The stereochemical course of addition of allyltrimethylsilane to protected L-alaninals and L-serinals in the presence of Lewis acids. Total synthesis of cis-(2R,3S)-3-hydroxyproline

Article abstract of DOI:10.1016/S0957-4166(02)00255-0

The influence of Lewis acids on the diastereoselectivity of the addition of allyltrimethylsilane to variously protected L-alaninals and L-serinals was investigated and high levels of asymmetric induction were achieved. Stereochemical models for rationalisation of the results obtained are proposed. cis-(2R,3S)-3-Hydroxyproline was synthesised starting from N-Cbz, O-TBS-L-serinal, in which the crucial step involves addition of allyltrimethylsilane to the aldehyde in the presence of SnCl₄.

Full text of DOI:10.1016/S0957-4166(02)00255-0

TETRAHEDRON:
ASYMMETRY
Pergamon
Tetrahedron: Asymmetry 13 (2002) 1103–1113
The stereochemical course of addition of allyltrimethylsilane to
protected -alaninals and -serinals in the presence of Lewis
acids. Total synthesis of cis-(2R,3S)-3-hydroxyproline†
L
L
D. Gryko,a P. Prokopowicza and J. Jurczaka,b,
*
aInstitute of Organic Chemistry, Polish Academy of Science, Kasprzaka 44/52, 01-224 Warsaw, Poland
bDepartment of Chemistry, Warsaw University, Pasteura 1, 02-093 Warsaw, Poland
Received 15 May 2002; accepted 22 May 2002
Abstract—The influence of Lewis acids on the diastereoselectivity of the addition of allyltrimethylsilane to variously protected
L
-alaninals and L-serinals was investigated and high levels of asymmetric induction were achieved. Stereochemical models for
rationalisation of the results obtained are proposed. cis-(2R,3S)-3-Hydroxyproline was synthesised starting from N-Cbz,O-TBS-L-
serinal, in which the crucial step involves addition of allyltrimethylsilane to the aldehyde in the presence of SnCl4. © 2002 Elsevier
Science Ltd. All rights reserved.
1. Introduction
interest in alleviating viral infections. The development
of improved synthetic methodologies for potential gly-
cosidase inhibitors is still a challenging field. Various
synthetic methodologies have been applied so far and
the subject of the total synthesis of amino sugars has
been covered in several general reviews.9–15 Very often,
the stereoselective C(3) elongation of the carbon skele-
ton is the crucial problem in the synthesis of amino
sugars from a-amino acids and their derivatives.16–18
We would like to propose a new pathway for the
synthesis of cis-(2R,3S)-3-hydroxyproline 1, a useful
building block. The retrosynthetic analysis, shown in
The amino acyl–proline cis/trans isomerisation is a rate
limiting step in protein folding1 and modulates the
biological activity of peptides.2 Therefore, proline plays
a particular role in protein structure and peptide con-
formation.3 cis-(2R,3S)-3-Hydroxyproline 1 is a struc-
tural unit present in some biologically important
compounds such as slaframine,4 castanospermine,5 and
detoxinine.6 They belong to the class of compounds
known as polyhydroxylated amino sugars, which
proved to be highly effective glycosidase inhibitors.7
Because glycoproteins mediate cell–cell recognition,8 it
is reasoned that inhibitors of glycosidases should be of
Scheme 1, suggests that N,O-protected-L-serinal and an
allyl reagent could serve as starting materials.19 Prelim-
Scheme 1.
* Corresponding author. E-mail: jurczak@icho.edu.pl
† Dedicated to Professor Marek Chmielewski on the occasion of his 60th birthday.
0957-4166/02/$ - see front matter © 2002 Elsevier Science Ltd. All rights reserved.
PII: S0957-4166(02)00255-0

1104
D. Gryko et al. / Tetrahedron: Asymmetry 13 (2002) 1103–1113
inary results20 concerning the diastereoselective addi-
tion of allyltrimethylsilane 2 to N,O-protected- -seri-
nals suggested that this process requires deeper studies
in order to explain the influence of the N-protecting
group and the O-protecting group and the Lewis acid
used as a promoter.
fact that an a-chelate cannot be formed and the addi-
tion occurs from the less-hindered side of the Felkin–
L
Anh model. We propose that in our case
a
seven-membered chelate is formed as a result of coordi-
nation of titanium by two oxygen atoms (Fig. 1). The
formation of a similar complex was proposed by Gar-
ner,24 who observed syn-diastereoselectivity in the case
of addition to N,N-diprotected derivative of serinal in
the presence of ZnCl2.
2. Results and discussion
Surprisingly, the addition of 2 to 3, in the presence of
SnCl4 (Table 1, entry 6), occurred with no diastereofa-
cial selectivity. Presumably because attack by the allyl
reagent occurs partially from the less hindered side of
the Felkin–Anh model and partially from the less hin-
dered side of the seven-membered chelate. The reaction
2.1. Addition of allyltrimethylsilane 2 to N-mono- and
N,N-diprotected-L-alaninals
With the aim of studying the influence of Lewis acids
and protecting groups on the addition of allyltrimethyl-
silane 2 to a-amino aldehydes, we decided to use the
of allyltrimethylsilane 2 with N-Bn,N-Cbz- -alaninal 3
L
derivatives of
Cbz- -Alaninal 3, N-Bn,N-Boc-
Bn,N-Ts- -alaninal 5 were selected as typical examples
of N,N-diprotected derivatives. Each of these alaninals
L
-alanine as the simplest case. N-Bn,N-
mediated by BF3·OEt2 was slow, but gave both high
yield and good anti-diastereoselectivity (12:88) (Table
1, entry 3). According to our expectations, we did not
obtain any allylated adduct derived from N-Bn,N-Boc-
L
L
-alaninal 4 and N-
L
was obtained from the respective L-alaninol, using the
TEMPO21 oxidation method, since it gives a-amino
aldehydes of high enantiomeric purity. Additions of
allyltrimethylsilane 2 were initially carried out in the
presence of Lewis acids such as TiCl4, SnCl4 (so-called
chelating Lewis acids) and BF3·OEt2, applying equimo-
lar acid and aldehyde (Scheme 2, Table 1).
L
-alaninal 4; this aldehyde is not stable enough under
the reaction conditions. Furthermore, we studied the
addition of 2 to N-Bn,N-Ts- -alaninal 5. In all reac-
L
tions, the anti-diastereoisomer 5b (Table 1, entries 2, 5
The reaction of N-Bn,N-Cbz-
L
-alaninal
3
with
allyltrimethylsilane 2 in the presence of TiCl4 gave the
syn-isomer 3a as the major product (Table 1, entry 1).
This result was unexpected because, in the case of
additions to N,N-diprotected
L-alaninals, anti-
diastereoselectivity22,23 is usually observed due to the
Figure 1.
Scheme 2.
Table 1. Addition of allyltrimethylsilane 2 to variously protected L-alaninals
Entry
Aldehyde
Lewis acid
Reaction time (h)
syn/anti ratio
Yield (%)
1
2
3
4
5
6
7
8
3
5
3
4
5
3
4
5
TiCl4
TiCl4
BF3·OEt2
BF3·OEt2
BF3·OEt2
SnCl4
25
22
48
18
24
2
84:16
32:68
12:88
–
7:93
50:50
–
62a
87
83b
0b
85a
96
SnCl4
SnCl4
18
1.5
0b
34:66
34b
a Unreacted starting material was recovered.
b By-products were formed.

D. Gryko et al. / Tetrahedron: Asymmetry 13 (2002) 1103–1113
1105
Table 2. The influence of the amount of TiCl4 used on
the addition of allyltrimethylsilane 2 to N-monoprotected
-alaninals
and 8) was formed as the major product and the best
result was obtained when BF3·OEt2 was used as a
promoter (entry 5). Depending on the N-protecting
group in the aldehyde and the promoter used, the anti-
or syn-adduct may be the major product of the reac-
tions of N,N-diprotected-L-alaninal and allyltrimethyl-
silane 2 catalysed by a Lewis acid.
L
Entry
Amount of
TiCl4
Aldehyde
syn/anti ratio Yield (%)
(equiv.)
