A practical synthesis of the phytosiderophore 2′-deoxymugineic acid: A key to the mechanistic study of iron acquisition by graminaceous plants

Article abstract of DOI:10.1002/anie.200702403

(Chemical Equation Presented) Ironing out plant uptake: The phytosiderophores mugineic acid (MA) and deoxymugineic acid (DMA) were synthesized in a few steps with minimum use of protecting groups and workup/purification procedures (see scheme; Boc: tert-butoxycarbonyl) and their potencies in transporter-mediated iron(III) acquisition were tested. The sufficient supply of these compounds will enable study of the molecular mechanism of iron acquisition and utilization by graminaceous plants.

Full text of DOI:10.1002/anie.200702403

Communications
DOI: 10.1002/anie.200702403
Iron Acquisition
A Practical Synthesis of the Phytosiderophore 2’-Deoxymugineic Acid:
A Key to the Mechanistic Study of Iron Acquisition by Graminaceous
Plants**
Kosuke Namba, Yoshiko Murata, Manabu Horikawa, Takashi Iwashita, and Shoichi Kusumoto*
Dedicated to Professor Yoshito Kishi on the occasion of his 70th birthday
Crops grown on alkaline soils, which cover about one-third of
the worldꢀs land, are prone to iron-deficiency stress because of
the low solubility of iron there. To acquire the insoluble iron
efficiently, some graminaceous plants have developed a
unique strategy characterized by the synthesis and secretion
of an iron-chelator phytosiderophore and by a system for the
specific uptake of iron(III) in its complex.[1] Mugineic acid
(MA, 1) was first identified as a phytosiderophore in barley[2]
and its analogues have since been isolated and identified from
various graminaceous species and cultivars;[3] they all form
water-soluble 1:1 complexes with iron(III). Owing to its
significant implication in plant physiology, this iron-uptake
system has been the subject of intensive research since its
discovery 30 years ago.[4] However, the mechanistic details of
the iron acquisition have not yet been fully elucidated. As a
part of our ongoing research program to clarify this mecha-
nism at the molecular level, we have recently identified and
characterized the MA·iron(III) transporter HvYS1 from
barley, which is the species most tolerant to iron deficiency
among graminaceous plants, and we have elucidated the
specific localization and function of the encoded protein in
the plant.[5] Large-scale expression of the transporter itself
and a more detailed mechanistic study thus became our next
targets.
The limited supply of phytosiderophores has been a
severe bottleneck, particularly for the study of the transport
mechanism. To drive the research forward, it is therefore
essential to establish an ample supply source of MA (1) and/
or 2’-deoxymugineic acid (DMA, 2), the latter of which is a
phytosiderophore for rice, wheat, and maize[6] with a similar
function of transporting iron(III). In addition, a supply of MA
and/or DMA will be potentially important in order to solve
the worldwide problem of shortage of food supplies because
the use of these compounds may make cultivation possible
even on alkaline soils that are not favorable for farming.[7]
Herein, we report an efficient, short synthesis of DMA (2),
which enables the use of this compound in large quantity for
biological studies. The potency of DMA in assisting specific
iron(III) transport through the transporter HvYS1 was
quantitatively estimated. A comparison of the potency of
synthetic DMA (2) with those of MA (1) and its diastereomer,
that is, 2’-epi-mugineic acid (2’-epi-MA, 3), which were
similarly synthesized in this work, clearly proved that these
three compounds exhibit the same level of iron-transport
ability. New access is, hence, open for the elucidation of the
mechanism of phytosiderophore-mediated iron acquisition by
plants by using synthetic 2 and its derivatives as probes. The
potential value of DMA as a fertilizer or iron supplement can
also now be investigated.
