The absolute stereochemistry of anachelins, siderophores from the cyanobacterium Anabaena cylindrica

Article abstract of DOI:10.1016/j.tet.2004.07.105

The absolute stereochemistry of anachelins (1 and 2), siderophores isolated from the freshwater cyanobacterium Anabaena cylindrica, was determined via the application of Boc-phenylglycine and Mosher's method. Consequently, it was revealed that a 1,1-dimethyl-3-amino-1,2,3,4-tetrahydro-6,7-dihydroxyquinolinium unit (Dmaq) has 3S (eq.) configuration, and a 6-amino-3,5,7-trihydroxyheptanoic acid unit (Atha) has 3R, 5S, 6S configuration. The 6S configuration of Atha suggested that l-Ser was a biosynthetic precursor of Atha.

Full text of DOI:10.1016/j.tet.2004.07.105

Tetrahedron 60 (2004) 9075–9080
The absolute stereochemistry of anachelins, siderophores from
the cyanobacterium Anabaena cylindrica
*
Yusai Ito, Keishi Ishida, Shigeru Okada and Masahiro Murakami
Laboratory of Marine Biochemistry, Graduate School of Agricultural and Life Sciences, The University of Tokyo, Bunkyo-ku,
Tokyo 113-8657, Japan
Received 7 June 2004; accepted 28 July 2004
Abstract—The absolute stereochemistry of anachelins (1 and 2), siderophores isolated from the freshwater cyanobacterium Anabaena
cylindrica, was determined via the application of Boc-phenylglycine and Mosher’s method. Consequently, it was revealed that a 1,1-
dimethyl-3-amino-1,2,3,4-tetrahydro-6,7-dihydroxyquinolinium unit (Dmaq) has 3S (eq.) configuration, and a 6-amino-3,5,7-trihydroxy-
heptanoic acid unit (Atha) has 3R, 5S, 6S configuration. The 6S configuration of Atha suggested that L-Ser was a biosynthetic precursor of
Atha.
q 2004 Elsevier Ltd. All rights reserved.
1. Introduction
Siderophores are low molecular weight and highly specific
Fe(III) chelating agents synthesized by microorganisms to
sequester iron from the iron-limited environment.1 The
structures of siderophores are diverse and several hundred
structures have been reported from various microorganisms.
Recently, the study of siderophore biosynthesis has been
eagerly pursued, because the ability of siderophores from
pathogenic microorganisms to efficiently assimilate Fe(III)
from host cells has proven to be an important virulence
factor. Therefore, proteins involved in the biosynthesis of
siderophores have potential to be targets of therapeutic
drugs.2 In the view of application for the designed
Anachelin and anachelin-2 (1 and 2) were isolated as the
biosynthesis of natural-product-like molecules, side-
rophores are quite attractive molecules because many of
them are peptide-like molecules3 that are nonetheless
biosynthesized non-ribosomally by large, multidomain
enzymes termed non-ribosomal peptide synthetases
(NRPS).2,4 Furthermore, it has been recently elucidated
that some siderophores, including yersiniabactin from
Yersinia pestis,5 are biosynthesized through the combi-
nation of NRPS and its biosynthetic cousin, polyketide
synthase (PKS).6
first genuine cyanobacterial siderophores from the fresh-
water cyanobacterium Anabaena cylindrica.7,8 The struc-
tures of 1 and 2 have the sequence of three hydrophilic
amino acids (L-Thr-D-Ser-L-Ser) in the middle of molecule
and it has been assumed that D-Ser plays a role to resist
hydrolysis by endogenous proteases. Two unique units
responsible for binding iron are attached through amide
bonds to each terminus of the sequence. One unit is a
6-amino-3,5,7-trihydroxyheptanoic acid unit (Atha). A
2-(2-hydroxyphenyl)-2-oxazoline ring of 1 and 2 functions
as an iron-binding group formed from cyclization of 6-NH2
and 7- or 5-OH of Atha, respectively, with salicylic acid
(Sal). Although the ring system has been found in other
siderophores such as the mycobactins of Mycobacteria,9 it
has usually been derived from cyclization of Ser or Thr.10 In
fact, the long-chain polyhydroxy unit Atha, which enhances
the hydrophilic nature of the molecule, is quite unique in
siderophores that usually consist of small endogenous
Keywords: Anachelin; Siderophore; Iron; Cyanobacteria; Anabaena
cylindrica.