1
2
3
4
5
6
7
8
0.50
0.75
1.00
1.25
1.50
2.00
0.50
0.75
1.00
1.25
1.50
2.00
6
6
6
6
6
6
7
7
7
7
7
7
60:40
58:42
58:42
55:45
52:48
52:48
50:50
–
40
99
99
92
86
71
39
0
Since the additions in the presence of TiCl4 did not give
the expected good results and there was some evidence
that this reaction depended on the amount of the Lewis
acid added, as reported by Kiyooka et al.,25 we consid-
ered it very interesting to study the influence of the
amount of TiCl4 used on the addition of allyltrimethyl-
silane 2 to N-mono- and N,N-diprotected-
L-alaninals.
9
50:50
–
–
41
0
0
For N-Cbz- -alaninal 6 and N-Ts- -alaninal 7, the
L
L
10
11
12
diastereoisomeric excess did not depend on the amount
of the Lewis acid applied and it ranged about zero; a
high yield was obtained when 1 equiv. of TiCl4 per 1
63:37
67
equiv. of N-Cbz- -alaninal 6 (Table 2, entry 3) and 2
L
equiv. for 1 equiv. N-Ts- -alaninal 7 were used (Table
L
Table 3. The influence of the amount of TiCl4 used on
the addition of allyltrimethylsilane 2 to N,N-diprotected
2, entry 12).
L
-alaninals
For N-Bn,N-Cbz- -alaninal 3, the dependence was
L
Entry
Amount of
TiCl4
(equiv.)
Aldehyde
syn/anti ratio Yield (%)
more evident, the highest diastereoisomeric ratio (90:10)
was obtained when 0.75 mmol of TiCl4 per 1 mmol of
aldehyde 3 was used (Table 3, entry 2). Increasing the
amount of Lewis acid added led to a decrease in the
syn-diastereoselection, but gave increased yields until
the ratio of TiCl4 reached 1.25 mmol per 1 mmol of
1
2
3
4
5
6
7
8
0.50
0.75
1.00
1.25
1.50
2.00
0.50
0.75
1.00
1.25
1.50
2.00
3
3
3
3
3
3
5
5
5
5
5
5
–
0
54
62
79
53
48
49
58
87
88
92
99
90:10
84:16
75:25
77:23
69:31
5:95
aldehyde 3 (entry 4). In the case of N-Bn,N-Ts-L-alan-
inal 5, the best result (5:95, 50% yield) was obtained
when 0.5 mmol of TiCl4 per 1 mmol of the aldehyde
was used (entry 7). The yield increases (reaching 99%)
with increasing amounts of TiCl4 (2 mmol per 1 mmol
of the aldehyde) (entry 12).
36:64
32:68
33:67
42:58
34:56
9
10
11
12
2.2. Addition of allyltrimethylsilane 2 to N-mono-
protected-O-protected-L-serinals
We have reported the preliminary results of the
diastereoselective addition of allyltrimethylsilane 2 to
N,O-Protected
of the respective alcohols using the TEMPO method.21
N-Cbz,O-TBS- -Serinal 8 and N-Bn,N-Cbz,O-TBS-
serinal 10 were obtained with virtually no racemisation.
Similar results were obtained for N-Cbz,O-BOM-
serinal 9 and N-Bn,N-Cbz,O-BOM- -serinal 11. Hav-
ing in hand the enantiomerically pure a-amino-
b-hydroxy aldehydes, we investigated the addition of
allyltrimethylsilane 2. The reaction of N-Cbz,O-
L-serinals were obtained via oxidation
N,O-protected-L-serinals mediated by TiCl4 and
SnCl4.20 It was observed that it is possible to control
the diastereoselectivity of this process by changing pro-
tecting groups as well as Lewis acids. Since C(3) elonga-
tion is the crucial step in the synthesis of
cis-(2R,3S)-3-hydroxyproline 1,19 we decided to study
L
L-
L
-
L
addition of allyltrimethylsilane 2 to N,O-protected-L-
serinals in detail (Scheme 3).
Scheme 3.

1106
D. Gryko et al. / Tetrahedron: Asymmetry 13 (2002) 1103–1113
TBS-
L
-serinal 8 with 2, irrespective of the Lewis acid
products 9a, 9b (entries 6–11). At room temperature,
the formation of by-products occurs (entries 7, 9 and
11), whereas lowered temperature results in quantitative
recovery of the substrate 9 (entries 6, 8 and 10).
used, gave the syn-diastereoisomer 8a as the major
product (Table 4).
This diastereoisomer can be obtained almost exclusively
when SnCl4 is applied as a mediator (Table 4, entry 1).
We did not observe any significant influence of the
amount of TiCl4 added on the diastereoselectivity of the
reaction of 2 with 8, which remained at the level of
8:2–7:3 (entries 2–4). Increasing the amount of TiCl4
caused an increase in the yield. In the case of less active
Lewis acids such as ZnBr2 or ZnCl2, the reaction tem-
perature had to be elevated in order to obtain the
desired allylated adducts (8a and 8b) but, as a conse-
quence, the decrease in the diastereoselectivity was
observed (entries 7 and 8). Reactions mediated by
MgBr2·OEt2 or AlEt3 did not lead to allylated adducts
(8a and 8b) at all (entries 9 and 10).
2.3. Addition of allyltrimethylsilane 2 to N,N-
diprotected-O-protected-L-serinals
The addition of allyltrimethylsilane 2 to N-Bn,N-CBz-
O-TBS- -serinal 10, mediated by TiCl4, gave the syn-
L
diastereoisomer 10a as the predominant product (Table
6, entries 2–4) and increasing the amount of TiCl4 gave
increases in both the diastereoselectivity and yield.
When the addition was carried out in the presence of
SnCl4 or BF3·OEt2, low to moderate anti-diastereose-
lectivity was observed (entries 1 and 6). At 4°C, weak
Lewis acids did not promote the reaction of 2 with 10
at all (entries 8, 10 and 12); at room temperature, only
low syn-diastereoselectivity and low yield were obtained
(entries 7, 9 and 11).
The addition of allyltrimethylsilane 2 to N-Cbz,O-
BOM- -serinal 9, in the presence of SnCl4, gave pre-
L
dominantly the syn-adduct 9a, but both the level of the
diastereoselectivity and yield decreased (Table 5, entry
1). Very interesting results were obtained when the
addition of 2 to 9 was carried out in the presence of
TiCl4 as a mediator (entries 2–4). Surprisingly, anti-
diastereoselectivity was observed regardless of the
amount of Lewis acid used. Reactions mediated by
ZnBr2, ZnCl2 and MgBr2·OEt2 failed to give the desired
N-Bn,N-CBz-O-BOM- -Serinal 11 reacted with
L
allyltrimethylsilane 2 to give the anti-adduct 11b as a
major product in the presence of all Lewis acids used
(Table 7). Usually, the diastereoselectivity was moder-
ate and the best result (18:82, 72% yield) was obtained
for the reaction carried out in the presence of TiCl4 (1.5
equiv.) (entry 4).
Table 4. Addition of allyltrimethylsilane 2 to N-Cbz-O-TBS-
L
-serinal 8
Temperature (°C)
Entry
Lewis acid
Reaction time (h)
syn/anti ratio
Yield (%)
1
2
3
4
5
6
7
8
SnCl4
1
22
20
20
20.75
2.2
26.25
25.75
23.5
15
−78
−78
−78
−78
−78
−48
20
20
20
20
98:2
80:20
70:30
72:28
58:42
79:21
85:15
80:20
–
88
40a
83
90
33
52
63
51
0b
TiCl4, 0.5 equiv.
TiCl4, 1.0 equiv.
TiCl4, 1.5 equiv.
BF3·OEt2
BF3·OEt2
ZnBr2
ZnCl2
MgBr2·OEt2
AlEt3
9
10
–
0a
a Unreacted starting material was recovered.
b By-products were formed.
