Investigations of a nucleophilic alaninol synthon derived from serine

Article abstract of DOI:10.1021/ja9906249

A nucleophilic synthon, (S)-(+)-4-(2-oxazolidonyl)-methyltriphenylphosphinyl iodide 6, available from L-serine in five steps (overall yield of 52%), reacts with aldehydes to produce alkenes in good to excellent yields (74-89%) and, in some cases, provides

Full text of DOI:10.1021/ja9906249

J. Am. Chem. Soc. 1999, 121, 7509-7516
7509
Investigations of a Nucleophilic Alaninol Synthon Derived from
Serine
Mukund P. Sibi,* Drew Rutherford,1 Paul A. Renhowe,2 and Biqin Li
Contribution from the Department of Chemistry, North Dakota State UniVersity,
Fargo, North Dakota 58105-5516
ReceiVed February 26, 1999
Abstract: A nucleophilic synthon, (S)-(+)-4-(2-oxazolidonyl)-methyltriphenylphosphinyl iodide 6, available
from L-serine in five steps (overall yield of 52%), reacts with aldehydes to produce alkenes in good to excellent
yields (74-89%) and, in some cases, provides excellent stereocontrol of the new double bond. The geometry
of the newly formed double bond is influenced by the nature of the aldehyde and reaction conditions. The
trends in olefin configuration are discussed. Application of this methodology allows for easy preparation of
molecules containing double bonds allylic to nitrogen, including oxazolidinones and â,γ-unsaturated amino
alcohols. Several of the unsaturated oxazolidinones are converted to â,γ-unsaturated amino alcohols in high
yields (75 to 90%).
Naturally occurring difunctional amino acids with their rich
array of functional groups and inherent chirality are ideal
building blocks for the preparation of useful chiral synthons.
Serine, an amino acid readily available in both enantiomeric
forms, has served as one such starting material. Differentially
protected serine and derivatives derived from the parent amino
acid are useful synthons in the preparation of natural and
unnatural amino acids and amino alcohols. The chemistry of
nucleophilic, electrophilic, and radical alanine or alaninol
equivalents derived from serine has been explored in several
laboratories,3 most notably by Baldwin,4 Garner,5 Jackson,6
Sasaki,7 Vederas8 and Viallefont.9 Work in our laboratory has
also focused on synthons derived from serine.10 The chemistry
of a nucleophilic alaninol synthon is the subject of this paper.11
A major emphasis of the utility of nucleophilic alanine
(alaninol) synthons has been in the development of new
methodologies for the syntheses of nonproteinogenic amino
acids. These amino acids are significant because of their wide
range of biological activities.12 In particular, â,γ-unsaturated
amino acids have attracted considerable interest because of their
(1) Taken in part from the M. S. thesis of Drew Rutherford (North Dakota
State University, 1994) and M. S. thesis of Biqin Li (North Dakota State
University, 1992).
(2) Undergraduate research participant 1988-1990, NSF-REU participant
(6) Selected examples of reactions of zinc reagents see: (a) Jackson,
R. F. W.; James, K.; Wythes, M. J.; Wood., A. J. Chem. Soc., Chem.
Commun. 1989, 644. (b) Jackson, R. F. W.; Wishart, N.; Wood, A.; James,
K.; Wythes, M. J. J. Org. Chem. 1992, 57, 3397. (c) Jackson, R. F. W.;
Turner, D.; Block, M. H. Synlett 1996, 862. (d) Jung, M. E.; Starkey, L. S.
Tetrahedron 1997, 53, 8815.
(7) Sulfones: (a) Sasaki, N. A.; Hashimoto, C.; Potier, P. Tetrahedron
Lett. 1987, 28, 6069. (b) Sasaki, N. A.; Hashimoto, C.; Pauly, R.
Tetrahedron Lett. 1989, 30, 1943.
(8) Electrophilic alaninol synthons. â-lactones: (a) Arnold, L. D.; May,
R. G.; Vederas, J. C. J. Am. Chem. Soc. 1988, 110, 2237 and references
therein. (b) Arnold, L. D.; Drover, J. C. G.; Vederas, J. C. J. Am. Chem.
Soc. 1987, 109, 4649.
(9) Electrophilic synthons: (a) Bajgrowicz, J. A.; El Hallaoui, A.;
Jacquier, R.; Pigiere, Ch.; Viallefont, Ph. Tetrahedron 1985, 41, 1833. (b)
El Marini, A.; Roumestant, M. L.; Viallefont, P.; Razafindrambou, D.;
Bonato, M.; Follet, M. Synthesis 1992, 1104. (c) Atmani, A.; El Hallaoui,
A.; El Hajji, S.; Roumestant, M. L.; Viallefont, P. Synth. Commun. 1991,
21, 2383.
1988.
(3) General: Williams, R. M. Synthesis of Optically ActiVe R-Amino
Acids; Pergamon: New York, 1989; Chapter 2 and references therein.
Duthaler, R. O. Tetrahedron 1994, 50, 1539. Nucleophilic synthons: (a)
Seebach, D.; Wasmuth, D. Angew. Chem., Int. Ed. Engl. 1981, 20, 971. (b)
Wolf, J.-P.; Rapoport, H. J. Org. Chem. 1989, 54, 3164. (c) Daddu, R.;
Eckhardt, M.; Furlong, M.; Knoess, H. P.; Berger, S.; Knochel, P.
Tetrahedron 1994, 50, 2415. (d) Ye, B.; Otaka, A.; Burke, T. R., Jr. Synlett
1996, 459. (e) Malan, C.; Morin, C. Synlett 1996, 167. (f) Duthaler, R. O.
Angew. Chem., Int. Ed. Engl. 1991, 30, 705. (g) Blanchard, P.; El Kortbi,
M. S.; Fourrey, J.-L.; Robert-Gero, M. Tetrahedron Lett. 1992, 33, 3319.
Electrophilic synthons: (h) Dureault, A.; Tranchepain I.; Depezay, J.-C. J.
Org. Chem. 1989, 54, 5324. (i) Sato, K.; Kozikowski, A. P. Tetrahedron
Lett. 1989, 30, 4073. Radical alaninol synthons: (j) Maria, E. J.; Da Silva,
A. D.; Fourrey, J.-L.; Machado, A.; Robert-Gero, M. Tetrahedron Lett. 1994,
35, 3301.
