Utilizing an iron(III)-chelation masking strategy to prepare mono- and bis-functionalized aerobactin analogues for targeting pathogenic bacteria

Article abstract of DOI:10.1039/c7cc05197b

A direct and selective method for the functionalization of aerobactin has been described. Selectivity was achieved by masking the functioning carboxylate via iron-chelation, leaving the two remaining carboxylates for direct modification. Both mono- and bis-functionalized aerobactin effectively targeted pathogenic bacteria, showing a facile method with prospective applications.

Full text of DOI:10.1039/c7cc05197b

ChemComm
View Article Online
View Journal | View Issue
COMMUNICATION
Utilizing an iron(III)-chelation masking strategy to
prepare mono- and bis-functionalized aerobactin
analogues for targeting pathogenic bacteria†
Cite this: Chem. Commun., 2017,
53, 9265
Received 5th July 2017,
Accepted 26th July 2017
a
b
Yu-Hin Ho,‡a Sheng-Yang Ho,‡a Cheng-Chih Hsu,
Jiun-Jie Shie
and
a
Tsung-Shing Andrew Wang
*
DOI: 10.1039/c7cc05197b
rsc.li/chemcomm
A direct and selective method for the functionalization of aerobactin has linked to one citric acid through two amide bonds. The two AHL
been described. Selectivity was achieved by masking the functioning hydroxamates and the central a-hydroxyl carboxylate of citrate
carboxylate via iron-chelation, leaving the two remaining carboxylates are the iron(III) chelating moieties. Epidemiological studies have
for direct modification. Both mono- and bis-functionalized aerobactin considered aerobactin to be a virulence factor with certain
effectively targeted pathogenic bacteria, showing a facile method with correlation with bacterial pathogenicity.6 In pathogenic strains
prospective applications.
such as hypervirulent K. pneumoniae7 and uropathogenic E. coli,8 the
occurrence of aerobactin is markedly higher than that in avirulent
Iron, which is a key component of many biological processes, is strains. On the WHO priority pathogens list in 2017,9 carbapenem-
essential to the survival of all beings, including bacteria. Iron resistant and third-generation cephalosporin-resistant Entero-
availability is limited by hosts as a nutrient barrier to prevent bacteriaceae, including K. pneumoniae and E. coli, have been
the proliferation of malicious invaders. To overcome this flagged as ‘‘critical’’. Therefore, it is worthwhile to develop
nutritional barrier during infection, siderophores, a class of functional probes targeting these pathogens.
iron chelators with exceptionally high affinity, are adopted by
Although aerobactin and its cognate receptor, IutA, are
microbial pathogens to scavenge iron from their hosts.1 The frequently found in many pathogenic bacteria, it remained unknown
necessity of siderophore production during infection and their whether aerobactin can be used as a bacterial targeting unit. More
ability to carry iron(III) across the membrane into the cytoplasm specifically, it was not clear how to modify aerobactin to prepare
provide an ideal target for therapeutic applications.2 In fact, functional conjugates that can be recognized by bacteria. Aerobactin
natural siderophore-toxin conjugates, also known as sideromycins, was first isolated in 1969,10 and the first total synthesis of aerobactin
such as microcin E492, albomycin and salmycin, have drawn as was reported in 1982.11 Despite previous synthetic efforts, aerobactin
much attention as their intriguing conjugate design.3 Artificial cannot be used to immediately prepare aerobactin conjugates or
sideromycin mimics have also been synthesized, and many of analogues. To date, no well-defined synthetic aerobactin conjugates
them have shown promising antimicrobial activities.4 Additionally, or analogues have been reported.
siderophore-based vaccine conjugates have recently been demon-
In aerobactin, the two AHL carboxylates are potential modifica-
strated as potential therapeutic alternatives to eliminate bacterial tion sites because they are not involved in iron chelation. However,
infections in animal models.5 However, the challenge of the the free central a-hydroxyl carboxylate in citrate poses a synthetic
general synthesis of diverse siderophore conjugates has prevented challenge for selective modification without a sophisticated protec-
their applications.
