Aminopyrrolic synthetic receptors for monosaccharides: A class of carbohydrate-binding agents endowed with antibiotic activity versus pathogenic yeasts

Article abstract of DOI:10.1002/chem.201103318

The biological activity of a set of structurally related aminopyrrolic synthetic receptors for monosaccharides has been tested versus yeast and yeast-like microorganisms and compared to their binding affinity toward mannosides. Antibiotic activity compara

Full text of DOI:10.1002/chem.201103318

DOI: 10.1002/chem.201103318
Aminopyrrolic Synthetic Receptors for Monosaccharides: A Class of
Carbohydrate-Binding Agents Endowed with Antibiotic Activity versus
Pathogenic Yeasts
Cristina Nativi,[b, c] Oscar Francesconi,[b] Gabriele Gabrielli,[b] Irene De Simone,[b]
Benedetta Turchetti,[d] Tommaso Mello,[e] Lorenzo Di Cesare Mannelli,[f]
Carla Ghelardini,[f] Pietro Buzzini,[d] and Stefano Roelens*[a]
Abstract: The biological activity of
a set of structurally related aminopyr-
rolic synthetic receptors for monosac-
charides has been tested versus yeast
and yeast-like microorganisms and
compared to their binding affinity
toward mannosides. Antibiotic activity
comparable to that of well-known poly-
ene (amphotericin B) or azole (ketoco-
nazole) drugs has been found for some
members of the family, along with
a general correlation with binding abili-
ties. A systematic structure–activity–af-
finity investigation shed light on the
structural and functional requirements
necessary for antibiotic activity and
identified the tripodal compound 1 as
the most potent compound of the set.
Together with toxicity tests and inhibi-
tor localization experiments performed
through fluorescence microscopy, these
studies led to the characterization of
a new class of carbohydrate binding
agents possessing antibiotic activity, in
which pyrrolic groups precisely struc-
tured on a tripodal architecture appear
to be responsible for permeability
through the cell wall of pathogens, as
well as for antibiotic activity inside the
cytoplasm.
Keywords: antibiotics
·
carbohy-
drates · microbiology · molecular
recognition · synthetic receptors
Introduction
Surface glycans are generally linked to proteins or lipids and
form different structures that are cell-type specific, giving
rise to a carbohydrate coating the diversity of which is es-
sential for discriminating the “self” from the “nonself” in
the interaction with other cells. Typically, the antigenic sur-
face of viruses and pathogenic microorganisms is covered by
a coating of glycolipids and glycoproteins, which trigger the
immune response based on self/nonself carbohydrate dis-
crimination.[2] On the other hand, in the case of viruses such
as HIV, the immune response is escaped because the carbo-
hydrates on the pathogen are themselves synthesized by the
host and therefore not recognized as nonself.[3] All these
phenomena rely on molecular recognition of carbohydrates
as a key step of the process, which explains the attention
that has been focused on this subject in the last decade.[4]
In this context, targeting the glycans of viral glycoproteins
with carbohydrate-binding agents (CBAs) is recently emerg-
ing as a promising strategy for antiviral therapies and vacci-
nes.[3a,5] Because mannose is frequently encountered as the
terminal sugar of glycans and is densely expressed in the
glycan shield of the viral envelope of several viral infections
of high health risk (HIV, HCV, etc.), terminal oligomanno-
sides have become attractive therapeutic targets.[6] For ex-
ample, the pradimicin A and the benanomicin A antibiotics,
which have been found to bind to the terminal mannosides
exposed on fungal membrane,[7] have antiviral activity
against HIV-1, with EC50 values in the micromolar range.[8]
Carbohydrates are abundant on the surface of living eukary-
otic cells, including microorganisms, in which they are in-
volved in a number of functions, from adhesion to infection,
from signal transmission to immune-system modulation.[1]
[a] Dr. S. Roelens
Istituto di Metodologie Chimiche (IMC)
Consiglio Nazionale delle Ricerche (CNR)
Dipartimento di Chimica, Polo Scientifico e Tecnologico
50019 Sesto Fiorentino, Firenze (Italy)
Fax : (+39)055-457-3570
E-mail: stefano.roelens@unifi.it
[b] Prof. C. Nativi, Dr. O. Francesconi, G. Gabrielli, I. D. Simone
Dipartimento di Chimica, Universitꢀ di Firenze
Polo Scientifico e Tecnologico, 50019 Sesto Fiorentino, Firenze (Italy)
[c] Prof. C. Nativi
FiorGen, Polo Scientifico e Tecnologico
50019 Sesto Fiorentino, Firenze (Italy)
[d] Dr. B. Turchetti, Prof. P. Buzzini
Dipartimento di Biologia Applicata and Industrial Yeasts
Collection DBVPG, Universitꢀ di Perugia, 06121 Perugia (Italy)
[e] Dr. T. Mello+
Dipartimento di Fisiopatologia Clinica
Universitꢀ di Firenze, 50139 Firenze (Italy)
[f] Dr. L. D. C. Mannelli, Prof. C. Ghelardini
Dipartimento di Farmacologia Preclinica e Clinica,
Universitꢀ di Firenze, 50139 Firenze (Italy)
[+] Contributed to this work with fluorescence microscopy photographs.
