN) ppm. Rf (Et2O): 0.21.
Both (S)- and (R)-1-naphthylglycine were accessible
through enzymatic resolution using enzymes from different
sources, i.e. Pseudomonas putida and Ochrobactrum
anthropi. These very useful amino acids could be effectively
transformed into the enantiomerically pure 1-naphthylbi-
s(oxazoline) ligand. (S)-1-Naphthyl-box was successfully
synthesized in 60% yield from (S)-1-naphthylglycine and in
37% yield from 1-naphthaldehyde.
Studies of the scope of this novel 1-naphthyl-substituted
ligand afforded several interesting results. Firstly, the influ-
ence of the enlarged 4-aryl substituent was not unambigu-
ous. In the Mukaiyama aldol reaction and the hetero-
DielsϪAlder reaction, the enantiomeric excesses were com-
parably moderate for the use of 1-naphthyl- and phenyl-box
as ligands. However, addition of a more bulky nucleophile,
such as the silyl enol ether 15, clearly showed the construct-
ive influence of enlargement of the aryl substituent, re-
sulting in a much higher enantioselectivity with 1-naphthyl-
box than with phenyl-box. In this new addition reaction to
methyl pyruvate, catalysis by 1-naphthyl- vs. tert-butyl-box
resulted in similar enantioselectivity, particularly in
CH2Cl2. We observed several cases of reversal of selectivity,
which are similar to previously reported results. Neverthe-
less, the explanation for this inversion remains ambiguous
since it appears to be dependent on a range of parameters,
such as ligand, metal, solvent and substrate. Optimization
studies on this reaction and follow-up chemistry are cur-
rently being performed.
1-Naphthylglycine Amide (3): Concentrated HCl (3.7 mL,
37.5 mmol, 0.98 equiv.) was added to a solution of 7 (11.0 g,
38.2 mmol) in acetone (500 mL). After stirring for 1 h at room tem-
perature, a precipitate was formed, which was collected by filtra-
tion, washed with acetone and dried in vacuo. The amide 6 was
isolated as its HCl salt (8.48 g, 94%). Extraction of an aqueous
solution of the hydrochloride (pH ഠ 10, 500 mL) with CH2Cl2 (3
ϫ 200 mL) afforded the HCl-free product as a white, crystalline
solid (8.01 g, 89% from 7). 3: IR: ν˜ ϭ 3263 cmϪ1 (CONH2), 3318,
3064, 1657 (CONH2), 1624 (CONH2). 1H NMR (CDCl3, 298 K,
200 MHz): δ ϭ 8.15 (d, J ϭ 7.6 Hz, 1 H, Ar-H), 7.84 (m, 2 H, Ar-
H), 7.58Ϫ7.41 (m, 4 H, Ar-H), 6.84 (br., 1 H, NH2), 5.69 (br., 1
H, NH2), 5.19 (s, 1 H, CH) ppm. 1H NMR (CD3OD, 298 K,
300 MHz): δ ϭ 8.23 (d, J ϭ 8.5 Hz, 1 H, Ar-H), 7.85 (m, 2 H, Ar-
H), 7.58Ϫ7.43 (m, 4 H, Ar-H), 5.21 (s, 1 H, CH) ppm. 13C NMR
(CD3OD, 298 K, 75 MHz): δ ϭ 178.90 (CϭO), 138.54, 132.45,
129.92, 129.57, 127.49, 126.84, 126.54, 126.16, 124.54 (Ar-C), 56.92
(CϪN) ppm. Rf (CHCl3/MeOH/NH3(25%), 60:45:20): 0.45.
HRMS(EIϩ): calculated for (C12H12N2O)ϩ 200.09496, found
200.09486.
Similarly, in the MukaiyamaϪMichael reaction the effect
of enlargement of the aryl substituent on the bis(oxazoline)
was clearly observed. We now expect that our ligand could
afford good results in reactions where the distance between
reaction center and coordination site is relatively large. In
this respect, the versatility of the 1-Np-box ligand is par-
ticularly enhanced by affording the opposite enantiomer
than the tBu-box-induced reaction.