* Corresponding author. Tel.: C81-3-5841-5302; fax: C81-3-5841-8166;
e-mail: aokada@mail.ecc.u-tokyo.ac.jp
0040–4020/$ - see front matter q 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tet.2004.07.105

9076
Y. Ito et al. / Tetrahedron 60 (2004) 9075–9080
organic moieties such as amino acids and diamines.
Furthermore, the structure of Atha strongly suggests a
hybrid of non-ribosomal peptide and polyketide bio-
synthetic origin. The other unit is a 1,1-dimethyl-3-amino-
1,2,3,4-tetrahydro-6,7-dihydroxyquinolinium unit (Dmaq),
of which a catechol group works as another iron-binding
group. The structure of Dmaq is similar to the chromophore
of the pyoverdines, the main siderophores of fluorescent
pseudomonads,11 but the dimethylated quaternary nitrogen
seen in Dmaq has never been reported in any siderophores,
and is reminiscent of quinoline alkaloids from plants.
(TMSCHN2) in dry MeOH to protect the catechol
group,14 2 was hydrolysed in acid condition. HPLC
purification of the hydrolysate furnished the 1,1-dimethyl-
3-amino-1,2,3,4-tetrahydro-6,7-dimethoxyquinolinium unit
(MeDmaq, 3), which was subsequently converted to (R)-
and (S)-BPG amides (4 and 5, respectively). 1H NMR
spectra of 4 and 5 were recorded for calculation of
anisotropic chemical shift differences (the DdZdRKdS)
for each proton. The Dd values for two NMe-1 and H-2 were
positive, while negative Dd values were observed for H-4,
H-5, and OMe-6, which indicated that C-3 possessed S
configuration (Fig. 2). Thus, the absolute configuration of
Dmaq was assigned as 3S (eq).
Thus, Atha and Dmaq units of anachelins (1 and 2) have
unique structural features and their biosynthetic pathway is
also interesting, but the stereochemistry of 1 and 2 remained
to be determined. Herein, we describe the absolute
stereochemistry of Atha and Dmaq of 1 and 2.
2. Results
Anachelins (1 and 2) were prepared from the iron-deficient
culture supernatant of A. cylindrica NIES-19 as described
previously.8 Previous study reported that the oxazoline ring
of 1 and 2 was sensitive to hydrolysis and easily opened to
afford a 5- or 7-salicyl ester of Atha in acidic conditions,
respectively.7,8 Treatment with dilute alkali finally led the
salicyl esters to a salicyl amide of Atha by the O/N acyl
shift.8,12 Since both salicyl amide derivatives prepared from
1 and 2 were identical in spectral analyses, including NMR
(data not shown), the configuration of all chiral centers of 1
and 2 were determined to be identical. Therefore, we used a
suitable sample of 1 or 2 for the determination of the
stereochemistry of each chiral center.
Figure 2. Dd values [Dd (in ppm)ZdR KdS] obtained for (R)- and (S)-BPG
amides of 3 (4 and 5, respectively).
2.2. Atha
To elucidate the relative configuration of C-3/C-5 of Atha, 1
was converted to a diacetonide derivative (6) with 2,2-
dimethoxypropane in DMF. NOESY spectrum of 6 showed
the cross peaks of H-3/H-5, H-3/H-4b, H-5/H-4b, H-3/Me (d
1.42), and H-5/Me (d 1.42) in Atha, indicating that the 1,3-
dioxane ring in Atha existed in chair configuration to show
the relative configuration of C-3/C-5 of Atha to be syn
(Fig. 3). The oxazoline ring of 2 seemed to be convenient to
determine the relative configuration of C-5/C-6 of Atha.