The molecular structure of DMA consists of three parts
corresponding to a-hydroxybutyric acid, a-aminobutyric acid,
and azetidine-2-carboxylic acid. For the preparation of a large
quantity of DMA, we chose reductive amination as a key
reaction in the assembly of the three amino acid derivatives,
because a judicious choice of building blocks and reaction
conditions was expected to enable sequential transformations
with minimal requirements for workup and purification
procedures. Several syntheses of MA and DMA have been
reported based on the same principle of reductive amina-
tion.[8] There were, however, two major setbacks in the
previous syntheses: the extensive use of protecting groups and
the cumbersome purification at each step. Another route was
also used to assemble the three building blocks, by peptide
coupling followed by reduction of the peptide bonds,[9] where
the efficiency of synthesis was not satisfactory either. We
[*] Dr. K. Namba, Dr. Y. Murata, Dr. M. Horikawa, Dr. T. Iwashita,
Prof. Dr. S. Kusumoto
Suntory Institute for Bioorganic Research
1-1-1 Wakayamadai, Shimamoto, Mishima
Osaka, 618-8503 (Japan)
Fax: (+81)75-962-2115
E-mail: skus@sunbor.or.jp
[**] We thankDr. Y. Ito of our institute for measuring the high-resolution
mass spectra and Prof. J. F. Ma of Okayama University for providing
natural mugineic acid. This workwas supported by Grants-in-Aid for
Scientific Research (Grant nos.: 18710191 and 18510200) from the
Ministry of Education, Culture, Sports, Science and Technology,
Japan. We acknowledge Suntory Co., Inc., for financial support.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
7060
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 7060 –7063

Angewandte
Chemie
therefore aimed to develop a truly practical synthesis of
DMA with minimal use of protecting groups.
We began the synthesis with commercially available Boc-
l-allylglycine (4; Scheme 1). After ozonolysis of 4 in meth-
favor of the unnatural a configuration of the hydroxy group.
To obtain an acceptable diastereoselectivity, the allylic
oxidation was investigated with several other allylglycine
derivatives. Among them, oxidation of Cbz-l-allylglycine tert-
butyl ester by selenium oxide resulted in the formation of an
approximately 2:1 mixture of the 2-hydroxy allylglycine 10 in
a modest yield (Scheme 2). Even though the diastereoselec-
Scheme 1. One-pot synthesis of a protected 2’-deoxymugineic acid (9)
from Boc-l-allylglycine (4) and deprotection to give free 2’-deoxymugi-
neic acid (DMA, 2). Boc=tert-butoxycarbonyl.
Scheme 2. Synthesis of mugineic acid (MA, 1) and its 2’-epimer (2’-epi-
MA, 3). Cbz=benzyloxycarbonyl.
anol at ꢀ788C, the mixture containing 5 was directly treated
with NaBH3CN and l-azetidine-2-carboxylic acid and
warmed to room temperature to give 6. After evaporation
of the reaction solvent (methanol), the residual 6 was treated
with dry hydrogen chloride in ethanol. Clean cleavage of the
Boc group proceeded in the presence of the remaining excess
NaBH3CN and the two carboxylic acid groups were con-
currently converted into ethyl esters to give 7. The reaction
mixture was evaporated to remove excess hydrogen chloride
and the residue was treated with a solution of aldehyde 8[10]
and NaBH3CN in methanol. The second reductive amination
also proceeded smoothly to give protected DMA 9. By
extraction of 9 with an organic solvent at this stage, all of the
water-soluble byproducts, including amino acids and the
excess reducing agent and its degradation products, can be
readily removed. Subsequent normal flash chromatography
on silica gel afforded the product in a pure state. The
protected DMA was obtained in a good overall yield from
commercially available 4 through a one-pot synthesis and
with just a single chromatography process at the final step.
The isolation/purification of any intermediates was not
required before that. Finally, the deprotection of 9 gave
DMA in quantitative yield. After deionization by ion-
exchange resin, recrystallization from EtOH/MeOH/H2O
gave free DMA (2).
tivity at the 2-position was still in favor of the unnatural
configuration, the mixed compound 10 was subjected to the
subsequent reactions toward the mugineic acid derivatives, as
shown in Scheme 2. Ozonolysis of 10 followed by reductive
amination with free l-azetidine-2-carboxylic acid gave the
desired compound 11 in a good yield as a mixture of
diastereomers. Hydrogenolytic deprotection of 11 and sub-
sequent direct reductive amination with aldehyde 8 after
removal of the palladium catalyst resulted in a mixture of 12
and 13 in high yield. The diastereomers 12 and 13 were then
separated by HPLC (CAPCELLPAK C18-UG80; 35%
CH3CN in H2O containing 1% AcOH). The isolated diaste-
reomers were separately treated with 6m hydrochloric acid to
give 2’-epi-MA (3) and MA (1), respectively, in quantitative
yields. Separation of the deprotected 2’-epimers, 1 and 3, was
difficult even by HPLC under various conditions, so the
separation had to be carried out before the final removal of
the tert-butyl esters.