Microvirin,
a cyanobacterial monovalent lectin, exerts
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201103318.
a potent inhibitory activity toward HIV-1 infection (IC50 =
5064
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 5064 – 5072

FULL PAPER
2–12 nm) by selectively recognizing the Mana
(1-2)Man
toward mannosides, were measured and compared to the
growth inhibition of some yeast and yeast-like pathogenic
target strains. Association constants were measured by
1H NMR spectroscopic titrations according to a previously
described protocol (see the Supporting Information for de-
tails), which consists of the simultaneous fit of the complex-
ation-induced shifts of all the available signals from both the
receptor and the glycoside to the appropriate association
model by nonlinear regression analysis. Minimum inhibitory
concentrations (MICs) were determined in RPMI 1640 cul-
ture medium (Sigma–Aldrich) by using 96-well microtiter
plates. After incubation at 358C for 48 h under aerobic con-
ditions, growth in the presence of two-fold increasing con-
centrations of each aminopyrrolic receptor was compared
visually with the growth observed in the absence of antimi-
crobial agent. The compounds selected for testing, including
the progenitor of the family 1 and the most effective recep-
tors 2 and 3, are depicted below (the R and S descriptors
motif of the high-mannose-type oligosaccharides of the
gp120 viral protein.[9] It is thus clear that having available ef-
fective and selective receptors for mannosides would be of
paramount importance for the development of CBAs.
In the course of our studies on molecular recognition of
carbohydrates, we developed a family of aminopyrrolic tri-
podal structures that proved to be effective receptors for
monosaccharides in polar organic solvents, showing various
levels of selectivity toward the biologically relevant mono-
saccharides.[10] Because of their binding ability, these mole-
cules may interact with cell-surface glycans and consequent-
ly interfere with the living processes of the cell. Intrigued by
this possibility, we were prompted to carry out an investiga-
tion on the biological activity of this class of synthetic recep-
tors; in particular, we were interested in finding out whether
any inhibitory activity could be associated to their binding
properties.
Searching for the appropriate target to test biological ac-
tivity, we resorted to the use of cells of yeast and yeast-like
microorganisms. Yeasts are among the most widely studied
eukaryotic microorganisms and are extremely important in
cell biology because they are similar to human cells in many
respects. In particular, expression and processing of glyco-
proteins is largely analogous to that occurring in mammalian
cells. Many of the secretory processes are indeed conserved
between yeasts and mammals, and for this reason yeasts are
considered a good model system to study glycoside process-
ing.[11] In addition, some yeasts which are opportunistic
pathogens and cause infections in humans, present a close
carbohydrate similarity between the mannan portion ex-
posed on the cell surface and the oligomannose glycans of
the HIV gp120 glycoprotein of the viral envelope.[3a] As
a matter of fact, the species Candida albicans, Candida tro-
picalis, and Candida parapsilosis were for many years the
most frequently isolated pathogens from clinical speci-
mens.[12] Furthermore, the emergence of AIDS and the ad-
vances in biomedical procedures (e.g., organ transplants, ag-
gressive chemotherapies, intravascular devices, etc.) in-
creased the risk of infections in immunocompromised
hosts.[13] Likewise, some yeast-like organisms are also associ-
ated with human and animal diseases, such as a few species
of the genus Prototheca (P. zopfii and P. whickeramii).[13c,14]
We report, herein, the results of an investigation on the
binding properties, antibiotic activity, and cytotoxicity of
a family of structurally related aminopyrrolic carbohydrate
receptors, together with a detailed analysis of the structure–
binding–activity relationship.
are used to designate the receptors of all-R and all-S abso-
lute configuration of the six stereocenters, respectively). The
corresponding affinities toward octyl a- and b-mannosides
(OctaMan, OctbMan), measured in a polar solvent (acetoni-
trile) and expressed as BC050 values,[15] and the MIC values
toward four different microorganisms (C. tropicalis, P. nor-
vegensis, P. wickerhamii, and P. zopfii) are reported in
Tables 1 and 2, respectively, and in Figure S1 in the Support-
ing Information. The MIC values for two well-known antibi-
otic drugs, namely, Amphotericin B (AmB) and Ketocona-
zole (Keto), measured toward the investigated cell lines, are
also reported in Table 2 for direct comparison.
Results and Discussion
Preliminary screening: In a preliminary screening, the affini-
ties of the most representative members of the aminopyrrol-
ic family of receptors, that is, the structures showing the
most interesting binding properties toward the glycosides of
the biologically relevant monosaccharides, in particular
Two main features were clearly emerging from this
screening: 1) with the exception of (R)-3, the activity of
which is uninteresting, all of the other compounds tested
Chem. Eur. J. 2012, 18, 5064 – 5072
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5065

S. Roelens et al.
Table 1. Binding affinities (BC500 [mm]) of synthetic receptors 1–3 for
high affinities for the mannosides (see Figure S1 in the Sup-
porting Information).
octyl mannosides in CD3CN.[a]
OctaMan
OctbMan
1
15.60
0.13
0.44
0.30
0.29
7.40
0.87
0.57
1.22
0.08
Structure–activity studies: On the basis of the results from
the preliminary screening and of the interesting antibiotic
activity observed toward pathogenic microorganisms, a de-
tailed structure–activity investigation, developed from the
structure of 1 as a lead compound, was carried out with the
main goal of ascertaining the structural and functional re-
quirements essential for antibiotic activity. Structure modifi-
cations on the parent tripodal aminopyrrolic architecture in-
volved: 1) the variation of number and substitution pattern
on the aromatic platform of the aminopyrrolic binding arms,
2) the introduction of additional functional groups on the
binding arms, 3) the variation of the size of the aromatic
platform, and 4) the dissection of the architecture into the
constituent structural elements. The new family of com-
pounds prepared through these modifications is depicted
below, the MIC values determined toward 14 different
strains of yeast and yeast-like microorganisms are reported
in Table 3, the corresponding affinities measured toward
octyl mannosides are reported in Table 4, and a graphical
comparison of the activity–affinity values is depicted in
Figure 1.
