Synthesis and Characterization of Anguibactin to Reveal Its Competence to Function as a Thermally Stable Surrogate Siderophore for a Gram-Negative Pathogen, Acinetobacter baumannii

Article abstract of DOI:10.1021/acs.orglett.8b02789

Total synthesis of anguibactin was accomplished for the first time, and the following biochemical characterizations allowed for the determination of its Fe(III) binding mode as well as the demonstration of its iron delivery capability for Acinetobacter ba

Full text of DOI:10.1021/acs.orglett.8b02789

Letter
Cite This: Org. Lett. XXXX, XXX, XXX−XXX
pubs.acs.org/OrgLett
Synthesis and Characterization of Anguibactin To Reveal Its
Competence To Function as a Thermally Stable Surrogate
Siderophore for a Gram-Negative Pathogen, Acinetobacter
baumannii
Haeun Lee, Woon Young Song, Minju Kim,†,‡ Min Wook Lee, Soojeung Kim, Ye Song Park,
†
,∥
†,∥
†
†
†
Kyungwon Kwak,† Man Hwan Oh, and Hak Joong Kim*
,‡
§
,†
†
Department of Chemistry, Korea University, 145 Anam-ro, Seongbuk-gu, Seoul 02841, Republic of Korea
‡
Center for Molecular Spectroscopy and Dynamics, Institute for Basic Science (IBS), Korea University, Seoul 02841, Republic of
Korea
§
Department of Nanobiomedical Science, Dankook University, Cheonan 31116, Republic of Korea
S Supporting Information
*
ABSTRACT: Total synthesis of anguibactin was accomplished for
the first time, and the following biochemical characterizations allowed
for the determination of its Fe(III) binding mode as well as the
demonstration of its iron delivery capability for Acinetobacter
baumannii. These properties, in addition to the thermal stability
over acinetobactin, render anguibactin as a competent surrogate
siderophore that can be useful for the future development of a
siderophore-based antibiotic delivery system against A. baumannii.
oncerns for the antibiotic resistance of the Gram-negative
C
pathogen Acinetobacter baumannii have been significantly
elevated over the past decades because of its rapid spread in
the healthcare units as well as the paucity of effective
1
,2
therapeutics. Apart from the development of β-lactamase
genes, one distinguishing drug resistance mechanism of A.
baumannii is the fortification of the drug permeability barrier,
particularly in preventing the penetration of antibiotic
3
,4
molecules across the outer membrane. As a reasonable
solution to address such permeability issues of drug-resistant A.
baumannii, a siderophore-based antibiotic delivery strategy was
5
−7
Figure 1. Structures of the siderophores investigated in this study.
proposed by us and others.
In this approach, the
siderophore uptake machineries, originally utilized for the
acquisition of iron, are exploited to intracellularly deliver the
antibiotic attached to a siderophore. Recently, some success of
this approach against Gram-negative pathogens has been
demonstrated by a number of independent research
5
characterization of acinetobactins, we proposed that 1 would
be more suitable to be utilized as an antibiotic delivery vector
because of its well-defined cellular uptake mechanism, enabling
the rational design of the potent conjugate structure (Figure
S1).
8
−13
groups.
One serious drawback of 1 for antibiotic delivery is its
thermal instability, for which it readily isomerizes to 2. Upon
contemplating suitable structural modifications of 1 to improve
its stability and at the same time maintain its interaction with
the acinetobactin uptake machineries encoded in a bau gene
set (Figure S1A), another natural siderophore, anguibactin (3),
from the fish pathogen Vibrio anguillarum 775 caught our
Among the three siderophores identified in A. bauman-
1
4−16
nii,
acinetobactin is the most frequently observed in the
17
clinically isolated A. baumannii strains, and thus, it has been
5
,18−20
the subject of extensive characterization.
was originally proposed to possess the oxazoline ring at its core
1, Ab-Oxa); however, later it was found that the isolated
Its structure
(
acinetobactin featured an isoxazolidinone moiety (2, Ab-Isox),
23
attention (Figure 1). Anguibactin was not only structurally
which was derived by the spontaneous isomerization of 1
2
1,22
(
Figure 1).
