Ferrichrome Has Found Its Match: Biomimetic Analogues with Diversified Activity Map Discrete Microbial Targets

Article abstract of DOI:10.1002/chem.201702647

Siderophores provide an established platform for studying molecular recognition principles in biological systems. Herein, the preparation of ferrichrome (FC) biomimetic analogues varying in length and polarity of the amino acid chain separating between the tripodal scaffold and the pendent Fe^III chelating hydroxamic acid groups was reported. Spectroscopic and potentiometric titrations determined their iron affinity to be within the range of efficient chelators. Microbial growth promotion and iron uptake studies were conducted on E. coli, P. putida and U. maydis. A wide range of siderophore activity was observed in the current series: from a rare case of a species-specific growth promotor in P. putida to an analogue matching FC in cross-phylum activity and uptake pathway. A fluorescent conjugate of the broad-range analogue visualized siderophore destination in bacteria (periplasmic space) vs. fungi (cytosol) mapping new therapeutic targets. Quantum dots (QDs) decorated with the most potent FC analogue provided a tool for immobilization of FC-recognizing bacteria. Bacterial clusters formed around QDs may provide a platform for their selection and concentration.

Full text of DOI:10.1002/chem.201702647

A Journal of
Accepted Article
Title: Ferrichrome has found its match: biomimetic analogs with
diversified activity map discrete microbial targets
Authors: Jenny Besserglick, Evgenia Olshvang, Agnieszka
Szebesczyk, Joseph Englander, Dana Levinson, Yitzhak
Hadar, Elzbieta Gumienna-Kontecka, and Abraham Shanzer
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Chem. Eur. J. 10.1002/chem.201702647
Link to VoR: http://dx.doi.org/10.1002/chem.201702647
Supported by

Chemistry - A European Journal
10.1002/chem.201702647
FULL PAPER
Ferrichrome has found its match: biomimetic analogs with
diversified activity map discrete microbial targets
Jenny Besserglick,[a] Evgenia Olshvang, Agnieszka Szebesczyk, Joseph Englander,[a] Dana Levinson,[c] Yitzhak Hadar,[c] Elzbieta
[a]
[b]
[
b]
[a]
Gumienna-Kontecka* , and Abraham Shanzer*
Abstract: Siderophores provide an established platform for studying
molecular recognition principles in biological systems. Herein we
describe the preparation of ferrichrome biomimetic analogs varying in
length and polarity of the amino acid chain separating between the
Ferrichrome[8] (FC) is an archetype of
a
natural
hydroxamate siderophore. Although of fungal origin, it can be
recognized by numerous bacterial species.[9] It is composed of a
cyclic hexapeptide ring template, non-symmetrically extended by
three identical L-ornithine derived arms each terminated by a
hydroxamate moiety. Together, the three bidentate hydroxamates
delineate an octahedral cavity matching the geometrical
requirement of the ferric ion.
Although over the years some elegant synthetic routes to
native siderophores were demonstrated,[10] they are still often
complex and tedious to synthesize.[11] Let alone to obtain their
conjugates,[12] due to lack of residual moieties for functionalization
with fluorescent markers, surface active molecules or drugs. In
this respect, FC is no different from other siderophores.
III
tripodal scaffold and the pendent Fe chelating hydroxamic acid
groups. Spectroscopic and potentiometric titrations determined their
iron affinity to be within the range of efficient chelators. Microbial
growth promotion and iron uptake studies were conducted on E.
coli, P. putida and U. maydis. A wide range of siderophore activity
was observed in the current series: from a rare case of a species-
specific growth promotor in P. putida to an analog matching
ferrichrome in cross-phylum activity and uptake pathway.
A
fluorescent conjugate of the broad-range analog visualized
siderophore destination in bacteria (periplasmic space) vs fungi
(
cytosol) mapping new therapeutic targets. Quantum Dots (QD)
An alternative approach attempts to reproduce the function
rather than the detailed structure of a biological substrate. Once
a biologically active scaffold is found, it can be further diversified
and tuned. This method, known as biomimetic chemistry, has
been applied for preparation of artificial siderophores,[9, 14] and in
our group we have focused on synthetic FC analogs.[15] Their
structure relies on four basic elements: (i) the C3 symmetric tris–
carboxylate template, replacing the native FC hexapeptide; (ii)
template extension with natural amino acid derived arms; (iii)
decorated with the most potent FC analog provided a tool for
immobilization of ferrichrome-recognizing bacteria. Bacterial clusters
formed around QDs may provide a platform for their selection and
concentration.
[13]
Introduction
retrohydroxamate ligand, with order reversed with respect to the
native hydroxamate moiety;[16] (iv) apical site for efficient
conjugation, that interferes neither with iron binding, nor with
receptor recognition.
This approach resulted in efficient preparation of multiple
FC analogs with biological activity and their enantio-pure iron
Siderophores are low molecular weight organic chelators, bio-
synthesized and secreted by microorganisms to sequester
environmental iron - a vital nutrient for a variety of essential
cellular processes.[1] Often the iron-siderophore complex is
recognized by specialized microbial outer-membrane receptors,
which trigger its uptake by a cascade of transporting events.[2]
Finally the iron is released in the cytosol and the siderophore is
usually recycled to acquire additional iron.[3] While the iron uptake
mechanism has been extensively studied[1b, 4] their recycling
processes mostly remain elusive.[5]
complexes
preferences.
for
exploring
transporters’
stereo-specific
[17] [18]
Moreover, analogs conjugated with surface
[19]
adhesive functionalities, drug candidates and fluorescent tags
for visualizing their ‘trafficking’ and compartmentalization have
been prepared. Owing to their modularity, tripodal FC analogs
Siderophores usually encapsulate the ferric ion by a variety
were successfully utilized beyond siderophore biology, for
example as molecular switches[20] and sensors.[21]
So far, this methodology did not overcome some inherent
limitations including: (i) low overall number of FC analogs
capable of promoting microbial growth, particularly in E. coli and
[
6]
of hard donor bidentate ligands, most common being catechols,
hydroxamates[ and α-hydroxy carboxylic acids.[4a]
7]
(
ii) the fact that activity across multiple species, similar to native
[
a]
Dr. J. Besserglick, Dr. E. Olshvang, Dr. J. Englander, Prof. Dr. A.
Shanzer; Department of Organic Chemistry,
The Weizmann Institute of Science, Rehovot 7610001, Israel.
E-mail: abraham.shanzer@weizmann.ac.il
Dr. A. Szebesczyk, Dr. E. Gumienna-Kontecka; Faculty of Chemistry,
University of Wrocław, F. Joliot-Curie 14, 50-383 Wrocław, Poland.
E-mail: elzbieta.gumienna-kontecka@chem.uni.wroc.pl
D. Levinson ,Prof. Dr. Y. Hadar;
Department of Plant Pathology and Microbiology, The R.H. Smith
Faculty of Agriculture Food and Environment
FC, was obtained only sporadically.
To overcome these limitations, we resorted to the detailed
structure of FC in its transporter of E. coli - Ferric Hydroxamate
Uptake protein component A (FhuA). FhuA is still among the few
siderophore transporters with a resolved crystal structure.[22] The
model derived from X-ray diffraction analysis designates tight
contacts between the protein transporter and the iron binding
moiety of FC (please refer to Supporting information, Figure S1).
We assumed that these contacts might have been disrupted by
bulky groups in the immediate vicinity to the metal chelating
[
[
b]
c]
The Hebrew University of Jerusalem, Rehovot 7610001, Israel.
Supporting information for this article is given via a link at the end of
the document.
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201702647
FULL PAPER
center in our previous α-amino acid analogs leading to
diminishing of transporter recognition as bulkiness increased.
Additionally, the crystallographic data indicate the presence of
two sets of H-bonding network dominating the recognition
event.[22b, 23] The first H-bonding network links between the
hydroxamate oxygens and proton-donors within the transporter;
while the second network connects the ring backbone in FC with
the transporter. Keeping α position unsubstituted and introducing
moieties capable of hydrogen bonding were expected to expend
the number of biologically active FC analogs. Thus, replacing α-
amino acid hydroxamates with their β and γ equivalents tackles
both complementary approaches simultaneously: introducing
polar substituents for enhanced hydrogen bonding to the
transporter, yet moving them further away from the metal binding
in FC, while the other contains chiral centers and possesses
peripheral functional groups.
Detailed structures of all the analogs in this series are
provided in Chart 1. The simplest β-amino acid, β-alanine, and
the simplest γ-amino acid, γ-aminobutyric acid (GABA), yielded
the structural skeleton of the series represented by analogs 1 and
2 respectively. Utilization of natural aspartic and glutamic amino
acids led to chiral structures with peripheral carboxylates (analogs
3 and 4). A non-natural chiral amino acid, homo-β-alanine,
incorporates a methyl group instead of a carboxylate at the β
position to the hydroxamate (analog 5). Replacement of
carboxylates by methyl groups allows the comparison of steric
effects at the β position in homo-β-alanine derived tripod 5 to
steric and electrostatic effects of the carboxylates in aspartic
derived tripod 3. All chiral analogs (3-5) were prepared as a pair
of enantiopure compounds.
site.
A
recent example of analogs with
β
positioned
[
24]
hydroxamates indicates a significant potential of this approach.