(R)-2
(S)-2
(R)-3
(S)-3
1
[a] Formation constants (log b) were measured by H NMR spectroscopy
(400 MHz) from titration experiments at T=298 K. Intrinsic median
binding concentration (BC050) values were calculated from the (log b)
values by using the “BC50 Calculator” Program (see reference [15]).
markedly inhibited the growth of all yeast and yeast-like
target species; the MIC values obtained, in particular those
for compound 1, are comparable to those measured for
AmB and lower than those obtained for Keto; and 2) al-
though in general binding to mannosides is associated to in-
hibition activity, which is noteworthy considering the widely
different media in which these properties were measured,
affinity and activity do not follow the same trend. Indeed,
compound 1 shows the highest activities associated with the
lowest affinities, whereas the poorly inhibiting (R)-3 shows
5066
www.chemeurj.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 5064 – 5072

Aminopyrrolic Synthetic Receptors for Monosaccharides
FULL PAPER
Table 2. Antibiotic activity (MIC [mgmLÀ1]) of compounds 1–3, AmB,
The above results raised the question as to whether the
whole structure was required for the growth inhibition activ-
ity expressed by 1. Dissection of the parent architecture into
the constituent elements gave a clear answer to the ques-
tion: compounds 13–15, prepared to clarify the individual
contributions of single functional groups, showed no inhibi-
tory activity toward the target strains, clearly demonstrating
that structuring all the constituent components into the tri-
podal architecture is essential for inhibiting the growth of
yeast and yeast-like microorganisms. In this context, it is
noteworthy that 12, the congener of 1 in which the three
ethyl groups have been removed, displays a dramatic loss of
activity with respect to 1, even though the removed substitu-
ents were not expected to have an effect on the activity.
Concerning the binding ability toward mannosides in rela-
tion to the inhibition properties, it is evident that for the set
of compounds investigated a direct correlation between
MIC and BC050 values does not hold, as anticipated from the
preliminary screening, nor may be expected, when consider-
ing that affinities were measured in acetonitrile, whereas an-
tibiotic activities are obtained from cultures in aqueous
media. Nevertheless, a broad correspondence exists, in that
compounds showing no affinity for mannosides also showed
no inhibitory activity, whereas, with few exceptions, com-
pounds exhibiting various levels of affinity are also endowed
with antibiotic activity. This feature can be better appreciat-
ed from the graphical representation of data depicted in
Figure 1, in which activities and affinities are compared
face-to-face. The evidence that the two properties do not
follow the same trend suggests that the origin of the antibi-
otic activity does not reside in the ability to bind to manno-
and Keto versus strains of yeast and yeast-like microorganisms.[a]
C. tropicalis
P. norvegensis[b]
P. wickerhamii
P. zopfii
3982[c]
6163[c]
8879[c]
8880[c]
1
4
8
8
64
16
2
2
4
4
64
16
1
2
4
4
64
16
2
8
4
4
128
16
1
(R)-2
(S)-2
(R)-3
(S)-3
AmB
Keto
32
8
16
32
[a] MIC values were determined after 48 h growth in RPMI 1640
medium. The experiments were performed as duplicates. [b] Formerly
known as Candida zeylanoides (see reference [13]). [c] DBVPG collec-
tion accession number.
The main conclusion that can be drawn from the analysis
of these data is that none of the modifications carried out
improved the antibiotic activity of the parent structure. This
quite unexpected outcome clearly shows that compound 1 is
a structure already optimized toward the set of strains
tested. Although sequentially removing the binding arms, as
in 4 and 5, predictably increases MIC values, introducing ad-
ditional binding arms on the platform, as in 6 and 7, does
not increase activities. Likewise, additional functional
groups on the binding arms, such as in 8–10, introduced on
the parent structure 1 to exert electronic effects and affect-
ing H-bonding ability, had a marked effect on binding (e.g.,
in 9) but did not produce increased antibiotic activity
toward any strain. Replacing the benzene platform with
a napthalene moiety in 11[16] led to higher MIC values, even
though the activity drop was less pronounced than that ob-
served for the homologous
Table 4. Binding affinities (BC050 [mm]) of compounds 1 and 4–15 for octyl mannosides in CD3CN.[a]
structure 6. Altogether, the sys-
tematic structural variations
studied induced loss of activity
in all cases, with the sole excep-
tion of 5 toward strain 6141, the
activity of which showed a sig-
nificant increase.
1
4
5
6
7
8
9
10
11
12
13
14
15
OctaMan
OctbMan
15.6
7.40
n.d.
n.d.