Table 5. Addition of allyltrimethylsilane 2 to N-Cbz-O-BOM-
L
-serinal 9
Entry
Lewis acid
Reaction time (h)
Temperature (°C)
syn/anti ratio
Yield (%)
1
2
3
4
5
6
7
8
SnCl4
1.5
19
16
16
23
20
20
20
40
40
40
−78
−78
−78
−78
−78
4
20
4
20
4
82:18
2:98
3:97
2:98
89:11
–
–
–
–
–
68
17a
90
82b
65
0a,b
0b
TiCl4, 0.5 equiv.
TiCl4, 1.0 equiv.
TiCl4, 1.5 equiv.
BF3·OEt2
ZnCl2
ZnCl2
ZnBr2
ZnBr2
MgBr2·OEt2
MgBr2·OEt2
0a
9
10
11
0b
0a
20
–
0a
a Unreacted starting material was recovered.
b By-products were formed.

D. Gryko et al. / Tetrahedron: Asymmetry 13 (2002) 1103–1113
Table 6. Addition of allyltrimethylsilane 2 to N-Bn-N-Cbz-O-TBS- -serinal 10
1107
L
Entry
Lewis acid
Reaction time (h)
Temperature (°C)
syn/anti ratio
Yield (%)
1
2
3
4
5
6
7
8
SnCl4
1.5
20
3
3
19
5
15
5
15
5
15
5
−78
−78
−78
−78
−78
−20
20
4
20
4
20
45:55
90:10
94:6
97:3
–
25:75
54:46
–
52:48
–
60:40
–
75
38a
70
79
0a
TiCl4, 0.5 equiv.
TiCl4, 1.0 equiv.
TiCl4, 1.5 equiv.
BF3·OEt2
BF3·OEt2
ZnCl2
ZnCl2
ZnBr2
ZnBr2
MgBr2·OEt2
MgBr2·OEt2
72
33b
0a
9
34b
0a
10
11
12
10b
0a
4
a Unreacted starting material was recovered.
b By-products were formed.
Table 7. Addition of allyltrimethylsilane 2 to N-Bn-N-Cbz-O-TBS-L-serinal 11
Entry
Lewis acid
Reaction time (h)
Temperature (°C)
syn/anti ratio
Yield (%)
1
2
3
4
5
6
7
8
9
SnCl4
3
20
3.5
2.5
15
21
36
36
36
−78
−78
−78
−78
−78
−20
20
21:79
27:73
26:74
18:82
–
27:73
24:76
44:55
73
30a
59
72
0a
TiCl4, 0.5 equiv.
TiCl4, 1.0 equiv.
TiCl4, 1.5 equiv.
BF3·OEt2
BF3·OEt2
ZnCl2
25a
23a
6a
ZnCl2
MgBr2·OEt2
20
20
0a
a Unreacted starting material was recovered.
taken for the major product from reaction of N-Cbz,O-
BOM- -serinal 9 with allyltrimethylsilane 2 mediated
by SnCl4. Therefore syn-configuration was assigned to
compound 9a and, consequently, anti-configuration
was assigned to adduct 9b.
2.4. Chemical correlation
L
Having determined the extent of asymmetric induction,
we studied the sense of induction by establishing the
configuration of the products using the chemical corre-
lation method. Adduct 8a, obtained in the reaction of
N-Cbz,O-TBS-
L
-serinal 8 with 2 in the presence of
Hydrogenation of syn-N-Cbz,O-TBS adduct 8a as well
as the N-Bn,N-Cbz,O-TBS-derivative 10a, being the
major product of the addition reaction performed in
the presence of TiCl4 (Table 6, entry 4), gave the same
N-deprotected compound 13 (Scheme 5), as determined
SnCl4 (Table 4, entry 1), was transformed into the
known cis-(2R,3S)-3-hydroxyproline 119 which has the
hydroxy and the hydroxymethyl groups in relation cis
(Scheme 7). Therefore, compound 8a bears the amino
and the hydroxy groups unequivocally in the syn-orien-
tation, this means that the C(3) stereogenic centre has
(S)-absolute configuration. The other diastereoisomer
8b possesses the opposite (R)-configuration.
1
by H and 13C NMR spectra. This indicates that N-
Bn,N-Cbz,O-TBS derivative 10a has the amino and
hydroxy groups in the syn-orientation.
Hydrogenation of syn-N-Cbz,O-BOM-adduct 9a gave
fully deprotected dihydroxyamine, which was immedi-
ately (without purification) subjected to reaction with
acetic anhydride, giving the triacetylated syn-product
14a. As a result of hydrogenation of the syn/anti 11a/
11b mixture (18:82) (Table 7, entry 4), in the presence
In order to assign the stereochemical course of the allyl
addition to N-Cbz,O-BOM- -serinal 9, the TBS group
L
in syn-8a was replaced by the BOM group giving
1
adduct 9a (Scheme 4). The H and 13C NMR spectra
taken for 9a obtained from 8a were identical to those
Scheme 4.

1108
D. Gryko et al. / Tetrahedron: Asymmetry 13 (2002) 1103–1113
bonyl groups—model B (Fig. 2), which resembles
model A. Therefore, the allyl reagent approaches from
the same side as in model A, giving predominantly
syn-8a. Similar effects16 were observed in the hetero-
Diels–Alder reaction of a-amino aldehydes. Since we
did not observe any anti-diastereoselectivity, the forma-
tion of b-chelate (model C) between the hydroxy group
(protected with the TBS group) and the carbonyl group
is presumably excluded due to the steric hindrance.
N-Cbz,O-BOM- -Serinal 9 can form a-chelates, (A) or
L
(B), and b-chelates (C) (Fig. 2) since the BOM group
does not influence strongly the coordination by the
ether oxygen atom. In the presence of SnCl4 and
BF3·OEt2 the addition of allyltrimethylsilane 2 to alde-
hyde 9 afforded syn-adduct 9a as a major product. In
both cases, the a-chelates should be formed because, for
SnCl4, the reaction proceeds via a-chelate (A). Since
BF3·OEt2 cannot form cyclic chelates, model B should
operate in this case. Although TiCl4 is regarded as a
strong Lewis acid, the reaction mediated by TiCl4 gave
predominantly the anti-product 9b. It is highly unlikely
that this reaction proceeds via the Felkin–Anh model,
therefore we assume that TiCl4 and aldehyde 9 form a
six-membered b-chelate (C) and attack by the allyl
reagent occurs from the side opposite of the amino
group leading to anti-diastereoisomer 9b.
Scheme 5.
of the Degussa catalyst, the benzyl and carbobenzoxy
groups were cleaved. Then the free amino and hydroxy
groups were protected by treating with acetic anhy-
dride. Further replacement of the BOM group with the
acetyl group resulted mainly in the formation of the
compound 14b. Since the H and 13C NMR spectra of
1
compounds 14a and 14b differ from each other, but
relate to compounds of the same structure, we conclude
that they have the opposite relative configurations.
Thus, in the addition reaction of allyltrimethylsilane 2
to N-Bn,N-Cbz,O-BOM- -serinal 11, the anti-
L
diastereoisomer 11b was formed as the major product
(Scheme 6).
The most interesting result was observed for the addi-
tion of 2 to N-Bn,N-Cbz,O-TBS- -serinal 10 when the
2.5. Stereochemical course of the additions
L
syn-adduct 10a was mainly formed in the presence of
TiCl4. Since the formation of a-chelate is excluded by
the presence of steric hindrance, there should be some
other interactions resulting in another stable conform-
ation. We propose that titanium is coordinated by two
carbonyl oxygen atoms (CHO and COOBn) forming a
seven-membered chelate—model D (Fig. 2) and, there-
fore, attack of the allyl reagent from the less hindered
side leads to the syn-product 10a. A similar chelate was
first proposed by Kunz,30 who proved the formation of
syn-Diastereoselectivity was observed for all additions
of 2 to N-Cbz,O-TBS- -serinal 8, irrespective of the
L
Lewis acid used. In the case of the chelating Lewis
acids,26–28 the direction of asymmetric induction can be
explained by the transition state A, as shown in Fig. 2.
Attack by the allyl reagent occurs from the less-hin-
dered side of the cyclic model A. Surprisingly, the
addition in the presence of non-chelating acids such as
26,27,29
BF3·OEt2
also resulted in syn-adduct 8a, as a
consequence of H-bonding between the NH and car-
Scheme 6.
Figure 2.