Despite the identification of the structure of 2,
the physiological relevance of these forms needed to be
Received: August 31, 2018
addressed. In our previous study concerning the functional
©
XXXX American Chemical Society
A
DOI: 10.1021/acs.orglett.8b02789
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
similar to acinetobactin, but the cognate uptake machineries
for these two siderophores also resembled each other (Figure
consistent with the reported data (Table S1). Along with 3, its
deoxy analogue (4) was also prepared in an analogous manner
(Figure S3, Table S3). Previously, we as well as Shapiro et al.
found that the 3-OH group of the phenyl ring was not required
2
4
S1). Particularly, the resistance of 3 toward isomerization
2
1
was confirmed by Wuest et al. These prior studies
collectively led us to hypothesize that this molecule could
serve as a thermally stable surrogate of Ab-Oxa (1) for A.
baumannii. In this regard, herein, we describe the first synthesis
of 3 as well as its functional characterization in comparison
with 1 as experimental evidence to support our hypothesis.
The synthesis of 3 began with construction of the thiazoline
core by condensation of L-cysteine and aryl nitrile 5 derived
5
,19
for siderophore function of 1.
Therefore, 3-deoxy-
anguibactin (4) was considered worth evaluating, particularly
because its synthesis and handling were less demanding than
those of 3.
With two thiazoline derivatives (3 and 4) in hand, their
Fe(III) binding properties were assessed by chrome azurol S
(CAS) assays. As shown in Figure 2A, the color change of the
2
5
from vanillin, yielding acid 8 (Scheme 1). Then, the N,O-
Scheme 1. Synthesis of Anguibactin (3) and 3-Deoxy-
anguibactin (4)
Figure 2. Spectral characterization of 3 and 4. (A) Chrome azurol S
assay results. Absorption (B) and fluorescence (C) spectra of 3 (100
μM), 4 (100 μM), and their corresponding Fe(III)-complexes
protected imidazole fragment 10, whose synthesis was
26
previously established, was treated with trifluoroacetic acid
to remove the N-Boc protecting group, and the resulting
benzyloxy amine 11 was subjected to a coupling with acid 8
using HATU in the presence of K CO . Unfortunately, in this
prepared by treatment of 50 μM Fe(acac) in phosphate buffered
3
saline (PBS, pH = 7.4). (D) Plots of the fluorescence titration results
for 3 (100 μM) and 4 (100 μM) with Fe(acac) in PBS buffer (insets,
3
2
3
Job plot analysis results; χ, [Fe(III)]/([3 or 4] + [Fe(III)]); I0,
fluorescence intensity without Fe(III) addition; and I, fluorescence
intensity at a specific χ).
initial trial, significant epimerization was observed at position
4
′ to give a near-racemic mixture. Since the analogous step
22,26
during acinetobactin synthesis was stereoselective,
this
result was disappointing. The epimerization appeared to be
due to the absence of a substituent at position 5′ in 8. In the
case of acinetobactin, the (R)-methyl group at that position
could secure the sterically stable trans geometry with the (S)-
configuration at C4′ (Figure 1). Indeed, similar epimerization
events were previously observed by us and Nelson et al. during
the syntheses of fimsbactin A and transvalencin Z,
CAS solution from deep blue to orange was evident at >50 μM
for 3 as well as 4 under the given conditions, confirming that
they were efficient Fe(III) binders. The spectral character-
ization of 3 and 4 showed conspicuous changes in the
absorption and fluorescence upon binding with Fe(III) as
depicted in Figure 2B and 2C. Among them, the fluorescence
quenching property was exploited to determine the binding
stoichiometry between these ligands and Fe(III). The
previously reported mass data as well as an X-ray crystal
2
7,28
respectively.
To resolve this stereoselectivity issue, a series
of reaction conditions varying the coupling agents and solvents
were screened (Figure S2, Table S2), and we found that the
use of HATU in dichloromethane with a free amine form 12
led to decent stereoselectivity (er = 98:2) in the formation of
23,29
structure have suggested a 1:1 complex formation,
but
such gas and solid phase results are not always guaranteed to
represent the behaviors in liquid phase. In this regard, Actis et
al. attempted to determine the binding stoichiometry by
1
3.