In this paper, we present a new series of improved FC
analogs. The series comprises eight new tripodal iron complexes,
designed to incorporate a number of variable structural elements
in order to investigate their influence on physico-chemical
properties, as well as on their microbial uptake process. In terms
of iron affinity, the pFe[25] values of most analogs are within the
range of typical siderophores. However, in terms of siderophore
activity more variability is observed: from a rare case of species-
specificity to a broad-spectrum activity similar to native FC. While
the first may find application in bacterial diagnostics the second,
with the broad-spectrum activity, may offer future therapeutic
advantages.
Chart 1. Structures of FC and its β- and γ- amino acid biomimetic analogs:
Fluorescently marked broad-activity analog was followed in
several mutants of FC transport cascade in E. coli and its pathway
proved to be indistinguishable from that of FC, thus making it a
[A] Structure of the native FC. [B] Schematic representation of the biomimetic
FC analogs. Analog 1 is derived from β-alanine amino acid, R=NH , R’=H, n=1;
2
Analog 2 - GABA amino acid, R=Me, R’=H, n=2; Analog 3- aspartic amino acid,
R=Me, R’= COOH, n=1; Analog 4- glutamic amino acid, R=Me, R’=COOH, n=2:
Analog 5- homo-β-alanine amino acid, R=Me, R’=Me, n=1.
‘
truly’ biomimetic analog. Same fluorescently labeled analog
provided a real-time imaging tool for distribution patterns inside
bacteria and fungi subsequent to iron uptake. These images
describe quite different ‘trafficking’ and compartmentalization,
identify discrete targets and suggest utilization of diverse drug-
conjugates for differential activities.
We also show that our most versatile FC analog attached to
Quantum Dots (QD) generates an adhesive and emissive
platform ready for tethering microorganisms possessing the FC
transporter. Coupled with other siderophores or their analogs, this
platform carries the potential to overcome the formidable need to
Most analogs in the series were prepared by a convergent
synthetic route combining three building blocks (Scheme 1). First,
the templates were assembled by reacting triols 6 or 7 with tert-
Butyl acrylate. When triol 7 was utilized, the apical amine was
protected with Alloc prior to hydrolysis of the t-Butyl esters, which
yields the deprotected triacids 12 and 13. The second building
block, namely N-methyl O-benzylhydroxylamine hydrochloride 18
was prepared as described elsewhere.[26] Next, it was coupled to
commercially available N-Boc protected amino acids 14-17 using
CDI to obtain the protected arms 19-22. After Boc removal these
arms were coupled to a suitable template to obtain O-Benzyl
protected tripods 27 and 29-31. Hydroxamic acids were finally
deprotected using hydrogenolysis to yield the final analogs 1-3
and 5. When triacid 12 was utilized Alloc was removed prior to
hydrogenolysis (compound 28).
find
distinct
growth
conditions
for
difficult-to-culture
microorganisms. The QD derived methodology allows ‘fishing’
bacteria, based on siderophore content, thus providing rapid
means for potential microbial diagnostics and separation.
Taken together the presented results not only shed light on the
insufficiently studied fate of siderophores subsequent to iron
delivery, but also provide a tool for identifying new therapeutic
targets, distinct for different microorganisms.
The convergent synthetic route failed in the case of analog
4. The coupling step between the corresponding deprotected arm
Results and Discussion
Design and synthesis. In the current series of FC analogs two
sets of compounds derived from β- and γ- amino acids were
prepared; one containing long unsubstituted arms similar to those
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201702647
FULL PAPER
Scheme 1. Convergent synthetic route to FC tripodal analogs:
Reagents and conditions: (a) NaOHaq, TBAI; (b) Allyl chloroformate, NaOH (aq.), THF, 0°C→r.t.; (c) 4N HCl in dioxane; (d) CDI, DIPEA, THF; (e) 4N HCl in
o
dioxane, 0 C; (f) CDI, DIPEA, THF; (g) 1.2 mol% Pd[Ph
3
P], PhSiH
3
, DCM; (h) H
2
, Pd/C, MeOH.
and triacid 13 led to intramolecular cyclization, forming a
pyroglutamic acid derivative instead of the desired tripod. Thus,
this analog was obtained by a linear synthetic route (Scheme 2)
starting from previously described glutamic acid derivative 32
with γ-carboxylate protected by trimethylsilyl ethanol (TMSE).
After coupling to template 13 TMSE was removed by TBAF to
yield triacid 34, which was immediately coupled to
hydroxylamine 18 using CDI to obtain the benzyl protected tripod
For additional studies analog 1 was either loaded on
CdSe/CdS/CdZnS/ZnS QDs (Scheme 3) or labeled with a
naphthalimide fluorescent dye 42 (Scheme 4). Synthesis of the
QD conjugate began from fully protected β-alanine derived tripod
27.
Scheme 3. Conjugation of β-alanine derived analog to CdSe/CdS QDs.
35. Finally, the benzyl groups were removed by hydrogenolysis
to yield the desired analog 4.
Scheme 2. Linear synthetic route to glutamic acid derived analog 4.
Reagents and conditions: (a) CDI, DIPEA, THF; (b) 1N TBAF in THF; (c)
Reagents and conditions: (a) 1.2 mol% Pd[Ph
DIPEA, THF. Yield: 66% for two steps; (c) H
eq Fe3+, MeOH. Quant. yield; (e) EDC, KOH, H
3
P], PhSiH
, Pd/C, MeOH. Quant. yield; (d)
O/MeOH.
3
, DCM; (b) NHS, DIC,
hydroxylamine 18, CDI, DIPEA, THF; (d) H
2
, Pd/C, MeOH.
2
1
2
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201702647
FULL PAPER
Scheme 4. Labeling of β-alanine derived analog with a naphthalimide fluorescent dye.
3 3 2
Reagents and conditions: (a) 1.2 mol% Pd[Ph P], PhSiH , DCM; (b) Succinic anhydride, TEA, THF; (c) HBTU, DIPEA, DCM: DMF 1:1; (d) H , Pd/C, MeOH.
Following Alloc removal it was coupled to N-Cbz protected
caproic acid, thus introducing a flexible linker in order to separate
between the analog and the QDs' surface. The obtained
intermediate 36 was subjected to hydrogenolysis to remove
protection from both the hydroxamic acids and the terminal
amine (compound 37). Next, the free hydroxamic acids were
complexed with ferric chloride (intermediate 38). The iron
complex served as a protecting group for hydroxamic acids and
later was used for further experiments with bacteria. Finally, the
iron complex 38 was loaded on the QDs by coupling its terminal
amine to multiple carboxylates on the surface of each particle to
obtain β-alanyl tripod decorated QDs (Compound 40).
and using known values of protonation constants of respective
ligands (Table S1), EDTA and stability constant of Fe-EDTA.[
For analogs 3 and 4 some instability of the ferric complexes was
observed before the equilibrium with EDTA could be attained,
and the constants were not reliable (not shown). Therefore, the
development of different method was necessary. A combination
of pH-dependent spectrophotometric and potentiometric
titrations was applied. The logKFeL are given in Table 1. Close
values of logKFeL of analogs 2 and 5 determined by EDTA
29]
competition
experiments
and
pH-dependent
spectrophotometric/potentiometric titrations (Table 1) indicate
the reliability of the second method.
Fluorescent labeling of β-alanine derived analog (Scheme
) started from deprotecting the apical amine of compound 27.
Table 1: Overall stability constants (logKFeL) and pFe values for studied
complexes with iron (III) ions[
a]
4
This amine was further reacted with succinic anhydride to
introduce a short linker for dye connection (compound 41). The
carboxylate of the linker was coupled to the piperazine of the
previously reported dye 42.[27] Finally, the labeled analog 1f was
obtained after deprotection of 43 by hydrogenolysis.
Overall Iron Complex Stability. Successful biomimetic
compounds must efficiently perform the function of the natural
entity they mimic. Siderophores encapsulate the ferric ion with
Analog
logKFeL
pFe
1
2
26.20(5)
28.7(2)[b]
23.3
23.7
28.05(1)
3
4
5
24.4(1)
17.2
19.8
23.3
extraordinary affinity in order to overcome extremely low soluble
24.65(9)
27.8(2)[b]
iron concentrations in the natural environment.[
28]
Thus,
determination of iron complex stabilities of siderophore
biomimetics is vital for their application under natural iron
concentrations and competition with other siderophores.
Results of physico-chemical analysis are summarized
below and in Supporting Information. All tripodal compounds
form mononuclear complexes with ferric ions, and the metal-to-
ligand 1:1 stoichiometry was confirmed by ESI-MS (Table S2)
and spectrophotometry (vide infra).
2
7.06(3)
FC
29.07[c]
25.2[c]
[
a]
[b]
III
Conditions: T = 298K, I = 0.10M NaClO
EDTA competition titration. Conditions: clig = cFe = 1.5·10 M, cEDTA = 0-1.5·10-
4
.
Constants determined from
-4
3
[c]
[30]
M, I = 0.10M NaClO
4
.
Reference.