Unfortunately, it was impossible to establish the configur-
ation from the 1H NMR spectrum of 2 recorded in DMSO-
d6, since the signals due to H-7 of Atha and H-3 of two Ser
residues overlapped. After the trial of several solvents,
however, the spectrum recorded in C5D5N gave the
2.1. Dmaq
The relative configuration at C-3 of Dmaq was elucidated by
the combination of J values and NOESY correlations. Large
vicinal coupling constants of H-2a/H-3 and H-3/H-4a (11.7
and 11.1 Hz, respectively) were observed in 1H NMR
spectrum of 1 and NOESY experiment of 1 showed the
cross peak between H-2a and H-4a, indicating that the
secondary amine group was equatorially connected to C-3
of Dmaq (Fig. 1). The absolute configuration at C-3 was
determined by a Boc-phenylglycine (BPG) method,13
because the BPG amides give rise to much higher Dd
values than those given by 2-methoxy-2-trifluoromethyl-2-
phenylacetyl (MTPA) ester of Mosher’s method. After
treatment of 2 with excess trimethylsilyldiazomethane
Figure 1. The relative stereochemistry of Dmaq of 1. Coupling constants
(Hz) and NOESY correlations are shown by plain lines and a dashed arrow,
respectively.
Figure 3. The relative stereochemistry of C3/C5 of Atha of a diacetonide
derivative of 1 (6). NOESY correlations are shown by dashed arrows.

Y. Ito et al. / Tetrahedron 60 (2004) 9075–9080
9077
favorable separation between the signal of H-7 of Atha and
other signals. The irradiation of the signal of H-7 of Atha
changed thatofH-6 tobea distinctdoublet (3JH-5/H-6Z6.4 Hz,
Fig. 4). Previous study described that in a phenyl-oxazoline
ring the large coupling constant (3JHHZ10.0–11.0 Hz)
corresponded to cis form and the small one (3JHHZ6.0–
7.0 Hz) to trans form.15 Therefore, the conformation of the
oxazoline ring of 2 was assigned to be trans, indicating the
relative stereochemistry of C-5/C-6 of Atha to be syn.
consistently fell into regions that lay to the left and right
of the secondary MTPA ester group at C-3 except for that of
Sal (Fig. 5), which was interpreted to be caused by turn of a
branched Sal amide toward C-3. Thus the absolute stereo-
chemistry at C-3 of Atha was confirmed to be R configuration.
Consequently, the absolute stereochemistry at C-3, C-5, and
C-6 of Atha unit were assigned as R, S, and S.
Figure 5. Dd values [Dd (in ppm)ZdSKdR] obtained for 3,7-bis[(S)- and
(R)-MTPA] esters of 9 (10 and 11, respectively).
3. Discussion
Figure 4. 1H NMR spectra of 2 in C5D5N. Top: non-irradiation, bottom:
irradiation of H-7 signal of Atha at dH 4.00.
The stereochemistry of anachelins (1 and 2) was completely
determined in this study. The heterocyclic oxazoline ring is
a class of iron binding group in siderophores and it is usually
derived from cyclization of L-Ser or L-Thr.10 The S
configuration at C-6 of Atha suggested that L-Ser was
incorporated as a precursor in the biosynthesis of Atha.
Furthermore, the structure of Atha also suggested that the
carbon chain from C-1 to C-4 of Atha was incorporated by
condensation of malonyl moieties by PKS. Recently, Walsh
and co-workers revealed yersiniabactin, a siderophore
containing three heterocycles (two thiazolines and a
thiazolidine) from the plague bacterium Yersinia pestis,5
to be produced by NRPS and PKS, which comprises
Sal, three L-Cys, three methyl units from S-adenosyl-
methionine, and a malonyl moiety.6 Yersiniabactin synthe-
tase was the first example of a NRPS/PKS hybrid assemble
and involves two switching points from NRPS to PKS
module. On the basis of suggested biosynthetic scheme of
yersiniabactin, we propose the pathway to 1 that has two
switching points from NRPS to PKS modules as follows
(Scheme 2). At the beginning, L-Ser condenses to Sal on the
first NRPS followed by cyclization to the oxazoline ring.