(4) Nucleophilic synthons: (a) Baldwin, J. E.; Moloney, M. G.; North,
M J. Chem. Soc., Perkin Trans. 1 1989, 833. Aspartate Derivatives: (b)
Baldwin, J. E.; Moloney, M. G.; North, M. Tetrahedron 1989, 45, 6309.
Electrophilic alanine/alaninol synthons. Aziridines: (c) Baldwin, J. E.;
Adlington, R. M.; Robinson, N. G. J. Chem. Soc., Chem. Commun. 1987,
153. Sulfamidates: (d) Baldwin, J. E.; Spivey, A. C.; Schofield, C. J.
Tetrahedron: Asymmetry 1990, 1, 881. Radical alanine synthon: (e)
Baldwin, J. E.; Adlington, R. M.; Birch, D. J.; Crawford, J. A.; Sweeney,
J. B. J. Chem. Soc., Chem. Commun. 1986, 1339.
(10) Nucleophilic synthons: (a) Sibi, M. P.; Christensen, J. W.; Li, B.;
Renhowe, P. A. J. Org. Chem. 1992, 57, 4329. (b) Sibi, M. P.; Li, B.
Tetrahedron Lett. 1992, 33, 4115. Electrophilic synthon: (c) Sibi, M. P.;
Rutherford, D.; Sharma, R. J. Chem. Soc., Perkin Trans. 1 1994, 1675.
Radical synthon: (d) Sibi, M. P.; Ji, J. J. Am. Chem. Soc. 1996, 118, 3063.
(11) For a preliminary account of this work see: Sibi, M. P.; Renhowe,
P. A. Tetrahedron Lett. 1990, 31, 7407.
(5) (a) Garner, P.; Park, J. M. J. Org. Chem. 1990, 55, 3772 and
references therein. For slight variants see: (b) Beaulieu, P. L.; Schiller, P.
W. Tetrahedron Lett. 1988, 29, 2019. (c) Blaskovich, M. A.; Lajoie, G. A.
J. Am. Chem. Soc. 1993, 115, 5021.
(12) (a) Barrett, G. C., Ed. Chemistry and Biochemistry of the Amino
Acids; Chapman and Hall: London, U.K., 1985. (b) Wagner, I.; Musso, H.
Angew. Chem., Int. Ed. Engl. 1983, 22, 816. (c) Williams, R. M.; Sinclair,
P. J.; Zhai, D.; Chen, J. J. Am. Chem. Soc. 1988, 110, 1547.
10.1021/ja9906249 CCC: $18.00 © 1999 American Chemical Society
Published on Web 08/11/1999

7510 J. Am. Chem. Soc., Vol. 121, No. 33, 1999
Sibi et al.
potential as suicide enzyme inhibitors,13 and several procedures
are available for their synthesis.14 Sasaki et al.7 have prepared
very useful synthons from L- and D-serine and evaluated them
as nucleophilic alaninol equivalents. Itaya and co-workers,15 in
their synthesis of wybutine, showed that a Wittig reagent derived
from serine undergoes olefination with little racemization, albeit
in very low yields (5-13%). A modification of this reagent
and Wittig reactions with slightly better efficiency was recently
reported by the same group.16 We were interested in preparing
a nucleophilic alaninol synthon from serine which could be used
for the generation of either E or Z â,γ-unsaturated amino
alcohols and amino acids, with good stereocontrol, and in high
chemical yields.17 This paper describes the synthesis and
reactions of one such synthon and presents a mechanistic
explanation for the observed stereoselectivities in the olefination.
Scheme 1
isolation of the product ester required a cumbersome lyophiliza-
tion due to its high solubility in water. In a modified and
optimized procedure, triphosgene can be successfully used as a
phosgene equivalent. The yield of this reaction is 95%, and a
nonaqueous work up allows for easy isolation of the product
oxazolidinone.19 The next step in the sequence was the reduction
of the methyl ester using sodium borohydride which furnished
a highly water-soluble alcohol 3 in excellent yields.20 The
alcohol is heat-sensitive and racemizes at high temperatures.
The alcohol 3 was then converted to the corresponding iodo
compound 5 in a two-step sequence. Tosylation of the primary
alcohol using TsCl/pyridine, followed by refluxing with sodium
iodide in acetone for 5 h, gave the iodide 5 (68% for two steps).
Several alternative one-step processes were evaluated for the
conversion of 3 to 5. These were either inferior to the two-step
procedure in overall yields or required chromatographic puri-
fication. The phosphonium salt 6 was prepared by stirring the
iodo compound with excess triphenylphosphine in DMF at 100
°C for 24 h. The phosphonium salt is stable and can be stored
in a desiccator indefinitely. Typically it was dried in an
abderhalden apparatus prior to use. The overall yield for the
synthesis of the phosphonium salt from serine methyl ester over
five steps is 52% and requires no chromatographic purification.
Having established an optimal procedure for the preparation
of the phosphonium salt, we were concerned about maintaining
the optical integrity of the chiral center during deprotonation
and formation of the ylide from 6. Encouraged by the studies
of Meyers and others,21 we surmised that dideprotonation of 6
should prevent formation of any â-eliminated product and also
that the oxazolidinone should reduce the acidity of the methine
at the chiral center, rendering it less prone to deprotonation (vide
infra). Seebach has also examined the structural features
necessary to prevent â-elimination during carbanion formation.22
Both alkyllithium and metal amide bases were evaluated for
the generation of the ylide from 6. These results are tabulated
in Table 1. Treatment of compound 6 with 1.9 equiv of base
(n-BuLi or LiHMDS) in THF or DME as solvents at -78 °C
resulted in an orange to red colored solution which was stirred
at that temperature for 1 h and then quenched with benzaldehyde
Results and Discussion
Our initial task was to establish optimal reaction conditions
for the preparation of the phosphonium salt 6 from readily
available L-serine methyl ester hydrochloride18 such that the
intermediates required very little or no chromatographic puri-
fication (Scheme 1). In our earlier reported methodology,
treatment of serine methyl ester hydrochloride with phosgene
(as a 20% toluene solution) in aqueous potassium carbonate (3
h at 0 °C) gave the best conversion to the oxazolidinone 2. The
(13) (a) Chang, M. N. T.; Walsh, C. J. Am. Chem. Soc. 1980, 102, 7368.
(b) Silverman, R. B. Mechanism-Based Enzyme InactiVation: Chemistry
and Enzymology; CRC Press: Boca Raton, FL, 1988. (c) Walsh, C.