Preparation of (S)-1-Naphthylglycine [(S)-2] and (R)-1-Naphthylgly-
cine Amide [(R)-3] Using Pseudomonas putida
l-Aminopeptidase: A
Eur. J. Org. Chem. 2003, 317Ϫ324
321

F. P. J. T. Rutjes et al.
solution of 3·HCl (2.28 g, 9.65 mmol) in H2O (135 mL) was J ϭ 7.5 Hz, 1 H, Ar-H), 7.79 (d, J ϭ 8.2 Hz, 1 H, Ar-H), 7.62Ϫ7.46
FULL PAPER
brought to pH 8.3 with aqueous KOH (5 ). After addition of
aqueous MnSO4 (11.5 mL, 10 m) and whole cells of Pseudomonas
putida ATCC 12633 (2.17 g), the mixture was shaken for 90 h at 37
°C. The mixture was cooled to room temperature and the aqueous
(m, 4 H, Ar-H), 4.92 (dd, J ϭ 3.9, 8.0 Hz, 1 H, CH), 3.96 (dd, J ϭ
3.9, 10.7 Hz, 1 H, CH2), 3.67 (dd, J ϭ 8.0, 10.7 Hz, 1 H, CH2)
ppm. 13C NMR (CD3OD, 298 K, 75 MHz): δ ϭ 138.83, 135.32,
132.43, 129.69, 128.75, 127.20, 126.56, 126.46, 124.16, 123.66 (Ar-
layer was extracted with CH2Cl2 (3 ϫ 75 mL). Both layers were C), 68.61 (CH2), 53.76 (CH) ppm. [α]2D5 ϭ ϩ83 (c ϭ 0.50, MeOH)
filtered through Hyflo to remove the enzyme residue. The aqueous
layer was changed to pH ഠ 5 and purified by ion-exchange chroma-
tography. Removal of the solvent afforded acid (S)-2 as an off-white
solid (834 mg, 43%). After removal of the solvent from the CH2Cl2
layer, amide (R)-3 was obtained as a light brown solid (791 mg,
41%). The ee was determined by 1H NMR spectroscopy to be 73%
for (S)-2 and 99% for (R)-3.[29]
[ref.[31] for (R)-8 [α]2D5 ϭ Ϫ85 (c ϭ 0.50, MeOH)]. Rf (CHCl3/
MeOH/NH3(25%), 60:45:20): 0.88.
(S,S)-N,NЈ-Bis[2-hydroxyl-1-(1-naphthyl)ethyl]-2,2-dimethyl-1,3-
propanediamide [(S)-9]: A solution of amino alcohol (S)-8 (1.20 g,
6.44 mmol) and Et3N (4.6 mL, 33.2 mmol, 5.2 equiv.) in CH2Cl2
(40 mL) was cooled to 0 °C. A solution of 2,2-dimethylmalonyl
dichloride[32] (0.562 g, 3.3 mmol, 0.52 equiv.) in CH2Cl2 (15 mL)
was then added with a cannula. After stirring for 90 min, CH2Cl2
(40 mL) was added and the mixture was washed with aqueous HCl
(150 mL, 1 ). The aqueous layer was extracted with CH2Cl2 (3 ϫ
100 mL). The combined organic layers were washed with saturated
aqueous NaHCO3 (150 mL) and the aqueous layer was extracted
with CH2Cl2 (100 mL). The combined organic layers were washed
with brine (150 mL) and the aqueous layer was extracted with
CH2Cl2 (100 mL). The combined organic layers were dried
(MgSO4) and concentrated in vacuo. The resulting brown solid was
recrystallized from EtOAc, yielding diamide (S)-9 as an off-white
solid (1.24 g, 79%). (S)-9: M.p. 99 °C. IR (solvent): ν˜ ϭ 3337 (OH),
Preparation of (S)-1-Naphthylglycine [(S)-2] and (R)-1-naphthylgly-
cine Amide [(R)-3] Using Ochrobactrum anthropi l-Amidase: A solu-
tion of 3·HCl (12.0 g, 50.7 mmol) in H2O (85 mL) was brought to
pH 6.5 with aqueous KOH (5 ). After addition of aqueous ZnSO4
(1.5 mL, 80 m) and a cell-free extract containing the -amidase
from Ochrobactrum anthropi NCIMB 40321 in water (2 mL), the
total weight of the mixture was brought to 120 g by addition of
water. The mixture was shaken for 70 h at 50 °C, after which an
extra 2 mL of cell-free extract was added and the mixture was
shaken for another 20 h. After cooling the mixture to room temper-
ature, the mixture was filtered through Hyflo. The resulting clear
solution was brought to pH ഠ 9 by addition of aqueous KOH (5
) and extracted with CH2Cl2 (3 ϫ 250 mL). The aqueous layer
was then changed to pH ഠ 5 and purified by ion-exchange chroma-
tography. Removal of the solvent resulted in isolation of acid (S)-
2 as an off-white solid (5.04 g, 49%). The ee was determined by 1H
NMR spectroscopy to be 99%.[29] (S)-2: IR: ν˜ ϭ 3047, 1695, 1618,
3047, 2971, 2933, 2876, 1642 (CϭO), 1522 cmϪ1 1H NMR
.