98.0
44.5
12.0
5.90
8.20
8.00
20.0
10.4
4.00
0.65
17.0
4.80
n.d.
n.d.
22.0
20.6
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
[a] Formation constants (log b) were measured by 1H NMR spectroscopy (400 MHz) from titration experi-
ments at T=298 K. BC050 values were calculated from the (log b) values by using the “BC50 Calculator” Pro-
gram (see reference [15]). n.d.=non-detectable.
Table 3. Antibiotic activity (MIC [mgmLÀ1]) of compounds 1 and 4–15 against different strains of yeast and yeast-like microorganisms.[a]
DBVPG[b] Species
1
4
5
6
7
8
9
10
11
256
128
128
64
12
13
14
15
6133
6157
3828
6140
6150
3982
6163
6142
6053
6782
6141
8879
8880
8830
Candida albicans
Candida albicans
Candida glabrata
64 128 512
16 128 512
>1024 512 256
256
64
128
64
>1024
512
>1024
512
>1024
>1024
>1024
512
>1024
256
>1024
>1024
>1024
>1024
>1024
512
>1024
>1024
>1024
>1024
>1024
>1024
>1024
>1024
>1024
>1024
>1024
>1024
>1024
>1024
>1024
>1024
>1024
>1024
>1024
>1024
>1024
>1024
>1024
>1024
>1024
>1024
>1024
>1024
1024 256 256
8
8
8
4
2
256 128
128 64
256 512
128
64
1024 256
1024 64
>1024 512
64
64
64
32
8
Meyerozyma guiliiermondii
Candida parapsilosis
Candida tropicalis
Pichia norvegensis
Clavispora lusitaniae
Yarrowia lipolytica
Pichia kudriavzevii
Kluyveromyces marxianus
Prototheca wickeramii
Prototheca zopfii
>1024
8
>1024 1024
64
16
512
64
32
8
128
32
64
8
512
128
128
32
64
128
128
8
128
32
128
8
256
256
256
64
512
32 128 256
128 512
16 128
>1024 256 128
512
256
128
128
256
1024
1024
>1024
1024
512
>1024
1024
>1024
1024
>1024
>1024
>1024
8
512
512
256
256
256
256
62 128
64
32
32 128
64 128
32
64
4
64
64
8
8
2
8
8
16
16
32 128
64 128
256
256
256
512
Prototheca zopfii
64
[a] MIC values were determined after 48 h growth in RPMI 1640 medium. The experiments were performed as duplicates. [b] DBVPG collection acces-
sion number.
Chem. Eur. J. 2012, 18, 5064 – 5072
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5067

S. Roelens et al.
Figure 1. Graph of the antibiotic activity (MIC [mgmLÀ1]) of compounds 1 and 4–15 versus 14 different strains of yeast and yeast-like microorganisms
(top) and the corresponding affinities (BC050 [mm]) for octyl mannosides (bottom).
sides; on the other hand, binding seems to be a prerequisite
to exert inhibition activity, which suggests that binding is in-
volved in the process at some stage. This may be compatible
with the hypothesis that the ability to bind to mannosides
would facilitate the approach of the compounds to the yeast
and yeast-like cells, triggering a process that would subse-
quently involve adhesion to the cell wall and penetration
across the membrane into the cytoplasm, in which they
might exert antibiotic activity by interacting with some spe-
cific target. This latter hypothesis is based on the very strict
structural requirements associated to the inhibition activity
expressed against yeast and yeast-like cells, which appear in-
tolerant of even small variations, including the presence of
apparently innocent components (such as the ethyl groups
on the aromatic platform), thus suggesting a very precise fit
of the molecule into a binding pocket.
Uptake and toxicity studies: To gain more information and
support for this hypothesis, we investigated the fate of the
aminopyrrolic inhibitors when incubated with the microor-
ganisms. Fluorescence appeared to be the convenient prop-
erty to follow for the localization of inhibitors in the cells.
However, monitoring 1 by fluorescence was unfeasible,
whereas 11 showed excitation and emission spectra that
could be exploited for fluorescence microscopy. On the
other hand, the lower activity of 11 with respect to 1 could
be compensated for by using a higher concentration of 11 in
cell cultures for fluorescence experiments, under the as-
Table 5. Viability [%] of human hepatic cell lines[a] after treating with increasing concentrations [mm] of inhibitor, and the corresponding lethal concen-
tration (LC50 [mm]).[b] [Please note: headings have been centered on columns]
Concentration [mm]
LC50 [mm]
0
0.03
0.1
0.3
1
3
10
30
100
300
1
11
16
AmB
Keto
100.0Æ3.4
100.0Æ2.7
100.0Æ2.6
100.0Æ4.6
100.0Æ2.8
100.1Æ2.9
98.1Æ1.4
105.2Æ4.9
106.7Æ3.7
100.7Æ1.8
104.3Æ4.4
102.4Æ2.9
109.9Æ6.8
97.7Æ2.2
103.6Æ3.7
107.4Æ1.1
105.9Æ3.5
106.7Æ3.6
98.5Æ3.4
103.7Æ2.4
104.7Æ2.9
105.8Æ2.0
103.1Æ4.3
99.0Æ4.4
97.2Æ0.7
25.1Æ3.5
18.3Æ2.8
94.0Æ3.9
93.5Æ3.6
34.7Æ4.3
44.9Æ5.4
17.4Æ3.9
42.6Æ1.4
90.0Æ6.2
16.5Æ4.4
24.3Æ6.4
17.3Æ3.3
25.2Æ0.4
65.3Æ3.9
14.8Æ2.1
15.6Æ5.1
8.3Æ0.1
2.6Æ0.8
27.2Æ2.1
189.4Æ3.2
8.2Æ1.0
101.0Æ4.0
96.9Æ4.8
91.0Æ1.8
72.5Æ5.6
21.3Æ3.8
41.8Æ6.5
8.2Æ0.2
8.0Æ3.2
9.0Æ1.3
[a] HuH-7 human hepatoma cell line. [b] Data are expressed as % of the basal value (0 mm). Viability was evaluated by MTT test. All treatments were
performed after 48 h incubation with the analyzed compounds in DMEM/HamF12 culture media with 10% serum and 1% DMSO.