tion strategy. Alternatively, we turned our focus on the well-studied
Aerobactin is a representative example of a hydroxamate– iron(III) coordination of aerobactin.12 Copper(II) chelation has been
citrate hybrid siderophore that is widely used by both terrestrial widely used to mask a-amino acids, rendering robust and selective
and marine bacteria. This hexadentate chelator is composed of side chain modifications.13 In our case, we envisioned that the
two N6-acetyl-N6-hydroxy-L-lysines (AHLs) that are symmetrically central a-hydroxyl carboxylate in the citrate moiety can be masked
by iron(III) chelation. In other words, the central carboxylate could be
a Department of Chemistry, National Taiwan University, Taipei 10617,
considered as being ‘‘protected’’ by iron(III). Using this chelation
strategy, we might be able to directly and selectively modify native
aerobactin.
To obtain native aerobactin, we revised the preparation of
aerobactin (Scheme 1). For the citric acid portion, the reported
t-butyl citrate protected at the central carboxylate14 was chosen
Taiwan. E-mail: wangts@ntu.edu.tw
b Institute of Chemistry, Academia Sinica, Taipei 11529, Taiwan
† Electronic supplementary information (ESI) available: Synthesis, purification and
characterization of compounds; photolysis; and bacterial imaging experiments. See
DOI: 10.1039/c7cc05197b
‡ These authors contributed equally to the work.
This journal is ©The Royal Society of Chemistry 2017
Chem. Commun., 2017, 53, 9265--9268 | 9265

View Article Online
Communication
ChemComm
Scheme 2 The iron(III)-chelation masking strategy for the synthesis of
mono- and bis-substituted aerobactin analogues. See the ESI† for the
extended substrate scope.
Scheme 1 Alternative chemical synthesis of aerobactin.
due to the easy removal of its by-product dicyclohexylurea
(DCU). The mono-substituted compound could not be obtained
to prevent undesired imide formation and subsequent amide as the major end-product by simply controlling the amounts of
shift.15 For the AHL portion, we started with L-lysine using coupling reagents or linkers. Instead, it could only be obtained
modified methods from previous analogous schemes.16 In by monitoring the reaction progress (Fig. S4, ESI†). Using
short, two amino groups were sequentially protected by Boc optimal iron(III)-masking conditions, we were able to prepare
and then by Cbz groups. The esterification was achieved by the mono- and bis-substituted aerobactin analogues 11a and
treatment with SOCl2 and EtOH at low temperatures; Boc can 11b in separate reactions. These analogues can be purified by
be removed at the same time. Then, without purification, the HPLC on the milligram scale with 490% purity (Fig. S2 and S3,
resulting product 3 was subjected to benzoyl peroxide (BPO) ESI†). Interestingly, the purified mono-substituted analogue 11a
oxidation17 under optimized buffer conditions to give the exhibited two close HPLC peaks with identical masses, which is
desired hydroxylamine derivatives 4. After acetylation, the fully consistent with mono-substitution at two AHL carboxylates
protected AHL 5 was obtained in good yield and after only four generating a pair of diastereomers (Fig. S1, ESI†). This result
separation steps.
strengthened the iron(III)-chelation masking strategy.
Using dicyclohexylcarbodiimide (DCC), the t-butyl citrate
To verify that the central a-hydroxyl carboxylate remains
was pre-activated with N-hydroxysuccinimide (HOSu) to give 7. intact after amide coupling, we observed the photoreactivity of
Cbz-deprotection of 5 was achieved with HBr in AcOH to afford the a-hydroxyl carboxylate in many siderophores, such as
6 for immediate use.16 The typical Cbz-deprotection by catalytic aerobactin, vibrioferrin, and petrobactin.18 Structural analysis
hydrogenation led to complex mixtures, where the intra-molecular by NMR was not feasible due to the strong paramagnetic iron(III)
transacetylation was the major side product. The construction of bound in the analogues. Iron removal was time-consuming and
aerobactin was accomplished by amide bond formation between 6 not necessary because iron-bound aerobactin analogues can be
and 7. To prevent amide shift, a slight excess of 6 was introduced. used directly in biological tests. Photolysis was facilitated by the
Finally, after global deprotection, aerobactin was obtained in 32% chelated iron(III), which was reduced to iron(II), dissociated and
overall yield from L-lysine.