With the sufficient quantities of MA (1), DMA (2), and 2’-
epi-MA (3) in hand, their capacities for iron(III)-complex
transport through the HvYS1 transporter expressed in
Xenopus laevis oocytes[5] were examined. We have reported
some results of structure–activity studies for this family of
phytosiderophores.[11] Yet the effect of the hydroxy group at
the C2’ position was not known. Thus, we evaluated the role
of this particular hydroxy group in iron acquisition by
measuring electrophysiological transport activity in Xenopus
laevis oocytes. HvYS1 is a specific phytosiderophore–iron(III)
transporter identified from iron-deficient barley roots and
belongs to the oligopeptide transporter (OPT) family.[12] The
open reading frame (ORF) region of HvYS1 cDNA was
amplified by PCR. In the oocytes injected with cRNA
Success in the above synthesis of DMA prompted us to
apply the same strategy to the synthesis of MA (1), the supply
of which has also been greatly required for functional studies.
The same starting material, Boc-l-allylglycine, was expected
to give 1 after conversion into 2-hydroxy-Boc-l-allylglycine
through allylic oxidation and subsequent application of the
same series of reactions.
Allylic oxidation of Boc-l-allylglycine ethyl ester by
selenium oxide proceeded smoothly. However, this oxidation
resulted in the formation of a 9:1 diastereomeric mixture in
Angew. Chem. Int. Ed. 2007, 46, 7060 –7063
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
7061

Communications
encoding the ORF region, currents were induced by iron(III)
Experimental Section
complexes of synthetic compounds 1–3 at a level similar to
that with natural MA–iron(III) (Figure 1). Currents were
absent in water- or noninjected control oocytes.
9: Ozone was bubbled through a solution of Boc-l-allylglycine (1.2 g,
5.7 mmol) in methanol (40 mL) at ꢀ788C until the color of the
solution changed to blue. After nitrogen was bubbled through until
the blue color had gone, l-azetidine-2-carboxylic acid (564 mg,
5.7 mmol) and NaBH3CN (351 mg, 5.7 mmol) were added to the
solution. The mixture was stirred for 2 h at room temperature and
then concentrated under reduced pressure. A suspension of the
residue in cooled anhydrous HCl/EtOH (prepared from acetyl
chloride (4.5 mL) and ethanol (100 mL)) was stirred for 2 h at 08C,
stirred for 15 h at room temperature, concentrated under reduced
pressure, dehydrated by toluene azeotropy, and dried under vacuum
for several hours. Aldehyde 8 (1.3 g, 5.7 mmol) and NaBH3CN
(351 mg, 5.7 mmol) were added to a mixture of the residue in
methanol (50 mL) at room temperature. The mixture was stirred for
4 h, quenched with sat. NaHCO3, and extracted with ethyl acetate (3 ”
200 mL). The combined organic layers were dried over anhydrous
MgSO4, filtered, and concentrated under reduced pressure. The
residue was purified by flash column chromatography on silica gel
(elution with hexane/ethyl acetate (2:1)!ethyl acetate including
0.1% triethylamine) to give protected DMA 9 (1.44 g, 55%) as
colorless oil: 1H NMR (400 MHz, CDCl3): d = 4.22–4.12 (m, 4H), 3.91
(t, J = 6.4 Hz, 1H), 3.58 (t, J = 8.8 Hz, 1H), 3.39 (brt, J = 6.0 Hz, 1H),
3.25 (brt, J = 6.0 Hz, 1H), 2.82 (q, J = 7.6 Hz, 1H), 2.75–2.61 (m, 2H),
2.59–2.55 (m, 2H), 2.32 (m, 1H), 2.21 (m, 1H), 1.80–1.58 (m, 4H),
1.45 (s, 9H), 1.27 (t, J = 7.2 Hz, 3H), 1.26 (t, J = 7.2 Hz, 3H), 1.16 ppm
(s, 9H); 13C NMR (100 MHz, CDCl3): d = 174.38, 173.36, 172.07,
80.07, 74.11, 69.68, 64.77, 60.16, 60.11, 59.21, 55.06, 50.68, 43.71, 33.93,
30.65, 27.56, 27.49, 20.96, 13.95, 13.86 ppm; HRMS: m/z calcd for
Figure 1. Iron-transporting activities of natural MA, synthetic MA, 2’-
epi-MA, and DMA. The activity was assessed by two-electrode voltage-
clamp analysis with Xenopus laevs oocytes. a) Electrogenic iron(III)-
complex transport by HvYS1. Each compound was added as the
iron(III) complex at a concentration of 50 mm (indicated by the black
bars). b) The currents relative to that with natural MA–iron(III) are
shown. The error bars show the standard deviation (number of
experiments: 5–8). n.d.: not detected.