5068
www.chemeurj.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 5064 – 5072

Aminopyrrolic Synthetic Receptors for Monosaccharides
FULL PAPER
sumption that the two inhibi-
tors would follow the same
pathway. To ascertain the exis-
tence of a correlation between
the inhibition activity and the
localization of the inhibitor in
the cell, the tetraminonaphtha-
lene derivative 16, the homo-
A direct conclusion from these experiments is that the
presence of pyrrolic groups seems to be responsible for the
permeability of the active inhibitor 11 through the yeast cell
wall; in the absence of pyrrolic groups, the aminic com-
pound 16 does not penetrate into the cytoplasm, in addition
to be biologically inactive and incapable of binding to man-
nosides. It is also evident that the inhibition activity is exert-
ed inside the cytoplasm, rather than at the level of the cell
surface. Whether pyrrolic moieties are also determinant for
the activity of the inhibitor or just for its permeability
cannot be assessed unambiguously from the present data;
nevertheless, this evidence appears to support the existence
of a chain of events in which binding to the oligomannosides
of the cell wall may play a crucial role in the internalization
process.
Toxicity is a critical property of compounds possessing an-
tibiotic activity. In particular the well known antifungal
drugs AmB and Keto are associated with the development
of hepatotoxicity and nephrotoxicity.[18] To characterize the
in vitro toxicity of the aminopyrrolic inhibitors, the viability
of a human hepatic cell line (HuH-7 human hepatoma cell
line) has been tested. HuH-7 cells were incubated in the
presence of increasing concentrations of compounds 1, 11,
and 16, and the cell viability was monitored after 24 and
48 h incubation. Results are reported in Table S2 in the Sup-
porting Information and Table 5, respectively, together with
the corresponding LC50 values. Cell viabilities obtained by
incubation with AmB and Keto and corresponding LC50
values are also reported for comparison. A graphical repre-
sentation of toxicity data is reported in the graph in
Figure 3.
logue of 11 devoid of pyrrolic groups, was prepared as an in-
active, fluorescent counterpart of 11. The latter was used as
the reference compound in blank experiments, in analogy to
compound 13, used as a reference for 1, which completely
lost its inhibition properties when the pyrrolic moieties had
been removed.[17]
Pichia norvegensis DBVPG 6163, which was the strain
most effectively inhibited by 11, was thus incubated a) in the
absence of any additive, b) in the presence of 11, and c) in
the presence of 16. The cultures were then observed by fluo-
rescence microscopy and the corresponding images are re-
ported in Figure 2. The images in Figure 2b and c clearly
show that both compounds have been absorbed by the yeast
cells, which exhibited an unequivocally stronger fluores-
cence than the background, concentrated in distinct areas of
the cell.
Figure 2. Fluorescence emitted by cells of Pichia norvegensis DBVPG
6163 after 48 h growth in RPMI 1640 medium a) in the absence of any
additive; b) in the presence of 11 (128 mgmLÀ1); c) in the presence of 16
(128 mgmLÀ1); d) as in (b), after washing with 0.9% NaCl sterile solu-
tion; and e) as in (c), after washing with 0.9% NaCl sterile solution.
Apparently, 11 and 16 were taken up from the culture in
much the same way; however, when cells were washed with
0.9% NaCl sterile solution, fluorescence persisted unaltered
in cells incubated with 11 (Figure 2d), whereas it turned off
to the background level in cells treated with 16 (Figure 2e),
which demonstrated that the latter was actually just “depos-
ited” on the cell surface, whereas the former had been inter-
nalized in the cytoplasm. This conclusion was confirmed by
fluorescence scanning focused on different planes, which
showed emission from the whole body of cells treated with
11, but only from the cell surface for those treated with 16.
Figure 3. Viability [%] of HuH-7 hepatocytes after 48 h incubation with
!
&
*
^
increasing concentration [mm] of 1 ( ), 11 ( ), 16 ( ), AmB ( ), and
~
Keto ( ).
Although up to 1 mm of all the tested compounds are non-
toxic, a steep decrease in cell viability is observed for
1 above 1 mm, for AmB above 3 mm, and for 11 above 10 mm;
Chem. Eur. J. 2012, 18, 5064 – 5072
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5069

S. Roelens et al.
for Keto, a quite shallow decrease is observed from 1 mm,
whereas for 16 a shallow decrease above 30 mm provides evi-
dence of an overall low toxicity. In terms of LC50, AmB and
Keto show similar values, whereas 1 exhibits a slightly small-
er value. It is worth noting that the decreasing toxicities of
compounds 1, 11, and 16, with LC50 values spanning nearly
two orders of magnitude, follow the same trend observed
for their overall inhibition activity toward the yeast and
yeast-like target strains, which suggests a possible common
mode of action.
