Synthesis and structural modeling of the amphiphilic siderophore rhizobactin-1021 and its analogs

Article abstract of DOI:10.1016/j.bmcl.2005.05.114

We describe two convenient syntheses of rhizobactin-1021 (Rz), a citrate-based siderophore amphiphile produced by the nitrogen-fixing root symbiont Rhizobium meliloti-1021, and several analogs. Our approach features a singly amidated, tert-butyl-protected

Full text of DOI:10.1016/j.bmcl.2005.05.114

Bioorganic & Medicinal Chemistry Letters 15 (2005) 3771–3774
Synthesis and structural modeling of the amphiphilic siderophore
rhizobactin-1021 and its analogs
Evgeny A. Fadeev, Minkui Luo and John T. Groves*
Department of Chemistry, Princeton University, Princeton, NJ 08544, USA
Received 13 April 2005; revised 13 May 2005; accepted 16 May 2005
Available online 28 June 2005
Abstract—We describe two convenient syntheses of rhizobactin-1021 (Rz), a citrate-based siderophore amphiphile produced by the
nitrogen-fixing root symbiont Rhizobium meliloti-1021, and several analogs. Our approach features a singly amidated, tert-butyl-
protected citrate intermediate that easily affords a variety of Rz analogs in the late stages of the synthesis. Structural modeling
and the monolayer behavior of Rz and its metal complexes are consistent with a structural reorganization upon Rz-mediated iron
chelation.
ꢀ 2005 Elsevier Ltd. All rights reserved.
The principal strategy for bacterial response to iron
restriction is to secrete siderophores, a family of small
organic compounds with a highly selective iron-chelat-
ing ability.1 These microbial iron chelators have attract-
ed attention not only due to their essential functions for
bacterial growth2 but also because of a variety of phar-
macological applications.3 Drug-conjugated sidero-
phores have been described as potential Trojan Horse
vehicles for antibiotic delivery;4 siderophore-based iron
scavengers have shown promising therapeutic activities
for breast cancer,5 reperfusion injury,6 and malaria,7
and siderophore-like chelators can be used for the treat-
ment of iron-overload diseases.8,9 Consequently, studies
in these fields still largely depend on the availability of
Figure 1. Structures of rhizobactin-1021 (Rz) and its analogs.
structurally diverse natural and artificial siderophores.
Rhizobactin-1021 (Rz) is a citrate-based siderophore
amphiphile produced by the nitrogen-fixing root symbi-
ont Rhizobium meliloti-1021.10 A unique aspect of the Rz
structure lies in its two non-equivalent hydroxamate
possible therapeutic applications of Rz stimulated us
to develop a convenient access to this compound and
its analogs.
subunits, with one long hydrocarbon chain (Fig. 1).
We recently showed that the overall structures of sider-
ophore amphiphiles play important roles in their mem-
In this paper, we describe two synthetic approaches to
Rz. The overall unsymmetrical structure of Rz makes
its synthesis more challenging than those of its symmet-
ric counterparts.11,15–22 The core strategy was to design
brane-interaction properties.11–13 The a,b-unsaturated
hydroxamate moiety in Rz is also found in acinetofer-
an unsymmetrical precursor suitable for the preparation
rin,14 mycobactins,2 and nannochelin A.15 Accordingly,
the structural features, biological properties, and
of various Rz analogs by a universal procedure. The
molecular conformations and monolayer behavior of
Rz and its metal complexes were also investigated.
Keywords: Amphiphile; Iron; Structure; Acinetoferrin; Mycobactin.
*
Scheme 1 shows our two synthetic approaches to race-
mic Rz and its analogs. The key step for the first
Corresponding author. Tel.: +1 609 258 3593; fax: +1 609 258
0348; e-mail: jtgroves@princeton.edu
0960-894X/$ - see front matter ꢀ 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2005.05.114

3772
E. A. Fadeev et al. / Bioorg. Med. Chem. Lett. 15 (2005) 3771–3774
Scheme 1. Synthesis of the citrate-based amphiphilic siderophore rhizobactin-1021 (Rz) and its analogs. Reagents and conditions:25 (i) a. DCC,
HOSu/dioxane; b. 10 + TEA/CH3CN; (ii) a. R1COCl/DCM, ꢀ78 ꢁC (R1 5 R2); b. R2COCl/DCM, reflux; (iii) DCC/DCM; (iv) 12/DCM, 25 ꢁC;
(v) a. DCC, HOSu/DCM, 25 ꢁC; b. 10 + TEA/DCM; (vi) R2COCl/DCM (R1@CH3), reflux; (vii) 5 N NaOH/CH3OH, 0 ꢁC; (viii) TFA/H2O, 19:1;
(ix) HCl gas; (x) Acetyl chloride/DCM, reflux.