The two-ring intermediate is elongated by condensation of
The absolute stereochemistry of Atha was elucidated by
Mosher’s method.16 It was found that acidic methanolysis of
1 gave a 7-O-salicyl-Atha methyl ester as a major
degradation product (7, Scheme 1). Then, the methanolysate
including 7 was treated in diluted NaOH solution to yield a
N-salicyl Atha amide (8) by the O/N acyl shift.12 Under
acidic conditions, 8 was spontaneously cyclized to a delta
lactone and therefore the application of Mosher’s method
for the secondary hydroxyl group at C-3 of the delta lactone
was undertaken. However, the expected MTPA diester was
not obtained because of the rapid dehydration between C-2
and C-3 of the delta lactone. Therefore, to prevent from the
spontaneous cyclization, 8 was converted to a methyl ester
(9) with TMSCHN2 again, which also methylated the
phenol group (Scheme 1). Then the methylester (9) was
treated with 2 equiv of each (R)- and (S)-MTPACl in the
presence of DMAP in dry C5H5N, followed by HPLC
purification to furnish the 3,7-bis[(S)- and (R)-MTPA] esters
of 9 (10 and 11, respectively, Scheme 1). The Dd values
(dSKdR) obtained from 1H NMR data of 10 and 11
Scheme 1.

9078
Y. Ito et al. / Tetrahedron 60 (2004) 9075–9080
Scheme 2. Propose biosynthetic scheme for anachelin (1).
two malonyl groups on an intervening PKS module.
Following ketone reduction, the molecule is transferred to
the second NRPS module to attach the peptide sequence
(L-Thr-D-Ser-L-Ser). However, one question arises from
the scheme proposed above; when is the oxazoline of
anachelin-2 (2) cyclized? If the carbon chain of Atha is
elongated by the condensation of malonyl moieties on PKS,
the oxazoline of 2 should be cyclized after the condensation
of the first malonyl unit to Sal-Ser and the following
reduction of its ketone to the hydroxyl group corresponding
to 5-OH of Atha. However, some studies showed that
cyclization to a phenyl-oxazoline or a phenyl-thiazoline
occurred soon after the condensation of Sal and amino acids
(L-Ser or L-Cys, respectively) on NRPS module.2,6 There-
fore, it is reasonable to surmise that 2 is an artificial
compound formed from 1 during culture or purification. In
our previous study, however, the conversion between 1 and
2 was not observed in the acidic solution used during HPLC
purification and the culture supernatant in the fourth day
showed the presence of both 1 and 2 in the ratio of 1:1.8
Moreover, 1 dissolved in culture medium (pH 8.5) was
incubated under a fluorescent light for 1 week, but the
conversion from 1 to 2 was not observed (data not shown).
Therefore, it is assumed that 2 is also a product of
biosynthesis of A. cylindrica NIES-19, but we cannot
exclude the possibility of the conversion between 1 and 2 in
cell and other biosynthetic schemes. The structure of Dmaq
is similar to the chromophore of pyoverdines, the main
siderophore of fluorescent pseudomonads.11 Feeding exper-
iments and genetic studies revealed that Tyr was a precursor
of the chromophore.17 Therefore, it is possible that Tyr is
also incorporated in Dmaq of 1 and 2, and if so, its form
could be L on the basis of the stereochemistry of the
resulting siderophores.
2.49 and dC 39.5. FAB-MS spectra were measured using
glycerol as matrix on a JEOL JMS SX-102 mass
spectrometer.
4.2. Material
Anachelins (1 and 2) were prepared from the iron-deficient
culture supernatant of A. cylindrica NIES-19 as described
previously.8 The obtained 1 and 2 were lyophilized
completely and stored as powder at K20 8C. Boc-
phenylglycine (BPG) was prepared from a commercial
phenylglycine.
4.2.1. MeDmaq (3). To a solution of 2 (11.2 mg) in dry
MeOH (100 mL) and dry benzene (500 mL), 8 equiv of
TMSCHN2 solution (68.8 mL) was added and stirred for
12 h under argon. The reacting mixture was dried in vacuo,
followed by hydrolysis in 6 N HCl at 110 8C for 10 h. The
solvent was removed in a stream of dry N2, and the residue
was applied to a YMC-ODS column (50!150 mm) and
eluted with 0, 25, 50, and 100% MeOH. The concentrated
H2O fraction was purified by HPLC (Cosmosil C18 MS
column, 10.0!250 mm; 0–20% MeCN containing 0.05%
TFA in 20 min; flow rate 2 mL/min, UV detection at
210 nm) to yield MeDmaq (3, 2.4 mg). Retention time
(min): 3 (25.6). HRFAB-MS m/z 237.1612 (MC) calculated
1
for C13H21N2O2 (DK0.8 mmu). H NMR (DMSO-d6), dH
3.56 (s, NMe-1a), 3.65 (s, NMe-1b), 3.66 (m, H-2a), 3.90
(m, H-2b), 4.45 (m, H-3), 2.75 (dd, JZ16.2, 11.1 Hz, H-4a),
2.98 (s, JZ16.2, 5.0 Hz, H-4b), 6.91 (s, H-5), 3.75 (s, OMe-
6), 3.80 (s, OMe-7), 7.40 (s, H-8).