(CD3Cl, 298 K, 200 MHz): δ ϭ 8.04 (m, 2 H, Ar-H), 7.89Ϫ7.76
(m, 4 H, Ar-H), 7.48 (m, 4 H, Ar-H), 7.35 (m, 4 H, Ar-H), 7.18
(d, J ϭ 7.9 Hz, 2 H, NH), 5.96 (m, 2 H, CH), 4.10Ϫ3.94 (m, 4 H,
CH2), 3.92 (br., 2 H, OH), 1.50 (s, 6 H, CH3) ppm. 13C NMR
(CD3Cl, 298 K, 50 MHz): δ ϭ 174.09 (2C, CϭO), 134.34, 134.03,
131.04, 129.08, 128.66, 126.79, 126.04, 125.34, 123.48, 122.91 (10C,
Ar-C), 65.59 (2C, CH2), 51.87 (2C, CH), 50.24 [C(CH3)], 23.85 (2C,
CH3) ppm. C29H30N2O4·1/2H2O: calcd. C 72.63, H 6.52, N 5.84;
found C 72.36, H 6.32, N 5.67. [α]2D5 ϭ Ϫ39 (c ϭ 0.50, MeOH). Rf
(EtOAc/MeOH, 95:5) ϭ 0.41.
1570, 1514 cmϪ1
.
1H NMR (D2O, 298 K, 200 MHz): δ ϭ
8.13Ϫ7.98 (m, 3 H, Ar-H), 7.66Ϫ7.54 (m, 4 H, Ar-H), 5.50 (s, 1
H, CH) ppm. 13C NMR (CD3OD, 298 K, 75 MHz): δ ϭ 173.85
(CϭO), 135.60, 133.72, 132.73, 130.64, 129.97, 127.96, 127.70,
127.16, 126.49, 124.32 (Ar-C), 56.45 (CϪN) ppm. [α]2D5 ϭ Ϫ7.9
(c ϭ 0.05, H2O) [ref.[30] for (R)-2 [α]2D5 ϭ ϩ8.0 (c ϭ 0.05, H2O)]. Rf
(CHCl3/MeOH/NH3(25%), 60:45:20): 0.65.
(S,S)-2,2-Bis{2-[4(S)-(1-naphthyl)-1,3-oxazolinyl]}propane [(S)-1]: A
solution of tosyl chloride (1.98 g, 10.4 mmol, 2.1 equiv.) in CH2Cl2
(40 mL) was added via cannula to a solution of (S)-9 (2.32 g,
4.93 mmol), DMAP (65 mg, 0.53 mmol, 11 mol %) and triethylam-
ine (3.5 mL, 25.2 mmol, 5.1 equiv.) in CH2Cl2 (125 mL). After stir-
ring at room temperature for two days, the mixture was washed
with saturated aqueous NH4Cl (200 mL) and the aqueous layer was
extracted with CH2Cl2 (3 ϫ 200 mL). The combined organic layers
were washed with saturated aqueous NaHCO3 (400 mL), and the
aqueous layer was extracted with CH2Cl2 (3 ϫ 300 mL). The com-
bined organic layers were dried (MgSO4) and concentrated in va-
cuo to yield the crude product as an off-white solid. Purification
by column chromatography (EtOAc/heptane, 4:6) resulted in isola-
tion of (S)-1 as a white solid (1.82 g, 85%). The product was puri-
fied by recrystallization (EtOAc/heptane) and was used as such for
catalysis. (S)-1: M.p. 59 °C. IR (solvent): ν˜ ϭ 3057, 2981, 2893,
After removal of the solvent from the CH2Cl2 layer, (R)-3 was ob-
tained as a white solid (5.59 g, 43%). The ee was determined by 1H
1
NMR to be 99%.[29] (R)-3: H NMR (CD3OD, 298 K, 300 MHz):
δ ϭ 8.24 (d, J ϭ 8.5 Hz, 1 H, Ar-H), 7.85 (m, 2 H, Ar-H),
7.56Ϫ7.41 (m, 4 H, Ar-H), 5.21 (s, 1 H, CH). [α]2D5 ϭ Ϫ114 (c ϭ
0.5, MeOH).