D. Gryko et al. / Tetrahedron: Asymmetry 13 (2002) 1103–1113
1109
such a complex by 13C NMR. Later, Garner24 assumed
that in the presence of ZnCl2 the cycloaddition to
carbon C(2). Treatment of 17 with NaBH3CN, followed
by selective deprotection of the primary hydroxy group,
afforded the pyrrolidine derivative 19 in 91% yield.
Oxidation of the hydroxy group in 19 with NaIO4 in
the presence of RuCl3 in a biphasic system33 resulted in
the protected amino acid 20. The silyl protective group
diprotected
L-serinal proceeds via such a seven-mem-
bered a-chelate.
In all cases of the addition to N-Bn,N-Cbz,O-BOM-L-
34
serinal 11, the major product results from the b-chela-
tion-controlled reaction (model E) that gives mainly the
1,2-anti-isomer 11a, showing that the O-BOM oxygen
atom is able to coordinate to Lewis acids such as SnCl4,
TiCl4, ZnCl2 and ZnBr2. Of course, we are aware of the
fact that attack of the allyl reagent from the less-hin-
dered side of the Felkin–Anh model gives the same
result as represented by the reaction mediated by
BF3·OEt2.
was removed using H2SiF6 to provide the correspond-
ing derivative 21 in 81% yield. Deprotection of 21 by
hydrogenolysis (H2, Pd/C) then gave the desired 1. For
independent verification of the cis-correlation of the
substituents, the free cis-(2R,3S)-3-hydroxyproline was
transformed into the corresponding hydrochloride 22.
1
Compound 22 had H and 13C NMR spectra consistent
with those reported in the literature.34
2.6. Total synthesis of cis-(2R,3S)-3-hydroxyproline, 1
3. Conclusions
Having in hand a very good method for the synthesis of
The diastereoselective synthesis of cis-(2R,3S)-3-
hydroxyproline 1 proceeds in 52% yield over eight
steps, starting from the suitably protected L-serinal 8.
diastereoisomerically pure syn-adduct 8a, we pursued
the synthesis of cis-(2R,3S)-3-hydroxyproline
1
(Scheme 7). Protection of the hydroxy group of the
syn-adduct 8a using triisopropylsilyl triflate, followed
by syn-dihydroxylation with NMO and OsO431 gave the
polyhydroxy amine 16, which, without purification, was
subjected to oxidative cleavage using sodium
periodate32 to afford the aldehyde 17 in 81% yield.
Analysis of 1H and 13C NMR data proved that 17
This synthetic strategy proves the utility of C(3)-elonga-
tion of a-amino aldehydes. During the course of our
studies we found that the addition of allyltrimethylsi-
lane 2 to a-amino aldehydes strongly depends on the
Lewis acid used and the N,O-protecting groups present
in the aldehyde. We have observed that when TiCl4 is
used as a mediator for the reaction of allyltrimethylsi-
lane 2 and a-amino aldehydes preferentially forms the
seven-membered chelates. Furthermore, we are able to
1
adopts the cyclic conformation 17a. In H NMR spec-
tra of 17, there are no signals for the CH6 O and NH6
protons, instead a broad signal characteristic of an OH
6
obtain the anti- and syn-adducts, derived from
L
-alan-
proton appears at 6.45–6.56 ppm as well as a doublet of
inal and
purity.
L-serinal, in good yields and enantiomeric
doublets at 5.36 ppm, which is assigned to the amine
Scheme 7.

1110
D. Gryko et al. / Tetrahedron: Asymmetry 13 (2002) 1103–1113
4. Experimental
4.2.3.
(1S,2S)-[1-(Benzyloxymethyl)-2-hydroxypent-4-
enyl]carbamic acid benzyl ester, 9a. 1H NMR (500
MHz, 363 K, toluene-d8) 2.16 (t, J=7.1, 2H), 3.54
(ABM/2, JAB=9.9 Hz, JBM=4.9 Hz, 1H), 3.59 (ABM/
2, JAB=10.3 Hz, JAM=5.9 Hz, 1H), 3.72 (dt, Jt=6.6
Hz, Jd=2.1 Hz, 1H), 3.8–3.9 (m, 1H), 4.41 (s, 2H), 4.49
(s. 2H), 4.9–5.0 (m, 2H), 5.05 (s, 2H), 5.11 (d, J=8.1
Hz, 1H), 5.74 (ddt, Jt=7.2 Hz, Jd=17.3, 10.2 Hz, 1H),
7.0–7.3 (m, 10H); 13C NMR (100 MHz, 363 K, toluene-
d8) 38.8, 54.1, 66.7, 69.3, 70.0, 70.9, 95.2, 117.2, 128.0,
128.1, 128.4, 128.6, 128.8, 137.3; LSIMSHR
C22H28NO5 (M+H)+ calcd 386.1967, found 386.1979;
[h]2D0 −1.5 (c 0.85, CHCl3).
4.1. General
All chemicals were used as received unless otherwise
noted. Reagent grade solvents (CHCl3, CH2Cl2, hex-
anes, AcOEt) were distilled prior use. All reported
NMR spectra were recorded with a Bruker spectrome-
ter at 500 (1H NMR) and 125 (13C NMR) MHz or a
Bruker spectrometer at 400 (1H NMR) and 100 (13C
NMR) MHz or a Varian Gemini spectrometer at 200
(1H NMR) and 50 (13C NMR) MHz. Chemical shifts
are reported as l values relative to TMS peak defined
at l=0.00 (1H NMR) or l=0.0 (13C NMR). IR spectra
were obtained on a Perkin–Elmer 1640 FTIR. Mass
spectra were obtained on an AMD-604 Intectra instru-
ment using the EI or LSIMS technique. Chromatogra-
phy was performed on silica (Kiesel gel 60, 200–400
mesh). Optical rotations were recorded using a JASCO
DIP-360 polarimeter with a thermally jacketed 10 cm
cell.
4.2.4.
(1S,2R)-[1-(Benzyloxymethyl)-2-hydroxypent-4-
enyl]carbamic acid benzyl ester, 9b. 1H NMR (400
MHz, 363 K, toluene-d8) 2.0–2.1 (m, 2H), 3.5–3.7 (m,
4H), 4.32 (s, 2H), 4.39 (s, 2H), 4.8(5)–4.9 (m, 3H), 4.95
(s, 2H), 5.66 (ddt, Jt=6.9 Hz, Jd=17.2, 10.3 Hz, 1H),
6.9–7.1(5) (m, 10H); 13C NMR (100 MHz, 363 K,
toluene-d8) 38.9, 55.2, 66.7, 67.6, 69.9, 71.8, 95.2, 117.2,
127.8, 128.0, 128.1, 128.8, 128.6, 128.8, 137.3;
LSIMSHR C22H28NO5 (M+H)+ calcd 386.1967, found
386.1964; the diastereoisomeric purity of this com-
pound was 90%.
4.2. Addition of allyltrimethylsilane 2 to a-amino
aldehydes, general procedure
The a-amino aldehyde (0.23 mmol) was dissolved in dry
CH2Cl2 (2 mL) under argon and cooled to −78°C, then
the Lewis acid (0.23 mmol) was added. After 5 min,
allyltrimethylsilane 2 (0.46 mmol) was added and the
reaction mixture was stirred for another 1.5 h at −78°C.
The mixture was diluted with saturated aqueous NH4F
and extracted with ether. The organic phase was dried
over MgSO4, rotary evaporated, and chromatographed
(silica, hexane/ethyl acetate).
4.2.5. (1S,2S)-Benzyl-[1-(tert-butyldimethylsilanyloxy-
methyl)-2-hydroxypent-4-enyl]carbamic acid benzyl ester,
1
10a. H NMR (400 MHz, 363 K, toluene-d8) −0.10 (s,
6H), 0.81 (s, 9H), 2.0–2.1 (m, 2H), 3.6–3.7 (m, 1H),
3.8–3.9 (m, 3H), 4.48 (d, J=15.6 Hz, 1H), 4.56 (d,
J=15.6 Hz, 1H), 4.86 (s, 1H), 4.9 (m, 2H), 4.97 (s, 1H),
4.98 (s, 1H), 5.7–5.8 (m, 1H), 6.9–7.2 (m, 10H); 13C
NMR (100 MHz, 363 K, toluene-d8) −5.7, 18.2, 25.9,
39.6, 52.5, 62.8, 65.0, 67.5, 70.7, 116.5, 127.1, 127.4,
127.9, 128.2, 128.4, 128.6, 128.8, 135.4; IR (film) 698,
777, 837, 1104, 1253, 1471, 1679, 1698, 2856, 2929,
3435; LSIMSHR C27H40NO4Si (M+H)+ calcd 470.2727,
found 470.2725; [h]2D0 +13.9 (c 1.6, CHCl3).