approach (1 ! 4 ! 6) is to differentiate the two carbox-
ylic acid groups of the citrate moiety. This goal was
accomplished via preparing the singly amidated citrate
derivative (4) by coupling 1-N-(acetyl-benzoyloxy)-1,3-
diaminopropane hydrochloride (12) with the tert-butyl-
protected cyclic citric anhydride (3). This anhydride
intermediate is readily obtained from the tert-butyl-pro-
tected citric acid (1)20 via DCC-facilitated intramolecu-
lar dehydration in 90% yield. The amine hydrochloride
(12) was synthesized via acetylation of N-(benzoyloxy)-
the steps 3 ! 5, we found it unnecessary to pursue the
rigorous purification of 4, since compound 13 did not re-
act under the subsequent conditions. Consequently, the
first column purification could be postponed until the
preparation of 5 with the overall yield of 51% (3 ! 5).
These tert-butyl-protected citrate intermediates com-
pletely avoid the undesirable imide formation via intra-
molecular condensation.18,22,24 The fully protected
precursors of Rz and its unsymmetrical analogs (6d–g)
were obtained by coupling with the corresponding
trans-2-alkenoyl chloride in 70%–80% yield. The depro-
tections of the benzoyl and tert-butyl groups (6d–
g ! 8d–g) were carried out with aqueous NaOH/MeOH
(80%–95%) and 95% TFA/water, respectively.11 Conse-
quently, compounds 8d–g were obtained with an overall
yield of 26%–35% with respect to 1. The 1H NMR spec-
trum of racemic compound 8e (Rz) obtained in this way
was identical to that of authentic rhizobactin-1021.10
3-(tert-butoxycarbonylamino)propyl-amine22,23
and
subsequent deprotection of the tert-butoxycarbonyl
group with dry HCl (9 ! 12, yield 96% and 82% for
the two steps, respectively). This early installation of
the acetyl hydroxamate moiety allows ready modifica-
tions for a variety of analogs (8c–f) simply by attaching
different trans-2-alkenoyl chains (6 ! 7). The hydro-
chloride salt was chosen for compound 12, since the
more typical procedure of using TFA afforded an oil
and facilitated the undesirable acyl transfer reaction
(12 ! 13).24 This rearrangement was further compensat-
ed by the use of 30% excess 12 in the subsequent cou-
pling reaction (3 ! 4). Here, the yield of 4 was
quantitative with respect to 3. Compound 4 was not eas-
ily separated from the acyl transfer byproduct 13
through column chromatography; however, compound
4 could be purified via repetitive washing using aqueous
HCl (pH 2). This compound was then converted into the
activated N-(hydroxyl)succinimidyl ester that underwent
facile coupling with 1-N-benzoyloxy-1,3-diaminopro-
pane dihydrochloride (10) to give compound 5.11 For
We also synthesized racemic Rz via another more
straightforward approach (1 ! 2 ! 6e in Scheme 1).
Intermediate 2 was prepared as we have described by
coupling tert-butyl-protected citric acid (1) with 1-N-
benzoyloxy-1,3-diaminopropane
dihydrochloride
(10).11 This symmetrical citrate diamide (2) was then
reacted with one equivalent of trans-2-decenoyl chloride
and then acetyl chloride. The resulting mixture consisted
of a nearly statistical ratio of compound 6e as the major
component with the two expected symmetrical products.
This target compound 6e was readily purified by
thin-layer chromatography. We found that the addition

E. A. Fadeev et al. / Bioorg. Med. Chem. Lett. 15 (2005) 3771–3774
3773
of the less reactive trans-2-decenoyl chloride before the
acetyl chloride maximized the yield of 6e. We also no-
ticed that running the first acylation reaction at
ꢀ78 ꢁC for 2 h gave the highest yield of the mono-
acyl-coupled intermediate. The second N-benzoyloxy-
amino moiety in 2 was then rapidly reacted with excess
acetyl chloride in refluxing DCM to prevent further acyl-
ation by trans-2-decenoyl chloride. This one-pot reaction
eventually gave an overall yield of 48% of 6e. Rz (8e) was
obtained through the same deprotection protocols as
described above. Here, we also prepared 8a–c, three sym-
metric analogs of Rz, by reacting 2 with the corresponding
acyl chlorides (80% yield for 2 ! 6a–c and 80%–95% for
6a–c ! 8a–c, 31%–36% overall from 2).