Tetrahedron 1982, 38, 871. (d) Rando, R. R. Science 1974, 185, 320. (d)
Andrews, M. D.; O’Callaghan, K. A.; Vederas, J. C. Tetrahedron 1997,
53, 8295.
(14) For the synthesis of â,γ-unsaturated amino acids see review:
Havlicek, L.; Hanus, J. Collect. Czech. Chem. Commun. 1991, 56, 1365.
For a recent effort in this area see: Petasis, N. A.; Zavialov, I. A. J. Am.
Chem. Soc. 1997, 119, 445 and references therein. Other selected work
see: (a) Baldwin, J. E.; Adlington, R. M.; O’Niel, I. A.; Schofield, C.;
Spivey, A. C.; Sweeney, J. B. J. Chem. Soc., Chem. Commun. 1989, 1852.
(b) Duthaler, R. O. Angew. Chem., Int. Ed. Engl. 1991, 30, 705. (c)
Castelhano, A. L.; Horne, S.; Taylor, G. J.; Billedeau, R.; Krantz, A.
Tetrahedron 1988, 44, 5451 and references therein. (d) Angst, C. Pure Appl.
Chem. 1987, 59, 373. (e) Crisp, G. T.; Glink, P. T. Tetrahedron 1992, 48,
3541. (e) Badorrey, R.; Cativiela, C.; Diaz-de-Villegas, M. D.; Ga´lvez, J.
A. Synthesis 1997, 747. (f) O’Donnell, M. J.; Li, M.; Bennett, W. D.; Grote,
T. Tetrahedron Lett. 1994, 35, 9383. (g) Pedersen, M. L.; Berkowitz, D.
B. J. Org. Chem. 1993, 58, 6966. (h) Berkowitz, D. B.; Smith, M. K.
Synthesis 1996, 39. (i) Clayden, J.; Collington, E. W.; Warren, S.
Tetrahedron Lett. 1993, 34, 1327. (j) Mehmandoust, M.; Petit, Y.;
Larcheveˆque, M. Tetrahedron Lett. 1992, 33, 4313.
(15) Itaya, T.; Shimomichi, M.; Ozasa, M. Tetrahedron Lett. 1988, 29,
4129.
(19) Sibi, M. P.; Harris, B. J.; Renhowe, P. A.; Rutherford, D.;
Christensen, J. W.; Li, B.; Lu, J. Organic Synth. under evaluation.
(20) Koellensperger, F. G.; Hartelben, Y.; Kretzschmar, R.; Neleter, B.
Ger. Patent. 2538424, 1977; Chem. Abstr. 1977, 87, 23257Z.
(21) (a) Linderman, R. J.; Meyers, A. I. Tetrahedron Lett. 1983, 24, 3043.
(b) Refs 4a and 7a. (c) Ref 3a. (d) McGarvey, G. J.; Hiner, R. N.; Matsubara,
Y.; Oh, T. Tetrahedron Lett. 1983, 24, 2733. (e) McGarvey, G. J.; Williams,
J. M.; Hiner, R. N.; Matsubara, Y.; Oh, T. J. Am. Chem. Soc. 1986, 108,
4943. (f) Jackson, R. F. W.; James, K.; Wythes, M. J.; Wood, A.
Tetrahedron Lett. 1989, 30, 5941.
(16) (a) Itaya, T.; Iida, T.; Shimuzu, S.; Mizutani, A.; Morisue, M.;
Sugimoto, Y.; Tachinaka, M. Chem. Pharm. Bull. Jpn. 1993, 41, 252. (b)
Itaya, T.; Kanai, T.; Iida, T. Tetrahedron Lett. 1997, 38, 1979.
(17) For applications of this synthon in synthesis see: (a) Refs 10a and
c. (b) Bertozzi, C.; Hoeprich, P. D., Jr.; Bednarski, M. D. J. Org. Chem.
1992, 57, 6092. (c) Bertozzi, C.; Cook, D. G.; Kobertz, W. R.; Gonzalez-
Scarano, F.; Bednarski, M. J. Am. Chem. Soc. 1992, 114, 10639.
(18) Serine methyl ester hydrochloride is commercially available or it
can be readily prepared in high yields from serine, thionyl chloride, and
methanol (Guttmann, S.; Boissonnas, R. A. HelV. Chim. Acta 1958, 41,
1852).
(22) Eyer, M.; Seebach, D. J. Am. Chem. Soc. 1985, 107, 3601 and
references therein.

A Nucleophilic Alaninol Synthon
J. Am. Chem. Soc., Vol. 121, No. 33, 1999 7511
Table 1. Wittig Reactions with Compound 6 and Benzaldehyde
Table 2. E:Z Ratios of Wittig Reaction Using Aromatic Aldehydes
yield
%
E:Z
a
entry
conditions
(8/9)b
1
2
3
4
5
6
7
8
n-BuLi, THF, -78 °C to rt
86
86
87
88
62
68
57
62
>99:1
3.6:1
>99:1
>99:1
3.2:1
1.9:1
2.5:1
3.2:1
n-BuLi, THF, -78 °C 2.5 h
n-BuLi, DME, -78 °C to rt
n-LiHMDS, THF, -78 °C to rt
n-LiHMDS, THF, -78 °C 2.5 h
n-BuLi, THF, 20% HMPA, -78 °C to rt
NaHMDS, THF, -78 °C 2.5 h
NaHMDS, THF, -78 °C to rt
temp. of
quench (°C)
yielda product
(%)
E:Z
(config.) ratiob
entry
1
RCHO
PhCHO
25
86
8 (E)
9 (Z)
100:0
a Isolated yield after column chromatography. Ratios determined
b
1
by H NMR.
2
3
-78
86
88
78:22
100:0
4-OMe-PhCHO
4-piperonal
2-furyl
25
10 (E)
11 (Z)
to produce alkene 8. The product identity and olefin stereo-
chemistry were established by H NMR. The optical purity of
1
4
5
-78
88
81
79:21
100:0
the olefin product was established by converting the product to
known N-BOC-homophenylalanine. No racemization was ob-
served during the olefination process.23 Addition of HMPA
increased the Z selectivity of the reaction, providing nearly equal
amounts of E:Z mixture (entry 6). The effect of the base
counterion on stereoselectivity was also examined. Use of
NaHMDS gave lower chemical yields as well as stereoselectivity
(entry 7).24 A change in temperature did not lead to further
improvements (entry 8). For further experiments, n-BuLi was
used as the base for deprotonation and formation of the ylide.
