A convenient synthesis of 2-fluoro- and 6-fluoro-(2S,3R)-threo-(3,4-dihydroxyphenyl)serine using Sharpless asymmetric aminohydroxylation

Article abstract of DOI:10.1016/S0040-4039(01)01816-0

Ring-fluorinated β-hydroxy α-amino methyl esters 3b,c were synthesized enantioselectively using Sharpless asymmetric aminohydroxylation. These were converted to 2-fluoro- and 6-fluoro-(2S,3R)-threo-(3,4-dihydroxyphenyl)serine 1b,c using our previously published procedure.

Full text of DOI:10.1016/S0040-4039(01)01816-0

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 42 (2001) 8401–8403
A convenient synthesis of 2-fluoro- and
6-fluoro-(2S,3R)-threo-(3,4-dihydroxyphenyl)serine using
Sharpless asymmetric aminohydroxylation
In Ho Kim and Kenneth L. Kirk*
Laboratory of Bioorganic Chemistry, National Institutes of Diabetes and Digestive and Kidney Diseases,
National Institutes of Health, Bethesda, MD 20892, USA
Received 5 September 2001; accepted 19 September 2001
Abstract—Ring-fluorinated b-hydroxy a-amino methyl esters 3b,c were synthesized enantioselectively using Sharpless asymmetric
aminohydroxylation. These were converted to 2-fluoro- and 6-fluoro-(2S,3R)-threo-(3,4-dihydroxyphenyl)serine 1b,c using our
previously published procedure. © 2001 Elsevier Science Ltd. All rights reserved.
There is substantial evidence that administered (2S,3R)-
threo-(3,4-dihydroxyphenyl)serine ( -threo-DOPS) 1a
Asymmetric aminohydroxylation of cinnamates nor-
mally produces b-amino a-hydroxy esters as the major
product. This reaction has been used effectively to
prepare such compounds.5 However, Sharpless recently
reported that, in certain cases, reversal of this regioselec-
tivity can be achieved by using the anthraquinone ligands
(DHQ)2AQN and (DHQD)2AQN.6 This observation
provided the basis for our work.
L
crosses the blood–brain barrier and is subsequently
decarboxylated to produce norepinephrine in the central
nervous system (CNS).1,2 Norepinephrine is an impor-
tant neurotransmitter in the CNS and is the principal
neurotransmitter of the sympathetic nervous system. We
have found that 2-fluoro- and 6-fluoronorepinephrine are
selective agonists for b- and a-adrenergic receptors,
respectively.3 These combined observations suggest that
Initially, we used benzylcarbamate as the amine source.
An advantage of this reagent would be that the tris-
O,N,N-benzyl protecting groups of the product could be
removed in one step. Unfortunately, the reaction of the
6-fluoro-cinnamate methyl ester 2c with benzylcarba-
mate in the presence of (DHQD)2AQN ligand afforded
the amino alcohols in only low yield (4%). However, the
desired regioisomer 3c was formed with good regioselec-
tivity (5:1). The reaction of 2c with t-butylcarbamate as
the amine source in CH3CN/H2O (1:1) gave the desired
regioisomer 3c in increased yield (22%), in good regiose-
lectivity (3.7:1) and good enantioselectivity (88% ee).
Although the chemical yield was less than satisfactory,
the fact that the product 3c is identical to the intermediate
prepared in our previous synthesis facilitated structural
determination of regioisomers and enantiomers. Encour-
aged by these results, we next examined the behavior of
2-fluoro-cinnamate methyl ester 2b in the same system.
In this case, the desired regioisomer 3b was formed with
only moderate regioselectivity (2.7:1) but with good
enantioselectivity (84%) and reasonable yield (33%).
2- and 6-fluoro- -threo-DOPS 1b,c may possess interest-
L
ing biological properties and therapeutic potential. In
recognition of this, we recently completed multi-step
syntheses of these analogues based on initial aldol
condensations using a chiral glycine synthon.4 In addi-
tion, we continued our on-going attempts to apply the
Sharpless asymmetric aminohydroxylation to this syn-
thetic problem, since we recognized this as providing a
more direct route with potential for scale-up. Despite
initial and repeated disappointing results, we now report
that, under carefully defined conditions, this procedure
indeed offers a convenient and efficient route to these
analogues (Scheme 1). Moreover, this demonstrates that
normal regiochemical bias can be overcome to provide
access to the important class of phenyl serines from
readily available cinnamates in a highly enantioselective
manner.
Keywords: 2-fluoro- and 6-fluoro-(2S,3R)-threo-(3,4-dihydrox-
yphenyl)serine; Sharpless asymmetric aminohydroxylation; regioselec-
tivity; enantioselectivity.
From these results, we chose to retain the t-butylcaba-
mate as the amine source and to examine effects of
* Corresponding author.
0040-4039/01/$ - see front matter © 2001 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(01)01816-0