Localization experiments on compounds 11 and 16 in cells
of human hepatocytes provided additional information clari-
fying some ambiguities found with pathogens. Fluorescence
microscopy images obtained on HuH-7 cell lines from ex-
periments analogous to those performed on Pichia norve-
gensis are shown in Figure 4.
count the different binding abilities of 11 and 16 toward
mannosides, this evidence reinforces the hypothesis that per-
meability of the inhibitor mediated by carbohydrate recogni-
tion on the cell wall of yeasts is a crucial step of a cascade
of events responsible for the antibiotic activity of these com-
pounds. It should be noted, however, that the carbohydrate
recognition-mediated entry into the yeast cell, characterizing
this family of aminopyrrolic structures, may not be limited
to their interaction with mannosides. Because monosacchari-
dic glycosides other than mannosides are also recognized, it
cannot be excluded that interaction with different glycans of
the cell wall may indeed be responsible for the permeability
features of the pyrrolic inhibitors.
A further aspect sheding light on the above ambiguities is
the activity–toxicity differences between 11 and 16. Both are
found inside the cytoplasm after incubating the hepatocytes
in their presence; yet they show a markedly different toxici-
ty (see Table 5 and Figure 3). Considering that the antibiotic
activity of this family follows the toxicity trend, it may be
expected that the lack of activity observed for 16 with
pathogenic yeast and yeast-like microorganisms would be
due to poor inhibiting activity towards microbial cell
growth, rather than to poor permeability through the cell
wall, due to the lack of pyrrolic groups. In turn this would
confirm that pyrrolic groups on the aminic scaffold are not
only responsible for the permeability of 11 (and of related
aminopyrrolic compounds of the family), but also for the
actual inhibition properties exerted in the cell cytoplasm.
Conclusion
Figure 4. Fluorescence emitted by cells of HuH-7 after 15 min incubation
a) in the absence of any additive; b) in the presence of 11 (128 mgmLÀ1);
c) in the presence of 16 (128 mgmLÀ1); d) as in (b), after washing with
Dulbeccoꢂs modified eagle medium (DMEM); and e) as in (c), after
washing with DMEM.
In this work, we have shown that a family of structurally re-
lated aminopyrrolic synthetic receptors for monosaccharides
represents a new class of carbohydrate binding agents pos-
sessing antibiotic activity against yeast and yeast-like patho-
gens, some of which exhibit potencies comparable to that of
well-known polyene (amphotericin B) or azole (ketocona-
zole) drugs. Summarizing the main conclusions emerging
from this work:
It can be seen that, as for yeast cells, fluorescence sub-
stantially stronger than that of the background is emitted
from the hepatocytes, when incubated with 11 and 16 (Fig-
ure 4b and c, respectively, in comparison to Figure 4a).
After 15 min incubation, both inhibitors were internalized
into the cytoplasm and appeared concentrated in small fluo-
rescent spots, most likely because of their inclusion into
small vesicles by a cell cytoplasmic process. This observation
was confirmed by subsequent washing of cells with DMEM
solution, which caused a somewhat attenuated fluorescence,
but did not remove emission from either one of the two
treated strains, showing that both compounds were largely
internalized into the cytoplasm (Figure 4d and e). It appears
that, in contrast to the yeast cell wall, the hepatocyte mem-
brane is permeable to both aminic compounds.[19] The differ-
ent permeability behavior of hepatocites compared to yeast
cells may be ascribed to the glycoconjugate coating of the
yeast cell wall. This may discriminate 11 from 16 by selec-
tively allowing carbohydrate recognition-mediated entry
into the yeast cell of the aminopyrrolic inhibitor but not of
its structurally related aminic counterpart. Taking into ac-
1) The tripodal aminopyrrolic compound 1 is the optimal
structure in terms of antibiotic activity within the family.
None of the modifications carried out improved the anti-
biotic activity of the parent structure. On the other hand,
structuring all the constituent elements into the tripodal
architecture turned out to be essential for inhibiting the
growth of yeast and yeast-like microorganisms.
2) Although a direct correlation between MICs and affini-
ties for mannosides was not found, a broad correspond-
ence exists, in that compounds showing no affinity for
mannosides also lack any inhibitory activity, whereas
compounds showing various levels of affinity are also en-
dowed with antibiotic activity. This suggests that the
origin of the antibiotic activity does not reside in the
ability to bind mannosides, which are densely expressed
on the cell wall of these pathogens, but that binding is in-
5070
www.chemeurj.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 5064 – 5072

Aminopyrrolic Synthetic Receptors for Monosaccharides
FULL PAPER
3629–3638; d) M. Mazik, Chem. Soc. Rev. 2009, 38, 935–956; e) S.
Kubik, Angew. Chem. 2009, 121, 1750–1753; Angew. Chem. Int. Ed.
