Synthesis of 1-N-[(2S,4S)- and (2S,4R)-5-amino-4-fluoro-2- hydroxypentanoyl]dibekacins (study on structure-toxicity relationships)

Article abstract of DOI:10.1016/S0008-6215(97)10080-5

(2S,4S)- and (2S,4R)-5-azido-2-O-benzyl-4-fluoro-2-hydroxypentanoic acids (15 and 19) have been prepared from L-malic acid (1), and coupled to the H₂N-1 group of 3,2',6'-tris(N-benzyloxycarbonyl)-3_-N- (trifluoroacetyl)dibekacin (23), to give, after reduction and deblocking, 1- N[(2S,4S)- and (2S,4R)-5-amino-4-fluoro-2-hydroxypentanoyl]dibekacins (26 and 27). The fluorinated arbekacin analogs showed almost the same antibacterial activities as that of arbekacin, but lower toxicity. Comparision of the toxicity between 26 (and 27) and the arbekacin analogs (28-30) with change of the 1 N-side-chain indicates that the observed decrease in toxicity was a function of the chain length rather than the introduction of flourine.

Full text of DOI:10.1016/S0008-6215(97)10080-5

Carbohydrate Research 306 (1998) 349±360
Synthesis of 1-N-[(2S,4S)- and (2S,4R)-5-amino-
4-¯uoro-2-hydroxypentanoyl]dibekacins (study
on structure±toxicity relationships)
Yoshiaki Takahashi *, Jun Kohno, Tsutomu Tsuchiya *
Institute of Bioorganic Chemistry, 3-34-17 Ida, Nakahara-ku, Kawasaki 211, Japan
Received 17 September 1997; accepted in revised form 1 November 1997
Abstract
(2S,4S)- and (2S,4R)-5-azido-2-O-benzyl-4-¯uoro-2-hydroxypentanoic acids (15 and 19) have
been prepared from l-malic acid (1), and coupled to the H2N-1 group of 3,20,60-tris(N-benzyloxy-
carbonyl)-300-N-(tri¯uoroacetyl)dibekacin (23), to give, after reduction and deblocking, 1-N-
[(2S,4S)- and (2S,4R)-5-amino-4-¯uoro-2-hydroxypentanoyl]dibekacins (26 and 27). The ¯uori-
nated arbekacin analogs showed almost the same antibacterial activities as that of arbekacin, but
lower toxicity. Comparision of the toxicity between 26 (and 27) and the arbekacin analogs (28±30)
with change of the 1N-side-chain indicates that the observed decrease in toxicity was a function of
the chain length rather than the introduction of ¯ourine. # 1998 Elsevier Science Ltd. All rights
reserved
Keywords: 5-Amino-4-¯uoro-2-hydroxypentanoic acid; Arbekacin; Fluorination; Toxicity
1. Introduction
containing analogs of kanamycins [3±5]. However,
contrary to our expectation, the results showed
that no such decrease of toxicity occurred. This
was explained [1] on the basis of a low decrease in
basicity of the H2N-4000 group compared to that of
arbekacin [pKa 10.2 (arbekacin)!8.7]; in other
words, the ÁpKa value (=1.5) did not reach 2.2
(pKa 8.0), the value expected to show a toxicity-
decreasing e€ect under physiological conditions
(pHꢀ7.0). The present study was designed to
examine this concept [1], and describes the syn-
thesis of dibekacin analogs containing 1-N-
[(2S,4S)- and (2S,4R)-5-amino-4-¯uoro-2-hydroxy-
pentanoyl] residues which, compared to AHB, have
a one-carbon extended chain.
In a previous paper [1] we reported the synthesis
of 1-N-[(2R,3R)- and (2R,3S)-4-amino-3-¯uoro-2-
hydroxybutanoyl] analogs of arbekacin {1-N-[(2S)-
4-amino-2-hydroxybutanoyl](AHB)dibekacin} [2]
in the hope of obtaining derivatives with decreased
toxicity. This was based on the hypothesis that
decreasing the basicity of the H2N-4000 group in
arbekacin by introducing ¯ourine, an electron-
withdrawing atom, proximal to the amino group
(in this case, at C-3000) would decrease the toxicity
of arbekacin, as observed in other ¯uorine-
* Corresponding authors.
0008-6215/98/$19.00 # 1998 Elsevier Science Ltd. All rights reserved
PII S0008-6215(97)10080-5

350
Y. Takahashi et al./Carbohydrate Research 306 (1998) 349±360
Scheme 1.
2. Results and discussion
ethereal HCl to give the chloromethyl ketone 4.
Attempted displacement of the terminal chlorine in
4 by an azide group by NaN3 in DMF (at room
temperature) or in hot CH3CN gave the corre-
sponding azidoketone 5 in only a trace amount.
However, treatment of 4 with NaN3±Et4NBr in
CH2Cl2±CH3CN, gave 5 in high yield. If, in this
reaction, CH2Cl2 was omitted, 5 was produced in
only a low yield; the reason is unknown. Next,
conversion of 5 into the corresponding 20-ol was
attempted with the N3 group remaining intact. As
the azido group is known to be comparatively
stable to NaBH4 [8±10], compound 5 was initially
Synthesis.ÐFirst of all, preparation of (2S,4R
and 2S,4S)-5-amino-4-¯uoro-2-hydroxypentanoic
acids was designed according to the retrosynthesis
shown (Scheme 1). After the vicinal carboxyl and
hydroxyl groups of l-malic acid (1) had been
doubly protected with chloral [6,7], the monoacid 2
was re¯uxed with thionyl chloride, giving the corre-
sponding acid chloride 3. One-carbon elongation
was performed by treatment of 3 with diazo-
methane to a€ord an intermediary diazoketone
(R-COC HN2+), which was then treated with