The next set of reaction conditions was developed to
distinguish what factors largely influenced the E/Z ratios for
the Wittig reaction of 6 with aldehydes. Preliminary reactions
showed that, besides the identity of the aldehyde, the temperature
of the reaction at the time of quenching was the most influential
on E/Z ratios. For the investigation, two standard reaction
conditions were selected. In the first method (method A), ylide
7 was generated from 6 using n-BuLi (1.95 equiv, -78 °C) as
the base and the yellow-orange solution was stirred for 1 h.
The aldehyde was added at -78 °C as a solution in THF, the
reaction stirred at -78 °C for 2.5 h then quenched with a
saturated ammonium chloride solution at -78 °C. The second
reaction (method B) was identical to the first, except that the
ammonium chloride quench was carried out after warming the
reaction to room temperature. The two reactions were conducted
side-by-side, and after identical workups, the crude reactions
were analyzed to determine the E/Z olefin ratio of the products
8 and 9.
25
12 (E)
13 (Z)
6
7
-78
73
88
81:19
92:8
25
14 (E)
15 (Z)
8
9
-78
77
88
81:19
100:0
2-thienyl
25
16 (E)
17 (Z)
10
-78
86
81:19
a Isolated yield after column chromatography. b Ratios determined
1
by H NMR.
Table 3. E:Z Ratios of Wittig Reaction Using Aliphatic Aldehydes
temp. of
quench (°C)
yielda product
E:Z
(config.) ratiob
entry
1
RCHO
(%)
CH3(CH2)4CHO
25
89
18 (E)
19 (Z)
31:69
2
3
-78
25
74
82
23:77
50:50
(H3C)2CHCHO
20 (E)
21 (Z)
4
5
-78
25
59
86
24:76
30:70
CH3(CH2)13CHO
22 (E)
23 (Z)
6
-78
87
31:69
a Isolated yield after column chromatography. b Ratios determined
1
by H NMR.
integration of the olefinic protons in the 1H NMR. In the
reactions of 7 with aromatic aldehydes quenched at -78 °C,
although the trans product still predominated, the low-temper-
ature quench increased the appearance of the cis isomer (Table
2, entries 2, 4, 6, 8, and 10). It is interesting to note that the
identity of the aromatic aldehyde had little effect on the E/Z
outcome of the reactions when comparing the E/Z ratios of either
reactions quenched at room temperature or those quenched at
-78 °C.
The reactions of 7 with aliphatic aldehydes are presented in
Table 3. The aliphatic aldehydes were chosen to explore steric
effects on the reaction: linear vs branched and short- vs long-
chain. Examination of the data in Table 3 showed a distinct
trend. Reactions of 7 with aliphatic aldehydes gave predomi-
nately Z products, regardless of the temperature at which the
reaction was quenched (except in Table 3, entry 3). This
observation is in direct contrast to the reactions of 7 with
aromatic aldehydes where the predominant unsaturated product
The reactions of 7 with five different aromatic aldehydes
using the standard reaction conditions are shown in Table 2.
The aldehydes were selected to encompass a range of func-
tionalities on and in the aromatic ring. This included the
heteroatomic aromatic aldehydes of thiophene and furan. A
general trend is clearly evident from the data. Reactions that
were quenched at room temperature gave almost exclusively E
products (Table 2, entries 1, 3, 5, 7, and 9). In all cases, except
the reaction with furyl-2-carboxaldehyde (Table 2, entry 7), no
Z product was detectable. The E/Z ratios were established by
(23) Several products from the Wittig olefination process have been
converted to known compounds. See, for example, refs 10a,b and 17.
(24) The reasons for the variation in stereoselectivity with changes in
the base are not apparent.

7512 J. Am. Chem. Soc., Vol. 121, No. 33, 1999
Sibi et al.
Table 4. E:Z Ratios of Wittig Reaction Using Aldehydes
Containing R- and â-Heteroatoms
E product observed when the oxygen was R was nearly twice
that when the same heteroatom was â.25
When oxygen and sulfur in the R position were compared, a
significant change in the E/Z ratio was observed. With oxygen
in the R position, the E/Z ratio for the reaction with 6 was similar
to the E/Z ratio seen with linear aliphatic aldehydes (Table 4,
entry 1 vs Table 3, entry 1 or 5): moderate Z selectivity.
However, when sulfur was substituted for the oxygen in the R
position, the selectivity was reversed: the E product predomi-
nates.
The highest cis selectivity that was observed with any type
of aldehyde was the reaction of D-glyceraldehyde acetonide with
the ylide 7 (Table 4, entries 7 and 8). The reaction gave only
1
the Z product; no E product was observed in the H NMR or
isolated by chromatography. Moving the chiral center from the
R position to the â position had a dramatic effect on the E/Z
ratio. With the chiral center closer to the reaction site, a single
product was produced (Table 4, entry 7); when the chiral center
was one carbon removed from the reaction site, a 40:60 Z/E
distribution was obtained (Table 4, entry 9).
The ylide 7 investigated in this work exhibits characteristics
indicative of both a stabilized and nonstabilized ylide and is
best classified as semistabilized.26 For this system, prediction
of E:Z outcome worked best by examination of the aldehyde
rather than the ylide. In systems where the aldehyde was
stabilized by resonance interaction with the aromatic ring,
predominately E configured olefins were produced (Table 2).
The comparison of reactions quenched at -78 °C with those
quenched at room temperature indicates that equilibration plays
a key role in determining product geometry. The low temper-
ature quenches give a kinetic product ratio, whereas room
temperature quench results in a thermodynamic outcome. In fact,
each of the aromatic aldehydes gave nearly identical E:Z ratios
when the different aldehydes were quenched at -78 °C. The E
selectivity observed with ylide 7 and aromatic aldehydes at
higher temperatures is in most part due to equilibration.