(R)-1-Naphthylglycine [(R)-2]: To a 5% solution of (R)-3 (100 mg,
0.50 mmol) in a buffer (pH 8: 500 mL of 0.1  NaH2PO4 ϩ 467 mL
of 0.1  NaOH) was added 50 mg of dried whole cells of Rhodoc-
cocus erythropolis NCIMB 11540 and the reaction mixture was
stirred at 37 °C for 48 h. The reaction mixture was filtered through
Hyflo, the filtrate was acidified and purified by ion-exchange chro-
matography to give (R)-2 (74 mg, 74%) as a white solid. Analytical
data were identical to the literature data.[30]
1657 (CϭN) cmϪ1 1H NMR (CDCl3, 298 K, 200 MHz): δ ϭ
.
(S)-1-Naphthylglycinol [(S)-8]: LiAlH4 (6.70 g, 176 mmol) was ad- 7.90Ϫ7.75 (m, 6 H, Ar-H), 7.60Ϫ7.40 (m, 8 H, Ar-H), 5.99 (dd,
ded in four portions to a suspension of acid (S)-2 (3.50 g, J ϭ 8.1, 10.1 Hz, 2 H, CH), 4.96 (dd, J ϭ 8.0, 10.2 Hz, 2 H, CH2),
17.4 mmol) in THF (250 mL). After refluxing for 17 h, the mixture 4.13 (t, J ϭ 8.1 Hz, 2 H, CH2), 1.80 (s, 6 H, CH3) ppm. 13C NMR
was cooled to 0 °C and quenched by addition of H2O (6 mL), aque-
ous NaOH (6 mL, 4 ) and H2O (15 mL) over a period of 15 min.
After stirring for 30 min at room temperature, the mixture was fil-
(CDCl3, 298 K, 300 MHz): δ ϭ 170.58 (2C, CϭN), 138.35, 133.82,
130.52, 129.02, 127.83, 126.23, 125.74, 125.56, 123.44, 122.69 (20C,
Ar-C), 75.06 (2C, CH2), 66.17 (2C, CH), 39.33 [C(CH3)2], 24.61
tered. The residue was washed with EtOAc (3 ϫ 150 mL) and the (2C, CH3) ppm. HRMS(EIϩ): calculated for (C29H26N2O2)ϩ
combined filtrates were concentrated in vacuo to yield amino alco-
hol (S)-8 as a yellow solid (2.94 g, 90%). (S)-8: H NMR (CDCl3,
434.1994, found 434.1996. C29H26N2O2·1/4H2O: calcd. C 79.34, H
6.08, N 6.38; found C 79.36, H 5.84, N 6.23. [α]2D5 ϭ ϩ180 (c ϭ
1
298 K, 300 MHz): δ ϭ 8.11 (d, J ϭ 8.9 Hz, 1 H, Ar-H), 7.88 (d, 0.50, CDCl3). Rf (EtOAc/heptane, 4:6) ϭ 0.23.
322 Eur. J. Org. Chem. 2003, 317Ϫ324

An Efficient Synthesis of 1-Naphthylbis(oxazoline)
FULL PAPER
A. Jørgensen, Angew. Chem. 2000, 112, 3702Ϫ3733; Angew.
Chem. Int. Ed. 2000, 39, 3558Ϫ3588.