Y. Takahashi et al./Carbohydrate Research 306 (1998) 349±360
351
treated with NaBH4 (NaBH3CN or Me4NBH4) in
MeOH (or THF), but a considerable amount of
undesirable by-products (ninhydrin-negative) were
produced. Next, Zn(BH4)2 in Et2O was tried,
whereupon the oxo group was selectively reduced
with simultaneous cyclization, giving a mixture of
Ç-lactones (6 (major) and 7). The chirality of 6 and
7 at C-4 was determined by NOE experiments; in 6
irradiation of the higher-®eld half of the H-5 pro-
tons caused a signal increase of H-2 together with
the H-3 (H-3a and 3b) protons, indicating the
R-con®guration at C-4. In 7, the S con®guration at
C-4 was deduced from a ROESY experiment for
10, because of the appearance, in 7, of the H-2 and
4 signals at closely proximal positions (Table 1).
Benzylation of the 2-hydroxyl group of 6 and 7 was
performed by a conventional method (PhCH2Br in
DMF in the presence of Ag2O), whereupon, how-
ever, partial epimerization occurred at C-2, giving
a mixture of 2-O-benzyl products (8 and 9 from 6;
10 and 11 from 7) with the 2R isomers pre-
dominating. Each of the mixtures was separated
Table 1
1H and 19F NMR data for compounds 3±13, 15, and 17±19 in CDCl3
Compound
Chemical shifts in ppm (J in Hz)
H-2
H-3a
Ð
H-3b
Ð
H-4
Ð
H-5a
H-5b
Ð
F-4
Ð
PhCH2
Other signals
3
4
5.91 d
2
4.88 dt
0
J5,1 a
Ð
3.52 dd (H-10a)
3.62 dd (H-10b)
J2,5
4
0
J5,1 b
4.89 dt
0
J5,1 a 3.5
0
0
4
J1 a,1 b 19
5.90 d
J2,5
Ð
Ð
Ð
Ð
Ð
Ð
Ð
Ð
Ð
Ð
Ð
Ð
3.23 dd (H-10a)
3.47 dd (H-10b)
2
0
J5,1 b
0
0
4
J1 a,1 b 19
4.14 s (H-30a,30b)
3.10 dd (H-10a)
3.29 dd (H-10b)
5
5.91 d
J2,5
4.88 dt
0
J5,1 a 3.5
2
0
J5,1 b
0
0
4
J1 a,1 b 19
4.05 s (H-30a,30b)
2.68 br s (OH)
6
4.66 dd
2.37 ddd
J3a,3b 13.5
J3a,4 8.5
2.09 dt
J3a,3b 13
J3a,4 10.5
2.28 ddd
J3a,3b 13.5
2.45 ddd
J3b,4 3.5
4.77 dq
J4,5a
J4,5b 3.5
4.57 ddt
3.49 dd
J5a,5b 13
3.69 dd
3.62 dd
3.63 dd
3.54 dd
3.26 dd
Ð
Ð
Ð
Ð
Ð
Ð
Ð
J2,3a
J2,3b
8
9
4
7
4.63 dd
J2,3a 10.5
J2,3b 8.5
4.27 dd
J2,3a 7.5
J2,3b 5.5
4.26 dd
2.68 ddd
6
3.51 dd
3.87 br s (OH)
J3b,4
J4,5a
J4,5b 3.5
4.76 dddd 3.42 dd
6
J5a,5b 13.5
8 and 11
9 and 10
12
2.33 ddd
J3b,4
4.69 d
4.93 d
Jgem 11.5
4.75 d
4.97 d
7
J4,5a
J4,5b 3.5
4
J5a,5b 13.5
J3a,4
2.10 dt
6
2.52 ddd
6
4.47 ddt
3.49 dd
J5a,5b 13
J2,3a 9
J2,3b 8.5
J3a,3b 13
J3a,4
1.96 ddd
J3b,4
J4,5a 6
J4,5b 4.5
4.01 m
9
Jgem 12
4.17 dd
5
1.99 dt
J3b,4 7.5
3.23 dd
J5a,5b 12.5
4.44 d 2.97 d (OH)
4.77 d J4,OH 2.5
Jgem 11 3.77 s (CO2Me)
J2,3a
J3a,3b 14.5
J3a,4 4
J4,5a
J4,5b
6
5
J2,3b 7.5
4.19 dd
J2,3a 11
13
1.90 dddd 2.20 dddd 4.90 dddt 3.37 ddd 3.46 ddd
188.4 dddt 4.45 d 3.77 s (CO2Me)
4.75 d
Jgem 11
J3a,3b 15
J3a,4 2.5
J3a,F 35
J3b,4 10
J3b,F 13
J4,5a
J4,5b
J4,F 49
6
3
J5a,5b 13.5 J5b,F 24
J5a,F 24
J2,3b
3
15
4.23 dd
J2,3a 11
1.93 dddd 2.28 dddd 4.89 dddt 3.39 ddd 3.46 ddd
188.6 dddt 4.51 d
4.79 d
Jgem 11
J3a,3b 15
J3a,4 2.5
J3a,F 35
1.87 ddd
J3a,3b 14.5
J3b,4 10
J3b,F 12.5
J4,5a
J4,5b
J4,F 49
3.98 m
6
3
J5a,5b 13.5 J5b,F
J5a,F 23.5
2
J2,3b
3
17
18
4.25 dd
J2,3a 8.5
1.94 ddd
3.24 dd
3.32 dd
Ð
4.43 d 2.42 d (OH)
4.77 d J4,OH 4
J3b,4 9.5
J4,5a 6.5 J5a,5b 12.5
J4,5b
2.26 dddd 4.87 dddt 3.35 ddd 3.38 ddd
J2,3b
4.10 t
4
J3a,4
3
4
Jgem 11 3.77 s (CO2Me)
4.43 d 3.78 s (CO2Me)
4.76 d
2.10 ddt
J3a,3b 14.5
J3a,4 5.5
J3a,F 25
185.8 ddq
J2,3a 5.5
6
J3b,4 7.5
J3b,F 17
J4,5a
J4,5b
6
3
J5a,5b 13.5 J5b,F 25
J5a,F 23.5
J2,3b
Jgem 11.5
J4,F 48
4.93 dddd 3.36 ddd 3.39 ddd
J4,5a J5a,5b 13.5 J5b,F 24.5
19
4.15 t
J2,3a
J2,3b
2.13 dddd 2.32 ddt
186.0 dddt 4.51 d
4.79 d
Jgem 11.5
6
6
J3a,3b 15
J3a,4 4.5
J3a,F 27
J3b,4
J3b,F 15
8
6
J4,5b 3.5 J5a,F 23.5
J4,F 48

352
Y. Takahashi et al./Carbohydrate Research 306 (1998) 349±360
chromatographically. The structures were con-
1
group, 8 and 10 were each treated with methanolic
HCl, whereupon the desired open-chain esters 12
and 17 were obtained in high yields as syrups
(however, long storage or further puri®cation
caused partial relactonization, giving the starting
materials). Fluorination of 12 or 17 with diethyla-
minosulfur tri¯uoride (DAST) [11] gave the corre-
sponding 4-¯uoro derivatives (13 and 18) with
®rmed by their H NMR spectra (Table 1) as well
as by changes in the speci®c rotations: a change of
6!8 or 7!10 gave a levorotatory change, whereas
a change of 6!9 or 7!11 gave a large dextro-
rotatory change. It is noteworthy that 8 and 11,
1
and 9 and 10, showed identical H NMR signals
and had opposite rotations. Next, to free the HO-4