Additional enhancements in selectivity may result from nucleo-
philic participation of the R-amino substituent and subsequent
equilibration of the oxaphosphetane intermediates (eq 2), and
a Isolated yield after column chromatography. b Ratios determined
by 1H NMR. c Isolated yield after Wittig reaction and BOC protection.
d E isomer 28 gave 30 after BOC protection. e Z isomer 29 gave 31
after BOC protection. f E isomer 33 gave 35 after BOC protection.g Z
isomer 34 gave 36 after BOC protection.
was in the E configuration. Reactions that were quenched at
-78 °C showed an increase in the amount of the Z isomer (Table
3, entries 2 and 4). The identity of the aliphatic aldehyde did
affect the E/Z product distribution to some extent. Reactions of
7 with the more sterically demanding valeraldehyde gave a
higher proportion of the E product when quenched at room
temperature (Table 3, entry 3) as compared to the two linear
aldehydes quenched in the same manner (Table 3, entries 1 and
5). Reactions quenched at -78 °C gave predominately Z
products in similar E/Z ratios regardless of the aliphatic aldehyde
used. E/Z ratios did not vary outside experimental error for short-
vs long-chain linear aliphatic aldehydes (Table 3, entries 1 vs
5 and 2 vs 6).
The third type of aldehyde that was examined in this study
was aliphatic aldehydes which contained heteroatoms R and â
to the carbonyl group. Reactions of 7 with these aldehydes are
presented in Table 4. One of the most striking features of these
data is that there is no difference between reactions that were
quenched at room temperature vs reactions that were quenched
at -78 °C. This is in contrast to that observed for reactions
with either aromatic or aliphatic aldehydes. Another intriguing
feature of these data was the difference in E/Z values when an
oxygen R to the carbonyl was compared to an oxygen that was
â to the carbonyl (Table 4, entry 1 vs entry 5). The amount of
(25) The exact reasons for this variation in selectivity with structural
modifications of the aldehydes are not apparent at this time. In an analogous
series, N-protected R-amino aldehydes react with 7 to provide Z-olefins
only. Sibi, M. P.; Christensen, J. W. J. Org. Chem. 1999, in press.
(26) (a) For comprehensive reviews see: (a) Vedejs, E.; Peterson. M. J.
In AdVances in Carbanion Chemistry; Snieckus, V., Ed.; Jai Press:
Greenwich, CT, 1996; Vol. 2, Chapter 1. (b) Maryanoff, B. E.; Reitz, A.
B. Chem. ReV. 1989, 89, 863. (c) Johnson, A. W. Ylides and Imines of
Phosphorus; Wiley: New York, 1993. (d) Enholm, E. J.; Satici, H.; Prasad,
G. J. Org. Chem. 1990, 55, 324.

A Nucleophilic Alaninol Synthon
J. Am. Chem. Soc., Vol. 121, No. 33, 1999 7513
Table 5. Yield E/Z Ratios of 12:13 as a Function of Base
Equivalents
equi
of base
temp. of
quench (°C)
yield
(%)
E:Zb
(12/13)
entry
base
1
2
3
4
5
6
n-BuLi
n-BuLi
n-BuLi
n-BuLi
LDA
1.95
1.0
3.0
3.0
1.95
3.0
25
25
25
-78
25
62a
29
40
50
85
49
100:0
100:0
100:0
91:9
100:0
100:0
Figure 1.
LDA
25
a Isolated yield after column chromatography. b Ratios determined
by H NMR.
partly due to salt effects.27 The observed E-selectivity with 7
and aromatic aldehydes agrees well with the results of Meyers,21
Maryanoff,28 and Mioskowski29 for oxido ylides.
1
In the preliminary report for the reactions with the nucleo-
philic synthon, the double deprotonation of the phosphonium
salt was proposed to have prevented the substrate from race-
mization during ylide formation. To test this theory, we
subjected the synthon to several reaction conditions that involved
manipulating the number of equivalents of the base. The
standard reaction used 1.95 equiv of n-BuLi as the base and
piperonal as the aldehyde. Yields and E/Z ratios for this reaction
were taken as the reference point.
In sharp contrast to the aromatic aldehydes, aliphatic alde-
hydes and aldehydes with R- or â-heteroatom showed no
evidence of equilibration, as the E:Z ratios of room-temperature
quenches vs quenches at -78 °C were essentially identical.
When reactions using aromatic aldehydes were conducted, no
precipitation, or very little, of the insoluble triphenylphosphine
oxide was observed. However, in reactions of aliphatic or
heteroatom containing aldehydes, upon addition of the aldehyde
to the ylide, instantaneous formation of triphenylphosphine oxide
was observed. The best selectivities were with aldehydes
containing a ring system R to the carbonyl (Table 4, entries 7
and 8).30 A working hypothesis for the observed selectivity with
heteroatom containing aldehydes is as follows (Figure 1). The
invariance of diastereomeric ratio with temperature indicates
that reactions of 7 with aliphatic and heteroatom containing
aliphatic aldehydes proceed under kinetic conditions. In these
cases, the initial approach of the aldehyde solely dictated the
product geometries. Assuming that the stereochemistry is
established in the transition state leading to the oxaphosphetanes,
the E:Z ratio differences within the series of alkoxy aldehydes
may be accounted for by considering the transition states shown
in Figure 1. The two set of compounds, that is, R- and â-alkoxy
aldehydes, differ in reactivity, with the former being slightly
more reactive than the latter. The highest Z-selectivity observed
was with glyceraldehyde acetonide (Model F, Figure 1). When
the structures E and F are compared, the R-alkoxy group is
oriented anti to the carbonyl oxygen to minimize dipole-dipole
interactions. In structure F, the ether oxygen may provide
additional stability by interaction with the lithium cation and
thus account for the observed high Z selectivity.31 When F and
G are compared, reactions with the less reactive â-alkoxy
aldehydes could proceed through a later transition state and
account for lower Z:E selectivity.
Two reactions were run side by side; in the first reaction,
1.95 equiv of n-BuLi was used, and in the second, only one
equivalent was used. Both reactions were quenched at room
temperature after a standard reaction sequence was performed
(Table 5). The reaction run under standard conditions gave the
E product exclusively in 62% yield (entry 1). The reaction with
one equivalent of base also gave the E product exclusively, but
in only 29% yield. No racemization was observed in either case.