8402
I. H. Kim, K. L. Kirk / Tetrahedron Letters 42 (2001) 8401–8403
Series a: R1 =R2 =H
b: R1 =F, R2 =H
c: R1 =H, R2 =F
R1
R1 OH
R1 NHBoc
CO2Me
BnO
BnO
CO2Me
BnO
BnO
CO2Me
BnO
BnO
i
+
NHBoc
OH
R2
R2
R2
4b,c
3b,c
2b,c
R1 OH
R1 OH
-
BnO
BnO
CO2Me
HO
HO
CO2
NH3+Cl-
iii,iv
ii
3b,c
NH3+Cl-
R2
R2
5b: 88%
5c: 82%
1b: 86%
1c: 67%
Scheme 1. Reagents and conditions: (i) K2OsO2(OH)4, t-BuOCl, 1N NaOH, BocNH2, (DHQD)2AQN, organic solvent/H2O (1:1),
rt, 2 h; (ii) HCl, EtOAc, rt, 4 h; (iii) 2N NaOH, MeOH, rt, 1 day; (iv) 10% Pd-C, H2, 3N HCl/MeOH (1:4), rt, 2 h for 1b; 10%
Pd-C, H2, MeOH, rt, 2 h for 1c.
of n-PrOH and CH3CN gave a more modest yield
compared to CH2Cl2 mixtures, but the enantioselectiv-
ity was significantly improved to 78% (entry 11). Dou-
bling the amount of each reagent relative to that used
in the previous entries gave an increased yield (36%)
but a decreased enantioselectivity (68% ee). However, if
this increased portion of reagents was introduced as a
second addition after a 1 h interval, the desired
regioisomer 3c was formed with much improved enan-
tioselectivity (82% ee) and increased yield (38%).
Because of the inconvenience that two separate genera-
tions of N-chloro-N-sodio t-butylcarbamate entails, we
elected to use the conditions described in entry 12
condition for scale-up because of the good yield (54%)
based on recovery of starting material. Since the enan-
tiomeric purity can be increased by subsequent recrys-
tallization, we also felt the enantioselectivity (68% ee)
was acceptable. After carrying out the aminohydroxyla-
tion according to the conditions in entry 12, 3c was
separated readily from its regioisomer 4c. This was
treated with gaseous HCl in ethyl acetate to give free
amino acid methyl ester 5c in 82% yield. Recrystalliza-
tion of the resulting amino acid methyl ester 5c
increased the enantiomeric purity (95% ee).9 Using our
changes in other reaction parameters. Because asym-
metric aminohydroxylation reactions are known to be
sensitive to solvent,7 we studied the effect of solvent
variation on the reaction of 2b with t-butylcarbamate.
From this, we determined that the reaction in PrOH/
H2O (1:1) gave greatly improved results. The desired
regioisomer 3b was obtained as the major isomer with
good regioselectivity (4.5:1), good enantioselectivity
(86%) and good yield (76%). At this point 3b was
separated readily from the regioisomer 3c by silica gel
column chromatography. Treatment of 3b with gaseous
HCl in ethyl acetate, as described in our previous work,
gave amino acid methyl ester 5b.8 Recrystallization of
5b from methanol, ethyl acetate, and ether led to
selective crystallization of the racemate, and the
(2S,3R)-5b recovered from the mother liquor had
increased in enantiomeric purity from 66 to 93% ee.9
We followed our previous procedure with the resulting
enantiomerically purified amino acid methyl ester 5b to
obtain
-threo-2FDOPS 1b,4 identical, except for mag-
L
nitude of optical rotation, with our previously prepared
material:4 ([h]2D5=−18.0 (c 0.14, MeOH) from asymmet-
ric aminohydroxylation; [h]2D5=−15.1 (c 0.65, MeOH)
from asymmetric aldol condensation).
previous procedure, 5c was converted to
L-threo-
With this success we concentrated on the optimization
of the asymmetric aminohydroxylation reaction of 6-
fluoro methyl cinnamate 2c. Unfortunately, the condi-
tions that were effective for the reaction with 2b gave
unsatisfactory results with 2c. The lower solubility of 2c
relative to 2b and extensive formation of the diol
byproduct (11–29%), a complication frequently encoun-
tered with this reaction, were apparent contributing
factors to the low yield of 2c. To improve solubility, we
added CH2Cl2 to the reaction mixture. As shown in
Table 1 (entries 8–10), this resulted in increased yields
but somewhat lower enantioselectivities. A 10:1 mixture
6FDOPS 1c, identical to our previously prepared sam-
ple,4 except for magnitude of optical rotation.
([h]2D5=−20.5 (c 0.073, MeOH) from asymmetric amino-
hydroxylation; [h]2D5=−22.6 (c 0.073, MeOH) from
asymmetric aldol condensation).
In conclusion, we have made both -threo-2FDOPS 1b
L
and -threo-6FDOPS 1c using the Sharpless asymmet-
L
ric aminohydroxylation.6 Although the conversion in
the case of 2c is only modest, the short reaction
sequence makes this the method of choice for the
synthesis of both 1b and 1c.