In order to eliminate the differences in ligand protonation
In order to determine the stability constants of the
constants, which may influence logKFeL, and compare the iron
trihydroxamate forms of discussed ferric complexes, FeL
1 5
- FeL
chelation efficacy between the studied ligands, and parental FC,
(
analogs 1-5), the spectrophotometric competition experiments
the free hexa-aqua-iron(III) concentrations were calculated (pFe
with EDTA were performed at pH 8.0. The overall stability
constants, logKFeL (Table 1), were calculated according to eq. S6,
III
=
-log[Fe (aq)] at pH = 7.4 with a total ligand concentration of
0-5 M and total FeIII concentration of 10-6 M, Table 1).[25, 31]
1
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201702647
FULL PAPER
Comparison of all obtained data reveals that the pFe values
obtained for analogs 1, 2 and 5 are very close, while these of
carboxylate derivatives 3 and 4 are significantly lower. The
analysis of pH-dependent UV-visible spectral profiles of all
compounds exposes an influence of the carboxylates on the
trihydroxamate complex formation. The pH-dependent spectral
characteristics of the ferric-analog 1, 2 and 5 systems clearly
indicate the presence of only two absorbing species along the
pH range studied (0.1-8), i.e. dihydroxamate[32] and
trihydroxamate (Plot S1, Table S4), with equilibria between the
two forms, represented as pKNHOH, ranging from 2.45 for 2 to
native FC was used as a positive control and represents the
maximal growth level in each experiment. Sterile double distilled
water, instead of an iron source, was used as a negative control.
Each enantiomer of chiral analogs 3-5 was tested separately.
Table 2 presents the results as relative growth promotion activity.
Actual growth curves are presented in the Supporting
information (Figure S2). All tested analogs demonstrated growth
promotion activity at least in one bacterial species and the
majority of the tripods were recognized by both bacterial strains
used in this study. Overall, P. putida was less selective and was
able to take advantage of more analogs than E. coli, although
both species perfectly recognize the native FC. Notably, both
enantiomers of chiral analogs 3-5 had similar growth promotion
and no chiral preference was observed in either of the bacterial
strains.
The majority of native siderophores incorporate chiral
building blocks, frequently derived from amino acids.[36]
Consequently, the asymmetric centers of the ligand also direct
the chirality of the octahedral iron(III) complex, which adopts
either Λ or Δ arrangement around the ferric ion, since the two
configurations become energetically nonequivalent. The outer
membrane receptor selectively recognizes and transports iron
complexes of predominantly one configuration.[37] In the case of
FC its outer receptor is built to recognize the Λ arrangement
around the iron, induced by its L-ornithine derived arms.[38]
2.76 for 5 (Table S3). This weak difference indicates that
shortening of the spacer between the hydroxamate group and
the amide moiety, or the presence of uncharged methyl group
nearby the hydroxamate only very slightly affect the binding
properties of the molecule. Presence of the carboxylate in close
proximity to the hydroxamate hampered the binding of third
hydroxamate group, with the pKNHOH of 7.3 and 6.75 for 3 and 4,
respectively (Table S3). Likely, one of the carboxylate groups is
involved in complex formation along with two hydroxamates.
Only after the deprotonation of the third hydroxamic group,
trihydroxamate complex formation is possible. The tridentate
coordination via hydroxamate and carboxylate oxygens was
earlier documented for iron(III) complexes of aspartic acid-- and
glutamic acid--hydroxamic acids.[
33]
In these sytems the
preferential binding of carboxylate in complexes of 1:1 and 1:2
III
Fe :ligand stoichiometry prevented the formation of complexes
Table 2: Bacterial growth promotion of the FC tripodal analogs in the current
with 1:3 ratio, typical for monohydroxamates. In addition, when
unbound the carboxylates might destabilize the ferric complexes
through induction of electrostatic repulsion and/or steric
constrains.
Overall, the pFe values calculated for analogs 1, 2 and 5
are less than two orders of magnitude lower than of FC. However,
pFe of analogs with carboxylate groups (3 and 4) are up to eight
orders of magnitude lower, indicating a destabilizing role of
carboxylates in encapsulation of ferric ions by the hydroxamates.
It is worth emphasizing that at pH 7.4, there are still mixed
carboxylate-hydroxamate complexes present (Plot S4).
Nevertheless, pFe values for all studied tripodal FC analogs
remain in the range for effective iron chelators.
series.
Iron sources
Amino acid [a]
Relative growth promotion
E. coli
P. putida
++++
++++
++++
++
FC- FeIII
++++
1
2
3
4
- FeIII
β-alanine
++++
- FeIII
GABA
+
- FeIII [b]
aspartic acid
glutamic acid
h- β-alanine
++
++
-
- FeIII [b]
+++
5- FeIII [b]
DDW
++++
-
Bacterial growth promotion studies. The ability of iron-
III
III
loaded analogs 1-Fe - 5-Fe (Table 2) to serve as a sole iron
source for bacteria was quantified as growth level in iron-
depleted MKB medium. To eliminate background iron, the
medium was treated with an iron chelating amidoxime resin
-
Relative growth promotion activity levels: ++++ full activity comparable to FC;
++ high activity; ++ moderate activity; + weak activity; - no activity. FC -
+
[
a]
‘
Purolite S910’ (“Purolite”). Two bacterial strains were used in
ferrichrome, DDW - Double Distilled Water; amino acid incorporated into the
iron binding arm of each analog, [b]both enantiomers promote equal growth
levels in both strains. Actual growth curves are presented in the Supporting
information (Figure S2).
this study: E. coli UT5600[
34]
and P. putida JM218.[19a] Both
strains are defective in siderophore production in order to
exclude the possibility of iron uptake through the action of
naturally secreted siderophores. Thus, iron complexes of the
analogs were added as a sole iron source and growth levels
were measured over time until the stationary phase. To ensure
equal amount of added iron source, stock solutions of all analogs
were calibrated using the Cu·CAS method.[ Iron complex of
Iron complexes of achiral siderophores, such as
desferrioxamine (DFO), exist as a racemic mixture of Λ and Δ
enantiomers.[39] However, unlike carbon-generated asymmetric
centers, Λ and Δ configurations are in dynamic equilibrium and
35]
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201702647
FULL PAPER
can interchange. The receptor still recognizes mainly one
structure, but it also shifts the equilibrium towards the uptaken
form. Subsequently, given enough time the whole population of
the siderophore can be utilized. Current series of FC analogs
comprises two achiral compounds, namely β-alanine derived
tripod 1 and GABA derived tripod 2, as well as three chiral
analogs derived from aspartic acid (tripod 3), glutamic acid
E. coli is demanding not only in terms of sterical hindrance,
but also in the number methylenes separating the template
bound amides and the hydroxamic acids. We have previously
established that the glycine derived tripod with one methylene
spacer was not recognized by E. coli. On the other hand, tripod
1 with two-methylene spacer had excellent growth promotion
activity, while analog 2 with three methylenes was only weakly
active. Therefore, in the given series E. coli demonstrates
optimal uptake for two methylene spacer arms. Yet again, the
addition of a carboxylate, this time to the γ-amino acid skeleton
(glutamic derived analog 4), had a positive effect on the
recognition process in E. coli.
Two main structural parameters seem to control FC analog
recognition by E. coli: (i) presence of substituents on the iron
chelating arms, (ii) linker length connecting the template to the
hydroxamates. Generally, substituents interfere with recognition,
except for carboxylates, which might play a role in hydrogen
bonding to the FhuA transporter. On the other hand, the linker
length has an optimum at two methylenes (analog 1). Following
these guidelines leads to broad-spectrum activity (analog 1).
Disregarding them results in species-specificity towards P.
putida (analog 5).
(
tripod 4) and homo- β-alanine (tripod 5). Bacterial growth
studies (Table 2) demonstrated equal siderophore activity for
both enantiomers of each chiral analog. Circular Dichroism (CD)
spectroscopy provided a plausible explanation for these results.
While enantiomers of β-amino acid, tripods 3 and 5, display a
distinct chirality pattern around the iron, for γ- amino acid analog
4
at the same concentration only a trace signal was observed in
the Ligand to Metal Charge Transfer (LMCT) region of the iron
complex (data not shown).[14d]
Such decrease in chiral
preference around the iron with increasing linker length has been
previously observed in transition from to oriented
α
β
hydroxamates as well.[24] The chiral center of β and γ amino
acids is decoupled from the hydroxamates by one or two
methylene spacers. As a result, the existing chiral center cannot
induce a strong chiral preference around the bound iron, like in
native siderophores, and both enantiomers of each analog can
be recognized by the outer membrane receptor, probably
through the equilibrium exchange between the Λ and Δ
configurations. For practical applications of β- and γ- amino acid
derived tripods the chirality of the utilized amino acid does not
seem to play significant role in bacterial recognition and both
enantiomers are probably useful in the preparation of biologically
active analogs.
Using artificial mimics, it is possible to compare between
bacteria in terms of selectivity for iron sources. FC is recognized
to the same extent by both species in this study. However,
comparison between the biomimetic analogs reveals hidden
differences, and finer organization levels of their iron uptake
systems become visible. While P. putida can recognize
practically any tris-hydroxamate ferric complex in the current
series, E. coli is more selective and sensitive to sterical
hindrance, even at the β-position to the iron binding site. This is
demonstrated by comparison between β-alanine and homo-β-
alanine derived tripods 1 and 5. While the former is fully
recognized by E. coli, one methyl at the β-position of the latter
switches off the recognition. In light of this result, the moderate
siderophore activity of the aspartic derived tripod 3 in E. coli
comes as a surprise, because this tripod has a carboxylate at the
β-position. In terms of sterical hindrance this functional group
has a more dramatic effect, yet E. coli still recognizes it as a
suitable iron source to a significant extent. P. putida behaves
differently towards analogs, which contain carboxylates. Both
tripods 3 and 4 demonstrate lower growth promotion in this strain
compared to other compounds in the series. The effect of polar
side groups at various positions relative to hydroxamates should
be further examined by constructing additional analogs.