These results show that dideprotonation is necessary for higher
yield of the olefin product. This observation is consistent with
a competitive deprotonation of the N-H versus the formation
of the ylide. The pKa values for the N-H and the Ph3P+CH2R
are similar (∼20.6 for N-H32 and ∼21-22 for Ph3P+CH2R)33
and will compete for the single equivalent of base. When the
nitrogen is deprotonated, it precludes the formation of the ylide
and the yield is lowered. A second sequence was carried out
with 3 equiv of n-BuLi (Table 5, entries 3 and 4). The reactions
gave the olefins in reasonable yields and in characteristic E/Z
ratios for the respective quenches. In this case, the yields were
lowered because the ylide was in direct competition with the
extra equivalent of n-BuLi for the aldehyde. The butylated
product was formed in 60% (Table 5, entry 3) and 53% (Table
5, entry 4) yield and accounted for the mass balance of the
aldehyde. Even in the presence of an extra equivalent of base,
the olefins showed no signs of racemization. A third reaction
sequence was performed with 3 equiv of nonnucleophilic LDA
as the base. The standard reaction with 1.95 equiv was compared
to one using 3 equiv (Table 5, entries 5 and 6). The standard
reaction (1.95 equiv) gave significantly higher yields than the
one with 3.0 equiv. The major byproducts observed in this
reaction were piperonyl alcohol (28%) and triphenylphosphine
(26%). The piperonyl alcohol presumably arose from the base-
catalyzed Cannizarro reaction of the aldehyde, whereas the
(27) For an excellent discussion on stereochemistry and mechanism of
Wittig reactions using oxido ylides see: Vedejs, E.; Peterson. M. J. In Topics
in Stereochemistry; Eliel, E. L., Wilen, S. H., Eds.; Wiley: New York,
1994, Vol. 21, Chapter 1.
(28) Maryanoff, B. E.; Reitz, A. B.; Duhl-Emswiler, B. A. J. Am. Chem.
Soc. 1985, 107, 217.
(29) Boubia, B.; Mann, A.; Bellamy, F. D.; Mioskowski, C. Angew.
Chem., Int. Ed. Engl. 1990, 29, 1454.
(30) For Z-selective olefinations with R-alkoxyaldehydes see: Takanashi,
T.; Miyazawa, M.; Ueno, H.; Tsuji, J. Tetrahedron Lett. 1986, 27, 3881.
(31) The observed high (Z)-selectivity for the glyceraldhyde acetonide
is mostly due to the reactivity of the aldehyde and the irreversibility of the
reaction. Formation of the alternate (E)-isomer places the rigid dioxolane
ring close to bulky phosphine.
(32) (a) Zhang, X.-M.; Bordwell, F. G. J. Org. Chem. 1994, 59, 6456.
(b) Zhang, X.-M.; Bordwell, F. G. J. Am. Chem. Soc. 1994, 116, 968.
(33) Bordwell, F. G. Acc. Chem. Res. 1988, 21, 456.

7514 J. Am. Chem. Soc., Vol. 121, No. 33, 1999
Sibi et al.
Scheme 2
Scheme 3
dianionic ylide 7 reacted with piperonal to give the E product
exclusively; however, the methylated ylide 40 gave a mixture
of Z and E products and the Z isomer predominated. The
substitution of the methyl group for the hydrogen in 39 has
nearly reversed the E:Z trend that was seen for 6. The results
from reactions, which were quenched at -78 °C, are shown in
entries 3 and 4 of Table 6. The temperature at which reactions
with 39 were quenched made little difference to E:Z ratios, but
chemical yields were lower when the reaction was stopped at
lower temperature (Table 6, entries 2 and 4). The low chemical
yields observed for reactions with 39 can be explained by
considering the electrophilicity of the carbamate carbonyl. When
two equivalents of base is added to 6, the acidic hydrogen is
abstracted, leaving a negative charge on the nitrogen which, by
resonance, is spread over the amide carbon and oxygen. This
drastically reduces the electrophilicity of the carbamate carbonyl,
protecting it from nucleophilic attack. Wittig salt 39 and ylide
40, without the anionic nitrogen, contain reactive carbonyls that
are susceptible to nucleophilic attack by n-BuLi. Thus a
competition between deprotonation to form the ylide and
nucleophilic attack may lower the yields of the olefin products
(vide infra).34 Increased chemical yields of olefins from 39 by
using a nonnucleophilic base (LDA, Table 6, entry 5) or more
equivalents of n-BuLi (Table 6, entry 6) were not realized.
The loss of the anionic nitrogen in 7 also opens up another
possibility for the decomposition of the ylide. When the ylide
was first investigated,11 it was theorized that the double
deprotonation of the phosphonium salt 6 was necessary to reduce
the possibility of â-elimination. The negative charge on the
nitrogen in 7 significantly decreases the likelihood of the
â-elimination pathway (Scheme 3). In order for the elimination
to occur, a double negative charge would reside on the nitrogen.
However, with the loss of the anionic nitrogen, as in 39, the
possibility of the â-elimination pathway is increased. In this
case, the â-elimination would leave a single negative charge
on a highly stabilized amide. This stabilization of the â-elimina-
tion pathway would lead to a much more unstable ylide and
would reduce the yields significantly. Evidence for this â-e-
limination pathway was observed by the partial racemization
of the olefin in Table 6, entry 2. The elimination mechanism
destroyed the stereocenter in 39, and when 39 was regenerated
by reclosure of the ring, a racemized Wittig reagent was formed.
Aldehyde condensation with the racemized Wittig reagent
produced racemic olefins.
Table 6. Comparison of Ylide 7 and Ylide 40 in Side-by-Side
Reactions
temp of
yield of
entry
P
eq n-BuLi quench (°C) olefin (%) E:Za ee (%)
1
2
3
4
5
6
H
CH3
H
CH3
CH3
CH3
1.95
1.0
25
25
-78
-78
25
78
27
67
14
34
14
100:0 >99
41:59
65b
1.95
1.0
81:19 >99c
47:53
48:52
1.0d
2.0
25
49:51
1
a Ratios determined by H NMR. b Determined by comparison of
optical rotation to authentic sample of (E) olefin 41 prepared by
alkylation of 12. c ee for the major isomer. d LDA was used as the base.
triphenylphosphine was a byproduct from decomposition of the
ylide. In neither case was any racemization observed. From this
study, the highest yields were obtained when 1.95 equiv of the
respective base was used. Too little base resulted in significant
reduction in yields, whereas excess base increased the number
of side reactions.
One of the goals of the present study was to determine what
role the anionic nitrogen in ylide 7 played in determining the
stereochemical outcome of the olefinic products. To better
understand this role, we prepared a phosphonium salt in which
the oxazolidinone nitrogen was methylated (Scheme 2). The
synthesis of the N-methylated phosphonium salt 39 was
straightforward. Tosylate 4 was alkylated with methyl iodide
in 85% yield, giving N-methyl tosylate 37 which was converted
to the N-methyl iodide 38 by displacement with sodium iodide
in refluxing acetone.