C24H44N2O7 [M+H]+: 473.3229; found: 473.3221; [a]2D0 = ꢀ92.8 (c =
+
1.0, CHCl3).
2: A suspension of 9 (600 mg, 1.2 mmol) in cooled 6m HCl
(20 mL) was stirred for 10 h at room temperature and concentrated
under reduced pressure. A mixure of the residue in 1m NaOH
(30 mL) was stirred for 10 h at room temperature. The mixture was
neutralized by 1m HCl and concentrated under reduced pressure. The
residue was purified by column chromatography on an ion-exchange
resin (Dowex 50W” 8) (elution with water!1m NH3) to give the
DMA ammonium salt (360 mg, 99%) as a white solid. Further
purification by recrystallization (from H2O/MeOH/EtOH) gave pure
2 (260 mg, 71%) as a white solid: 1H NMR (300 MHz, D2O): d = 4.63
(t, J = 9.6 Hz, 1H), 4.01 (dd, J = 6.6, 4.5 Hz, 1H), 3.97 (td, J = 9.6,
4.2 Hz, 1H), 3.83 (app. q, J = 9.6 Hz, 1H), 3.64 (dd, J = 7.8, 3.6 Hz,
1H), 3.38–3.18 (m, 2H), 3.14–2.96 (m, 2H), 2.60 (qd, J = 9.6, 4.2 Hz,
1H), 2.41 (dt, J = 11.7, 9.3 Hz, 1H), 2.18–1.70 ppm (m, 4H); 13C NMR
(75 MHz, D2O): d = 180.05, 173.21, 172.17, 70.32, 66.90, 59.61, 51.24,
50.50, 44.25, 30.45, 24.54, 21.12 ppm; HRMS: m/z calcd for
These results clearly show that iron complexes of these
synthetic compounds possess essentially the same transport
activity as that of natural MA. The 2’-hydroxy group of MA is
hardly responsible for the iron-chelating or transport activity.
We also obtained similar results with oocytes injected with
cRNA encoding ZmYS1, which was found in maize, was
reported as the first example of an iron(III)–phytosidero-
phore transporter,[13] and is known to transport various
phytosiderophore-bound metals including Zn2+, Cu2+, and
Ni2+ (see the Supporting Information).[14]
The above result that the presence or absence of the 2’-
hydroxy group has no definite effect on the iron-transport
ability of these phytosiderophores means that DMA, which is
now available in sufficient amounts, can be used for the study
of the physiological role and mechanism of iron transport by
particular transporter proteins. Iron acquisition in the pres-
ence of synthetic DMA by a recombinant plant expressing the
transporters is already in progress in our laboratory and the
results will be soon published elsewhere. Another point to be
mentioned is that the 2’-hydroxy group can serve as a
potential position for the labeling of mugineic acid analogues
for functional studies. The introduction of functionality for
affinity labeling on the MA skeleton has, so far, never been
successful. The main reason was that all of the previously
prepared labeled products lost the ability to form iron(III)
complexes owing to the structural modification. We are
attempting to avoid this problem by introducing a label at the
2’-hydroxy group of mugineic or 2’-epi-mugineic acid.