Y. Takahashi et al./Carbohydrate Research 306 (1998) 349±360
353
Table 2
13C NMR chemical shifts (ꢀ, ppm) and coupling constants (JC,F, Hz) for compounds 6±13, 15, 17±19, 21, and 22 in CDCl3
Compound
C-1
C-2
C-3
C-4
C-5
C-6
Other signals
6
7
177.3
67.0
33.0
33.5
32.5
32.3
36.9
76.0
75.1
76.2
75.0
69.1
54.2
53.6
54.0
53.9
56.4
176.4
173.9
173.8
172.4
68.1
72.9
72.9
76.4
8 and 11
9 and 10
12
72.3a
72.4a
52.2b
72.8a
13
15
17
172.5
176.9
172.8
74.0 d (3.7) 35.7 d (21.0)
73.7 d (3.7) 35.5 d (21.1)
89.0 d (174.2)
88.8 d (174.5)
67.6
54.4 d (21.0)
54.3 d (21.0)
56.8
52.2b 73.1a
73.4a
75.1
36.9
52.1b
72.9a
18
19
21
22
172.1
176.1 d (3.0)
73.7 d (5.6) 35.3 d (21.5)
73.3 d (4.9) 35.0 d (21.2)
89.3 d (173.2)
89.1 d (173.6)
33.8 d (20.9)
34.5 d (22.4)
54.1 d (21.3)
54.0 d (21.5)
85.1 d (173.2)
84.2 d (177.3)
52.2b 72.5a
72.8a
171.3
171.6
71.1 d (5.6)
70.9 d (6.3)
46.4 d (24.3)
45.5 d (26.2)
73.5a
72.7a
a,b
Signals for PhCH2 and COOCH3, respectively.
inversion, accompanied by a small amount of 4-
eno compound 14 (this was a mixture of E and Z
isomers). Alkaline hydrolysis of 13 (or 18) gave the
respective free acids 15 (or 19). To ascertain the
structures of 15 and 19, they were each reduced
catalytically and the amino acids obtained (16 and
20) were cyclized according to the procedure of
Pellegata and coworkers [12], with 1,1,1,3,3,3-hexa-
methyldisilazane to give the corresponding ꢀ-lac-
tams (21 and 22). The structures were con®rmed by
1H NMR spectra and NOE experiments: in 21,
large coupling constants J3,4ax, J4ax,f, and J6ax,f
indicate that BnO-3 (equatorial) and F-5 (axial) are
in opposite faces, and this was supported by the
NOE experiments: irradiation of F-5 caused
indicate that both compounds had activities almost
identical to that of arbekacin. This indicates that the
activity is not in¯uenced by the one-carbon elon-
gation or the orientation of F-4000. It is noteworthy
that the more-active counterpart, 1-N-[(2R,3R)- [not
Table 3
13C NMR chemical shifts (ꢀ, ppm) and coupling constants
(JC,F, Hz) for compounds 26±30 and arbekacin (ABK) in DCl±
D2O (pD 3)
26
27
28
29
30
ABKa
C-1
C-2
C-3
C-4
C-5
C-6
49.6
31.2
49.8
78.6
75.7
81.0
96.0
49.7
21.4
26.2
66.9
43.5
98.9
68.8
56.0
66.4
49.7
31.1
49.8
78.6
75.7
80.7
96.0
49.7
21.4
26.2
66.8
43.4
98.7
68.8
56.0
66.4
49.5 49.6 49.6
31.1 31.2 31.3
49.9 49.9 49.9
78.6 78.6 78.6
75.8 75.7 75.7
81.7 80.9 80.9
96.0 96.0 95.9
49.7 49.7 49.8
21.5 21.5 21.5
26.3 26.3 26.3
66.9 66.9 66.9
43.5 43.5 43.6
99.2 98.8 98.7
68.9 68.9 68.9
56.1 56.1 56.1
66.5 66.5 66.5
73.1 73.1 73.1
60.8 60.8 60.8
49.6
31.2
49.8
78.5
75.7
81.1
95.9
49.7
21.5
26.3
66.9
43.5
98.9
68.8
56.1
66.5
73.0
60.7
4
increases of the H-3ax and H-4eq signals. A H5
4
C-10
C-20
C-30
C-40
C-500
C-600
C-100
C-200
C-300
C-400
C-500
C-6000
C-1
C-2000
C-3000
C-4000
C-5000
C-6000
JC-2
JC-3
JC-4
JC-5
(or S5) structure was proposed as the most prob-
able conformation for 21. In 22, a structure with
the BnO-3 and F-5 in the same face was con®rmed
on account of the observation of an NOE between
H-3 and H-5, although the correct conformation is
not clear. These results indicate that 15 and 19
have the expected structures (Table 2).
Coupling of 15 (or 19) with 3,20,60-tris[N-benzyl-
oxycarbonyl(=Z)-300-N-tri¯uoroacetyl]dibekacin 23
having the H2N-1 group free was performed utiliz-
ing the active ester 24 (or 25) of 15 (or 19) accord-
ing to conventional methodology [1,13]. The
condensed 1-N-acyl derivatives were successively
de(tri¯uoroacetyl)ated, reduced (N3-5000), and
hydrogenolyzed (three Z groups) to give the ®nal
products (26 and 27).
73.0
60.7
73.0
60.7
176.6
68.2d
37.0 d
88.4 d
43.9 d
176.4
68.9 d
36.8 d
89.4 d
43.7 d
173.8 177.1 177.5 176.3
68.7 71.8 72.1
42.9 31.3 33.9
24.1 22.8
70.5
31.7
37.9
40.1 27.3
40.3
000,F
3.5
19.5
170.4
20.3
ꢀ4
000,F
000,F
000,F
19.4
169.6
19.6
Biological activity.ÐAntibacterial activities of 26
and 27 as compared with arbekacin (Table 4)
a
Ref. [1].

354
Y. Takahashi et al./Carbohydrate Research 306 (1998) 349±360
(2R,3S)]-4-amino-3-¯uoro-2-hydroxybutanoyl]di-
bekacin reported previously [1] has an F-3-d
structure. In terms of acute toxicity, both 26
(LD50ꢀ130 mg/kg, mice, intravenous injection)
and 27 (LD50ꢀ125) showed almost the same value,
but this was considerably lower than that of arbe-
kacin (LD50ꢀ75). This indicates that antibacterial
activity and toxicity do not necessarily change in
parallel [4]. This lower toxicity of 26 and 27 as
compared with arbekacin, however, is not to be
ascribed to the decrease in basicity of H2N-5000 that
results from the induction of the F-4000 atom, as had
been predicted in our previous work and was
explained in detail in [1] (for example, arbekacin
and an arbekacin analog having the (2R,3R)-4-
amino-3-¯uoro-2-hydroxybutanoyl residue have
similar toxicities). This unexpected decrease in
toxicity in the present derivatives 26 and 27 must,
therefore, be ascribed to the di€erence in the chain-
length or the position of the ¯uorine atom intro-
duced.
To clarify the problem, three arbekacin analogs
with varying lengths of the side chain, without
¯uorine, were conventionally prepared, namely,