The replacement of the acidic hydrogen with a methyl group
(6 vs 39) significantly altered the reactivity of the phosphonium
salt. The methylated Wittig salt 39 was subjected to reaction
conditions that normally gave high yields of exclusively E
olefinic products when using the nonmethylated Wittig salt 6
(aromatic aldehyde and method B). Wittig salts 6 and 39 were
converted to their corresponding ylides 7 and 40, using 1.95
equiv of n-BuLi for 6 and 1.0 equiv of n-BuLi for 39. The results
are summarized in Table 6. There were two major differences
between the two reactions. First, the reaction using 39 provided
extremely low yields of the olefinic products; only a 27% yield
was observed (Table 6, entry 2). The nonmethylated Wittig salt
6 gave a similar olefin in nearly triple the yield (Table 6, entry
1). Second, the E:Z ratios were drastically different. The
Conversion of Oxazolidinones to â,γ-unsaturated Amino
Alcohols. Amino acids and amino alcohols containing unsat-
uration of defined geometry at the â-position have attracted
considerable attention. The conversion of oxazolidin-2-ones into
the corresponding amino alcohols has been investigated. The
best method available for this transformation, developed by
Kunieda,35 involved a two-step process. The first step was the
protection of the oxazolidinone nitrogen as the tert-butyloxy-
(34) We believe that the major decomposition pathway is through ring
opening of the ylide and nucleophilic attack at the carbamate carbonyl is
minor. For an example of ring opening of an oxazolidinone by BuLi see:
Mirzaei, Y. R.; Balasubramaniam, T. N.; Lefler, B. J.; Natale, N. R. J.
Heterocycl. Chem. 1990, 27, 2000.
(35) Ishizuka, T.; Kunieda, T. Tetrahedron Lett. 1987, 28, 4185.

A Nucleophilic Alaninol Synthon
J. Am. Chem. Soc., Vol. 121, No. 33, 1999 7515
Table 7. E:Z Ratios of Wittig Reaction Using Aldehydes
Containing R and â-Heteroatoms
sodium benzophenone/ketyl prior to use. Chloroform, hexane, and CH2-
Cl2 were distilled from calcium hydride. Standard benchtop techniques
were employed for handling air-sensitive reagents, and all reactions
were carried out under nitrogen. Flash column chromatography was
performed using Merck 60 silica gel, 230-400 mesh. 1H and 13C NMR
spectra were recorded in CDCl3 at 270 (400) and 65 MHz, respectively.
Gas chromatography was performed on a DB-5 capillary column with
FID detector. Chiral HPLC analyses were performed using a Chiralcel
OD column (Chiral Technologies, Inc.). The minor (cis) isomers from
the Wittig condensation of aromatic aldehydes with ylide 7 were
obtained by column chromatography of reactions quenched at low
temperature (-78 °C).
(S)-(-)-4-Iodomethyl-2-oxazolidinone (5). A mixture of 6.50 g
(24.0 mmol) of tosylate 4 and 18.2 g (112 mmol) of sodium iodide in
100 mL of freshly distilled dry acetone was refluxed under dry nitrogen
until the reaction was judged to be complete by reverse-phase TLC
[Rf(tosylate) ) 0.71, Rf(iodide)) 0.64 in 50:50 EtOAc/hexanes] and/
1
or H NMR (5 h). The reaction was cooled to room temperature, the
excess sodium iodide was filtered off, and the solids were washed with
ethyl acetate. The combined filtrate was concentrated on a rotary
evaporator. The crude yellow solid was dissolved in EtOAc and washed
with a saturated sodium sulfite solution until the layers became colorless.
The aqueous layer was extracted with EtOAc (5 × 50 mL). The
combined organic layers were dried over anhydrous MgSO4 and filtered,
and the solvent was removed under reduced pressure to give 5.72 g of
a tan solid. Column chromatography using 75:25 EtOAc/hexanes
yielded 5.14 g of white solid (95%). The solids may be recrystallized
from an EtOAc/hexanes solution: mp 68-71 °C; Rf 0.53 (100%
a Isolated yield after column chromatography. b Ratios determined
by H NMR.
1
carbonyl carbamate using BOC anhydride. The second step was
the catalytic hydrolysis of the cyclic carbamate by cesium
carbonate (eq 4). This methodology was applied to a sample of
olefins available from the Wittig reaction, with the results shown
in Table 7. The methodology allowed for smooth conversion
of oxazolidinone with either E or Z configurations, with
applicability to simple (Table 7, entry 1) or complex (Table 7,
entry 4) side chains in excellent yields. No isomerization of
the olefin was noticed with either reaction. To verify that there
was no racemization during the Wittig reaction and subsequent
transformations, we converted the primary alcohols to their
corresponding Mosher esters and established their optical purity
to be >97% by 1H and 19F NMR.36 The allylic amino alcohols
serve as precursors for the synthesis of â,γ-unsaturated amino
acids, compounds with interesting biological profiles. Several
methods have been reported for the above transformation.37
1
EtOAc); H NMR (400 MHz, CDCl3) δ 3.24 (m, 2H), 4.12 (m, 2H),
4.53 (app. t, J ) 7, 8 Hz, 1H), 6.43 (s, 1H); 13C NMR (100 MHz,
CDCl3) δ 8.4, 53.0, 70.4, 159.5; IR (CHCl3) 3448, 3255, 1767 cm-1
;
MS (GC-MS) m/e 227 (M+, 4), 100 (80), 86 (100); [R]2D5 -30.2 (c 1.6,
CH2Cl2). Anal. calcd for C4H6INO2: C, 21.16; H, 2.66; N, 6.17.
Found: C, 21.25; H, 2.63; N, 5.95.
(S)-(+)-4-(2-Oxazolidonyl)-methyltriphenylphosphonyl Iodide (6).