General Method for the Preparation of Methyl 3-(2,2-Dimethyl-6-
oxo-6H-[1,3]dioxin-4-yl)-2-hydroxy-2-methylpropionate (16): (2,2-
dimethyl-6-methylene-6H-[1,3]dioxin-4-yloxy)trimethylsilane (15)
was prepared according to the method published by Krohn and
Schafer.[33] Box ligand (11 mol %) and Cu(OTf)2 (7 mg, 10 mol %)
were weighted under Ar. After stirring for 1 h under Ar, solvent
(2 mL) was added and the stirring under Ar was continued for over-
night. After cooling the solution to Ϫ78 °C, methyl pyruvate (19
µL, 0.21 mmol) was added. After stirring for 30 min at Ϫ78 °C,
enol 15 (50 µL, 1.1 equiv.) was added and the mixture was allowed
to warm to room temperature overnight and then stirred for an-
other night. After addition of saturated aqueous NaHCO3 (10 mL)
and solvent (10 mL), the layers were separated and the organic
layer was washed with brine (10 mL), dried (MgSO4) and concen-
trated in vacuo. This afforded a mixture of addition product 16
and its TMS-protected analogue.
[4]
[5]
For a recent example, see: D. A. Evans, J. M. Janey, Org. Lett.
2001, 3, 2101Ϫ2103.
For recent examples, see:
[5a]
G. Desimoni, G. Faita, S. Filip-
pone, M. Mella, M. G. Zampori, M. Zema, Tetrahedron 2001,
[5b]
57, 10203Ϫ10212.
D. A. Evans, K. A. Scheidt, J. N. John-
ston, M. C. Willis, J. Am. Chem. Soc. 2001, 123, 4480Ϫ4491.
[6a] W. Zhuang, N. Gathergood, R. G. Hazell, K. A. Jørgensen,
[6]
[6b]
J. Org. Chem. 2001, 66, 1009Ϫ1013.
N. Gathergood, W.
Zhuang, K. A. Jørgensen, J. Am. Chem. Soc. 2000, 122,
12517Ϫ12522.
[7]
I. W. Davies, R. J. Deeth, R. D. Larsen, P. J. Reider, Tetrahedron
Lett. 1999, 40, 1233Ϫ1236.
[8] [8a]
G. Desimoni, G. Faita, M. Mella, Tetrahedron 1996, 52,
[8b]
13649Ϫ13654.
J. Ji, D. M. Barnes, J. Zhang, S. A. King, S.
J. Wittenberger, H. E. Morton, J. Am. Chem. Soc. 1999, 121,
10215Ϫ10216.
[9]
See, for example: A. Datta Gupta, D. Bhuniya, V. K. Singh,
Tetrahedron 1994, 50, 13725Ϫ13730.
Optional Procedure: TBAF (0.2 mL, 1.0 equiv., 1  in THF) was
added to a solution of the crude mixture (0.21 mmol) in THF
(10 mL) at Ϫ78 °C and the orange solution immediately became
green. After stirring for four days at room temp., saturated aqueous
NH4Cl (5 mL) was added and the color of the solution immediately
changed to orange/brown. The mixture was extracted with Et2O (3
ϫ 10 mL), dried (MgSO4) and concentrated in vacuo. This afforded
unprotected alcohol 16 as a yellow oil. Purification by column
chromatography (EtOAc/heptane, 1:4) resulted in alcohol 16 as a
colorless oil. Yields were determined after catalysis (without treat-
ment with TBAF). The TMS/H ratio was determined by NMR
[10] [10a] P. von Matt, A. Pfaltz, Angew. Chem. 1993, 105, 614Ϫ615;
[10b]
Angew. Chem. Int. Ed. Engl. 1993, 32, 566Ϫ568.
J. Sprinz,
[10c]
G. Helmchen, Tetrahedron Lett. 1993, 34, 1769Ϫ1772.
G.
J. Dawson, C. G. Frost, J. M. J. Williams, Tetrahedron Lett.
1993, 34, 3149Ϫ3150.
M. Johannsen, K. A. Jørgensen, J. Org. Chem. 1995, 60,
5757Ϫ5762.
D. A. Evans, J. S. Johnson, C. S. Burgey, K. R. Campos, Tetra-
hedron Lett. 1999, 40, 2879Ϫ2882.