Figure 3. Measurements of mean molecular areas (Mma) using
Langmuir–Blodgett techniques. The variations of the surface pressure
were recorded by pressing the monolayers after a 30-min incubation of
Rz onto the air/subphase interface (HEPES buffered solution, at pH
7.4 and 20 ꢁC, in the presence and absence of 5 mM FAC for Rz and
Fe-Rz, respectively). The headgroup sizes of Rz and Fe-Rz were
obtained by extrapolating the linear region to zero surface pressure
The likely molecular conformations of Rz and its metal
complex (Fig. 2) were derived from the NMR structures
that we have reported for acinetoferrin (Af) and schiz-
okinen.11,12 All diastereomers of Fe-Rz adopt similar
3-D orientations except chirality. Similar to the symmet-
rical citrate siderophores,11,12 the Rz iron complex has
most of its polar residues buried inside and its hydrocar-
bon skeleton exposed to the outside. The negative char-
ge of Rz metal complex is delocalized over the metal
center and its six coordinating oxygens. This conforma-
tional change of Fe-Rz would be expected to facilitate its
membrane flip-flop.12 The N-cis-cis conformation of Rz
metal complex makes its terminal methyl and trans-2-
decenoyl chain adopt an anti-parallel orientation with
the angle of 130ꢁ.11 However, the overall structure of
the Rz metal complex is much less extended than that
of its Af analog, because of the short methyl group in
the former vs the long octenoyl chain in the latter.
2
2
˚
with Mma 28 A for Rz and 46 A for Fe-Rz, respectively.
˚
coordination of Fe-Rz and Fe-Af.11 In this coordination
geometry the extended side chain orientation for Fe-Af
2
would account for the 68 A increase relative to Fe-Rz
˚
due to the incommensurate arrangement of the second
hydrocarbon chain with respect to the parallel packing
of the phospholipid side chains.11,12 The headgroup size
2
˚
of Fe-Rz (46 A ) reported here is in good agreement
with the structural models of Rz metal complexes de-
scribed above. As we have suggested, changes in head-
group size and molecular conformation upon binding
iron are likely important determinants of membrane
binding and permeability.12
The headgroup sizes of Rz and its iron complex were
measured via Langmuir–Blodgett techniques to further
understand the conformational changes upon Rz-medi-
ated iron chelation. As shown in Figure 3, Rz and Fe-
Rz form well-behaved Langmuir monolayers. Head-
This work reports the first synthesis of Rz (8e) and its
unsymmetrical analogs (8d and 8f–g). Each of the two
approaches has its respective merits. The first one
(1 ! 4 ! 6) provides ready access to a variety of Rz-
based homologs via the common intermediate 5 with-
out major change of the procedure. Furthermore, this
strategy may be adopted for large-scale or combinato-
rial solid-phase synthesis and to prepare fluorophore-
conjugated analogs by substituting the hydrocarbon
moieties with hydrophobic fluorescent probes. Such
derivatives are useful for siderophore-mediated intra-
cellular trafficking studies. The second approach
(1 ! 2 ! 6) provides a straightforward route to obtain
a few hundred milligrams of compound in a short peri-
od of time. In addition, both approaches keep the
iron-chelating moieties protected until the final stage
and therefore minimize the iron contamination. Iron
chelation by Rz causes an expansion of the size of
2
˚
group sizes of 28 and 46 A were obtained for Rz and
Fe-Rz, respectively, by extrapolating the solid-monolay-
2
er region to zero surface pressure. The 18 A increase of
˚
Rz headgroup size indicates that iron chelation causes a
structural reorganization in Rz. The smaller headgroup
2
size of Fe-Rz (46 A ) in contrast to the mean molecular
˚
2
area of Fe-Af (114 A ) is consistent with the N-cis-cis
˚
2
˚
the headgroup by 18 A , consistent with the structural
models of Rz and its metal complex. The molecular
structures and monolayer properties of Rz and its met-
al complex further argue that this amphiphile can
interact with biological membranes. Studies of iron
acquisition and membrane behavior of Rz and its ana-
logs are ongoing.