C12H19N2O7 [MꢀH]ꢀ: 303.1200; found: 303.1198; [a]2D2 = ꢀ69.2 (c =
ꢀ
1.0, H2O, pH 6.5; literature value: ꢀ70.5[6]).
1: 1H NMR (400 MHz, D2O, pH 4.53 adjusted by the addition of
1m DCl): d = 4.88 (t, J = 9.5 Hz, 1H), 4.44 (dt, J = 9.3, 2.9 Hz, 1H),
4.17 (dd, J = 4.6, 7.3 Hz, 1H), 4.10 (dt, J = 4.2, 10.1 Hz, 1H), 4.03 (app.
q, J = 9.7 Hz, 1H), 3.85 (d, J = 2.9 Hz, 1H), 3.56 (dd, J = 9.5, 13.7 Hz,
1H), 3.43 (dd, J = 2.7, 13.7 Hz, 1H), 3.33–3.25 (m, 1H), 3.25–3.16 (m,
1H), 2.77–2.66 (m, 1H), 2.62–2.52 (m, 1H), 2.22–2.14 (m, 1H), 2.09–
1.99 ppm (m, 1H); 13C NMR (100 MHz, D2O, pH 4.53 adjusted by the
addition of 1m DCl): d = 182.3, 175.8, 171.8, 73.2, 70.6, 67.4 (2C), 59.0,
+
53.9, 47.9, 32.9, 24.6 ppm; HRMS: m/z calcd for C12H21N2O8
[M+H]+: 321.1298; found: 321.1292; [a]D24 = ꢀ63.5 (c = 0.31, H2O;
literature values: ꢀ70.7,[2] ꢀ64.6[8h]).
3: 1H NMR (400 MHz, D2O, pH 4.53 adjusted by the addition of
1m DCl): d = 4.89 (t, J = 9.5 Hz, 1H), 4.22 (ddd, J = 2.7, 8.3, 9.8 Hz,
1H), 4.16 (dd, J = 4.4, 7.6 Hz, 1H), 4.12–4.07 (m, 1H), 4.06 (app. q,
J = 9.7 Hz, 1H), 3.63 (d, J = 8.3 Hz, 1H), 3.57 (dd, J = 2.7, 13.2 Hz,
1H), 3.41 (dd, J = 9.8, 13.2 Hz, 1H), 3.30 (dt, J = 12.7, 7.1 Hz, 1H),
3.16 (dt, J = 12.7, 7.1 Hz, 1H), 2.75–2.65 (m, 1H), 2.61–2.51 (m, 1H),
2.20–2.10 (m, 1H), 2.05–1.96 ppm (m, 1H); 13C NMR (100 MHz,
7062 www.angewandte.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 7060 –7063

Angewandte
Chemie
D2O, pH 4.50 adjusted by the addition of 1m DCl): d = 182.1, 175.7,
[2] T. Takemoto, K. Nomoto, S. Fushiya, R. Ouchi, H. Kusano, H.
172.4, 73.2, 69.9, 67.5, 67.4, 60.0, 54.4, 48.1, 32.7, 24.5 ppm; HRMS:
Hikino, S. Takagi, Y. Matsuura, M. Kakudo, Proc. Jpn. Acad.
1978, 54, 469 – 473.
[3] J. F. Ma, K. Nomoto, Physiol. Plant. 1996, 97, 609 – 617.
[4] S. Takagi, Soil Sci. Plant Nutr. 1976, 22, 423 – 433.
[5] Y. Murata, J. F. Ma, N. Yamaji, D. Ueno, K. Nomoto, T. Iwashita,
Plant J. 2006, 46, 563 – 572.
m/z calcd for C12H21N2O8 [M+H]+: 321.1298; found: 321.1298;
+
[a]2D4 = ꢀ48.8 (c = 0.57, H2O).