R. Knoch, A. Wilk, K. J. Wannowius, D. Reinen, H. Elias,
Inorg. Chem. 1990, 29, 3799Ϫ3805.
J. Thorhauge, M. Roberson, R. G. Hazell, K. A. Jørgensen,
Chem. Eur. J. 2002, 8, 1888Ϫ1898.
S. Crosignani, G. Desimoni, G. Faita, P. P. Righetti, Tetrahed-
ron 1998, 54, 15721Ϫ15730.
[11]
[12]
[13]
[14]
[15]
[16]
[17]
1
analysis of the methoxy signal H NMR (300 MHz, CDCl3): δ ϭ
3.81 (OH), 3.73 ppm [OSi(CH3)3]. 16: IR (solvent): ν˜ ϭ 3491 (OH),
2997, 2953, 2251, 1730 cmϪ1. 1H NMR (CDCl3, 298 K, 300 MHz):
δ ϭ 5.24 (s, 1 H, α-CH), 3.74 (s, 3 H, CO2CH3), 3.50 (br. s, 1 H,
OH), 2.79 (d, J ϭ 14.7 Hz, 1 H, CH2), 2.52 (d, J ϭ 14.7 Hz, 1 H,
CH2), 1.59 (s, 6 H, 2 ϫ CH3), 1.42 (s, 3 H, CH3) ppm. 13C NMR
(CDCl3, 298 K, 75 MHz): δ ϭ 175.56 (CϭO), 166.60 (CϭO),
160.42 (CϭCϪO), 106.49 [C(Me)2], 95.87 (CH), 72.76 (CϪOH),
52.86 (CH2), 43.50 (OϪCH3), 27.00 (CH3), 25.77 (CH3), 24.04
(CH3) ppm. The enantiomeric excess was determined by HPLC
analysis on a chiralpak AD column (heptane/2-propanol, 95:5,
0.5 mL/min, 35 °C), tR ϭ 37.5 min, 46.6 min.
D. A. Evans, S. J. Miller, T. Lectka, P. von Matt, J. Am. Chem.
Soc. 1999, 121, 7559Ϫ7573.
Optimizations were performed with Spartan (version 5.1.3).
Calculations were carried out without counterions or solvent
molecules.
[18] [18a]
L. B. Wolf, T. Sonke, K. C. M. F. Tjen, B. Kaptein, Q. B.
Broxterman, H. E. Schoemaker, F. P. J. T. Rutjes, Adv. Synth.
[18b]
Catal. 2001, 343, 662Ϫ674.
T. Sonke, B. Kaptein, W. H. J.
Boesten, Q. B. Broxterman, J. Kamphuis, H. E. Schoemaker,
F. Formaggio, C. Toniolo, F. P. J. T. Rutjes, in Stereoselective
Biocatalysis (Ed.: R. N. Patel), Marcel Dekker, New York,
2000, p. 23Ϫ58.
W. H. J. Boesten (DSM/Stamicarbon) British Patent 1548032,
Chem. Abstr. 1997, 87, 39839b.
W. J. J. van den Tweel, T. J. G. M. van Dooren, P. H. de Jonge,
B. Kaptein, A. L. L. Duchateau, J. Kamphuis, Appl. Microbiol.
Biotechnol. 1993, 39, 296Ϫ300.
[19]
[20]
Acknowledgments
DSM is kindly acknowledged for providing the enzymes and (S,S)-
dibenzoyl tartaric anhydride.
[21]
W. H. J. Boesten, M. J. H. Cals (DSM/Stamicarbon) US Patent
4705752, Chem. Abstr. 1987, 105, 170617k.
[22] [22a]
D. A. Evans, G. S. Peterson, J. S. Johnson, D. M. Barnes,
K. R. Campos, K. A. Woerpel, J. Org. Chem. 1998, 63,
4541Ϫ4544. [22b] D. A. Evans, K. A. Woerpel, M. M. Hinman,
M. M. Faul, J. Am. Chem. Soc. 1993, 113, 726Ϫ728.
This ligand was kindly provided by L. Thijs (University of
Nijmegen); [α]2D5 ϭ Ϫ120 (CH2Cl2); for synthesis see: G. Desi-
moni, G. Faita, M. Mella, Tetrahedron 1996, 52, 13649Ϫ13654.