2009, 48, 1722–1725; f) A. P. Davis, T. D. James in Functional Syn-
thetic Receptors (Eds.: T. Schrader, A. D. Hamilton), Wiley-VCH,
Weinheim, 2005, pp. 45–109.
volved in the process at some stage, possibly facilitating
the approach of the inhibitor and its permeation through
the cell wall.
3) By following, through fluorescence microscopy, the fate
of the inhibitors when incubated with the microorgan-
isms it was found that aminopyrrolic compounds were in-
ternalized into the cytoplasm, whereas the corresponding
structures devoid of pyrrolic groups did not pass through
the cell wall. While supporting the previous conclusion,
this evidence indicates that pyrrolic groups are required
for inhibitor permeability.
4) The toxicity of this family of inhibitors parallels their an-
tibiotic activity, in that the most active compounds are
also the most toxic toward human hepatocytes. However,
the toxicity of the most effective inhibitors is comparable
to that of amphotericin B and ketoconazole.
5) Localization experiments through fluorescence microsco-
py on human hepatocytes showed that both the (inac-
tive) amino- and the (active) aminopyrrolic compounds
were internalized into the cytoplasm, which demonstrat-
ed that pyrrolic groups are responsible not only for the
permeability, but also for the antibiotic activity and tox-
icity of the inhibitors. Furthermore, this evidence reinfor-
ces the hypothesis that inhibitor permeation mediated by
carbohydrate recognition on the cell wall of yeasts is
a crucial step of a cascade of events responsible for the
antibiotic activity of these compounds.
[5] J. Balzarini, Nat. Rev. Microbiol. 2007, 5, 583–597.
[6] S.-K. Wang, P.-H. Liang, R. D. Astronomo, T.-L. Hsu, S.-L. Hsieh,
D. R. Burton, C.-H. Wong, Proc. Natl. Acad. Sci. USA 2008, 105,
3690–3695.
[7] a) Y. Nakagawa, Y. Masuda, K. Yamada, T. Doi, K. Takegoshi, Y.
Igarashi, Y. Ito, Angew. Chem. 2011, 123, 6208–6212; Angew. Chem.
Int. Ed. 2011, 50, 6084–6088; b) T. J. Walsh, N. Giri, Eur. J. Clin. Mi-
crobiol. Infect. Dis. 1997, 16, 93–97; c) M. Watanabe, H. Tohyama,
T. Hiratami, S. Inoue, S. Kondo, T. Takeuchi, H. Yamaguchi, J. Anti-
biot. 1997, 50, 1042–1051; d) T. Ueki, M. Oka, Y. Fukagawa, T. Oki,
J. Antibiot. 1993, 46, 465–477.
[8] a) J. Balzarini, K. Van Laethem, D. Daelemans, S. Hatse, A. Bugatti,
M. Rusnati, Y. Igarashi, T. Oki, D. Schols, J. Virol. 2007, 81, 362–
373; b) A. Tanabe-Tochikura, T. S. Tochikura, O. Yoshida, T. Oki, N.
Yamamoto, Virology 1990, 176, 467–473; c) H. Hoshino, J. Seki, T.
Takeuchi, J. Antibiot. 1989, 42, 344–346.
[9] S. Shahzad-ul-Hussan, E. Gustchina, R. Ghirlando, G. M. Clore,
C. A. Bewley, J. Biol. Chem. 2011, 286, 20788–20796.
[10] a) C. Nativi, O. Francesconi, G. Gabrielli, A. Vacca, S. Roelens,
Chem. Eur. J. 2011, 17, 4814–4820; b) A. Ardꢄ, F. J. CaÇada, C.
Nativi, O. Francesconi, G. Gabrielli, A. Ienco, J. Jimꢅnez-Barbero, S.
Roelens, Chem. Eur. J. 2011, 17, 4821–4829; c) A. Ardꢄ, C. Venturi,
C. Nativi, O. Francesconi, G. Gabrielli, F. J. CaÇada, J. Jimꢅnez-Bar-
bero, S. Roelens, Chem. Eur. J. 2010, 16, 414–418; d) C. Nativi, M.
Cacciarini, O. Francesconi, G. Moneti, S. Roelens, Org. Lett. 2007, 9,
4685–4688; e) C. Nativi, M. Cacciarini, O. Francesconi, A. Vacca, G.
Moneti, A. Ienco, S. Roelens, J. Am. Chem. Soc. 2007, 129, 4377–
4385.
[11] T. R. Gemmill, R. B. Trimble, Biochim. BioPhys. Acta. 1999, 1426,
227–237.
[12] C. R. Cooper Jr., in The Yeasts: A Taxonomic Study, Vol. 1, 5th ed.
(Eds.: C. P. Kurtzman, J. W. Fell, T. Boekhout), Elsevier Science,
Amsterdam, 2011, pp. 9–19.
As a closing remark, quoting from the recent ACS Webi-
nar on the future of antimicrobial drug discovery: “Antimi-
crobial resistance is growing at an alarming rate and The In-
fectious Disease Society of America has issued the challenge
to develop ten new antibiotics by 2020. How difficult will
this be? What are the hurdles?”. In this context, it appears
that the aminopyrrolic inhibitors of yeast and yeast-like pro-
liferation presented in this paper may constitute a new tool
against pathogenic microorganisms and may have an impact
on the development of new drugs circumventing antibiotic
resistance.