4.2.1. (1S,2S)-[1-(tert-Butyldimethylsilanyloxymethyl)-2-
hydroxypent-4-enyl]carbamic acid benzyl ester, 8a. 1H
NMR (400 MHz, 363 K, toluene-d8) −0.08 (s, 6H), 0.80
(s, 9H), 2.0–2.1 (m, 3H), 3.5–3.6 (m, 1H), 3.6–3.7 (m,
2H), 3.73 (dt, Jt=5.3 Hz, Jd=3.3 Hz, 1H), 4.8–5.0 (m,
5H), 5.58 (ddt, Jt=6.5 Hz, Jd=17.9 Hz, 1H), 6.9–7.2
(m, 5H); 13C NMR (100 MHz, 363 K, toluene-d8) −5.7,
25.8, 38.9, 55.2, 64.9, 66.6, 71.1, 117.2, 128.4, 128.6,
128.8, 137.3; IR (film) 697, 778, 837, 1028, 1099, 1216,
1258, 1471, 1509, 1703, 2856, 2929, 3068, 3437;
LSIMSHR C20H34NO4Si (M+H)+ calcd 380.2257,
found 380.2256; [h]2D0 +20.5 (c 1.47, CHCl3).
4.2.6. (1S,2R)-Benzyl-[1-(tert-butyldimethylsilanyloxy-
methyl)-2-hydroxypent-4-enyl]carbamic acid benzyl ester,
1
10b. H NMR (400 MHz, 363 K, toluene-d8) −0.07 (s,
3H), −0.07 (s, 3H), 0.81 (s, 1H), 0.83 (s, 9H), 2.0–2.1
(m, 2H), 3.5–3.6 (m, 1H), 3.9–4.1 (m, 3H), 4.35 (d,
J=16.8 Hz, 1H), 4.67 (d, J=16.8 Hz, 1H), 4.8–4.9 (m,
2H), 4.99 (s, 2H), 5.54 (ddt, Jt=6.3 Hz, Jd=16.5, 9.8
Hz, 1H), 6.9–7.1 (m, 10H); 13C NMR (100 MHz, 363
K, toluene-d8) −5.7, 18.2, 25.9, 40.0, 67.4, 72.6, 116.7,
127.4, 127.7, 127.9, 128.2, 128.4, 128.6, 135.2, 136.3; IR
(film) 699, 777, 837, 1005, 1101, 1253, 1471, 1679, 1698,
2929, 3447; LSIMSHR C27H40NO4Si (M+H)+ calcd
470.2726, found 470.2725; [h]2D0 −10.8 (c 0.9, CHCl3).
4.2.2. (1S,2R)-[1-(tert-Butyldimethylsilanyloxymethyl)-
1
2-hydroxypent-4-enyl]carbamic acid benzyl ester, 8b. H
NMR (500 MHz, 363 K, toluene-d8) −0.03 (s, 3H),
−0.03 (s, 3H), 0.85 (s, 9H), 2.1–2.2 (m, 3H), 3.5–3.6 (m,
2H), 3.67 (ABM/2, JAB=10.3 Hz, JAM=3.9 Hz, 1H),
3.79 (ABM/2, JAB=10.3 Hz, JBM=3.5 Hz, 1H), 4.9–5.0
(m, 2H), 5.03 (s, 2H), 5.04 (m, 1H), 5.76 (ddt, Jt=7.0
Hz, Jd=17.2, 10.3 Hz, 1H), 7.0–7.2 (m, 5H); 13C NMR
(100 MHz, 363 K, toluene-d8) −5.4, 18.5, 26.1, 39.5,
56.3, 63.2, 67.0, 72.3, 117.5, 128.5, 128.7, 128.9, 137.6,
156.4; IR (film) 697, 779, 837, 1045, 1084, 1216, 1255,
1470, 1510, 1705, 2857, 2930, 3440; LSIMSHR
C20H34NO4Si (M+H)+ calcd 380.2257, found 380.2248;
[h]2D0 +27.3 (c 1.56, CHCl3).
4.2.7. (1S,2R)-Benzyl-[1-(benzyloxymethyl)-2-hydroxy-
1
pent-4-enyl]-carbamic acid benzyl ester, 11b. H NMR
(500 MHz, 363 K, toluene-d8) 2.14 (s, 2H), 3.6–3.7 (m,
1H), 3.89 (s, 1H), 3.97 (dd, J=10.3, 4.0 Hz, 1H), 4.06
(s, 1H), 4.38 (d, J=5.1 Hz, 1H), 4.42 (d, J=4.5 Hz,
2H), 4.44 (m, 1H), 4.51 (dd, J=19.8, 6.3 Hz, 2H), 4.69
(d, J=15.4 Hz, 1H), 4.8–4.9 (m, 2H), 5.06 (s, 2H), 5.58
(ddt, Jt=7.0 Hz, Jd=17.2, 10.3 Hz, 1H), 7.0–7.2(5) (m,

D. Gryko et al. / Tetrahedron: Asymmetry 13 (2002) 1103–1113
1111
15H); 13C NMR (100 MHz, 363 K, toluene-d8) 40.2,
66.6, 67.8, 69.9, 73.1, 95.3, 117.3, 127.5, 127.8, 128.1,
128.2, 128.3, 128.5, 128.6, 128.7, 128.8, 128.9, 129.1,
135.5, 137.3, 138.9, 157.3; IR (film) 698, 736, 1048,
1114, 1239, 1419, 1454, 1678, 1696, 2887, 2938, 3031,
3065, 3451; LSIMSHR C29H34NO5 (M+H)+ calcd
476.2437, found 476.2436; [h]2D0 −10.1 (c 1.76, CHCl3).
Diastereoisomer 11a was not isolated in pure form.
hydrogenation in the presence of Degussa catalyst (50%
of H2O) was dissolved in CH2Cl2 and treated with an
excess of Ac2O and NEt3 and a catalytic amount of
DMAP. The reaction mixture was stirred for 2 h. After
standard work-up the reaction mixture was again
hydrogenated in the presence of 10% Pd/C followed by
reaction with acetic anhydride. The reaction mixture
was stirred for 2 h at rt and then extracted with
CH2Cl2. Column chromatography (silica, hexanes/
1
AcOEt) gave 14b as the main product. H NMR (200
4.2.8. (1S,2S)-(2-Hydroxy-1-hydroxymethylpent-4-enyl)-
carbamic acid benzyl ester, 12. To a solution of adduct
8a (646 mg, 1.70 mmol) in THF (7 mL) was added
Bu4NF·3H2O (1.02 g, 3.23 mmol). After 1.5 h, the
reaction mixture was diluted with saturated aqueous
NaHCO3 and extracted with ether. The chromatogra-
phy (silica, hexane/ethyl acetate, gradually from 8:2 to
1:1) gave 354 mg (82%) of the desired product. 1H
NMR (500 MHz, 363 K, toluene-d8) 2.1–2.1(6) (m, 2H),
3.51 (ABM/2, JAB=10.9 Hz, JBM=4.6 Hz, 1H), 3.56
(ABM/2, JAB=10.9 Hz, JAM=5.2 Hz, 1H), 3.6–3.6(4)
(m, 1H), 3.67 (dt, Jt=6.8 Hz, Jd=2.1 Hz, 1H), 4.9(6)–
5.0 (m, 2H), 5.04 (s, 2H), 5.31 (d, J=6.0 Hz, 1H), 5.70
(ddt, Jt=7.1 Hz, Jd=17.0, 11.5 Hz, 1H), 7.0–7.2(5) (m,
5H); 13C NMR (125 MHz, D2O) 39.2, 55.8, 64.7, 67.2,
71.7, 117.7, 127.8, 128.0, 128.2, 128.7, 128.9, 129.1,
134.9, 157.2; IR (film) 698, 1028, 1066, 1252, 1340,
1527, 1698, 2941, 3401 cm−1; LSIMSHR C14H20NO4
(M+H)+ calcd 266.1392, found 266.1392; [h]2D0 +3.0 (c
2.9, CHCl3).