D. A. Evans, C. S. Burgey, M. C. Kozlowski, S. W. Tregay, J.
Am. Chem. Soc. 1999, 121, 686Ϫ699.
S. Yao, M. Johannsen, H. Audrain, R. G. Hazell, K. A.
Jørgensen, J. Am. Chem. Soc. 1998, 120, 8599Ϫ8605.
For examples of enantioselective syntheses of similar diox-
enones, see: J. Krüger, E. M. Carreira, J. Am. Chem. Soc. 1998,
120, 837Ϫ838.
[1]
[1a]
For reviews of box ligands, see:
A. Pfaltz, Acc. Chem. Res.
1993, 26, 339Ϫ345. [1b] A. K. Ghosh, P. Mathivanan, J. Cappi-
[1c]
[23]
ello, Tetrahedron: Asymmetry 1998, 9, 1Ϫ45.
K. A.
Jørgensen, M. Johannsen, S. Yao, H. Audrain, J. Thorhauge,
Acc. Chem. Res. 1999, 32, 605Ϫ613. [1d] D. A. Evans, T. Rovis,
[1e]
[24]
[25]
[26]
J. S. Johnson, Pure Appl. Chem. 1999, 71, 1407Ϫ1415.
J. S.
Johnson, D. A. Evans, Acc. Chem. Res. 2000, 33, 325Ϫ335.
[2]
[3]
For recent examples, see: [2a] T. Rovis, D. A. Evans, Prog. Inorg.
[2b]
Chem. 2001, 50, 1Ϫ150.
J. M. Fraile, J. I. Garcia, M. A.
Harmer, C. I. Herrerias, J. A. Mayoral, J. Mol. Chem. A-Chem.
2001, 165, 211Ϫ218. [2c] D. Carmona, M. P. Lamata, L. A. Oro,
Coord. Chem. Rev. 2000, 200, 717Ϫ772.
For recent examples, see: [3a] A. K. Ghosh, M. Shirai, Tetrahed-
For examples of the use of similar dioxenones in follow-up
chemistry, see: S. Benetti, R. Romagnoli, C. Derisi, G. Spalluto,
V. Zanirato, Chem. Rev. 1995, 95, 1065Ϫ1114.
[27]
[3b]
ron Lett. 2001, 42, 6231Ϫ6233.
D. A. Evans, J. S. Johnson,
[3c]
E. J. Olhava, J. Am. Chem. Soc. 2000, 122, 1635Ϫ1649.
K.
Eur. J. Org. Chem. 2003, 317Ϫ324
323

F. P. J. T. Rutjes et al.
FULL PAPER
[28]
[31]
[32]
D. A. Evans, T. Rovis, M. C. Kozlowski, C. W. Downey, J. S.
J. M. Takacs, M. R. Jaber, A. S. Vellekoop, J. Org. Chem. 1998,
Tedrow, J. Am. Chem. Soc. 2000, 122, 9134Ϫ9142.
(S,S)-Dibenzoyl tartaric anhydride (DBTAAN, 0.5 mL, 15
63, 2742Ϫ2748.
Prepared according to literature:
[29]
[32a]
D. A. Evans, G. S.
µmol, 1.1 equiv., 50 m in CD3CN) was added in three por-
tions to a solution of the acid or amide (14 µmol) in a Na2B4O7
buffer (0.5 mL, 0.4  in D2O); after each portion the sample
was shaken vigorously for 5 min. The sample was left overnight
before the H NMR spectrum was measured.
R. M. Williams, J. A. Hendrix, J. Org. Chem. 1990, 55,
3723Ϫ3728.
Peterson, J. S. Johnson, D. M. Barnes, K. R. Campos, K. A.
Woerpel, J. Org. Chem. 1998, 63, 4541Ϫ4544. [32b] D. A. Evans,
R. Martin, G. A. Taylor, C. H. M. Yap, J. Chem. Soc., Perkin
Trans. 1 1983, 12, 1635Ϫ1640.
K. Krohn, G. Schafer, Liebigs Ann. 1996, 265Ϫ270.
Received August 5, 2002
1
[33]
[30]
[O02454]
324
Eur. J. Org. Chem. 2003, 317Ϫ324