[13] Among these yeasts (labeled as emerging pathogens), the literature
reports cases of cutaneous or systemic mycoses mainly caused by
strains of Candida glabrata, Meyerozyma guilliermondii (teleomor-
phic state of Candida guilliermondii), Pichia kudriavzevii (teleomor-
phic state of Candida krusei), Pichia norvegensis (former Candida
zeylanoides), and Clavispora lusitaniae, see: a) M. A. Lachance in
The Yeasts: A Taxonomic Study, Vol. 2, 5th ed. (Eds.: C. P. Kurtz-
man, J. W. Fell, T. Boekhout), Elsevier Science, Amsterdam, 2011,
pp. 349–353; b) C. P. Kurtzman in The Yeasts: A Taxonomic Study,
Vol. 2, 5th ed. (Eds.: C. P. Kurtzman, J. W. Fell, T. Boekhout), Elsev-
ier Science, Amsterdam, 2011, pp. 685–707; c) R. S. Pore in The
Yeasts: A Taxonomic Study, Vol. 3, 5th ed. (Eds.: C. P. Kurtzman,
J. W. Fell, T. Boekhout), Elsevier Science, Amsterdam, 2011,
pp. 2071–2080.
Acknowledgements
A research grant to Dr. Oscar Francesconi (grant no. 2008.1475) from
Ente Cassa di Risparmio di Firenze is gratefully acknowledged.
[14] U. Roesler, A. Mçller, A. Hensel, D. Baumann, U. Truyen, Int. J.
Syst. Evol. Microbiol. 2006, 56, 1419–1425.
[15] The BC50 (median binding concentration) parameter, which can be
calculated from the measured association constants, is a generalized
affinity descriptor univocally defining on a common scale the overall
binding ability of receptors for ligands in chemical systems involving
any number of complex species. The BC50 is defined as the total con-
centration of receptor necessary for binding 50% of the ligand and
directly provides a quantitative measure of binding ability as, in
analogy to the IC50 parameter, the lower the value, the higher the
affinity. The BC050 (intrinsic median binding concentration) is the
value assumed by the BC50 when the fraction of bound receptor is
zero, that is, when forming the first complex molecule; this is a re-
ceptor specific constant value, independent of concentration, which
coincides with the dissociation constant Kd when the 1:1 association
is the only equilibrium present in solution. A detailed description of
the treatment of the BC50 descriptor has been reported (see referen-
[1] a) The Sugar Code: Fundamentals of Glycosciences (Ed.: H.-J.
Gabius), Wiley-VCH, Weinheim, 2009; b) B. Ernst, W. Hart, P.
Sinaꢃ, Carbohydrates in Chemistry and Biology, Part I.2 and
Part II.4, Wiley-VCH, Weinheim, 2000; c) T. K. Lindhorst, Essentials
of Carbohydrate Chemistry and Biochemistry, 1st ed., Wiley-VCH,
Weinheim, 2000.
[2] R. Medzhitov, C. A. Janeway, Jr., Science 2002, 296, 298–300.
[3] a) C. N. Scanlan, J. Offer, N. Zitzmann, R. A. Dwek, Nature 2007,
446, 1038–1045; b) N. L. Letvin, Nat. Rev. Immunol. 2006, 6, 930–
939.
[4] a) S. Jin, Y. Cheng, S. Reid, M. Li, B. Wang, Med. Res. Rev. 2010, 30,
171–257; b) D. B. Walker, G. Joshi, A. P. Davis, Cell. Mol. Life Sci.
2009, 66, 3177–3191; c) A. P. Davis, Org. Biomol. Chem. 2009, 7,
Chem. Eur. J. 2012, 18, 5064 – 5072
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5071

S. Roelens et al.
ce [10a], and references therein). To expedite the calculation of the
BC50 parameter, a computer program has been developed for PC
platforms (“BC50 Calculator”) and is available for free upon request
from the corresponding author. A detailed description of the equa-
tions used by the program for computing BC50 and BC050 values is re-
ported in the above references.
[18] a) G. P. Patel, C. W. Crank, J. B. Leikin, J. Med. Toxicol. 2011, 7, 12–
15; b) L. A. Garcꢆa Rodrꢆguez, A. Duque, J. Castellsague, S. Perez-
Gutthann, B. H. Stricker, Br. J. Clin. Pharmacol. 1999, 48, 847–852;
c) R. N. Chien, L. J. Yang, P. Y. Lin, Y. F. Liaw, Hepatology 1997, 25,
103–107; d) F. J. Abajo, A. J. Carcas, Br. Med. J. 1986, 293, 1243.
[19] Permeability of human hepatic cell membrane to protonated aminic
compounds has been reported, see: O. Hayashida, C. Eguchi, K.
Kimura, Y. Oyama, T. Nakashima, K. Shioji, Chem. Lett. 2010, 39,
1321–1322.
[16] Replacement of the benzene platform with
a tetrahydropyrene
moiety was also tested, but the resulting MIC values were unrelia-
ble, likely because of the interference of the inhibitor with the cul-
ture medium, which turned dark blue on addition of the inhibitor.
[17] Compound 16 gave MIC values >1024 mgmLÀ1 for all the strains
tested. No evidence of binding to mannosides could be detected.
Received: October 20, 2011
Revised: December 28, 2011
Published online: March 5, 2012
5072
www.chemeurj.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 5064 – 5072