Figure 2. Likely 3-D structures of Rz and its metal complex (a
representative 3-D structure). These structures were derived by
analogy to acinetoferrin and its gallium complex.11 There is
structural reorganization in the headgroup region of Rz upon iron
chelation.
a

3774
E. A. Fadeev et al. / Bioorg. Med. Chem. Lett. 15 (2005) 3771–3774
20. Milewska, M. J.; Chimiak, A.; Glowacki, Z. J. Prakt.
Chem. 1987, 329, 447.
Acknowledgments
21. Miller, M. J.; Maurer, P. J. J. Am. Chem. Soc. 1982, 104,
3096.
22. Wang, Q. X. H.; Phanstiel, O. J. Org. Chem. 1998, 63,
1491.
23. Wang, Q. X. H.; King, J.; Phanstiel, O. J. Org. Chem.
1997, 62, 8104.
Support of this work by the National Science Founda-
tion (CHE-0316301 and CHE-0221978) through the
Environmental Molecular Science Institute (CEBIC) at
Princeton University is gratefully acknowledged.
24. Ghosh, A.; Miller, M. J. J. Org. Chem. 1993, 58, 7652.
25. Compound 11, 1H NMR: (300 MHz, CDCl3): d 1.41 (s,
9H, CH3), 1.79 (m, 2H, CH2), 2.06 (s, 3H, COCH3), 2.88
(t, 2H, BocN–CH2), 3.21 (m, 2H, CON–CH2), 5.08 (br s,
1H, NH), 7.54 (m, 2H, aromatic), 7.70 (m, 1H, aromatic),
8.11 (m, 2H, aromatic). Compound 12, 1H NMR
(400 MHz, CD3OD): d 2.01 (q, 2H, CH2), 2.08 (s, 3H,
CH3), 3.08 (t, 2H, N-CH2), 3.96 (t, 2H, CON–CH2), 7.60
(t, 2H, aromatic), 7.77 (t, 1H, aromatic), 8.14 (d, 2H,
aromatic). Anal Calcd for C12H17ClN2O3: C, 52.85; H,
6.28; Cl, 13.00; N, 10.27, found: C, 52.94; H, 6.92; Cl,
14.55; N, 10.75. Compound 3, 1H NMR (400 MHz,
CDCl3): d 1.48 (s, 9H, CH3), 2.96 (q, 4H, CH2). Anal
Calcd for C10H14O6: C, 52.17; H, 6.13, found: C, 52.22; H,
Supplementary data
Supplementary data associated with this article can be
found in the online version at doi:10.1016/
j.bmcl.2005.05.114.
References and notes
1. Albrecht-Gary, A. M.; Crumbliss, A. L. Met. Ions Biol.
Syst. 1998, 35, 239.
2. Ratledge, C.; Dover, L. G. Annu. Rev. Microbiol. 2000, 54,
881.
1
6.08. Compound 5, H NMR (400 MHz, CDCl3): d 1.41
3. Roosenberg, J. M.; Lin, Y. M.; Lu, Y.; Miller, M. J. Curr.
Med. Chem. 2000, 7, 159.
(s, 9H, 3 · CH3), 1.77 (m, 4H, 2 · C–CH2–C), 2.00 (s, 3H,
COCH3), 2.59 (m, 4H, 2 · COCH2), 3.15 (m, 2H, BzON–
CH2), 3.28 (m, 4H, 2 · CON–CH2), 3.82 (m, 2H,
CON(BzO)–CH2), 6.86 (m, 1H, NH), 6.95 (m, 1H, NH),
7.44 (m, 4H, aromatic), 7.63 (m, 2H, aromatic), 7.95 (m,
2H, aromatic), 8.03 (m, 2H, aromatic). HRESIMS Calcd
for C32H42N4O10 + Na+ 665.2799, found 665.2787. Anal
Calcd for C32H42N4O10: C, 59.80; H, 6.59; N, 8.72, found:
C, 59.35; H, 6.43; N, 8.49. Compound 6e, 1H NMR
(400 MHz, CD3OD): d 0.86 (t, 3H, C9–CH3), 1.28 (m,
10H, 5 · –CH2–), 1.44 (s, 9H, 3 · CH3), 1.88 (m, 4H,
2 · C–CH2-C), 2.06 (br s, 3H, COCH3), 2.18 (m, 2H,