β-alanine derived tripod
1 showed superior growth
promotion in both tested strains. To follow its distribution after
iron uptake it was labeled with naphthalimide dye 42. This dye is
quenched in the presence of iron and lights up only in the free
chelator. Therefore, it is used as a turn-on sensor for iron
extraction from siderophores. Both E. coli and P. putida were
incubated with the iron complex of the labeled analog 1f for
several minutes and observed using fluorescent microscopy.
Figure 1A demonstrates a typical result of such experiment. After
iron uptake, the green fluorescence of the free chelator 1f
concentrates in the periphery of the bacterial cell, while the
cytosol remains darker. Thus, secretion of the free analog from
the cytosol after iron extraction is probably a fast process. The
experiment demonstrated in Figure 1A does not provide enough
information about the exact location of the secreted analog 1f. It
could theoretically reside in the periplasmic space or bind to
bacterial cells from outside. An elegant solution, which allows
distinguishing between the two possibilities, was found in
literature. It was previously reported that in gram-negative cells
under osmotic stress the cytosol shrinks and the periplasm
deforms and relocates to the opposite poles of the cell.[40] When
osmotic stress was induced by adding glycerol to live bacterial
cells, incubated with iron complex of 1f the fluorescence
relocation was consistent with this of the periplasmic space and
was clearly observed in the two poles of both E. coli and P. putida
strains (Figure 1B, E. coli demonstrated similar results). Thus,
after iron extraction the free chelator is rapidly excreted to the
periplasmic space and remains there for a significant amount of
time. These findings are in line with a report on antimicrobial
activity in gram-negative bacteria of
siderophore, conjugated to a β-lactam antibiotic. Alternatively,
a hydroxamate-type
[
41]
it might be reasonable to target both membranes of gram-
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201702647
FULL PAPER
negative bacteria from within the periplasmic space through
conjugation with surface active compounds, e.g. antimicrobial
peptides.
be detected. In ΔfhuB strain residual fluorescence could be
observed, consistent with the possibility of partial iron extraction
in the periplasm, described in literature.[37] These results
emphasize the necessity of all FC uptake system components
for efficient processing of 1f-FeIII by E. coli. Using a suitable
‘
cascade of mutants’ we were able to demonstrate that
III
compound 1f-Fe utilizes the same pathway as the iron-loaded
FC, making the β-alanyl tripod a ‘truly’ biomimetic analog.
FC is an endogenous siderophore to many fungal species,
including the extensively studied Ustilago maydis.[47] In fungi two
distinct mechanisms of iron acquisition from siderophores exist.
Some siderophores donate their iron at the membrane without
entering the cell, while others are actively transported into the
cytoplasm along with the ferric ion. FC is known to be
internalized via the active transport mechanism.[36]
U. maydis was incubated with 1f-FeIII (Figure 1C-E). Within
minutes, intense green fluorescence of the free chelator could
be observed in the cytosol (Figure 1D). Therefore, the iron
complex of β-alanine derived analog is efficiently recognized by
U. maydis and is transported to the cytoplasm, similarly to native
FC. Membrane dye FM4-64[48] (Figure 1D, orange-red) was
added as well to follow vesicle trafficking. Analog 1f is distributed
mostly outside the labeled vesicles and displays a punctate
pattern.
Overall, fluorescence microscopy results in P. putida, E.
coli and U. maydis emphasize the cross-phylum siderophore
activity of β-alanyl derived analog, similarly to native FC.
Previously, another tripodal FC analog displayed uptake and
utilization by U. maydis.[49] However, it could not be recognized
by E. coli, thus diverging from native FC.
Figure 1. Distribution of analog 1f in bacteria and fungi following
microbial iron uptake from the supplied 1f-Fe complex: Analog 1f is
III
quenched by iron and becomes fluorescent upon iron extraction by bacteria or
fungi. [A] Distribution pattern in live P. putida. The fluorescence of the iron-
free analog concentrates in the periphery of the bacterial cells. [B] Osmotic
shock was induced in P. putida, incubated with iron-loaded tripod 1f. The
fluorescence shifts to the poles of the bacterial cell, consistent with the
periplasmic space translocation under osmotic shock. [C-E] Light and
fluorescence images of the fungus U. maydis, labeled with FM4-64 membrane
dye and incubated with analog 1f. [C] DIC image. [D] two-channel fluorescence
image: green- analog 1f, orange-red- FM4-64. Iron free analog 1f is mostly
present outside the labeled vesicles and its internalization is consistent with
the uptake mechanism of native FC. [E] merge of DIC and fluorescence
images.
In gram-negative bacteria iron loaded siderophores are
[
4a, 42]
transported into the cell via a chain of specific receptors.
In
the case of FC uptake in E. coli,[
4c, 43]
iron loaded FC is first
recognized by the outer membrane receptor FhuA, which
transports it into the periplasmic space using the proton motive
force, supplied by TonB.[44] In the periplasm the siderophore-iron
complex binds to the FhuD shuttle protein. FhuD brings it to the
inner membrane receptor FhuB, which transfers the iron
complex across the cytoplasmic membrane, using energy
provided by the adjacent ATPase FhuC.[45] We were interested
to estimate whether the iron complex of β-alanyl derived analog
Figure 2. Bacterial cluster formed around emissive QDs coated with the
iron-loaded β-alanyl derived analog: [A] Bright field image of a bacterial
aggregate formed after application of the coated QDs. [B] Fluorescence image
of the bacterial cluster. Red emission is observed from within the bacterial
aggregate.
1f enters the bacterial cell via the same pathway as FC. Thus,
we repeated the fluorescence microscopy experiment with
several E. coli knockouts, each lacking a component of the FC
uptake system, namely deletion mutants: ΔfhuA,[ ΔtonB, ΔfhuD,
ΔfhuB and ΔfhuC (the Keio collection)[46]. After several minutes
of incubation with 1f-FeIII fluorescence was monitored in each
strain and compared to that of E. coli ΔentC,[46] which has intact
FC uptake system. In all but one mutant no fluorescence could
Due to the highly stabile association between siderophores
and their outer-membrane transporters, it is possible to capture
microorganisms by immobilized siderophores. Given the vast
assortment of available siderophores and their analogs they can
be used for differentiation between microorganisms and for
9]
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201702647
FULL PAPER
pathogen detection. Indeed, in recent years such prototype
systems were built using chips, functionalized silica and even
gold nanoparticles.[50]
on siderophore activity in the case of E. coli, it might be
interesting to examine additional analogs with polar, but not
charged or chelating groups.
Iron complex of β-alanine derived analog 1 was bound to
emissive QDs in order to examine the possibility of bacterial
trapping on their surface. To ensure proper interactions between
the analog and its outer membrane receptor, a long flexible linker
was introduced between the tripod and the surface of the QDs.
The iron complex decorated QDs were incubated with bacteria,
grown under iron starvation. Figure 2A shows a bright field image
of a bacterial aggregate formed after several minutes. Unlike the
naphthalimide fluorescent label 42, these particular QDs are not
quenched by iron. The fluorescence image of the same cluster
Attaching a fluorescent tag to the analog most similar to FC
(tripod 1) enabled following its fate post iron utilization in real-
time. While in P. putida and E. coli the periplasmic space was
identified as the cellular destination and the site of accumulation
for the iron-free analog, in U. maydis, cellular targets within the
cytosol were visualized and recorded. These results also
emphasize the biomimetic nature of analog 1f, acting cross
phylum, like native FC. Furthermore, tripod 1f utilizes the uptake
pathway of FC as demonstrated by FC transport mutants in E.
coli. Coupled with high iron affinity, these data mark the β-
alanine derived tripod a ‘true’ biomimetic FC analog.
(
Figure 2B) clearly shows the QDs fluorescence, which is
emitted from within the bacterial cluster. Such cluster formation
behavior could lead to development of bacterial cell
concentration methods, using various siderophore analogs.
Species-specific artificial analogs could play a crucial role in
development of methodologies for selective bacterial
concentration from diluted samples, especially when culturability
issues arise. Emissive QDs are readily available in a wide range
of colors. Coated each with a different analog or a native
siderophore, such array could provide a platform for microbial
separation based on diversity of their iron uptake systems.
Emissive QDs provided both fluorescence and a surface
for coating with multiple molecules of analog 1 iron complex. This
surface proved to be well suited for immobilization of FC-
recognizing bacteria. Such bacterial clusters, formed around
QDs, might provide a platform for future bacterial selection and
concentration, based on siderophore diversity. This could be
especially useful in cases of impaired culturability. Different
siderophores or their analogs, coupled to a palette of emissive
QDs might be applicable in various sorting techniques.