MHz, 273 K, CDCl3) 0.92 (t, J=7.2 Hz, 3H), 1.2–1.7
(m, 4H), 2.02 (s, 3H), 2.08 (s, 3H), 2.09 (s, 3H), 4.08
(ABM/2, JAB=11.4 Hz, JAM=3.9 Hz, 1H), 4.27 (ABM/
2, JAB=11.4 Hz, JBM=5.9 Hz, 1H), 4.3–4.5 (m, 1H),
4.94 (dt, Jt=5.3 Hz, Jd=7.7 Hz, 1H), 5.91 (d, J=8.5
Hz, 1H); 13C NMR (125 MHz, 273 K, CDCl3) 13.9,
16.7, 20.9, 21.1, 23.5, 33.7, 50.6, 62.7, 73.8; LSIMSHR
C12H22NO5 (M+H)+ calcd 260.1498, found 260.1499.
4.2.12. (1S,2S)-[1-(tert-Butyldimethylsilanyloxymethy)-
2-triisopropylsilanyloxypent-4-enyl]carbamic acid benzyl
ester, 15. To the solution of adduct 8a (3.3 g, 8.7 mmol)
in CH2Cl2 (20 mL) at 0°C under argon, 2,6-lutidine (2.1
mL, 18.1 mmol) and triisopropylsilyl trifluoromethane-
sulfonate (3.1 mL, 12 mmol) were added. After stirring
for 18 h at 20°C, the reaction mixture was diluted with
water and extracted with ether to give an oil (4.7 g),
which was used without purification for the next reac-
tion. For analytical purposes, a sample of the crude
reaction mixture was chromatographed (silica, hexane/
1
4.2.9.
(2S,3S)-[2-Amino-1-(tert-butyldimethylsilanyl-
ethyl acetate, 9:1) H NMR (500 MHz, 363 K, toluene-
oxy)hex-5-en]-3-ol, 13. Hydrogenation of the adduct 8a
or 10a in the presence of Pd/C (10%) gave the title
compound. 1H NMR (500 MHz, 363 K, toluene-d8)
0.04 (s, 6H), 0.87 (s, 9H), 0.91 (t, J=7.0 Hz, 3H),
1.3–1.5 (m, 5H), 2.14 (s, 2H), 2.66 (dt, Jt=4.2 Hz,
Jd=5.9 Hz, 1H), 3.48 (dt, Jt=4.2 Hz, Jd=8.3 Hz, 1H),
3.58 (ABM/2, JAB=10.0 Hz, JAM=5.9 Hz, 1H), 3.65
(ABM/2, JAB=10.0 Hz, JBM=4.2 Hz, 1H); 13C NMR
(125 MHz, 273 K, CDCl3) −5.6, −5.5, 14.0, 18.1, 18.9,
25.8, 36.5, 56.0, 66.4, 71.3; IR (film) 777, 837, 1097,
1256, 1472, 1583, 2858, 2956, 3360 cm−1; LSIMSHR
C12H30NO2Si (M+H)+ calcd 248.2046, found 248.2044;
[h]2D0 −15.9 (c 1.28, CHCl3).
d8) 0.07 (s, 3H), 0.09 (s, 3H), 0.93 (s, 9H), 1.08 (s, 18H),
1.09 (s, 3H), 2.3–2.4 (m, 1H), 2.4–2.5 (m, 1H), 3.69 (t,
J=9.2 Hz, 1H), 3.7–3.8 (m, 1H), 3.9–4.1 (m, 1H),
4.3–4.4 (m, 1H), 5.0–5.1 (m, 5H), 5.7–5.8 (m, 1H),
7.0–7.4 (m, 5H); 13C NMR (125 MHz, 363 K, toluene-
d8) −5.2, −5.1, 12.8, 13.4, 18.3, 18.4, 18.5, 26.1, 26.2,
39.7, 55.4, 62.6, 66.9, 70.9, 118.0, 125.5, 128.3, 128.4,
128.6, 134.4; IR (film) 680, 777, 839, 1028, 1107, 1211,
1251, 1463, 1497, 1728, 2867, 2947, 3447; LSIMSHR
C29H54NO4Si2 (M+H)+ calcd 536.3591, found 536.3584;
[h]2D0 +5.6 (c 1.7, CHCl3).
4.2.13. (1S,2S,4S,R)-[1-(tert-Butyldimethylsilanyloxy-
methyl) - 4,5 - dihydroxy - 2 - triisopropylsilanyloxypentyl]-
carbamic acid benzyl ester, 16. Compound 15 (4.7 g, 8.7
mmol) was dissolved in a mixture of acetone (50 mL)
and water (7 mL) then treated sequentially with 4-
methylmorpholine N-oxide (1.66 g, 14.2 mmol) and a
solution of OsO4 in tert-butanol (6.1 mL, 0.05 M). The
reaction mixture was stirred overnight at room temper-
ature. Saturated aqueous KHSO3 was added and, after
10 min, the reaction mixture was extracted with ethyl
acetate. Chromatography (silica, hexane/ethyl acetate,
gradually from 9:1 to 6:4) gave 16 as a mixture of
4.2.10. (2S,3S)-1,3-Diacetoxy-2-acetamidohexane, 14a.
Compound 9a after hydrogenation in the presence of
Pd/C (10%) was dissolved in CH2Cl2, and treated with
an excess of Ac2O and NEt3 and a catalytic amount of
DMAP. The reaction mixture was stirred for 2 h.
Extraction followed by chromatography (silica, hex-
anes/AcOEt) gave 14a as a colourless oil. 1H NMR
(200 MHz, 273 K, CDCl3) 0.9–1.0 (m, 3H), 1.2–1.6 (m,
4H), 2.05 (s, 3H), 2.07 (s, 3H), 2.13 (s, 3H), 4.2–4.3 (m,
2H), 5.0–5.3 (m, 2H), 5.68 (d, J=8.5 Hz, 1H); 13C
NMR (50 MHz, 273 K, CDCl3) 13.9, 18.4, 18.5, 20.8,
20.9, 33.7, 53.1, 61.3, 72.1; LSIMSHR C12H22NO5 (M+
H)+ calcd 260.1498, found 260.1500.
1
two epimers (4 g, 81%). H NMR (500 MHz, 363 K,
toluene-d8) 0.06 (s, 3H), 0.07 (s, 3H), 0.08 (s, 3H),
0.10 (s, 3H), 0.92 (s, 9H), 0.95 (s, 9H), 1.08 (s, 18H),
1.08 (s, 3H), 1.09 (s, 18H), 1.09 (s, 3H), 1.6–1.7 (m,
2H), 1.80 (dt, Jt=8.3 Hz, Jd=14.4 Hz, 1H), 2.10 (brs,
1H), 2.53 (brs, 1H), 3.13 (dd, J=10.7, 7.0 Hz,
4.2.11. (2S,3R)-1,3-Diacetoxy-2-acetamidohexane, 14b.
A mixture of compounds 11a and 11b (18:82) following

1112
D. Gryko et al. / Tetrahedron: Asymmetry 13 (2002) 1103–1113
1H), 3.24 (dd, J=10.7, 6.6 Hz, 1H), 3.28 (dd, J=10.8,
3.6 Hz, 1H), 3.38 (dd, J=10.8, 3.9 Hz, 1H), 3.7–3.8 (m,
8H), 4.0–4.1(5) (m, 2H), 4.39 (ddd, J=8.2, 4.8, 2.2 Hz,
1H), 4.60 (ddd, J=9.3, 4.5, 1.5 Hz, 1H), 5.0–5.1 (m,
6H), 5.24 (d, J=9.0 Hz, 1H), 7.0–7.1 (m, 10H); 13C
NMR (125 MHz, 363 K, toluene-d8) −5.2, −5.2, −5.1,
−5.1, 13.5, 18.5, 18.5, 26.2, 26.2, 38.3, 39.0, 55.7, 57.0,
62.7, 63.2, 67.1, 67.2, 67.5, 67.5, 69.0, 69.5, 69.9, 70.4,
128.2, 128.3, 128.5, 128.8, 128.9, 129.1, 137.7;
LSIMSHR C29H56NO6Si2 (M+H)+ calcd 570.3646,
found 570.3653.