Electrophysiological studies for Xenopus laevis oocytes with
synthetic mugineic acid derivative–iron(III) complexes: Two-elec-
trode voltage-clamp analysis of the HvYS1 transporter in Xenopus
laevis oocytes was performed as described previously.[5] The ORF
region of HvYS1 cDNA (DNA data base of Japan (DDBJ) under the
accession number AB214183) was amplified by PCR with the forward
primer 5’-GCTCTAGACCACCATGGACATCG-3’ and the reverse
primer 5’-CGCGGGATCCTTAGGCAGCAGGTAG-3’; it was then
inserted into the XbaI,BamHIsite of Xenopus expression vector
pSP64 polyA. The plasmid was linearized with BamH1 and then the
cRNAwas transcribed in vitro with SP6 RNA polymerase. The cRNA
solution (50 nL, 0.05 mgmLꢀ1) was injected into Xenopus oocytes,
which then were incubated for 2–4 days at 168c in ND96 buffer
(pH 7.6) containing 96 mm NaCl, 2 mm KCl, 1.8 mm CaCl2, 1 mm
MgCl2, and 5 mm 2-[4-(2-hydroxyethyl)-1-piperazinyl]ethanesulfonic
acid (HEPES). The oocytes were voltage clamped at ꢀ60 mV (Oocite
clamp OC-725c, Warner Instruments) and steady-state currents were
obtained in response to the addition of the MA–, 2’-epi-MA–, or
DMA–iron(III) complexes (10 mL, 7.5 mm; final concentration:
50 mm). The transport activity was analyzed with Origin 6.1 software
(Microcal Software).
[6] K. Nomoto, H. Yoshioka, M. Arima, S. Fushiya, S. Takagi, T.
Takemoto, Chimia 1981, 35, 249 – 250.
[7] S. Mori, Curr. Opin. Plant Biol. 1999, 2, 250 – 253.
[8] DMA synthesis: a) Y. Ohfune, M. Tomita, K. Nomoto, J. Am.
Chem. Soc. 1981, 103, 2409 – 2410; b) S. Fushiya, Y. Sato, S.
Nakatsuyama, N. Kanuma, S. Nozoe, Chem. Lett. 1981, 909 –
912; c) F. Matsuura, Y. Hamada, T. Shioiri, Tetrahedron 1994, 50,
9457 – 9470; d) T. Shioiri, N. Irako, S. Sakakibara, F. Matsuura, Y.
Hamada, Heterocycles 1997, 44, 519 – 530; e) S. Singh, G. Cross-
ley, S. Ghosal, Y. Lefievre, M. W. Pennington, Tetrahedron Lett.
2005, 46, 1419 – 1421. For MA synthesis: f) F. Matsuura, Y.
Hamada, T. Shioiri, Tetrahedron Lett. 1992, 33, 7917 – 7920; g) F.
Matsuura, Y. Hamada, T. Shioiri, Tetrahedron 1993, 49, 8211 –
8222; h) Y. Jung, K. Hoon, M. Chang, Arch. Pharmacal Res.
1999, 22, 624 – 628.
[9] a) S. S. Klair, H. R. Mohan, T. Kitahara, Tetrahedron Lett. 1998,
39, 89 – 92; b) K. Miyakoshi, J. Oshita, T. Kitahara, Tetrahedron
2001, 57, 3355 – 3360.
[10] T. Wakamiya, M. Inoue, E. Hirai, R. Shimizu, Y. Yamaguchi, T.
Iwashita, K. Nomoto, Pept. Sci. 1999, 35, 465 – 468. A facile
synthesis of aldehyde 8 is described in the Supporting Informa-
tion.
Received: June 3, 2007
Published online: August 9, 2007
[11] J. F. Ma, G. Kusano, S. Kimura, K. Nomoto, Phyochemistry 1993,
34, 599 – 603.
[12] M. R. Yen, Y. H. Tseng, M. H. Saier, Jr., Microbiology 2001, 147,
2881 – 2883.
[13] C. Curie, Z. Panaviene, C. Loulergue, S. L. Dellaporta, J. F.
Briat, E. L. Walker, Nature 2001, 409, 346 – 349.
Keywords: amino acids · iron · phytosiderophores · plant uptake ·
synthetic methods
.
[1] H. Marschner, V. Rꢁmheld, M. Kissel, J. Plant Nutr. 1986, 9,
[14] G. Schaaf, U. Ludewig, B. E. Erenoglu, S. Mori, T. Kitahara, N.
von Wierꢂn, J. Biol. Chem. 2004, 279, 9091 – 9096.
695 – 713.
Angew. Chem. Int. Ed. 2007, 46, 7060 –7063
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 7063