The power of biomimetic chemistry is clearly demonstrated
in mimicking FC by starting with simplified structures. Only a few
reiteration steps were sufficient to reach maturity and obtain
mimics that fulfill all the functions reserved to native FC. On the
way, FC mimics with unique properties were identified leading to
remarkable abilities awaiting utilization. The biomimetic
approach is not limited to a given initial structure and therefore
other starting templates and their reiteration, modifications and
proper tune-up should equally be expected to reach similar
results and mimics with new capabilities collected on the road.
The chemical versatility allows starting with modified
Conclusions
This work addresses itself to identify structural motifs in FC
mimics that control siderophore activity. This paper describes a
series of biomimetic analogs to FC designed to identify structural
motifs governing broad-spectrum activity on one hand and
promoting species-specificity on the other. A set of guidelines to
both effects have been formulated and discussed. The major
contribution to obtaining additional structurally diverse analogs
came from utilization of β- and λ- amino acid derived
hydroxamates.
structures ready to induce additional functionalization.
combination of elementary properties like ‘recognition’ and
signaling’ capabilities, described in the second part of the
A
‘
manuscript, resulted in obtaining an arsenal of theoretical and
practical tools for visualizing iron ‘trafficking’ and
compartmentalization, thus distinguishing between bacteria and
fungi. In general, this methodology allows mapping of
addressable targets. Iron transporters are solely reserved to the
microbial world while mammalian cells resort to entirely different
mechanisms for their iron supply. A long sought antifungal agent
that does not interact with mammalian cells, primarily of liver or
kidneys, due to their lack of siderophore transporters, seems a
likely outcome worth perusing.
Growth promotion activity of the current mimics presented
a dramatic variation, from an analog specific exclusively to P.
putida, but not to E. coli (tripod 5) to an analog with FC-like
activity causing maximal growth promotion in both E. coli and P.
putida (tripod 1). Other analogs fall in between, exhibiting some
activity in both strains (tripods 2-4). No difference in activity could
be observed in enantiomers of analogs 3-5, most probably due
to weak induction of chiral preference around the iron by the β-
and γ- positioned chiral center.
Although different analogs vary in iron binding efficiency
over a substantial pFe range, all of them fall within the active
siderophore values. Correspondingly, these data emphasize the
unfavorable influence of carboxylate substituents on iron affinity
Experimental Section
(
analogs 3 and 4), probably due to combination of electrostatic
repulsion and competition with hydroxamates over ferric ion
coordination. Due to beneficial effect of carboxylate substituents
Materials and methods. Commercially available amino acids, coupling
and protecting group reagents were purchased from "ChemImpex
international”. Other fine chemicals were purchased from "Sigma Aldrich".
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201702647
FULL PAPER
1
Amidoxime iron (III) chelating resin Purolite® S910 was obtained from
Purolite® SRL. Magnesium sulfate heptahydrate for the preparation of
MKB medium of Bio Ultra grade (min. 99.5%) was purchased from Sigma
Aldrich. Commercial reagents were used for synthesis without further
purification. Flash chromatography was performed using "Merck"
Kieselgel (40-63 µm) silica gel. NMR spectra were recorded on "Brucker
Avance" 250 MHz, 300 MHz, 400 MHz and 500 MHz machines equipped
with QNP probe. Chemical shifts are reported in ppm on the δ scale
downfield from TMS. ESI-MS spectra of synthetic intermediates were
recorded in a ZQ-4000 ("Micromass", Manchester, UK) instrument.
Bacterial growth was measured using a BioMate-3 spectrophotometer
obtained as a white solid (200 mg, 57%). H NMR (400 MHz, CDCl
3
) δ
1.33 (quint, J = 6.7 Hz, 2H), 1.49 (quint, J = 7.0 Hz, 2H), 1.58 (quint, J =
7.1Hz, 2H), 2.18 (t, J = 7.1 Hz, 2H), 2.37 (t, J = 5.6, Hz, 6H), 2.60 (bs,
6H), 3.18 (m, 11H), 3.47 (d, J = 5.5 Hz, 6H), 3.64 (t, J = 5.6Hz, 6H), 3.70
(s, 6H), 4.80 (s, 6H), 5.07 (s, 2H), 5.23 (bs, 1H), 6.49 (s, 1H), 6.91 (bs,
3H), 7.37 (m, 20H). 13C DEPTQ NMR (400 MHz, CDCl
3
) δ 25.10, 26.11,
29.49, 32.21, 33.35, 34.52, 36.61, 36.80, 40.82, 59.71, 66.43, 67.34,
69.20, 76.24, 127.96, 128.03, 128.43, 128.72, 129.07, 129.28, 134.08,
136.72, 156.46, 171.26, 173.61, 173.82. ESI-MS: calc.: 1154.59, found:
1154.99.
(
‘Thermo Electron Corporation’, Wisconsin US). For physico-chemical
studies all solutions were prepared in doubly distilled water. Water was
de-oxygenated by CO - and O -free argon. All stock solutions were
Coupling of tripod 38 to QDs: CdSe/CdS/CdZnS/ZnS QDs were
fabricated by Successive Ionic Layer Adsorption and Reaction (SILAR)
method.[51] The QDs were coated with 3-Mercaptopropionic acid (MPA)
2
2
prepared using a R200D Sartorius analytical balance (precision 0.01 mg).
Iron(III) perchlorate stock solution was prepared immediately before use
to provide surface carboxylates for coupling to tripod 38, each QD having
about 400 carboxylates on average. Carboxylate coated QDs were kindly
provided by Prof. U. Banin from the Hebrew University in Jerusalem. For
storage, QDs were divided to aliquots (1x10-9 mol each), two volumes of
from Fe(ClO
AppliChem p.a. 70%) and standardized by ICP-AES. HClO
titrated by standardized NaOH (0.1 ‘Fluka’ standard solution).
4
)
3 2
·xH O (‘Aldrich’, chlorides < 0.005%) in 0.01 M HClO
4
(
4
solution was
M
acetone were added to each aliquot, precipitated QDs were span down
by centrifugation (6500 rpm, 2 min) and the solvent was discarded. The
QD pellet was stored at -20 °C. Protected tripod 36 (80 mg, 0.07 mmol)
was dissolved in MeOH (10 mL), Pd cat. (10% on carbon, 10 mg) was
added and the reaction was stirred for 4.5 hrs under hydrogen
atmosphere. Pd was filtered off, and MeOH was evaporated to provide
the final deprotected compound 37 (48 mg, 99%), which was immediately
used. To HEPES buffer (10 mM, pH=7.4, 240 μL) tripod 37 (0.02 M in
MeOH, 30 μL) was added, followed by ferric chloride (0.02M in MeOH,
30 μL, 1 eq) to form the stock solution of iron complex 38. QDs (1x10-9
mol) were reconstituted in HEPES buffer (10 mM, pH=7.4, 250 μL),
containing EDC (10 mM) and incubated for a few minutes. Stock solution
of 38 (10 μL, 20 eq) was added to the activated QDs and the reaction
mixture was shaken for 1 hr. Excess reagents were separated from
functionalized QDs 40 by dialysis against DDW.
Carbonate-free NaOH solution (0.1 M ‘Fluka’ standard solution) was
standardized by titration with potassium hydrogen phthalate (‘Merck’.
1
p.a.). For detailed synthetic procedures and characterization ( H-NMR,
13C-NMR and ESI-MS) of all new compounds please refer to the
Supporting information. Selected examples are presented below.
Amine tripod 28: In a three-necked round bottom flask tripod 27 (0.64 g,
0.64 mmol) was dissolved in dry DCM (10 mL). The flask was sealed with
a rubber septum and phenylsilane (160 μL, 1.28 mmol, 2eq) was added
through. The reaction mixture was degassed by bubbling dry nitrogen gas
into the reaction mixture while vigorously stirring for 10 minutes. Next,
tetrakis(triphenylphosphine)palladium(0) (15 mg, 2% W/W) dispersed in
DCM (1 mL) was added to the flask and degassing continued for
additional 10 minutes. Nitrogen stream was closed and the reaction
mixture was stirred in the dark for 2 hrs. Once the reaction was complete
(
followed by TLC with ninhydrin staining) solids were filtered off and
Protected naphthalimide tripod 43: Tripod 41 (403 mg, 0.4 mmol) was
dissolved in DCM:DMF (1:1, 5 mL) and cooled on an ice bath. A few drops
of DIPEA were added to bring the pH to around 7. HBTU (350 mg, 0.92
mmol) was added and the reaction mixture was stirred for 30 min on ice.
Meanwhile, the naphtalimide dye 42[27a, 52] (250 mg, 0.7 mmol) was
dissolved in DCM (5 mL). It was added in one portion to the reaction
mixture, which was stirred on ice for additional 30 min. The ice bath was
removed and the reaction was stirred at RT for 72 hrs. The volatiles were
evaporated. The residue was dissolved in EtOAc (50 mL) and washed
first with 0.5N HCl (30 mL x 3) and then with water (30 mL x 2). The
aqueous washings were extracted with EtOAc. The combined organic
volatiles were evaporated. Final purification was performed using silica
gel chromatography with EtOAc: MeOH (9:1) as eluent to yield the
desired product as viscous colorless oil (0.55 g, 95% yield). H-NMR (300
1
MHz, CDCl
H), 3.38 (s, 6H), 3.48 (dd, J1=11.7 Hz, J2=5.7 Hz, 6H), 3.66 (t, J= 5.7
Hz, 6H), 4.79 (s, 6H), 7.02 (bs, 3H), 7.36 (m, 15H). 13C-NMR (300 MHz,
CDCl ): δ 32.13, 33.32, 34.49, 36.86, 55.98, 67.41, 72.55, 76.18, 128.85,
29.19, 129.42, 134.07, 171.29, 173.88. ESI-MS: calc.: 907.47, found:
07.59.