Y. Takahashi et al./Carbohydrate Research 306 (1998) 349±360
355
Table 4
Minimal inhibitory concentrationa (ꢁg/mL) of compounds 26±30 and arbekacin (ABK)
Test organismb
26
27
28
29
0.10
<0.05
>100
30
0.20
0.10
>100
ABK
0.20
0.05
>100
St. a. FDA 209 P
St. a. Smith
St. a. Ap01c
0.10
<0.05
>100
0.10
<0.05
>100
0.39
0.05
>100
Micr. l. FDA16
Micr. l. PCI 1001
B. c. ATCC 10702
Coryn. b. 1810
E. c. NIHJ
E. c. K-12 R5d
E. c. K-12 ML1629e
E. c. K-12 ML1410 R81e
E. c. K-12 LA290 R55f
E. c. K-12 LA290 R64
E. c. W677
E. c. JR66/W677f,g
E. c. JR225h
Kl. p. PCI602
Kl. p. 22#3038f,g
Sh. s. JS11746
Sal. e. 1891
3.13
3.13
0.78
1.56
0.39
1.56
1.56
0.78
0.39
0.39
3.13
3.13
1.56
0.78
0.39
6.25
6.25
1.56
3.13
0.39
>100
>100
3.13
3.13
1.56
0.39
0.39
1.56
50
1.56
>100
6.25
25
3.13
1.56
1.56
6.25
3.13
3.13
3.13
3.13
50
100
3.13
>100
>100
>100
>100
>100
0.78
1.56
3.13
0.39
0.39
1.56
0.78
0.39
0.78
0.78
3.13
25
1.56
6.25
25
0.39
1.56
1.56
1.56
3.13
0.39
0.39
1.56
1.56
0.78
0.78
0.78
3.13
3.13
3.13
1.56
1.56
1.56
1.56
0.39
0.39
1.56
0.78
1.56
0.78
0.78
3.13
3.13
3.13
3.13
1.56
1.56
0.78
0.39
0.39
1.56
0.78
0.78
0.78
0.78
6.25
12.5
1.56
6.25
25
0.39
1.56
1.56
1.56
1.56
0.39
0.39
1.56
0.78
0.78
0.78
0.78
3.13
6.25
6.25
1.56
6.25
0.39
1.56
6.25
Serr. marc.
Prot. r. GN311
Prov. sp. Pv 16i
Prov. sp. 2991i
Ps. aerug. A3
Ps. aerug. H9g
Ps. aerug. GN315d
50
>100
0.78
6.25
>100
12.5
25
0.39
1.56
6.25
0.78
1.56
6.25
25
25
Judged by the agar dilution streak method (Mueller±Hinton agar, 37 ꢁC, 18 h).
a
b Abbreviations: St. a., Staphylococcus aureus; Micr. l., Micrococcus luteus; B. c., Bacillus cereus; Coryn. b., Corynebacterium bovis;
E. c., Escherichia coli; Kl. p., Klebsiella pneumoniae; Sh. s., Shigella sonnei; Sal. e., Salmonella enteritidis; Serr. mar., Serratia mar-
cescens; Prot. r., Proteus rettgeri; Prov., Providencia; Ps. aerug., Pseudomonas aeruginosa.
c Resistant strain producing AAD(40).
AAC(60).
d
e APH(30)-I.
AAD(200).
f
APH(30)-II.
AAC(3).
g
h
0
i
AAC(2 ).
1-N-[(2S)-3-amino-2-hydroxypropanoyl]dibekacin
(28) [14], 1-N-[(2S)-5-amino-2-hydroxypentanoyl]-
dibekacin (29), and 1-N-[(2S)-6-amino-2-hydroxy-
hexanoyl]dibekacin (30) (Table 3). The antibacterial
activities (Table 4) of compounds 28, 29, and
arbekacin were found to be almost the same,
whereas 30 was less active. However, it was estab-
lished that the acute toxicity of these compounds
was signi®cantly in¯uenced by the chain-length,
with decrease in toxicity with increase in length (28:
ꢀ65, 29: ꢀ120, 30: ꢀ125 mg/kg). A speci®c feature
is that 26, 27, and 29 with the same chain-length,
had analogous toxicities, indicating that the
decrease in toxicity of 26 (and 27) compared to
arbekacin must be ascribed to the di€erence in
chain length, and not to the introduction of ¯uor-
ine at C-4000.
3. Experimental
General methods.ÐMelting points were deter-
mined on a Ko¯er block and are uncorrected.
Optical rotations were determined with a Perkin±
Elmer 241 polarimeter. IR spectra were measured
with a Jasco A-202 grating spectrophotometer. 1H,
13C, and 19F NMR spectra were recorded at 250
(1H, for 3±5 and 7) and 235.3 MHz (19F) with a
Bruker WM 250 spectrometer, and at 500 (1H),
125.8 (13C), and 470.6 MHz (19F, for 21) with an
1
AMX 500 spectrometer. Chemical shifts (ꢀ) of H,
13C, and 19F spectra were measured down®eld
from internal Me4Si (for H and 13C) and internal
1
Freon 11 (for 19F), unless otherwise stated, and
were con®rmed, when necessary, by shift-correlated
2D spectra. Thin-layer chromatography (TLC) was

356
Y. Takahashi et al./Carbohydrate Research 306 (1998) 349±360
performed on Kieselgel 60 F254 (Merck), and col-
umn chromatography, on Wakogel C-200, unless
stated otherwise. HPLC was performed on a
C.I.G. pre-packed silica-gel column (Kusano
Kagakukikai Co., Japan) using 1:3 hexane±EtOAc.
(2R,5S)-5-(2-Chloro-2-oxoethyl)-2-trichloro-
methyl-1,3-dioxolan-4-one (3).ÐA mixture of 2
5-Azido-3,5-dideoxy-l-erythro-pentono-1,4-lactone
(6) and 5-azido-3,5-dideoxy-d-threo-pentono-1,4-
lactone (7).ÐTo a cold ( 10 ꢁC) solution of 5
(1.95 g, 6.45 mmol) in THF (60 mL) was added
ꢀ0.07 M ethereal Zn(BH4)2 (®rst 30 mL, and then
another 30 mL after 1 h; ꢀ4 mmol in total), and the
mixture was kept for 2 h in the cold. TLC (1:1
toluene±EtOAc) of the organic layer showed spots
at Rf 0.25 (6), 0.2 (7), and 0 (cf. 5: Rf 0.7). After
addition of AcOH (ꢀ0.8 mL), the mixture was
extracted with EtOAc. The extracts were washed
with aq Na2SO4 (satd), dried (Na2SO4), and con-
centrated to a syrup. HPLC of the syrup a€orded 6
[6,7] [11.9 g, 45 mmol; [ꢂ]d +38.5ꢁ (c 5, EtOH),
23
[ꢂ]d +35ꢁ (c 1, CHCl3); lit [6], [ꢂ]d +39.1ꢁ (c
5.14, EtOH), lit [7], [ꢂ]d20+33.03 (c 1.005, CHCl3)]
and SOCl2 (20 mL, 270 mmol) was gently re¯uxed
for 60 h. Evaporation of SOCl2 gave a residue that
crystallized from hexane to give 3 as needles
(11.8 g, 93%); mp 74±75 ꢁC (lit [6] 70±72 ꢁC),
23
28
(565 mg, 56%) and 7 (357 mg, 35%) as syrups.
137ꢁ (c 1, CHCl3); IR
[ꢂ]d24+24ꢁ (c 1, CHCl3); IR (KBr): 1780 (CˆO),
Compound 6 had [ꢂ]d
22
1
1
1820 cm
C6H4Cl4O4: C, 25.56; H, 1.43; Cl, 50.30. Found: C,
25.79; H, 1.50; Cl, 50.19.
[CˆO (lactone)]. Anal. Calcd for
(neat): 1780 (CˆO), 2110 cm (N3); NOE di€er-
ence spectroscopy: irradiation of H-5a, H-2
(0.4%), H-3a (0.6%), H-3b (1.2%), and H-4
(1.8%) signals were increased. Anal. Calcd for
C5H7N3O3: C, 38.22; H, 4.49; N, 26.74. Found: C,
(2R,5S)-5-(3-Chloro-2-oxopropyl)-2-trichloro-
methyl-1,3-dioxolan-4-one (4).ÐTo a solution of 3
(10.0 g, 35 mmol) in Et2O (100 mL) was added
0.25 M ethereal CH2N2 (280 mL, 70 mmol), and the
solution was kept for 30 min at room temperature.
Ethereal 1.2 M HCl (70 mL) was added, and the
mixture was kept for a further 30 min. TLC (6:1
toluene±EtOAc) of the solution showed a spot at
Rf 0.55. After concentration, the residue was chro-
matographed (6:1 toluene±EtOAc) to give 4 as a
crystalline solid (9.8 g, 93%); mp 75±76 ꢁC
(toluene±hexane), [ꢂ]d23+33ꢁ (c 1, CHCl3); IR
22
38.32; H, 4.31; N, 26.41. Compound 7 had [ꢂ]d
+59ꢁ (c 1, CHCl3); IR (neat): 1780, 2110 cm .
1
Anal. Calcd for C5H7N3O3: C, 38.22; H, 4.49; N,
26.74. Found: C, 38.23; H, 4.75; N, 26.63.
5-Azido-2-O-benzyl-3,5-dideoxy-l-erythro-pent-
ono-1,4-lactone (8) and 5-azido-2-O-benzyl-3,5-
dideoxy-l-threo-pentono-1,4-lactone (9).ÐTo
a
solution of 6 (395 mg) in DMF (8 mL) were added
Ag2O (1.16 g) and excess PhCH2Br (0.6 mL), and
the mixture was stirred for 3 h at room tempera-
ture; TLC (4:1 toluene±EtOAc) of the organic layer
showed two spots at Rf 0.55 (8, major) and 0.45 (9)
(cf. 6: Rf 0.05). Filtration followed by concentra-
tion in vacuo gave a syrup, which was extracted
with CHCl3. The extracts were washed with 0.05 M
aq HCl and water, dried (Na2SO4), and con-
centrated. The residue was chromatographed (8:1
toluene±EtOAc) to give 8 (503 mg, 81%) and 9
1
(KBr): 1735 (CˆO), 1815 cm [CˆO (lactone)];
13C NMR (CDCl3; con®rmed by the 1H-13C
HMBC): ꢀ 198.8 (C-20), 170.3 (C-4), 105.2 (C-2),
97.7 (CCl3), 70.5 (C-5), 47.5 (C-30), 40.7 (C-10).
Anal. Calcd for C7H6Cl4O4: C, 28.41; H, 2.04; Cl,
47.92. Found: C, 28.67; H, 1.87; Cl, 47.98.
(2R,5S)-5-(3-Azido-2-oxopropyl)-2-trichloro-
methyl-1,3-dioxolan-4-one (5).ÐTo a cold (0 ꢁC)
solution of 4 (6.0 g, 20 mmol) in 6:5 CH2Cl2±
CH3CN (110 mL) were added NaN3 (1.7 g,
26 mmol) and Et4NBr (4.5 g, 21 mmol), and the
mixture was stirred for 3 h in the cold; TLC (6:1
toluene±EtOAc) of the organic layer showed a spot
at Rf 0.45. After dilution with CHCl3 (600 mL), the
solution was washed with water, dried (Na2SO4),
and concentrated. The residue was puri®ed by
chromatography (6:1 toluene±EtOAc) to give 5 as
a crystalline solid (5.28 g, 86%), mp 105±106 ꢁC
(toluene±hexane), [ꢂ]d23+38ꢁ (c 1, CHCl3); IR
(KBr): 1725, 1830, 2110 cm 1 (N3). Anal. Calcd for
C7H6Cl3N3O4: C, 27.79; H, 2.00; Cl, 35.16; N,
13.89. Found: C, 28.09; H, 2.24; Cl, 35.37; N, 13.71.
21
(68 mg, 11%) as syrups. Compound 8 had [ꢂ]d
160ꢁ (c 1, CHCl3). Anal. Calcd for C12H13N3O3:
C, 58.29; H, 5.30; N, 16.99. Found: C, 58.17; H,
5.31; N, 16.75. Compound 9 had [ꢂ]d
5ꢁ (c 1,
20
CHCl3). Anal. Calcd for C12H13N3O3: C, 58.29; H,
5.30; N, 16.99. Found: C, 58.05; H, 5.26; N, 16.77.
5-Azido-2-O-benzyl-3,5-dideoxy-d-threo-pentono-
1,4-lactone (10) and 5-azido-2-O-benzyl-3,5-dideoxy-
d-erythro-pentono-1,4-lactone (11).ÐTo a solution
of 7 (1.08 g) in DMF (22 mL) were added Ag2O
(3.20 g) and excess PhCH2Br (1.64 mL), and the
mixture was stirred for 3 h at room temperature.
TLC (4:1 toluene±EtOAc) of the organic layer
showed two spots at Rf 0.45 (10, major) and 0.55