4.2.16. (2S,3S)-[2-(Hydroxymethyl)-3-triisopropylsilanyl-
oxy]pyrrolidine-1-carboxylic acid benzyl ester 19. Com-
pound 18 (3.8 g) was dissolved in a mixture of concen-
trated acetic acid (10 mL) and methanol (30 mL), and
the resulting solution was heated under reflux for 1 h,
then cooled to room temperature, diluted with water,
and neutralised with solid NaHCO3. Extraction with
CH2Cl2 followed by chromatography (silica, hexane/
ethyl acetate, 9:1) gave a colourless oil (2.6 g, 91%
1
starting from 16). H NMR (500 MHz, 363 K, toluene-
d8) 0.9–1.0 (m, 21H), 1.6–1.8 (m, 2H), 3.1–3.2 (m, 1H),
3.3–3.4 (m, 1H), 3.8–3.8(6) (m, 1H), 3.89 (m, 1H), 4.02
(s, 1H), 4.21 (dd, J=7.3, 6.9 Hz, 1H), 5.02 (AB/2,
J=12.4 Hz, 1H), 5.10 (AB/2, J=12.4 Hz, 1H), 6.9(5)–
7.2(5) (m, 5H); 13C NMR (125 MHz, 363 K, toluene-d8)
12.8, 18.2, 33.4, 44.1, 62.9, 67.3, 73.5, 128.7, 128.8,
129.1, 137.6; IR (film) 684, 882, 999, 1088, 1131, 1359,
1418, 1703, 2866, 2944, 3446 cm−1; LSIMSHR
C22H38NO4Si (M+H)+ calcd 408.2570, found 408.2566.
4.2.14. (1S,2S)-[1-(tert-Butyldimethylsilanyloxymethyl)-
4-oxo-2-triisopropylsilanyloxybutyl]carbamic acid benzyl
ester, 17. To a magnetically stirred suspension of silica
gel (14 g) in CH2Cl2 (112 mL) was added a solution of
NaIO4 (1.95 g, 9.1 mmol) in water (14 mL) dropwise,
followed by a solution of diol 16 (4.0 g, 7.0 mmol) in
CH2Cl2 (17 mL) and the resulting mixture was stirred
overnight at room temperature. Silica gel was removed
by filtration and the filtrate was washed with water and
brine, and dried over MgSO4. The crude product (3.8 g)
was used for the next reaction. For analytical purposes,
a sample of the crude reaction mixture was chro-
matographed (silica, hexane/ethyl acetate, 9:1). 1H
NMR (500 MHz, 363 K, toluene-d8) 0.0–0.1 (m, 6H),
0.9–1.0(5) (m, 30H), 2.19 (dt, Jt=7.2 Hz, Jd=12.8 Hz,
1H), 2.3–2.4 (m, 1H), 3.9–3.9(6) (m, 2H), 4.02 (d,
J=3.3 Hz, 1H), 4.80 (dt, Jt=7.4 Hz, Jd=10.7 Hz, 1H),
5.0–5.0(5) (m, 1H), 5.11 (s, 1H), 5.36 (dd, J=12.1, 6.1
Hz, 1H), 5.5–5.7 (s, 1H), 7.0–7.3 (m, 5H); 13C NMR
(125 MHz, 363 K, toluene-d8) −5.3, −5.2, 12.8, 12.8,
18.3, 18.3, 26.2, 26.4, 62.3, 67.1, 70.6, 80.9, 128.6,
1238.9, 129.1; IR (film) 875, 1111, 1255, 1410, 1715,
2880, 2960, 3480 cm−1; LSIMSHR C28H52NO5Si2 (M+
H)+ calcd 560.3204, found 560.3212.
4.2.17. (2R,3S)-3-Triisopropylsilanyloxypyrrolidine-1,2-
dicarboxylic acid-1-benzyl ester, 20. Alcohol 19 (770 mg,
1.9 mmol) was dissolved in the mixture of CCl4 (3.8
mL) and acetonitrile (3.8 mL) then water (5.7 mL),
NaIO4 (1.67 g, 7.8 mmol) and RuCl3·xH2O (10 mg,
0.05 mmol) were added. The resulting biphasic mixture
was vigorously stirred overnight at rt. CH2Cl2 (10 mL)
was added and the phases were separated. The upper
aqueous phase was extracted with CH2Cl2 and the
combined organic extracts were dried (MgSO4) and
concentrated to afford the crude product (790 mg). For
analytical purposes, a sample of the crude reaction
mixture was chromatographed (silica, hexane/ethyl ace-
tate). 1H NMR (500 MHz, 363 K, toluene-d8) 1.0–
1.0(7) (m, 21H), 1.7–1.7(6) (m, 1H), 2.0–2.1 (m, 1H),
3.2–3.3 (m, 1H), 3.6–3.7 (m, 1H), 4.36 (q, J=7.4 Hz,
1H), 4.44 (s, 1H), 5.07 (AB/2, J=12.4 Hz, 1H), 5.12
(AB/2, J=12.4 Hz, 1H), 7.0–7.3 (m, 5H); 13C NMR
(125 MHz, 363 K, toluene-d8) 12.8, 18.2, 33.0, 44.4,
63.6, 67.4, 73.4, 128.3, 128.8, 129.1, 154.7, 173.8; IR
(film) 870, 1122, 1352, 1410, 1710, 2880, 2960, 2500–
3600 m−1; LSIMSHR C22H36NO5Si (M+H)+ calcd
422.2363, found 422.2364; [h]2D0 −7.3 (c 2.3, CHCl3).
4.2.15. (2S,3S)-[2-(tert-Butyldimethylsilanyloxymethyl)-
3-triisopropylsilanyloxy]pyrrolidine-1-carboxylic
acid
benzyl ester, 18. To the solution of compound 17 (3.8 g,
crude) in a mixture of concentrated acetic acid (38 mL)
and methanol (38 mL) was added NaBH3CN (8 g, 127
mmol) and the temperature was maintained below
30°C. When the reaction was complete by TLC, the
reaction mixture was diluted with water and neutralised
with solid NaHCO3. Extraction with ethyl acetate
afforded an oil (3.8 g), which was used in the next
reaction without purification. For analytical purposes,
a sample of the crude reaction mixture was chro-
matographed (silica, hexane/ethyl acetate, 9:1). 1H
NMR (500 MHz, 363 K, toluene-d8) 0.03 (s, 3H), 0.05
(s, 3H), 0.92 (s, 9H), 1.0–1.1 (m, 21H), 1.78 (ddt,
Jt=7.4 Hz, Jd=12.3, 2.6 Hz, 1H), 3.2–3.4 (m, 1H),
3.5–3.5(6) (m, 1H), 3.82 (s, 1H), 3.97 (d, J=9.2 Hz,
1H), 4.08 (s, 1H), 4.25 (dt, Jt=7.2 Hz, Jd=9.6 Hz, 1H),
5.1–5.2 (m, 3H), 7.0–7.1 (m, 5H); 13C NMR (125 MHz,
363 K, toluene-d8) −5.3, 12.8, 12.9, 18.3, 26.3, 33.1,
44.3, 61.6, 66.9, 72.5, 128.6, 128.7, 128.9, 154.9; IR
(film) 829, 1085, 1110, 1410, 1710, 2880, 2960 cm−1;
LSIMSHR C22H52NO4Si2 (M+H)+ calcd 522.3435,
found 522.3428; [h]2D0 27.6 (c 1.5, CHCl3).