3
): δ 2.40 (t, J = 5.7 Hz, 6H), 2.60 (bt, J = 5.3 Hz, 6H), 3.17 (s,
9
3
1
9
Protected caproic acid linked tripod 36: Cbz protected 6-
aminohexanoic acid (105 mg, 0.4 mmol) was dissolved in DCM (3 mL),
and stirred at RT for a few minutes. N-hydroxysuccinimide (55 mg, 1.2
eq) was dissolved in THF (2 mL) and added to the reaction mixture. pH
of the reaction was brought to alkaline using DIPEA, followed by addition
of DIC (130 µL, 2.1 eq). The mixture was left to stir for 3 hrs at RT.
Compound 28 (152 mg, 0.3 mmol) was added and the reaction was
stirred at RT for 24 hrs. All volatiles were evaporated and the crude was
dissolved in EtOAc. The solution was kept in -18 °C for a few hours and
the white precipitate was filtered off. The organic layer was washed with
3
fraction was washed again with 5% NaHCO (30 mL x 3) and with water.
These washes were back extracted with EtOAc as well. Finally, the
combined organic layers were washed three times with brine and dried
over sodium sulfate. EtOAc was evaporated and the residue was purified
on a silica gel column. The column was washed with two portions of
3
EtOAc:Hex (4:1), while the product was eluted later using CHCl :MeOH
(95:1). The final protected tripod was obtained as a bright yellow highly
1
viscous oil (398 mg, 75%). H-NMR (400 MHz, CDCl
3
): δ 0.96 (t, J= 7.34,
3H), 1.40-1.45 (m, 6H), 1.70 (quint, J = 8.0 Hz, 2H), 2.39 (t, J = 5.7, 6H),
2.61-2.62 (m, 8H), 2.69 (t, J = 6.3, 2H), 3.18-3.23 (m, 13H), 3.48 (dd, J1=
11.3, J2=5.6 Hz, 6H), 3.66 (t, J = 5.7, 6H), 3.71 (s, 6H), 3.80 (bs, 2H),
3.88 (bs, 2H), 4.16 (t, J =8.0 Hz, 2H), 4.80 (s, 6H), 6.74 (s, 1H), 6.99 (bs,
3H), 7.19 (d, J = 8.0 Hz, 1H), 7.37 (m, 15H), 7.71 (t, J =8.0 Hz, 1H), 8.39
(d, J = 8.4 Hz, 1H), 8.50 (d, J =8.0 Hz, 1H), 8.58 (d, J =7.0 Hz, 1H). 13C-
1N HCl x3, followed by water, 5% NaHCO
3
(aq) x3 and brine. The
resulting organic layer was dried over Na SO
2
4
, and the solvent was
evaporated. The residue was adsorbed to silica and loaded onto a silica
gel column, equilibrated with Hex: Acetone (9:1); the final product was
eluted with a gradient of Hex: Acetone (9:1 to 8:2). The product was
3
NMR (400 MHz, CDCl ): δ 13.89, 20.43, 27.99, 28.30, 31.42, 32.29,
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201702647
FULL PAPER
3
6
1
1
3.42, 33.45, 34.64, 36.95, 40.16, 41.86, 45.48, 52.97, 59.96, 67.47,
9.29, 76.33, 115.38, 117.62, 123.45, 126.08, 126.28, 128.78, 129.34,
29.82, 131.21, 132.36, 134.20, 155.17, 163.94, 164.40, 170.69, 171.35,
72.69, 173.94. ESI-MS: calc.: 1326.65, found: 1327.00.
from spectrophotometric measurements, the fixed spectra of ferric ion in
solution in pH 0.6 (for the first series of experiments) or 2.0 (for the
second series of experiments) were included.[59]
Bacterial strains and growth media. P. putida JM218 strain and U.
-
pH-dependent UV-Visible and Potentiometric Titrations. Proposed
tripodal biomimetic analogs bind ferric ions starting from pH around 0 and
the complexation is essentially complete much below pH of 2. Therefore,
the stability constants of the first species formed in solution were
determined from the spectrophotometric pH-dependent batch titrations
carried out in pH range of around 0–2 (Plot S1a, Table S3), by following
the changes in the intense LMCT band of the metal complex. These
values were further used as constant values in spectrophotometric or
potentiometric titrations carried out in pH range from 2 to 8 (Plot S1b,
Table S3) to get the overall stability constants, logKFeL (Table 1).
maydis sid were provided by Y. Hadar. E. coli UT5600, ΔentC, ΔfhuA,
ΔtonB, ΔfhuD, ΔfhuB and ΔfhuC[46] were purchased from The E. coli
Genetic Stock Center at Yale. Bacteria were kept as glycerol stubs at -
78ºC. Before each experiment bacteria were grown in standard LB
medium overnight at 30ºC. MKB[60] medium was inoculated with the
culture from LB and incubated for 24 hrs. To prepare iron depleted MKB
for growth promotion experiments Purolite S910 resin (10g/L) was added
to the medium. After stirring at RT for 5 hrs the resin was filtered off and
the medium was autoclaved. Trace iron(III) concentration levels were
confirmed by ICP-MS. U. maydis were grown under iron deficient
conditions as described previously.[49]
In the first series of experiments the stock solution of iron complexes was
divided into various batches, with constant total volume of 2 mL, cFe =
Bacterial growth measurement. Each growth promotion experiment
was conducted in triplicates. Bacteria were diluted in 30 mL of sterile iron
depleted MKB medium to match the starting OD of 0.01 for P. putida
JM218 and of 0.02 for E.coli UT5600. Iron complexes of FC or the tested
analogs were added to the final concentration of 1.0·10 M. Bacteria were
incubated at 30°C and their growth was monitored by measuring OD at
600 nm every 1-2 hrs until stationary phase.
0
.0002 M and metal to ligand molar ratio 1:1. In pH range from 0.1 to 2,
the pH was controlled by the concentration of the perchloric acid. Ionic
strength was adjusted to 1 M with addition of NaClO . After preparation,
each solution was allowed to equilibrate for 50-60 min, and then its visible
4
-
6
spectrum was recorded using
spectrophotometer.
a
Varian CARY 300 UV/Vis
In the second set of experiments, 15 mL of solution containing a 1:1 molar
ratio of Fe :tripod with ferric concentration around 0.0002 M, were
introduced into a jacketed cell (‘Metrohm’) maintained at 25.0±0.2 °C. The
4
initial pH was adjusted to around 2 with HClO , and the titration of the
complex (2 < pH < 11) was then carried out by addition of known volumes
of 0.1 M sodium hydroxide by the Titrando 905 (‘Metrohm’) titrator.
Simultaneous pH and UV-visible measurements (250 nm - 800 nm) were
recorded using a Varian CARY 50 UV/Vis spectrophotometer fitted with
C-Technologies optical fibers and immersion probe made of quartz (‘C-
Technologies’) or metal (‘Hellma’ Ultra-Mini Immersion Probe). Both
simultaneous with spectroscopic measurements and separate
potentiometric titrations were performed using an automatic titrator
system Titrando 905 (‘Metrohm’) with a combined glass electrode
Analog 1f uptake monitoring: Bacterial strains were taken from a
glycerol stub (2–3 μL) and cultivated in sterile LB (3 mL with 0.1 mM final
concentration of kanamycin) for 10 hours at 30º C. Then, about 2–3 μL of
this bacterial stock were transferred into sterile MKB medium (3 mL with
0.1mM final concentration of kanamycin) and grown under the same
conditions for another 24 hours. Before introduction of our analog
bacterial samples (1 mL) were washed several times by addition of fresh
MKB and centrifugation. FC uptake deletion mutants: ΔfhuA, ΔtonB,
ΔfhuD, ΔfhuB and ΔfhuC produce enterobactin, which could scavenge
iron from analog 1f. They were washed thoroughly. After washings, the
pellet was resuspended in fresh MKB (1 mL). Stock solution of
fluorescent analog 1f was dissolved in DDW to 4·10-4 M, iron was added
III
in a 1:1 ratio from FeCl
3
stock solution (1.73·10-2 M; calibrated using ICP-
(
‘Mettler Toledo InLab Semi micro’) with polymer filling (XEROLYT®
EXTRA). The combined glass electrode was calibrated as a hydrogen
concentration probe by titrating known amounts of HClO (0.1 M) with
CO -free NaOH solutions (0.1 M).[53] The cell was thermostated at 25.0
0.2 °C. A stream of Argon, pre-saturated with water vapor, was passed
MS) in HCl (1 N). This freshly prepared iron complex (10 μL) was added
to the bacterial sample at final concentration of 4·10-6 M. For imaging, the
bacterial cells (10 μL) were immobilized on a coverslip with 1% agarose
(30 μL). The coverslip was imaged on a DeltaVision microscope (‘GE’)
using EGFP and TL Brightfield channels at x100. Epi-fluorescent
excitation was performed using a DAPI filter at 488 nm and emission was
recorded using a GFP filter at 509 nm. All images were acquired under
identical conditions. Images were subsequently analyzed using ZEN 2
software.