A solution of the iodide 5 (15.00 g, 66.08 mmol) and triphenylphosphine
(175.04 g, 668.1 mmol) in 30 mL of dry DMF was stirred at 100 °C
for 61 h. The DMF was removed under vacuum, and the resulting
residue was triturated with dry THF to remove excess triphenylphos-
phine, followed by repeated washings with ether to afford a yellow
solid. This solid was crystallized from MeOH/EtOAc and dried in an
Abderhalden, giving 27.52 g (85%) of a white solid: mp 214-218 °C
1
(decomp); H NMR (400 MHz, D2O) δ 3.79 (m, 1H), 3.96 (m, 1H),
4.04 (app. dd, J ) 9, 5 Hz, 1H), 4.38 (app. dt, J ) 9, 2 Hz, 1H), 4.54
(m, 1H), 7.85 (m, 15H); 13C NMR (65 MHz, D2O) δ 28.3 (d), 47.6,
69.2, 116.9 (d), 130.6 (d), 133.7 (d), 135.4, 157.8; IR (CHCl3) 3228,
3174, 1771 cm-1; [R]2D5 +34.9 (c 2.28, MeOH). Anal. calcd for
C22H21INO2P: C, 54.00; H, 4.32; N, 2.86. Found: C, 53.84; H, 4.41;
N, 2.65.
Conclusions
In summary, we have shown that a new nucleophilic alaninol
synthon is easily prepared from serine and that it undergoes
condensations with aldehydes smoothly and with high stereo-
selectivity. The resulting olefins can be easily converted to
allylic amino alcohols without any racemization of the chiral
center. Additionally, the requirement for a dianionic species to
prevent â-elimination was also established. An application of
the synthon in the enantiospecific synthesis of indolizidine
alkaloid slaframine and its analogues is described in a recent
paper.25
Wittig Reaction: General Procedure. (E)-(R)-(+)-4-(2′-Phenyl-
1′-ethylenyl)-2-oxazolidinone (8). Phosphonium salt 6 (0.489 g, 1.00
mmol) was placed in a flame-dried, 50 mL, three-necked, round-bottom
flask, and the flask was evacuated and filled with an atmosphere of
dry nitrogen (3×). Freshly distilled dry THF (10 mL) was added and
the suspension cooled to -78 °C; n-BuLi (0.79 mL, 1.9 mmol) was
added over a 10 min period to give an orange-colored solution. This
solution was stirred at -78 °C for 1 h, and benzaldehyde (0.093 mL,
0.91 mmol) was added slowly over 5 min. The solution was stirred for
an additional 2.5 h, the dry ice/2-propanol bath was removed, and the
reaction was allowed to warm to room temperature. The reaction was
quenched with a saturated NH4Cl solution (∼10 mL) and stirred for
10 min. THF was removed by rotary evaporation and the aqueous layer
extracted with EtOAc (3 × 15 mL), and the combined organic layers
were dried with anhydrous MgSO4. The drying agent was filtered off
and the organic solvent removed under reduced pressure to leave a
thick yellow oil. Column chromatography using 60:40 EtOAc/hexanes
yielded 0.1480 g of a white solid (86%) as a single isomer (E): mp
Experimental
All reagents were used as received from the supplier. Compounds
2-4 have been previously described (see Supporting Information).10c
Tetrahydrofuran, ether, and 1,2-dimethoxyethane were distilled from
(36) Esterification of the primary alcohol with Mosher acid see: Svatos,
A.; Valterova, I.; Saman, D.; Vrkoc, J. Collect. Czech. Chem. Commun.
1990, 55, 485.
(37) For selected methods for the conversion of amino alcohols to amino
acids: (a) Anelli, P. L.; Biffi, C.; Montanari, F.; Quici, S. J. Org. Chem.
1987, 52, 2559. (b) Pasto, M.; Moyano, A.; Pericas, M. A.; Riera, A.
Tetrahedron: Asymmetry 1995, 6, 2329. (c) Burgess, K.; Liu, L. T.; Pal,
B. J. Org. Chem. 1993, 58, 4758. (d) Beaulieu, P. L.; Duceppe, J.-S.;
Johnson, C. J. Org. Chem. 1991, 56, 4196. (e) See refs 7 and 17b.
1
140-145 °C; Rf 0.47 (50:50 hexanes/EtOAc); H NMR (400 MHz,
CDCl3) δ 4.11-4.15 (m, 1H), 4.52-4.60 (m, 2H), 6.09-6.15 (m, 2H),
6.60 (d, J ) 16 Hz, 1H), 7.26-7.38 (m, 5H); 13C NMR (100 MHz,
CDCl3) δ 55.1, 70.1, 126.4, 126.6, 128.4, 128.7, 133.7, 135.4, 159.6;

7516 J. Am. Chem. Soc., Vol. 121, No. 33, 1999
Sibi et al.
IR (CHCl3) 3459, 1755, 1658, 1602 cm-1; MS (DIP/EI) m/e 189 (M+,
36), 144 (25), 130 (100); [R]2D5 +19.3 (c 1.945, CH2Cl2). Anal. calcd
for C11H11NO2: C, 69.83; H, 5.86; N, 7.40. Found: C, 69.63; H, 5.84;
N, 7.32.
Acknowledgment. We thank NSF (OSR-9108770 and OSR-
9452892), NSF-REU, and North Dakota State University for
providing financial support for this work and the Metabolism
Research Laboratories (USDA-Fargo) for the use of their
polarimeter. Partial support for this work was provided by the
NSF's Instrumentation and Laboratory Improvement Program
through Grant No. USE-9152532.
General Procedure for Hydrolysis of N-(tert-Butoxycarbonyl)-
2-oxazolidinones to N-(tert-Butoxycarbonyl)-amino Alcohols. A 1.0
mmol quantity of the N-BOC-2-oxazolidinone was dissolved in 10 mL
of distilled methanol and stirred at room temperature. A catalytic amount
of cesium carbonate (0.3 mmol) was added, and the reaction was stirred
until TLC indicated that the reaction was complete (typically 3-6 h).
The methanol was removed by rotary evaporation and the crude residue
dissolved in 10 mL of CH2Cl2. The organic solution was placed in a
separatory funnel and washed with 15 mL of a 10% citric acid solution.
The aqueous layer was back extracted with CH2Cl2 (2 × 10 mL), the
combined organics were dried over MgSO4 and filtered, and the organic
solvent was removed by rotary evaporation. This residue was column
chromatographed to give the N-BOC-amino alcohol.
Supporting Information Available: Characterization data
for compounds 9-27 and 37-39 and data for compounds
pertaining to reactions reported in Tables 5 and 6. This material
is available free of charge via the Internet at http://pubs.acs.org.
JA9906249