The final-step removal of the two ether-type protecting
23
groups was also problematic. Initially, a variety of Lewis acids
were tried, but none were found effective, particularly in
cleaving the O-benzyl ether. Finally, it was found that inclusion
of tetrabutylammonium iodide (TBAI) to boron trichloride
absorption titration in liquid phase, but they failed to gain a
conclusive result because of the interference by ready
formation of colloidal Fe(III) hydroxide at a neutral pH,
suggesting that the selection of the Fe(III) source would be
critical for successful titration. Analogous to their observation,
in our initial titration trial, the use of FeCl3 could not
completely quench the fluorescence of 3 and 4, even at a 2-fold
(
BCl ) led to clean removal of both benzyl and methyl groups
3
to successfully generate 3 in a moderate yield. The exact role of
TBAI is uncertain, but it is likely to assist the cleavage of the
C−O bond by providing a strong nucleophile, an iodide anion.
excess. However, we discovered that Fe(acac) could serve as a
3
1
The H NMR spectrum of the synthesized anguibactin was
suitable ferric ion source, by which the fluorescence quenching
B
DOI: 10.1021/acs.orglett.8b02789
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
of 3 and 4 was efficient and reproducible. As shown in Figure
2
D, the fluorescence of both compounds reached the lowest
level upon the addition of 0.5 equiv of Fe(acac) , and the Job
3
plot analysis (Figure 2D, insets) where the maxima were
observed at the Fe(III) molar fraction (χ) of ca. 0.33
unequivocally showed a 2:1 complex formation between
these two compounds and Fe(III).
The determined 2:1 binding stoichiometry of 3 was identical
5
,18
to that of 1,
which is consistent with our hypothesis of the
functional similarity between these two siderophores. The next
key question to be addressed was whether 3 could deliver the
iron into A. baumannii. To address this question, first, a growth
promotion assay was conducted with A. baumannii ΔbasD, a
30
mutant whose acinetobactin biosynthetic gene was disrupted.
As shown in Figure 3A, 3 and 4 were capable of stimulating the
Figure 4. Density functional theory calculation results. (A) Structures
of the model substrates (15, 16), transition states (TS-O, TS-S), and
products (17, 18) used in the calculation. (B) Calculated enthalpic
energy profile for the isomerization reactions from 15 and 16. The
predicted transition state structures at TS-O and TS-S levels were
depicted in a ball-and-stick model (gold, carbon; white, hydrogen;
red, oxygen; blue, nitrogen; and yellow, sulfur). All calculations were
done considering gas phase.
Figure 3. Growth promotion assay results using A. baumannii ΔbasD
structural optimization revealed that the energy gap between
(
A) and ΔbauA (B) strains. The numbers on the x-axis indicate the
corresponding compounds.
the nonbonding orbital (n ) of the attacking oxyanion and the
O
accepting σ antibonding orbital (σ*) of 16 was smaller than
that of 15, suggesting that the isomerization of 16 should be
more amenable (Figure S4). Because this prediction was
contradictory to the observed isomerization trend, we turned
our attention to the influence of the kinetic barrier. In this
regard, a full reaction energy profile for each isomerization
reaction was calculated as shown in Figure 4B, in which the
enthalpic activation energy for the isomerization for 15 was
found to be smaller than that for 16 by 3.83 kcal/mol,
corresponding to the 640-fold longer half-life of the latter
process. This result is consistent with the observed resistance
of 3 to isomerization. Analysis of the predicted transition state
models suggested that the high activation energy for 16 could
be due to the unfavorable angle between the dissociating C−S
bond and the attacking oxygen atom (∠S1−C5−O = 146° vs
∠O1−C5−O = 172° for 15), which prevents effective orbital
growth of this test strain under iron-deficient conditions
(
Luria−Bertani broth containing 200 μM 2,2′-bipyridyl) as
potently as 1 could. While a similar observation was reported
24
by Dorsey et al. with an agar plate-based assay, a hitherto
unaddressed, yet important, issue was whether 3 would exploit
the bau transport system for the entry into A. baumannii
(
Figure S1). In this regard, another growth promotion assay
was carried out with A. baumannii ΔbauA, where the bauA
gene encoding the outer membrane receptor was disrupted.
30
As shown in Figure 3B, 3 and 4 could not stimulate the growth
of this mutant in contrast to 2, which clearly demonstrated that
the pathway for the iron delivery to A. baumannii by 3 was
analogous to that of 1 in terms of absolute dependence on
BauA. The observed attenuation in the activities for A.
baumannii ΔbasD at high concentrations of 3 and 4 were
unexpected. While full elucidation of such an unusual behavior
would require additional in-depth analysis, complete immobi-
lization of free Fe(III) in the medium by high concentrations
of 3 and 4 as well as their suboptimal interactions with the bau
transport machineries compared to Ab-Oxa seem to prevent
the progression of the A. baumannii growth from the early lag
phase to the exponential log phase. Nonetheless, it is worth
noting that 3 and 4 maintained moderate iron delivery activity
at pharmacologically relevant single-digit micromolar ranges.