Y. Takahashi et al./Carbohydrate Research 306 (1998) 349±360
(11). The mixture was treated as ®rst described 1.5 Hz. Compound 13, [ꢂ]d
357
74ꢁ (c 1, CHCl3).
24
above to give 10 (1.36 g, 80%) and 11 (160 mg, 9%)
as syrups. Compound 10 had [ꢂ]d +4ꢁ (c 1,
CHCl3); the H NMR spectrum was the same as
Anal. Calcd for C13H16FN3O3: C, 55.51; H, 5.73;
F, 6.75; N, 14.94. Found: C, 55.42; H, 5.81; F,
6.92; N, 14.93.
24
1
that of 9; in ROESY, a cross peak was observed
between H-2 and H-4. Anal. Calcd for
C12H13N3O3: C, 58.29; H, 5.30; N, 16.99. Found:
(2S,4S)-5-Azido-2-O-benzyl-4-¯uoro-2-hydroxy-
pentanoic acid (15).ÐTo a solution of 13 (1.80 g)
in MeOH (36 mL) was added M aq NaOH (9 mL),
and the solution was kept for 1 h at room tem-
perature. TLC (3:1 CHCl3±MeOH) of the solution
showed a single spot at Rf 0.5 (cf. 13: Rf 0.9). The
solution was concentrated to a low volume, diluted
with water, and acidi®ed with M aq HCl to pH ꢀ1
under cooling. Extraction of the product with
EtOAc followed by concentration gave 15 as a
C, 58.54; H, 5.37; N, 17.23. Compound 11 had
ꢁ
20
1
[ꢂ]d +161 (c 1, CHCl3); the H NMR spectrum
was the same as that of 8. Anal. Calcd for
C12H13N3O3: C, 58.29; H, 5.30; N, 16.99. Found:
C, 58.00; H, 5.52; N, 16.69.
Methyl (2S,4R)-5-azido-2-O-benzyl-2,4-dihydroxy-
pentanoate (12).ÐA solution of 8 (4.58 g) in 0.1 M
methanolic HCl (75 mL) was kept for 10 min at
room temperature. TLC (4:1 toluene±EtOAc) of
the solution showed two spots at Rf 0.35 (12) and
0.55 (8, minor). After addition of CHCl3 (500 mL),
the solution was washed with water, dried
(Na2SO4), and concentrated. Chromatography (6:1
toluene±EtOAc) of the syrup gave 12 as a syrup
(4.25 g, 82%) together with 8 recovered (640 mg).
Compound 12 had IR (neat): 1740 (CˆO),
crystalline solid (1.70 g, 99%), mp 99±100 ꢁC
24
(CHCl3±hexane), [ꢂ]d
81ꢁ (c 1, CHCl3). Anal.
Calcd for C12H14FN3O3: C, 53.93; H, 5.28; F, 7.11;
N, 15.72. Found: C, 53.74; H, 5.18; F, 6.86; N,
15.92.
Methyl (2S,4S)-5-azido-2-O-benzyl-2,4-dihydroxy-
pentanoate (17).ÐA solution of 10 (1.56 g) in
0.1 M methanolic HCl (25.5 mL) was kept for
10 min at room temperature. TLC (4:1 toluene±
EtOAc) of the solution showed two spots at Rf 0.4
(17) and 0.45 (10, minor). Processing as described
for 12 gave a syrup, chromatography (6:1 toluene±
EtOAc) of which gave 17 as a syrup (1.41 g, 80%)
together with 10 recovered (230 mg). Compound 17
had IR (neat): 1740 (CˆO), 2110 cm 1 (N3).
Methyl (2S,4R)-5-azido-2-O-benzyl-4-¯uoro-2-
hydroxypentanoate (18).ÐTo a solution of DAST
(1.4 mL) in 8:1 benzene±pyridine (30 mL) was
added a solution of 17 (1.00 g) in benzene (20 mL),
and the mixture was treated as described for 13.
TLC (10:2:1 cyclohexane±CHCl3±acetone) of the
solution showed spots at Rf 0.45 (14), 0.35 (18,
major), 0.2 (trace), and 0.1 (a mixture of 17 and
several by-products). Similar puri®cation as
described for 13 gave 18 as a syrup (545 mg, 54%),
1
2110 cm (N3).
Methyl (2S,4S)-5-azido-2-O-benzyl-4-¯uoro-2-
hydroxypentanoate (13) and methyl (S)-5-azido-2-
O-benzyl-2-hydroxy-4-pentenoate (14).ÐTo a solu-
tion of DAST (2.0 mL, 15.1 mmol) in 8:1 benzene±
pyridine (45 mL) was added a solution of 12
(1.50 g, 5.37 mmol) in benzene (30 mL), and the
mixture was kept for 1 h at room temperature and
then for 2 h at 60 ꢁC. TLC (10:2:1 cyclohexane±
CHCl3±acetone) of the solution showed spots at Rf
0.45 (14), 0.35 (13, major), 0.2 (trace), and 0.1. The
mixture was poured into aq NaHCO3 (satd,
300 mL) and toluene (300 mL), and after shaking
for 10 min, the organic layer that separated was
washed with water, dried (Na2SO4), and con-
centrated. Chromatography (10:2:1 cyclohexane±
CHCl3±acetone) of the residue gave 13 as a syrup
together with 14 (125 mg). Compound 18 had
60ꢁ (c 1, CHCl3). Anal. Calcd for
1
23
(1.16 g, 77%), along with syrupy 14 (84 mg), H
[ꢂ]d
NMR (CDCl3): (the ratio of the E and Z isomers
was ꢀ1:2) ꢀ 6.23 [dt, 1 H, H-5 (Z)], 5.94 [dt, 1 H,
H-5 (E)], 5.34 [dt, 1 H, H-4 (E)], 4.94 [q, 1 H, H-4
(Z)], 4.44 and 4.71 [each d of 1 H, J 12 Hz, PhCH2
(Z)], 4.43 and 4.72 [each d of 1 H, J 12 Hz, PhCH2
(E)], 3.98 [dd, 1 H, H-2 (Z)], 3.95 [dd, 1 H, H-2
(E)], 3.75 (s, COOMe), 2.58 [dddd, 1 H, H-3b (Z)],
2.54 [ddt, 1 H, H-3a (Z)], 2.49 [m, 2 H, H-3a,3b
(E)]; J (E isomer): J2,3a 7, J2,3b 5.5, J3a,4&J3b,4 7.5,
J4,5 13.5, J3a,5&J3b,5ꢀ1.5 Hz; (Z isomer): J2,3a 7,
J2,3b 5.5, J3a,3b 15, J3a,4&J3b,4&J4,5 7.5, J3a,5&J3b,5
C13H16FN3O3: C, 55.51; H, 5.73; F, 6.75; N, 14.94.
Found: C, 55.63; H, 5.79; F, 6.98; N, 15.15.
(2S,4R)-5-Azido-2-O-benzyl-4-¯uoro-2-hydroxy-
pentanoic acid (19).ÐTo a solution of 18 (350 mg)
in MeOH (7 mL) was added M aq NaOH (1.8 mL),
and the solution was kept for 1 h at room tem-
perature. TLC (3:1 CHCl3±MeOH) of the solution
showed a single spot at Rf 0.5 (cf. 18: Rf 0.9).
Similar treatment as described for 15 gave 19 as a
syrup (315 mg, 95%). An analytical sample (syrup)
was prepared by column chromatography using