oxidized back to iron(III) by air before being re-chelated with the
We then investigated the feasibility of functionalization by product (Fig. 1a). Only the iron-bound a-hydroxyl carboxylate
iron(III)-chelation masking (Scheme 2). To test this idea, a set of can undergo photolysis, unlike the AHL carboxylates. 10, 11a,
amino linkers with different functional groups was examined. and 11b were irradiated with 419 nm light and then subjected to
In the typical coupling reactions, aerobactin was mixed with HPLC/MS analysis. All three compounds gave the corresponding
different amounts of linkers and coupling regents (Table S1, photolysis products (Fig. 1 and Fig. S5, S6, ESI†). LC/MS/MS
ESI†). The reactions were monitored directly by LC/MS. Notably, analysis for the mono- and bis-substituted 11a and 11b was also
in the presence of 2 equivalents of iron(III), only the mono- and conducted to verify the regioselectivity of the amide coupling. The
bis-substituted amide products were detected, even when excess results were consistent with photolysis (Fig. S7 and S8, ESI†).
linkers (10–40 equiv.) and coupling reagents (100 equiv.) were
We next examined whether the mono- and bis-substitution
used; no tri-substituted products were observed. In contrast, in affected the iron-uptake ability of aerobactin. Neither substitution
the absence of iron(III), tri-substituted products were easily affected the iron binding, as shown in the characteristic red
observed, and similar results were observed when iron(III) was iron-bound complex due to the charge-transfer absorption. To
insufficient (0.5 equiv.). These results clearly indicated that determine whether the mono- or bis-substitution affected bacterial
iron(III) masked one carboxylate from amide coupling. When binding, we prepared aerobactin–fluorophore conjugates by click
excess coupling reagents were used, DCC was a better option chemistry. Although direct functionalization required fewer steps,
9266 | Chem. Commun., 2017, 53, 9265--9268
This journal is ©The Royal Society of Chemistry 2017

View Article Online
ChemComm
Communication
Fig. 2 (a) Functionalization of the azido aerobactin analogues 11a and 11b
with FL-DIBO by SPAAC. (b) Microscopy images of K. pneumoniae treated
with mono-FL-Aero and bis-FL-Aero under different iron supplementations
(inset scale bar: 1 mm).
Fig. 1 Photolysis reaction of bis-C5N3 Aero 11b. (a) The general photolysis
scheme on the a-hydroxyl carboxylate moiety in the presence of iron(III).
(b) HPLC chromatograms for the photolysis of bis-C5N3 Aero at 210 nm
and 500 nm. (c) MS spectra of bis-C5N3 Aero 11b and its corresponding
photolysis product 11b*.
With the gene cluster located at the pathogenicity islands,
click chemistry was superior due to its synthetic flexibility. Both enteric bacteria secrete aerobactin as their virulence factor. To
clickable azido analogues 11a and 11b (Scheme 2) were prepared examine the selective targeting of our aerobactin conjugates, we
to conduct copper(I)-catalyzed azide–alkyne cycloaddition (CuAAC) surveyed several species of bacteria (Fig. 3). Four pathogenic
or strain-promoted azide–alkyne cycloaddition (SPAAC). The Gram-negative bacteria, including E. coli O157:H7, E. coli
reactivity of the iron-bound aerobactin analogues in CuAAC CFT073, S. enterica and V. cholerae, were tested. E. coli and S. enterica
varied, and it was difficult to find proper click pairs and exhibited high labeling efficiencies by both mono-FL-Aero and
conditions. Instead, SPAAC gave more reliable click products. bis-FL-Aero. Consistent with a previous report, V. cholerae was
Thus, a recently reported symmetric fluorogenic FL-DIBO was not labeled by either of our probes.20 Moreover, two Gram-
synthesized19 and then clicked by SPAAC to prepare two fluorescent positive bacteria, B. subtilis and S. epidermidis, were tested as
conjugates, mono-FL-Aero and bis-FL-Aero, both with purities negative controls because aerobactin is thought to be utilized
490% by HPLC (Fig. S2 and S3, ESI†).§
only in Gram-negative bacteria. As expected, these two bacteria
The bacterial targeting abilities of mono-FL-Aero and bis-FL- showed no labeling by either probe. When it comes to the
Aero were first tested in pathogenic K. pneumoniae. The bacteria epidemiologically important pathogens, P. aeruginosa and S. aureus,
were cultured in low iron (1 mM) minimal media (MM9) to which were also tested, the results were inconclusive because their
induce the siderophore-mediated iron-uptake pathway before autofluorescence overlapped with the FL-DIBO emission. Never-
incubation with aerobactin–fluorophore conjugates. mono-FL-Aero theless, our results demonstrated the good bacterial targeting
was able to label K. pneumoniae with high efficiency, suggesting that feature of aerobactin in many pathogens.