Collectively, these experiments show that both 3 and 4 were
functionally competent surrogates of Ab-Oxa, yet devoid of the
thermal stability issue.
interactions between nO and σ*
during nucleophilic
C−S
substitution. While the use of the simplified models as well
as the consideration of gas phase are the limitations of this
calculation study, the information gained here is still worth
noting in that it provides some valuable clues for a better
understanding of the effects of a single element difference (O
vs S) on the isomerization propensity of acinetobactin and
anguibactin.
In summary, here, we reported the first successful synthesis
of anguibactin (3), which allowed for its in-depth character-
ization. The spectral analysis confirmed its capability as an
efficient Fe(III) binder, and fluorescence titrations enabled an
unequivocal determination of the 2:1 binding stoichiometry
between 3 and Fe(III) in liquid phase for the first time. The
growth promotion assay under iron-deficient conditions using
two A. baumannii mutant strains, ΔbasD and ΔbauA, indicated
that 3 depended on the bau transport system for cellular entry.
These characteristics were analogous to those of Ab-Oxa (1),
but not Ab-Isox (2), and strongly supported our hypothesis
To better understand the origin for the thermal stability of
anguibactin, density functional theory (DFT) calculations were
performed with simplified compounds (15 and 16) that model
1
and 3, respectively, in consideration of the protonation states
of their hydroxamate moieties at neutral pH (Figure 4A).
Natural bond orbital analysis (NBO) after ground state
C
DOI: 10.1021/acs.orglett.8b02789
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
that 3 is a thermally stable surrogate of Ab-Oxa for A.
baumannii.
(9) Zheng, T.; Nolan, E. M. J. Am. Chem. Soc. 2014, 136, 9677.
(
4
(
10) Chairatana, P.; Zheng, T.; Nolan, E. M. Chem. Sci. 2015, 6,
458.
11) Ghosh, M.; Miller, P. A.; Mollmann, U.; Claypool, W. D.;
̈
The current efforts to develop siderophore-based antibiotic
delivery systems have been primarily focused on targeting
drug-resistant Gram-negative pathogens including A. bauman-
Schroeder, V. A.; Wolter, W. R.; Suckow, M.; Yu, H.; Li, S.; Huang,
W.; Zajicek, J.; Miller, M. J. J. Med. Chem. 2017, 60, 4577.
11,12
nii. Recent reports by Miller and Nolan laboratories
are
(
12) Neumann, W.; Sassone-Corsi, M.; Raffatellu, M.; Nolan, E. M.
J. Am. Chem. Soc. 2018, 140, 5193.
13) Ghosh, M.; Lin, Y.-M.; Miller, P. A.; Mollmann, U.; Boggess,
̈
particularly worth paying attention to in that they have
successfully demonstrated the notable activity of cytoplasm-
targeting siderophore−antibiotic conjugates in Gram-negative
bacteria for the first time. While the antibiotic delivery
capability of anguibactin (3) and its deoxy derivative (4)
awaits experimental validation, it would be an intriguing
inquiry to be addressed whether they could serve as effective
vehicles to deliver antibiotic molecules inside drug-resistant A.
baumannii.
(
W. C.; Miller, M. J. ACS Infect. Dis. 2018, ASAP, DOI: 10.1021/
acsinfecdis.8b00150.
(14) Yamamoto, S.; Okujo, N.; Sakakibara, Y. Arch. Microbiol. 1994,
1
(
62, 249.
15) Proschak, A.; Lubuta, P.; Gru
Berardinis, V.; Bode, H. B. ChemBioChem 2013, 14, 633.
16) Penwell, W. F.; DeGrace, N.; Tentarelli, S.; Gauthier, L.;
̈
n, P.; Lohr, F.; Wilharm, G.; De
̈
(
Gilbert, C. M.; Arivett, B. A.; Miller, A. A.; Durand-Reville, T. F.;
Joubran, C.; Actis, L. A. ChemBioChem 2015, 16, 1896.
(17) Antunes, L. C. S.; Imperi, F.; Towner, K. J.; Visca, P. Res.
Microbiol. 2011, 162, 279.