4
2
±
over the surface of the solution. Before every experiment the exact
concentration of ligand solution was determined by potentiometric
titration of 3 mL of sample the ligand, with clig ~ 0.001 M. In separate
potentiometric titrations metal to ligand ratio was 1:1 and the
concentration of metal was 0.003 M. The ionic strength was fixed at I =
4
0.1 M with NaClO (‘Fluka’, purum p.a.).
Funding Sources
Analysis and Processing of the Data. The spectrophotometric data
were analysed with Specfit program.[54] The potentiometric data (about
We are grateful to the Polish National Science Center (NCN,
UMO-2011/03/B/ST5/01057) for financial support.
1
40 points collected over the pH range 2-11) were refined with the
[
55]
[56]
Superquad
least-squares methods.[
or Hyperquad 2006
programs which use non-linear
57]
In the calculations of complex stability
constants the protonation constants of free ligands and the hydrolysis
We would like to express our gratitude to the Gerhardt M. J.
Schmidt Minnerva Center on Supermolecular Architecture, the
Helen and Martin Kimmel Center for Molecular Design and the
constants related to Fe(OH) (log11 = -2.56), Fe(OH)2+ (log12= -6.20),
2
+
4+
2
Fe (OH)
2
(log22= -2.84) and Fe(OH) (log13 = -11.41) species were
3
[
58]
taken into account. In the calculation of complexes stability constants
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201702647
FULL PAPER
Kimmelman Center for Macromolecular Assemblies for financial
support.
[
14] a) T. Lifa, W. Tieu, R. K. Hocking, R. Codd, Inorg. Chem. 2015, 54,
3573-3583; b) S. Heggemann, U. Mollmann, P. Gebhardt, L. Heinisch,
Biometals 2003, 16, 539-551; c) S. K. Sharma, M. J. Miller, S. M. Payne,
J. Med.Chem. 1989, 32, 357-367; d) S. Dhungana, J. M. Harrington, P.
Geblhardt, U. Mollmann, A. L. Crumbliss, Inorg. Chem. 2007, 46, 8362-
8371; e) B. H. Lee, M. J. Miller, C. A. Prody, J. B. Neilands, J. Med.
Chem. 1985, 28, 317-323; f) B. H. Lee, M. J. Miller, C. A. Prody, J. B.
Neilands, J. Med. Chem. 1985, 28, 323-327; g) K. Matsumoto, T.
Ozawa, K. Jitsukawa, H. Masuda, Inorg. Chem. 2004, 43, 8538-8546.
15] a) A. Shanzer, J. Libman, R. Lazar, Y. Tor, Pure Appl. Chem. 1989, 61,
Acknowledgements
We thank Prof. U. Banin and his team, from the Institute of
Chemistry, The Hebrew University of Jerusalem for kindly
supplying us with carboxylate functionalized QDs.
[
[
1529-1534; b) A. Shanzer, J. Libman, R. Lazar, Y. Tor, T. Emery,
Biochem. Bioph. Res. Co. 1988, 157, 389-394; c) E. Jurkevitch, Y.
Hadar, Y. Chen, J. Libman, A. Shanzer, J Bacteriol 1992, 174, 78-83.
16] a) T. Emery, L. Emery, R. K. Olsen, Biochem. Bioph. Res. Co. 1984,
Keywords: Biomimetic chemistry • Siderophores •
Ferrichrome• Fluorescent conjugates • Quantum Dots
119, 1191-1197; b) R. K. Olsen, K. Ramasamy, J. Org. Chem. 1985, 50,
2
264-2271.
[
[
1]
2]
a) J. B. Neilands, Annu. Rev. Nutr. 1981, 1, 27-46; b) C. Wandersman,
P. Delepelaire, Annual Review of Microbiology 2004, 58, 611-647; c) M.
L. Guerinot, Annual Review of Microbiology 1994, 48, 743-772.
a) M. Miethke, M. A. Marahiel, Microbiol. Mol. Biol. Rev. 2007, 71, 413-
[
[
[
17] A. Shanzer, J. Libman, P. Yakirevitch, Y. Hadar, Y. Chen, E. Jurkevitch,
Chirality 1993, 5, 359-365.
18] E. Jurkevitch, Y. Hadar, Y. Chen, J. Libman, A. Shanzer, J. Bacteriol.
1992, 174, 78-83.
451; b) D. E. Crowley, Y. C. Wang, C. P. P. Reid, P. J. Szaniszlo, Plant
19] a) H. Weizman, O. Ardon, B. Mester, J. Libman, O. Dwir, Y. Hadar, Y.
Chen, A. Shanzer, J. Am. Chem. Soc. 1996, 118, 12368-12375; b) R.
Nudelman, O. Ardon, Y. Hadar, Y. N. Chen, J. Libman, A. Shanzer, J.
Med. Chem. 1998, 41, 1671-1678.
Soil 1991, 130, 179-198; c) B. C. Chu, A. Garcia-Herrero, T. H.
Johanson, K. D. Krewulak, C. K. Lau, R. S. Peacock, Z. Slavinskaya, H.
J. Vogel, Biometals 2010, 23, 601-611; d) F. Pattus, M. A. Abdallah, J.
Chin. Chem. Soc. 2000, 47, 1-20.
[
[
20] L. Zelikovich, J. Libman, A. Shanzer, Nature 1995, 374, 790-792.
21] I. Rubinstein, S. Steinberg, Y. Tor, A. Shanzer, J. Sagiv, Nature 1988,
[
[
3]
4]
M. Hannauer, Y. Barda, G. L. A. Mislin, A. Shanzer, I. J. Schalk, J
Bacteriol 2010, 192, 1212-1220.
332, 426-429.
a) D. J. Raines, T. J. Sanderson, A. K. Duhme-Klair, E. Wilde, in
Elsevier Reference Module in Chemistry, Molecular Sciences and
Chemical Engineering, Elsevier, 2015; b) R. Chakraborty, in Iron
Uptake in Bacteria with Emphasis on E. coli and Pseudomonas (Eds.:
R. Chakraborty, V. Braun, K. Hantke, P. Cornelis), Springer
Netherlands, Dordrecht, 2013, pp. 1-29; c) V. Braun, Front Biosci 2003,
[
22] a) A. D. Ferguson, E. Hofmann, J. W. Coulton, K. Diederichs, W. Welte,
Science 1998, 282, 2215-2220; b) K. P. Locher, B. Rees, R. Koebnik,
A. Mitschler, L. Moulinier, J. P. Rosenbusch, D. Moras, Cell 1998, 95,
7
71-778.
23] K. D. Krewulak, H. J. Vogel, Biochim. Biophys. Acta 2008, 1778, 1781-
804.
[
[
[
[
[
1
8
, 1409-1421; d) D. van der Helm, R. Chakraborty, in Microbial
Transport Systems, Wiley-VCH Verlag GmbH & Co. KGaA, 2003, pp.
61-287.
24] E. Olshvang, A. Szebesczyk, H. Kozlowski, Y. Hadar, E. Gumienna-
Kontecka, A. Shanzer, Dalton Trans. 2015, 44, 20850-20858.
2
25] W. R. Harris, C. J. Carrano, K. N. Raymond, J. Am. Chem. Soc. 1979,
[
[
[
[
5]
6]
7]
8]
I. J. Schalk, L. Guillon, Amino Acids 2013, 44, 1267-1277.
H. Boukhalfa, A. L. Crumbliss, Biometals 2002, 15, 325-339.
R. Codd, Coord. Chem. Rev. 2008, 252, 1387-1408.
101, 2722-2727.
26] a) B. H. Lee, M. J. Miller, J. Org. Chem. 1983, 48, 24-31; b) R. Grigg, Z.
Rankovic, M. Thoroughgood, Tetrahedron 2000, 56, 8025-8032.
a) J. B. Neilands, J. Am. Chem. Soc. 1952, 74, 4846-4847; b) J. B.
Neilands, Science 1967, 156, 1443-1447; c) S. J. Rogers, R. A. Warren,
J. B. Neilands, Nature 1963, 200, 167.
27] a) M. M. Meijler, R. Arad-Yellin, Z. I. Cabantchik, A. Shanzer, J. Am.
Chem. Soc. 2002, 124, 12666-12667; b) R. Nudelman, O. Ardon, Y.
Hadar, Y. Chen, J. Libman, A. Shanzer, J. Med. Chem. 1998, 41, 1671-
[
[
9]
M. Gaspar, M. A. Santos, K. Krauter, G. Winkelmann, Biometals 1999,
1678.
12, 209-218.
[
[
28] J. B. Neilands, J. Biol. Chem. 1995, 270, 26723-26726.
29] L. D. Pettit, H. K. Powell, Timble, York, UK, 2001, (www.acadsoft.co.uk,
version 5.6).
10] a) A. Shanzer, J. Libman, J. Chem. Soc., Chem. Commun. 1983, 846-
8
47; b) E. J. Corey, S. Bhattacharyya, Tetrahedron Letters 1977, 3919-
922; c) J. M. Roosenberg, Y. M. Lin, Y. Lu, M. J. Miller, Curr. Med.