4.2.2. (R)-BPG amide of 3 (4). To solution of 3 (1.0 mg) in
dehydrated DMF (1 mL), 1-ethyl-3-(3-dimethylamino-
propyl)carbodiimide (2.0 mg), 1-hydroxybenzotriazole
(HOBt, 2.0 mg) and (R)-BPG (2.0 mg) were added at 0 8C
and stirred at room temperature for 10 h. The reacting
mixture was dried in vacuo and redissolved in MeOH, and
subjected to HPLC (Cosmosil C18 MS column, 10.0!
250 mm; 32–37% MeCN containing 0.05% TFA in 10 min;
flow rate 2 mL/min, UV detection at 210 nm) to afford (R)-
BPG amide of 3 (4, 1.2 mg). Retention time (min): 4 (12.0).
HRFAB-MS m/z 470.2647 (MC) calculated for
4. Experimental
4.1. Instrumentation
NMR spectra were recorded on a JEOL JNM-A600
1
spectrometer at 27 8C. H and 13C NMR chemical shifts
were referenced to residual solvent peaks of DMSO-d6 at dH

Y. Ito et al. / Tetrahedron 60 (2004) 9075–9080
9079
1
C26H36N3O5 (DK0.8 mmu). H NMR (DMSO-d6), MeD-
210 nm) to yield 8 (1.5 mg). Retention time (min): 8 (18.4).
HRFAB-MS m/z 314.1263 (MCH)C calculated for
C14H20NO7 (DC2.3 mmu). H NMR (DMSO-d6), Sal, dH
6.88 (m, H-4), 7.35 (t, JZ7.7 Hz, H-5), 6.88 (m, H-6), 7.91
(d, JZ7.7 Hz, H-7), Atha, 2.16 (dd, JZ15.0, 8.1 Hz, H-2a),
2.32 (dd, JZ15.0, 3.9 Hz, H-2b), 4.00 (m, H-3), 1.48 (m,
H-4a), 1.53 (m, H-4b), 3.98 (m, H-5), 3.99 (m, H-6), 3.44
(m, H-7a), 3.52 (m, H-7b), 8.34 (d, JZ8.5 Hz, NH).
maq, dH 3.56 (s, NMe-1a), 3.65 (s, NMe-1b), 3.66 (m, H-2a),
3.90 (brd, JZ12.0 Hz, H-2b), 4.45 (m, H-3), 2.78 (dd, JZ
16.2, 11.1 Hz, H-4a), 2.95 (dd, JZ16.2, 5.0 Hz, H-4b), 6.91
(s, H-5), 3.75 (s, OMe-6), 3.80 (s, OMe-7), 7.40 (s, H-8),
7.40 (m, NH), BPG, 5.21 (d, JZ4.2 Hz, H-2), 7.42 (d, JZ
7.5 Hz, H-4), 7.36 (t, JZ7.5 Hz, H-5), 7.32 (t, JZ7.5 Hz,
H-6), 8.61 (d, JZ4.2 Hz, NH), 1.40 (s, Me).
1
4.2.3. (S)-BPG amide of 3 (5). (S)-BPG amide of 3 (5,
1.0 mg) was prepared by the same manner as described
above. Retention time (min): 5 (13.6). HRFAB-MS m/z
470.2652 (MC) calculated for C26H36N3O5 (DK0.3 mmu).