358
Y. Takahashi et al./Carbohydrate Research 306 (1998) 349±360
the lower phase of 20:1:1 CHCl3±MeOH±20% aq
49ꢁ (c 1, CHCl3). Anal. Calcd for
C12H14FN3O3: C, 53.93; H, 5.28; F, 7.11; N, 15.72.
Found: C, 54.15; H, 5.37; F, 7.12; N, 15.87.
J5,F 49, J5,6a 4, J5,6b 5 Hz, J5,NH ꢂ 0.5, H-5; irra-
diation of NH caused collapse to ddddd), 4.73 and
4.99 (each d of 1 H, J 12 Hz, PhCH2), 3.88 (dd, 1
H, J3,4a 9, J3,4b 7 Hz, H-3), 3.54 (ddddd, 1 H, J6b,4a
1, J6b,5 5, J6b,F 14, J6b,6a 13.5, J6b,NH 4 Hz, H-6b),
3.43 (ddddd, 1 H, J6a,4b ꢂ 1, J6a,5 4, J6a,6b 13.5,
J6a,F 24, J6a,NH 3 Hz, H-6a), 2.52 (ddddd, 1 H, J4b,3
7, J4b,4a 14.5, J4b,5 6, J4b,6a ꢂ 1, J4b,F 22 Hz, H-4b),
2.22 (ddddd, 1 H, J4a,3 9, J4a,4b 14.5, J4a,5 5.5, J4a,6b
1, J4a,F 21 Hz, H-4a); NOE di€erence spectroscopy:
irradiation of H-3 increased the signals of H-4b
(3.3%), H-5 (1.9%), and H-6a (2.2 %). 19F NMR
(CDCl3): ꢀ 178.3 (dddq; JF,NH ꢀ1.5 Hz). Anal.
Calcd for C12H14FNO2: C, 64.56; H, 6.32; F, 8.51;
N, 6.28. Found: C, 64.52; H, 6.29; F, 8.62; N, 6.35.
N-Hydroxysuccinimide esters (24 and 25) of 15
and 19.ÐA mixture of 15 (or 19) (1.0 mmol), N-
hydroxysuccinimide (1.02 mmol), and N,N0-dicy-
clohexylcarbodiimide (1.0 mmol) in dry EtOAc
(5.5 mL) was stirred for 1 h at room temperature.
The resultant precipitate was ®ltered o€, washed
with EtOAc, and the combined ®ltrate and wash-
ings were concentrated to give syrupy 24 (or 25),
which showed a single spot at Rf 0.45 in TLC (4:1
toluene±EtOAc), and was used without further
puri®cation.
22
AcOH, [ꢂ]d
(3S,5S) - 3 - Benzyloxy - 5 - ¯uoro - 2 - piperidinone
(21).ÐA solution of 15 (200 mg) in 3:1 MeOH±
H2O (8 mL) was hydrogenated under H2 in the
presence of Pd-black for 40 min at room tempera-
ture. TLC (1:1 CHCl3±MeOH) of the solution
showed a single spot at Rf 0.2 (cf. 15: Rf 0.65).
Filtration followed by concentration gave 16 as a
pale-yellow solid (168 mg, 93%). A mixture of the
solid in CH3CN (1.7 mL), 1,1,1,3,3,3-hexa-
methyldisilazane (3.1 mL), and 1.2 M ethereal
HCl (0.6 mL; prepared by introducing HCl vapor
into Et2O) was re¯uxed overnight. TLC (1:3
toluene±EtOAc) of the solution showed a spot at
Rf 0.25. Addition of MeOH (8 mL) followed by
concentration gave a residue, which was puri®ed
by chromatography (1:3 toluene±EtOAc) to give
21 as a crystalline solid (90 mg, 58%), mp 119±
120 ꢁC (toluene±hexane), [ꢂ]d
115ꢁ (c 1,
22
1
1
CHCl3); IR (KBr): 1630 cm (amide); H NMR
(CDCl3): ꢀ 5.95 (br s, 1 H, NH), 5.07 (dddq, 1 H,
J5,4ax 3.5, J5,4eq 5, J5,F 48, J5,6ax 4, J5,6eq 3, J5,NH
ꢀ1 Hz, H-5), 4.72 and 5.02 (each d of 1 H, J 12 Hz,
PhCH2), 4.13 (dd, 1 H, J3,4ax 9, J3,4eq 5.5 Hz, H-3),
3.62 (dddd, 1 H, J6ax,5 4, J6ax,F 34, J6ax,6eq 13.5,
J6ax,NH 2 Hz, H-6ax), 3.52 (dddt, 1 H, J6eq,4eq 2,
J6eq,5 3, J6eq,F 18, J6eq,6ax 13.5, J6eq,NH 3 Hz, H-
6eq), 2.47 (ddddd, 1 H, J4eq,3 5.5, J4eq,4ax 14.5,
J4eq,5 5, J4eq,6eq 2, J4eq,F 12.5 Hz, H-4eq), 2.19
(dddd, 1 H, J4ax,3 9, J4ax,4eq 14.5, J4ax,5 3.5, J4ax,F
34 Hz, H-4ax); NOE di€erence spectroscopy: irra-
diation of F increased the signals of H-3 (27%), H-
4eq (21%), H-5 (100%; taken as the reference for
the increases), and H-6eq (23%). 19F NMR
1-N-[(2S,4S)-5-Amino-4-¯uoro-2-hydroxypenta-
noyl]dibekacin (26).ÐTo a solution of 23 [13]
(716 mg, 0.75 mmol) in 2:1 THF±H2O (40 mL) was
added 24 (580 mg, ꢀ1.6 mmol) in THF (10 mL)
and, after the pH had been adjusted to ꢀ8 by
addition of aq NaHCO3 (satd), the solution was
kept for 1 h at room temperature. Concentration
gave a residue, which was washed with water and
EtOAc. The solid obtained (632 mg) was dissolved
in M NH3 in 2:1 THF±H2O (36 mL) and the solu-
tion was kept for 40 h at room temperature [de(tri-
¯uoroacetyl)ation]. Concentration gave a residue,
which was dissolved in 30:15: 1 1,4-dioxane±H2O±
AcOH (28 mL) and hydrogenated in the presence
of Pd-black for 5 h. After ®ltration, the ®ltrate was
concentrated, and the solid was chromatographed
on a CM Sephadex C-25 column (aq 0!0.15 M
NH3) to give 26 as a solid, which was dried
thoroughly in vacuo (0.2ꢀ1 mmHg) in a desiccator
(CDCl3): ꢀ
182.8 (dddt). Anal. Calcd for
C12H14FNO2: C, 64.56; H, 6.32; F, 8.51; N, 6.28.
Found: C, 64.77; H, 6.24; F, 8.30; N, 6.43.
(3S,5R) - 3 - Benzyloxy - 5 - ¯uoro - 2 - piperidinone
(22).ÐA solution of 19 (150 mg) in 3:1 MeOH±
H2O (6 mL) was hydrogenated in a similar manner
as described for 21. TLC (1:1 CHCl3±MeOH) of
the solution showed a single spot at Rf 0.2 (cf. 19:
Rf 0.65). Post-treatment as described for 16 gave 20
as a pale-yellow solid (128 mg, 95%). The solid was
then treated as described for 21 to give 22 as a
for 3 days in the presence of P2O5 (225 mg, 43%);
ꢁ
23
1
[ꢂ]d +71 (c 1, H2O); H NMR (DCl-D2O, pD
3): ꢀ (signals relating to the side chain are shown
mainly) 5.83 (d, 1 H, H-10), 5.22 (d, 1 H, H-100),
5.11 (double multiplets, 1 H, H-4000), 4.41 (dd, 1 H,
H-2000), 2.28 (dddd, 1 H, H-3000b), 1.93 (dddd, 1 H,
crystalline solid (73 mg, 62%); mp 85.5±86.5 ꢁC
23
(toluene±hexane), [ꢂ]d
83ꢁ (c 1, CHCl3); IR
1
1
(KBr): 1645 cm (amide); H NMR (CDCl3): ꢀ
6.09 (br s, 1 H, NH), 4.94 (m, 1 H, J5,4a 5.5, J5,4b 6,
H-3000a); J1 ,2 &J1 ,2 3.8, J2 ,3 a 11, J2 ,3 b 3,
0
0
00 00
000 000
000 000