the non-chelating AHL carboxylate was a proper conjugation site
In this study, we have developed an iron(III)-chelation masking
(Fig. 2b). Surprisingly, bis-FL-Aero also labeled K. pneumoniae with strategy to prepare mono- and bis-functionalized analogues of
good efficiency, implying that neither AHL carboxylates interfered native aerobactin without using complicated protecting protocols.
with receptor binding. Subsequently, high levels of iron (100 mM) This strategy could be generalized to other native siderophores
were introduced during culture preparation to turn off the side- carrying non-chelating functional groups, such as staphyloferrin A
rophore pathway. No significant fluorescence labeling was observed from S. aureus.21 Currently, the reactions were limited to the few
for either conjugate, indicating that the uptake process was due to milligram scale by HPLC, but their expansion to a larger scale is
the iron-dependent siderophore system. This also indicated that the possible after optimizing the separation schemes. Both mono- and
clicked FL-DIBO did not bind to the bacteria non-specifically.
bis-functionalized aerobactin conjugates were able to label a wide
This journal is ©The Royal Society of Chemistry 2017
Chem. Commun., 2017, 53, 9265--9268 | 9267

View Article Online
Communication
ChemComm
Notes and references
§ Commercially available dibenzocyclooctyl (DBCO) reagents gave
two regioisomers by SPAAC, which complicated the mono- and bis-
functionalized probe preparations.
1 (a) C. P. Doherty, J. Nutr., 2007, 137, 1341–1344; (b) A. M. Ganzoni
and M. Puschmann, Ann. Hematol., 1975, 31, 313–322; (c) E. P. Skaar,
PLoS Pathog., 2010, 6, e1000949.
´
´
2 (a) C. Ji, R. E. Juarez-Hernandez and M. J. Miller, Future Med. Chem.,
¨
2012, 4, 297–313; (b) U. Mollmann, L. Heinisch, A. Bauernfeind,
¨
T. Kohler and D. Ankel-Fuchs, BioMetals, 2009, 22, 615–624;
(c) M. Miethke and M. A. Marahiel, Microbiol. Mol. Biol. Rev., 2007,
71, 413–451; (d) M. Ballouche, P. Cornelis and C. Baysse, Recent Pat.
Anti-Infect. Drug Discovery, 2009, 4, 190–205.
¨
3 (a) V. Braun, A. Pramanik, T. Gwinner, M. Koberle and E. Bohn,
BioMetals, 2009, 22, 3; (b) G. Gause, Br. Med. J., 1955, 2, 1177; (c) V. De
Lorenzo and A. Pugsley, Antimicrob. Agents Chemother., 1985, 27,
´
666–669; (d) X. Thomas, D. Destoumieux-Garzon, J. Peduzzi,
C. Afonso, A. Blond, N. Birlirakis, C. Goulard, L. Dubost, R. Thai
and J.-C. Tabet, J. Biol. Chem., 2004, 279, 28233–28242.
¨
4 (a) T. A. Wencewicz, T. E. Long, U. Mollmann and M. J. Miller,
Bioconjugate Chem., 2013, 24, 473–486; (b) T. Zheng and E. M. Nolan,
J. Am. Chem. Soc., 2014, 136, 9677–9691; (c) M. J. Miller and
F. Malouin, Acc. Chem. Res., 1993, 26, 241–249; (d) A. A. Lee,
Y. C. S. Chen, E. Ekalestari, S. Y. Ho, N. S. Hsu, T. F. Kuo and
T. S. A. Wang, Angew. Chem., Int. Ed., 2016, 55, 12338–12342.
5 (a) M. Sassone-Corsi, P. Chairatana, T. Zheng, A. Perez-Lopez,
R. A. Edwards, M. D. George, E. M. Nolan and M. Raffatellu, Proc.
Natl. Acad. Sci. U. S. A., 2016, 113, 13462–13467; (b) L. A. Mike,
S. N. Smith, C. A. Sumner, K. A. Eaton and H. L. Mobley, Proc. Natl.
Acad. Sci. U. S. A., 2016, 113, 13468–13473.