1H NMR (DMSO-d6), MeDmaq, dH 3.54 (s, NMe-1a), 3.53
(s, NMe-1b), 3.51 (m, H-2a), 3.74 (brd, JZ12.0 Hz, H-2b),
4.50 (m, H-3), 2.86 (dd, JZ16.2, 11.1 Hz, H-4a), 3.06 (dd,
JZ16.2, 5.1 Hz, H-4b), 6.95 (s, H-5), 3.78 (s, OMe-6), 3.80
(s, OMe-7), 7.40 (s, H-8). 7.40 (m, NH), BPG, 5.21 (d, JZ
4.2 Hz, H-2), 7.41 (d, JZ7.5 Hz, H-4), 7.36 (t, JZ7.5 Hz,
H-5), 7.32 (t, JZ7.5 Hz, H-6), 8.61 (d, JZ4.2 Hz, NH),
1.40 (s, Me).
4.2.6. Methylester of 8 (9). To solution of 8 (1.5 mg) in
dehydrated MeOH (1.0 mL), excess TMSCHN2 (100 mL,
2.0 M solution) was added and stirred at room temperature
for 24 h. The reacting mixture was directly subjected to
HPLC (Cosmosil C18 MS column, 10.0!250 mm; 30–60%
MeCN containing 0.05% TFA in 15 min; flow rate 2 mL/
min, UV detection at 210 nm) to afford a methylester (9,
1.4 mg). Retention time (min): 9 (18.4). HRFAB-MS m/z
342.1570 (MCH)C calculated for C16H24NO7 (DC
1.8 mmu). 1H NMR (CD3OD), Sal, dH 3.98 (s, OMe),
7.15 (d, JZ7.7 Hz, H-4), 7.50 (t, JZ7.7 Hz, H-5), 7.06 (t,
JZ7.7 Hz, H-6), 7.98 (d, JZ7.7 Hz, H-7), Atha, 3.61 (s,
OMe), 2.43 (dd, JZ15.2, 4.7 Hz, H-2a), 2.52 (dd, JZ15.2,
8.1 Hz, H-2b), 4.24 (m, H-3), 1.66 (dt, JZ14.1, 8.1 Hz, H-
4a), 1.75 (dt, JZ14.1, 5.1 Hz, H-4b), 4.21 (m, H-5), 4.12
(dt, JZ1.7, 6.4 Hz, H-6), 3.70 (d, JZ6.4 Hz, H-7, 2H).
4.2.4. Diacetonide derivative of 1 (6). To solution of 1
(10.4 mg) in dehydrated DMF (1.0 mL), p-TsOH (11.8 mg)
and excess amount of 2,2-dimethoxypropane (500 mL) were
added and stirred at room temperature for 24 h. The reacting
mixture was dried in vacuo and dissolved in H2O, and
subjected to ODS open column chromatography, and eluted
with 0, 20, 50 and 100% MeOH. The 50% MeOH fraction
was lyophilized to obtain a diacetonide derivative (6,
7.5 mg). HRFAB-MS m/z 841.4024 (MC) calculated for
C41H57N6O13 (DC4.0 mmu). 1H NMR (DMSO-d6), Sal, dH
6.96 (d, JZ7.7 Hz, H-4), 7.42 (t, JZ7.7 Hz, H-5), 6.91 (t,
JZ7.7 Hz, H-6), 7.58 (d, JZ7.7 Hz, H-7), Atha, 2.06 (m,
H-2a), 2.35 (m, H-2b), 4.25 (m, H-3), 1.26 (m, H-4a), 1.46
(m, H-4b), 4.12 (m, H-5), 4.33 (m, H-6), 4.33 (m, H-7a),
4.46 (m, H-7b), 1.21 (s, Me), 1.42 (s, Me), Thr, 4.33 (m,
H-2), 3.97 (m, H-3), 1.27 (m, H-4), 7.44 (m, NH), Ser (1),
4.16 (m, H-2), 3.65 (m, H-3a), 3.73 (m, H-3b), 8.50 (d, JZ
5.7 Hz, NH), Ser (2), 4.22 (m, H-2), 3.58 (m, H-3a), 3.64 (m,
H-3b), 8.96 (d, JZ5.6 Hz, NH), Dmaq, 3.51 (s, NMe-1a),
3.60 (s, NMe-1b), 2.97 (m, H-2a), 3.82 (m, H-2b), 4.50 (br,
H-3), 2.85 (dd, JZ15.4, 11.1 Hz, H-4a), 2.95 (dd, JZ15.3,
5.1 Hz, H-4b), 6.56 (s, H-5), 7.27 (s, H-8), 7.83 (d, JZ
7.7 Hz, NH), 1.13 (s, Me), 1.29 (s, Me).
4.2.7. 3,7-Bis[(S)-MTPA] ester of 9 (10). The methylester
9 (400 mg: 1.17 mmol) was dried in a reaction tube which
was made by cutting off a NMR tube (5 mm). And the
solution of DMAP in dehydrated pyridine (100 mL,
3.15 mg/mL, 2.58 mmol) was added to the reaction tube.