C@C–CH2), 2.62 (ab-quartet, 4H, 2 · OC–CH2), 3.28 (m,
4H, 2 · CON–CH2), 3.90 (m, 4H, 2 · CON(BzO)–CH2),
6.18 (d, 1H, OC-CH@), 6.98 (m, 1H, @CH–C), 7.62 (m,
4H, aromatic), 7.72 (m, 2H, aromatic), 7.95 (m, 4H,
aromatic), 8.16 (m, 4H, aromatic). HRESIMS Calcd for
C42H58N4O11 + Na+ 817.4000, found 817.3976. Com-
pound 8b. 1H NMR (400 MHz, CDCl3): d 0.82 (t, 6H,
C7–CH3), 1.24 (m, 16H, 8 · –CH2–), 1.51 (m, 4H, 2 · CO–
C–CH2), 1.71 (m, 4H, 2 · C–CH2–C), 2.37 (m, 4H,
2 · N(BzO)–CO–CH2–), 2.60 (ab-quartet, 4H, 2 · OC–
CH2), 3.10 (m, 4H, 2 · CON–CH2), 3.56 (m, 4H,
2 · CON(BzO)–CH2). 13C NMR (400 MHz, CD3OD): d
14.16, 23.83, 26.12, 27.72, 30.35, 30.35, 30.65, 33.08, 33.46,
37.81, 45.21, 46.89, 75.23, 172.27, 176.46, 177.01. HRE-
SIMS Calcd for C28H521N4O9 + Na+ 611.3632, found
611.3626. Compound 8e, H NMR (400 MHz, CD3OD):
d 0.83 (t, 3H, C9–CH3), 1.34 (m, 8H, 4 · –CH2–), 1.42 (m,
2H, C@C–C–CH2), 1.76 (m, 4H, 2 · C–CH2–C), 2.02 (s,
3H, COCH3), 2.16 (m, 2H, C@C–CH2), 2.60 (ab-quartet,
4H, 2 · COCH2), 3.12 (m, 4H, 2 · CON–CH2), 3.60 (m,
4H, 2 · CON(O)–CH2), 6.52 (d, 1H, COCH@), 6.76 (m,
1H, @CH–). HRESIMS Calcd for C24H42N4O9 + H+
531.3030, found 531.3021.
4. Durham, T. B.; Miller, M. J. J. Org. Chem. 2003, 68, 27.
5. Pahl, P. M. B.; Horwitz, M. A.; Horwitz, K. B.; Horwitz,
L. D. Breast Cancer Res. Treat. 2001, 69, 69.
6. Horwitz, L. D.; Sherman, N. A.; Kong, Y. N.; Pike, A.
W.; Gobin, J.; Fennessey, P. V.; Horwitz, M. A. Proc.
Natl. Acad. Sci. U.S.A. 1998, 95, 5263.
7. Golenser, J.; Tsafack, A.; Amichai, Y.; Libman, J.;
Shanzer, A.; Cabantchik, Z. I. Antimicrob. Agents Che-
mother. 1995, 39, 61.
8. Turcot, I.; Stintzi, A.; Xu, J. D.; Raymond, K. N. J. Biol.
Inorg. Chem. 2000, 5, 634.
9. Bergeron, R. J.; Wiegand, J.; McManis, J. S.; Bussenius,
J.; Smith, R. E.; Weimar, W. R. J. Med. Chem. 2003, 46,
1470.
10. Persmark, M.; Pittman, P.; Buyer, J. S.; Schwyn, B.; Gill,
P. R.; Neilands, J. B. J. Am. Chem. Soc. 1993, 115, 3950.
11. Fadeev, E. A.; Luo, M.; Groves, J. T. J. Am. Chem. Soc.
2004, 126, 12065.
12. Luo, M.; Fadeev, E. A.; Groves, J. T. J. Am. Chem. Soc.
2005, 127, 1726.
13. Xu, G. F.; Martinez, J. S.; Groves, J. T.; Butler, A. J. Am.
Chem. Soc. 2002, 124, 13408.
14. Okujo, N.; Sakakibara, Y.; Yoshida, T.; Yamamoto, S.
Biometals 1994, 7, 170.
15. Bergeron, R. J.; Phanstiel, O. J. Org. Chem. 1992, 57,
7140.
16. Bergeron, R. J.; Huang, G. F.; Smith, R. E.; Bharti, N.;
McManis, J. S.; Butler, A. Tetrahedron 2003, 59, 2007.
17. Gardner, R. A.; Ghobrial, G.; Naser, S. A.; Phanstiel, O.
J. Med. Chem. 2004, 47, 4933.
18. Guo, H. Y.; Naser, S. A.; Ghobrial, G.; Phanstiel, O.
J. Med. Chem. 2002, 45, 2056.
19. Lee, B. H.; Miller, M. J. J. Org. Chem. 1983, 48, 24.