3
[
[
[
30] G. Anderegg, F. L'Eplattenier, G. Schwarzenbach, Helv. Chim. Acta
Chem. 2000, 7, 159-197; d) R. Mashiach, M. M. Meijler, Org. Lett. 2013,
5, 1702-1705; e) M. Meyer, J. R. Telford, S. M. Cohen, D. J. White, J.
1963, 46, 1409-1422.
1
31] W. R. Harris, C. J. Carrano, S. R. Cooper, S. R. Sofen, A. E. Avdeef, J.
V. Mcardle, K. N. Raymond, J. Am. Chem. Soc. 1979, 101, 6097-6104.
32] a) A. M.,Albrecht-Gary, A.L. Crumbliss, Vol. 35 (Ed.: A. Sigel, H. Sigel),
Marcel Dekker, New York, 1998, p. 239; b) A. L. Crumbliss, Handbook
of Microbial Iron Chelates, (Ed.: G. Winkelmann). CRC Press: Boca
Raton, Florida, 1991, pp. 177.
Xu, K. N. Raymond, J. Am. Chem. Soc. 1997, 119, 10093-10103; f) J.
Hu, M. J. Miller, J. Am. Chem. Soc. 1997, 119, 3462-3468; g) J. Deng,
Y. Hamada, T. Shioiri, J. Am. Chem. Soc. 1995, 117, 7824-7825; h) P.
J. Maurer, M. J. Miller, J. Am. Chem. Soc. 1982, 104, 3096-3101; i) J.
L. H. Madsen, T. C. Johnstone, E. M. Nolan, J. Am. Chem. Soc. 2015,
137, 9117-9127; j) D. Seebach, H.-M. Müller, H. M. Bürger, D. A.
[
33] a) E. Farkas, P. Buglyo, J. Chem. Soc., Dalton Trans. 1990, 1549-1551;
b) E. Farkas, D. A. Brown, R. Cittaro, W. K. Glass, J. Chem. Soc.,
Dalton Trans. 1993, 2803-2807.
Plattner, Ache, Angew. Chem. Int. Ed. Engl 1992, 31, 434-435; k) R. J.
Bergeron, J. J. Pegram, J. Org. Chem. 1988, 53, 3131-3134.
[
[
11] A. Ino, A. Murabayashi, Tetrahedron 2001, 57, 1897-1902.
12] T. Zheng, J. L. Bullock, E. M. Nolan, J. Am. Chem. Soc. 2012, 134,
[
[
34] M. E. Elish, J. R. Pierce, C. F. Earhart, J. Gen. Microbiol. 1988, 134,
1355-1364.
18388-18400.
35] B. Schwyn, J. B. Neilands, Anal. Biochem. 1987, 160, 47-56.
[
13] R. Breslow, Chem. Soc. Rev 1972, 1, 553-580.
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201702647
FULL PAPER
[
[
36] R. C. Hider, X. Kong, Nat. Prod. Rep. 2010, 27, 637-657.
Israel, 1989 (Eds.: Y. Chen, Y. Hadar), Springer Netherlands, Dordrecht,
1991, pp. 289-297.
37] I. J. Schalk, G. L. A. Mislin, K. Brillet, Curr. Top. Membr .2012, 69, 37-
66.
[
[
38] a) G. Winkelmann, V. Braun, Fems Microbiol. Lett. 1981, 11, 237-241;
b) H. U. Naegeli, W. Kellerschierlein, Helv. Chim. Acta 1978, 61, 2088-
2
095.
39] S. Dhungana, P. S. White, A. L. Crumbliss, J. Biol. Inorg. Chem. 2001,
, 810-818.
6
[
[
40] H. Schwarz, A. L. Koch, Microbiology 1995, 141, 3161-3170.
41] E. K. Dolence, A. A. Minnick, M. J. Miller, J. Med. Chem. 1990, 33, 461-
464.
[
[
[
[
[
42] V. Braun, K. Hantke, in Microbial transport systems (Ed.: G.
Winkelmann), Wiley-VCH, Weinheim, 2002, pp. 289-312.
43] R. Chakraborty, E. Storey, D. van der Helm, Biometals 2007, 20, 263-
274.
44] a) K. Postle, R. A. Larsen, Biometals 2007, 20, 453-465; b) M. C.
Wiener, Curr Opin Struct Biol 2005, 15, 394-400.
45] K. N. Raymond, B. E. Allred, A. K. Sia, Acc. Chem. Res. 2015, 48, 2496-
2505.
46] T. Baba, T. Ara, M. Hasegawa, Y. Takai, Y. Okumura, M. Baba, K. A.
Datsenko, M. Tomita, B. L. Wanner, H. Mori, Mol. Syst. Biol. 2006, 1-
11.
[
47] B. Winterberg, S. Uhlmann, U. Linne, F. Lessing, M. A. Marahiel, H.
Eichhorn, R. Kahmann, J. Schirawski, Mol. Microbiol. 2010, 75, 1260-
1
271.
48] M. W. Lewis, I. V. Robalino, N. O. Keyhani, Microbiology 2009, 155,
110-3120.
[
[
[
3
49] O. Ardon, R. Nudelman, C. Caris, J. Libman, A. Shanzer, Y. Chen, Y.
Hadar, J. Bacteriol. 1998, 180, 2021-2026.
50] a) D. D. Doorneweerd, W. A. Henne, R. G. Reifenberger, P. S. Low,
Langmuir 2010, 26, 15424-15429; b) T. Inomata, T. Murase, H. Ido, T.
Ozawa, H. Masuda, Chem. Lett. 2014, 43, 1146-1148; c) T. Inomata, H.
Eguchi, Y. Funahashi, T. Ozawa, H. Masuda, Langmuir 2012, 28, 1611-
1617; d) T. Inomata, H. Eguchi, K. Matsumoto, Y. Funahashi, T. Ozawa,
H. Masuda, Biosens. Bioelectron. 2007, 23, 751-755; e) Y. Kim, D. P.
Lyvers, A. Wei, R. G. Reifenberger, P. S. Low, Lab Chip 2012, 12, 971-
9
76; f) T. Zheng, E. M. Nolan, Metallomics 2012, 4, 866-880.
51] a) S. Ruhle, M. Shalom, A. Zaban, Chemphyschem 2010, 11, 2290-
304; b) H. Tsukigase, Y. Suzuki, M. H. Berger, T. Sagawa, S.
[
[
2
Yoshikawa, J. Nanosci. Nanotechnol. 2011, 11, 1914-1922.
52] C. Y. Li, X. B. Zhang, L. Qiao, Y. Zhao, C. M. He, S. Y. Huan, L. M. Lu,
L. X. Jian, G. L. Shen, R. Q. Yu, Anal. Chem. 2009, 81, 9993-10001.
53] P. Gans, B. O'Sullivan, Talanta 2000, 51, 33-37.
[
[
54] a) H. Gampp, M. Maeder, C. J. Meyer, A. D. Zuberbuehler, Talanta
1986, 33, 943-951; b) H. Gampp, M. Maeder, C. J. Meyer, A. D.
Zuberbuhler, Talanta 1985, 32, 95-101; c) H. Gampp, M. Maeder, C. J.
Meyer, A. D. Zuberbuhler, Talanta 1985, 32, 257-264; d) F. J. Rossotti,
H. S. Rossotti, R. J. Whewell, J. Inorg. Nucl. Chem. 1971, 33, 2051-
2065.
[
[
[
55] P. Gans, A. Sabatini, A. Vacca, J. Chem. Soc. Dalton 1985, 1195-1200.
56] P. Gans, A. Sabatini, A. Vacca, Talanta 1996, 43, 1739-1753.
57] P. Gans, Data Fitting in the Chemical Sciences: By the Method of Least
Squares, Wiley, Chichester, 1992.
[
[
[
58] C. F. Baes, R. E. Mesmer, The hydrolysis of cations, Wiley, New York,
1
976, pp. 489
59] a) R. M. Milburn, W. C. Vosburgh, J. Am. Chem. Soc. 1955, 77, 1352-
355; b) R. M. Milburn, J. Am. Chem. Soc. 1957, 79, 537-540.
1
60] M. Höfte, K. Y. Seong, E. Jurkevitch, W. Verstraete, in Iron Nutrition
and Interactions in Plants: “Proceedings of the Fifth International
Symposium on Iron Nutrition and Interactions in Plants”, Jerusalem,
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201702647
FULL PAPER
FULL PAPER
A ‘true’ analog matching Ferrichrome
Jenny Besserglick, Evgenia Olshvang,
(
FC) in cross-phylum activity, uptake
Agnieszka
Szebesczyk,
Joseph
pathway and iron affinity is described.
This fluorescently labeled analog
sheds light on the fate of
siderophores following iron delivery
and provides a tool for identifying
microbial targets. Quantum Dots
Englander, Dana Levinson, Yitzhak
Hadar, Elzbieta Gumienna-Kontecka*,
and Abraham Shanzer*
Page No. – Page No.
Ferrichrome has found its match:
biomimetic analogs with diversified
activity map discrete microbial
targets
(
QD) decorated with this analog
trigger cluster formation by FC-
recognizing bacteria, suggesting a
method for microbial concentration
and separation.
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201702647
FULL PAPER
This article is protected by copyright. All rights reserved.