6 (a) V. De Lorenzo and J. Martinez, Eur. J. Clin. Microbiol. Infect. Dis.,
1988, 7, 621–629; (b) X. Nassif and P. J. Sansonetti, Infect. Immun.,
1986, 54, 603–608.
7 T. A. Russo, R. Olson, U. MacDonald, J. Beanan and B. A. Davidson,
Infect. Immun., 2015, 83, 3325–3333.
8 S. M. Payne and I. Neilands, CRC Crit. Rev. Microbiol., 1988, 16,
81–111.
9 C. Willyard, Nature, 2017, 543, 15.
Fig. 3 Bacteria labeling experiments with the mono-FL-Aero and bis-FL-Aero
conjugates under iron-limiting conditions (left: fluorescence channel and right:
merge) (20 mM of either probe, cultured in MM9 minimal medium, more details
are provided in the ESI†) (inset scale bar: 1 mm).
10 F. Gibson and D. Magrath, Biochim. Biophys. Acta, 1969, 192, 175–184.
11 P. J. Maurer and M. J. Miller, J. Am. Chem. Soc., 1982, 104, 3096–3101.
12 W. R. Harris, C. J. Carrano and K. N. Raymond, J. Am. Chem. Soc.,
1979, 101, 2722–2727.
13 S. Wiejak, E. Masiukiewicz and B. Rzeszotarska, Chem. Pharm. Bull.,
1999, 47, 1489–1490.
14 R. A. Gardner, G. Ghobrial, S. A. Naser and O. Phanstiel IV, J. Med.
Chem., 2004, 47, 4933–4940.
15 A. Ghosh and M. J. Miller, J. Org. Chem., 1993, 58, 7652–7659.
16 K. A. Fennell and M. J. Miller, Org. Lett., 2007, 9, 1683–1685.
17 Q. X. Wang, J. King and O. Phanstiel IV, J. Org. Chem., 1997, 62,
8104–8108.
18 (a) F. C. Ku¨pper, C. J. Carrano, J.-U. Kuhn and A. Butler, Inorg. Chem., 2006,
45, 6028–6033; (b) S. A. Amin, D. H. Green, F. C. Ku¨pper and C. J. Carrano,
Inorg. Chem., 2009, 48, 11451–11458; (c) K. Barbeau, G. Zhang, D. H. Live
and A. Butler, J. Am. Chem. Soc., 2002, 124, 378–379; (d) A. Butler and
R. M. Theisen, Coord. Chem. Rev., 2010, 254, 288–296.
range of pathogenic bacteria. Our results also suggested the
possibility of delivering more than one cargo by aerobactin at
the same time. Aerobactin conjugates carrying two synergistic
toxins might promote enhanced bacterial killing effects. For
future therapeutic applications, the cellular locations of mono- and
bis-functionalized conjugates after uptake should be confirmed as
aerobactin analogues might be acquired into bacteria or stuck on
surface receptors.
We thank the Ministry of Science and Technology, Taiwan
(MOST 104-2119-M-002-021 and 105-2113-M-002-003 to T.-S. A. W.;
MOST 105-2113-M-002-004-MY2 to C.-C. H.) and the Center for
Emerging Material and Advanced Devices, National Taiwan
University (NTU-ERP-106R880210) for financial support.
19 F. Friscourt, C. J. Fahrni and G.-J. Boons, J. Am. Chem. Soc., 2012,
134, 18809–18815.
20 T. Funahashi, T. Tanabe, H. Aso, H. Nakao, Y. Fujii, K. Okamoto,
S. Narimatsu and S. Yamamoto, Microbiology, 2003, 149, 1217–1225.
¨
21 S. Konetschny-Rapp, G. Jung, J. Meiwes and H. Zahner, Eur.
J. Biochem., 1990, 191, 65–74.
9268 | Chem. Commun., 2017, 53, 9265--9268
This journal is ©The Royal Society of Chemistry 2017