4.2.18. cis-(2R,3S)-3-Hydroxyproline, 1. Compound 20
(131 mg, 0.31 mmol) was dissolved in acetonitrile (1.5
mL) in a Teflon Erlenmayer flask. Hydrofluoric acid
(0.4 mL, 40%) and silica gel (40 mg) were added, and
the resulting mixture was stirred for 50 min at 50–60°C.
The excess hydrofluoric acid was decomposed by addi-
tion of silica gel (250 mg). The solvents were removed
and the remaining silica gel was loaded onto the top of
a silica column for chromatography. Chromatography
(silica, CH2Cl2/methanol, solvent gradient from 95:5 to
70:30 with the addition of 0.3% HCOOH from the
beginning) afforded the desired product 21 (67 mg,
81%). Hydrogenation of the hydroxy acid 21 (49 mg,
0.2 mmol) in methanol (1 mL) in the presence of Pd/C
(10%) for 18 h followed by filtration through Celite and
1
evaporation gave white crystals of 1 (24 mg, 99%). H
NMR (500 MHz, D2O) 2.0–2.1 (m, 1H), 2.1(4)–2.2 (m,
1H), 3.33 (dt, Jt=11.7 Hz, Jd=2.8 Hz, 1H), 3.49 (dt,
Jt=10.8 Hz, Jd=7.7 Hz, 1H), 3.99 (d, J=4.1 Hz, 1H),

D. Gryko et al. / Tetrahedron: Asymmetry 13 (2002) 1103–1113
1113
4.63 (m, 1H); 13C NMR (125 MHz, D2O) 36.4, 46.6,
70.5, 74.4; IR (film) 870, 1122, 1352, 1410, 1710, 2880,
2960, 2500–3600 m−1; EI MSHR C5H9NO3 (M)+ calcd
131.0582, found 131.0583; [h]2D0 +89 (c 0.7, H2O), mp
225–235°C {lit. 220–230°C}.
10. Pelyves, I. F.; Monneret, C.; Herczegh, P. Synthetic
Aspects of Aminodeoxy Sugars of Antibiotics; Spring:
Heidelberg, 1988.
11. Jurczak, J.; Golebiowski, A. Chem. Rev. 1989, 89, 149–
164.
12. Jurczak, J.; Golebiowski, A. Studies in Natural Product
Chemistry, Atta-ur-Rehman, Ed.; Elsevier: Amsterdam,
1989; Vol. 4, p. 111.
13. Golebiowski, A.; Jurczak, J. In Recent Progress in the
Chemical Synthesis of Antibiotics; Lukas, G.; Ohno, M.,
Eds.; Springer: Heidelberg, 1990; p. 365.
14. Dondoni, A. In Modern Synthetic Methods; Scheffold,
R., Ed.; Verlag Helvetica Chimica Acta: Basel, 1992; p.
377.
15. Jurczak, J.; Golebiowski, A. In Antibiotics and Antiviral
Compounds Chemical Synthesis and Modification; Krohn,
K.; Kirst, H.; Maag, H., Eds.; VCH: Weinheim, 1993; p.
343.
16. Golebiowski, A.; Jurczak, J. Synlett 1993, 241–245.
17. Kiciak, K.; Jakobsson, U.; Golebiowski, A.; Jurczak, J.
Polish J. Chem. 1994, 68, 199–238.
18. Jurczak, J. In Preparative Carbohydrate Chemistry;
Hanessian, S., Ed.; Dekker: New York, 1997; p. 593.
19. Jurczak, J.; Prokopowicz, P.; Golebiowski, A. Tetra-
hedron Lett. 1993, 34, 7107–7110.
20. Jurczak, J.; Prokopowicz, P. Tetrahedron Lett. 1998, 39,
9835–9838.
4.2.19. cis-(2R,3S)-3-Hydroxyproline hydrochloride 22.
To a solution of compound 18 (100 mg, 0.25 mmol) in
THF (1 mL) was added Bu4NF·3H2O (150 mg, 0.48
mmol) and the reaction mixture was stirred for 1.5 h at
room temperature. Aqueous NaHCO3 was added and
the reaction mixture was extracted with ether. Column
chromatography (silica, hexane/ethyl acetate, from 1:1
to 3:7) afforded the deprotected compound (57 mg,
92%). Hydrogenation in the presence of Pd/C (10%)
followed by the addition of AcCl (0.1 mL, 1.4 mmol)
1
gave compound 22 (31 mg, 90%). H NMR (200 MHz,
D2O) 1.9–2.4 (m, 2H), 3.4–4.1 (m, 6H); 13C NMR (50
MHz, D2O) 33.4, 44.0, 58.4, 66.1, 70.6; IR (film) 1058,
1395, 1462, 1674, 2743, 2958, 3361 cm−1; EI MSHR for
free amine C5H11NO2 (M)+ calcd 117.0790, found
117.0790; [h]2D0 +16.9 (c 1.1, MeOH).
Acknowledgements
21. Jurczak, J.; Gryko, D.; Kobrzycka, E.; Gruza, H.; Proko-
powicz, P. Tetrahedron 1998, 54, 6051–6064.
22. Gryko, D.; Urbanczyk-Lipkowska, Z.; Jurczak, J. Tetra-
hedron 1997, 39, 13373–13382.
This work was supported by the Polish State Commit-
tee for Scientific Research, Grant PBZ 6.05/T09/1999.
23. Gryko, D.; Jurczak, J. Helv. Chim. Acta 2000, 83, 2705–
2711.
24. Garner, P.; Ramakanth, S. J. Org. Chem. 1986, 51,
2609–2612.
25. Kiyooka, S.; Nakano, M.; Shiota, F.; Fujiyama, R. J.
Org. Chem. 1989, 54, 5409–5411.
References
1. Stein, T. Adv. Protein Chem. 1993, 44, 1–23.
2. Be`lec, L.; Slaninova, J.; Lubell, W. D. J. Med. Chem.
2000, 43, 1448–1455.
26. Keck, G. E.; Boden, E. P. Tetrahedron Lett. 1984, 25,
1879–1883.
3. Weiss, M. S.; Jabs, A.; Hilgenfeld, R. Nat. Struct. Biol.
1998, 5, 676–676.
27. Khan, S. D.; Keck, G. E.; Hehre, W. J. Tetrahedron Lett.
1987, 28, 279–281.
4. Gardiner, R. A.; Rinehart, K. L.; Snyder, J. J.; Broquist,
H. P. J. Am. Chem. Soc. 1968, 90, 5639–5640.
5. Hohenschutz, L. D.; Bell, E. A.; Jewess, P. J.; Leworthy,
D. P.; Pryce, R. J.; Arnold, E.; Clardy, J. Phytochemistry
1981, 20, 811–814.
28. Shambayati, S.; Schreiber, S. L. In Lewis Acid Carbonyl
Complexation in Comprehensive Organic Synthesis; Trost,
B. M., Ed.; Pergamon Press: Oxford, 1991; p. 283.
29. Reetz, M. T. Angew. Chem. 1984, 96, 542–555.
30. Kunz, T.; Janowitz, A.; Reissig, H. Chem. Ber. 1989, 122,
2165–2175.
31. Jarosz, S. Polish J. Chem. 1992, 66, 1853–1858.
32. Daumas, M.; VO-Quang, L.; Le Goffic, F. Synthesis
1989, 64–65.
6. Kakinuma, K.; Otake, N.; Yonehara, H. Tetrahedron
Lett. 1972, 2509–2512.
7. Ha¨usler, H.; Kawakami, R. P.; Mlaker, E.; Severn, W.
B.; Wrodnigg, T. M.; Stu¨tz, A. E. J. Carbohydr. Chem.
2000, 19, 435–449.
8. Provencher, L.; Steensma, D. H.; Wong, C.-H. Bioorg.
33. Carlsen, H. J.; Katsuki, T.; Martin, V. S.; Sharpless, K.
Med. Chem. 1994, 2, 1179–1188.
B. J. Org. Chem. 1981, 46, 3936–3938.
34. Thaning, M.; Wistrand, L. G. Acta Chem. Scand. 1989,
43, 290–295.
9. Coppola, G. M.; Schuster, H. H. Asymmetric Synthesis:
Construction of Chiral Molecules Using Amino Acids;
Wiley: New York, 1987.