ASSOCIATED CONTENT
Supporting Information
■
*
S
(18) Shapiro, J. A.; Wencewicz, T. A. ACS Infect. Dis. 2016, 2, 157.
(19) Shapiro, J. A.; Wencewicz, T. A. Metallomics 2017, 9, 463.
(20) Bohac, T. J.; Shapiro, J. A.; Wencewicz, T. A. ACS Infect. Dis.
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.or-
glett.8b02789.
2017, 3, 802.
Detailed experimental procedures for physicochemical
characterization, biological activity assay, energy calcu-
lation, chemical synthesis of compounds, and NMR
spectral data (PDF)
(21) Wuest, W. M.; Sattely, E. S.; Walsh, C. T. J. Am. Chem. Soc.
2009, 131, 5056.
(
22) Takeuchi, Y.; Ozaki, S.; Satoh, M.; Mimura, K.-I.; Hara, S.-I.;
Abe, H.; Nishioka, H.; Harayama, T. Chem. Pharm. Bull. 2010, 58,
552.
23) Actis, L. A.; Fish, W.; Crosa, J. H.; Kellerman, K.; Ellenberger,
1
(
AUTHOR INFORMATION
Corresponding Author
S. R.; Hauser, F. M.; Sanders-Loehr, J. J. Bacteriol. 1986, 167, 57.
(24) Dorsey, C. W. Microbiology 2004, 150, 3657.
■
*
(
(
25) Zamri, A.; Abdallah, M. A. Tetrahedron 2000, 56, 249.
26) Kim, J.; Lee, J. E.; Ree, H.; Kim, H. J. Bull. Korean Chem. Soc.
E-mail: hakkim@korea.ac.kr.
ORCID
2015, 36, 439.
27) Ree, H.; Kim, J.; Song, W. Y.; Lee, J. E.; Kim, H. J. Bull. Korean
Chem. Soc. 2015, 36, 1520.
28) Nelson, K. M.; Salomon, C. E.; Aldrich, C. C. J. Nat. Prod.
012, 75, 1037.
29) Jalal, M. A. F.; Hossain, M. B.; van der Helm, D.; Sanders-
(
Hak Joong Kim: 0000-0002-3213-2230
Author Contributions
∥H.L. and Y.W.S. contributed equally.
(
2
(
Notes
Loehr, J.; Actis, L. A.; Crosa, J. H. J. Am. Chem. Soc. 1989, 111, 292.
(30) Oh, M. H.; Lee, J. C.; Kim, J.; Choi, C. H.; Han, K. Appl.
Environ. Microbiol. 2015, 81, 3357.
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
This study was supported by grants from the National
Research Foundation of Korea (NRF-2015R1D1A1A01-
■
0
56815, NRF2010-0020209) and the Korea Health Technol-
ogy R&D Project through the Korea Health Industry
Development Institute (KHIDI), funded by the Ministry of
Health & Welfare, Republic of Korea (Grant Number:
HI17C1657). K.K. is also thankful for the support from IBS-
R023-D1.
REFERENCES
■
(
1) Wong, D.; Nielsen, T. B.; Bonomo, R. A.; Pantapalangkoor, P.;
Luna, B.; Spellberg, B. Clin. Microbiol. Rev. 2017, 30, 409.
(
2) Dijkshoorn, L.; Nemec, A.; Seifert, H. Nat. Rev. Microbiol. 2007,
, 939.
3) Vila, J.; Marti, S.; Sanchez-Cespedes, J. J. Antimicrob. Chemother.
007, 59, 1210.
4) Zgurskaya, H. I.; Lop
015, 1, 512.
5) Song, W. Y.; Jeong, D.; Kim, J.; Lee, M. W.; Oh, M. H.; Kim, H.
J. Org. Lett. 2017, 19, 500.
6) Wilson, B. R.; Bogdan, A. R.; Miyazawa, M.; Hashimoto, K.;
Tsuji, Y. Trends Mol. Med. 2016, 22, 1077.
5
(
2
(
2
́
ez, C. A.; Gnanakaran, S. ACS Infect. Dis.
(
(
(
(
7) Schalk, I. J.; Mislin, G. L. A. J. Med. Chem. 2017, 60, 4573.
8) Wencewicz, T. A.; Miller, M. J. J. Med. Chem. 2013, 56, 4044.
D
DOI: 10.1021/acs.orglett.8b02789
Org. Lett. XXXX, XXX, XXX−XXX