(R)-MTPACl solution (0.48 mL, 2.58 mmol) was carefully
added and placed for 2 h at room temperature. The reaction
mixture was lyophilized and dissolved in MeOH and
subjected to HPLC (Cosmosil C18 MS column, 10.0!
250 mm; 76–100% MeCN containing 0.05% TFA in
12 min; flow rate 2 mL/min, UV detection at 210 nm) to
afford a 3,7-bis[(S)-MTPA] ester derivative (10, 500 mg).
Retention time (min): 10 (18.4), 5,7-bis[(S)-MTPA] ester
(18.8), and 3,5,7-tris[(S)-MTPA] ester (22.4). HRFAB-MS
m/z 774.2365 (MCH)C calculated for C36H38NO11 (DC
1.6 mmu). 1H NMR (CD3OD), Sal, dH 3.900 (s, OMe),
7.130 (d, JZ7.7 Hz, H-4), 7.517 (t, JZ7.7 Hz, H-5), 7.052
(t, JZ7.7 Hz, H-6), 7.851 (d, JZ7.7 Hz, H-7), Atha, 3.516
(s, OMe), 2.626 (dd, JZ16.2, 9.0 Hz, H-2a), 2.715 (dd, JZ
16.2, 3.8 Hz, H-2b), 5.635 (m, H-3), 1.807 (m, H-4a), 1.995
(m, H-4b), 3.885 (m, H-5), 4.460 (m, H-6), 4.382 (dd, JZ
11.1, 5.6 Hz, H-7a), 4.614 (dd, JZ11.1, 7.7 Hz, H-7b),
8.540 (br, NH).
4.2.5. N-Salicyl-Atha amide (8). 10% HCl–MeOH solution
(3.0 mL) containing 1 (10 mg) was heated to 100 8C in a
sealed tube for 10 h and then cooled. The reaction mixture
was evaporated and lyophilized to remove traces of HCl.
The HPLC and following spectral analyses of a part of the
resultant products showed that a 7-O-salicyl Atha methyl
ester (7) was a major degradation compound 7; FABMS m/z
328 (MCH)C, 1H NMR (DMSO-d6). Sal, dH 6.88 (m, H-4),
7.35 (t, JZ7.7 Hz, H-5), 6.88 (m, H-6), 7.91 (d, JZ7.7 Hz,
H-7), Atha, 2.16 (dd, JZ15.0, 8.1 Hz, H-2a), 2.30 (dd, JZ
15.0, 3.9 Hz, H-2b), 4.00 (m, H-3), 1.48 (m, H-4a), 1.53 (m,
H-4b), 3.98 (m, H-5), 3.62 (m, H-6), 4.44 (m, H-7a), 4.52
(m, H-7b). Then, the resultant compound was dissolved in
0.02 N NaOH solution and incubated at room temperature
for 1 h, followed by resolution by HPLC (Cosmosil C18 MS
column, 10.0!250 mm; 20–30% MeCN containing 0.05%
TFA in 10 min; flow rate 2 mL/min, UV detection at
4.2.8. 3,7-Bis[(R)-MTPA] ester of 9 (11). A 3,7-bis[(R)-
MTPA] ester derivative (11, 500 mg) was prepared using
(S)-MTPACl by the same procedure as described above.