Y. Takahashi et al./Carbohydrate Research 306 (1998) 349±360
359
000
000
000
000
000
000
000
000
J3 a,3 b 15, J3 a,4 2.5, J3 b,4 10, J3 a,F 37, J3 b,F
000
3.34 mmol) as described for 28 to give 29 as a solid
(1.45 g, 67%); [ꢂ]d +82ꢁ (c 0.5, H2O); H NMR
(DCl±D2O, pD 3): ꢀ 5.83 (d, 1 H, H-10), 5.22 (d, 1
H, H-100), 4.20 (dd, 1 H, H-2000), 3.09 (slightly br t, 2
H, J 7.5 Hz, H-5000a,5000b), 2.29 (dt, 1 H, H-2eq),
ꢀ1.93 (H-3000b), 1.91 (q, 1 H, H-2ax), 1.77±1.95 (H-
4000a,4000b), 1.62±1.74 (m, 2 H, H-40ax, 3000a);
24
1
13, J4 ,F ꢀ51 Hz. 19F NMR (DCl-D2O, Freon 11
as the external reference, pD 3): ꢀ 192.5 (m).
.
.
Anal. Calcd for C23H45FN6O10 1.5 H2CO3 H2O:
C, 42.30; H, 7.24; F, 2.73; N, 12.08. Found: C,
42.11; H, 7.30; F, 2.78; N, 11.98.
1-N-[(2S,4R)-5-Amino-4-¯uoro-2-hydroxy-
pentanoyl]dibekacin (27).ÐCompound 23 (530 mg,
0.56 mmol) was treated with 25 (410 mg, 1.1 mmol)
0
0
00 00
000 000
000 000
J1 ,2 &J1 ,2 3.8, J2 ,3 a 9, J2 ,3 b 3.5 Hz. Anal.
.
.
Calcd for C23H46N6O10 H2CO3 H2O: C, 44.57; H,
7.79; N, 13.00. Found: C, 44.47; H, 7.92; N, 12.55.
1-N-[(S)-6-Amino-2-hydroxyhexanoyl]dibekacin
(30).Ð(S)-6-(Benzyloxycarbonyl)amino-2-hydroxy-
hexanoic acid [15,18] (1.50 g, 5.33 mmol) was trea-
in the manner described for 26 to give 27 as a solid
ꢁ
22
1
(178 mg, 49%); [ꢂ]d +81 (c 1, H2O); H NMR
(DCl-D2O, pD 3): ꢀ 5.82 (d, 1 H, H-10), 5.22 (d, 1
H, H-100), 5.15 (double multiplets, 1 H, H-4000), 4.42
(dd, 1 H, H-2000), 2.25 (ddt, 1 H, H-3000b), 2.18 (ddt,
ted
with
N-hydroxysuccinimide
(625 mg,
1 H, H-3000a); J1 ,2 &J1 ,2 3.8, J2 ,3 a 7.5, J2 ,3 b
5.43 mmol) in the manner described for 24 to give
the active ester as a syrup. Subsequently the syrup
was treated with 23 (3.20 g, 3.37 mmol) as descri-
0
0
00 00
000 000
000 000
000
000
000
000
000
000
000
4.5, J3 a,3 b 15, J3 a,4 7.5, J3 b,4 4.5, J3 a,F 18.5,
000
J3 b,F 30, J4 ,F ꢀ50 Hz. 19F NMR (DCl±D2O,
000
25
Freon-11 as the external reference, pD 3): ꢀ 190.3
bed for 28 to give 30 as a solid (1.40 g, 63%), [ꢂ]d
(m). Anal. Calcd for C23H45FN6O10 H2CO3 0.5
H2O: C, 43.96; H, 7.38; F, 2.90; N, 12.82. Found:
C, 43.94; H, 7.51; F, 3.04; N, 13.07.
+80ꢁ (c 1, H2O); H NMR (DCl±D2O, pD 3): ꢀ
1
.
.
5.85 (d, 1 H, H-10), 5.24 (d, 1 H, H-100), 4.20 (dd, 1
H, H-2000), 3.07 (slightly br t, 2 H, J 7.5 Hz, H-
6000a,6000b), 2.31 (dt, 1 H, H-2eq), 1.93 (q, 1 H, H-
2ax), 1.88 (dddd, 1 H, H-3000b), 1.76 (m, 2 H, H-
5000a,5000b), ꢀ1.68 (H-3000a), 1.55 (m, 2 H, H-
1-N-[(S)-3-Amino-2-hydroxypropanoyl]dibekacin
(28).ÐA mixture of (S)-N-(benzyloxycarbonyl)-
isoserine [15,16] (1.20 g, 5.02 mmol), N-hydroxy-
succinimide (580 mg, 5.04 mmol), and N,N0-dicy-
clohexylcarbodiimide (1.04 g, 5.04 mmol) in THF
(25 mL) was stirred for 2 h at room temperature.
The precipitate was ®ltered o€ and washed
thoroughly with THF. The combined ®ltrate and
washings were concentrated to give the active ester
of the protected isoserine as a syrup. Subsequently
the syrup was treated with 23 (2.74 g, 2.88 mmol) as
described for 26 to give a solid, which was sub-
jected to chromatography on a column of CM-
Sephadex C-25 (aq 0!0.5 M NH3) to give 28 as a
solid (920 mg, 53%), [ꢂ]2d5+89ꢁ (c 1, H2O) (lit
4000a,4000b); J1 ,2 &J1 ,2 3.8, J2 ,3 a 9, J2 ,3 b 4,
0
0
00 00
000 000
000 000
000
000
000
000
000
000
J3 a,3 b 14, J3 b,4 a 6.5 (or 10), J3 b,4 b 10 (or
.
6.5) Hz. Anal. Calcd for C24H48N6O10 H2-
.
CO3 H2O: C, 45.45; H, 7.93; N, 12.72. Found: C,
45.71; H, 7.73; N, 12.52.
Acknowledgements
The authors express deep thanks to both Dr.
Tomio Takeuchi of Institute of Microbial Chem-
istry and Meiji Seika Kaisha, Ltd. for the measure-
ment of acute toxicities. The authors are also
grateful to Dr. Masa Hamada of the Institute of
Microbial Chemistry for the measurement of anti-
bacterial spectra.
1
[14], no data reported); H NMR (DCl±D2O, pD
3): ꢀ 5.84 (d, 1 H, H-10), 5.21 (d, 1 H, H-100), 4.52
(dd, 1 H, H-2000), 3.44 (dd, 1 H, H-3000b), 3.28 (dd, 1
H, H-3000a), 2.32 (dt, 1 H, H-2eq), 1.94 (q, 1 H, H-
2ax); J1,2ax&J2ax,2eq&J2ax,3 13, J1,2eq&J2eq,3 4.5,
0
0
00 00
000 000
000 000
000
000
J1 ,2 &J1 ,2 3.8, J2 ,3 a 8.5, J2 ,3 b 4, J3 a,3 b
.
13 Hz. Anal. Calcd for C21H42N6O10 H2CO3: C,
References
43.70; H, 8.00; N, 13.90. Found: C, 43.40; H, 7.68;
N, 14.10.
[1] Y. Takahashi, C. Ueda, T. Tsuchiya, and Y.
Kobayashi, Carbohydr. Res., 249 (1993) 57±76.
[2] S. Kondo, K. Iinuma, H. Yamamoto, K. Maeda,
and H. Umezawa, J. Antibiot., 26 (1973) 412±415.
[3] Y. Kobayashi, T. Tsuchiya, T. Ohgi, N. Taneichi,
and Y. Koyama, Carbohydr. Res., 230 (1992) 89±
105.
1-N-[(S)-5-Amino-2-hydroxypentanoyl]dibekacin
(29).Ð(S)-5-(Benzyloxycarbonyl)amino-2-hydroxy-
pentanoic acid [15,17,18] (1.50 g, 5.61 mmol) was
treated with N-hydroxysuccinimide (660 mg,
5.73 mmol) in the manner described for 24 to give
the corresponding active ester as a syrup. Subse-
quently the syrup was treated with 23 (3.17 g,
[4] T. Shitara, Y. Kobayashi, T. Tsuchiya, and S.
Umezawa, Carbohydr. Res., 232 (1992) 273±290.