I. H. Kim, K. L. Kirk / Tetrahedron Letters 42 (2001) 8401–8403
Table 1. Optimization of Sharpless asymmetric aminohydroxylation
8403
Entry
Substrate
Solventa
Yield (%)
3b:4b or 3c:4cb
ee of 3b,c (%)c
1
2
3
4
5
6
7
8
2b
2b
2b
2c
2c
2c
2c
2c
2c
2c
2c
2c
2c
2c
2c
n-PrOH
CH3CN
t-BuOH
n-PrOH
CH3CN
n-BuOH
t-BuOH
n-PrOH:CH2Cl2 (10:1)
n-PrOH:CH2Cl2 (1:1)
n-PrOH:CH2Cl2 (1:10)
n-PrOH:CH3CN (10:1)
n-PrOH:CH3CN (10:1)e
n-PrOH:CH3CN (10:1)f
n-PrOH:CH3CN (1:1)
n-PrOH:CH3CN (1:10)
76
33
4.5:1
2.7:1
4.8:1
3.9:1
3.7:1
4.3:1
3.9:1
4.6:1
4.1:1
6.1:1
4.3:1
3.7:1
4.1:1
4.0:1
4.4:1
86
84
60
68
87.6
44.4
73
60
66
65.6
77.8
67.8
82.4
57.4
84
22 (29)d
24 (56)d
22
35
6 (25)d
38
9
28
10
11
12
13
14
15
40 (42)d
24 (61)d
36 (54)d
38 (66)d
11 (26)d
9 (29)d
a The actual experimental solvent system is this solvent/H2O (1:1).
b This regioselectivity was determined by 300 MHz 1H NMR analysis.
c This enantioselectivity was determined by HPLC [Chiralcel OD; n-hexane/i-PrOH (70:30); flow rate 0.15 mL/min; retention time (2S,3R)-enan-
tiomer of 3b=43.02, (2R,3S)-enantiomer of 3b=38.88, (2S,3R)-enantiomer of 3c=50.40, (2R,3S)-enantiomer of 3c=42.68].
d Yield based on recovered starting material.
e Amounts of regent used in entry 11 were doubled.
f Reagent quantities used in entry 12 were added in two portions at a 1 h interval.
References
2813–2817; (c) Laxma Reddy, K.; Sharpless, K. B. J. Am.
Chem. Soc. 1998, 120, 1207–1217.
1. Thomas, S. A.; Marck, B. T.; Palmiter, R. D.; Matsumoto,
A. M. J. Neurochem. 1998, 70, 2468–2476.
2. Kikuchi, K.; Nashino, K.; Ohyu, H. Brain Res. 2000, 860,
130–135.
3. Lu, S.-f.; Herbert, B.; Haufe, G.; Laue, K. W.; Padgett, W.
L.; Oshunleti, O.; Daly, J. W.; Kirk, K. L. J. Med. Chem.
2000, 43, 1611–1619.
4. Herbert, B.; Kim, I. H.; Kirk, K. L. J. Org. Chem. 2001,
66, 4892–4897.
5. (a) Li, G.; Chang, H.-T.; Sharpless, K. B. Angew. Chem.,
Int. Ed. Engl. 1996, 35, 451–454; (b) Li, G.; Angert, H. H.;
Sharpless, K. B. Angew. Chem., Int. Ed. Engl. 1996, 35,
6. Tao, B.; Schlingloff, G.; Sharpless, K. B. Tetrahedron Lett.
1998, 39, 2507–2510.
7. Rudolph, J.; Sennhenn, P. C.; Vlaar, C. P.; Sharpless, K.
B. Angew. Chem., Int. Ed. Engl. 1996, 35, 2810–2813.
8. Gibson, F. S.; Bergmeier, S. C.; Rapoport, H. J. Org.
Chem. 1994, 39, 3216–3218.
9. This enantiomeric purity was determined by HPLC [Chi-
ralpak AD; 1:1 n-hexane/EtOH+0.2% diethylamine; flow
rate 1 mL/min; retention time (2S,3R)-enantiomer of 5b=
16.14 min, (2R,3S)-enantiomer of 5b=10.38 min, (2S,3R)-
enantiomer of 5c=16.22 min, (2R,3S)-enantiomer of
5c=11.86 min].