HRFAB-MS m/z 774.2360 (MCH)C calculated for
1
C36H38NO11 (DC1.1 mmu). H NMR (CD3OD), Sal, dH
3.923 (s, OMe), 7.156 (d, JZ7.7 Hz, H-4), 7.525 (t, JZ
7.7 Hz, H-5), 7.077 (t, JZ7.7 Hz, H-6), 7.894 (d, JZ
7.7 Hz, H-7), Atha, 3.516 (s, OMe), 2.687 (dd, JZ16.2,
9.0 Hz, H-2a), 2.757 (dd, JZ16.2, 3.8 Hz, H-2b), 5.623 (m,
H-3), 1.695 (m, H-4a), 1.900 (m, H-4b), 3.685 (m, H-5),
4.388 (m, H-6), 4.223 (dd, JZ11.1, 5.6 Hz, H-7a), 4.573
(dd, JZ11.1, 7.7 Hz, H-7b), 8.425 (br, NH).

9080
Y. Ito et al. / Tetrahedron 60 (2004) 9075–9080
7. Beiderbeck, H.; Taraz, K.; Budzikiewicz, H.; Walsby,
Acknowledgements
A. E. Z. Naturforsch. 2000, 55c, 681–687.
8. Itou, Y.; Okada, S.; Murakami, M. Tetrahedron 2001, 57,
9093–9099.
We thank Dr. N. Oku (National Cancer Institute at
Frederick) for his advice and discussion of organic
reactions. This work was partially supported by a Grant-
in-Aid for Scientific Research from the Ministry of
Education, Culture, Sports, Science and Technology of
Japan. Y. Ito is financially supported by Research Fellow-
ship of Japan Society for the Promotion of Science for
Young Scientists.
9. Vergue, A. F.; Walz, A. J.; Miller, M. J. Nat. Prod. Rep. 2000,
17, 99–116.
10. (a) Ong, S. A.; Peterson, T.; Neilands, J. B. Agrobactin.
J. Biol. Chem. 1979, 254, 1860–1865. (b) Peterson, T.;
Neilands, J. B. Parabactin. Tetrahedron Lett. 1979, 50,
4805–4808. (c) Griffiths, G. L.; Sigel, S. P.; Payne, S. S.;
Neilands, J. B. Vibriobactin. J. Biol. Chem. 1984, 259,
383–385.
11. Budzikiewicz, H. FEMS Microbiol. Rev. 1993, 104, 209–228.
12. Elliott, D. F. J. Chem. Soc. 1949, 589–594.
References and notes
ˇ ´
13. Seco, J. M.; Quinoa, E.; Riguera, R. J. Org. Chem. 1999, 64,
4669–4675.
1. Neilands, J. B. J. Biol. Chem. 1995, 270, 26723–26726.
2. Quadri, L. E. N.; Sello, J.; Keating, T. A.; Weinreb, P. H.; Walsh,
C. T. Chem. Biol. 1998, 5, 631–645.
14. Hashimoto, N.; Aoyama, T.; Shioiri, T. Chem. Pharm. Bull.
1981, 29, 1475–1478.
15. Futagawa, S.; Inui, T.; Shiba, T. Bull. Chem. Soc. Jpn 1973,
46, 3308–3310.
3. Drechsel, H.; Jung, G. J. Peptide Sci. 1998, 4, 147–181.
4. (a) Gehring, A. M.; Mori, I.; Walsh, C. T. Enterobactin.
Biochemistry 1998, 37, 2648–2659. (b) Keating, T. A.;
Marshall, C. G.; Walsh, C. T. Vibriobactin. Biochemistry
2000, 39, 15513–15521.
16. Ohtani, I.; Kusumi, T.; Kashman, Y.; Kakisawa, H. J. Am.
Chem. Soc. 1991, 113, 4092–4096.
¨
17. (a) Bockmann, M.; Taraz, K.; Budzikiewicz, H. Z. Natur-
forsch. 1997, 52c, 319–324. (b) Mossaialos, D.; Ochsner, U.;
Baysse, C.; Chablain, P.; Pirnay, J.-P.; Koedam, N.;
5. Perry, R. D.; Balbo, P. B.; Jones, H. A.; Fetherstone, J. D.;
Demoll, E. Microbiology 1999, 15, 1181–1190.
´
Budzikiewicz, H.; Fernandez, D. U.; Schafer, M.; Ravel, J.;
¨
6. Miller, D. A.; Luo, L.; Hillson, N.; Keating, T. A.; Walsh, C. T.
Chem. Biol. 2002, 9, 333–344.
Cornelis, P. Mol. Microbiol. 2002, 45, 1673–1685.