360
Y. Takahashi et al./Carbohydrate Research 306 (1998) 349±360
[12] R. Pellegata, M. Pinza, and G. Pi€eri, Synthesis,
[5] T. Tsuchiya, T. Shitara, S. Umezawa, T. Takeuchi,
M. Hamada, N. Tomono, and E. Umemura, Car-
bohydr. Res., 240 (1993) 307±312.
[6] H. Eggerer and C. Grunewalder, Liebigs Ann.
Chem., 677 (1964) 200±208.
[7] M. Baron, J. Mol. Structure, 12 (1972) 71±78.
[8] Y. Ali and A.C. Richardson, Carbohydr. Res., 5
(1967) 441±448.
[9] T. Sheradsky, in S. Patai (Ed.), The Chemistry of
the Azido Group, Interscience, New York, 1971,
Chap. 6, pp 313±395.
[10] K. Soai, S. Yokoyama, and A. Ookawa, Synthesis,
(1987) 48±49.
[11] L.N. Markovskij, V.E. Pashinnik, and A.V. Kirsa-
nov, Synthesis, (1973) 787±789; M. Hudlicky, Org.
React., 35 (1988) 513±637; J.A. Wilkinson, Chem.
Rev., 92 (1992) 505±519.
(1978) 614±616.
[13] T. Tsuchiya, Y. Takagi, and S. Umezawa, Tetra-
hedron Lett., (1979) 4951±4954.
[14] H. Umezawa, K. Maeda, S. Kondo, and S. Ume-
zawa, Chem. Abstr., 82 (1975) 86637r.
[15] T.H. Haskell, R. Rodebaugh, N. Plessas, D. Wat-
son, and R.D. Westland, Carbohydr. Res., 28
(1973) 263±280.
[16] U. Schmidt, R. Meyer, V. Leitenberger, F. Stabler,
and A. Lieberknecht, Synthesis, (1991) 409±413.
[17] H. Saeki, Y. Shimada, Y. Ohashi, S. Sugawara, and
E. Ohki, Chem. Pharm. Bull., 22 (1974) 1151±1158.
[18] D.S. Karanewsky, M.C. Badia, D.W. Cushman,
J.M. DeForrest, T. Dejneka, M.J. Loots, M.G.
Perri, E.W. Petrillo, Jr., and J.R. Powell, J. Med.
Chem., 31 (1988) 204±212.